diff options
Diffstat (limited to 'library/alloc/src')
76 files changed, 45569 insertions, 0 deletions
diff --git a/library/alloc/src/alloc.rs b/library/alloc/src/alloc.rs new file mode 100644 index 000000000..cc8da7bcc --- /dev/null +++ b/library/alloc/src/alloc.rs @@ -0,0 +1,444 @@ +//! Memory allocation APIs + +#![stable(feature = "alloc_module", since = "1.28.0")] + +#[cfg(not(test))] +use core::intrinsics; +use core::intrinsics::{min_align_of_val, size_of_val}; + +use core::ptr::Unique; +#[cfg(not(test))] +use core::ptr::{self, NonNull}; + +#[stable(feature = "alloc_module", since = "1.28.0")] +#[doc(inline)] +pub use core::alloc::*; + +use core::marker::Destruct; + +#[cfg(test)] +mod tests; + +extern "Rust" { + // These are the magic symbols to call the global allocator. rustc generates + // them to call `__rg_alloc` etc. if there is a `#[global_allocator]` attribute + // (the code expanding that attribute macro generates those functions), or to call + // the default implementations in libstd (`__rdl_alloc` etc. in `library/std/src/alloc.rs`) + // otherwise. + // The rustc fork of LLVM 14 and earlier also special-cases these function names to be able to optimize them + // like `malloc`, `realloc`, and `free`, respectively. + #[rustc_allocator] + #[rustc_allocator_nounwind] + fn __rust_alloc(size: usize, align: usize) -> *mut u8; + #[cfg_attr(not(bootstrap), rustc_deallocator)] + #[rustc_allocator_nounwind] + fn __rust_dealloc(ptr: *mut u8, size: usize, align: usize); + #[cfg_attr(not(bootstrap), rustc_reallocator)] + #[rustc_allocator_nounwind] + fn __rust_realloc(ptr: *mut u8, old_size: usize, align: usize, new_size: usize) -> *mut u8; + #[cfg_attr(not(bootstrap), rustc_allocator_zeroed)] + #[rustc_allocator_nounwind] + fn __rust_alloc_zeroed(size: usize, align: usize) -> *mut u8; +} + +/// The global memory allocator. +/// +/// This type implements the [`Allocator`] trait by forwarding calls +/// to the allocator registered with the `#[global_allocator]` attribute +/// if there is one, or the `std` crate’s default. +/// +/// Note: while this type is unstable, the functionality it provides can be +/// accessed through the [free functions in `alloc`](self#functions). +#[unstable(feature = "allocator_api", issue = "32838")] +#[derive(Copy, Clone, Default, Debug)] +#[cfg(not(test))] +pub struct Global; + +#[cfg(test)] +pub use std::alloc::Global; + +/// Allocate memory with the global allocator. +/// +/// This function forwards calls to the [`GlobalAlloc::alloc`] method +/// of the allocator registered with the `#[global_allocator]` attribute +/// if there is one, or the `std` crate’s default. +/// +/// This function is expected to be deprecated in favor of the `alloc` method +/// of the [`Global`] type when it and the [`Allocator`] trait become stable. +/// +/// # Safety +/// +/// See [`GlobalAlloc::alloc`]. +/// +/// # Examples +/// +/// ``` +/// use std::alloc::{alloc, dealloc, handle_alloc_error, Layout}; +/// +/// unsafe { +/// let layout = Layout::new::<u16>(); +/// let ptr = alloc(layout); +/// if ptr.is_null() { +/// handle_alloc_error(layout); +/// } +/// +/// *(ptr as *mut u16) = 42; +/// assert_eq!(*(ptr as *mut u16), 42); +/// +/// dealloc(ptr, layout); +/// } +/// ``` +#[stable(feature = "global_alloc", since = "1.28.0")] +#[must_use = "losing the pointer will leak memory"] +#[inline] +pub unsafe fn alloc(layout: Layout) -> *mut u8 { + unsafe { __rust_alloc(layout.size(), layout.align()) } +} + +/// Deallocate memory with the global allocator. +/// +/// This function forwards calls to the [`GlobalAlloc::dealloc`] method +/// of the allocator registered with the `#[global_allocator]` attribute +/// if there is one, or the `std` crate’s default. +/// +/// This function is expected to be deprecated in favor of the `dealloc` method +/// of the [`Global`] type when it and the [`Allocator`] trait become stable. +/// +/// # Safety +/// +/// See [`GlobalAlloc::dealloc`]. +#[stable(feature = "global_alloc", since = "1.28.0")] +#[inline] +pub unsafe fn dealloc(ptr: *mut u8, layout: Layout) { + unsafe { __rust_dealloc(ptr, layout.size(), layout.align()) } +} + +/// Reallocate memory with the global allocator. +/// +/// This function forwards calls to the [`GlobalAlloc::realloc`] method +/// of the allocator registered with the `#[global_allocator]` attribute +/// if there is one, or the `std` crate’s default. +/// +/// This function is expected to be deprecated in favor of the `realloc` method +/// of the [`Global`] type when it and the [`Allocator`] trait become stable. +/// +/// # Safety +/// +/// See [`GlobalAlloc::realloc`]. +#[stable(feature = "global_alloc", since = "1.28.0")] +#[must_use = "losing the pointer will leak memory"] +#[inline] +pub unsafe fn realloc(ptr: *mut u8, layout: Layout, new_size: usize) -> *mut u8 { + unsafe { __rust_realloc(ptr, layout.size(), layout.align(), new_size) } +} + +/// Allocate zero-initialized memory with the global allocator. +/// +/// This function forwards calls to the [`GlobalAlloc::alloc_zeroed`] method +/// of the allocator registered with the `#[global_allocator]` attribute +/// if there is one, or the `std` crate’s default. +/// +/// This function is expected to be deprecated in favor of the `alloc_zeroed` method +/// of the [`Global`] type when it and the [`Allocator`] trait become stable. +/// +/// # Safety +/// +/// See [`GlobalAlloc::alloc_zeroed`]. +/// +/// # Examples +/// +/// ``` +/// use std::alloc::{alloc_zeroed, dealloc, Layout}; +/// +/// unsafe { +/// let layout = Layout::new::<u16>(); +/// let ptr = alloc_zeroed(layout); +/// +/// assert_eq!(*(ptr as *mut u16), 0); +/// +/// dealloc(ptr, layout); +/// } +/// ``` +#[stable(feature = "global_alloc", since = "1.28.0")] +#[must_use = "losing the pointer will leak memory"] +#[inline] +pub unsafe fn alloc_zeroed(layout: Layout) -> *mut u8 { + unsafe { __rust_alloc_zeroed(layout.size(), layout.align()) } +} + +#[cfg(not(test))] +impl Global { + #[inline] + fn alloc_impl(&self, layout: Layout, zeroed: bool) -> Result<NonNull<[u8]>, AllocError> { + match layout.size() { + 0 => Ok(NonNull::slice_from_raw_parts(layout.dangling(), 0)), + // SAFETY: `layout` is non-zero in size, + size => unsafe { + let raw_ptr = if zeroed { alloc_zeroed(layout) } else { alloc(layout) }; + let ptr = NonNull::new(raw_ptr).ok_or(AllocError)?; + Ok(NonNull::slice_from_raw_parts(ptr, size)) + }, + } + } + + // SAFETY: Same as `Allocator::grow` + #[inline] + unsafe fn grow_impl( + &self, + ptr: NonNull<u8>, + old_layout: Layout, + new_layout: Layout, + zeroed: bool, + ) -> Result<NonNull<[u8]>, AllocError> { + debug_assert!( + new_layout.size() >= old_layout.size(), + "`new_layout.size()` must be greater than or equal to `old_layout.size()`" + ); + + match old_layout.size() { + 0 => self.alloc_impl(new_layout, zeroed), + + // SAFETY: `new_size` is non-zero as `old_size` is greater than or equal to `new_size` + // as required by safety conditions. Other conditions must be upheld by the caller + old_size if old_layout.align() == new_layout.align() => unsafe { + let new_size = new_layout.size(); + + // `realloc` probably checks for `new_size >= old_layout.size()` or something similar. + intrinsics::assume(new_size >= old_layout.size()); + + let raw_ptr = realloc(ptr.as_ptr(), old_layout, new_size); + let ptr = NonNull::new(raw_ptr).ok_or(AllocError)?; + if zeroed { + raw_ptr.add(old_size).write_bytes(0, new_size - old_size); + } + Ok(NonNull::slice_from_raw_parts(ptr, new_size)) + }, + + // SAFETY: because `new_layout.size()` must be greater than or equal to `old_size`, + // both the old and new memory allocation are valid for reads and writes for `old_size` + // bytes. Also, because the old allocation wasn't yet deallocated, it cannot overlap + // `new_ptr`. Thus, the call to `copy_nonoverlapping` is safe. The safety contract + // for `dealloc` must be upheld by the caller. + old_size => unsafe { + let new_ptr = self.alloc_impl(new_layout, zeroed)?; + ptr::copy_nonoverlapping(ptr.as_ptr(), new_ptr.as_mut_ptr(), old_size); + self.deallocate(ptr, old_layout); + Ok(new_ptr) + }, + } + } +} + +#[unstable(feature = "allocator_api", issue = "32838")] +#[cfg(not(test))] +unsafe impl Allocator for Global { + #[inline] + fn allocate(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError> { + self.alloc_impl(layout, false) + } + + #[inline] + fn allocate_zeroed(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError> { + self.alloc_impl(layout, true) + } + + #[inline] + unsafe fn deallocate(&self, ptr: NonNull<u8>, layout: Layout) { + if layout.size() != 0 { + // SAFETY: `layout` is non-zero in size, + // other conditions must be upheld by the caller + unsafe { dealloc(ptr.as_ptr(), layout) } + } + } + + #[inline] + unsafe fn grow( + &self, + ptr: NonNull<u8>, + old_layout: Layout, + new_layout: Layout, + ) -> Result<NonNull<[u8]>, AllocError> { + // SAFETY: all conditions must be upheld by the caller + unsafe { self.grow_impl(ptr, old_layout, new_layout, false) } + } + + #[inline] + unsafe fn grow_zeroed( + &self, + ptr: NonNull<u8>, + old_layout: Layout, + new_layout: Layout, + ) -> Result<NonNull<[u8]>, AllocError> { + // SAFETY: all conditions must be upheld by the caller + unsafe { self.grow_impl(ptr, old_layout, new_layout, true) } + } + + #[inline] + unsafe fn shrink( + &self, + ptr: NonNull<u8>, + old_layout: Layout, + new_layout: Layout, + ) -> Result<NonNull<[u8]>, AllocError> { + debug_assert!( + new_layout.size() <= old_layout.size(), + "`new_layout.size()` must be smaller than or equal to `old_layout.size()`" + ); + + match new_layout.size() { + // SAFETY: conditions must be upheld by the caller + 0 => unsafe { + self.deallocate(ptr, old_layout); + Ok(NonNull::slice_from_raw_parts(new_layout.dangling(), 0)) + }, + + // SAFETY: `new_size` is non-zero. Other conditions must be upheld by the caller + new_size if old_layout.align() == new_layout.align() => unsafe { + // `realloc` probably checks for `new_size <= old_layout.size()` or something similar. + intrinsics::assume(new_size <= old_layout.size()); + + let raw_ptr = realloc(ptr.as_ptr(), old_layout, new_size); + let ptr = NonNull::new(raw_ptr).ok_or(AllocError)?; + Ok(NonNull::slice_from_raw_parts(ptr, new_size)) + }, + + // SAFETY: because `new_size` must be smaller than or equal to `old_layout.size()`, + // both the old and new memory allocation are valid for reads and writes for `new_size` + // bytes. Also, because the old allocation wasn't yet deallocated, it cannot overlap + // `new_ptr`. Thus, the call to `copy_nonoverlapping` is safe. The safety contract + // for `dealloc` must be upheld by the caller. + new_size => unsafe { + let new_ptr = self.allocate(new_layout)?; + ptr::copy_nonoverlapping(ptr.as_ptr(), new_ptr.as_mut_ptr(), new_size); + self.deallocate(ptr, old_layout); + Ok(new_ptr) + }, + } + } +} + +/// The allocator for unique pointers. +#[cfg(all(not(no_global_oom_handling), not(test)))] +#[lang = "exchange_malloc"] +#[inline] +unsafe fn exchange_malloc(size: usize, align: usize) -> *mut u8 { + let layout = unsafe { Layout::from_size_align_unchecked(size, align) }; + match Global.allocate(layout) { + Ok(ptr) => ptr.as_mut_ptr(), + Err(_) => handle_alloc_error(layout), + } +} + +#[cfg_attr(not(test), lang = "box_free")] +#[inline] +#[rustc_const_unstable(feature = "const_box", issue = "92521")] +// This signature has to be the same as `Box`, otherwise an ICE will happen. +// When an additional parameter to `Box` is added (like `A: Allocator`), this has to be added here as +// well. +// For example if `Box` is changed to `struct Box<T: ?Sized, A: Allocator>(Unique<T>, A)`, +// this function has to be changed to `fn box_free<T: ?Sized, A: Allocator>(Unique<T>, A)` as well. +pub(crate) const unsafe fn box_free<T: ?Sized, A: ~const Allocator + ~const Destruct>( + ptr: Unique<T>, + alloc: A, +) { + unsafe { + let size = size_of_val(ptr.as_ref()); + let align = min_align_of_val(ptr.as_ref()); + let layout = Layout::from_size_align_unchecked(size, align); + alloc.deallocate(From::from(ptr.cast()), layout) + } +} + +// # Allocation error handler + +#[cfg(not(no_global_oom_handling))] +extern "Rust" { + // This is the magic symbol to call the global alloc error handler. rustc generates + // it to call `__rg_oom` if there is a `#[alloc_error_handler]`, or to call the + // default implementations below (`__rdl_oom`) otherwise. + fn __rust_alloc_error_handler(size: usize, align: usize) -> !; +} + +/// Abort on memory allocation error or failure. +/// +/// Callers of memory allocation APIs wishing to abort computation +/// in response to an allocation error are encouraged to call this function, +/// rather than directly invoking `panic!` or similar. +/// +/// The default behavior of this function is to print a message to standard error +/// and abort the process. +/// It can be replaced with [`set_alloc_error_hook`] and [`take_alloc_error_hook`]. +/// +/// [`set_alloc_error_hook`]: ../../std/alloc/fn.set_alloc_error_hook.html +/// [`take_alloc_error_hook`]: ../../std/alloc/fn.take_alloc_error_hook.html +#[stable(feature = "global_alloc", since = "1.28.0")] +#[rustc_const_unstable(feature = "const_alloc_error", issue = "92523")] +#[cfg(all(not(no_global_oom_handling), not(test)))] +#[cold] +pub const fn handle_alloc_error(layout: Layout) -> ! { + const fn ct_error(_: Layout) -> ! { + panic!("allocation failed"); + } + + fn rt_error(layout: Layout) -> ! { + unsafe { + __rust_alloc_error_handler(layout.size(), layout.align()); + } + } + + unsafe { core::intrinsics::const_eval_select((layout,), ct_error, rt_error) } +} + +// For alloc test `std::alloc::handle_alloc_error` can be used directly. +#[cfg(all(not(no_global_oom_handling), test))] +pub use std::alloc::handle_alloc_error; + +#[cfg(all(not(no_global_oom_handling), not(test)))] +#[doc(hidden)] +#[allow(unused_attributes)] +#[unstable(feature = "alloc_internals", issue = "none")] +pub mod __alloc_error_handler { + use crate::alloc::Layout; + + // called via generated `__rust_alloc_error_handler` + + // if there is no `#[alloc_error_handler]` + #[rustc_std_internal_symbol] + pub unsafe fn __rdl_oom(size: usize, _align: usize) -> ! { + panic!("memory allocation of {size} bytes failed") + } + + // if there is an `#[alloc_error_handler]` + #[rustc_std_internal_symbol] + pub unsafe fn __rg_oom(size: usize, align: usize) -> ! { + let layout = unsafe { Layout::from_size_align_unchecked(size, align) }; + extern "Rust" { + #[lang = "oom"] + fn oom_impl(layout: Layout) -> !; + } + unsafe { oom_impl(layout) } + } +} + +/// Specialize clones into pre-allocated, uninitialized memory. +/// Used by `Box::clone` and `Rc`/`Arc::make_mut`. +pub(crate) trait WriteCloneIntoRaw: Sized { + unsafe fn write_clone_into_raw(&self, target: *mut Self); +} + +impl<T: Clone> WriteCloneIntoRaw for T { + #[inline] + default unsafe fn write_clone_into_raw(&self, target: *mut Self) { + // Having allocated *first* may allow the optimizer to create + // the cloned value in-place, skipping the local and move. + unsafe { target.write(self.clone()) }; + } +} + +impl<T: Copy> WriteCloneIntoRaw for T { + #[inline] + unsafe fn write_clone_into_raw(&self, target: *mut Self) { + // We can always copy in-place, without ever involving a local value. + unsafe { target.copy_from_nonoverlapping(self, 1) }; + } +} diff --git a/library/alloc/src/alloc/tests.rs b/library/alloc/src/alloc/tests.rs new file mode 100644 index 000000000..7d560964d --- /dev/null +++ b/library/alloc/src/alloc/tests.rs @@ -0,0 +1,30 @@ +use super::*; + +extern crate test; +use crate::boxed::Box; +use test::Bencher; + +#[test] +fn allocate_zeroed() { + unsafe { + let layout = Layout::from_size_align(1024, 1).unwrap(); + let ptr = + Global.allocate_zeroed(layout.clone()).unwrap_or_else(|_| handle_alloc_error(layout)); + + let mut i = ptr.as_non_null_ptr().as_ptr(); + let end = i.add(layout.size()); + while i < end { + assert_eq!(*i, 0); + i = i.offset(1); + } + Global.deallocate(ptr.as_non_null_ptr(), layout); + } +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn alloc_owned_small(b: &mut Bencher) { + b.iter(|| { + let _: Box<_> = Box::new(10); + }) +} diff --git a/library/alloc/src/borrow.rs b/library/alloc/src/borrow.rs new file mode 100644 index 000000000..904a53bb4 --- /dev/null +++ b/library/alloc/src/borrow.rs @@ -0,0 +1,495 @@ +//! A module for working with borrowed data. + +#![stable(feature = "rust1", since = "1.0.0")] + +use core::cmp::Ordering; +use core::hash::{Hash, Hasher}; +use core::ops::Deref; +#[cfg(not(no_global_oom_handling))] +use core::ops::{Add, AddAssign}; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::borrow::{Borrow, BorrowMut}; + +use crate::fmt; +#[cfg(not(no_global_oom_handling))] +use crate::string::String; + +use Cow::*; + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, B: ?Sized> Borrow<B> for Cow<'a, B> +where + B: ToOwned, + <B as ToOwned>::Owned: 'a, +{ + fn borrow(&self) -> &B { + &**self + } +} + +/// A generalization of `Clone` to borrowed data. +/// +/// Some types make it possible to go from borrowed to owned, usually by +/// implementing the `Clone` trait. But `Clone` works only for going from `&T` +/// to `T`. The `ToOwned` trait generalizes `Clone` to construct owned data +/// from any borrow of a given type. +#[cfg_attr(not(test), rustc_diagnostic_item = "ToOwned")] +#[stable(feature = "rust1", since = "1.0.0")] +pub trait ToOwned { + /// The resulting type after obtaining ownership. + #[stable(feature = "rust1", since = "1.0.0")] + type Owned: Borrow<Self>; + + /// Creates owned data from borrowed data, usually by cloning. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s: &str = "a"; + /// let ss: String = s.to_owned(); + /// + /// let v: &[i32] = &[1, 2]; + /// let vv: Vec<i32> = v.to_owned(); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use = "cloning is often expensive and is not expected to have side effects"] + fn to_owned(&self) -> Self::Owned; + + /// Uses borrowed data to replace owned data, usually by cloning. + /// + /// This is borrow-generalized version of [`Clone::clone_from`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s: String = String::new(); + /// "hello".clone_into(&mut s); + /// + /// let mut v: Vec<i32> = Vec::new(); + /// [1, 2][..].clone_into(&mut v); + /// ``` + #[stable(feature = "toowned_clone_into", since = "1.63.0")] + fn clone_into(&self, target: &mut Self::Owned) { + *target = self.to_owned(); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ToOwned for T +where + T: Clone, +{ + type Owned = T; + fn to_owned(&self) -> T { + self.clone() + } + + fn clone_into(&self, target: &mut T) { + target.clone_from(self); + } +} + +/// A clone-on-write smart pointer. +/// +/// The type `Cow` is a smart pointer providing clone-on-write functionality: it +/// can enclose and provide immutable access to borrowed data, and clone the +/// data lazily when mutation or ownership is required. The type is designed to +/// work with general borrowed data via the `Borrow` trait. +/// +/// `Cow` implements `Deref`, which means that you can call +/// non-mutating methods directly on the data it encloses. If mutation +/// is desired, `to_mut` will obtain a mutable reference to an owned +/// value, cloning if necessary. +/// +/// If you need reference-counting pointers, note that +/// [`Rc::make_mut`][crate::rc::Rc::make_mut] and +/// [`Arc::make_mut`][crate::sync::Arc::make_mut] can provide clone-on-write +/// functionality as well. +/// +/// # Examples +/// +/// ``` +/// use std::borrow::Cow; +/// +/// fn abs_all(input: &mut Cow<[i32]>) { +/// for i in 0..input.len() { +/// let v = input[i]; +/// if v < 0 { +/// // Clones into a vector if not already owned. +/// input.to_mut()[i] = -v; +/// } +/// } +/// } +/// +/// // No clone occurs because `input` doesn't need to be mutated. +/// let slice = [0, 1, 2]; +/// let mut input = Cow::from(&slice[..]); +/// abs_all(&mut input); +/// +/// // Clone occurs because `input` needs to be mutated. +/// let slice = [-1, 0, 1]; +/// let mut input = Cow::from(&slice[..]); +/// abs_all(&mut input); +/// +/// // No clone occurs because `input` is already owned. +/// let mut input = Cow::from(vec![-1, 0, 1]); +/// abs_all(&mut input); +/// ``` +/// +/// Another example showing how to keep `Cow` in a struct: +/// +/// ``` +/// use std::borrow::Cow; +/// +/// struct Items<'a, X: 'a> where [X]: ToOwned<Owned = Vec<X>> { +/// values: Cow<'a, [X]>, +/// } +/// +/// impl<'a, X: Clone + 'a> Items<'a, X> where [X]: ToOwned<Owned = Vec<X>> { +/// fn new(v: Cow<'a, [X]>) -> Self { +/// Items { values: v } +/// } +/// } +/// +/// // Creates a container from borrowed values of a slice +/// let readonly = [1, 2]; +/// let borrowed = Items::new((&readonly[..]).into()); +/// match borrowed { +/// Items { values: Cow::Borrowed(b) } => println!("borrowed {b:?}"), +/// _ => panic!("expect borrowed value"), +/// } +/// +/// let mut clone_on_write = borrowed; +/// // Mutates the data from slice into owned vec and pushes a new value on top +/// clone_on_write.values.to_mut().push(3); +/// println!("clone_on_write = {:?}", clone_on_write.values); +/// +/// // The data was mutated. Let's check it out. +/// match clone_on_write { +/// Items { values: Cow::Owned(_) } => println!("clone_on_write contains owned data"), +/// _ => panic!("expect owned data"), +/// } +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "Cow")] +pub enum Cow<'a, B: ?Sized + 'a> +where + B: ToOwned, +{ + /// Borrowed data. + #[stable(feature = "rust1", since = "1.0.0")] + Borrowed(#[stable(feature = "rust1", since = "1.0.0")] &'a B), + + /// Owned data. + #[stable(feature = "rust1", since = "1.0.0")] + Owned(#[stable(feature = "rust1", since = "1.0.0")] <B as ToOwned>::Owned), +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized + ToOwned> Clone for Cow<'_, B> { + fn clone(&self) -> Self { + match *self { + Borrowed(b) => Borrowed(b), + Owned(ref o) => { + let b: &B = o.borrow(); + Owned(b.to_owned()) + } + } + } + + fn clone_from(&mut self, source: &Self) { + match (self, source) { + (&mut Owned(ref mut dest), &Owned(ref o)) => o.borrow().clone_into(dest), + (t, s) => *t = s.clone(), + } + } +} + +impl<B: ?Sized + ToOwned> Cow<'_, B> { + /// Returns true if the data is borrowed, i.e. if `to_mut` would require additional work. + /// + /// # Examples + /// + /// ``` + /// #![feature(cow_is_borrowed)] + /// use std::borrow::Cow; + /// + /// let cow = Cow::Borrowed("moo"); + /// assert!(cow.is_borrowed()); + /// + /// let bull: Cow<'_, str> = Cow::Owned("...moo?".to_string()); + /// assert!(!bull.is_borrowed()); + /// ``` + #[unstable(feature = "cow_is_borrowed", issue = "65143")] + #[rustc_const_unstable(feature = "const_cow_is_borrowed", issue = "65143")] + pub const fn is_borrowed(&self) -> bool { + match *self { + Borrowed(_) => true, + Owned(_) => false, + } + } + + /// Returns true if the data is owned, i.e. if `to_mut` would be a no-op. + /// + /// # Examples + /// + /// ``` + /// #![feature(cow_is_borrowed)] + /// use std::borrow::Cow; + /// + /// let cow: Cow<'_, str> = Cow::Owned("moo".to_string()); + /// assert!(cow.is_owned()); + /// + /// let bull = Cow::Borrowed("...moo?"); + /// assert!(!bull.is_owned()); + /// ``` + #[unstable(feature = "cow_is_borrowed", issue = "65143")] + #[rustc_const_unstable(feature = "const_cow_is_borrowed", issue = "65143")] + pub const fn is_owned(&self) -> bool { + !self.is_borrowed() + } + + /// Acquires a mutable reference to the owned form of the data. + /// + /// Clones the data if it is not already owned. + /// + /// # Examples + /// + /// ``` + /// use std::borrow::Cow; + /// + /// let mut cow = Cow::Borrowed("foo"); + /// cow.to_mut().make_ascii_uppercase(); + /// + /// assert_eq!( + /// cow, + /// Cow::Owned(String::from("FOO")) as Cow<str> + /// ); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn to_mut(&mut self) -> &mut <B as ToOwned>::Owned { + match *self { + Borrowed(borrowed) => { + *self = Owned(borrowed.to_owned()); + match *self { + Borrowed(..) => unreachable!(), + Owned(ref mut owned) => owned, + } + } + Owned(ref mut owned) => owned, + } + } + + /// Extracts the owned data. + /// + /// Clones the data if it is not already owned. + /// + /// # Examples + /// + /// Calling `into_owned` on a `Cow::Borrowed` returns a clone of the borrowed data: + /// + /// ``` + /// use std::borrow::Cow; + /// + /// let s = "Hello world!"; + /// let cow = Cow::Borrowed(s); + /// + /// assert_eq!( + /// cow.into_owned(), + /// String::from(s) + /// ); + /// ``` + /// + /// Calling `into_owned` on a `Cow::Owned` returns the owned data. The data is moved out of the + /// `Cow` without being cloned. + /// + /// ``` + /// use std::borrow::Cow; + /// + /// let s = "Hello world!"; + /// let cow: Cow<str> = Cow::Owned(String::from(s)); + /// + /// assert_eq!( + /// cow.into_owned(), + /// String::from(s) + /// ); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_owned(self) -> <B as ToOwned>::Owned { + match self { + Borrowed(borrowed) => borrowed.to_owned(), + Owned(owned) => owned, + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_deref", issue = "88955")] +impl<B: ?Sized + ToOwned> const Deref for Cow<'_, B> +where + B::Owned: ~const Borrow<B>, +{ + type Target = B; + + fn deref(&self) -> &B { + match *self { + Borrowed(borrowed) => borrowed, + Owned(ref owned) => owned.borrow(), + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized> Eq for Cow<'_, B> where B: Eq + ToOwned {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized> Ord for Cow<'_, B> +where + B: Ord + ToOwned, +{ + #[inline] + fn cmp(&self, other: &Self) -> Ordering { + Ord::cmp(&**self, &**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, 'b, B: ?Sized, C: ?Sized> PartialEq<Cow<'b, C>> for Cow<'a, B> +where + B: PartialEq<C> + ToOwned, + C: ToOwned, +{ + #[inline] + fn eq(&self, other: &Cow<'b, C>) -> bool { + PartialEq::eq(&**self, &**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, B: ?Sized> PartialOrd for Cow<'a, B> +where + B: PartialOrd + ToOwned, +{ + #[inline] + fn partial_cmp(&self, other: &Cow<'a, B>) -> Option<Ordering> { + PartialOrd::partial_cmp(&**self, &**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized> fmt::Debug for Cow<'_, B> +where + B: fmt::Debug + ToOwned<Owned: fmt::Debug>, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match *self { + Borrowed(ref b) => fmt::Debug::fmt(b, f), + Owned(ref o) => fmt::Debug::fmt(o, f), + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized> fmt::Display for Cow<'_, B> +where + B: fmt::Display + ToOwned<Owned: fmt::Display>, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match *self { + Borrowed(ref b) => fmt::Display::fmt(b, f), + Owned(ref o) => fmt::Display::fmt(o, f), + } + } +} + +#[stable(feature = "default", since = "1.11.0")] +impl<B: ?Sized> Default for Cow<'_, B> +where + B: ToOwned<Owned: Default>, +{ + /// Creates an owned Cow<'a, B> with the default value for the contained owned value. + fn default() -> Self { + Owned(<B as ToOwned>::Owned::default()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<B: ?Sized> Hash for Cow<'_, B> +where + B: Hash + ToOwned, +{ + #[inline] + fn hash<H: Hasher>(&self, state: &mut H) { + Hash::hash(&**self, state) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + ToOwned> AsRef<T> for Cow<'_, T> { + fn as_ref(&self) -> &T { + self + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_add", since = "1.14.0")] +impl<'a> Add<&'a str> for Cow<'a, str> { + type Output = Cow<'a, str>; + + #[inline] + fn add(mut self, rhs: &'a str) -> Self::Output { + self += rhs; + self + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_add", since = "1.14.0")] +impl<'a> Add<Cow<'a, str>> for Cow<'a, str> { + type Output = Cow<'a, str>; + + #[inline] + fn add(mut self, rhs: Cow<'a, str>) -> Self::Output { + self += rhs; + self + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_add", since = "1.14.0")] +impl<'a> AddAssign<&'a str> for Cow<'a, str> { + fn add_assign(&mut self, rhs: &'a str) { + if self.is_empty() { + *self = Cow::Borrowed(rhs) + } else if !rhs.is_empty() { + if let Cow::Borrowed(lhs) = *self { + let mut s = String::with_capacity(lhs.len() + rhs.len()); + s.push_str(lhs); + *self = Cow::Owned(s); + } + self.to_mut().push_str(rhs); + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_add", since = "1.14.0")] +impl<'a> AddAssign<Cow<'a, str>> for Cow<'a, str> { + fn add_assign(&mut self, rhs: Cow<'a, str>) { + if self.is_empty() { + *self = rhs + } else if !rhs.is_empty() { + if let Cow::Borrowed(lhs) = *self { + let mut s = String::with_capacity(lhs.len() + rhs.len()); + s.push_str(lhs); + *self = Cow::Owned(s); + } + self.to_mut().push_str(&rhs); + } + } +} diff --git a/library/alloc/src/boxed.rs b/library/alloc/src/boxed.rs new file mode 100644 index 000000000..c1ceeb0de --- /dev/null +++ b/library/alloc/src/boxed.rs @@ -0,0 +1,2087 @@ +//! A pointer type for heap allocation. +//! +//! [`Box<T>`], casually referred to as a 'box', provides the simplest form of +//! heap allocation in Rust. Boxes provide ownership for this allocation, and +//! drop their contents when they go out of scope. Boxes also ensure that they +//! never allocate more than `isize::MAX` bytes. +//! +//! # Examples +//! +//! Move a value from the stack to the heap by creating a [`Box`]: +//! +//! ``` +//! let val: u8 = 5; +//! let boxed: Box<u8> = Box::new(val); +//! ``` +//! +//! Move a value from a [`Box`] back to the stack by [dereferencing]: +//! +//! ``` +//! let boxed: Box<u8> = Box::new(5); +//! let val: u8 = *boxed; +//! ``` +//! +//! Creating a recursive data structure: +//! +//! ``` +//! #[derive(Debug)] +//! enum List<T> { +//! Cons(T, Box<List<T>>), +//! Nil, +//! } +//! +//! let list: List<i32> = List::Cons(1, Box::new(List::Cons(2, Box::new(List::Nil)))); +//! println!("{list:?}"); +//! ``` +//! +//! This will print `Cons(1, Cons(2, Nil))`. +//! +//! Recursive structures must be boxed, because if the definition of `Cons` +//! looked like this: +//! +//! ```compile_fail,E0072 +//! # enum List<T> { +//! Cons(T, List<T>), +//! # } +//! ``` +//! +//! It wouldn't work. This is because the size of a `List` depends on how many +//! elements are in the list, and so we don't know how much memory to allocate +//! for a `Cons`. By introducing a [`Box<T>`], which has a defined size, we know how +//! big `Cons` needs to be. +//! +//! # Memory layout +//! +//! For non-zero-sized values, a [`Box`] will use the [`Global`] allocator for +//! its allocation. It is valid to convert both ways between a [`Box`] and a +//! raw pointer allocated with the [`Global`] allocator, given that the +//! [`Layout`] used with the allocator is correct for the type. More precisely, +//! a `value: *mut T` that has been allocated with the [`Global`] allocator +//! with `Layout::for_value(&*value)` may be converted into a box using +//! [`Box::<T>::from_raw(value)`]. Conversely, the memory backing a `value: *mut +//! T` obtained from [`Box::<T>::into_raw`] may be deallocated using the +//! [`Global`] allocator with [`Layout::for_value(&*value)`]. +//! +//! For zero-sized values, the `Box` pointer still has to be [valid] for reads +//! and writes and sufficiently aligned. In particular, casting any aligned +//! non-zero integer literal to a raw pointer produces a valid pointer, but a +//! pointer pointing into previously allocated memory that since got freed is +//! not valid. The recommended way to build a Box to a ZST if `Box::new` cannot +//! be used is to use [`ptr::NonNull::dangling`]. +//! +//! So long as `T: Sized`, a `Box<T>` is guaranteed to be represented +//! as a single pointer and is also ABI-compatible with C pointers +//! (i.e. the C type `T*`). This means that if you have extern "C" +//! Rust functions that will be called from C, you can define those +//! Rust functions using `Box<T>` types, and use `T*` as corresponding +//! type on the C side. As an example, consider this C header which +//! declares functions that create and destroy some kind of `Foo` +//! value: +//! +//! ```c +//! /* C header */ +//! +//! /* Returns ownership to the caller */ +//! struct Foo* foo_new(void); +//! +//! /* Takes ownership from the caller; no-op when invoked with null */ +//! void foo_delete(struct Foo*); +//! ``` +//! +//! These two functions might be implemented in Rust as follows. Here, the +//! `struct Foo*` type from C is translated to `Box<Foo>`, which captures +//! the ownership constraints. Note also that the nullable argument to +//! `foo_delete` is represented in Rust as `Option<Box<Foo>>`, since `Box<Foo>` +//! cannot be null. +//! +//! ``` +//! #[repr(C)] +//! pub struct Foo; +//! +//! #[no_mangle] +//! pub extern "C" fn foo_new() -> Box<Foo> { +//! Box::new(Foo) +//! } +//! +//! #[no_mangle] +//! pub extern "C" fn foo_delete(_: Option<Box<Foo>>) {} +//! ``` +//! +//! Even though `Box<T>` has the same representation and C ABI as a C pointer, +//! this does not mean that you can convert an arbitrary `T*` into a `Box<T>` +//! and expect things to work. `Box<T>` values will always be fully aligned, +//! non-null pointers. Moreover, the destructor for `Box<T>` will attempt to +//! free the value with the global allocator. In general, the best practice +//! is to only use `Box<T>` for pointers that originated from the global +//! allocator. +//! +//! **Important.** At least at present, you should avoid using +//! `Box<T>` types for functions that are defined in C but invoked +//! from Rust. In those cases, you should directly mirror the C types +//! as closely as possible. Using types like `Box<T>` where the C +//! definition is just using `T*` can lead to undefined behavior, as +//! described in [rust-lang/unsafe-code-guidelines#198][ucg#198]. +//! +//! # Considerations for unsafe code +//! +//! **Warning: This section is not normative and is subject to change, possibly +//! being relaxed in the future! It is a simplified summary of the rules +//! currently implemented in the compiler.** +//! +//! The aliasing rules for `Box<T>` are the same as for `&mut T`. `Box<T>` +//! asserts uniqueness over its content. Using raw pointers derived from a box +//! after that box has been mutated through, moved or borrowed as `&mut T` +//! is not allowed. For more guidance on working with box from unsafe code, see +//! [rust-lang/unsafe-code-guidelines#326][ucg#326]. +//! +//! +//! [ucg#198]: https://github.com/rust-lang/unsafe-code-guidelines/issues/198 +//! [ucg#326]: https://github.com/rust-lang/unsafe-code-guidelines/issues/326 +//! [dereferencing]: core::ops::Deref +//! [`Box::<T>::from_raw(value)`]: Box::from_raw +//! [`Global`]: crate::alloc::Global +//! [`Layout`]: crate::alloc::Layout +//! [`Layout::for_value(&*value)`]: crate::alloc::Layout::for_value +//! [valid]: ptr#safety + +#![stable(feature = "rust1", since = "1.0.0")] + +use core::any::Any; +use core::async_iter::AsyncIterator; +use core::borrow; +use core::cmp::Ordering; +use core::convert::{From, TryFrom}; +use core::fmt; +use core::future::Future; +use core::hash::{Hash, Hasher}; +#[cfg(not(no_global_oom_handling))] +use core::iter::FromIterator; +use core::iter::{FusedIterator, Iterator}; +use core::marker::{Destruct, Unpin, Unsize}; +use core::mem; +use core::ops::{ + CoerceUnsized, Deref, DerefMut, DispatchFromDyn, Generator, GeneratorState, Receiver, +}; +use core::pin::Pin; +use core::ptr::{self, Unique}; +use core::task::{Context, Poll}; + +#[cfg(not(no_global_oom_handling))] +use crate::alloc::{handle_alloc_error, WriteCloneIntoRaw}; +use crate::alloc::{AllocError, Allocator, Global, Layout}; +#[cfg(not(no_global_oom_handling))] +use crate::borrow::Cow; +use crate::raw_vec::RawVec; +#[cfg(not(no_global_oom_handling))] +use crate::str::from_boxed_utf8_unchecked; +#[cfg(not(no_global_oom_handling))] +use crate::vec::Vec; + +#[unstable(feature = "thin_box", issue = "92791")] +pub use thin::ThinBox; + +mod thin; + +/// A pointer type for heap allocation. +/// +/// See the [module-level documentation](../../std/boxed/index.html) for more. +#[lang = "owned_box"] +#[fundamental] +#[stable(feature = "rust1", since = "1.0.0")] +// The declaration of the `Box` struct must be kept in sync with the +// `alloc::alloc::box_free` function or ICEs will happen. See the comment +// on `box_free` for more details. +pub struct Box< + T: ?Sized, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +>(Unique<T>, A); + +impl<T> Box<T> { + /// Allocates memory on the heap and then places `x` into it. + /// + /// This doesn't actually allocate if `T` is zero-sized. + /// + /// # Examples + /// + /// ``` + /// let five = Box::new(5); + /// ``` + #[cfg(all(not(no_global_oom_handling)))] + #[inline(always)] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn new(x: T) -> Self { + #[rustc_box] + Box::new(x) + } + + /// Constructs a new box with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let mut five = Box::<u32>::new_uninit(); + /// + /// let five = unsafe { + /// // Deferred initialization: + /// five.as_mut_ptr().write(5); + /// + /// five.assume_init() + /// }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + #[inline] + pub fn new_uninit() -> Box<mem::MaybeUninit<T>> { + Self::new_uninit_in(Global) + } + + /// Constructs a new `Box` with uninitialized contents, with the memory + /// being filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let zero = Box::<u32>::new_zeroed(); + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[inline] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed() -> Box<mem::MaybeUninit<T>> { + Self::new_zeroed_in(Global) + } + + /// Constructs a new `Pin<Box<T>>`. If `T` does not implement [`Unpin`], then + /// `x` will be pinned in memory and unable to be moved. + /// + /// Constructing and pinning of the `Box` can also be done in two steps: `Box::pin(x)` + /// does the same as <code>[Box::into_pin]\([Box::new]\(x))</code>. Consider using + /// [`into_pin`](Box::into_pin) if you already have a `Box<T>`, or if you want to + /// construct a (pinned) `Box` in a different way than with [`Box::new`]. + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "pin", since = "1.33.0")] + #[must_use] + #[inline(always)] + pub fn pin(x: T) -> Pin<Box<T>> { + (#[rustc_box] + Box::new(x)) + .into() + } + + /// Allocates memory on the heap then places `x` into it, + /// returning an error if the allocation fails + /// + /// This doesn't actually allocate if `T` is zero-sized. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// let five = Box::try_new(5)?; + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn try_new(x: T) -> Result<Self, AllocError> { + Self::try_new_in(x, Global) + } + + /// Constructs a new box with uninitialized contents on the heap, + /// returning an error if the allocation fails + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// let mut five = Box::<u32>::try_new_uninit()?; + /// + /// let five = unsafe { + /// // Deferred initialization: + /// five.as_mut_ptr().write(5); + /// + /// five.assume_init() + /// }; + /// + /// assert_eq!(*five, 5); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[inline] + pub fn try_new_uninit() -> Result<Box<mem::MaybeUninit<T>>, AllocError> { + Box::try_new_uninit_in(Global) + } + + /// Constructs a new `Box` with uninitialized contents, with the memory + /// being filled with `0` bytes on the heap + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// let zero = Box::<u32>::try_new_zeroed()?; + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[inline] + pub fn try_new_zeroed() -> Result<Box<mem::MaybeUninit<T>>, AllocError> { + Box::try_new_zeroed_in(Global) + } +} + +impl<T, A: Allocator> Box<T, A> { + /// Allocates memory in the given allocator then places `x` into it. + /// + /// This doesn't actually allocate if `T` is zero-sized. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let five = Box::new_in(5, System); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[must_use] + #[inline] + pub const fn new_in(x: T, alloc: A) -> Self + where + A: ~const Allocator + ~const Destruct, + { + let mut boxed = Self::new_uninit_in(alloc); + unsafe { + boxed.as_mut_ptr().write(x); + boxed.assume_init() + } + } + + /// Allocates memory in the given allocator then places `x` into it, + /// returning an error if the allocation fails + /// + /// This doesn't actually allocate if `T` is zero-sized. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let five = Box::try_new_in(5, System)?; + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn try_new_in(x: T, alloc: A) -> Result<Self, AllocError> + where + T: ~const Destruct, + A: ~const Allocator + ~const Destruct, + { + let mut boxed = Self::try_new_uninit_in(alloc)?; + unsafe { + boxed.as_mut_ptr().write(x); + Ok(boxed.assume_init()) + } + } + + /// Constructs a new box with uninitialized contents in the provided allocator. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let mut five = Box::<u32, _>::new_uninit_in(System); + /// + /// let five = unsafe { + /// // Deferred initialization: + /// five.as_mut_ptr().write(5); + /// + /// five.assume_init() + /// }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[cfg(not(no_global_oom_handling))] + #[must_use] + // #[unstable(feature = "new_uninit", issue = "63291")] + pub const fn new_uninit_in(alloc: A) -> Box<mem::MaybeUninit<T>, A> + where + A: ~const Allocator + ~const Destruct, + { + let layout = Layout::new::<mem::MaybeUninit<T>>(); + // NOTE: Prefer match over unwrap_or_else since closure sometimes not inlineable. + // That would make code size bigger. + match Box::try_new_uninit_in(alloc) { + Ok(m) => m, + Err(_) => handle_alloc_error(layout), + } + } + + /// Constructs a new box with uninitialized contents in the provided allocator, + /// returning an error if the allocation fails + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let mut five = Box::<u32, _>::try_new_uninit_in(System)?; + /// + /// let five = unsafe { + /// // Deferred initialization: + /// five.as_mut_ptr().write(5); + /// + /// five.assume_init() + /// }; + /// + /// assert_eq!(*five, 5); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + pub const fn try_new_uninit_in(alloc: A) -> Result<Box<mem::MaybeUninit<T>, A>, AllocError> + where + A: ~const Allocator + ~const Destruct, + { + let layout = Layout::new::<mem::MaybeUninit<T>>(); + let ptr = alloc.allocate(layout)?.cast(); + unsafe { Ok(Box::from_raw_in(ptr.as_ptr(), alloc)) } + } + + /// Constructs a new `Box` with uninitialized contents, with the memory + /// being filled with `0` bytes in the provided allocator. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let zero = Box::<u32, _>::new_zeroed_in(System); + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[cfg(not(no_global_oom_handling))] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub const fn new_zeroed_in(alloc: A) -> Box<mem::MaybeUninit<T>, A> + where + A: ~const Allocator + ~const Destruct, + { + let layout = Layout::new::<mem::MaybeUninit<T>>(); + // NOTE: Prefer match over unwrap_or_else since closure sometimes not inlineable. + // That would make code size bigger. + match Box::try_new_zeroed_in(alloc) { + Ok(m) => m, + Err(_) => handle_alloc_error(layout), + } + } + + /// Constructs a new `Box` with uninitialized contents, with the memory + /// being filled with `0` bytes in the provided allocator, + /// returning an error if the allocation fails, + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let zero = Box::<u32, _>::try_new_zeroed_in(System)?; + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + pub const fn try_new_zeroed_in(alloc: A) -> Result<Box<mem::MaybeUninit<T>, A>, AllocError> + where + A: ~const Allocator + ~const Destruct, + { + let layout = Layout::new::<mem::MaybeUninit<T>>(); + let ptr = alloc.allocate_zeroed(layout)?.cast(); + unsafe { Ok(Box::from_raw_in(ptr.as_ptr(), alloc)) } + } + + /// Constructs a new `Pin<Box<T, A>>`. If `T` does not implement [`Unpin`], then + /// `x` will be pinned in memory and unable to be moved. + /// + /// Constructing and pinning of the `Box` can also be done in two steps: `Box::pin_in(x, alloc)` + /// does the same as <code>[Box::into_pin]\([Box::new_in]\(x, alloc))</code>. Consider using + /// [`into_pin`](Box::into_pin) if you already have a `Box<T, A>`, or if you want to + /// construct a (pinned) `Box` in a different way than with [`Box::new_in`]. + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[must_use] + #[inline(always)] + pub const fn pin_in(x: T, alloc: A) -> Pin<Self> + where + A: 'static + ~const Allocator + ~const Destruct, + { + Self::into_pin(Self::new_in(x, alloc)) + } + + /// Converts a `Box<T>` into a `Box<[T]>` + /// + /// This conversion does not allocate on the heap and happens in place. + #[unstable(feature = "box_into_boxed_slice", issue = "71582")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + pub const fn into_boxed_slice(boxed: Self) -> Box<[T], A> { + let (raw, alloc) = Box::into_raw_with_allocator(boxed); + unsafe { Box::from_raw_in(raw as *mut [T; 1], alloc) } + } + + /// Consumes the `Box`, returning the wrapped value. + /// + /// # Examples + /// + /// ``` + /// #![feature(box_into_inner)] + /// + /// let c = Box::new(5); + /// + /// assert_eq!(Box::into_inner(c), 5); + /// ``` + #[unstable(feature = "box_into_inner", issue = "80437")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn into_inner(boxed: Self) -> T + where + Self: ~const Destruct, + { + *boxed + } +} + +impl<T> Box<[T]> { + /// Constructs a new boxed slice with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let mut values = Box::<[u32]>::new_uninit_slice(3); + /// + /// let values = unsafe { + /// // Deferred initialization: + /// values[0].as_mut_ptr().write(1); + /// values[1].as_mut_ptr().write(2); + /// values[2].as_mut_ptr().write(3); + /// + /// values.assume_init() + /// }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit_slice(len: usize) -> Box<[mem::MaybeUninit<T>]> { + unsafe { RawVec::with_capacity(len).into_box(len) } + } + + /// Constructs a new boxed slice with uninitialized contents, with the memory + /// being filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let values = Box::<[u32]>::new_zeroed_slice(3); + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [0, 0, 0]) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed_slice(len: usize) -> Box<[mem::MaybeUninit<T>]> { + unsafe { RawVec::with_capacity_zeroed(len).into_box(len) } + } + + /// Constructs a new boxed slice with uninitialized contents. Returns an error if + /// the allocation fails + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// let mut values = Box::<[u32]>::try_new_uninit_slice(3)?; + /// let values = unsafe { + /// // Deferred initialization: + /// values[0].as_mut_ptr().write(1); + /// values[1].as_mut_ptr().write(2); + /// values[2].as_mut_ptr().write(3); + /// values.assume_init() + /// }; + /// + /// assert_eq!(*values, [1, 2, 3]); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn try_new_uninit_slice(len: usize) -> Result<Box<[mem::MaybeUninit<T>]>, AllocError> { + unsafe { + let layout = match Layout::array::<mem::MaybeUninit<T>>(len) { + Ok(l) => l, + Err(_) => return Err(AllocError), + }; + let ptr = Global.allocate(layout)?; + Ok(RawVec::from_raw_parts_in(ptr.as_mut_ptr() as *mut _, len, Global).into_box(len)) + } + } + + /// Constructs a new boxed slice with uninitialized contents, with the memory + /// being filled with `0` bytes. Returns an error if the allocation fails + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// let values = Box::<[u32]>::try_new_zeroed_slice(3)?; + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [0, 0, 0]); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn try_new_zeroed_slice(len: usize) -> Result<Box<[mem::MaybeUninit<T>]>, AllocError> { + unsafe { + let layout = match Layout::array::<mem::MaybeUninit<T>>(len) { + Ok(l) => l, + Err(_) => return Err(AllocError), + }; + let ptr = Global.allocate_zeroed(layout)?; + Ok(RawVec::from_raw_parts_in(ptr.as_mut_ptr() as *mut _, len, Global).into_box(len)) + } + } +} + +impl<T, A: Allocator> Box<[T], A> { + /// Constructs a new boxed slice with uninitialized contents in the provided allocator. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let mut values = Box::<[u32], _>::new_uninit_slice_in(3, System); + /// + /// let values = unsafe { + /// // Deferred initialization: + /// values[0].as_mut_ptr().write(1); + /// values[1].as_mut_ptr().write(2); + /// values[2].as_mut_ptr().write(3); + /// + /// values.assume_init() + /// }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit_slice_in(len: usize, alloc: A) -> Box<[mem::MaybeUninit<T>], A> { + unsafe { RawVec::with_capacity_in(len, alloc).into_box(len) } + } + + /// Constructs a new boxed slice with uninitialized contents in the provided allocator, + /// with the memory being filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::alloc::System; + /// + /// let values = Box::<[u32], _>::new_zeroed_slice_in(3, System); + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [0, 0, 0]) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed_slice_in(len: usize, alloc: A) -> Box<[mem::MaybeUninit<T>], A> { + unsafe { RawVec::with_capacity_zeroed_in(len, alloc).into_box(len) } + } +} + +impl<T, A: Allocator> Box<mem::MaybeUninit<T>, A> { + /// Converts to `Box<T, A>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the value + /// really is in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let mut five = Box::<u32>::new_uninit(); + /// + /// let five: Box<u32> = unsafe { + /// // Deferred initialization: + /// five.as_mut_ptr().write(5); + /// + /// five.assume_init() + /// }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const unsafe fn assume_init(self) -> Box<T, A> { + let (raw, alloc) = Box::into_raw_with_allocator(self); + unsafe { Box::from_raw_in(raw as *mut T, alloc) } + } + + /// Writes the value and converts to `Box<T, A>`. + /// + /// This method converts the box similarly to [`Box::assume_init`] but + /// writes `value` into it before conversion thus guaranteeing safety. + /// In some scenarios use of this method may improve performance because + /// the compiler may be able to optimize copying from stack. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let big_box = Box::<[usize; 1024]>::new_uninit(); + /// + /// let mut array = [0; 1024]; + /// for (i, place) in array.iter_mut().enumerate() { + /// *place = i; + /// } + /// + /// // The optimizer may be able to elide this copy, so previous code writes + /// // to heap directly. + /// let big_box = Box::write(big_box, array); + /// + /// for (i, x) in big_box.iter().enumerate() { + /// assert_eq!(*x, i); + /// } + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn write(mut boxed: Self, value: T) -> Box<T, A> { + unsafe { + (*boxed).write(value); + boxed.assume_init() + } + } +} + +impl<T, A: Allocator> Box<[mem::MaybeUninit<T>], A> { + /// Converts to `Box<[T], A>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the values + /// really are in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// let mut values = Box::<[u32]>::new_uninit_slice(3); + /// + /// let values = unsafe { + /// // Deferred initialization: + /// values[0].as_mut_ptr().write(1); + /// values[1].as_mut_ptr().write(2); + /// values[2].as_mut_ptr().write(3); + /// + /// values.assume_init() + /// }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[inline] + pub unsafe fn assume_init(self) -> Box<[T], A> { + let (raw, alloc) = Box::into_raw_with_allocator(self); + unsafe { Box::from_raw_in(raw as *mut [T], alloc) } + } +} + +impl<T: ?Sized> Box<T> { + /// Constructs a box from a raw pointer. + /// + /// After calling this function, the raw pointer is owned by the + /// resulting `Box`. Specifically, the `Box` destructor will call + /// the destructor of `T` and free the allocated memory. For this + /// to be safe, the memory must have been allocated in accordance + /// with the [memory layout] used by `Box` . + /// + /// # Safety + /// + /// This function is unsafe because improper use may lead to + /// memory problems. For example, a double-free may occur if the + /// function is called twice on the same raw pointer. + /// + /// The safety conditions are described in the [memory layout] section. + /// + /// # Examples + /// + /// Recreate a `Box` which was previously converted to a raw pointer + /// using [`Box::into_raw`]: + /// ``` + /// let x = Box::new(5); + /// let ptr = Box::into_raw(x); + /// let x = unsafe { Box::from_raw(ptr) }; + /// ``` + /// Manually create a `Box` from scratch by using the global allocator: + /// ``` + /// use std::alloc::{alloc, Layout}; + /// + /// unsafe { + /// let ptr = alloc(Layout::new::<i32>()) as *mut i32; + /// // In general .write is required to avoid attempting to destruct + /// // the (uninitialized) previous contents of `ptr`, though for this + /// // simple example `*ptr = 5` would have worked as well. + /// ptr.write(5); + /// let x = Box::from_raw(ptr); + /// } + /// ``` + /// + /// [memory layout]: self#memory-layout + /// [`Layout`]: crate::Layout + #[stable(feature = "box_raw", since = "1.4.0")] + #[inline] + #[must_use = "call `drop(from_raw(ptr))` if you intend to drop the `Box`"] + pub unsafe fn from_raw(raw: *mut T) -> Self { + unsafe { Self::from_raw_in(raw, Global) } + } +} + +impl<T: ?Sized, A: Allocator> Box<T, A> { + /// Constructs a box from a raw pointer in the given allocator. + /// + /// After calling this function, the raw pointer is owned by the + /// resulting `Box`. Specifically, the `Box` destructor will call + /// the destructor of `T` and free the allocated memory. For this + /// to be safe, the memory must have been allocated in accordance + /// with the [memory layout] used by `Box` . + /// + /// # Safety + /// + /// This function is unsafe because improper use may lead to + /// memory problems. For example, a double-free may occur if the + /// function is called twice on the same raw pointer. + /// + /// + /// # Examples + /// + /// Recreate a `Box` which was previously converted to a raw pointer + /// using [`Box::into_raw_with_allocator`]: + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let x = Box::new_in(5, System); + /// let (ptr, alloc) = Box::into_raw_with_allocator(x); + /// let x = unsafe { Box::from_raw_in(ptr, alloc) }; + /// ``` + /// Manually create a `Box` from scratch by using the system allocator: + /// ``` + /// #![feature(allocator_api, slice_ptr_get)] + /// + /// use std::alloc::{Allocator, Layout, System}; + /// + /// unsafe { + /// let ptr = System.allocate(Layout::new::<i32>())?.as_mut_ptr() as *mut i32; + /// // In general .write is required to avoid attempting to destruct + /// // the (uninitialized) previous contents of `ptr`, though for this + /// // simple example `*ptr = 5` would have worked as well. + /// ptr.write(5); + /// let x = Box::from_raw_in(ptr, System); + /// } + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [memory layout]: self#memory-layout + /// [`Layout`]: crate::Layout + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const unsafe fn from_raw_in(raw: *mut T, alloc: A) -> Self { + Box(unsafe { Unique::new_unchecked(raw) }, alloc) + } + + /// Consumes the `Box`, returning a wrapped raw pointer. + /// + /// The pointer will be properly aligned and non-null. + /// + /// After calling this function, the caller is responsible for the + /// memory previously managed by the `Box`. In particular, the + /// caller should properly destroy `T` and release the memory, taking + /// into account the [memory layout] used by `Box`. The easiest way to + /// do this is to convert the raw pointer back into a `Box` with the + /// [`Box::from_raw`] function, allowing the `Box` destructor to perform + /// the cleanup. + /// + /// Note: this is an associated function, which means that you have + /// to call it as `Box::into_raw(b)` instead of `b.into_raw()`. This + /// is so that there is no conflict with a method on the inner type. + /// + /// # Examples + /// Converting the raw pointer back into a `Box` with [`Box::from_raw`] + /// for automatic cleanup: + /// ``` + /// let x = Box::new(String::from("Hello")); + /// let ptr = Box::into_raw(x); + /// let x = unsafe { Box::from_raw(ptr) }; + /// ``` + /// Manual cleanup by explicitly running the destructor and deallocating + /// the memory: + /// ``` + /// use std::alloc::{dealloc, Layout}; + /// use std::ptr; + /// + /// let x = Box::new(String::from("Hello")); + /// let p = Box::into_raw(x); + /// unsafe { + /// ptr::drop_in_place(p); + /// dealloc(p as *mut u8, Layout::new::<String>()); + /// } + /// ``` + /// + /// [memory layout]: self#memory-layout + #[stable(feature = "box_raw", since = "1.4.0")] + #[inline] + pub fn into_raw(b: Self) -> *mut T { + Self::into_raw_with_allocator(b).0 + } + + /// Consumes the `Box`, returning a wrapped raw pointer and the allocator. + /// + /// The pointer will be properly aligned and non-null. + /// + /// After calling this function, the caller is responsible for the + /// memory previously managed by the `Box`. In particular, the + /// caller should properly destroy `T` and release the memory, taking + /// into account the [memory layout] used by `Box`. The easiest way to + /// do this is to convert the raw pointer back into a `Box` with the + /// [`Box::from_raw_in`] function, allowing the `Box` destructor to perform + /// the cleanup. + /// + /// Note: this is an associated function, which means that you have + /// to call it as `Box::into_raw_with_allocator(b)` instead of `b.into_raw_with_allocator()`. This + /// is so that there is no conflict with a method on the inner type. + /// + /// # Examples + /// Converting the raw pointer back into a `Box` with [`Box::from_raw_in`] + /// for automatic cleanup: + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let x = Box::new_in(String::from("Hello"), System); + /// let (ptr, alloc) = Box::into_raw_with_allocator(x); + /// let x = unsafe { Box::from_raw_in(ptr, alloc) }; + /// ``` + /// Manual cleanup by explicitly running the destructor and deallocating + /// the memory: + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::{Allocator, Layout, System}; + /// use std::ptr::{self, NonNull}; + /// + /// let x = Box::new_in(String::from("Hello"), System); + /// let (ptr, alloc) = Box::into_raw_with_allocator(x); + /// unsafe { + /// ptr::drop_in_place(ptr); + /// let non_null = NonNull::new_unchecked(ptr); + /// alloc.deallocate(non_null.cast(), Layout::new::<String>()); + /// } + /// ``` + /// + /// [memory layout]: self#memory-layout + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn into_raw_with_allocator(b: Self) -> (*mut T, A) { + let (leaked, alloc) = Box::into_unique(b); + (leaked.as_ptr(), alloc) + } + + #[unstable( + feature = "ptr_internals", + issue = "none", + reason = "use `Box::leak(b).into()` or `Unique::from(Box::leak(b))` instead" + )] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + #[doc(hidden)] + pub const fn into_unique(b: Self) -> (Unique<T>, A) { + // Box is recognized as a "unique pointer" by Stacked Borrows, but internally it is a + // raw pointer for the type system. Turning it directly into a raw pointer would not be + // recognized as "releasing" the unique pointer to permit aliased raw accesses, + // so all raw pointer methods have to go through `Box::leak`. Turning *that* to a raw pointer + // behaves correctly. + let alloc = unsafe { ptr::read(&b.1) }; + (Unique::from(Box::leak(b)), alloc) + } + + /// Returns a reference to the underlying allocator. + /// + /// Note: this is an associated function, which means that you have + /// to call it as `Box::allocator(&b)` instead of `b.allocator()`. This + /// is so that there is no conflict with a method on the inner type. + #[unstable(feature = "allocator_api", issue = "32838")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn allocator(b: &Self) -> &A { + &b.1 + } + + /// Consumes and leaks the `Box`, returning a mutable reference, + /// `&'a mut T`. Note that the type `T` must outlive the chosen lifetime + /// `'a`. If the type has only static references, or none at all, then this + /// may be chosen to be `'static`. + /// + /// This function is mainly useful for data that lives for the remainder of + /// the program's life. Dropping the returned reference will cause a memory + /// leak. If this is not acceptable, the reference should first be wrapped + /// with the [`Box::from_raw`] function producing a `Box`. This `Box` can + /// then be dropped which will properly destroy `T` and release the + /// allocated memory. + /// + /// Note: this is an associated function, which means that you have + /// to call it as `Box::leak(b)` instead of `b.leak()`. This + /// is so that there is no conflict with a method on the inner type. + /// + /// # Examples + /// + /// Simple usage: + /// + /// ``` + /// let x = Box::new(41); + /// let static_ref: &'static mut usize = Box::leak(x); + /// *static_ref += 1; + /// assert_eq!(*static_ref, 42); + /// ``` + /// + /// Unsized data: + /// + /// ``` + /// let x = vec![1, 2, 3].into_boxed_slice(); + /// let static_ref = Box::leak(x); + /// static_ref[0] = 4; + /// assert_eq!(*static_ref, [4, 2, 3]); + /// ``` + #[stable(feature = "box_leak", since = "1.26.0")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + #[inline] + pub const fn leak<'a>(b: Self) -> &'a mut T + where + A: 'a, + { + unsafe { &mut *mem::ManuallyDrop::new(b).0.as_ptr() } + } + + /// Converts a `Box<T>` into a `Pin<Box<T>>`. If `T` does not implement [`Unpin`], then + /// `*boxed` will be pinned in memory and unable to be moved. + /// + /// This conversion does not allocate on the heap and happens in place. + /// + /// This is also available via [`From`]. + /// + /// Constructing and pinning a `Box` with <code>Box::into_pin([Box::new]\(x))</code> + /// can also be written more concisely using <code>[Box::pin]\(x)</code>. + /// This `into_pin` method is useful if you already have a `Box<T>`, or you are + /// constructing a (pinned) `Box` in a different way than with [`Box::new`]. + /// + /// # Notes + /// + /// It's not recommended that crates add an impl like `From<Box<T>> for Pin<T>`, + /// as it'll introduce an ambiguity when calling `Pin::from`. + /// A demonstration of such a poor impl is shown below. + /// + /// ```compile_fail + /// # use std::pin::Pin; + /// struct Foo; // A type defined in this crate. + /// impl From<Box<()>> for Pin<Foo> { + /// fn from(_: Box<()>) -> Pin<Foo> { + /// Pin::new(Foo) + /// } + /// } + /// + /// let foo = Box::new(()); + /// let bar = Pin::from(foo); + /// ``` + #[stable(feature = "box_into_pin", since = "1.63.0")] + #[rustc_const_unstable(feature = "const_box", issue = "92521")] + pub const fn into_pin(boxed: Self) -> Pin<Self> + where + A: 'static, + { + // It's not possible to move or replace the insides of a `Pin<Box<T>>` + // when `T: !Unpin`, so it's safe to pin it directly without any + // additional requirements. + unsafe { Pin::new_unchecked(boxed) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T: ?Sized, A: Allocator> Drop for Box<T, A> { + fn drop(&mut self) { + // FIXME: Do nothing, drop is currently performed by compiler. + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Default> Default for Box<T> { + /// Creates a `Box<T>`, with the `Default` value for T. + fn default() -> Self { + #[rustc_box] + Box::new(T::default()) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_default_impls", issue = "87864")] +impl<T> const Default for Box<[T]> { + fn default() -> Self { + let ptr: Unique<[T]> = Unique::<[T; 0]>::dangling(); + Box(ptr, Global) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "default_box_extra", since = "1.17.0")] +#[rustc_const_unstable(feature = "const_default_impls", issue = "87864")] +impl const Default for Box<str> { + fn default() -> Self { + // SAFETY: This is the same as `Unique::cast<U>` but with an unsized `U = str`. + let ptr: Unique<str> = unsafe { + let bytes: Unique<[u8]> = Unique::<[u8; 0]>::dangling(); + Unique::new_unchecked(bytes.as_ptr() as *mut str) + }; + Box(ptr, Global) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone, A: Allocator + Clone> Clone for Box<T, A> { + /// Returns a new box with a `clone()` of this box's contents. + /// + /// # Examples + /// + /// ``` + /// let x = Box::new(5); + /// let y = x.clone(); + /// + /// // The value is the same + /// assert_eq!(x, y); + /// + /// // But they are unique objects + /// assert_ne!(&*x as *const i32, &*y as *const i32); + /// ``` + #[inline] + fn clone(&self) -> Self { + // Pre-allocate memory to allow writing the cloned value directly. + let mut boxed = Self::new_uninit_in(self.1.clone()); + unsafe { + (**self).write_clone_into_raw(boxed.as_mut_ptr()); + boxed.assume_init() + } + } + + /// Copies `source`'s contents into `self` without creating a new allocation. + /// + /// # Examples + /// + /// ``` + /// let x = Box::new(5); + /// let mut y = Box::new(10); + /// let yp: *const i32 = &*y; + /// + /// y.clone_from(&x); + /// + /// // The value is the same + /// assert_eq!(x, y); + /// + /// // And no allocation occurred + /// assert_eq!(yp, &*y); + /// ``` + #[inline] + fn clone_from(&mut self, source: &Self) { + (**self).clone_from(&(**source)); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_slice_clone", since = "1.3.0")] +impl Clone for Box<str> { + fn clone(&self) -> Self { + // this makes a copy of the data + let buf: Box<[u8]> = self.as_bytes().into(); + unsafe { from_boxed_utf8_unchecked(buf) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialEq, A: Allocator> PartialEq for Box<T, A> { + #[inline] + fn eq(&self, other: &Self) -> bool { + PartialEq::eq(&**self, &**other) + } + #[inline] + fn ne(&self, other: &Self) -> bool { + PartialEq::ne(&**self, &**other) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialOrd, A: Allocator> PartialOrd for Box<T, A> { + #[inline] + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + PartialOrd::partial_cmp(&**self, &**other) + } + #[inline] + fn lt(&self, other: &Self) -> bool { + PartialOrd::lt(&**self, &**other) + } + #[inline] + fn le(&self, other: &Self) -> bool { + PartialOrd::le(&**self, &**other) + } + #[inline] + fn ge(&self, other: &Self) -> bool { + PartialOrd::ge(&**self, &**other) + } + #[inline] + fn gt(&self, other: &Self) -> bool { + PartialOrd::gt(&**self, &**other) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Ord, A: Allocator> Ord for Box<T, A> { + #[inline] + fn cmp(&self, other: &Self) -> Ordering { + Ord::cmp(&**self, &**other) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Eq, A: Allocator> Eq for Box<T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Hash, A: Allocator> Hash for Box<T, A> { + fn hash<H: Hasher>(&self, state: &mut H) { + (**self).hash(state); + } +} + +#[stable(feature = "indirect_hasher_impl", since = "1.22.0")] +impl<T: ?Sized + Hasher, A: Allocator> Hasher for Box<T, A> { + fn finish(&self) -> u64 { + (**self).finish() + } + fn write(&mut self, bytes: &[u8]) { + (**self).write(bytes) + } + fn write_u8(&mut self, i: u8) { + (**self).write_u8(i) + } + fn write_u16(&mut self, i: u16) { + (**self).write_u16(i) + } + fn write_u32(&mut self, i: u32) { + (**self).write_u32(i) + } + fn write_u64(&mut self, i: u64) { + (**self).write_u64(i) + } + fn write_u128(&mut self, i: u128) { + (**self).write_u128(i) + } + fn write_usize(&mut self, i: usize) { + (**self).write_usize(i) + } + fn write_i8(&mut self, i: i8) { + (**self).write_i8(i) + } + fn write_i16(&mut self, i: i16) { + (**self).write_i16(i) + } + fn write_i32(&mut self, i: i32) { + (**self).write_i32(i) + } + fn write_i64(&mut self, i: i64) { + (**self).write_i64(i) + } + fn write_i128(&mut self, i: i128) { + (**self).write_i128(i) + } + fn write_isize(&mut self, i: isize) { + (**self).write_isize(i) + } + fn write_length_prefix(&mut self, len: usize) { + (**self).write_length_prefix(len) + } + fn write_str(&mut self, s: &str) { + (**self).write_str(s) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_for_ptrs", since = "1.6.0")] +impl<T> From<T> for Box<T> { + /// Converts a `T` into a `Box<T>` + /// + /// The conversion allocates on the heap and moves `t` + /// from the stack into it. + /// + /// # Examples + /// + /// ```rust + /// let x = 5; + /// let boxed = Box::new(5); + /// + /// assert_eq!(Box::from(x), boxed); + /// ``` + fn from(t: T) -> Self { + Box::new(t) + } +} + +#[stable(feature = "pin", since = "1.33.0")] +#[rustc_const_unstable(feature = "const_box", issue = "92521")] +impl<T: ?Sized, A: Allocator> const From<Box<T, A>> for Pin<Box<T, A>> +where + A: 'static, +{ + /// Converts a `Box<T>` into a `Pin<Box<T>>`. If `T` does not implement [`Unpin`], then + /// `*boxed` will be pinned in memory and unable to be moved. + /// + /// This conversion does not allocate on the heap and happens in place. + /// + /// This is also available via [`Box::into_pin`]. + /// + /// Constructing and pinning a `Box` with <code><Pin<Box\<T>>>::from([Box::new]\(x))</code> + /// can also be written more concisely using <code>[Box::pin]\(x)</code>. + /// This `From` implementation is useful if you already have a `Box<T>`, or you are + /// constructing a (pinned) `Box` in a different way than with [`Box::new`]. + fn from(boxed: Box<T, A>) -> Self { + Box::into_pin(boxed) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_slice", since = "1.17.0")] +impl<T: Copy> From<&[T]> for Box<[T]> { + /// Converts a `&[T]` into a `Box<[T]>` + /// + /// This conversion allocates on the heap + /// and performs a copy of `slice`. + /// + /// # Examples + /// ```rust + /// // create a &[u8] which will be used to create a Box<[u8]> + /// let slice: &[u8] = &[104, 101, 108, 108, 111]; + /// let boxed_slice: Box<[u8]> = Box::from(slice); + /// + /// println!("{boxed_slice:?}"); + /// ``` + fn from(slice: &[T]) -> Box<[T]> { + let len = slice.len(); + let buf = RawVec::with_capacity(len); + unsafe { + ptr::copy_nonoverlapping(slice.as_ptr(), buf.ptr(), len); + buf.into_box(slice.len()).assume_init() + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_cow", since = "1.45.0")] +impl<T: Copy> From<Cow<'_, [T]>> for Box<[T]> { + /// Converts a `Cow<'_, [T]>` into a `Box<[T]>` + /// + /// When `cow` is the `Cow::Borrowed` variant, this + /// conversion allocates on the heap and copies the + /// underlying slice. Otherwise, it will try to reuse the owned + /// `Vec`'s allocation. + #[inline] + fn from(cow: Cow<'_, [T]>) -> Box<[T]> { + match cow { + Cow::Borrowed(slice) => Box::from(slice), + Cow::Owned(slice) => Box::from(slice), + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_slice", since = "1.17.0")] +impl From<&str> for Box<str> { + /// Converts a `&str` into a `Box<str>` + /// + /// This conversion allocates on the heap + /// and performs a copy of `s`. + /// + /// # Examples + /// + /// ```rust + /// let boxed: Box<str> = Box::from("hello"); + /// println!("{boxed}"); + /// ``` + #[inline] + fn from(s: &str) -> Box<str> { + unsafe { from_boxed_utf8_unchecked(Box::from(s.as_bytes())) } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_cow", since = "1.45.0")] +impl From<Cow<'_, str>> for Box<str> { + /// Converts a `Cow<'_, str>` into a `Box<str>` + /// + /// When `cow` is the `Cow::Borrowed` variant, this + /// conversion allocates on the heap and copies the + /// underlying `str`. Otherwise, it will try to reuse the owned + /// `String`'s allocation. + /// + /// # Examples + /// + /// ```rust + /// use std::borrow::Cow; + /// + /// let unboxed = Cow::Borrowed("hello"); + /// let boxed: Box<str> = Box::from(unboxed); + /// println!("{boxed}"); + /// ``` + /// + /// ```rust + /// # use std::borrow::Cow; + /// let unboxed = Cow::Owned("hello".to_string()); + /// let boxed: Box<str> = Box::from(unboxed); + /// println!("{boxed}"); + /// ``` + #[inline] + fn from(cow: Cow<'_, str>) -> Box<str> { + match cow { + Cow::Borrowed(s) => Box::from(s), + Cow::Owned(s) => Box::from(s), + } + } +} + +#[stable(feature = "boxed_str_conv", since = "1.19.0")] +impl<A: Allocator> From<Box<str, A>> for Box<[u8], A> { + /// Converts a `Box<str>` into a `Box<[u8]>` + /// + /// This conversion does not allocate on the heap and happens in place. + /// + /// # Examples + /// ```rust + /// // create a Box<str> which will be used to create a Box<[u8]> + /// let boxed: Box<str> = Box::from("hello"); + /// let boxed_str: Box<[u8]> = Box::from(boxed); + /// + /// // create a &[u8] which will be used to create a Box<[u8]> + /// let slice: &[u8] = &[104, 101, 108, 108, 111]; + /// let boxed_slice = Box::from(slice); + /// + /// assert_eq!(boxed_slice, boxed_str); + /// ``` + #[inline] + fn from(s: Box<str, A>) -> Self { + let (raw, alloc) = Box::into_raw_with_allocator(s); + unsafe { Box::from_raw_in(raw as *mut [u8], alloc) } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_array", since = "1.45.0")] +impl<T, const N: usize> From<[T; N]> for Box<[T]> { + /// Converts a `[T; N]` into a `Box<[T]>` + /// + /// This conversion moves the array to newly heap-allocated memory. + /// + /// # Examples + /// + /// ```rust + /// let boxed: Box<[u8]> = Box::from([4, 2]); + /// println!("{boxed:?}"); + /// ``` + fn from(array: [T; N]) -> Box<[T]> { + #[rustc_box] + Box::new(array) + } +} + +#[stable(feature = "boxed_slice_try_from", since = "1.43.0")] +impl<T, const N: usize> TryFrom<Box<[T]>> for Box<[T; N]> { + type Error = Box<[T]>; + + /// Attempts to convert a `Box<[T]>` into a `Box<[T; N]>`. + /// + /// The conversion occurs in-place and does not require a + /// new memory allocation. + /// + /// # Errors + /// + /// Returns the old `Box<[T]>` in the `Err` variant if + /// `boxed_slice.len()` does not equal `N`. + fn try_from(boxed_slice: Box<[T]>) -> Result<Self, Self::Error> { + if boxed_slice.len() == N { + Ok(unsafe { Box::from_raw(Box::into_raw(boxed_slice) as *mut [T; N]) }) + } else { + Err(boxed_slice) + } + } +} + +impl<A: Allocator> Box<dyn Any, A> { + /// Attempt to downcast the box to a concrete type. + /// + /// # Examples + /// + /// ``` + /// use std::any::Any; + /// + /// fn print_if_string(value: Box<dyn Any>) { + /// if let Ok(string) = value.downcast::<String>() { + /// println!("String ({}): {}", string.len(), string); + /// } + /// } + /// + /// let my_string = "Hello World".to_string(); + /// print_if_string(Box::new(my_string)); + /// print_if_string(Box::new(0i8)); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn downcast<T: Any>(self) -> Result<Box<T, A>, Self> { + if self.is::<T>() { unsafe { Ok(self.downcast_unchecked::<T>()) } } else { Err(self) } + } + + /// Downcasts the box to a concrete type. + /// + /// For a safe alternative see [`downcast`]. + /// + /// # Examples + /// + /// ``` + /// #![feature(downcast_unchecked)] + /// + /// use std::any::Any; + /// + /// let x: Box<dyn Any> = Box::new(1_usize); + /// + /// unsafe { + /// assert_eq!(*x.downcast_unchecked::<usize>(), 1); + /// } + /// ``` + /// + /// # Safety + /// + /// The contained value must be of type `T`. Calling this method + /// with the incorrect type is *undefined behavior*. + /// + /// [`downcast`]: Self::downcast + #[inline] + #[unstable(feature = "downcast_unchecked", issue = "90850")] + pub unsafe fn downcast_unchecked<T: Any>(self) -> Box<T, A> { + debug_assert!(self.is::<T>()); + unsafe { + let (raw, alloc): (*mut dyn Any, _) = Box::into_raw_with_allocator(self); + Box::from_raw_in(raw as *mut T, alloc) + } + } +} + +impl<A: Allocator> Box<dyn Any + Send, A> { + /// Attempt to downcast the box to a concrete type. + /// + /// # Examples + /// + /// ``` + /// use std::any::Any; + /// + /// fn print_if_string(value: Box<dyn Any + Send>) { + /// if let Ok(string) = value.downcast::<String>() { + /// println!("String ({}): {}", string.len(), string); + /// } + /// } + /// + /// let my_string = "Hello World".to_string(); + /// print_if_string(Box::new(my_string)); + /// print_if_string(Box::new(0i8)); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn downcast<T: Any>(self) -> Result<Box<T, A>, Self> { + if self.is::<T>() { unsafe { Ok(self.downcast_unchecked::<T>()) } } else { Err(self) } + } + + /// Downcasts the box to a concrete type. + /// + /// For a safe alternative see [`downcast`]. + /// + /// # Examples + /// + /// ``` + /// #![feature(downcast_unchecked)] + /// + /// use std::any::Any; + /// + /// let x: Box<dyn Any + Send> = Box::new(1_usize); + /// + /// unsafe { + /// assert_eq!(*x.downcast_unchecked::<usize>(), 1); + /// } + /// ``` + /// + /// # Safety + /// + /// The contained value must be of type `T`. Calling this method + /// with the incorrect type is *undefined behavior*. + /// + /// [`downcast`]: Self::downcast + #[inline] + #[unstable(feature = "downcast_unchecked", issue = "90850")] + pub unsafe fn downcast_unchecked<T: Any>(self) -> Box<T, A> { + debug_assert!(self.is::<T>()); + unsafe { + let (raw, alloc): (*mut (dyn Any + Send), _) = Box::into_raw_with_allocator(self); + Box::from_raw_in(raw as *mut T, alloc) + } + } +} + +impl<A: Allocator> Box<dyn Any + Send + Sync, A> { + /// Attempt to downcast the box to a concrete type. + /// + /// # Examples + /// + /// ``` + /// use std::any::Any; + /// + /// fn print_if_string(value: Box<dyn Any + Send + Sync>) { + /// if let Ok(string) = value.downcast::<String>() { + /// println!("String ({}): {}", string.len(), string); + /// } + /// } + /// + /// let my_string = "Hello World".to_string(); + /// print_if_string(Box::new(my_string)); + /// print_if_string(Box::new(0i8)); + /// ``` + #[inline] + #[stable(feature = "box_send_sync_any_downcast", since = "1.51.0")] + pub fn downcast<T: Any>(self) -> Result<Box<T, A>, Self> { + if self.is::<T>() { unsafe { Ok(self.downcast_unchecked::<T>()) } } else { Err(self) } + } + + /// Downcasts the box to a concrete type. + /// + /// For a safe alternative see [`downcast`]. + /// + /// # Examples + /// + /// ``` + /// #![feature(downcast_unchecked)] + /// + /// use std::any::Any; + /// + /// let x: Box<dyn Any + Send + Sync> = Box::new(1_usize); + /// + /// unsafe { + /// assert_eq!(*x.downcast_unchecked::<usize>(), 1); + /// } + /// ``` + /// + /// # Safety + /// + /// The contained value must be of type `T`. Calling this method + /// with the incorrect type is *undefined behavior*. + /// + /// [`downcast`]: Self::downcast + #[inline] + #[unstable(feature = "downcast_unchecked", issue = "90850")] + pub unsafe fn downcast_unchecked<T: Any>(self) -> Box<T, A> { + debug_assert!(self.is::<T>()); + unsafe { + let (raw, alloc): (*mut (dyn Any + Send + Sync), _) = + Box::into_raw_with_allocator(self); + Box::from_raw_in(raw as *mut T, alloc) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Display + ?Sized, A: Allocator> fmt::Display for Box<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Debug + ?Sized, A: Allocator> fmt::Debug for Box<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized, A: Allocator> fmt::Pointer for Box<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + // It's not possible to extract the inner Uniq directly from the Box, + // instead we cast it to a *const which aliases the Unique + let ptr: *const T = &**self; + fmt::Pointer::fmt(&ptr, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_box", issue = "92521")] +impl<T: ?Sized, A: Allocator> const Deref for Box<T, A> { + type Target = T; + + fn deref(&self) -> &T { + &**self + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_box", issue = "92521")] +impl<T: ?Sized, A: Allocator> const DerefMut for Box<T, A> { + fn deref_mut(&mut self) -> &mut T { + &mut **self + } +} + +#[unstable(feature = "receiver_trait", issue = "none")] +impl<T: ?Sized, A: Allocator> Receiver for Box<T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<I: Iterator + ?Sized, A: Allocator> Iterator for Box<I, A> { + type Item = I::Item; + fn next(&mut self) -> Option<I::Item> { + (**self).next() + } + fn size_hint(&self) -> (usize, Option<usize>) { + (**self).size_hint() + } + fn nth(&mut self, n: usize) -> Option<I::Item> { + (**self).nth(n) + } + fn last(self) -> Option<I::Item> { + BoxIter::last(self) + } +} + +trait BoxIter { + type Item; + fn last(self) -> Option<Self::Item>; +} + +impl<I: Iterator + ?Sized, A: Allocator> BoxIter for Box<I, A> { + type Item = I::Item; + default fn last(self) -> Option<I::Item> { + #[inline] + fn some<T>(_: Option<T>, x: T) -> Option<T> { + Some(x) + } + + self.fold(None, some) + } +} + +/// Specialization for sized `I`s that uses `I`s implementation of `last()` +/// instead of the default. +#[stable(feature = "rust1", since = "1.0.0")] +impl<I: Iterator, A: Allocator> BoxIter for Box<I, A> { + fn last(self) -> Option<I::Item> { + (*self).last() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<I: DoubleEndedIterator + ?Sized, A: Allocator> DoubleEndedIterator for Box<I, A> { + fn next_back(&mut self) -> Option<I::Item> { + (**self).next_back() + } + fn nth_back(&mut self, n: usize) -> Option<I::Item> { + (**self).nth_back(n) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<I: ExactSizeIterator + ?Sized, A: Allocator> ExactSizeIterator for Box<I, A> { + fn len(&self) -> usize { + (**self).len() + } + fn is_empty(&self) -> bool { + (**self).is_empty() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<I: FusedIterator + ?Sized, A: Allocator> FusedIterator for Box<I, A> {} + +#[stable(feature = "boxed_closure_impls", since = "1.35.0")] +impl<Args, F: FnOnce<Args> + ?Sized, A: Allocator> FnOnce<Args> for Box<F, A> { + type Output = <F as FnOnce<Args>>::Output; + + extern "rust-call" fn call_once(self, args: Args) -> Self::Output { + <F as FnOnce<Args>>::call_once(*self, args) + } +} + +#[stable(feature = "boxed_closure_impls", since = "1.35.0")] +impl<Args, F: FnMut<Args> + ?Sized, A: Allocator> FnMut<Args> for Box<F, A> { + extern "rust-call" fn call_mut(&mut self, args: Args) -> Self::Output { + <F as FnMut<Args>>::call_mut(self, args) + } +} + +#[stable(feature = "boxed_closure_impls", since = "1.35.0")] +impl<Args, F: Fn<Args> + ?Sized, A: Allocator> Fn<Args> for Box<F, A> { + extern "rust-call" fn call(&self, args: Args) -> Self::Output { + <F as Fn<Args>>::call(self, args) + } +} + +#[unstable(feature = "coerce_unsized", issue = "27732")] +impl<T: ?Sized + Unsize<U>, U: ?Sized, A: Allocator> CoerceUnsized<Box<U, A>> for Box<T, A> {} + +#[unstable(feature = "dispatch_from_dyn", issue = "none")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> DispatchFromDyn<Box<U>> for Box<T, Global> {} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "boxed_slice_from_iter", since = "1.32.0")] +impl<I> FromIterator<I> for Box<[I]> { + fn from_iter<T: IntoIterator<Item = I>>(iter: T) -> Self { + iter.into_iter().collect::<Vec<_>>().into_boxed_slice() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_slice_clone", since = "1.3.0")] +impl<T: Clone, A: Allocator + Clone> Clone for Box<[T], A> { + fn clone(&self) -> Self { + let alloc = Box::allocator(self).clone(); + self.to_vec_in(alloc).into_boxed_slice() + } + + fn clone_from(&mut self, other: &Self) { + if self.len() == other.len() { + self.clone_from_slice(&other); + } else { + *self = other.clone(); + } + } +} + +#[stable(feature = "box_borrow", since = "1.1.0")] +impl<T: ?Sized, A: Allocator> borrow::Borrow<T> for Box<T, A> { + fn borrow(&self) -> &T { + &**self + } +} + +#[stable(feature = "box_borrow", since = "1.1.0")] +impl<T: ?Sized, A: Allocator> borrow::BorrowMut<T> for Box<T, A> { + fn borrow_mut(&mut self) -> &mut T { + &mut **self + } +} + +#[stable(since = "1.5.0", feature = "smart_ptr_as_ref")] +impl<T: ?Sized, A: Allocator> AsRef<T> for Box<T, A> { + fn as_ref(&self) -> &T { + &**self + } +} + +#[stable(since = "1.5.0", feature = "smart_ptr_as_ref")] +impl<T: ?Sized, A: Allocator> AsMut<T> for Box<T, A> { + fn as_mut(&mut self) -> &mut T { + &mut **self + } +} + +/* Nota bene + * + * We could have chosen not to add this impl, and instead have written a + * function of Pin<Box<T>> to Pin<T>. Such a function would not be sound, + * because Box<T> implements Unpin even when T does not, as a result of + * this impl. + * + * We chose this API instead of the alternative for a few reasons: + * - Logically, it is helpful to understand pinning in regard to the + * memory region being pointed to. For this reason none of the + * standard library pointer types support projecting through a pin + * (Box<T> is the only pointer type in std for which this would be + * safe.) + * - It is in practice very useful to have Box<T> be unconditionally + * Unpin because of trait objects, for which the structural auto + * trait functionality does not apply (e.g., Box<dyn Foo> would + * otherwise not be Unpin). + * + * Another type with the same semantics as Box but only a conditional + * implementation of `Unpin` (where `T: Unpin`) would be valid/safe, and + * could have a method to project a Pin<T> from it. + */ +#[stable(feature = "pin", since = "1.33.0")] +#[rustc_const_unstable(feature = "const_box", issue = "92521")] +impl<T: ?Sized, A: Allocator> const Unpin for Box<T, A> where A: 'static {} + +#[unstable(feature = "generator_trait", issue = "43122")] +impl<G: ?Sized + Generator<R> + Unpin, R, A: Allocator> Generator<R> for Box<G, A> +where + A: 'static, +{ + type Yield = G::Yield; + type Return = G::Return; + + fn resume(mut self: Pin<&mut Self>, arg: R) -> GeneratorState<Self::Yield, Self::Return> { + G::resume(Pin::new(&mut *self), arg) + } +} + +#[unstable(feature = "generator_trait", issue = "43122")] +impl<G: ?Sized + Generator<R>, R, A: Allocator> Generator<R> for Pin<Box<G, A>> +where + A: 'static, +{ + type Yield = G::Yield; + type Return = G::Return; + + fn resume(mut self: Pin<&mut Self>, arg: R) -> GeneratorState<Self::Yield, Self::Return> { + G::resume((*self).as_mut(), arg) + } +} + +#[stable(feature = "futures_api", since = "1.36.0")] +impl<F: ?Sized + Future + Unpin, A: Allocator> Future for Box<F, A> +where + A: 'static, +{ + type Output = F::Output; + + fn poll(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> { + F::poll(Pin::new(&mut *self), cx) + } +} + +#[unstable(feature = "async_iterator", issue = "79024")] +impl<S: ?Sized + AsyncIterator + Unpin> AsyncIterator for Box<S> { + type Item = S::Item; + + fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> { + Pin::new(&mut **self).poll_next(cx) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (**self).size_hint() + } +} diff --git a/library/alloc/src/boxed/thin.rs b/library/alloc/src/boxed/thin.rs new file mode 100644 index 000000000..649ccfcaa --- /dev/null +++ b/library/alloc/src/boxed/thin.rs @@ -0,0 +1,273 @@ +// Based on +// https://github.com/matthieu-m/rfc2580/blob/b58d1d3cba0d4b5e859d3617ea2d0943aaa31329/examples/thin.rs +// by matthieu-m +use crate::alloc::{self, Layout, LayoutError}; +use core::fmt::{self, Debug, Display, Formatter}; +use core::marker::PhantomData; +#[cfg(not(no_global_oom_handling))] +use core::marker::Unsize; +use core::mem; +use core::ops::{Deref, DerefMut}; +use core::ptr::Pointee; +use core::ptr::{self, NonNull}; + +/// ThinBox. +/// +/// A thin pointer for heap allocation, regardless of T. +/// +/// # Examples +/// +/// ``` +/// #![feature(thin_box)] +/// use std::boxed::ThinBox; +/// +/// let five = ThinBox::new(5); +/// let thin_slice = ThinBox::<[i32]>::new_unsize([1, 2, 3, 4]); +/// +/// use std::mem::{size_of, size_of_val}; +/// let size_of_ptr = size_of::<*const ()>(); +/// assert_eq!(size_of_ptr, size_of_val(&five)); +/// assert_eq!(size_of_ptr, size_of_val(&thin_slice)); +/// ``` +#[unstable(feature = "thin_box", issue = "92791")] +pub struct ThinBox<T: ?Sized> { + // This is essentially `WithHeader<<T as Pointee>::Metadata>`, + // but that would be invariant in `T`, and we want covariance. + ptr: WithOpaqueHeader, + _marker: PhantomData<T>, +} + +/// `ThinBox<T>` is `Send` if `T` is `Send` because the data is owned. +#[unstable(feature = "thin_box", issue = "92791")] +unsafe impl<T: ?Sized + Send> Send for ThinBox<T> {} + +/// `ThinBox<T>` is `Sync` if `T` is `Sync` because the data is owned. +#[unstable(feature = "thin_box", issue = "92791")] +unsafe impl<T: ?Sized + Sync> Sync for ThinBox<T> {} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T> ThinBox<T> { + /// Moves a type to the heap with its `Metadata` stored in the heap allocation instead of on + /// the stack. + /// + /// # Examples + /// + /// ``` + /// #![feature(thin_box)] + /// use std::boxed::ThinBox; + /// + /// let five = ThinBox::new(5); + /// ``` + #[cfg(not(no_global_oom_handling))] + pub fn new(value: T) -> Self { + let meta = ptr::metadata(&value); + let ptr = WithOpaqueHeader::new(meta, value); + ThinBox { ptr, _marker: PhantomData } + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<Dyn: ?Sized> ThinBox<Dyn> { + /// Moves a type to the heap with its `Metadata` stored in the heap allocation instead of on + /// the stack. + /// + /// # Examples + /// + /// ``` + /// #![feature(thin_box)] + /// use std::boxed::ThinBox; + /// + /// let thin_slice = ThinBox::<[i32]>::new_unsize([1, 2, 3, 4]); + /// ``` + #[cfg(not(no_global_oom_handling))] + pub fn new_unsize<T>(value: T) -> Self + where + T: Unsize<Dyn>, + { + let meta = ptr::metadata(&value as &Dyn); + let ptr = WithOpaqueHeader::new(meta, value); + ThinBox { ptr, _marker: PhantomData } + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized + Debug> Debug for ThinBox<T> { + fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result { + Debug::fmt(self.deref(), f) + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized + Display> Display for ThinBox<T> { + fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result { + Display::fmt(self.deref(), f) + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized> Deref for ThinBox<T> { + type Target = T; + + fn deref(&self) -> &T { + let value = self.data(); + let metadata = self.meta(); + let pointer = ptr::from_raw_parts(value as *const (), metadata); + unsafe { &*pointer } + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized> DerefMut for ThinBox<T> { + fn deref_mut(&mut self) -> &mut T { + let value = self.data(); + let metadata = self.meta(); + let pointer = ptr::from_raw_parts_mut::<T>(value as *mut (), metadata); + unsafe { &mut *pointer } + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized> Drop for ThinBox<T> { + fn drop(&mut self) { + unsafe { + let value = self.deref_mut(); + let value = value as *mut T; + self.with_header().drop::<T>(value); + } + } +} + +#[unstable(feature = "thin_box", issue = "92791")] +impl<T: ?Sized> ThinBox<T> { + fn meta(&self) -> <T as Pointee>::Metadata { + // Safety: + // - NonNull and valid. + unsafe { *self.with_header().header() } + } + + fn data(&self) -> *mut u8 { + self.with_header().value() + } + + fn with_header(&self) -> &WithHeader<<T as Pointee>::Metadata> { + // SAFETY: both types are transparent to `NonNull<u8>` + unsafe { &*((&self.ptr) as *const WithOpaqueHeader as *const WithHeader<_>) } + } +} + +/// A pointer to type-erased data, guaranteed to either be: +/// 1. `NonNull::dangling()`, in the case where both the pointee (`T`) and +/// metadata (`H`) are ZSTs. +/// 2. A pointer to a valid `T` that has a header `H` directly before the +/// pointed-to location. +#[repr(transparent)] +struct WithHeader<H>(NonNull<u8>, PhantomData<H>); + +/// An opaque representation of `WithHeader<H>` to avoid the +/// projection invariance of `<T as Pointee>::Metadata`. +#[repr(transparent)] +struct WithOpaqueHeader(NonNull<u8>); + +impl WithOpaqueHeader { + #[cfg(not(no_global_oom_handling))] + fn new<H, T>(header: H, value: T) -> Self { + let ptr = WithHeader::new(header, value); + Self(ptr.0) + } +} + +impl<H> WithHeader<H> { + #[cfg(not(no_global_oom_handling))] + fn new<T>(header: H, value: T) -> WithHeader<H> { + let value_layout = Layout::new::<T>(); + let Ok((layout, value_offset)) = Self::alloc_layout(value_layout) else { + // We pass an empty layout here because we do not know which layout caused the + // arithmetic overflow in `Layout::extend` and `handle_alloc_error` takes `Layout` as + // its argument rather than `Result<Layout, LayoutError>`, also this function has been + // stable since 1.28 ._. + // + // On the other hand, look at this gorgeous turbofish! + alloc::handle_alloc_error(Layout::new::<()>()); + }; + + unsafe { + // Note: It's UB to pass a layout with a zero size to `alloc::alloc`, so + // we use `layout.dangling()` for this case, which should have a valid + // alignment for both `T` and `H`. + let ptr = if layout.size() == 0 { + // Some paranoia checking, mostly so that the ThinBox tests are + // more able to catch issues. + debug_assert!( + value_offset == 0 && mem::size_of::<T>() == 0 && mem::size_of::<H>() == 0 + ); + layout.dangling() + } else { + let ptr = alloc::alloc(layout); + if ptr.is_null() { + alloc::handle_alloc_error(layout); + } + // Safety: + // - The size is at least `aligned_header_size`. + let ptr = ptr.add(value_offset) as *mut _; + + NonNull::new_unchecked(ptr) + }; + + let result = WithHeader(ptr, PhantomData); + ptr::write(result.header(), header); + ptr::write(result.value().cast(), value); + + result + } + } + + // Safety: + // - Assumes that either `value` can be dereferenced, or is the + // `NonNull::dangling()` we use when both `T` and `H` are ZSTs. + unsafe fn drop<T: ?Sized>(&self, value: *mut T) { + unsafe { + let value_layout = Layout::for_value_raw(value); + // SAFETY: Layout must have been computable if we're in drop + let (layout, value_offset) = Self::alloc_layout(value_layout).unwrap_unchecked(); + + // We only drop the value because the Pointee trait requires that the metadata is copy + // aka trivially droppable. + ptr::drop_in_place::<T>(value); + + // Note: Don't deallocate if the layout size is zero, because the pointer + // didn't come from the allocator. + if layout.size() != 0 { + alloc::dealloc(self.0.as_ptr().sub(value_offset), layout); + } else { + debug_assert!( + value_offset == 0 && mem::size_of::<H>() == 0 && value_layout.size() == 0 + ); + } + } + } + + fn header(&self) -> *mut H { + // Safety: + // - At least `size_of::<H>()` bytes are allocated ahead of the pointer. + // - We know that H will be aligned because the middle pointer is aligned to the greater + // of the alignment of the header and the data and the header size includes the padding + // needed to align the header. Subtracting the header size from the aligned data pointer + // will always result in an aligned header pointer, it just may not point to the + // beginning of the allocation. + let hp = unsafe { self.0.as_ptr().sub(Self::header_size()) as *mut H }; + debug_assert!(hp.is_aligned()); + hp + } + + fn value(&self) -> *mut u8 { + self.0.as_ptr() + } + + const fn header_size() -> usize { + mem::size_of::<H>() + } + + fn alloc_layout(value_layout: Layout) -> Result<(Layout, usize), LayoutError> { + Layout::new::<H>().extend(value_layout) + } +} diff --git a/library/alloc/src/collections/binary_heap.rs b/library/alloc/src/collections/binary_heap.rs new file mode 100644 index 000000000..197e7aaac --- /dev/null +++ b/library/alloc/src/collections/binary_heap.rs @@ -0,0 +1,1720 @@ +//! A priority queue implemented with a binary heap. +//! +//! Insertion and popping the largest element have *O*(log(*n*)) time complexity. +//! Checking the largest element is *O*(1). Converting a vector to a binary heap +//! can be done in-place, and has *O*(*n*) complexity. A binary heap can also be +//! converted to a sorted vector in-place, allowing it to be used for an *O*(*n* * log(*n*)) +//! in-place heapsort. +//! +//! # Examples +//! +//! This is a larger example that implements [Dijkstra's algorithm][dijkstra] +//! to solve the [shortest path problem][sssp] on a [directed graph][dir_graph]. +//! It shows how to use [`BinaryHeap`] with custom types. +//! +//! [dijkstra]: https://en.wikipedia.org/wiki/Dijkstra%27s_algorithm +//! [sssp]: https://en.wikipedia.org/wiki/Shortest_path_problem +//! [dir_graph]: https://en.wikipedia.org/wiki/Directed_graph +//! +//! ``` +//! use std::cmp::Ordering; +//! use std::collections::BinaryHeap; +//! +//! #[derive(Copy, Clone, Eq, PartialEq)] +//! struct State { +//! cost: usize, +//! position: usize, +//! } +//! +//! // The priority queue depends on `Ord`. +//! // Explicitly implement the trait so the queue becomes a min-heap +//! // instead of a max-heap. +//! impl Ord for State { +//! fn cmp(&self, other: &Self) -> Ordering { +//! // Notice that the we flip the ordering on costs. +//! // In case of a tie we compare positions - this step is necessary +//! // to make implementations of `PartialEq` and `Ord` consistent. +//! other.cost.cmp(&self.cost) +//! .then_with(|| self.position.cmp(&other.position)) +//! } +//! } +//! +//! // `PartialOrd` needs to be implemented as well. +//! impl PartialOrd for State { +//! fn partial_cmp(&self, other: &Self) -> Option<Ordering> { +//! Some(self.cmp(other)) +//! } +//! } +//! +//! // Each node is represented as a `usize`, for a shorter implementation. +//! struct Edge { +//! node: usize, +//! cost: usize, +//! } +//! +//! // Dijkstra's shortest path algorithm. +//! +//! // Start at `start` and use `dist` to track the current shortest distance +//! // to each node. This implementation isn't memory-efficient as it may leave duplicate +//! // nodes in the queue. It also uses `usize::MAX` as a sentinel value, +//! // for a simpler implementation. +//! fn shortest_path(adj_list: &Vec<Vec<Edge>>, start: usize, goal: usize) -> Option<usize> { +//! // dist[node] = current shortest distance from `start` to `node` +//! let mut dist: Vec<_> = (0..adj_list.len()).map(|_| usize::MAX).collect(); +//! +//! let mut heap = BinaryHeap::new(); +//! +//! // We're at `start`, with a zero cost +//! dist[start] = 0; +//! heap.push(State { cost: 0, position: start }); +//! +//! // Examine the frontier with lower cost nodes first (min-heap) +//! while let Some(State { cost, position }) = heap.pop() { +//! // Alternatively we could have continued to find all shortest paths +//! if position == goal { return Some(cost); } +//! +//! // Important as we may have already found a better way +//! if cost > dist[position] { continue; } +//! +//! // For each node we can reach, see if we can find a way with +//! // a lower cost going through this node +//! for edge in &adj_list[position] { +//! let next = State { cost: cost + edge.cost, position: edge.node }; +//! +//! // If so, add it to the frontier and continue +//! if next.cost < dist[next.position] { +//! heap.push(next); +//! // Relaxation, we have now found a better way +//! dist[next.position] = next.cost; +//! } +//! } +//! } +//! +//! // Goal not reachable +//! None +//! } +//! +//! fn main() { +//! // This is the directed graph we're going to use. +//! // The node numbers correspond to the different states, +//! // and the edge weights symbolize the cost of moving +//! // from one node to another. +//! // Note that the edges are one-way. +//! // +//! // 7 +//! // +-----------------+ +//! // | | +//! // v 1 2 | 2 +//! // 0 -----> 1 -----> 3 ---> 4 +//! // | ^ ^ ^ +//! // | | 1 | | +//! // | | | 3 | 1 +//! // +------> 2 -------+ | +//! // 10 | | +//! // +---------------+ +//! // +//! // The graph is represented as an adjacency list where each index, +//! // corresponding to a node value, has a list of outgoing edges. +//! // Chosen for its efficiency. +//! let graph = vec![ +//! // Node 0 +//! vec![Edge { node: 2, cost: 10 }, +//! Edge { node: 1, cost: 1 }], +//! // Node 1 +//! vec![Edge { node: 3, cost: 2 }], +//! // Node 2 +//! vec![Edge { node: 1, cost: 1 }, +//! Edge { node: 3, cost: 3 }, +//! Edge { node: 4, cost: 1 }], +//! // Node 3 +//! vec![Edge { node: 0, cost: 7 }, +//! Edge { node: 4, cost: 2 }], +//! // Node 4 +//! vec![]]; +//! +//! assert_eq!(shortest_path(&graph, 0, 1), Some(1)); +//! assert_eq!(shortest_path(&graph, 0, 3), Some(3)); +//! assert_eq!(shortest_path(&graph, 3, 0), Some(7)); +//! assert_eq!(shortest_path(&graph, 0, 4), Some(5)); +//! assert_eq!(shortest_path(&graph, 4, 0), None); +//! } +//! ``` + +#![allow(missing_docs)] +#![stable(feature = "rust1", since = "1.0.0")] + +use core::fmt; +use core::iter::{FromIterator, FusedIterator, InPlaceIterable, SourceIter, TrustedLen}; +use core::mem::{self, swap, ManuallyDrop}; +use core::ops::{Deref, DerefMut}; +use core::ptr; + +use crate::collections::TryReserveError; +use crate::slice; +use crate::vec::{self, AsVecIntoIter, Vec}; + +use super::SpecExtend; + +#[cfg(test)] +mod tests; + +/// A priority queue implemented with a binary heap. +/// +/// This will be a max-heap. +/// +/// It is a logic error for an item to be modified in such a way that the +/// item's ordering relative to any other item, as determined by the [`Ord`] +/// trait, changes while it is in the heap. This is normally only possible +/// through [`Cell`], [`RefCell`], global state, I/O, or unsafe code. The +/// behavior resulting from such a logic error is not specified, but will +/// be encapsulated to the `BinaryHeap` that observed the logic error and not +/// result in undefined behavior. This could include panics, incorrect results, +/// aborts, memory leaks, and non-termination. +/// +/// # Examples +/// +/// ``` +/// use std::collections::BinaryHeap; +/// +/// // Type inference lets us omit an explicit type signature (which +/// // would be `BinaryHeap<i32>` in this example). +/// let mut heap = BinaryHeap::new(); +/// +/// // We can use peek to look at the next item in the heap. In this case, +/// // there's no items in there yet so we get None. +/// assert_eq!(heap.peek(), None); +/// +/// // Let's add some scores... +/// heap.push(1); +/// heap.push(5); +/// heap.push(2); +/// +/// // Now peek shows the most important item in the heap. +/// assert_eq!(heap.peek(), Some(&5)); +/// +/// // We can check the length of a heap. +/// assert_eq!(heap.len(), 3); +/// +/// // We can iterate over the items in the heap, although they are returned in +/// // a random order. +/// for x in &heap { +/// println!("{x}"); +/// } +/// +/// // If we instead pop these scores, they should come back in order. +/// assert_eq!(heap.pop(), Some(5)); +/// assert_eq!(heap.pop(), Some(2)); +/// assert_eq!(heap.pop(), Some(1)); +/// assert_eq!(heap.pop(), None); +/// +/// // We can clear the heap of any remaining items. +/// heap.clear(); +/// +/// // The heap should now be empty. +/// assert!(heap.is_empty()) +/// ``` +/// +/// A `BinaryHeap` with a known list of items can be initialized from an array: +/// +/// ``` +/// use std::collections::BinaryHeap; +/// +/// let heap = BinaryHeap::from([1, 5, 2]); +/// ``` +/// +/// ## Min-heap +/// +/// Either [`core::cmp::Reverse`] or a custom [`Ord`] implementation can be used to +/// make `BinaryHeap` a min-heap. This makes `heap.pop()` return the smallest +/// value instead of the greatest one. +/// +/// ``` +/// use std::collections::BinaryHeap; +/// use std::cmp::Reverse; +/// +/// let mut heap = BinaryHeap::new(); +/// +/// // Wrap values in `Reverse` +/// heap.push(Reverse(1)); +/// heap.push(Reverse(5)); +/// heap.push(Reverse(2)); +/// +/// // If we pop these scores now, they should come back in the reverse order. +/// assert_eq!(heap.pop(), Some(Reverse(1))); +/// assert_eq!(heap.pop(), Some(Reverse(2))); +/// assert_eq!(heap.pop(), Some(Reverse(5))); +/// assert_eq!(heap.pop(), None); +/// ``` +/// +/// # Time complexity +/// +/// | [push] | [pop] | [peek]/[peek\_mut] | +/// |---------|---------------|--------------------| +/// | *O*(1)~ | *O*(log(*n*)) | *O*(1) | +/// +/// The value for `push` is an expected cost; the method documentation gives a +/// more detailed analysis. +/// +/// [`core::cmp::Reverse`]: core::cmp::Reverse +/// [`Ord`]: core::cmp::Ord +/// [`Cell`]: core::cell::Cell +/// [`RefCell`]: core::cell::RefCell +/// [push]: BinaryHeap::push +/// [pop]: BinaryHeap::pop +/// [peek]: BinaryHeap::peek +/// [peek\_mut]: BinaryHeap::peek_mut +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "BinaryHeap")] +pub struct BinaryHeap<T> { + data: Vec<T>, +} + +/// Structure wrapping a mutable reference to the greatest item on a +/// `BinaryHeap`. +/// +/// This `struct` is created by the [`peek_mut`] method on [`BinaryHeap`]. See +/// its documentation for more. +/// +/// [`peek_mut`]: BinaryHeap::peek_mut +#[stable(feature = "binary_heap_peek_mut", since = "1.12.0")] +pub struct PeekMut<'a, T: 'a + Ord> { + heap: &'a mut BinaryHeap<T>, + sift: bool, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: Ord + fmt::Debug> fmt::Debug for PeekMut<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("PeekMut").field(&self.heap.data[0]).finish() + } +} + +#[stable(feature = "binary_heap_peek_mut", since = "1.12.0")] +impl<T: Ord> Drop for PeekMut<'_, T> { + fn drop(&mut self) { + if self.sift { + // SAFETY: PeekMut is only instantiated for non-empty heaps. + unsafe { self.heap.sift_down(0) }; + } + } +} + +#[stable(feature = "binary_heap_peek_mut", since = "1.12.0")] +impl<T: Ord> Deref for PeekMut<'_, T> { + type Target = T; + fn deref(&self) -> &T { + debug_assert!(!self.heap.is_empty()); + // SAFE: PeekMut is only instantiated for non-empty heaps + unsafe { self.heap.data.get_unchecked(0) } + } +} + +#[stable(feature = "binary_heap_peek_mut", since = "1.12.0")] +impl<T: Ord> DerefMut for PeekMut<'_, T> { + fn deref_mut(&mut self) -> &mut T { + debug_assert!(!self.heap.is_empty()); + self.sift = true; + // SAFE: PeekMut is only instantiated for non-empty heaps + unsafe { self.heap.data.get_unchecked_mut(0) } + } +} + +impl<'a, T: Ord> PeekMut<'a, T> { + /// Removes the peeked value from the heap and returns it. + #[stable(feature = "binary_heap_peek_mut_pop", since = "1.18.0")] + pub fn pop(mut this: PeekMut<'a, T>) -> T { + let value = this.heap.pop().unwrap(); + this.sift = false; + value + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone> Clone for BinaryHeap<T> { + fn clone(&self) -> Self { + BinaryHeap { data: self.data.clone() } + } + + fn clone_from(&mut self, source: &Self) { + self.data.clone_from(&source.data); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord> Default for BinaryHeap<T> { + /// Creates an empty `BinaryHeap<T>`. + #[inline] + fn default() -> BinaryHeap<T> { + BinaryHeap::new() + } +} + +#[stable(feature = "binaryheap_debug", since = "1.4.0")] +impl<T: fmt::Debug> fmt::Debug for BinaryHeap<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.iter()).finish() + } +} + +impl<T: Ord> BinaryHeap<T> { + /// Creates an empty `BinaryHeap` as a max-heap. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// heap.push(4); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn new() -> BinaryHeap<T> { + BinaryHeap { data: vec![] } + } + + /// Creates an empty `BinaryHeap` with at least the specified capacity. + /// + /// The binary heap will be able to hold at least `capacity` elements without + /// reallocating. This method is allowed to allocate for more elements than + /// `capacity`. If `capacity` is 0, the binary heap will not allocate. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::with_capacity(10); + /// heap.push(4); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn with_capacity(capacity: usize) -> BinaryHeap<T> { + BinaryHeap { data: Vec::with_capacity(capacity) } + } + + /// Returns a mutable reference to the greatest item in the binary heap, or + /// `None` if it is empty. + /// + /// Note: If the `PeekMut` value is leaked, the heap may be in an + /// inconsistent state. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// assert!(heap.peek_mut().is_none()); + /// + /// heap.push(1); + /// heap.push(5); + /// heap.push(2); + /// { + /// let mut val = heap.peek_mut().unwrap(); + /// *val = 0; + /// } + /// assert_eq!(heap.peek(), Some(&2)); + /// ``` + /// + /// # Time complexity + /// + /// If the item is modified then the worst case time complexity is *O*(log(*n*)), + /// otherwise it's *O*(1). + #[stable(feature = "binary_heap_peek_mut", since = "1.12.0")] + pub fn peek_mut(&mut self) -> Option<PeekMut<'_, T>> { + if self.is_empty() { None } else { Some(PeekMut { heap: self, sift: false }) } + } + + /// Removes the greatest item from the binary heap and returns it, or `None` if it + /// is empty. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::from([1, 3]); + /// + /// assert_eq!(heap.pop(), Some(3)); + /// assert_eq!(heap.pop(), Some(1)); + /// assert_eq!(heap.pop(), None); + /// ``` + /// + /// # Time complexity + /// + /// The worst case cost of `pop` on a heap containing *n* elements is *O*(log(*n*)). + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop(&mut self) -> Option<T> { + self.data.pop().map(|mut item| { + if !self.is_empty() { + swap(&mut item, &mut self.data[0]); + // SAFETY: !self.is_empty() means that self.len() > 0 + unsafe { self.sift_down_to_bottom(0) }; + } + item + }) + } + + /// Pushes an item onto the binary heap. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// heap.push(3); + /// heap.push(5); + /// heap.push(1); + /// + /// assert_eq!(heap.len(), 3); + /// assert_eq!(heap.peek(), Some(&5)); + /// ``` + /// + /// # Time complexity + /// + /// The expected cost of `push`, averaged over every possible ordering of + /// the elements being pushed, and over a sufficiently large number of + /// pushes, is *O*(1). This is the most meaningful cost metric when pushing + /// elements that are *not* already in any sorted pattern. + /// + /// The time complexity degrades if elements are pushed in predominantly + /// ascending order. In the worst case, elements are pushed in ascending + /// sorted order and the amortized cost per push is *O*(log(*n*)) against a heap + /// containing *n* elements. + /// + /// The worst case cost of a *single* call to `push` is *O*(*n*). The worst case + /// occurs when capacity is exhausted and needs a resize. The resize cost + /// has been amortized in the previous figures. + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push(&mut self, item: T) { + let old_len = self.len(); + self.data.push(item); + // SAFETY: Since we pushed a new item it means that + // old_len = self.len() - 1 < self.len() + unsafe { self.sift_up(0, old_len) }; + } + + /// Consumes the `BinaryHeap` and returns a vector in sorted + /// (ascending) order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// + /// let mut heap = BinaryHeap::from([1, 2, 4, 5, 7]); + /// heap.push(6); + /// heap.push(3); + /// + /// let vec = heap.into_sorted_vec(); + /// assert_eq!(vec, [1, 2, 3, 4, 5, 6, 7]); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "binary_heap_extras_15", since = "1.5.0")] + pub fn into_sorted_vec(mut self) -> Vec<T> { + let mut end = self.len(); + while end > 1 { + end -= 1; + // SAFETY: `end` goes from `self.len() - 1` to 1 (both included), + // so it's always a valid index to access. + // It is safe to access index 0 (i.e. `ptr`), because + // 1 <= end < self.len(), which means self.len() >= 2. + unsafe { + let ptr = self.data.as_mut_ptr(); + ptr::swap(ptr, ptr.add(end)); + } + // SAFETY: `end` goes from `self.len() - 1` to 1 (both included) so: + // 0 < 1 <= end <= self.len() - 1 < self.len() + // Which means 0 < end and end < self.len(). + unsafe { self.sift_down_range(0, end) }; + } + self.into_vec() + } + + // The implementations of sift_up and sift_down use unsafe blocks in + // order to move an element out of the vector (leaving behind a + // hole), shift along the others and move the removed element back into the + // vector at the final location of the hole. + // The `Hole` type is used to represent this, and make sure + // the hole is filled back at the end of its scope, even on panic. + // Using a hole reduces the constant factor compared to using swaps, + // which involves twice as many moves. + + /// # Safety + /// + /// The caller must guarantee that `pos < self.len()`. + unsafe fn sift_up(&mut self, start: usize, pos: usize) -> usize { + // Take out the value at `pos` and create a hole. + // SAFETY: The caller guarantees that pos < self.len() + let mut hole = unsafe { Hole::new(&mut self.data, pos) }; + + while hole.pos() > start { + let parent = (hole.pos() - 1) / 2; + + // SAFETY: hole.pos() > start >= 0, which means hole.pos() > 0 + // and so hole.pos() - 1 can't underflow. + // This guarantees that parent < hole.pos() so + // it's a valid index and also != hole.pos(). + if hole.element() <= unsafe { hole.get(parent) } { + break; + } + + // SAFETY: Same as above + unsafe { hole.move_to(parent) }; + } + + hole.pos() + } + + /// Take an element at `pos` and move it down the heap, + /// while its children are larger. + /// + /// # Safety + /// + /// The caller must guarantee that `pos < end <= self.len()`. + unsafe fn sift_down_range(&mut self, pos: usize, end: usize) { + // SAFETY: The caller guarantees that pos < end <= self.len(). + let mut hole = unsafe { Hole::new(&mut self.data, pos) }; + let mut child = 2 * hole.pos() + 1; + + // Loop invariant: child == 2 * hole.pos() + 1. + while child <= end.saturating_sub(2) { + // compare with the greater of the two children + // SAFETY: child < end - 1 < self.len() and + // child + 1 < end <= self.len(), so they're valid indexes. + // child == 2 * hole.pos() + 1 != hole.pos() and + // child + 1 == 2 * hole.pos() + 2 != hole.pos(). + // FIXME: 2 * hole.pos() + 1 or 2 * hole.pos() + 2 could overflow + // if T is a ZST + child += unsafe { hole.get(child) <= hole.get(child + 1) } as usize; + + // if we are already in order, stop. + // SAFETY: child is now either the old child or the old child+1 + // We already proven that both are < self.len() and != hole.pos() + if hole.element() >= unsafe { hole.get(child) } { + return; + } + + // SAFETY: same as above. + unsafe { hole.move_to(child) }; + child = 2 * hole.pos() + 1; + } + + // SAFETY: && short circuit, which means that in the + // second condition it's already true that child == end - 1 < self.len(). + if child == end - 1 && hole.element() < unsafe { hole.get(child) } { + // SAFETY: child is already proven to be a valid index and + // child == 2 * hole.pos() + 1 != hole.pos(). + unsafe { hole.move_to(child) }; + } + } + + /// # Safety + /// + /// The caller must guarantee that `pos < self.len()`. + unsafe fn sift_down(&mut self, pos: usize) { + let len = self.len(); + // SAFETY: pos < len is guaranteed by the caller and + // obviously len = self.len() <= self.len(). + unsafe { self.sift_down_range(pos, len) }; + } + + /// Take an element at `pos` and move it all the way down the heap, + /// then sift it up to its position. + /// + /// Note: This is faster when the element is known to be large / should + /// be closer to the bottom. + /// + /// # Safety + /// + /// The caller must guarantee that `pos < self.len()`. + unsafe fn sift_down_to_bottom(&mut self, mut pos: usize) { + let end = self.len(); + let start = pos; + + // SAFETY: The caller guarantees that pos < self.len(). + let mut hole = unsafe { Hole::new(&mut self.data, pos) }; + let mut child = 2 * hole.pos() + 1; + + // Loop invariant: child == 2 * hole.pos() + 1. + while child <= end.saturating_sub(2) { + // SAFETY: child < end - 1 < self.len() and + // child + 1 < end <= self.len(), so they're valid indexes. + // child == 2 * hole.pos() + 1 != hole.pos() and + // child + 1 == 2 * hole.pos() + 2 != hole.pos(). + // FIXME: 2 * hole.pos() + 1 or 2 * hole.pos() + 2 could overflow + // if T is a ZST + child += unsafe { hole.get(child) <= hole.get(child + 1) } as usize; + + // SAFETY: Same as above + unsafe { hole.move_to(child) }; + child = 2 * hole.pos() + 1; + } + + if child == end - 1 { + // SAFETY: child == end - 1 < self.len(), so it's a valid index + // and child == 2 * hole.pos() + 1 != hole.pos(). + unsafe { hole.move_to(child) }; + } + pos = hole.pos(); + drop(hole); + + // SAFETY: pos is the position in the hole and was already proven + // to be a valid index. + unsafe { self.sift_up(start, pos) }; + } + + /// Rebuild assuming data[0..start] is still a proper heap. + fn rebuild_tail(&mut self, start: usize) { + if start == self.len() { + return; + } + + let tail_len = self.len() - start; + + #[inline(always)] + fn log2_fast(x: usize) -> usize { + (usize::BITS - x.leading_zeros() - 1) as usize + } + + // `rebuild` takes O(self.len()) operations + // and about 2 * self.len() comparisons in the worst case + // while repeating `sift_up` takes O(tail_len * log(start)) operations + // and about 1 * tail_len * log_2(start) comparisons in the worst case, + // assuming start >= tail_len. For larger heaps, the crossover point + // no longer follows this reasoning and was determined empirically. + let better_to_rebuild = if start < tail_len { + true + } else if self.len() <= 2048 { + 2 * self.len() < tail_len * log2_fast(start) + } else { + 2 * self.len() < tail_len * 11 + }; + + if better_to_rebuild { + self.rebuild(); + } else { + for i in start..self.len() { + // SAFETY: The index `i` is always less than self.len(). + unsafe { self.sift_up(0, i) }; + } + } + } + + fn rebuild(&mut self) { + let mut n = self.len() / 2; + while n > 0 { + n -= 1; + // SAFETY: n starts from self.len() / 2 and goes down to 0. + // The only case when !(n < self.len()) is if + // self.len() == 0, but it's ruled out by the loop condition. + unsafe { self.sift_down(n) }; + } + } + + /// Moves all the elements of `other` into `self`, leaving `other` empty. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// + /// let mut a = BinaryHeap::from([-10, 1, 2, 3, 3]); + /// let mut b = BinaryHeap::from([-20, 5, 43]); + /// + /// a.append(&mut b); + /// + /// assert_eq!(a.into_sorted_vec(), [-20, -10, 1, 2, 3, 3, 5, 43]); + /// assert!(b.is_empty()); + /// ``` + #[stable(feature = "binary_heap_append", since = "1.11.0")] + pub fn append(&mut self, other: &mut Self) { + if self.len() < other.len() { + swap(self, other); + } + + let start = self.data.len(); + + self.data.append(&mut other.data); + + self.rebuild_tail(start); + } + + /// Clears the binary heap, returning an iterator over the removed elements + /// in heap order. If the iterator is dropped before being fully consumed, + /// it drops the remaining elements in heap order. + /// + /// The returned iterator keeps a mutable borrow on the heap to optimize + /// its implementation. + /// + /// Note: + /// * `.drain_sorted()` is *O*(*n* \* log(*n*)); much slower than `.drain()`. + /// You should use the latter for most cases. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(binary_heap_drain_sorted)] + /// use std::collections::BinaryHeap; + /// + /// let mut heap = BinaryHeap::from([1, 2, 3, 4, 5]); + /// assert_eq!(heap.len(), 5); + /// + /// drop(heap.drain_sorted()); // removes all elements in heap order + /// assert_eq!(heap.len(), 0); + /// ``` + #[inline] + #[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] + pub fn drain_sorted(&mut self) -> DrainSorted<'_, T> { + DrainSorted { inner: self } + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all elements `e` for which `f(&e)` returns + /// `false`. The elements are visited in unsorted (and unspecified) order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(binary_heap_retain)] + /// use std::collections::BinaryHeap; + /// + /// let mut heap = BinaryHeap::from([-10, -5, 1, 2, 4, 13]); + /// + /// heap.retain(|x| x % 2 == 0); // only keep even numbers + /// + /// assert_eq!(heap.into_sorted_vec(), [-10, 2, 4]) + /// ``` + #[unstable(feature = "binary_heap_retain", issue = "71503")] + pub fn retain<F>(&mut self, mut f: F) + where + F: FnMut(&T) -> bool, + { + let mut first_removed = self.len(); + let mut i = 0; + self.data.retain(|e| { + let keep = f(e); + if !keep && i < first_removed { + first_removed = i; + } + i += 1; + keep + }); + // data[0..first_removed] is untouched, so we only need to rebuild the tail: + self.rebuild_tail(first_removed); + } +} + +impl<T> BinaryHeap<T> { + /// Returns an iterator visiting all values in the underlying vector, in + /// arbitrary order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let heap = BinaryHeap::from([1, 2, 3, 4]); + /// + /// // Print 1, 2, 3, 4 in arbitrary order + /// for x in heap.iter() { + /// println!("{x}"); + /// } + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter(&self) -> Iter<'_, T> { + Iter { iter: self.data.iter() } + } + + /// Returns an iterator which retrieves elements in heap order. + /// This method consumes the original heap. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(binary_heap_into_iter_sorted)] + /// use std::collections::BinaryHeap; + /// let heap = BinaryHeap::from([1, 2, 3, 4, 5]); + /// + /// assert_eq!(heap.into_iter_sorted().take(2).collect::<Vec<_>>(), [5, 4]); + /// ``` + #[unstable(feature = "binary_heap_into_iter_sorted", issue = "59278")] + pub fn into_iter_sorted(self) -> IntoIterSorted<T> { + IntoIterSorted { inner: self } + } + + /// Returns the greatest item in the binary heap, or `None` if it is empty. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// assert_eq!(heap.peek(), None); + /// + /// heap.push(1); + /// heap.push(5); + /// heap.push(2); + /// assert_eq!(heap.peek(), Some(&5)); + /// + /// ``` + /// + /// # Time complexity + /// + /// Cost is *O*(1) in the worst case. + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn peek(&self) -> Option<&T> { + self.data.get(0) + } + + /// Returns the number of elements the binary heap can hold without reallocating. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::with_capacity(100); + /// assert!(heap.capacity() >= 100); + /// heap.push(4); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn capacity(&self) -> usize { + self.data.capacity() + } + + /// Reserves the minimum capacity for at least `additional` elements more than + /// the current length. Unlike [`reserve`], this will not + /// deliberately over-allocate to speculatively avoid frequent allocations. + /// After calling `reserve_exact`, capacity will be greater than or equal to + /// `self.len() + additional`. Does nothing if the capacity is already + /// sufficient. + /// + /// [`reserve`]: BinaryHeap::reserve + /// + /// # Panics + /// + /// Panics if the new capacity overflows [`usize`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// heap.reserve_exact(100); + /// assert!(heap.capacity() >= 100); + /// heap.push(4); + /// ``` + /// + /// [`reserve`]: BinaryHeap::reserve + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve_exact(&mut self, additional: usize) { + self.data.reserve_exact(additional); + } + + /// Reserves capacity for at least `additional` elements more than the + /// current length. The allocator may reserve more space to speculatively + /// avoid frequent allocations. After calling `reserve`, + /// capacity will be greater than or equal to `self.len() + additional`. + /// Does nothing if capacity is already sufficient. + /// + /// # Panics + /// + /// Panics if the new capacity overflows [`usize`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// heap.reserve(100); + /// assert!(heap.capacity() >= 100); + /// heap.push(4); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve(&mut self, additional: usize) { + self.data.reserve(additional); + } + + /// Tries to reserve the minimum capacity for at least `additional` elements + /// more than the current length. Unlike [`try_reserve`], this will not + /// deliberately over-allocate to speculatively avoid frequent allocations. + /// After calling `try_reserve_exact`, capacity will be greater than or + /// equal to `self.len() + additional` if it returns `Ok(())`. + /// Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the collection more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: BinaryHeap::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BinaryHeap; + /// use std::collections::TryReserveError; + /// + /// fn find_max_slow(data: &[u32]) -> Result<Option<u32>, TryReserveError> { + /// let mut heap = BinaryHeap::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// heap.try_reserve_exact(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// heap.extend(data.iter()); + /// + /// Ok(heap.pop()) + /// } + /// # find_max_slow(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve_2", since = "1.63.0")] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.data.try_reserve_exact(additional) + } + + /// Tries to reserve capacity for at least `additional` elements more than the + /// current length. The allocator may reserve more space to speculatively + /// avoid frequent allocations. After calling `try_reserve`, capacity will be + /// greater than or equal to `self.len() + additional` if it returns + /// `Ok(())`. Does nothing if capacity is already sufficient. + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BinaryHeap; + /// use std::collections::TryReserveError; + /// + /// fn find_max_slow(data: &[u32]) -> Result<Option<u32>, TryReserveError> { + /// let mut heap = BinaryHeap::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// heap.try_reserve(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// heap.extend(data.iter()); + /// + /// Ok(heap.pop()) + /// } + /// # find_max_slow(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve_2", since = "1.63.0")] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.data.try_reserve(additional) + } + + /// Discards as much additional capacity as possible. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap: BinaryHeap<i32> = BinaryHeap::with_capacity(100); + /// + /// assert!(heap.capacity() >= 100); + /// heap.shrink_to_fit(); + /// assert!(heap.capacity() == 0); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn shrink_to_fit(&mut self) { + self.data.shrink_to_fit(); + } + + /// Discards capacity with a lower bound. + /// + /// The capacity will remain at least as large as both the length + /// and the supplied value. + /// + /// If the current capacity is less than the lower limit, this is a no-op. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap: BinaryHeap<i32> = BinaryHeap::with_capacity(100); + /// + /// assert!(heap.capacity() >= 100); + /// heap.shrink_to(10); + /// assert!(heap.capacity() >= 10); + /// ``` + #[inline] + #[stable(feature = "shrink_to", since = "1.56.0")] + pub fn shrink_to(&mut self, min_capacity: usize) { + self.data.shrink_to(min_capacity) + } + + /// Returns a slice of all values in the underlying vector, in arbitrary + /// order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(binary_heap_as_slice)] + /// use std::collections::BinaryHeap; + /// use std::io::{self, Write}; + /// + /// let heap = BinaryHeap::from([1, 2, 3, 4, 5, 6, 7]); + /// + /// io::sink().write(heap.as_slice()).unwrap(); + /// ``` + #[must_use] + #[unstable(feature = "binary_heap_as_slice", issue = "83659")] + pub fn as_slice(&self) -> &[T] { + self.data.as_slice() + } + + /// Consumes the `BinaryHeap` and returns the underlying vector + /// in arbitrary order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let heap = BinaryHeap::from([1, 2, 3, 4, 5, 6, 7]); + /// let vec = heap.into_vec(); + /// + /// // Will print in some order + /// for x in vec { + /// println!("{x}"); + /// } + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "binary_heap_extras_15", since = "1.5.0")] + pub fn into_vec(self) -> Vec<T> { + self.into() + } + + /// Returns the length of the binary heap. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let heap = BinaryHeap::from([1, 3]); + /// + /// assert_eq!(heap.len(), 2); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn len(&self) -> usize { + self.data.len() + } + + /// Checks if the binary heap is empty. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::new(); + /// + /// assert!(heap.is_empty()); + /// + /// heap.push(3); + /// heap.push(5); + /// heap.push(1); + /// + /// assert!(!heap.is_empty()); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_empty(&self) -> bool { + self.len() == 0 + } + + /// Clears the binary heap, returning an iterator over the removed elements + /// in arbitrary order. If the iterator is dropped before being fully + /// consumed, it drops the remaining elements in arbitrary order. + /// + /// The returned iterator keeps a mutable borrow on the heap to optimize + /// its implementation. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::from([1, 3]); + /// + /// assert!(!heap.is_empty()); + /// + /// for x in heap.drain() { + /// println!("{x}"); + /// } + /// + /// assert!(heap.is_empty()); + /// ``` + #[inline] + #[stable(feature = "drain", since = "1.6.0")] + pub fn drain(&mut self) -> Drain<'_, T> { + Drain { iter: self.data.drain(..) } + } + + /// Drops all items from the binary heap. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let mut heap = BinaryHeap::from([1, 3]); + /// + /// assert!(!heap.is_empty()); + /// + /// heap.clear(); + /// + /// assert!(heap.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) { + self.drain(); + } +} + +/// Hole represents a hole in a slice i.e., an index without valid value +/// (because it was moved from or duplicated). +/// In drop, `Hole` will restore the slice by filling the hole +/// position with the value that was originally removed. +struct Hole<'a, T: 'a> { + data: &'a mut [T], + elt: ManuallyDrop<T>, + pos: usize, +} + +impl<'a, T> Hole<'a, T> { + /// Create a new `Hole` at index `pos`. + /// + /// Unsafe because pos must be within the data slice. + #[inline] + unsafe fn new(data: &'a mut [T], pos: usize) -> Self { + debug_assert!(pos < data.len()); + // SAFE: pos should be inside the slice + let elt = unsafe { ptr::read(data.get_unchecked(pos)) }; + Hole { data, elt: ManuallyDrop::new(elt), pos } + } + + #[inline] + fn pos(&self) -> usize { + self.pos + } + + /// Returns a reference to the element removed. + #[inline] + fn element(&self) -> &T { + &self.elt + } + + /// Returns a reference to the element at `index`. + /// + /// Unsafe because index must be within the data slice and not equal to pos. + #[inline] + unsafe fn get(&self, index: usize) -> &T { + debug_assert!(index != self.pos); + debug_assert!(index < self.data.len()); + unsafe { self.data.get_unchecked(index) } + } + + /// Move hole to new location + /// + /// Unsafe because index must be within the data slice and not equal to pos. + #[inline] + unsafe fn move_to(&mut self, index: usize) { + debug_assert!(index != self.pos); + debug_assert!(index < self.data.len()); + unsafe { + let ptr = self.data.as_mut_ptr(); + let index_ptr: *const _ = ptr.add(index); + let hole_ptr = ptr.add(self.pos); + ptr::copy_nonoverlapping(index_ptr, hole_ptr, 1); + } + self.pos = index; + } +} + +impl<T> Drop for Hole<'_, T> { + #[inline] + fn drop(&mut self) { + // fill the hole again + unsafe { + let pos = self.pos; + ptr::copy_nonoverlapping(&*self.elt, self.data.get_unchecked_mut(pos), 1); + } + } +} + +/// An iterator over the elements of a `BinaryHeap`. +/// +/// This `struct` is created by [`BinaryHeap::iter()`]. See its +/// documentation for more. +/// +/// [`iter`]: BinaryHeap::iter +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Iter<'a, T: 'a> { + iter: slice::Iter<'a, T>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for Iter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Iter").field(&self.iter.as_slice()).finish() + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for Iter<'_, T> { + fn clone(&self) -> Self { + Iter { iter: self.iter.clone() } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for Iter<'a, T> { + type Item = &'a T; + + #[inline] + fn next(&mut self) -> Option<&'a T> { + self.iter.next() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + #[inline] + fn last(self) -> Option<&'a T> { + self.iter.last() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for Iter<'a, T> { + #[inline] + fn next_back(&mut self) -> Option<&'a T> { + self.iter.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for Iter<'_, T> { + fn is_empty(&self) -> bool { + self.iter.is_empty() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Iter<'_, T> {} + +/// An owning iterator over the elements of a `BinaryHeap`. +/// +/// This `struct` is created by [`BinaryHeap::into_iter()`] +/// (provided by the [`IntoIterator`] trait). See its documentation for more. +/// +/// [`into_iter`]: BinaryHeap::into_iter +/// [`IntoIterator`]: core::iter::IntoIterator +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Clone)] +pub struct IntoIter<T> { + iter: vec::IntoIter<T>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for IntoIter<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("IntoIter").field(&self.iter.as_slice()).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Iterator for IntoIter<T> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.iter.next() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> DoubleEndedIterator for IntoIter<T> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.iter.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for IntoIter<T> { + fn is_empty(&self) -> bool { + self.iter.is_empty() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for IntoIter<T> {} + +// In addition to the SAFETY invariants of the following three unsafe traits +// also refer to the vec::in_place_collect module documentation to get an overview +#[unstable(issue = "none", feature = "inplace_iteration")] +#[doc(hidden)] +unsafe impl<T> SourceIter for IntoIter<T> { + type Source = IntoIter<T>; + + #[inline] + unsafe fn as_inner(&mut self) -> &mut Self::Source { + self + } +} + +#[unstable(issue = "none", feature = "inplace_iteration")] +#[doc(hidden)] +unsafe impl<I> InPlaceIterable for IntoIter<I> {} + +unsafe impl<I> AsVecIntoIter for IntoIter<I> { + type Item = I; + + fn as_into_iter(&mut self) -> &mut vec::IntoIter<Self::Item> { + &mut self.iter + } +} + +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[unstable(feature = "binary_heap_into_iter_sorted", issue = "59278")] +#[derive(Clone, Debug)] +pub struct IntoIterSorted<T> { + inner: BinaryHeap<T>, +} + +#[unstable(feature = "binary_heap_into_iter_sorted", issue = "59278")] +impl<T: Ord> Iterator for IntoIterSorted<T> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.inner.pop() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let exact = self.inner.len(); + (exact, Some(exact)) + } +} + +#[unstable(feature = "binary_heap_into_iter_sorted", issue = "59278")] +impl<T: Ord> ExactSizeIterator for IntoIterSorted<T> {} + +#[unstable(feature = "binary_heap_into_iter_sorted", issue = "59278")] +impl<T: Ord> FusedIterator for IntoIterSorted<T> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T: Ord> TrustedLen for IntoIterSorted<T> {} + +/// A draining iterator over the elements of a `BinaryHeap`. +/// +/// This `struct` is created by [`BinaryHeap::drain()`]. See its +/// documentation for more. +/// +/// [`drain`]: BinaryHeap::drain +#[stable(feature = "drain", since = "1.6.0")] +#[derive(Debug)] +pub struct Drain<'a, T: 'a> { + iter: vec::Drain<'a, T>, +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T> Iterator for Drain<'_, T> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.iter.next() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T> DoubleEndedIterator for Drain<'_, T> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.iter.next_back() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T> ExactSizeIterator for Drain<'_, T> { + fn is_empty(&self) -> bool { + self.iter.is_empty() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Drain<'_, T> {} + +/// A draining iterator over the elements of a `BinaryHeap`. +/// +/// This `struct` is created by [`BinaryHeap::drain_sorted()`]. See its +/// documentation for more. +/// +/// [`drain_sorted`]: BinaryHeap::drain_sorted +#[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] +#[derive(Debug)] +pub struct DrainSorted<'a, T: Ord> { + inner: &'a mut BinaryHeap<T>, +} + +#[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] +impl<'a, T: Ord> Drop for DrainSorted<'a, T> { + /// Removes heap elements in heap order. + fn drop(&mut self) { + struct DropGuard<'r, 'a, T: Ord>(&'r mut DrainSorted<'a, T>); + + impl<'r, 'a, T: Ord> Drop for DropGuard<'r, 'a, T> { + fn drop(&mut self) { + while self.0.inner.pop().is_some() {} + } + } + + while let Some(item) = self.inner.pop() { + let guard = DropGuard(self); + drop(item); + mem::forget(guard); + } + } +} + +#[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] +impl<T: Ord> Iterator for DrainSorted<'_, T> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.inner.pop() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let exact = self.inner.len(); + (exact, Some(exact)) + } +} + +#[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] +impl<T: Ord> ExactSizeIterator for DrainSorted<'_, T> {} + +#[unstable(feature = "binary_heap_drain_sorted", issue = "59278")] +impl<T: Ord> FusedIterator for DrainSorted<'_, T> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T: Ord> TrustedLen for DrainSorted<'_, T> {} + +#[stable(feature = "binary_heap_extras_15", since = "1.5.0")] +impl<T: Ord> From<Vec<T>> for BinaryHeap<T> { + /// Converts a `Vec<T>` into a `BinaryHeap<T>`. + /// + /// This conversion happens in-place, and has *O*(*n*) time complexity. + fn from(vec: Vec<T>) -> BinaryHeap<T> { + let mut heap = BinaryHeap { data: vec }; + heap.rebuild(); + heap + } +} + +#[stable(feature = "std_collections_from_array", since = "1.56.0")] +impl<T: Ord, const N: usize> From<[T; N]> for BinaryHeap<T> { + /// ``` + /// use std::collections::BinaryHeap; + /// + /// let mut h1 = BinaryHeap::from([1, 4, 2, 3]); + /// let mut h2: BinaryHeap<_> = [1, 4, 2, 3].into(); + /// while let Some((a, b)) = h1.pop().zip(h2.pop()) { + /// assert_eq!(a, b); + /// } + /// ``` + fn from(arr: [T; N]) -> Self { + Self::from_iter(arr) + } +} + +#[stable(feature = "binary_heap_extras_15", since = "1.5.0")] +impl<T> From<BinaryHeap<T>> for Vec<T> { + /// Converts a `BinaryHeap<T>` into a `Vec<T>`. + /// + /// This conversion requires no data movement or allocation, and has + /// constant time complexity. + fn from(heap: BinaryHeap<T>) -> Vec<T> { + heap.data + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord> FromIterator<T> for BinaryHeap<T> { + fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> BinaryHeap<T> { + BinaryHeap::from(iter.into_iter().collect::<Vec<_>>()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> IntoIterator for BinaryHeap<T> { + type Item = T; + type IntoIter = IntoIter<T>; + + /// Creates a consuming iterator, that is, one that moves each value out of + /// the binary heap in arbitrary order. The binary heap cannot be used + /// after calling this. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BinaryHeap; + /// let heap = BinaryHeap::from([1, 2, 3, 4]); + /// + /// // Print 1, 2, 3, 4 in arbitrary order + /// for x in heap.into_iter() { + /// // x has type i32, not &i32 + /// println!("{x}"); + /// } + /// ``` + fn into_iter(self) -> IntoIter<T> { + IntoIter { iter: self.data.into_iter() } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> IntoIterator for &'a BinaryHeap<T> { + type Item = &'a T; + type IntoIter = Iter<'a, T>; + + fn into_iter(self) -> Iter<'a, T> { + self.iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord> Extend<T> for BinaryHeap<T> { + #[inline] + fn extend<I: IntoIterator<Item = T>>(&mut self, iter: I) { + <Self as SpecExtend<I>>::spec_extend(self, iter); + } + + #[inline] + fn extend_one(&mut self, item: T) { + self.push(item); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +impl<T: Ord, I: IntoIterator<Item = T>> SpecExtend<I> for BinaryHeap<T> { + default fn spec_extend(&mut self, iter: I) { + self.extend_desugared(iter.into_iter()); + } +} + +impl<T: Ord> SpecExtend<Vec<T>> for BinaryHeap<T> { + fn spec_extend(&mut self, ref mut other: Vec<T>) { + let start = self.data.len(); + self.data.append(other); + self.rebuild_tail(start); + } +} + +impl<T: Ord> SpecExtend<BinaryHeap<T>> for BinaryHeap<T> { + fn spec_extend(&mut self, ref mut other: BinaryHeap<T>) { + self.append(other); + } +} + +impl<T: Ord> BinaryHeap<T> { + fn extend_desugared<I: IntoIterator<Item = T>>(&mut self, iter: I) { + let iterator = iter.into_iter(); + let (lower, _) = iterator.size_hint(); + + self.reserve(lower); + + iterator.for_each(move |elem| self.push(elem)); + } +} + +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, T: 'a + Ord + Copy> Extend<&'a T> for BinaryHeap<T> { + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iter: I) { + self.extend(iter.into_iter().cloned()); + } + + #[inline] + fn extend_one(&mut self, &item: &'a T) { + self.push(item); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} diff --git a/library/alloc/src/collections/binary_heap/tests.rs b/library/alloc/src/collections/binary_heap/tests.rs new file mode 100644 index 000000000..5a05215ae --- /dev/null +++ b/library/alloc/src/collections/binary_heap/tests.rs @@ -0,0 +1,489 @@ +use super::*; +use crate::boxed::Box; +use std::iter::TrustedLen; +use std::panic::{catch_unwind, AssertUnwindSafe}; +use std::sync::atomic::{AtomicU32, Ordering}; + +#[test] +fn test_iterator() { + let data = vec![5, 9, 3]; + let iterout = [9, 5, 3]; + let heap = BinaryHeap::from(data); + let mut i = 0; + for el in &heap { + assert_eq!(*el, iterout[i]); + i += 1; + } +} + +#[test] +fn test_iter_rev_cloned_collect() { + let data = vec![5, 9, 3]; + let iterout = vec![3, 5, 9]; + let pq = BinaryHeap::from(data); + + let v: Vec<_> = pq.iter().rev().cloned().collect(); + assert_eq!(v, iterout); +} + +#[test] +fn test_into_iter_collect() { + let data = vec![5, 9, 3]; + let iterout = vec![9, 5, 3]; + let pq = BinaryHeap::from(data); + + let v: Vec<_> = pq.into_iter().collect(); + assert_eq!(v, iterout); +} + +#[test] +fn test_into_iter_size_hint() { + let data = vec![5, 9]; + let pq = BinaryHeap::from(data); + + let mut it = pq.into_iter(); + + assert_eq!(it.size_hint(), (2, Some(2))); + assert_eq!(it.next(), Some(9)); + + assert_eq!(it.size_hint(), (1, Some(1))); + assert_eq!(it.next(), Some(5)); + + assert_eq!(it.size_hint(), (0, Some(0))); + assert_eq!(it.next(), None); +} + +#[test] +fn test_into_iter_rev_collect() { + let data = vec![5, 9, 3]; + let iterout = vec![3, 5, 9]; + let pq = BinaryHeap::from(data); + + let v: Vec<_> = pq.into_iter().rev().collect(); + assert_eq!(v, iterout); +} + +#[test] +fn test_into_iter_sorted_collect() { + let heap = BinaryHeap::from(vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]); + let it = heap.into_iter_sorted(); + let sorted = it.collect::<Vec<_>>(); + assert_eq!(sorted, vec![10, 9, 8, 7, 6, 5, 4, 3, 2, 2, 1, 1, 0]); +} + +#[test] +fn test_drain_sorted_collect() { + let mut heap = BinaryHeap::from(vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]); + let it = heap.drain_sorted(); + let sorted = it.collect::<Vec<_>>(); + assert_eq!(sorted, vec![10, 9, 8, 7, 6, 5, 4, 3, 2, 2, 1, 1, 0]); +} + +fn check_exact_size_iterator<I: ExactSizeIterator>(len: usize, it: I) { + let mut it = it; + + for i in 0..it.len() { + let (lower, upper) = it.size_hint(); + assert_eq!(Some(lower), upper); + assert_eq!(lower, len - i); + assert_eq!(it.len(), len - i); + it.next(); + } + assert_eq!(it.len(), 0); + assert!(it.is_empty()); +} + +#[test] +fn test_exact_size_iterator() { + let heap = BinaryHeap::from(vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]); + check_exact_size_iterator(heap.len(), heap.iter()); + check_exact_size_iterator(heap.len(), heap.clone().into_iter()); + check_exact_size_iterator(heap.len(), heap.clone().into_iter_sorted()); + check_exact_size_iterator(heap.len(), heap.clone().drain()); + check_exact_size_iterator(heap.len(), heap.clone().drain_sorted()); +} + +fn check_trusted_len<I: TrustedLen>(len: usize, it: I) { + let mut it = it; + for i in 0..len { + let (lower, upper) = it.size_hint(); + if upper.is_some() { + assert_eq!(Some(lower), upper); + assert_eq!(lower, len - i); + } + it.next(); + } +} + +#[test] +fn test_trusted_len() { + let heap = BinaryHeap::from(vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]); + check_trusted_len(heap.len(), heap.clone().into_iter_sorted()); + check_trusted_len(heap.len(), heap.clone().drain_sorted()); +} + +#[test] +fn test_peek_and_pop() { + let data = vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]; + let mut sorted = data.clone(); + sorted.sort(); + let mut heap = BinaryHeap::from(data); + while !heap.is_empty() { + assert_eq!(heap.peek().unwrap(), sorted.last().unwrap()); + assert_eq!(heap.pop().unwrap(), sorted.pop().unwrap()); + } +} + +#[test] +fn test_peek_mut() { + let data = vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]; + let mut heap = BinaryHeap::from(data); + assert_eq!(heap.peek(), Some(&10)); + { + let mut top = heap.peek_mut().unwrap(); + *top -= 2; + } + assert_eq!(heap.peek(), Some(&9)); +} + +#[test] +fn test_peek_mut_pop() { + let data = vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]; + let mut heap = BinaryHeap::from(data); + assert_eq!(heap.peek(), Some(&10)); + { + let mut top = heap.peek_mut().unwrap(); + *top -= 2; + assert_eq!(PeekMut::pop(top), 8); + } + assert_eq!(heap.peek(), Some(&9)); +} + +#[test] +fn test_push() { + let mut heap = BinaryHeap::from(vec![2, 4, 9]); + assert_eq!(heap.len(), 3); + assert!(*heap.peek().unwrap() == 9); + heap.push(11); + assert_eq!(heap.len(), 4); + assert!(*heap.peek().unwrap() == 11); + heap.push(5); + assert_eq!(heap.len(), 5); + assert!(*heap.peek().unwrap() == 11); + heap.push(27); + assert_eq!(heap.len(), 6); + assert!(*heap.peek().unwrap() == 27); + heap.push(3); + assert_eq!(heap.len(), 7); + assert!(*heap.peek().unwrap() == 27); + heap.push(103); + assert_eq!(heap.len(), 8); + assert!(*heap.peek().unwrap() == 103); +} + +#[test] +fn test_push_unique() { + let mut heap = BinaryHeap::<Box<_>>::from(vec![Box::new(2), Box::new(4), Box::new(9)]); + assert_eq!(heap.len(), 3); + assert!(**heap.peek().unwrap() == 9); + heap.push(Box::new(11)); + assert_eq!(heap.len(), 4); + assert!(**heap.peek().unwrap() == 11); + heap.push(Box::new(5)); + assert_eq!(heap.len(), 5); + assert!(**heap.peek().unwrap() == 11); + heap.push(Box::new(27)); + assert_eq!(heap.len(), 6); + assert!(**heap.peek().unwrap() == 27); + heap.push(Box::new(3)); + assert_eq!(heap.len(), 7); + assert!(**heap.peek().unwrap() == 27); + heap.push(Box::new(103)); + assert_eq!(heap.len(), 8); + assert!(**heap.peek().unwrap() == 103); +} + +fn check_to_vec(mut data: Vec<i32>) { + let heap = BinaryHeap::from(data.clone()); + let mut v = heap.clone().into_vec(); + v.sort(); + data.sort(); + + assert_eq!(v, data); + assert_eq!(heap.into_sorted_vec(), data); +} + +#[test] +fn test_to_vec() { + check_to_vec(vec![]); + check_to_vec(vec![5]); + check_to_vec(vec![3, 2]); + check_to_vec(vec![2, 3]); + check_to_vec(vec![5, 1, 2]); + check_to_vec(vec![1, 100, 2, 3]); + check_to_vec(vec![1, 3, 5, 7, 9, 2, 4, 6, 8, 0]); + check_to_vec(vec![2, 4, 6, 2, 1, 8, 10, 3, 5, 7, 0, 9, 1]); + check_to_vec(vec![9, 11, 9, 9, 9, 9, 11, 2, 3, 4, 11, 9, 0, 0, 0, 0]); + check_to_vec(vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]); + check_to_vec(vec![10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]); + check_to_vec(vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 0, 0, 1, 2]); + check_to_vec(vec![5, 4, 3, 2, 1, 5, 4, 3, 2, 1, 5, 4, 3, 2, 1]); +} + +#[test] +fn test_in_place_iterator_specialization() { + let src: Vec<usize> = vec![1, 2, 3]; + let src_ptr = src.as_ptr(); + let heap: BinaryHeap<_> = src.into_iter().map(std::convert::identity).collect(); + let heap_ptr = heap.iter().next().unwrap() as *const usize; + assert_eq!(src_ptr, heap_ptr); + let sink: Vec<_> = heap.into_iter().map(std::convert::identity).collect(); + let sink_ptr = sink.as_ptr(); + assert_eq!(heap_ptr, sink_ptr); +} + +#[test] +fn test_empty_pop() { + let mut heap = BinaryHeap::<i32>::new(); + assert!(heap.pop().is_none()); +} + +#[test] +fn test_empty_peek() { + let empty = BinaryHeap::<i32>::new(); + assert!(empty.peek().is_none()); +} + +#[test] +fn test_empty_peek_mut() { + let mut empty = BinaryHeap::<i32>::new(); + assert!(empty.peek_mut().is_none()); +} + +#[test] +fn test_from_iter() { + let xs = vec![9, 8, 7, 6, 5, 4, 3, 2, 1]; + + let mut q: BinaryHeap<_> = xs.iter().rev().cloned().collect(); + + for &x in &xs { + assert_eq!(q.pop().unwrap(), x); + } +} + +#[test] +fn test_drain() { + let mut q: BinaryHeap<_> = [9, 8, 7, 6, 5, 4, 3, 2, 1].iter().cloned().collect(); + + assert_eq!(q.drain().take(5).count(), 5); + + assert!(q.is_empty()); +} + +#[test] +fn test_drain_sorted() { + let mut q: BinaryHeap<_> = [9, 8, 7, 6, 5, 4, 3, 2, 1].iter().cloned().collect(); + + assert_eq!(q.drain_sorted().take(5).collect::<Vec<_>>(), vec![9, 8, 7, 6, 5]); + + assert!(q.is_empty()); +} + +#[test] +fn test_drain_sorted_leak() { + static DROPS: AtomicU32 = AtomicU32::new(0); + + #[derive(Clone, PartialEq, Eq, PartialOrd, Ord)] + struct D(u32, bool); + + impl Drop for D { + fn drop(&mut self) { + DROPS.fetch_add(1, Ordering::SeqCst); + + if self.1 { + panic!("panic in `drop`"); + } + } + } + + let mut q = BinaryHeap::from(vec![ + D(0, false), + D(1, false), + D(2, false), + D(3, true), + D(4, false), + D(5, false), + ]); + + catch_unwind(AssertUnwindSafe(|| drop(q.drain_sorted()))).ok(); + + assert_eq!(DROPS.load(Ordering::SeqCst), 6); +} + +#[test] +fn test_extend_ref() { + let mut a = BinaryHeap::new(); + a.push(1); + a.push(2); + + a.extend(&[3, 4, 5]); + + assert_eq!(a.len(), 5); + assert_eq!(a.into_sorted_vec(), [1, 2, 3, 4, 5]); + + let mut a = BinaryHeap::new(); + a.push(1); + a.push(2); + let mut b = BinaryHeap::new(); + b.push(3); + b.push(4); + b.push(5); + + a.extend(&b); + + assert_eq!(a.len(), 5); + assert_eq!(a.into_sorted_vec(), [1, 2, 3, 4, 5]); +} + +#[test] +fn test_append() { + let mut a = BinaryHeap::from(vec![-10, 1, 2, 3, 3]); + let mut b = BinaryHeap::from(vec![-20, 5, 43]); + + a.append(&mut b); + + assert_eq!(a.into_sorted_vec(), [-20, -10, 1, 2, 3, 3, 5, 43]); + assert!(b.is_empty()); +} + +#[test] +fn test_append_to_empty() { + let mut a = BinaryHeap::new(); + let mut b = BinaryHeap::from(vec![-20, 5, 43]); + + a.append(&mut b); + + assert_eq!(a.into_sorted_vec(), [-20, 5, 43]); + assert!(b.is_empty()); +} + +#[test] +fn test_extend_specialization() { + let mut a = BinaryHeap::from(vec![-10, 1, 2, 3, 3]); + let b = BinaryHeap::from(vec![-20, 5, 43]); + + a.extend(b); + + assert_eq!(a.into_sorted_vec(), [-20, -10, 1, 2, 3, 3, 5, 43]); +} + +#[allow(dead_code)] +fn assert_covariance() { + fn drain<'new>(d: Drain<'static, &'static str>) -> Drain<'new, &'new str> { + d + } +} + +#[test] +fn test_retain() { + let mut a = BinaryHeap::from(vec![100, 10, 50, 1, 2, 20, 30]); + a.retain(|&x| x != 2); + + // Check that 20 moved into 10's place. + assert_eq!(a.clone().into_vec(), [100, 20, 50, 1, 10, 30]); + + a.retain(|_| true); + + assert_eq!(a.clone().into_vec(), [100, 20, 50, 1, 10, 30]); + + a.retain(|&x| x < 50); + + assert_eq!(a.clone().into_vec(), [30, 20, 10, 1]); + + a.retain(|_| false); + + assert!(a.is_empty()); +} + +// old binaryheap failed this test +// +// Integrity means that all elements are present after a comparison panics, +// even if the order might not be correct. +// +// Destructors must be called exactly once per element. +// FIXME: re-enable emscripten once it can unwind again +#[test] +#[cfg(not(target_os = "emscripten"))] +fn panic_safe() { + use rand::{seq::SliceRandom, thread_rng}; + use std::cmp; + use std::panic::{self, AssertUnwindSafe}; + use std::sync::atomic::{AtomicUsize, Ordering}; + + static DROP_COUNTER: AtomicUsize = AtomicUsize::new(0); + + #[derive(Eq, PartialEq, Ord, Clone, Debug)] + struct PanicOrd<T>(T, bool); + + impl<T> Drop for PanicOrd<T> { + fn drop(&mut self) { + // update global drop count + DROP_COUNTER.fetch_add(1, Ordering::SeqCst); + } + } + + impl<T: PartialOrd> PartialOrd for PanicOrd<T> { + fn partial_cmp(&self, other: &Self) -> Option<cmp::Ordering> { + if self.1 || other.1 { + panic!("Panicking comparison"); + } + self.0.partial_cmp(&other.0) + } + } + let mut rng = thread_rng(); + const DATASZ: usize = 32; + // Miri is too slow + let ntest = if cfg!(miri) { 1 } else { 10 }; + + // don't use 0 in the data -- we want to catch the zeroed-out case. + let data = (1..=DATASZ).collect::<Vec<_>>(); + + // since it's a fuzzy test, run several tries. + for _ in 0..ntest { + for i in 1..=DATASZ { + DROP_COUNTER.store(0, Ordering::SeqCst); + + let mut panic_ords: Vec<_> = + data.iter().filter(|&&x| x != i).map(|&x| PanicOrd(x, false)).collect(); + let panic_item = PanicOrd(i, true); + + // heapify the sane items + panic_ords.shuffle(&mut rng); + let mut heap = BinaryHeap::from(panic_ords); + let inner_data; + + { + // push the panicking item to the heap and catch the panic + let thread_result = { + let mut heap_ref = AssertUnwindSafe(&mut heap); + panic::catch_unwind(move || { + heap_ref.push(panic_item); + }) + }; + assert!(thread_result.is_err()); + + // Assert no elements were dropped + let drops = DROP_COUNTER.load(Ordering::SeqCst); + assert!(drops == 0, "Must not drop items. drops={}", drops); + inner_data = heap.clone().into_vec(); + drop(heap); + } + let drops = DROP_COUNTER.load(Ordering::SeqCst); + assert_eq!(drops, DATASZ); + + let mut data_sorted = inner_data.into_iter().map(|p| p.0).collect::<Vec<_>>(); + data_sorted.sort(); + assert_eq!(data_sorted, data); + } + } +} diff --git a/library/alloc/src/collections/btree/append.rs b/library/alloc/src/collections/btree/append.rs new file mode 100644 index 000000000..b6989afb6 --- /dev/null +++ b/library/alloc/src/collections/btree/append.rs @@ -0,0 +1,107 @@ +use super::merge_iter::MergeIterInner; +use super::node::{self, Root}; +use core::alloc::Allocator; +use core::iter::FusedIterator; + +impl<K, V> Root<K, V> { + /// Appends all key-value pairs from the union of two ascending iterators, + /// incrementing a `length` variable along the way. The latter makes it + /// easier for the caller to avoid a leak when a drop handler panicks. + /// + /// If both iterators produce the same key, this method drops the pair from + /// the left iterator and appends the pair from the right iterator. + /// + /// If you want the tree to end up in a strictly ascending order, like for + /// a `BTreeMap`, both iterators should produce keys in strictly ascending + /// order, each greater than all keys in the tree, including any keys + /// already in the tree upon entry. + pub fn append_from_sorted_iters<I, A: Allocator + Clone>( + &mut self, + left: I, + right: I, + length: &mut usize, + alloc: A, + ) where + K: Ord, + I: Iterator<Item = (K, V)> + FusedIterator, + { + // We prepare to merge `left` and `right` into a sorted sequence in linear time. + let iter = MergeIter(MergeIterInner::new(left, right)); + + // Meanwhile, we build a tree from the sorted sequence in linear time. + self.bulk_push(iter, length, alloc) + } + + /// Pushes all key-value pairs to the end of the tree, incrementing a + /// `length` variable along the way. The latter makes it easier for the + /// caller to avoid a leak when the iterator panicks. + pub fn bulk_push<I, A: Allocator + Clone>(&mut self, iter: I, length: &mut usize, alloc: A) + where + I: Iterator<Item = (K, V)>, + { + let mut cur_node = self.borrow_mut().last_leaf_edge().into_node(); + // Iterate through all key-value pairs, pushing them into nodes at the right level. + for (key, value) in iter { + // Try to push key-value pair into the current leaf node. + if cur_node.len() < node::CAPACITY { + cur_node.push(key, value); + } else { + // No space left, go up and push there. + let mut open_node; + let mut test_node = cur_node.forget_type(); + loop { + match test_node.ascend() { + Ok(parent) => { + let parent = parent.into_node(); + if parent.len() < node::CAPACITY { + // Found a node with space left, push here. + open_node = parent; + break; + } else { + // Go up again. + test_node = parent.forget_type(); + } + } + Err(_) => { + // We are at the top, create a new root node and push there. + open_node = self.push_internal_level(alloc.clone()); + break; + } + } + } + + // Push key-value pair and new right subtree. + let tree_height = open_node.height() - 1; + let mut right_tree = Root::new(alloc.clone()); + for _ in 0..tree_height { + right_tree.push_internal_level(alloc.clone()); + } + open_node.push(key, value, right_tree); + + // Go down to the right-most leaf again. + cur_node = open_node.forget_type().last_leaf_edge().into_node(); + } + + // Increment length every iteration, to make sure the map drops + // the appended elements even if advancing the iterator panicks. + *length += 1; + } + self.fix_right_border_of_plentiful(); + } +} + +// An iterator for merging two sorted sequences into one +struct MergeIter<K, V, I: Iterator<Item = (K, V)>>(MergeIterInner<I>); + +impl<K: Ord, V, I> Iterator for MergeIter<K, V, I> +where + I: Iterator<Item = (K, V)> + FusedIterator, +{ + type Item = (K, V); + + /// If two keys are equal, returns the key-value pair from the right source. + fn next(&mut self) -> Option<(K, V)> { + let (a_next, b_next) = self.0.nexts(|a: &(K, V), b: &(K, V)| K::cmp(&a.0, &b.0)); + b_next.or(a_next) + } +} diff --git a/library/alloc/src/collections/btree/borrow.rs b/library/alloc/src/collections/btree/borrow.rs new file mode 100644 index 000000000..016f139a5 --- /dev/null +++ b/library/alloc/src/collections/btree/borrow.rs @@ -0,0 +1,47 @@ +use core::marker::PhantomData; +use core::ptr::NonNull; + +/// Models a reborrow of some unique reference, when you know that the reborrow +/// and all its descendants (i.e., all pointers and references derived from it) +/// will not be used any more at some point, after which you want to use the +/// original unique reference again. +/// +/// The borrow checker usually handles this stacking of borrows for you, but +/// some control flows that accomplish this stacking are too complicated for +/// the compiler to follow. A `DormantMutRef` allows you to check borrowing +/// yourself, while still expressing its stacked nature, and encapsulating +/// the raw pointer code needed to do this without undefined behavior. +pub struct DormantMutRef<'a, T> { + ptr: NonNull<T>, + _marker: PhantomData<&'a mut T>, +} + +unsafe impl<'a, T> Sync for DormantMutRef<'a, T> where &'a mut T: Sync {} +unsafe impl<'a, T> Send for DormantMutRef<'a, T> where &'a mut T: Send {} + +impl<'a, T> DormantMutRef<'a, T> { + /// Capture a unique borrow, and immediately reborrow it. For the compiler, + /// the lifetime of the new reference is the same as the lifetime of the + /// original reference, but you promise to use it for a shorter period. + pub fn new(t: &'a mut T) -> (&'a mut T, Self) { + let ptr = NonNull::from(t); + // SAFETY: we hold the borrow throughout 'a via `_marker`, and we expose + // only this reference, so it is unique. + let new_ref = unsafe { &mut *ptr.as_ptr() }; + (new_ref, Self { ptr, _marker: PhantomData }) + } + + /// Revert to the unique borrow initially captured. + /// + /// # Safety + /// + /// The reborrow must have ended, i.e., the reference returned by `new` and + /// all pointers and references derived from it, must not be used anymore. + pub unsafe fn awaken(self) -> &'a mut T { + // SAFETY: our own safety conditions imply this reference is again unique. + unsafe { &mut *self.ptr.as_ptr() } + } +} + +#[cfg(test)] +mod tests; diff --git a/library/alloc/src/collections/btree/borrow/tests.rs b/library/alloc/src/collections/btree/borrow/tests.rs new file mode 100644 index 000000000..56a8434fc --- /dev/null +++ b/library/alloc/src/collections/btree/borrow/tests.rs @@ -0,0 +1,19 @@ +use super::DormantMutRef; + +#[test] +fn test_borrow() { + let mut data = 1; + let mut stack = vec![]; + let mut rr = &mut data; + for factor in [2, 3, 7].iter() { + let (r, dormant_r) = DormantMutRef::new(rr); + rr = r; + assert_eq!(*rr, 1); + stack.push((factor, dormant_r)); + } + while let Some((factor, dormant_r)) = stack.pop() { + let r = unsafe { dormant_r.awaken() }; + *r *= factor; + } + assert_eq!(data, 42); +} diff --git a/library/alloc/src/collections/btree/dedup_sorted_iter.rs b/library/alloc/src/collections/btree/dedup_sorted_iter.rs new file mode 100644 index 000000000..60bf83b83 --- /dev/null +++ b/library/alloc/src/collections/btree/dedup_sorted_iter.rs @@ -0,0 +1,47 @@ +use core::iter::Peekable; + +/// A iterator for deduping the key of a sorted iterator. +/// When encountering the duplicated key, only the last key-value pair is yielded. +/// +/// Used by [`BTreeMap::bulk_build_from_sorted_iter`]. +pub struct DedupSortedIter<K, V, I> +where + I: Iterator<Item = (K, V)>, +{ + iter: Peekable<I>, +} + +impl<K, V, I> DedupSortedIter<K, V, I> +where + I: Iterator<Item = (K, V)>, +{ + pub fn new(iter: I) -> Self { + Self { iter: iter.peekable() } + } +} + +impl<K, V, I> Iterator for DedupSortedIter<K, V, I> +where + K: Eq, + I: Iterator<Item = (K, V)>, +{ + type Item = (K, V); + + fn next(&mut self) -> Option<(K, V)> { + loop { + let next = match self.iter.next() { + Some(next) => next, + None => return None, + }; + + let peeked = match self.iter.peek() { + Some(peeked) => peeked, + None => return Some(next), + }; + + if next.0 != peeked.0 { + return Some(next); + } + } + } +} diff --git a/library/alloc/src/collections/btree/fix.rs b/library/alloc/src/collections/btree/fix.rs new file mode 100644 index 000000000..91b612180 --- /dev/null +++ b/library/alloc/src/collections/btree/fix.rs @@ -0,0 +1,179 @@ +use super::map::MIN_LEN; +use super::node::{marker, ForceResult::*, Handle, LeftOrRight::*, NodeRef, Root}; +use core::alloc::Allocator; + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + /// Stocks up a possibly underfull node by merging with or stealing from a + /// sibling. If successful but at the cost of shrinking the parent node, + /// returns that shrunk parent node. Returns an `Err` if the node is + /// an empty root. + fn fix_node_through_parent<A: Allocator + Clone>( + self, + alloc: A, + ) -> Result<Option<NodeRef<marker::Mut<'a>, K, V, marker::Internal>>, Self> { + let len = self.len(); + if len >= MIN_LEN { + Ok(None) + } else { + match self.choose_parent_kv() { + Ok(Left(mut left_parent_kv)) => { + if left_parent_kv.can_merge() { + let parent = left_parent_kv.merge_tracking_parent(alloc); + Ok(Some(parent)) + } else { + left_parent_kv.bulk_steal_left(MIN_LEN - len); + Ok(None) + } + } + Ok(Right(mut right_parent_kv)) => { + if right_parent_kv.can_merge() { + let parent = right_parent_kv.merge_tracking_parent(alloc); + Ok(Some(parent)) + } else { + right_parent_kv.bulk_steal_right(MIN_LEN - len); + Ok(None) + } + } + Err(root) => { + if len > 0 { + Ok(None) + } else { + Err(root) + } + } + } + } + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + /// Stocks up a possibly underfull node, and if that causes its parent node + /// to shrink, stocks up the parent, recursively. + /// Returns `true` if it fixed the tree, `false` if it couldn't because the + /// root node became empty. + /// + /// This method does not expect ancestors to already be underfull upon entry + /// and panics if it encounters an empty ancestor. + pub fn fix_node_and_affected_ancestors<A: Allocator + Clone>(mut self, alloc: A) -> bool { + loop { + match self.fix_node_through_parent(alloc.clone()) { + Ok(Some(parent)) => self = parent.forget_type(), + Ok(None) => return true, + Err(_) => return false, + } + } + } +} + +impl<K, V> Root<K, V> { + /// Removes empty levels on the top, but keeps an empty leaf if the entire tree is empty. + pub fn fix_top<A: Allocator + Clone>(&mut self, alloc: A) { + while self.height() > 0 && self.len() == 0 { + self.pop_internal_level(alloc.clone()); + } + } + + /// Stocks up or merge away any underfull nodes on the right border of the + /// tree. The other nodes, those that are not the root nor a rightmost edge, + /// must already have at least MIN_LEN elements. + pub fn fix_right_border<A: Allocator + Clone>(&mut self, alloc: A) { + self.fix_top(alloc.clone()); + if self.len() > 0 { + self.borrow_mut().last_kv().fix_right_border_of_right_edge(alloc.clone()); + self.fix_top(alloc); + } + } + + /// The symmetric clone of `fix_right_border`. + pub fn fix_left_border<A: Allocator + Clone>(&mut self, alloc: A) { + self.fix_top(alloc.clone()); + if self.len() > 0 { + self.borrow_mut().first_kv().fix_left_border_of_left_edge(alloc.clone()); + self.fix_top(alloc); + } + } + + /// Stocks up any underfull nodes on the right border of the tree. + /// The other nodes, those that are neither the root nor a rightmost edge, + /// must be prepared to have up to MIN_LEN elements stolen. + pub fn fix_right_border_of_plentiful(&mut self) { + let mut cur_node = self.borrow_mut(); + while let Internal(internal) = cur_node.force() { + // Check if right-most child is underfull. + let mut last_kv = internal.last_kv().consider_for_balancing(); + debug_assert!(last_kv.left_child_len() >= MIN_LEN * 2); + let right_child_len = last_kv.right_child_len(); + if right_child_len < MIN_LEN { + // We need to steal. + last_kv.bulk_steal_left(MIN_LEN - right_child_len); + } + + // Go further down. + cur_node = last_kv.into_right_child(); + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::KV> { + fn fix_left_border_of_left_edge<A: Allocator + Clone>(mut self, alloc: A) { + while let Internal(internal_kv) = self.force() { + self = internal_kv.fix_left_child(alloc.clone()).first_kv(); + debug_assert!(self.reborrow().into_node().len() > MIN_LEN); + } + } + + fn fix_right_border_of_right_edge<A: Allocator + Clone>(mut self, alloc: A) { + while let Internal(internal_kv) = self.force() { + self = internal_kv.fix_right_child(alloc.clone()).last_kv(); + debug_assert!(self.reborrow().into_node().len() > MIN_LEN); + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::KV> { + /// Stocks up the left child, assuming the right child isn't underfull, and + /// provisions an extra element to allow merging its children in turn + /// without becoming underfull. + /// Returns the left child. + fn fix_left_child<A: Allocator + Clone>( + self, + alloc: A, + ) -> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + let mut internal_kv = self.consider_for_balancing(); + let left_len = internal_kv.left_child_len(); + debug_assert!(internal_kv.right_child_len() >= MIN_LEN); + if internal_kv.can_merge() { + internal_kv.merge_tracking_child(alloc) + } else { + // `MIN_LEN + 1` to avoid readjust if merge happens on the next level. + let count = (MIN_LEN + 1).saturating_sub(left_len); + if count > 0 { + internal_kv.bulk_steal_right(count); + } + internal_kv.into_left_child() + } + } + + /// Stocks up the right child, assuming the left child isn't underfull, and + /// provisions an extra element to allow merging its children in turn + /// without becoming underfull. + /// Returns wherever the right child ended up. + fn fix_right_child<A: Allocator + Clone>( + self, + alloc: A, + ) -> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + let mut internal_kv = self.consider_for_balancing(); + let right_len = internal_kv.right_child_len(); + debug_assert!(internal_kv.left_child_len() >= MIN_LEN); + if internal_kv.can_merge() { + internal_kv.merge_tracking_child(alloc) + } else { + // `MIN_LEN + 1` to avoid readjust if merge happens on the next level. + let count = (MIN_LEN + 1).saturating_sub(right_len); + if count > 0 { + internal_kv.bulk_steal_left(count); + } + internal_kv.into_right_child() + } + } +} diff --git a/library/alloc/src/collections/btree/map.rs b/library/alloc/src/collections/btree/map.rs new file mode 100644 index 000000000..cacbd54b6 --- /dev/null +++ b/library/alloc/src/collections/btree/map.rs @@ -0,0 +1,2423 @@ +use crate::vec::Vec; +use core::borrow::Borrow; +use core::cmp::Ordering; +use core::fmt::{self, Debug}; +use core::hash::{Hash, Hasher}; +use core::iter::{FromIterator, FusedIterator}; +use core::marker::PhantomData; +use core::mem::{self, ManuallyDrop}; +use core::ops::{Index, RangeBounds}; +use core::ptr; + +use crate::alloc::{Allocator, Global}; + +use super::borrow::DormantMutRef; +use super::dedup_sorted_iter::DedupSortedIter; +use super::navigate::{LazyLeafRange, LeafRange}; +use super::node::{self, marker, ForceResult::*, Handle, NodeRef, Root}; +use super::search::SearchResult::*; +use super::set_val::SetValZST; + +mod entry; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use entry::{Entry, OccupiedEntry, OccupiedError, VacantEntry}; + +use Entry::*; + +/// Minimum number of elements in a node that is not a root. +/// We might temporarily have fewer elements during methods. +pub(super) const MIN_LEN: usize = node::MIN_LEN_AFTER_SPLIT; + +// A tree in a `BTreeMap` is a tree in the `node` module with additional invariants: +// - Keys must appear in ascending order (according to the key's type). +// - Every non-leaf node contains at least 1 element (has at least 2 children). +// - Every non-root node contains at least MIN_LEN elements. +// +// An empty map is represented either by the absence of a root node or by a +// root node that is an empty leaf. + +/// An ordered map based on a [B-Tree]. +/// +/// B-Trees represent a fundamental compromise between cache-efficiency and actually minimizing +/// the amount of work performed in a search. In theory, a binary search tree (BST) is the optimal +/// choice for a sorted map, as a perfectly balanced BST performs the theoretical minimum amount of +/// comparisons necessary to find an element (log<sub>2</sub>n). However, in practice the way this +/// is done is *very* inefficient for modern computer architectures. In particular, every element +/// is stored in its own individually heap-allocated node. This means that every single insertion +/// triggers a heap-allocation, and every single comparison should be a cache-miss. Since these +/// are both notably expensive things to do in practice, we are forced to at very least reconsider +/// the BST strategy. +/// +/// A B-Tree instead makes each node contain B-1 to 2B-1 elements in a contiguous array. By doing +/// this, we reduce the number of allocations by a factor of B, and improve cache efficiency in +/// searches. However, this does mean that searches will have to do *more* comparisons on average. +/// The precise number of comparisons depends on the node search strategy used. For optimal cache +/// efficiency, one could search the nodes linearly. For optimal comparisons, one could search +/// the node using binary search. As a compromise, one could also perform a linear search +/// that initially only checks every i<sup>th</sup> element for some choice of i. +/// +/// Currently, our implementation simply performs naive linear search. This provides excellent +/// performance on *small* nodes of elements which are cheap to compare. However in the future we +/// would like to further explore choosing the optimal search strategy based on the choice of B, +/// and possibly other factors. Using linear search, searching for a random element is expected +/// to take B * log(n) comparisons, which is generally worse than a BST. In practice, +/// however, performance is excellent. +/// +/// It is a logic error for a key to be modified in such a way that the key's ordering relative to +/// any other key, as determined by the [`Ord`] trait, changes while it is in the map. This is +/// normally only possible through [`Cell`], [`RefCell`], global state, I/O, or unsafe code. +/// The behavior resulting from such a logic error is not specified, but will be encapsulated to the +/// `BTreeMap` that observed the logic error and not result in undefined behavior. This could +/// include panics, incorrect results, aborts, memory leaks, and non-termination. +/// +/// Iterators obtained from functions such as [`BTreeMap::iter`], [`BTreeMap::values`], or +/// [`BTreeMap::keys`] produce their items in order by key, and take worst-case logarithmic and +/// amortized constant time per item returned. +/// +/// [B-Tree]: https://en.wikipedia.org/wiki/B-tree +/// [`Cell`]: core::cell::Cell +/// [`RefCell`]: core::cell::RefCell +/// +/// # Examples +/// +/// ``` +/// use std::collections::BTreeMap; +/// +/// // type inference lets us omit an explicit type signature (which +/// // would be `BTreeMap<&str, &str>` in this example). +/// let mut movie_reviews = BTreeMap::new(); +/// +/// // review some movies. +/// movie_reviews.insert("Office Space", "Deals with real issues in the workplace."); +/// movie_reviews.insert("Pulp Fiction", "Masterpiece."); +/// movie_reviews.insert("The Godfather", "Very enjoyable."); +/// movie_reviews.insert("The Blues Brothers", "Eye lyked it a lot."); +/// +/// // check for a specific one. +/// if !movie_reviews.contains_key("Les Misérables") { +/// println!("We've got {} reviews, but Les Misérables ain't one.", +/// movie_reviews.len()); +/// } +/// +/// // oops, this review has a lot of spelling mistakes, let's delete it. +/// movie_reviews.remove("The Blues Brothers"); +/// +/// // look up the values associated with some keys. +/// let to_find = ["Up!", "Office Space"]; +/// for movie in &to_find { +/// match movie_reviews.get(movie) { +/// Some(review) => println!("{movie}: {review}"), +/// None => println!("{movie} is unreviewed.") +/// } +/// } +/// +/// // Look up the value for a key (will panic if the key is not found). +/// println!("Movie review: {}", movie_reviews["Office Space"]); +/// +/// // iterate over everything. +/// for (movie, review) in &movie_reviews { +/// println!("{movie}: \"{review}\""); +/// } +/// ``` +/// +/// A `BTreeMap` with a known list of items can be initialized from an array: +/// +/// ``` +/// use std::collections::BTreeMap; +/// +/// let solar_distance = BTreeMap::from([ +/// ("Mercury", 0.4), +/// ("Venus", 0.7), +/// ("Earth", 1.0), +/// ("Mars", 1.5), +/// ]); +/// ``` +/// +/// `BTreeMap` implements an [`Entry API`], which allows for complex +/// methods of getting, setting, updating and removing keys and their values: +/// +/// [`Entry API`]: BTreeMap::entry +/// +/// ``` +/// use std::collections::BTreeMap; +/// +/// // type inference lets us omit an explicit type signature (which +/// // would be `BTreeMap<&str, u8>` in this example). +/// let mut player_stats = BTreeMap::new(); +/// +/// fn random_stat_buff() -> u8 { +/// // could actually return some random value here - let's just return +/// // some fixed value for now +/// 42 +/// } +/// +/// // insert a key only if it doesn't already exist +/// player_stats.entry("health").or_insert(100); +/// +/// // insert a key using a function that provides a new value only if it +/// // doesn't already exist +/// player_stats.entry("defence").or_insert_with(random_stat_buff); +/// +/// // update a key, guarding against the key possibly not being set +/// let stat = player_stats.entry("attack").or_insert(100); +/// *stat += random_stat_buff(); +/// +/// // modify an entry before an insert with in-place mutation +/// player_stats.entry("mana").and_modify(|mana| *mana += 200).or_insert(100); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "BTreeMap")] +#[rustc_insignificant_dtor] +pub struct BTreeMap< + K, + V, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + root: Option<Root<K, V>>, + length: usize, + /// `ManuallyDrop` to control drop order (needs to be dropped after all the nodes). + pub(super) alloc: ManuallyDrop<A>, + // For dropck; the `Box` avoids making the `Unpin` impl more strict than before + _marker: PhantomData<crate::boxed::Box<(K, V)>>, +} + +#[stable(feature = "btree_drop", since = "1.7.0")] +unsafe impl<#[may_dangle] K, #[may_dangle] V, A: Allocator + Clone> Drop for BTreeMap<K, V, A> { + fn drop(&mut self) { + drop(unsafe { ptr::read(self) }.into_iter()) + } +} + +// FIXME: This implementation is "wrong", but changing it would be a breaking change. +// (The bounds of the automatic `UnwindSafe` implementation have been like this since Rust 1.50.) +// Maybe we can fix it nonetheless with a crater run, or if the `UnwindSafe` +// traits are deprecated, or disarmed (no longer causing hard errors) in the future. +#[stable(feature = "btree_unwindsafe", since = "1.64.0")] +impl<K, V, A: Allocator + Clone> core::panic::UnwindSafe for BTreeMap<K, V, A> +where + A: core::panic::UnwindSafe, + K: core::panic::RefUnwindSafe, + V: core::panic::RefUnwindSafe, +{ +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Clone, V: Clone, A: Allocator + Clone> Clone for BTreeMap<K, V, A> { + fn clone(&self) -> BTreeMap<K, V, A> { + fn clone_subtree<'a, K: Clone, V: Clone, A: Allocator + Clone>( + node: NodeRef<marker::Immut<'a>, K, V, marker::LeafOrInternal>, + alloc: A, + ) -> BTreeMap<K, V, A> + where + K: 'a, + V: 'a, + { + match node.force() { + Leaf(leaf) => { + let mut out_tree = BTreeMap { + root: Some(Root::new(alloc.clone())), + length: 0, + alloc: ManuallyDrop::new(alloc), + _marker: PhantomData, + }; + + { + let root = out_tree.root.as_mut().unwrap(); // unwrap succeeds because we just wrapped + let mut out_node = match root.borrow_mut().force() { + Leaf(leaf) => leaf, + Internal(_) => unreachable!(), + }; + + let mut in_edge = leaf.first_edge(); + while let Ok(kv) = in_edge.right_kv() { + let (k, v) = kv.into_kv(); + in_edge = kv.right_edge(); + + out_node.push(k.clone(), v.clone()); + out_tree.length += 1; + } + } + + out_tree + } + Internal(internal) => { + let mut out_tree = + clone_subtree(internal.first_edge().descend(), alloc.clone()); + + { + let out_root = out_tree.root.as_mut().unwrap(); + let mut out_node = out_root.push_internal_level(alloc.clone()); + let mut in_edge = internal.first_edge(); + while let Ok(kv) = in_edge.right_kv() { + let (k, v) = kv.into_kv(); + in_edge = kv.right_edge(); + + let k = (*k).clone(); + let v = (*v).clone(); + let subtree = clone_subtree(in_edge.descend(), alloc.clone()); + + // We can't destructure subtree directly + // because BTreeMap implements Drop + let (subroot, sublength) = unsafe { + let subtree = ManuallyDrop::new(subtree); + let root = ptr::read(&subtree.root); + let length = subtree.length; + (root, length) + }; + + out_node.push( + k, + v, + subroot.unwrap_or_else(|| Root::new(alloc.clone())), + ); + out_tree.length += 1 + sublength; + } + } + + out_tree + } + } + } + + if self.is_empty() { + BTreeMap::new_in((*self.alloc).clone()) + } else { + clone_subtree(self.root.as_ref().unwrap().reborrow(), (*self.alloc).clone()) // unwrap succeeds because not empty + } + } +} + +impl<K, Q: ?Sized, A: Allocator + Clone> super::Recover<Q> for BTreeMap<K, SetValZST, A> +where + K: Borrow<Q> + Ord, + Q: Ord, +{ + type Key = K; + + fn get(&self, key: &Q) -> Option<&K> { + let root_node = self.root.as_ref()?.reborrow(); + match root_node.search_tree(key) { + Found(handle) => Some(handle.into_kv().0), + GoDown(_) => None, + } + } + + fn take(&mut self, key: &Q) -> Option<K> { + let (map, dormant_map) = DormantMutRef::new(self); + let root_node = map.root.as_mut()?.borrow_mut(); + match root_node.search_tree(key) { + Found(handle) => Some( + OccupiedEntry { + handle, + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + } + .remove_kv() + .0, + ), + GoDown(_) => None, + } + } + + fn replace(&mut self, key: K) -> Option<K> { + let (map, dormant_map) = DormantMutRef::new(self); + let root_node = + map.root.get_or_insert_with(|| Root::new((*map.alloc).clone())).borrow_mut(); + match root_node.search_tree::<K>(&key) { + Found(mut kv) => Some(mem::replace(kv.key_mut(), key)), + GoDown(handle) => { + VacantEntry { + key, + handle: Some(handle), + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + } + .insert(SetValZST::default()); + None + } + } + } +} + +/// An iterator over the entries of a `BTreeMap`. +/// +/// This `struct` is created by the [`iter`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`iter`]: BTreeMap::iter +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Iter<'a, K: 'a, V: 'a> { + range: LazyLeafRange<marker::Immut<'a>, K, V>, + length: usize, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for Iter<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A mutable iterator over the entries of a `BTreeMap`. +/// +/// This `struct` is created by the [`iter_mut`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`iter_mut`]: BTreeMap::iter_mut +#[stable(feature = "rust1", since = "1.0.0")] +pub struct IterMut<'a, K: 'a, V: 'a> { + range: LazyLeafRange<marker::ValMut<'a>, K, V>, + length: usize, + + // Be invariant in `K` and `V` + _marker: PhantomData<&'a mut (K, V)>, +} + +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for IterMut<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let range = Iter { range: self.range.reborrow(), length: self.length }; + f.debug_list().entries(range).finish() + } +} + +/// An owning iterator over the entries of a `BTreeMap`. +/// +/// This `struct` is created by the [`into_iter`] method on [`BTreeMap`] +/// (provided by the [`IntoIterator`] trait). See its documentation for more. +/// +/// [`into_iter`]: IntoIterator::into_iter +/// [`IntoIterator`]: core::iter::IntoIterator +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_insignificant_dtor] +pub struct IntoIter< + K, + V, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + range: LazyLeafRange<marker::Dying, K, V>, + length: usize, + /// The BTreeMap will outlive this IntoIter so we don't care about drop order for `alloc`. + alloc: A, +} + +impl<K, V, A: Allocator + Clone> IntoIter<K, V, A> { + /// Returns an iterator of references over the remaining items. + #[inline] + pub(super) fn iter(&self) -> Iter<'_, K, V> { + Iter { range: self.range.reborrow(), length: self.length } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: Debug, V: Debug, A: Allocator + Clone> Debug for IntoIter<K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.iter()).finish() + } +} + +/// An iterator over the keys of a `BTreeMap`. +/// +/// This `struct` is created by the [`keys`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`keys`]: BTreeMap::keys +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Keys<'a, K, V> { + inner: Iter<'a, K, V>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: fmt::Debug, V> fmt::Debug for Keys<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// An iterator over the values of a `BTreeMap`. +/// +/// This `struct` is created by the [`values`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`values`]: BTreeMap::values +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Values<'a, K, V> { + inner: Iter<'a, K, V>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K, V: fmt::Debug> fmt::Debug for Values<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A mutable iterator over the values of a `BTreeMap`. +/// +/// This `struct` is created by the [`values_mut`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`values_mut`]: BTreeMap::values_mut +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "map_values_mut", since = "1.10.0")] +pub struct ValuesMut<'a, K, V> { + inner: IterMut<'a, K, V>, +} + +#[stable(feature = "map_values_mut", since = "1.10.0")] +impl<K, V: fmt::Debug> fmt::Debug for ValuesMut<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.inner.iter().map(|(_, val)| val)).finish() + } +} + +/// An owning iterator over the keys of a `BTreeMap`. +/// +/// This `struct` is created by the [`into_keys`] method on [`BTreeMap`]. +/// See its documentation for more. +/// +/// [`into_keys`]: BTreeMap::into_keys +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +pub struct IntoKeys<K, V, A: Allocator + Clone = Global> { + inner: IntoIter<K, V, A>, +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K: fmt::Debug, V, A: Allocator + Clone> fmt::Debug for IntoKeys<K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.inner.iter().map(|(key, _)| key)).finish() + } +} + +/// An owning iterator over the values of a `BTreeMap`. +/// +/// This `struct` is created by the [`into_values`] method on [`BTreeMap`]. +/// See its documentation for more. +/// +/// [`into_values`]: BTreeMap::into_values +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +pub struct IntoValues< + K, + V, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + inner: IntoIter<K, V, A>, +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V: fmt::Debug, A: Allocator + Clone> fmt::Debug for IntoValues<K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.inner.iter().map(|(_, val)| val)).finish() + } +} + +/// An iterator over a sub-range of entries in a `BTreeMap`. +/// +/// This `struct` is created by the [`range`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`range`]: BTreeMap::range +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "btree_range", since = "1.17.0")] +pub struct Range<'a, K: 'a, V: 'a> { + inner: LeafRange<marker::Immut<'a>, K, V>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for Range<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self.clone()).finish() + } +} + +/// A mutable iterator over a sub-range of entries in a `BTreeMap`. +/// +/// This `struct` is created by the [`range_mut`] method on [`BTreeMap`]. See its +/// documentation for more. +/// +/// [`range_mut`]: BTreeMap::range_mut +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "btree_range", since = "1.17.0")] +pub struct RangeMut<'a, K: 'a, V: 'a> { + inner: LeafRange<marker::ValMut<'a>, K, V>, + + // Be invariant in `K` and `V` + _marker: PhantomData<&'a mut (K, V)>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<K: fmt::Debug, V: fmt::Debug> fmt::Debug for RangeMut<'_, K, V> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let range = Range { inner: self.inner.reborrow() }; + f.debug_list().entries(range).finish() + } +} + +impl<K, V> BTreeMap<K, V> { + /// Makes a new, empty `BTreeMap`. + /// + /// Does not allocate anything on its own. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// + /// // entries can now be inserted into the empty map + /// map.insert(1, "a"); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + #[must_use] + pub const fn new() -> BTreeMap<K, V> { + BTreeMap { root: None, length: 0, alloc: ManuallyDrop::new(Global), _marker: PhantomData } + } +} + +impl<K, V, A: Allocator + Clone> BTreeMap<K, V, A> { + /// Clears the map, removing all elements. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, "a"); + /// a.clear(); + /// assert!(a.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) { + // avoid moving the allocator + mem::drop(BTreeMap { + root: mem::replace(&mut self.root, None), + length: mem::replace(&mut self.length, 0), + alloc: self.alloc.clone(), + _marker: PhantomData, + }); + } + + /// Makes a new empty BTreeMap with a reasonable choice for B. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// # #![feature(allocator_api)] + /// # #![feature(btreemap_alloc)] + /// use std::collections::BTreeMap; + /// use std::alloc::Global; + /// + /// let mut map = BTreeMap::new_in(Global); + /// + /// // entries can now be inserted into the empty map + /// map.insert(1, "a"); + /// ``` + #[unstable(feature = "btreemap_alloc", issue = "32838")] + pub fn new_in(alloc: A) -> BTreeMap<K, V, A> { + BTreeMap { root: None, length: 0, alloc: ManuallyDrop::new(alloc), _marker: PhantomData } + } +} + +impl<K, V, A: Allocator + Clone> BTreeMap<K, V, A> { + /// Returns a reference to the value corresponding to the key. + /// + /// The key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// assert_eq!(map.get(&1), Some(&"a")); + /// assert_eq!(map.get(&2), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get<Q: ?Sized>(&self, key: &Q) -> Option<&V> + where + K: Borrow<Q> + Ord, + Q: Ord, + { + let root_node = self.root.as_ref()?.reborrow(); + match root_node.search_tree(key) { + Found(handle) => Some(handle.into_kv().1), + GoDown(_) => None, + } + } + + /// Returns the key-value pair corresponding to the supplied key. + /// + /// The supplied key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// assert_eq!(map.get_key_value(&1), Some((&1, &"a"))); + /// assert_eq!(map.get_key_value(&2), None); + /// ``` + #[stable(feature = "map_get_key_value", since = "1.40.0")] + pub fn get_key_value<Q: ?Sized>(&self, k: &Q) -> Option<(&K, &V)> + where + K: Borrow<Q> + Ord, + Q: Ord, + { + let root_node = self.root.as_ref()?.reborrow(); + match root_node.search_tree(k) { + Found(handle) => Some(handle.into_kv()), + GoDown(_) => None, + } + } + + /// Returns the first key-value pair in the map. + /// The key in this pair is the minimum key in the map. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// assert_eq!(map.first_key_value(), None); + /// map.insert(1, "b"); + /// map.insert(2, "a"); + /// assert_eq!(map.first_key_value(), Some((&1, &"b"))); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn first_key_value(&self) -> Option<(&K, &V)> + where + K: Ord, + { + let root_node = self.root.as_ref()?.reborrow(); + root_node.first_leaf_edge().right_kv().ok().map(Handle::into_kv) + } + + /// Returns the first entry in the map for in-place manipulation. + /// The key of this entry is the minimum key in the map. + /// + /// # Examples + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// map.insert(2, "b"); + /// if let Some(mut entry) = map.first_entry() { + /// if *entry.key() > 0 { + /// entry.insert("first"); + /// } + /// } + /// assert_eq!(*map.get(&1).unwrap(), "first"); + /// assert_eq!(*map.get(&2).unwrap(), "b"); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn first_entry(&mut self) -> Option<OccupiedEntry<'_, K, V, A>> + where + K: Ord, + { + let (map, dormant_map) = DormantMutRef::new(self); + let root_node = map.root.as_mut()?.borrow_mut(); + let kv = root_node.first_leaf_edge().right_kv().ok()?; + Some(OccupiedEntry { + handle: kv.forget_node_type(), + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + }) + } + + /// Removes and returns the first element in the map. + /// The key of this element is the minimum key that was in the map. + /// + /// # Examples + /// + /// Draining elements in ascending order, while keeping a usable map each iteration. + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// map.insert(2, "b"); + /// while let Some((key, _val)) = map.pop_first() { + /// assert!(map.iter().all(|(k, _v)| *k > key)); + /// } + /// assert!(map.is_empty()); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn pop_first(&mut self) -> Option<(K, V)> + where + K: Ord, + { + self.first_entry().map(|entry| entry.remove_entry()) + } + + /// Returns the last key-value pair in the map. + /// The key in this pair is the maximum key in the map. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "b"); + /// map.insert(2, "a"); + /// assert_eq!(map.last_key_value(), Some((&2, &"a"))); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn last_key_value(&self) -> Option<(&K, &V)> + where + K: Ord, + { + let root_node = self.root.as_ref()?.reborrow(); + root_node.last_leaf_edge().left_kv().ok().map(Handle::into_kv) + } + + /// Returns the last entry in the map for in-place manipulation. + /// The key of this entry is the maximum key in the map. + /// + /// # Examples + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// map.insert(2, "b"); + /// if let Some(mut entry) = map.last_entry() { + /// if *entry.key() > 0 { + /// entry.insert("last"); + /// } + /// } + /// assert_eq!(*map.get(&1).unwrap(), "a"); + /// assert_eq!(*map.get(&2).unwrap(), "last"); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn last_entry(&mut self) -> Option<OccupiedEntry<'_, K, V, A>> + where + K: Ord, + { + let (map, dormant_map) = DormantMutRef::new(self); + let root_node = map.root.as_mut()?.borrow_mut(); + let kv = root_node.last_leaf_edge().left_kv().ok()?; + Some(OccupiedEntry { + handle: kv.forget_node_type(), + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + }) + } + + /// Removes and returns the last element in the map. + /// The key of this element is the maximum key that was in the map. + /// + /// # Examples + /// + /// Draining elements in descending order, while keeping a usable map each iteration. + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// map.insert(2, "b"); + /// while let Some((key, _val)) = map.pop_last() { + /// assert!(map.iter().all(|(k, _v)| *k < key)); + /// } + /// assert!(map.is_empty()); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn pop_last(&mut self) -> Option<(K, V)> + where + K: Ord, + { + self.last_entry().map(|entry| entry.remove_entry()) + } + + /// Returns `true` if the map contains a value for the specified key. + /// + /// The key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// assert_eq!(map.contains_key(&1), true); + /// assert_eq!(map.contains_key(&2), false); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn contains_key<Q: ?Sized>(&self, key: &Q) -> bool + where + K: Borrow<Q> + Ord, + Q: Ord, + { + self.get(key).is_some() + } + + /// Returns a mutable reference to the value corresponding to the key. + /// + /// The key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// if let Some(x) = map.get_mut(&1) { + /// *x = "b"; + /// } + /// assert_eq!(map[&1], "b"); + /// ``` + // See `get` for implementation notes, this is basically a copy-paste with mut's added + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get_mut<Q: ?Sized>(&mut self, key: &Q) -> Option<&mut V> + where + K: Borrow<Q> + Ord, + Q: Ord, + { + let root_node = self.root.as_mut()?.borrow_mut(); + match root_node.search_tree(key) { + Found(handle) => Some(handle.into_val_mut()), + GoDown(_) => None, + } + } + + /// Inserts a key-value pair into the map. + /// + /// If the map did not have this key present, `None` is returned. + /// + /// If the map did have this key present, the value is updated, and the old + /// value is returned. The key is not updated, though; this matters for + /// types that can be `==` without being identical. See the [module-level + /// documentation] for more. + /// + /// [module-level documentation]: index.html#insert-and-complex-keys + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// assert_eq!(map.insert(37, "a"), None); + /// assert_eq!(map.is_empty(), false); + /// + /// map.insert(37, "b"); + /// assert_eq!(map.insert(37, "c"), Some("b")); + /// assert_eq!(map[&37], "c"); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(&mut self, key: K, value: V) -> Option<V> + where + K: Ord, + { + match self.entry(key) { + Occupied(mut entry) => Some(entry.insert(value)), + Vacant(entry) => { + entry.insert(value); + None + } + } + } + + /// Tries to insert a key-value pair into the map, and returns + /// a mutable reference to the value in the entry. + /// + /// If the map already had this key present, nothing is updated, and + /// an error containing the occupied entry and the value is returned. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(map_try_insert)] + /// + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// assert_eq!(map.try_insert(37, "a").unwrap(), &"a"); + /// + /// let err = map.try_insert(37, "b").unwrap_err(); + /// assert_eq!(err.entry.key(), &37); + /// assert_eq!(err.entry.get(), &"a"); + /// assert_eq!(err.value, "b"); + /// ``` + #[unstable(feature = "map_try_insert", issue = "82766")] + pub fn try_insert(&mut self, key: K, value: V) -> Result<&mut V, OccupiedError<'_, K, V, A>> + where + K: Ord, + { + match self.entry(key) { + Occupied(entry) => Err(OccupiedError { entry, value }), + Vacant(entry) => Ok(entry.insert(value)), + } + } + + /// Removes a key from the map, returning the value at the key if the key + /// was previously in the map. + /// + /// The key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// assert_eq!(map.remove(&1), Some("a")); + /// assert_eq!(map.remove(&1), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn remove<Q: ?Sized>(&mut self, key: &Q) -> Option<V> + where + K: Borrow<Q> + Ord, + Q: Ord, + { + self.remove_entry(key).map(|(_, v)| v) + } + + /// Removes a key from the map, returning the stored key and value if the key + /// was previously in the map. + /// + /// The key may be any borrowed form of the map's key type, but the ordering + /// on the borrowed form *must* match the ordering on the key type. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(1, "a"); + /// assert_eq!(map.remove_entry(&1), Some((1, "a"))); + /// assert_eq!(map.remove_entry(&1), None); + /// ``` + #[stable(feature = "btreemap_remove_entry", since = "1.45.0")] + pub fn remove_entry<Q: ?Sized>(&mut self, key: &Q) -> Option<(K, V)> + where + K: Borrow<Q> + Ord, + Q: Ord, + { + let (map, dormant_map) = DormantMutRef::new(self); + let root_node = map.root.as_mut()?.borrow_mut(); + match root_node.search_tree(key) { + Found(handle) => Some( + OccupiedEntry { + handle, + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + } + .remove_entry(), + ), + GoDown(_) => None, + } + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all pairs `(k, v)` for which `f(&k, &mut v)` returns `false`. + /// The elements are visited in ascending key order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<i32, i32> = (0..8).map(|x| (x, x*10)).collect(); + /// // Keep only the elements with even-numbered keys. + /// map.retain(|&k, _| k % 2 == 0); + /// assert!(map.into_iter().eq(vec![(0, 0), (2, 20), (4, 40), (6, 60)])); + /// ``` + #[inline] + #[stable(feature = "btree_retain", since = "1.53.0")] + pub fn retain<F>(&mut self, mut f: F) + where + K: Ord, + F: FnMut(&K, &mut V) -> bool, + { + self.drain_filter(|k, v| !f(k, v)); + } + + /// Moves all elements from `other` into `self`, leaving `other` empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, "a"); + /// a.insert(2, "b"); + /// a.insert(3, "c"); + /// + /// let mut b = BTreeMap::new(); + /// b.insert(3, "d"); + /// b.insert(4, "e"); + /// b.insert(5, "f"); + /// + /// a.append(&mut b); + /// + /// assert_eq!(a.len(), 5); + /// assert_eq!(b.len(), 0); + /// + /// assert_eq!(a[&1], "a"); + /// assert_eq!(a[&2], "b"); + /// assert_eq!(a[&3], "d"); + /// assert_eq!(a[&4], "e"); + /// assert_eq!(a[&5], "f"); + /// ``` + #[stable(feature = "btree_append", since = "1.11.0")] + pub fn append(&mut self, other: &mut Self) + where + K: Ord, + A: Clone, + { + // Do we have to append anything at all? + if other.is_empty() { + return; + } + + // We can just swap `self` and `other` if `self` is empty. + if self.is_empty() { + mem::swap(self, other); + return; + } + + let self_iter = mem::replace(self, Self::new_in((*self.alloc).clone())).into_iter(); + let other_iter = mem::replace(other, Self::new_in((*self.alloc).clone())).into_iter(); + let root = self.root.get_or_insert_with(|| Root::new((*self.alloc).clone())); + root.append_from_sorted_iters( + self_iter, + other_iter, + &mut self.length, + (*self.alloc).clone(), + ) + } + + /// Constructs a double-ended iterator over a sub-range of elements in the map. + /// The simplest way is to use the range syntax `min..max`, thus `range(min..max)` will + /// yield elements from min (inclusive) to max (exclusive). + /// The range may also be entered as `(Bound<T>, Bound<T>)`, so for example + /// `range((Excluded(4), Included(10)))` will yield a left-exclusive, right-inclusive + /// range from 4 to 10. + /// + /// # Panics + /// + /// Panics if range `start > end`. + /// Panics if range `start == end` and both bounds are `Excluded`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::ops::Bound::Included; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(3, "a"); + /// map.insert(5, "b"); + /// map.insert(8, "c"); + /// for (&key, &value) in map.range((Included(&4), Included(&8))) { + /// println!("{key}: {value}"); + /// } + /// assert_eq!(Some((&5, &"b")), map.range(4..).next()); + /// ``` + #[stable(feature = "btree_range", since = "1.17.0")] + pub fn range<T: ?Sized, R>(&self, range: R) -> Range<'_, K, V> + where + T: Ord, + K: Borrow<T> + Ord, + R: RangeBounds<T>, + { + if let Some(root) = &self.root { + Range { inner: root.reborrow().range_search(range) } + } else { + Range { inner: LeafRange::none() } + } + } + + /// Constructs a mutable double-ended iterator over a sub-range of elements in the map. + /// The simplest way is to use the range syntax `min..max`, thus `range(min..max)` will + /// yield elements from min (inclusive) to max (exclusive). + /// The range may also be entered as `(Bound<T>, Bound<T>)`, so for example + /// `range((Excluded(4), Included(10)))` will yield a left-exclusive, right-inclusive + /// range from 4 to 10. + /// + /// # Panics + /// + /// Panics if range `start > end`. + /// Panics if range `start == end` and both bounds are `Excluded`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, i32> = + /// [("Alice", 0), ("Bob", 0), ("Carol", 0), ("Cheryl", 0)].into(); + /// for (_, balance) in map.range_mut("B".."Cheryl") { + /// *balance += 100; + /// } + /// for (name, balance) in &map { + /// println!("{name} => {balance}"); + /// } + /// ``` + #[stable(feature = "btree_range", since = "1.17.0")] + pub fn range_mut<T: ?Sized, R>(&mut self, range: R) -> RangeMut<'_, K, V> + where + T: Ord, + K: Borrow<T> + Ord, + R: RangeBounds<T>, + { + if let Some(root) = &mut self.root { + RangeMut { inner: root.borrow_valmut().range_search(range), _marker: PhantomData } + } else { + RangeMut { inner: LeafRange::none(), _marker: PhantomData } + } + } + + /// Gets the given key's corresponding entry in the map for in-place manipulation. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut count: BTreeMap<&str, usize> = BTreeMap::new(); + /// + /// // count the number of occurrences of letters in the vec + /// for x in ["a", "b", "a", "c", "a", "b"] { + /// count.entry(x).and_modify(|curr| *curr += 1).or_insert(1); + /// } + /// + /// assert_eq!(count["a"], 3); + /// assert_eq!(count["b"], 2); + /// assert_eq!(count["c"], 1); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn entry(&mut self, key: K) -> Entry<'_, K, V, A> + where + K: Ord, + { + let (map, dormant_map) = DormantMutRef::new(self); + match map.root { + None => Vacant(VacantEntry { + key, + handle: None, + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + }), + Some(ref mut root) => match root.borrow_mut().search_tree(&key) { + Found(handle) => Occupied(OccupiedEntry { + handle, + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + }), + GoDown(handle) => Vacant(VacantEntry { + key, + handle: Some(handle), + dormant_map, + alloc: (*map.alloc).clone(), + _marker: PhantomData, + }), + }, + } + } + + /// Splits the collection into two at the given key. Returns everything after the given key, + /// including the key. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, "a"); + /// a.insert(2, "b"); + /// a.insert(3, "c"); + /// a.insert(17, "d"); + /// a.insert(41, "e"); + /// + /// let b = a.split_off(&3); + /// + /// assert_eq!(a.len(), 2); + /// assert_eq!(b.len(), 3); + /// + /// assert_eq!(a[&1], "a"); + /// assert_eq!(a[&2], "b"); + /// + /// assert_eq!(b[&3], "c"); + /// assert_eq!(b[&17], "d"); + /// assert_eq!(b[&41], "e"); + /// ``` + #[stable(feature = "btree_split_off", since = "1.11.0")] + pub fn split_off<Q: ?Sized + Ord>(&mut self, key: &Q) -> Self + where + K: Borrow<Q> + Ord, + A: Clone, + { + if self.is_empty() { + return Self::new_in((*self.alloc).clone()); + } + + let total_num = self.len(); + let left_root = self.root.as_mut().unwrap(); // unwrap succeeds because not empty + + let right_root = left_root.split_off(key, (*self.alloc).clone()); + + let (new_left_len, right_len) = Root::calc_split_length(total_num, &left_root, &right_root); + self.length = new_left_len; + + BTreeMap { + root: Some(right_root), + length: right_len, + alloc: self.alloc.clone(), + _marker: PhantomData, + } + } + + /// Creates an iterator that visits all elements (key-value pairs) in + /// ascending key order and uses a closure to determine if an element should + /// be removed. If the closure returns `true`, the element is removed from + /// the map and yielded. If the closure returns `false`, or panics, the + /// element remains in the map and will not be yielded. + /// + /// The iterator also lets you mutate the value of each element in the + /// closure, regardless of whether you choose to keep or remove it. + /// + /// If the iterator is only partially consumed or not consumed at all, each + /// of the remaining elements is still subjected to the closure, which may + /// change its value and, by returning `true`, have the element removed and + /// dropped. + /// + /// It is unspecified how many more elements will be subjected to the + /// closure if a panic occurs in the closure, or a panic occurs while + /// dropping an element, or if the `DrainFilter` value is leaked. + /// + /// # Examples + /// + /// Splitting a map into even and odd keys, reusing the original map: + /// + /// ``` + /// #![feature(btree_drain_filter)] + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<i32, i32> = (0..8).map(|x| (x, x)).collect(); + /// let evens: BTreeMap<_, _> = map.drain_filter(|k, _v| k % 2 == 0).collect(); + /// let odds = map; + /// assert_eq!(evens.keys().copied().collect::<Vec<_>>(), [0, 2, 4, 6]); + /// assert_eq!(odds.keys().copied().collect::<Vec<_>>(), [1, 3, 5, 7]); + /// ``` + #[unstable(feature = "btree_drain_filter", issue = "70530")] + pub fn drain_filter<F>(&mut self, pred: F) -> DrainFilter<'_, K, V, F, A> + where + K: Ord, + F: FnMut(&K, &mut V) -> bool, + { + let (inner, alloc) = self.drain_filter_inner(); + DrainFilter { pred, inner, alloc } + } + + pub(super) fn drain_filter_inner(&mut self) -> (DrainFilterInner<'_, K, V>, A) + where + K: Ord, + { + if let Some(root) = self.root.as_mut() { + let (root, dormant_root) = DormantMutRef::new(root); + let front = root.borrow_mut().first_leaf_edge(); + ( + DrainFilterInner { + length: &mut self.length, + dormant_root: Some(dormant_root), + cur_leaf_edge: Some(front), + }, + (*self.alloc).clone(), + ) + } else { + ( + DrainFilterInner { + length: &mut self.length, + dormant_root: None, + cur_leaf_edge: None, + }, + (*self.alloc).clone(), + ) + } + } + + /// Creates a consuming iterator visiting all the keys, in sorted order. + /// The map cannot be used after calling this. + /// The iterator element type is `K`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(2, "b"); + /// a.insert(1, "a"); + /// + /// let keys: Vec<i32> = a.into_keys().collect(); + /// assert_eq!(keys, [1, 2]); + /// ``` + #[inline] + #[stable(feature = "map_into_keys_values", since = "1.54.0")] + pub fn into_keys(self) -> IntoKeys<K, V, A> { + IntoKeys { inner: self.into_iter() } + } + + /// Creates a consuming iterator visiting all the values, in order by key. + /// The map cannot be used after calling this. + /// The iterator element type is `V`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, "hello"); + /// a.insert(2, "goodbye"); + /// + /// let values: Vec<&str> = a.into_values().collect(); + /// assert_eq!(values, ["hello", "goodbye"]); + /// ``` + #[inline] + #[stable(feature = "map_into_keys_values", since = "1.54.0")] + pub fn into_values(self) -> IntoValues<K, V, A> { + IntoValues { inner: self.into_iter() } + } + + /// Makes a `BTreeMap` from a sorted iterator. + pub(crate) fn bulk_build_from_sorted_iter<I>(iter: I, alloc: A) -> Self + where + K: Ord, + I: IntoIterator<Item = (K, V)>, + { + let mut root = Root::new(alloc.clone()); + let mut length = 0; + root.bulk_push(DedupSortedIter::new(iter.into_iter()), &mut length, alloc.clone()); + BTreeMap { root: Some(root), length, alloc: ManuallyDrop::new(alloc), _marker: PhantomData } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V, A: Allocator + Clone> IntoIterator for &'a BTreeMap<K, V, A> { + type Item = (&'a K, &'a V); + type IntoIter = Iter<'a, K, V>; + + fn into_iter(self) -> Iter<'a, K, V> { + self.iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K: 'a, V: 'a> Iterator for Iter<'a, K, V> { + type Item = (&'a K, &'a V); + + fn next(&mut self) -> Option<(&'a K, &'a V)> { + if self.length == 0 { + None + } else { + self.length -= 1; + Some(unsafe { self.range.next_unchecked() }) + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (self.length, Some(self.length)) + } + + fn last(mut self) -> Option<(&'a K, &'a V)> { + self.next_back() + } + + fn min(mut self) -> Option<(&'a K, &'a V)> { + self.next() + } + + fn max(mut self) -> Option<(&'a K, &'a V)> { + self.next_back() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for Iter<'_, K, V> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K: 'a, V: 'a> DoubleEndedIterator for Iter<'a, K, V> { + fn next_back(&mut self) -> Option<(&'a K, &'a V)> { + if self.length == 0 { + None + } else { + self.length -= 1; + Some(unsafe { self.range.next_back_unchecked() }) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> ExactSizeIterator for Iter<'_, K, V> { + fn len(&self) -> usize { + self.length + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> Clone for Iter<'_, K, V> { + fn clone(&self) -> Self { + Iter { range: self.range.clone(), length: self.length } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V, A: Allocator + Clone> IntoIterator for &'a mut BTreeMap<K, V, A> { + type Item = (&'a K, &'a mut V); + type IntoIter = IterMut<'a, K, V>; + + fn into_iter(self) -> IterMut<'a, K, V> { + self.iter_mut() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> Iterator for IterMut<'a, K, V> { + type Item = (&'a K, &'a mut V); + + fn next(&mut self) -> Option<(&'a K, &'a mut V)> { + if self.length == 0 { + None + } else { + self.length -= 1; + Some(unsafe { self.range.next_unchecked() }) + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (self.length, Some(self.length)) + } + + fn last(mut self) -> Option<(&'a K, &'a mut V)> { + self.next_back() + } + + fn min(mut self) -> Option<(&'a K, &'a mut V)> { + self.next() + } + + fn max(mut self) -> Option<(&'a K, &'a mut V)> { + self.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> DoubleEndedIterator for IterMut<'a, K, V> { + fn next_back(&mut self) -> Option<(&'a K, &'a mut V)> { + if self.length == 0 { + None + } else { + self.length -= 1; + Some(unsafe { self.range.next_back_unchecked() }) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> ExactSizeIterator for IterMut<'_, K, V> { + fn len(&self) -> usize { + self.length + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for IterMut<'_, K, V> {} + +impl<'a, K, V> IterMut<'a, K, V> { + /// Returns an iterator of references over the remaining items. + #[inline] + pub(super) fn iter(&self) -> Iter<'_, K, V> { + Iter { range: self.range.reborrow(), length: self.length } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V, A: Allocator + Clone> IntoIterator for BTreeMap<K, V, A> { + type Item = (K, V); + type IntoIter = IntoIter<K, V, A>; + + fn into_iter(self) -> IntoIter<K, V, A> { + let mut me = ManuallyDrop::new(self); + if let Some(root) = me.root.take() { + let full_range = root.into_dying().full_range(); + + IntoIter { + range: full_range, + length: me.length, + alloc: unsafe { ManuallyDrop::take(&mut me.alloc) }, + } + } else { + IntoIter { + range: LazyLeafRange::none(), + length: 0, + alloc: unsafe { ManuallyDrop::take(&mut me.alloc) }, + } + } + } +} + +#[stable(feature = "btree_drop", since = "1.7.0")] +impl<K, V, A: Allocator + Clone> Drop for IntoIter<K, V, A> { + fn drop(&mut self) { + struct DropGuard<'a, K, V, A: Allocator + Clone>(&'a mut IntoIter<K, V, A>); + + impl<'a, K, V, A: Allocator + Clone> Drop for DropGuard<'a, K, V, A> { + fn drop(&mut self) { + // Continue the same loop we perform below. This only runs when unwinding, so we + // don't have to care about panics this time (they'll abort). + while let Some(kv) = self.0.dying_next() { + // SAFETY: we consume the dying handle immediately. + unsafe { kv.drop_key_val() }; + } + } + } + + while let Some(kv) = self.dying_next() { + let guard = DropGuard(self); + // SAFETY: we don't touch the tree before consuming the dying handle. + unsafe { kv.drop_key_val() }; + mem::forget(guard); + } + } +} + +impl<K, V, A: Allocator + Clone> IntoIter<K, V, A> { + /// Core of a `next` method returning a dying KV handle, + /// invalidated by further calls to this function and some others. + fn dying_next( + &mut self, + ) -> Option<Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV>> { + if self.length == 0 { + self.range.deallocating_end(self.alloc.clone()); + None + } else { + self.length -= 1; + Some(unsafe { self.range.deallocating_next_unchecked(self.alloc.clone()) }) + } + } + + /// Core of a `next_back` method returning a dying KV handle, + /// invalidated by further calls to this function and some others. + fn dying_next_back( + &mut self, + ) -> Option<Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV>> { + if self.length == 0 { + self.range.deallocating_end(self.alloc.clone()); + None + } else { + self.length -= 1; + Some(unsafe { self.range.deallocating_next_back_unchecked(self.alloc.clone()) }) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V, A: Allocator + Clone> Iterator for IntoIter<K, V, A> { + type Item = (K, V); + + fn next(&mut self) -> Option<(K, V)> { + // SAFETY: we consume the dying handle immediately. + self.dying_next().map(unsafe { |kv| kv.into_key_val() }) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (self.length, Some(self.length)) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V, A: Allocator + Clone> DoubleEndedIterator for IntoIter<K, V, A> { + fn next_back(&mut self) -> Option<(K, V)> { + // SAFETY: we consume the dying handle immediately. + self.dying_next_back().map(unsafe { |kv| kv.into_key_val() }) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V, A: Allocator + Clone> ExactSizeIterator for IntoIter<K, V, A> { + fn len(&self) -> usize { + self.length + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V, A: Allocator + Clone> FusedIterator for IntoIter<K, V, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> Iterator for Keys<'a, K, V> { + type Item = &'a K; + + fn next(&mut self) -> Option<&'a K> { + self.inner.next().map(|(k, _)| k) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } + + fn last(mut self) -> Option<&'a K> { + self.next_back() + } + + fn min(mut self) -> Option<&'a K> { + self.next() + } + + fn max(mut self) -> Option<&'a K> { + self.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> DoubleEndedIterator for Keys<'a, K, V> { + fn next_back(&mut self) -> Option<&'a K> { + self.inner.next_back().map(|(k, _)| k) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> ExactSizeIterator for Keys<'_, K, V> { + fn len(&self) -> usize { + self.inner.len() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for Keys<'_, K, V> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> Clone for Keys<'_, K, V> { + fn clone(&self) -> Self { + Keys { inner: self.inner.clone() } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> Iterator for Values<'a, K, V> { + type Item = &'a V; + + fn next(&mut self) -> Option<&'a V> { + self.inner.next().map(|(_, v)| v) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } + + fn last(mut self) -> Option<&'a V> { + self.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, K, V> DoubleEndedIterator for Values<'a, K, V> { + fn next_back(&mut self) -> Option<&'a V> { + self.inner.next_back().map(|(_, v)| v) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> ExactSizeIterator for Values<'_, K, V> { + fn len(&self) -> usize { + self.inner.len() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for Values<'_, K, V> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> Clone for Values<'_, K, V> { + fn clone(&self) -> Self { + Values { inner: self.inner.clone() } + } +} + +/// An iterator produced by calling `drain_filter` on BTreeMap. +#[unstable(feature = "btree_drain_filter", issue = "70530")] +pub struct DrainFilter< + 'a, + K, + V, + F, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> where + F: 'a + FnMut(&K, &mut V) -> bool, +{ + pred: F, + inner: DrainFilterInner<'a, K, V>, + /// The BTreeMap will outlive this IntoIter so we don't care about drop order for `alloc`. + alloc: A, +} +/// Most of the implementation of DrainFilter are generic over the type +/// of the predicate, thus also serving for BTreeSet::DrainFilter. +pub(super) struct DrainFilterInner<'a, K, V> { + /// Reference to the length field in the borrowed map, updated live. + length: &'a mut usize, + /// Buried reference to the root field in the borrowed map. + /// Wrapped in `Option` to allow drop handler to `take` it. + dormant_root: Option<DormantMutRef<'a, Root<K, V>>>, + /// Contains a leaf edge preceding the next element to be returned, or the last leaf edge. + /// Empty if the map has no root, if iteration went beyond the last leaf edge, + /// or if a panic occurred in the predicate. + cur_leaf_edge: Option<Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>>, +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<K, V, F, A: Allocator + Clone> Drop for DrainFilter<'_, K, V, F, A> +where + F: FnMut(&K, &mut V) -> bool, +{ + fn drop(&mut self) { + self.for_each(drop); + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<K, V, F> fmt::Debug for DrainFilter<'_, K, V, F> +where + K: fmt::Debug, + V: fmt::Debug, + F: FnMut(&K, &mut V) -> bool, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("DrainFilter").field(&self.inner.peek()).finish() + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<K, V, F, A: Allocator + Clone> Iterator for DrainFilter<'_, K, V, F, A> +where + F: FnMut(&K, &mut V) -> bool, +{ + type Item = (K, V); + + fn next(&mut self) -> Option<(K, V)> { + self.inner.next(&mut self.pred, self.alloc.clone()) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } +} + +impl<'a, K, V> DrainFilterInner<'a, K, V> { + /// Allow Debug implementations to predict the next element. + pub(super) fn peek(&self) -> Option<(&K, &V)> { + let edge = self.cur_leaf_edge.as_ref()?; + edge.reborrow().next_kv().ok().map(Handle::into_kv) + } + + /// Implementation of a typical `DrainFilter::next` method, given the predicate. + pub(super) fn next<F, A: Allocator + Clone>(&mut self, pred: &mut F, alloc: A) -> Option<(K, V)> + where + F: FnMut(&K, &mut V) -> bool, + { + while let Ok(mut kv) = self.cur_leaf_edge.take()?.next_kv() { + let (k, v) = kv.kv_mut(); + if pred(k, v) { + *self.length -= 1; + let (kv, pos) = kv.remove_kv_tracking( + || { + // SAFETY: we will touch the root in a way that will not + // invalidate the position returned. + let root = unsafe { self.dormant_root.take().unwrap().awaken() }; + root.pop_internal_level(alloc.clone()); + self.dormant_root = Some(DormantMutRef::new(root).1); + }, + alloc.clone(), + ); + self.cur_leaf_edge = Some(pos); + return Some(kv); + } + self.cur_leaf_edge = Some(kv.next_leaf_edge()); + } + None + } + + /// Implementation of a typical `DrainFilter::size_hint` method. + pub(super) fn size_hint(&self) -> (usize, Option<usize>) { + // In most of the btree iterators, `self.length` is the number of elements + // yet to be visited. Here, it includes elements that were visited and that + // the predicate decided not to drain. Making this upper bound more tight + // during iteration would require an extra field. + (0, Some(*self.length)) + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<K, V, F> FusedIterator for DrainFilter<'_, K, V, F> where F: FnMut(&K, &mut V) -> bool {} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, K, V> Iterator for Range<'a, K, V> { + type Item = (&'a K, &'a V); + + fn next(&mut self) -> Option<(&'a K, &'a V)> { + self.inner.next_checked() + } + + fn last(mut self) -> Option<(&'a K, &'a V)> { + self.next_back() + } + + fn min(mut self) -> Option<(&'a K, &'a V)> { + self.next() + } + + fn max(mut self) -> Option<(&'a K, &'a V)> { + self.next_back() + } +} + +#[stable(feature = "map_values_mut", since = "1.10.0")] +impl<'a, K, V> Iterator for ValuesMut<'a, K, V> { + type Item = &'a mut V; + + fn next(&mut self) -> Option<&'a mut V> { + self.inner.next().map(|(_, v)| v) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } + + fn last(mut self) -> Option<&'a mut V> { + self.next_back() + } +} + +#[stable(feature = "map_values_mut", since = "1.10.0")] +impl<'a, K, V> DoubleEndedIterator for ValuesMut<'a, K, V> { + fn next_back(&mut self) -> Option<&'a mut V> { + self.inner.next_back().map(|(_, v)| v) + } +} + +#[stable(feature = "map_values_mut", since = "1.10.0")] +impl<K, V> ExactSizeIterator for ValuesMut<'_, K, V> { + fn len(&self) -> usize { + self.inner.len() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for ValuesMut<'_, K, V> {} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> Iterator for IntoKeys<K, V, A> { + type Item = K; + + fn next(&mut self) -> Option<K> { + self.inner.next().map(|(k, _)| k) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } + + fn last(mut self) -> Option<K> { + self.next_back() + } + + fn min(mut self) -> Option<K> { + self.next() + } + + fn max(mut self) -> Option<K> { + self.next_back() + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> DoubleEndedIterator for IntoKeys<K, V, A> { + fn next_back(&mut self) -> Option<K> { + self.inner.next_back().map(|(k, _)| k) + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> ExactSizeIterator for IntoKeys<K, V, A> { + fn len(&self) -> usize { + self.inner.len() + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> FusedIterator for IntoKeys<K, V, A> {} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> Iterator for IntoValues<K, V, A> { + type Item = V; + + fn next(&mut self) -> Option<V> { + self.inner.next().map(|(_, v)| v) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } + + fn last(mut self) -> Option<V> { + self.next_back() + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> DoubleEndedIterator for IntoValues<K, V, A> { + fn next_back(&mut self) -> Option<V> { + self.inner.next_back().map(|(_, v)| v) + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> ExactSizeIterator for IntoValues<K, V, A> { + fn len(&self) -> usize { + self.inner.len() + } +} + +#[stable(feature = "map_into_keys_values", since = "1.54.0")] +impl<K, V, A: Allocator + Clone> FusedIterator for IntoValues<K, V, A> {} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, K, V> DoubleEndedIterator for Range<'a, K, V> { + fn next_back(&mut self) -> Option<(&'a K, &'a V)> { + self.inner.next_back_checked() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for Range<'_, K, V> {} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<K, V> Clone for Range<'_, K, V> { + fn clone(&self) -> Self { + Range { inner: self.inner.clone() } + } +} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, K, V> Iterator for RangeMut<'a, K, V> { + type Item = (&'a K, &'a mut V); + + fn next(&mut self) -> Option<(&'a K, &'a mut V)> { + self.inner.next_checked() + } + + fn last(mut self) -> Option<(&'a K, &'a mut V)> { + self.next_back() + } + + fn min(mut self) -> Option<(&'a K, &'a mut V)> { + self.next() + } + + fn max(mut self) -> Option<(&'a K, &'a mut V)> { + self.next_back() + } +} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, K, V> DoubleEndedIterator for RangeMut<'a, K, V> { + fn next_back(&mut self) -> Option<(&'a K, &'a mut V)> { + self.inner.next_back_checked() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<K, V> FusedIterator for RangeMut<'_, K, V> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Ord, V> FromIterator<(K, V)> for BTreeMap<K, V> { + fn from_iter<T: IntoIterator<Item = (K, V)>>(iter: T) -> BTreeMap<K, V> { + let mut inputs: Vec<_> = iter.into_iter().collect(); + + if inputs.is_empty() { + return BTreeMap::new(); + } + + // use stable sort to preserve the insertion order. + inputs.sort_by(|a, b| a.0.cmp(&b.0)); + BTreeMap::bulk_build_from_sorted_iter(inputs, Global) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Ord, V, A: Allocator + Clone> Extend<(K, V)> for BTreeMap<K, V, A> { + #[inline] + fn extend<T: IntoIterator<Item = (K, V)>>(&mut self, iter: T) { + iter.into_iter().for_each(move |(k, v)| { + self.insert(k, v); + }); + } + + #[inline] + fn extend_one(&mut self, (k, v): (K, V)) { + self.insert(k, v); + } +} + +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, K: Ord + Copy, V: Copy, A: Allocator + Clone> Extend<(&'a K, &'a V)> + for BTreeMap<K, V, A> +{ + fn extend<I: IntoIterator<Item = (&'a K, &'a V)>>(&mut self, iter: I) { + self.extend(iter.into_iter().map(|(&key, &value)| (key, value))); + } + + #[inline] + fn extend_one(&mut self, (&k, &v): (&'a K, &'a V)) { + self.insert(k, v); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Hash, V: Hash, A: Allocator + Clone> Hash for BTreeMap<K, V, A> { + fn hash<H: Hasher>(&self, state: &mut H) { + state.write_length_prefix(self.len()); + for elt in self { + elt.hash(state); + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, V> Default for BTreeMap<K, V> { + /// Creates an empty `BTreeMap`. + fn default() -> BTreeMap<K, V> { + BTreeMap::new() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: PartialEq, V: PartialEq, A: Allocator + Clone> PartialEq for BTreeMap<K, V, A> { + fn eq(&self, other: &BTreeMap<K, V, A>) -> bool { + self.len() == other.len() && self.iter().zip(other).all(|(a, b)| a == b) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Eq, V: Eq, A: Allocator + Clone> Eq for BTreeMap<K, V, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: PartialOrd, V: PartialOrd, A: Allocator + Clone> PartialOrd for BTreeMap<K, V, A> { + #[inline] + fn partial_cmp(&self, other: &BTreeMap<K, V, A>) -> Option<Ordering> { + self.iter().partial_cmp(other.iter()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Ord, V: Ord, A: Allocator + Clone> Ord for BTreeMap<K, V, A> { + #[inline] + fn cmp(&self, other: &BTreeMap<K, V, A>) -> Ordering { + self.iter().cmp(other.iter()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K: Debug, V: Debug, A: Allocator + Clone> Debug for BTreeMap<K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_map().entries(self.iter()).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<K, Q: ?Sized, V, A: Allocator + Clone> Index<&Q> for BTreeMap<K, V, A> +where + K: Borrow<Q> + Ord, + Q: Ord, +{ + type Output = V; + + /// Returns a reference to the value corresponding to the supplied key. + /// + /// # Panics + /// + /// Panics if the key is not present in the `BTreeMap`. + #[inline] + fn index(&self, key: &Q) -> &V { + self.get(key).expect("no entry found for key") + } +} + +#[stable(feature = "std_collections_from_array", since = "1.56.0")] +impl<K: Ord, V, const N: usize> From<[(K, V); N]> for BTreeMap<K, V> { + /// Converts a `[(K, V); N]` into a `BTreeMap<(K, V)>`. + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let map1 = BTreeMap::from([(1, 2), (3, 4)]); + /// let map2: BTreeMap<_, _> = [(1, 2), (3, 4)].into(); + /// assert_eq!(map1, map2); + /// ``` + fn from(mut arr: [(K, V); N]) -> Self { + if N == 0 { + return BTreeMap::new(); + } + + // use stable sort to preserve the insertion order. + arr.sort_by(|a, b| a.0.cmp(&b.0)); + BTreeMap::bulk_build_from_sorted_iter(arr, Global) + } +} + +impl<K, V, A: Allocator + Clone> BTreeMap<K, V, A> { + /// Gets an iterator over the entries of the map, sorted by key. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::new(); + /// map.insert(3, "c"); + /// map.insert(2, "b"); + /// map.insert(1, "a"); + /// + /// for (key, value) in map.iter() { + /// println!("{key}: {value}"); + /// } + /// + /// let (first_key, first_value) = map.iter().next().unwrap(); + /// assert_eq!((*first_key, *first_value), (1, "a")); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter(&self) -> Iter<'_, K, V> { + if let Some(root) = &self.root { + let full_range = root.reborrow().full_range(); + + Iter { range: full_range, length: self.length } + } else { + Iter { range: LazyLeafRange::none(), length: 0 } + } + } + + /// Gets a mutable iterator over the entries of the map, sorted by key. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map = BTreeMap::from([ + /// ("a", 1), + /// ("b", 2), + /// ("c", 3), + /// ]); + /// + /// // add 10 to the value if the key isn't "a" + /// for (key, value) in map.iter_mut() { + /// if key != &"a" { + /// *value += 10; + /// } + /// } + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter_mut(&mut self) -> IterMut<'_, K, V> { + if let Some(root) = &mut self.root { + let full_range = root.borrow_valmut().full_range(); + + IterMut { range: full_range, length: self.length, _marker: PhantomData } + } else { + IterMut { range: LazyLeafRange::none(), length: 0, _marker: PhantomData } + } + } + + /// Gets an iterator over the keys of the map, in sorted order. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(2, "b"); + /// a.insert(1, "a"); + /// + /// let keys: Vec<_> = a.keys().cloned().collect(); + /// assert_eq!(keys, [1, 2]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn keys(&self) -> Keys<'_, K, V> { + Keys { inner: self.iter() } + } + + /// Gets an iterator over the values of the map, in order by key. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, "hello"); + /// a.insert(2, "goodbye"); + /// + /// let values: Vec<&str> = a.values().cloned().collect(); + /// assert_eq!(values, ["hello", "goodbye"]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn values(&self) -> Values<'_, K, V> { + Values { inner: self.iter() } + } + + /// Gets a mutable iterator over the values of the map, in order by key. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// a.insert(1, String::from("hello")); + /// a.insert(2, String::from("goodbye")); + /// + /// for value in a.values_mut() { + /// value.push_str("!"); + /// } + /// + /// let values: Vec<String> = a.values().cloned().collect(); + /// assert_eq!(values, [String::from("hello!"), + /// String::from("goodbye!")]); + /// ``` + #[stable(feature = "map_values_mut", since = "1.10.0")] + pub fn values_mut(&mut self) -> ValuesMut<'_, K, V> { + ValuesMut { inner: self.iter_mut() } + } + + /// Returns the number of elements in the map. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// assert_eq!(a.len(), 0); + /// a.insert(1, "a"); + /// assert_eq!(a.len(), 1); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + pub const fn len(&self) -> usize { + self.length + } + + /// Returns `true` if the map contains no elements. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut a = BTreeMap::new(); + /// assert!(a.is_empty()); + /// a.insert(1, "a"); + /// assert!(!a.is_empty()); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + pub const fn is_empty(&self) -> bool { + self.len() == 0 + } +} + +#[cfg(test)] +mod tests; diff --git a/library/alloc/src/collections/btree/map/entry.rs b/library/alloc/src/collections/btree/map/entry.rs new file mode 100644 index 000000000..b6eecf9b0 --- /dev/null +++ b/library/alloc/src/collections/btree/map/entry.rs @@ -0,0 +1,555 @@ +use core::fmt::{self, Debug}; +use core::marker::PhantomData; +use core::mem; + +use crate::alloc::{Allocator, Global}; + +use super::super::borrow::DormantMutRef; +use super::super::node::{marker, Handle, NodeRef}; +use super::BTreeMap; + +use Entry::*; + +/// A view into a single entry in a map, which may either be vacant or occupied. +/// +/// This `enum` is constructed from the [`entry`] method on [`BTreeMap`]. +/// +/// [`entry`]: BTreeMap::entry +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "BTreeEntry")] +pub enum Entry< + 'a, + K: 'a, + V: 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + /// A vacant entry. + #[stable(feature = "rust1", since = "1.0.0")] + Vacant(#[stable(feature = "rust1", since = "1.0.0")] VacantEntry<'a, K, V, A>), + + /// An occupied entry. + #[stable(feature = "rust1", since = "1.0.0")] + Occupied(#[stable(feature = "rust1", since = "1.0.0")] OccupiedEntry<'a, K, V, A>), +} + +#[stable(feature = "debug_btree_map", since = "1.12.0")] +impl<K: Debug + Ord, V: Debug, A: Allocator + Clone> Debug for Entry<'_, K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match *self { + Vacant(ref v) => f.debug_tuple("Entry").field(v).finish(), + Occupied(ref o) => f.debug_tuple("Entry").field(o).finish(), + } + } +} + +/// A view into a vacant entry in a `BTreeMap`. +/// It is part of the [`Entry`] enum. +#[stable(feature = "rust1", since = "1.0.0")] +pub struct VacantEntry< + 'a, + K, + V, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + pub(super) key: K, + /// `None` for a (empty) map without root + pub(super) handle: Option<Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>>, + pub(super) dormant_map: DormantMutRef<'a, BTreeMap<K, V, A>>, + + /// The BTreeMap will outlive this IntoIter so we don't care about drop order for `alloc`. + pub(super) alloc: A, + + // Be invariant in `K` and `V` + pub(super) _marker: PhantomData<&'a mut (K, V)>, +} + +#[stable(feature = "debug_btree_map", since = "1.12.0")] +impl<K: Debug + Ord, V, A: Allocator + Clone> Debug for VacantEntry<'_, K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("VacantEntry").field(self.key()).finish() + } +} + +/// A view into an occupied entry in a `BTreeMap`. +/// It is part of the [`Entry`] enum. +#[stable(feature = "rust1", since = "1.0.0")] +pub struct OccupiedEntry< + 'a, + K, + V, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + pub(super) handle: Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::KV>, + pub(super) dormant_map: DormantMutRef<'a, BTreeMap<K, V, A>>, + + /// The BTreeMap will outlive this IntoIter so we don't care about drop order for `alloc`. + pub(super) alloc: A, + + // Be invariant in `K` and `V` + pub(super) _marker: PhantomData<&'a mut (K, V)>, +} + +#[stable(feature = "debug_btree_map", since = "1.12.0")] +impl<K: Debug + Ord, V: Debug, A: Allocator + Clone> Debug for OccupiedEntry<'_, K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_struct("OccupiedEntry").field("key", self.key()).field("value", self.get()).finish() + } +} + +/// The error returned by [`try_insert`](BTreeMap::try_insert) when the key already exists. +/// +/// Contains the occupied entry, and the value that was not inserted. +#[unstable(feature = "map_try_insert", issue = "82766")] +pub struct OccupiedError<'a, K: 'a, V: 'a, A: Allocator + Clone = Global> { + /// The entry in the map that was already occupied. + pub entry: OccupiedEntry<'a, K, V, A>, + /// The value which was not inserted, because the entry was already occupied. + pub value: V, +} + +#[unstable(feature = "map_try_insert", issue = "82766")] +impl<K: Debug + Ord, V: Debug, A: Allocator + Clone> Debug for OccupiedError<'_, K, V, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_struct("OccupiedError") + .field("key", self.entry.key()) + .field("old_value", self.entry.get()) + .field("new_value", &self.value) + .finish() + } +} + +#[unstable(feature = "map_try_insert", issue = "82766")] +impl<'a, K: Debug + Ord, V: Debug, A: Allocator + Clone> fmt::Display + for OccupiedError<'a, K, V, A> +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!( + f, + "failed to insert {:?}, key {:?} already exists with value {:?}", + self.value, + self.entry.key(), + self.entry.get(), + ) + } +} + +impl<'a, K: Ord, V, A: Allocator + Clone> Entry<'a, K, V, A> { + /// Ensures a value is in the entry by inserting the default if empty, and returns + /// a mutable reference to the value in the entry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// assert_eq!(map["poneyland"], 12); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn or_insert(self, default: V) -> &'a mut V { + match self { + Occupied(entry) => entry.into_mut(), + Vacant(entry) => entry.insert(default), + } + } + + /// Ensures a value is in the entry by inserting the result of the default function if empty, + /// and returns a mutable reference to the value in the entry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, String> = BTreeMap::new(); + /// let s = "hoho".to_string(); + /// + /// map.entry("poneyland").or_insert_with(|| s); + /// + /// assert_eq!(map["poneyland"], "hoho".to_string()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn or_insert_with<F: FnOnce() -> V>(self, default: F) -> &'a mut V { + match self { + Occupied(entry) => entry.into_mut(), + Vacant(entry) => entry.insert(default()), + } + } + + /// Ensures a value is in the entry by inserting, if empty, the result of the default function. + /// This method allows for generating key-derived values for insertion by providing the default + /// function a reference to the key that was moved during the `.entry(key)` method call. + /// + /// The reference to the moved key is provided so that cloning or copying the key is + /// unnecessary, unlike with `.or_insert_with(|| ... )`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// + /// map.entry("poneyland").or_insert_with_key(|key| key.chars().count()); + /// + /// assert_eq!(map["poneyland"], 9); + /// ``` + #[inline] + #[stable(feature = "or_insert_with_key", since = "1.50.0")] + pub fn or_insert_with_key<F: FnOnce(&K) -> V>(self, default: F) -> &'a mut V { + match self { + Occupied(entry) => entry.into_mut(), + Vacant(entry) => { + let value = default(entry.key()); + entry.insert(value) + } + } + } + + /// Returns a reference to this entry's key. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// assert_eq!(map.entry("poneyland").key(), &"poneyland"); + /// ``` + #[stable(feature = "map_entry_keys", since = "1.10.0")] + pub fn key(&self) -> &K { + match *self { + Occupied(ref entry) => entry.key(), + Vacant(ref entry) => entry.key(), + } + } + + /// Provides in-place mutable access to an occupied entry before any + /// potential inserts into the map. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// + /// map.entry("poneyland") + /// .and_modify(|e| { *e += 1 }) + /// .or_insert(42); + /// assert_eq!(map["poneyland"], 42); + /// + /// map.entry("poneyland") + /// .and_modify(|e| { *e += 1 }) + /// .or_insert(42); + /// assert_eq!(map["poneyland"], 43); + /// ``` + #[stable(feature = "entry_and_modify", since = "1.26.0")] + pub fn and_modify<F>(self, f: F) -> Self + where + F: FnOnce(&mut V), + { + match self { + Occupied(mut entry) => { + f(entry.get_mut()); + Occupied(entry) + } + Vacant(entry) => Vacant(entry), + } + } +} + +impl<'a, K: Ord, V: Default, A: Allocator + Clone> Entry<'a, K, V, A> { + #[stable(feature = "entry_or_default", since = "1.28.0")] + /// Ensures a value is in the entry by inserting the default value if empty, + /// and returns a mutable reference to the value in the entry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, Option<usize>> = BTreeMap::new(); + /// map.entry("poneyland").or_default(); + /// + /// assert_eq!(map["poneyland"], None); + /// ``` + pub fn or_default(self) -> &'a mut V { + match self { + Occupied(entry) => entry.into_mut(), + Vacant(entry) => entry.insert(Default::default()), + } + } +} + +impl<'a, K: Ord, V, A: Allocator + Clone> VacantEntry<'a, K, V, A> { + /// Gets a reference to the key that would be used when inserting a value + /// through the VacantEntry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// assert_eq!(map.entry("poneyland").key(), &"poneyland"); + /// ``` + #[stable(feature = "map_entry_keys", since = "1.10.0")] + pub fn key(&self) -> &K { + &self.key + } + + /// Take ownership of the key. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// + /// if let Entry::Vacant(v) = map.entry("poneyland") { + /// v.into_key(); + /// } + /// ``` + #[stable(feature = "map_entry_recover_keys2", since = "1.12.0")] + pub fn into_key(self) -> K { + self.key + } + + /// Sets the value of the entry with the `VacantEntry`'s key, + /// and returns a mutable reference to it. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, u32> = BTreeMap::new(); + /// + /// if let Entry::Vacant(o) = map.entry("poneyland") { + /// o.insert(37); + /// } + /// assert_eq!(map["poneyland"], 37); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(self, value: V) -> &'a mut V { + let out_ptr = match self.handle { + None => { + // SAFETY: There is no tree yet so no reference to it exists. + let map = unsafe { self.dormant_map.awaken() }; + let mut root = NodeRef::new_leaf(self.alloc.clone()); + let val_ptr = root.borrow_mut().push(self.key, value) as *mut V; + map.root = Some(root.forget_type()); + map.length = 1; + val_ptr + } + Some(handle) => match handle.insert_recursing(self.key, value, self.alloc.clone()) { + (None, val_ptr) => { + // SAFETY: We have consumed self.handle. + let map = unsafe { self.dormant_map.awaken() }; + map.length += 1; + val_ptr + } + (Some(ins), val_ptr) => { + drop(ins.left); + // SAFETY: We have consumed self.handle and dropped the + // remaining reference to the tree, ins.left. + let map = unsafe { self.dormant_map.awaken() }; + let root = map.root.as_mut().unwrap(); // same as ins.left + root.push_internal_level(self.alloc).push(ins.kv.0, ins.kv.1, ins.right); + map.length += 1; + val_ptr + } + }, + }; + // Now that we have finished growing the tree using borrowed references, + // dereference the pointer to a part of it, that we picked up along the way. + unsafe { &mut *out_ptr } + } +} + +impl<'a, K: Ord, V, A: Allocator + Clone> OccupiedEntry<'a, K, V, A> { + /// Gets a reference to the key in the entry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// assert_eq!(map.entry("poneyland").key(), &"poneyland"); + /// ``` + #[must_use] + #[stable(feature = "map_entry_keys", since = "1.10.0")] + pub fn key(&self) -> &K { + self.handle.reborrow().into_kv().0 + } + + /// Take ownership of the key and value from the map. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// if let Entry::Occupied(o) = map.entry("poneyland") { + /// // We delete the entry from the map. + /// o.remove_entry(); + /// } + /// + /// // If now try to get the value, it will panic: + /// // println!("{}", map["poneyland"]); + /// ``` + #[stable(feature = "map_entry_recover_keys2", since = "1.12.0")] + pub fn remove_entry(self) -> (K, V) { + self.remove_kv() + } + + /// Gets a reference to the value in the entry. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// if let Entry::Occupied(o) = map.entry("poneyland") { + /// assert_eq!(o.get(), &12); + /// } + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get(&self) -> &V { + self.handle.reborrow().into_kv().1 + } + + /// Gets a mutable reference to the value in the entry. + /// + /// If you need a reference to the `OccupiedEntry` that may outlive the + /// destruction of the `Entry` value, see [`into_mut`]. + /// + /// [`into_mut`]: OccupiedEntry::into_mut + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// assert_eq!(map["poneyland"], 12); + /// if let Entry::Occupied(mut o) = map.entry("poneyland") { + /// *o.get_mut() += 10; + /// assert_eq!(*o.get(), 22); + /// + /// // We can use the same Entry multiple times. + /// *o.get_mut() += 2; + /// } + /// assert_eq!(map["poneyland"], 24); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get_mut(&mut self) -> &mut V { + self.handle.kv_mut().1 + } + + /// Converts the entry into a mutable reference to its value. + /// + /// If you need multiple references to the `OccupiedEntry`, see [`get_mut`]. + /// + /// [`get_mut`]: OccupiedEntry::get_mut + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// assert_eq!(map["poneyland"], 12); + /// if let Entry::Occupied(o) = map.entry("poneyland") { + /// *o.into_mut() += 10; + /// } + /// assert_eq!(map["poneyland"], 22); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_mut(self) -> &'a mut V { + self.handle.into_val_mut() + } + + /// Sets the value of the entry with the `OccupiedEntry`'s key, + /// and returns the entry's old value. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// if let Entry::Occupied(mut o) = map.entry("poneyland") { + /// assert_eq!(o.insert(15), 12); + /// } + /// assert_eq!(map["poneyland"], 15); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(&mut self, value: V) -> V { + mem::replace(self.get_mut(), value) + } + + /// Takes the value of the entry out of the map, and returns it. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeMap; + /// use std::collections::btree_map::Entry; + /// + /// let mut map: BTreeMap<&str, usize> = BTreeMap::new(); + /// map.entry("poneyland").or_insert(12); + /// + /// if let Entry::Occupied(o) = map.entry("poneyland") { + /// assert_eq!(o.remove(), 12); + /// } + /// // If we try to get "poneyland"'s value, it'll panic: + /// // println!("{}", map["poneyland"]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn remove(self) -> V { + self.remove_kv().1 + } + + // Body of `remove_entry`, probably separate because the name reflects the returned pair. + pub(super) fn remove_kv(self) -> (K, V) { + let mut emptied_internal_root = false; + let (old_kv, _) = + self.handle.remove_kv_tracking(|| emptied_internal_root = true, self.alloc.clone()); + // SAFETY: we consumed the intermediate root borrow, `self.handle`. + let map = unsafe { self.dormant_map.awaken() }; + map.length -= 1; + if emptied_internal_root { + let root = map.root.as_mut().unwrap(); + root.pop_internal_level(self.alloc); + } + old_kv + } +} diff --git a/library/alloc/src/collections/btree/map/tests.rs b/library/alloc/src/collections/btree/map/tests.rs new file mode 100644 index 000000000..4c372b1d6 --- /dev/null +++ b/library/alloc/src/collections/btree/map/tests.rs @@ -0,0 +1,2338 @@ +use super::super::testing::crash_test::{CrashTestDummy, Panic}; +use super::super::testing::ord_chaos::{Cyclic3, Governed, Governor}; +use super::super::testing::rng::DeterministicRng; +use super::Entry::{Occupied, Vacant}; +use super::*; +use crate::boxed::Box; +use crate::fmt::Debug; +use crate::rc::Rc; +use crate::string::{String, ToString}; +use crate::vec::Vec; +use std::cmp::Ordering; +use std::convert::TryFrom; +use std::iter::{self, FromIterator}; +use std::mem; +use std::ops::Bound::{self, Excluded, Included, Unbounded}; +use std::ops::RangeBounds; +use std::panic::{catch_unwind, AssertUnwindSafe}; +use std::sync::atomic::{AtomicUsize, Ordering::SeqCst}; + +// Minimum number of elements to insert, to guarantee a tree with 2 levels, +// i.e., a tree who's root is an internal node at height 1, with edges to leaf nodes. +// It's not the minimum size: removing an element from such a tree does not always reduce height. +const MIN_INSERTS_HEIGHT_1: usize = node::CAPACITY + 1; + +// Minimum number of elements to insert in ascending order, to guarantee a tree with 3 levels, +// i.e., a tree who's root is an internal node at height 2, with edges to more internal nodes. +// It's not the minimum size: removing an element from such a tree does not always reduce height. +const MIN_INSERTS_HEIGHT_2: usize = 89; + +// Gathers all references from a mutable iterator and makes sure Miri notices if +// using them is dangerous. +fn test_all_refs<'a, T: 'a>(dummy: &mut T, iter: impl Iterator<Item = &'a mut T>) { + // Gather all those references. + let mut refs: Vec<&mut T> = iter.collect(); + // Use them all. Twice, to be sure we got all interleavings. + for r in refs.iter_mut() { + mem::swap(dummy, r); + } + for r in refs { + mem::swap(dummy, r); + } +} + +impl<K, V> BTreeMap<K, V> { + // Panics if the map (or the code navigating it) is corrupted. + fn check_invariants(&self) { + if let Some(root) = &self.root { + let root_node = root.reborrow(); + + // Check the back pointers top-down, before we attempt to rely on + // more serious navigation code. + assert!(root_node.ascend().is_err()); + root_node.assert_back_pointers(); + + // Check consistency of `length` with what navigation code encounters. + assert_eq!(self.length, root_node.calc_length()); + + // Lastly, check the invariant causing the least harm. + root_node.assert_min_len(if root_node.height() > 0 { 1 } else { 0 }); + } else { + assert_eq!(self.length, 0); + } + + // Check that `assert_strictly_ascending` will encounter all keys. + assert_eq!(self.length, self.keys().count()); + } + + // Panics if the map is corrupted or if the keys are not in strictly + // ascending order, in the current opinion of the `Ord` implementation. + // If the `Ord` implementation violates transitivity, this method does not + // guarantee that all keys are unique, just that adjacent keys are unique. + fn check(&self) + where + K: Debug + Ord, + { + self.check_invariants(); + self.assert_strictly_ascending(); + } + + // Returns the height of the root, if any. + fn height(&self) -> Option<usize> { + self.root.as_ref().map(node::Root::height) + } + + fn dump_keys(&self) -> String + where + K: Debug, + { + if let Some(root) = self.root.as_ref() { + root.reborrow().dump_keys() + } else { + String::from("not yet allocated") + } + } + + // Panics if the keys are not in strictly ascending order. + fn assert_strictly_ascending(&self) + where + K: Debug + Ord, + { + let mut keys = self.keys(); + if let Some(mut previous) = keys.next() { + for next in keys { + assert!(previous < next, "{:?} >= {:?}", previous, next); + previous = next; + } + } + } + + // Transform the tree to minimize wasted space, obtaining fewer nodes that + // are mostly filled up to their capacity. The same compact tree could have + // been obtained by inserting keys in a shrewd order. + fn compact(&mut self) + where + K: Ord, + { + let iter = mem::take(self).into_iter(); + if !iter.is_empty() { + self.root.insert(Root::new(*self.alloc)).bulk_push(iter, &mut self.length, *self.alloc); + } + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Immut<'a>, K, V, marker::LeafOrInternal> { + fn assert_min_len(self, min_len: usize) { + assert!(self.len() >= min_len, "node len {} < {}", self.len(), min_len); + if let node::ForceResult::Internal(node) = self.force() { + for idx in 0..=node.len() { + let edge = unsafe { Handle::new_edge(node, idx) }; + edge.descend().assert_min_len(MIN_LEN); + } + } + } +} + +// Tests our value of MIN_INSERTS_HEIGHT_2. Failure may mean you just need to +// adapt that value to match a change in node::CAPACITY or the choices made +// during insertion, otherwise other test cases may fail or be less useful. +#[test] +fn test_levels() { + let mut map = BTreeMap::new(); + map.check(); + assert_eq!(map.height(), None); + assert_eq!(map.len(), 0); + + map.insert(0, ()); + while map.height() == Some(0) { + let last_key = *map.last_key_value().unwrap().0; + map.insert(last_key + 1, ()); + } + map.check(); + // Structure: + // - 1 element in internal root node with 2 children + // - 6 elements in left leaf child + // - 5 elements in right leaf child + assert_eq!(map.height(), Some(1)); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_1, "{}", map.dump_keys()); + + while map.height() == Some(1) { + let last_key = *map.last_key_value().unwrap().0; + map.insert(last_key + 1, ()); + } + map.check(); + // Structure: + // - 1 element in internal root node with 2 children + // - 6 elements in left internal child with 7 grandchildren + // - 42 elements in left child's 7 grandchildren with 6 elements each + // - 5 elements in right internal child with 6 grandchildren + // - 30 elements in right child's 5 first grandchildren with 6 elements each + // - 5 elements in right child's last grandchild + assert_eq!(map.height(), Some(2)); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_2, "{}", map.dump_keys()); +} + +// Ensures the testing infrastructure usually notices order violations. +#[test] +#[should_panic] +fn test_check_ord_chaos() { + let gov = Governor::new(); + let map = BTreeMap::from([(Governed(1, &gov), ()), (Governed(2, &gov), ())]); + gov.flip(); + map.check(); +} + +// Ensures the testing infrastructure doesn't always mind order violations. +#[test] +fn test_check_invariants_ord_chaos() { + let gov = Governor::new(); + let map = BTreeMap::from([(Governed(1, &gov), ()), (Governed(2, &gov), ())]); + gov.flip(); + map.check_invariants(); +} + +#[test] +fn test_basic_large() { + let mut map = BTreeMap::new(); + // Miri is too slow + let size = if cfg!(miri) { MIN_INSERTS_HEIGHT_2 } else { 10000 }; + let size = size + (size % 2); // round up to even number + assert_eq!(map.len(), 0); + + for i in 0..size { + assert_eq!(map.insert(i, 10 * i), None); + assert_eq!(map.len(), i + 1); + } + + assert_eq!(map.first_key_value(), Some((&0, &0))); + assert_eq!(map.last_key_value(), Some((&(size - 1), &(10 * (size - 1))))); + assert_eq!(map.first_entry().unwrap().key(), &0); + assert_eq!(map.last_entry().unwrap().key(), &(size - 1)); + + for i in 0..size { + assert_eq!(map.get(&i).unwrap(), &(i * 10)); + } + + for i in size..size * 2 { + assert_eq!(map.get(&i), None); + } + + for i in 0..size { + assert_eq!(map.insert(i, 100 * i), Some(10 * i)); + assert_eq!(map.len(), size); + } + + for i in 0..size { + assert_eq!(map.get(&i).unwrap(), &(i * 100)); + } + + for i in 0..size / 2 { + assert_eq!(map.remove(&(i * 2)), Some(i * 200)); + assert_eq!(map.len(), size - i - 1); + } + + for i in 0..size / 2 { + assert_eq!(map.get(&(2 * i)), None); + assert_eq!(map.get(&(2 * i + 1)).unwrap(), &(i * 200 + 100)); + } + + for i in 0..size / 2 { + assert_eq!(map.remove(&(2 * i)), None); + assert_eq!(map.remove(&(2 * i + 1)), Some(i * 200 + 100)); + assert_eq!(map.len(), size / 2 - i - 1); + } + map.check(); +} + +#[test] +fn test_basic_small() { + let mut map = BTreeMap::new(); + // Empty, root is absent (None): + assert_eq!(map.remove(&1), None); + assert_eq!(map.len(), 0); + assert_eq!(map.get(&1), None); + assert_eq!(map.get_mut(&1), None); + assert_eq!(map.first_key_value(), None); + assert_eq!(map.last_key_value(), None); + assert_eq!(map.keys().count(), 0); + assert_eq!(map.values().count(), 0); + assert_eq!(map.range(..).next(), None); + assert_eq!(map.range(..1).next(), None); + assert_eq!(map.range(1..).next(), None); + assert_eq!(map.range(1..=1).next(), None); + assert_eq!(map.range(1..2).next(), None); + assert_eq!(map.height(), None); + assert_eq!(map.insert(1, 1), None); + assert_eq!(map.height(), Some(0)); + map.check(); + + // 1 key-value pair: + assert_eq!(map.len(), 1); + assert_eq!(map.get(&1), Some(&1)); + assert_eq!(map.get_mut(&1), Some(&mut 1)); + assert_eq!(map.first_key_value(), Some((&1, &1))); + assert_eq!(map.last_key_value(), Some((&1, &1))); + assert_eq!(map.keys().collect::<Vec<_>>(), vec![&1]); + assert_eq!(map.values().collect::<Vec<_>>(), vec![&1]); + assert_eq!(map.insert(1, 2), Some(1)); + assert_eq!(map.len(), 1); + assert_eq!(map.get(&1), Some(&2)); + assert_eq!(map.get_mut(&1), Some(&mut 2)); + assert_eq!(map.first_key_value(), Some((&1, &2))); + assert_eq!(map.last_key_value(), Some((&1, &2))); + assert_eq!(map.keys().collect::<Vec<_>>(), vec![&1]); + assert_eq!(map.values().collect::<Vec<_>>(), vec![&2]); + assert_eq!(map.insert(2, 4), None); + assert_eq!(map.height(), Some(0)); + map.check(); + + // 2 key-value pairs: + assert_eq!(map.len(), 2); + assert_eq!(map.get(&2), Some(&4)); + assert_eq!(map.get_mut(&2), Some(&mut 4)); + assert_eq!(map.first_key_value(), Some((&1, &2))); + assert_eq!(map.last_key_value(), Some((&2, &4))); + assert_eq!(map.keys().collect::<Vec<_>>(), vec![&1, &2]); + assert_eq!(map.values().collect::<Vec<_>>(), vec![&2, &4]); + assert_eq!(map.remove(&1), Some(2)); + assert_eq!(map.height(), Some(0)); + map.check(); + + // 1 key-value pair: + assert_eq!(map.len(), 1); + assert_eq!(map.get(&1), None); + assert_eq!(map.get_mut(&1), None); + assert_eq!(map.get(&2), Some(&4)); + assert_eq!(map.get_mut(&2), Some(&mut 4)); + assert_eq!(map.first_key_value(), Some((&2, &4))); + assert_eq!(map.last_key_value(), Some((&2, &4))); + assert_eq!(map.keys().collect::<Vec<_>>(), vec![&2]); + assert_eq!(map.values().collect::<Vec<_>>(), vec![&4]); + assert_eq!(map.remove(&2), Some(4)); + assert_eq!(map.height(), Some(0)); + map.check(); + + // Empty but root is owned (Some(...)): + assert_eq!(map.len(), 0); + assert_eq!(map.get(&1), None); + assert_eq!(map.get_mut(&1), None); + assert_eq!(map.first_key_value(), None); + assert_eq!(map.last_key_value(), None); + assert_eq!(map.keys().count(), 0); + assert_eq!(map.values().count(), 0); + assert_eq!(map.range(..).next(), None); + assert_eq!(map.range(..1).next(), None); + assert_eq!(map.range(1..).next(), None); + assert_eq!(map.range(1..=1).next(), None); + assert_eq!(map.range(1..2).next(), None); + assert_eq!(map.remove(&1), None); + assert_eq!(map.height(), Some(0)); + map.check(); +} + +#[test] +fn test_iter() { + // Miri is too slow + let size = if cfg!(miri) { 200 } else { 10000 }; + let mut map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + fn test<T>(size: usize, mut iter: T) + where + T: Iterator<Item = (usize, usize)>, + { + for i in 0..size { + assert_eq!(iter.size_hint(), (size - i, Some(size - i))); + assert_eq!(iter.next().unwrap(), (i, i)); + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + } + test(size, map.iter().map(|(&k, &v)| (k, v))); + test(size, map.iter_mut().map(|(&k, &mut v)| (k, v))); + test(size, map.into_iter()); +} + +#[test] +fn test_iter_rev() { + // Miri is too slow + let size = if cfg!(miri) { 200 } else { 10000 }; + let mut map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + fn test<T>(size: usize, mut iter: T) + where + T: Iterator<Item = (usize, usize)>, + { + for i in 0..size { + assert_eq!(iter.size_hint(), (size - i, Some(size - i))); + assert_eq!(iter.next().unwrap(), (size - i - 1, size - i - 1)); + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + } + test(size, map.iter().rev().map(|(&k, &v)| (k, v))); + test(size, map.iter_mut().rev().map(|(&k, &mut v)| (k, v))); + test(size, map.into_iter().rev()); +} + +// Specifically tests iter_mut's ability to mutate the value of pairs in-line. +fn do_test_iter_mut_mutation<T>(size: usize) +where + T: Copy + Debug + Ord + TryFrom<usize>, + <T as TryFrom<usize>>::Error: Debug, +{ + let zero = T::try_from(0).unwrap(); + let mut map = BTreeMap::from_iter((0..size).map(|i| (T::try_from(i).unwrap(), zero))); + + // Forward and backward iteration sees enough pairs (also tested elsewhere) + assert_eq!(map.iter_mut().count(), size); + assert_eq!(map.iter_mut().rev().count(), size); + + // Iterate forwards, trying to mutate to unique values + for (i, (k, v)) in map.iter_mut().enumerate() { + assert_eq!(*k, T::try_from(i).unwrap()); + assert_eq!(*v, zero); + *v = T::try_from(i + 1).unwrap(); + } + + // Iterate backwards, checking that mutations succeeded and trying to mutate again + for (i, (k, v)) in map.iter_mut().rev().enumerate() { + assert_eq!(*k, T::try_from(size - i - 1).unwrap()); + assert_eq!(*v, T::try_from(size - i).unwrap()); + *v = T::try_from(2 * size - i).unwrap(); + } + + // Check that backward mutations succeeded + for (i, (k, v)) in map.iter_mut().enumerate() { + assert_eq!(*k, T::try_from(i).unwrap()); + assert_eq!(*v, T::try_from(size + i + 1).unwrap()); + } + map.check(); +} + +#[derive(Clone, Copy, Debug, Eq, PartialEq, PartialOrd, Ord)] +#[repr(align(32))] +struct Align32(usize); + +impl TryFrom<usize> for Align32 { + type Error = (); + + fn try_from(s: usize) -> Result<Align32, ()> { + Ok(Align32(s)) + } +} + +#[test] +fn test_iter_mut_mutation() { + // Check many alignments and trees with roots at various heights. + do_test_iter_mut_mutation::<u8>(0); + do_test_iter_mut_mutation::<u8>(1); + do_test_iter_mut_mutation::<u8>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<u8>(MIN_INSERTS_HEIGHT_2); + do_test_iter_mut_mutation::<u16>(1); + do_test_iter_mut_mutation::<u16>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<u16>(MIN_INSERTS_HEIGHT_2); + do_test_iter_mut_mutation::<u32>(1); + do_test_iter_mut_mutation::<u32>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<u32>(MIN_INSERTS_HEIGHT_2); + do_test_iter_mut_mutation::<u64>(1); + do_test_iter_mut_mutation::<u64>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<u64>(MIN_INSERTS_HEIGHT_2); + do_test_iter_mut_mutation::<u128>(1); + do_test_iter_mut_mutation::<u128>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<u128>(MIN_INSERTS_HEIGHT_2); + do_test_iter_mut_mutation::<Align32>(1); + do_test_iter_mut_mutation::<Align32>(MIN_INSERTS_HEIGHT_1); + do_test_iter_mut_mutation::<Align32>(MIN_INSERTS_HEIGHT_2); +} + +#[test] +fn test_values_mut() { + let mut a = BTreeMap::from_iter((0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i))); + test_all_refs(&mut 13, a.values_mut()); + a.check(); +} + +#[test] +fn test_values_mut_mutation() { + let mut a = BTreeMap::new(); + a.insert(1, String::from("hello")); + a.insert(2, String::from("goodbye")); + + for value in a.values_mut() { + value.push_str("!"); + } + + let values = Vec::from_iter(a.values().cloned()); + assert_eq!(values, [String::from("hello!"), String::from("goodbye!")]); + a.check(); +} + +#[test] +fn test_iter_entering_root_twice() { + let mut map = BTreeMap::from([(0, 0), (1, 1)]); + let mut it = map.iter_mut(); + let front = it.next().unwrap(); + let back = it.next_back().unwrap(); + assert_eq!(front, (&0, &mut 0)); + assert_eq!(back, (&1, &mut 1)); + *front.1 = 24; + *back.1 = 42; + assert_eq!(front, (&0, &mut 24)); + assert_eq!(back, (&1, &mut 42)); + assert_eq!(it.next(), None); + assert_eq!(it.next_back(), None); + map.check(); +} + +#[test] +fn test_iter_descending_to_same_node_twice() { + let mut map = BTreeMap::from_iter((0..MIN_INSERTS_HEIGHT_1).map(|i| (i, i))); + let mut it = map.iter_mut(); + // Descend into first child. + let front = it.next().unwrap(); + // Descend into first child again, after running through second child. + while it.next_back().is_some() {} + // Check immutable access. + assert_eq!(front, (&0, &mut 0)); + // Perform mutable access. + *front.1 = 42; + map.check(); +} + +#[test] +fn test_iter_mixed() { + // Miri is too slow + let size = if cfg!(miri) { 200 } else { 10000 }; + + let mut map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + fn test<T>(size: usize, mut iter: T) + where + T: Iterator<Item = (usize, usize)> + DoubleEndedIterator, + { + for i in 0..size / 4 { + assert_eq!(iter.size_hint(), (size - i * 2, Some(size - i * 2))); + assert_eq!(iter.next().unwrap(), (i, i)); + assert_eq!(iter.next_back().unwrap(), (size - i - 1, size - i - 1)); + } + for i in size / 4..size * 3 / 4 { + assert_eq!(iter.size_hint(), (size * 3 / 4 - i, Some(size * 3 / 4 - i))); + assert_eq!(iter.next().unwrap(), (i, i)); + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + } + test(size, map.iter().map(|(&k, &v)| (k, v))); + test(size, map.iter_mut().map(|(&k, &mut v)| (k, v))); + test(size, map.into_iter()); +} + +#[test] +fn test_iter_min_max() { + let mut a = BTreeMap::new(); + assert_eq!(a.iter().min(), None); + assert_eq!(a.iter().max(), None); + assert_eq!(a.iter_mut().min(), None); + assert_eq!(a.iter_mut().max(), None); + assert_eq!(a.range(..).min(), None); + assert_eq!(a.range(..).max(), None); + assert_eq!(a.range_mut(..).min(), None); + assert_eq!(a.range_mut(..).max(), None); + assert_eq!(a.keys().min(), None); + assert_eq!(a.keys().max(), None); + assert_eq!(a.values().min(), None); + assert_eq!(a.values().max(), None); + assert_eq!(a.values_mut().min(), None); + assert_eq!(a.values_mut().max(), None); + a.insert(1, 42); + a.insert(2, 24); + assert_eq!(a.iter().min(), Some((&1, &42))); + assert_eq!(a.iter().max(), Some((&2, &24))); + assert_eq!(a.iter_mut().min(), Some((&1, &mut 42))); + assert_eq!(a.iter_mut().max(), Some((&2, &mut 24))); + assert_eq!(a.range(..).min(), Some((&1, &42))); + assert_eq!(a.range(..).max(), Some((&2, &24))); + assert_eq!(a.range_mut(..).min(), Some((&1, &mut 42))); + assert_eq!(a.range_mut(..).max(), Some((&2, &mut 24))); + assert_eq!(a.keys().min(), Some(&1)); + assert_eq!(a.keys().max(), Some(&2)); + assert_eq!(a.values().min(), Some(&24)); + assert_eq!(a.values().max(), Some(&42)); + assert_eq!(a.values_mut().min(), Some(&mut 24)); + assert_eq!(a.values_mut().max(), Some(&mut 42)); + a.check(); +} + +fn range_keys(map: &BTreeMap<i32, i32>, range: impl RangeBounds<i32>) -> Vec<i32> { + Vec::from_iter(map.range(range).map(|(&k, &v)| { + assert_eq!(k, v); + k + })) +} + +#[test] +fn test_range_small() { + let size = 4; + + let all = Vec::from_iter(1..=size); + let (first, last) = (vec![all[0]], vec![all[size as usize - 1]]); + let map = BTreeMap::from_iter(all.iter().copied().map(|i| (i, i))); + + assert_eq!(range_keys(&map, (Excluded(0), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Excluded(0), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Excluded(0), Included(size))), all); + assert_eq!(range_keys(&map, (Excluded(0), Unbounded)), all); + assert_eq!(range_keys(&map, (Included(0), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Included(0), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Included(0), Included(size))), all); + assert_eq!(range_keys(&map, (Included(0), Unbounded)), all); + assert_eq!(range_keys(&map, (Included(1), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Included(1), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Included(1), Included(size))), all); + assert_eq!(range_keys(&map, (Included(1), Unbounded)), all); + assert_eq!(range_keys(&map, (Unbounded, Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Unbounded, Included(size + 1))), all); + assert_eq!(range_keys(&map, (Unbounded, Included(size))), all); + assert_eq!(range_keys(&map, ..), all); + + assert_eq!(range_keys(&map, (Excluded(0), Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Excluded(0), Included(0))), vec![]); + assert_eq!(range_keys(&map, (Included(0), Included(0))), vec![]); + assert_eq!(range_keys(&map, (Included(0), Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Unbounded, Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Unbounded, Included(0))), vec![]); + assert_eq!(range_keys(&map, (Excluded(0), Excluded(2))), first); + assert_eq!(range_keys(&map, (Excluded(0), Included(1))), first); + assert_eq!(range_keys(&map, (Included(0), Excluded(2))), first); + assert_eq!(range_keys(&map, (Included(0), Included(1))), first); + assert_eq!(range_keys(&map, (Included(1), Excluded(2))), first); + assert_eq!(range_keys(&map, (Included(1), Included(1))), first); + assert_eq!(range_keys(&map, (Unbounded, Excluded(2))), first); + assert_eq!(range_keys(&map, (Unbounded, Included(1))), first); + assert_eq!(range_keys(&map, (Excluded(size - 1), Excluded(size + 1))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Included(size + 1))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Included(size))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Unbounded)), last); + assert_eq!(range_keys(&map, (Included(size), Excluded(size + 1))), last); + assert_eq!(range_keys(&map, (Included(size), Included(size + 1))), last); + assert_eq!(range_keys(&map, (Included(size), Included(size))), last); + assert_eq!(range_keys(&map, (Included(size), Unbounded)), last); + assert_eq!(range_keys(&map, (Excluded(size), Excluded(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Excluded(size), Included(size))), vec![]); + assert_eq!(range_keys(&map, (Excluded(size), Unbounded)), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Excluded(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Included(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Unbounded)), vec![]); + + assert_eq!(range_keys(&map, ..3), vec![1, 2]); + assert_eq!(range_keys(&map, 3..), vec![3, 4]); + assert_eq!(range_keys(&map, 2..=3), vec![2, 3]); +} + +#[test] +fn test_range_height_1() { + // Tests tree with a root and 2 leaves. We test around the middle of the + // keys because one of those is the single key in the root node. + let map = BTreeMap::from_iter((0..MIN_INSERTS_HEIGHT_1 as i32).map(|i| (i, i))); + let middle = MIN_INSERTS_HEIGHT_1 as i32 / 2; + for root in middle - 2..=middle + 2 { + assert_eq!(range_keys(&map, (Excluded(root), Excluded(root + 1))), vec![]); + assert_eq!(range_keys(&map, (Excluded(root), Included(root + 1))), vec![root + 1]); + assert_eq!(range_keys(&map, (Included(root), Excluded(root + 1))), vec![root]); + assert_eq!(range_keys(&map, (Included(root), Included(root + 1))), vec![root, root + 1]); + + assert_eq!(range_keys(&map, (Excluded(root - 1), Excluded(root))), vec![]); + assert_eq!(range_keys(&map, (Included(root - 1), Excluded(root))), vec![root - 1]); + assert_eq!(range_keys(&map, (Excluded(root - 1), Included(root))), vec![root]); + assert_eq!(range_keys(&map, (Included(root - 1), Included(root))), vec![root - 1, root]); + } +} + +#[test] +fn test_range_large() { + let size = 200; + + let all = Vec::from_iter(1..=size); + let (first, last) = (vec![all[0]], vec![all[size as usize - 1]]); + let map = BTreeMap::from_iter(all.iter().copied().map(|i| (i, i))); + + assert_eq!(range_keys(&map, (Excluded(0), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Excluded(0), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Excluded(0), Included(size))), all); + assert_eq!(range_keys(&map, (Excluded(0), Unbounded)), all); + assert_eq!(range_keys(&map, (Included(0), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Included(0), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Included(0), Included(size))), all); + assert_eq!(range_keys(&map, (Included(0), Unbounded)), all); + assert_eq!(range_keys(&map, (Included(1), Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Included(1), Included(size + 1))), all); + assert_eq!(range_keys(&map, (Included(1), Included(size))), all); + assert_eq!(range_keys(&map, (Included(1), Unbounded)), all); + assert_eq!(range_keys(&map, (Unbounded, Excluded(size + 1))), all); + assert_eq!(range_keys(&map, (Unbounded, Included(size + 1))), all); + assert_eq!(range_keys(&map, (Unbounded, Included(size))), all); + assert_eq!(range_keys(&map, ..), all); + + assert_eq!(range_keys(&map, (Excluded(0), Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Excluded(0), Included(0))), vec![]); + assert_eq!(range_keys(&map, (Included(0), Included(0))), vec![]); + assert_eq!(range_keys(&map, (Included(0), Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Unbounded, Excluded(1))), vec![]); + assert_eq!(range_keys(&map, (Unbounded, Included(0))), vec![]); + assert_eq!(range_keys(&map, (Excluded(0), Excluded(2))), first); + assert_eq!(range_keys(&map, (Excluded(0), Included(1))), first); + assert_eq!(range_keys(&map, (Included(0), Excluded(2))), first); + assert_eq!(range_keys(&map, (Included(0), Included(1))), first); + assert_eq!(range_keys(&map, (Included(1), Excluded(2))), first); + assert_eq!(range_keys(&map, (Included(1), Included(1))), first); + assert_eq!(range_keys(&map, (Unbounded, Excluded(2))), first); + assert_eq!(range_keys(&map, (Unbounded, Included(1))), first); + assert_eq!(range_keys(&map, (Excluded(size - 1), Excluded(size + 1))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Included(size + 1))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Included(size))), last); + assert_eq!(range_keys(&map, (Excluded(size - 1), Unbounded)), last); + assert_eq!(range_keys(&map, (Included(size), Excluded(size + 1))), last); + assert_eq!(range_keys(&map, (Included(size), Included(size + 1))), last); + assert_eq!(range_keys(&map, (Included(size), Included(size))), last); + assert_eq!(range_keys(&map, (Included(size), Unbounded)), last); + assert_eq!(range_keys(&map, (Excluded(size), Excluded(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Excluded(size), Included(size))), vec![]); + assert_eq!(range_keys(&map, (Excluded(size), Unbounded)), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Excluded(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Included(size + 1))), vec![]); + assert_eq!(range_keys(&map, (Included(size + 1), Unbounded)), vec![]); + + fn check<'a, L, R>(lhs: L, rhs: R) + where + L: IntoIterator<Item = (&'a i32, &'a i32)>, + R: IntoIterator<Item = (&'a i32, &'a i32)>, + { + assert_eq!(Vec::from_iter(lhs), Vec::from_iter(rhs)); + } + + check(map.range(..=100), map.range(..101)); + check(map.range(5..=8), vec![(&5, &5), (&6, &6), (&7, &7), (&8, &8)]); + check(map.range(-1..=2), vec![(&1, &1), (&2, &2)]); +} + +#[test] +fn test_range_inclusive_max_value() { + let max = usize::MAX; + let map = BTreeMap::from([(max, 0)]); + assert_eq!(Vec::from_iter(map.range(max..=max)), &[(&max, &0)]); +} + +#[test] +fn test_range_equal_empty_cases() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + assert_eq!(map.range((Included(2), Excluded(2))).next(), None); + assert_eq!(map.range((Excluded(2), Included(2))).next(), None); +} + +#[test] +#[should_panic] +fn test_range_equal_excluded() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + let _ = map.range((Excluded(2), Excluded(2))); +} + +#[test] +#[should_panic] +fn test_range_backwards_1() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + let _ = map.range((Included(3), Included(2))); +} + +#[test] +#[should_panic] +fn test_range_backwards_2() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + let _ = map.range((Included(3), Excluded(2))); +} + +#[test] +#[should_panic] +fn test_range_backwards_3() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + let _ = map.range((Excluded(3), Included(2))); +} + +#[test] +#[should_panic] +fn test_range_backwards_4() { + let map = BTreeMap::from_iter((0..5).map(|i| (i, i))); + let _ = map.range((Excluded(3), Excluded(2))); +} + +#[test] +fn test_range_finding_ill_order_in_map() { + let mut map = BTreeMap::new(); + map.insert(Cyclic3::B, ()); + // Lacking static_assert, call `range` conditionally, to emphasise that + // we cause a different panic than `test_range_backwards_1` does. + // A more refined `should_panic` would be welcome. + if Cyclic3::C < Cyclic3::A { + let _ = map.range(Cyclic3::C..=Cyclic3::A); + } +} + +#[test] +fn test_range_finding_ill_order_in_range_ord() { + // Has proper order the first time asked, then flips around. + struct EvilTwin(i32); + + impl PartialOrd for EvilTwin { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + Some(self.cmp(other)) + } + } + + static COMPARES: AtomicUsize = AtomicUsize::new(0); + impl Ord for EvilTwin { + fn cmp(&self, other: &Self) -> Ordering { + let ord = self.0.cmp(&other.0); + if COMPARES.fetch_add(1, SeqCst) > 0 { ord.reverse() } else { ord } + } + } + + impl PartialEq for EvilTwin { + fn eq(&self, other: &Self) -> bool { + self.0.eq(&other.0) + } + } + + impl Eq for EvilTwin {} + + #[derive(PartialEq, Eq, PartialOrd, Ord)] + struct CompositeKey(i32, EvilTwin); + + impl Borrow<EvilTwin> for CompositeKey { + fn borrow(&self) -> &EvilTwin { + &self.1 + } + } + + let map = BTreeMap::from_iter((0..12).map(|i| (CompositeKey(i, EvilTwin(i)), ()))); + let _ = map.range(EvilTwin(5)..=EvilTwin(7)); +} + +#[test] +fn test_range_1000() { + // Miri is too slow + let size = if cfg!(miri) { MIN_INSERTS_HEIGHT_2 as u32 } else { 1000 }; + let map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + fn test(map: &BTreeMap<u32, u32>, size: u32, min: Bound<&u32>, max: Bound<&u32>) { + let mut kvs = map.range((min, max)).map(|(&k, &v)| (k, v)); + let mut pairs = (0..size).map(|i| (i, i)); + + for (kv, pair) in kvs.by_ref().zip(pairs.by_ref()) { + assert_eq!(kv, pair); + } + assert_eq!(kvs.next(), None); + assert_eq!(pairs.next(), None); + } + test(&map, size, Included(&0), Excluded(&size)); + test(&map, size, Unbounded, Excluded(&size)); + test(&map, size, Included(&0), Included(&(size - 1))); + test(&map, size, Unbounded, Included(&(size - 1))); + test(&map, size, Included(&0), Unbounded); + test(&map, size, Unbounded, Unbounded); +} + +#[test] +fn test_range_borrowed_key() { + let mut map = BTreeMap::new(); + map.insert("aardvark".to_string(), 1); + map.insert("baboon".to_string(), 2); + map.insert("coyote".to_string(), 3); + map.insert("dingo".to_string(), 4); + // NOTE: would like to use simply "b".."d" here... + let mut iter = map.range::<str, _>((Included("b"), Excluded("d"))); + assert_eq!(iter.next(), Some((&"baboon".to_string(), &2))); + assert_eq!(iter.next(), Some((&"coyote".to_string(), &3))); + assert_eq!(iter.next(), None); +} + +#[test] +fn test_range() { + let size = 200; + // Miri is too slow + let step = if cfg!(miri) { 66 } else { 1 }; + let map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + for i in (0..size).step_by(step) { + for j in (i..size).step_by(step) { + let mut kvs = map.range((Included(&i), Included(&j))).map(|(&k, &v)| (k, v)); + let mut pairs = (i..=j).map(|i| (i, i)); + + for (kv, pair) in kvs.by_ref().zip(pairs.by_ref()) { + assert_eq!(kv, pair); + } + assert_eq!(kvs.next(), None); + assert_eq!(pairs.next(), None); + } + } +} + +#[test] +fn test_range_mut() { + let size = 200; + // Miri is too slow + let step = if cfg!(miri) { 66 } else { 1 }; + let mut map = BTreeMap::from_iter((0..size).map(|i| (i, i))); + + for i in (0..size).step_by(step) { + for j in (i..size).step_by(step) { + let mut kvs = map.range_mut((Included(&i), Included(&j))).map(|(&k, &mut v)| (k, v)); + let mut pairs = (i..=j).map(|i| (i, i)); + + for (kv, pair) in kvs.by_ref().zip(pairs.by_ref()) { + assert_eq!(kv, pair); + } + assert_eq!(kvs.next(), None); + assert_eq!(pairs.next(), None); + } + } + map.check(); +} + +#[should_panic(expected = "range start is greater than range end in BTreeMap")] +#[test] +fn test_range_panic_1() { + let mut map = BTreeMap::new(); + map.insert(3, "a"); + map.insert(5, "b"); + map.insert(8, "c"); + + let _invalid_range = map.range((Included(&8), Included(&3))); +} + +#[should_panic(expected = "range start and end are equal and excluded in BTreeMap")] +#[test] +fn test_range_panic_2() { + let mut map = BTreeMap::new(); + map.insert(3, "a"); + map.insert(5, "b"); + map.insert(8, "c"); + + let _invalid_range = map.range((Excluded(&5), Excluded(&5))); +} + +#[should_panic(expected = "range start and end are equal and excluded in BTreeMap")] +#[test] +fn test_range_panic_3() { + let mut map: BTreeMap<i32, ()> = BTreeMap::new(); + map.insert(3, ()); + map.insert(5, ()); + map.insert(8, ()); + + let _invalid_range = map.range((Excluded(&5), Excluded(&5))); +} + +#[test] +fn test_retain() { + let mut map = BTreeMap::from_iter((0..100).map(|x| (x, x * 10))); + + map.retain(|&k, _| k % 2 == 0); + assert_eq!(map.len(), 50); + assert_eq!(map[&2], 20); + assert_eq!(map[&4], 40); + assert_eq!(map[&6], 60); +} + +mod test_drain_filter { + use super::*; + + #[test] + fn empty() { + let mut map: BTreeMap<i32, i32> = BTreeMap::new(); + map.drain_filter(|_, _| unreachable!("there's nothing to decide on")); + assert_eq!(map.height(), None); + map.check(); + } + + // Explicitly consumes the iterator, where most test cases drop it instantly. + #[test] + fn consumed_keeping_all() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + assert!(map.drain_filter(|_, _| false).eq(iter::empty())); + map.check(); + } + + // Explicitly consumes the iterator, where most test cases drop it instantly. + #[test] + fn consumed_removing_all() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs.clone()); + assert!(map.drain_filter(|_, _| true).eq(pairs)); + assert!(map.is_empty()); + map.check(); + } + + // Explicitly consumes the iterator and modifies values through it. + #[test] + fn mutating_and_keeping() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + assert!( + map.drain_filter(|_, v| { + *v += 6; + false + }) + .eq(iter::empty()) + ); + assert!(map.keys().copied().eq(0..3)); + assert!(map.values().copied().eq(6..9)); + map.check(); + } + + // Explicitly consumes the iterator and modifies values through it. + #[test] + fn mutating_and_removing() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + assert!( + map.drain_filter(|_, v| { + *v += 6; + true + }) + .eq((0..3).map(|i| (i, i + 6))) + ); + assert!(map.is_empty()); + map.check(); + } + + #[test] + fn underfull_keeping_all() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| false); + assert!(map.keys().copied().eq(0..3)); + map.check(); + } + + #[test] + fn underfull_removing_one() { + let pairs = (0..3).map(|i| (i, i)); + for doomed in 0..3 { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i == doomed); + assert_eq!(map.len(), 2); + map.check(); + } + } + + #[test] + fn underfull_keeping_one() { + let pairs = (0..3).map(|i| (i, i)); + for sacred in 0..3 { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i != sacred); + assert!(map.keys().copied().eq(sacred..=sacred)); + map.check(); + } + } + + #[test] + fn underfull_removing_all() { + let pairs = (0..3).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| true); + assert!(map.is_empty()); + map.check(); + } + + #[test] + fn height_0_keeping_all() { + let pairs = (0..node::CAPACITY).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| false); + assert!(map.keys().copied().eq(0..node::CAPACITY)); + map.check(); + } + + #[test] + fn height_0_removing_one() { + let pairs = (0..node::CAPACITY).map(|i| (i, i)); + for doomed in 0..node::CAPACITY { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i == doomed); + assert_eq!(map.len(), node::CAPACITY - 1); + map.check(); + } + } + + #[test] + fn height_0_keeping_one() { + let pairs = (0..node::CAPACITY).map(|i| (i, i)); + for sacred in 0..node::CAPACITY { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i != sacred); + assert!(map.keys().copied().eq(sacred..=sacred)); + map.check(); + } + } + + #[test] + fn height_0_removing_all() { + let pairs = (0..node::CAPACITY).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| true); + assert!(map.is_empty()); + map.check(); + } + + #[test] + fn height_0_keeping_half() { + let mut map = BTreeMap::from_iter((0..16).map(|i| (i, i))); + assert_eq!(map.drain_filter(|i, _| *i % 2 == 0).count(), 8); + assert_eq!(map.len(), 8); + map.check(); + } + + #[test] + fn height_1_removing_all() { + let pairs = (0..MIN_INSERTS_HEIGHT_1).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| true); + assert!(map.is_empty()); + map.check(); + } + + #[test] + fn height_1_removing_one() { + let pairs = (0..MIN_INSERTS_HEIGHT_1).map(|i| (i, i)); + for doomed in 0..MIN_INSERTS_HEIGHT_1 { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i == doomed); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_1 - 1); + map.check(); + } + } + + #[test] + fn height_1_keeping_one() { + let pairs = (0..MIN_INSERTS_HEIGHT_1).map(|i| (i, i)); + for sacred in 0..MIN_INSERTS_HEIGHT_1 { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i != sacred); + assert!(map.keys().copied().eq(sacred..=sacred)); + map.check(); + } + } + + #[test] + fn height_2_removing_one() { + let pairs = (0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i)); + for doomed in (0..MIN_INSERTS_HEIGHT_2).step_by(12) { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i == doomed); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_2 - 1); + map.check(); + } + } + + #[test] + fn height_2_keeping_one() { + let pairs = (0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i)); + for sacred in (0..MIN_INSERTS_HEIGHT_2).step_by(12) { + let mut map = BTreeMap::from_iter(pairs.clone()); + map.drain_filter(|i, _| *i != sacred); + assert!(map.keys().copied().eq(sacred..=sacred)); + map.check(); + } + } + + #[test] + fn height_2_removing_all() { + let pairs = (0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i)); + let mut map = BTreeMap::from_iter(pairs); + map.drain_filter(|_, _| true); + assert!(map.is_empty()); + map.check(); + } + + #[test] + fn drop_panic_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut map = BTreeMap::new(); + map.insert(a.spawn(Panic::Never), ()); + map.insert(b.spawn(Panic::InDrop), ()); + map.insert(c.spawn(Panic::Never), ()); + + catch_unwind(move || drop(map.drain_filter(|dummy, _| dummy.query(true)))).unwrap_err(); + + assert_eq!(a.queried(), 1); + assert_eq!(b.queried(), 1); + assert_eq!(c.queried(), 0); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); + assert_eq!(c.dropped(), 1); + } + + #[test] + fn pred_panic_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut map = BTreeMap::new(); + map.insert(a.spawn(Panic::Never), ()); + map.insert(b.spawn(Panic::InQuery), ()); + map.insert(c.spawn(Panic::InQuery), ()); + + catch_unwind(AssertUnwindSafe(|| drop(map.drain_filter(|dummy, _| dummy.query(true))))) + .unwrap_err(); + + assert_eq!(a.queried(), 1); + assert_eq!(b.queried(), 1); + assert_eq!(c.queried(), 0); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 0); + assert_eq!(c.dropped(), 0); + assert_eq!(map.len(), 2); + assert_eq!(map.first_entry().unwrap().key().id(), 1); + assert_eq!(map.last_entry().unwrap().key().id(), 2); + map.check(); + } + + // Same as above, but attempt to use the iterator again after the panic in the predicate + #[test] + fn pred_panic_reuse() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut map = BTreeMap::new(); + map.insert(a.spawn(Panic::Never), ()); + map.insert(b.spawn(Panic::InQuery), ()); + map.insert(c.spawn(Panic::InQuery), ()); + + { + let mut it = map.drain_filter(|dummy, _| dummy.query(true)); + catch_unwind(AssertUnwindSafe(|| while it.next().is_some() {})).unwrap_err(); + // Iterator behaviour after a panic is explicitly unspecified, + // so this is just the current implementation: + let result = catch_unwind(AssertUnwindSafe(|| it.next())); + assert!(matches!(result, Ok(None))); + } + + assert_eq!(a.queried(), 1); + assert_eq!(b.queried(), 1); + assert_eq!(c.queried(), 0); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 0); + assert_eq!(c.dropped(), 0); + assert_eq!(map.len(), 2); + assert_eq!(map.first_entry().unwrap().key().id(), 1); + assert_eq!(map.last_entry().unwrap().key().id(), 2); + map.check(); + } +} + +#[test] +fn test_borrow() { + // make sure these compile -- using the Borrow trait + { + let mut map = BTreeMap::new(); + map.insert("0".to_string(), 1); + assert_eq!(map["0"], 1); + } + + { + let mut map = BTreeMap::new(); + map.insert(Box::new(0), 1); + assert_eq!(map[&0], 1); + } + + { + let mut map = BTreeMap::new(); + map.insert(Box::new([0, 1]) as Box<[i32]>, 1); + assert_eq!(map[&[0, 1][..]], 1); + } + + { + let mut map = BTreeMap::new(); + map.insert(Rc::new(0), 1); + assert_eq!(map[&0], 1); + } + + #[allow(dead_code)] + fn get<T: Ord>(v: &BTreeMap<Box<T>, ()>, t: &T) { + let _ = v.get(t); + } + + #[allow(dead_code)] + fn get_mut<T: Ord>(v: &mut BTreeMap<Box<T>, ()>, t: &T) { + let _ = v.get_mut(t); + } + + #[allow(dead_code)] + fn get_key_value<T: Ord>(v: &BTreeMap<Box<T>, ()>, t: &T) { + let _ = v.get_key_value(t); + } + + #[allow(dead_code)] + fn contains_key<T: Ord>(v: &BTreeMap<Box<T>, ()>, t: &T) { + let _ = v.contains_key(t); + } + + #[allow(dead_code)] + fn range<T: Ord>(v: &BTreeMap<Box<T>, ()>, t: T) { + let _ = v.range(t..); + } + + #[allow(dead_code)] + fn range_mut<T: Ord>(v: &mut BTreeMap<Box<T>, ()>, t: T) { + let _ = v.range_mut(t..); + } + + #[allow(dead_code)] + fn remove<T: Ord>(v: &mut BTreeMap<Box<T>, ()>, t: &T) { + v.remove(t); + } + + #[allow(dead_code)] + fn remove_entry<T: Ord>(v: &mut BTreeMap<Box<T>, ()>, t: &T) { + v.remove_entry(t); + } + + #[allow(dead_code)] + fn split_off<T: Ord>(v: &mut BTreeMap<Box<T>, ()>, t: &T) { + v.split_off(t); + } +} + +#[test] +fn test_entry() { + let xs = [(1, 10), (2, 20), (3, 30), (4, 40), (5, 50), (6, 60)]; + + let mut map = BTreeMap::from(xs); + + // Existing key (insert) + match map.entry(1) { + Vacant(_) => unreachable!(), + Occupied(mut view) => { + assert_eq!(view.get(), &10); + assert_eq!(view.insert(100), 10); + } + } + assert_eq!(map.get(&1).unwrap(), &100); + assert_eq!(map.len(), 6); + + // Existing key (update) + match map.entry(2) { + Vacant(_) => unreachable!(), + Occupied(mut view) => { + let v = view.get_mut(); + *v *= 10; + } + } + assert_eq!(map.get(&2).unwrap(), &200); + assert_eq!(map.len(), 6); + map.check(); + + // Existing key (take) + match map.entry(3) { + Vacant(_) => unreachable!(), + Occupied(view) => { + assert_eq!(view.remove(), 30); + } + } + assert_eq!(map.get(&3), None); + assert_eq!(map.len(), 5); + map.check(); + + // Inexistent key (insert) + match map.entry(10) { + Occupied(_) => unreachable!(), + Vacant(view) => { + assert_eq!(*view.insert(1000), 1000); + } + } + assert_eq!(map.get(&10).unwrap(), &1000); + assert_eq!(map.len(), 6); + map.check(); +} + +#[test] +fn test_extend_ref() { + let mut a = BTreeMap::new(); + a.insert(1, "one"); + let mut b = BTreeMap::new(); + b.insert(2, "two"); + b.insert(3, "three"); + + a.extend(&b); + + assert_eq!(a.len(), 3); + assert_eq!(a[&1], "one"); + assert_eq!(a[&2], "two"); + assert_eq!(a[&3], "three"); + a.check(); +} + +#[test] +fn test_zst() { + let mut m = BTreeMap::new(); + assert_eq!(m.len(), 0); + + assert_eq!(m.insert((), ()), None); + assert_eq!(m.len(), 1); + + assert_eq!(m.insert((), ()), Some(())); + assert_eq!(m.len(), 1); + assert_eq!(m.iter().count(), 1); + + m.clear(); + assert_eq!(m.len(), 0); + + for _ in 0..100 { + m.insert((), ()); + } + + assert_eq!(m.len(), 1); + assert_eq!(m.iter().count(), 1); + m.check(); +} + +// This test's only purpose is to ensure that zero-sized keys with nonsensical orderings +// do not cause segfaults when used with zero-sized values. All other map behavior is +// undefined. +#[test] +fn test_bad_zst() { + #[derive(Clone, Copy, Debug)] + struct Bad; + + impl PartialEq for Bad { + fn eq(&self, _: &Self) -> bool { + false + } + } + + impl Eq for Bad {} + + impl PartialOrd for Bad { + fn partial_cmp(&self, _: &Self) -> Option<Ordering> { + Some(Ordering::Less) + } + } + + impl Ord for Bad { + fn cmp(&self, _: &Self) -> Ordering { + Ordering::Less + } + } + + let mut m = BTreeMap::new(); + + for _ in 0..100 { + m.insert(Bad, Bad); + } + m.check(); +} + +#[test] +fn test_clear() { + let mut map = BTreeMap::new(); + for &len in &[MIN_INSERTS_HEIGHT_1, MIN_INSERTS_HEIGHT_2, 0, node::CAPACITY] { + for i in 0..len { + map.insert(i, ()); + } + assert_eq!(map.len(), len); + map.clear(); + map.check(); + assert_eq!(map.height(), None); + } +} + +#[test] +fn test_clear_drop_panic_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + + let mut map = BTreeMap::new(); + map.insert(a.spawn(Panic::Never), ()); + map.insert(b.spawn(Panic::InDrop), ()); + map.insert(c.spawn(Panic::Never), ()); + + catch_unwind(AssertUnwindSafe(|| map.clear())).unwrap_err(); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); + assert_eq!(c.dropped(), 1); + assert_eq!(map.len(), 0); + + drop(map); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); + assert_eq!(c.dropped(), 1); +} + +#[test] +fn test_clone() { + let mut map = BTreeMap::new(); + let size = MIN_INSERTS_HEIGHT_1; + assert_eq!(map.len(), 0); + + for i in 0..size { + assert_eq!(map.insert(i, 10 * i), None); + assert_eq!(map.len(), i + 1); + map.check(); + assert_eq!(map, map.clone()); + } + + for i in 0..size { + assert_eq!(map.insert(i, 100 * i), Some(10 * i)); + assert_eq!(map.len(), size); + map.check(); + assert_eq!(map, map.clone()); + } + + for i in 0..size / 2 { + assert_eq!(map.remove(&(i * 2)), Some(i * 200)); + assert_eq!(map.len(), size - i - 1); + map.check(); + assert_eq!(map, map.clone()); + } + + for i in 0..size / 2 { + assert_eq!(map.remove(&(2 * i)), None); + assert_eq!(map.remove(&(2 * i + 1)), Some(i * 200 + 100)); + assert_eq!(map.len(), size / 2 - i - 1); + map.check(); + assert_eq!(map, map.clone()); + } + + // Test a tree with 2 semi-full levels and a tree with 3 levels. + map = BTreeMap::from_iter((1..MIN_INSERTS_HEIGHT_2).map(|i| (i, i))); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_2 - 1); + assert_eq!(map, map.clone()); + map.insert(0, 0); + assert_eq!(map.len(), MIN_INSERTS_HEIGHT_2); + assert_eq!(map, map.clone()); + map.check(); +} + +fn test_clone_panic_leak(size: usize) { + for i in 0..size { + let dummies = Vec::from_iter((0..size).map(|id| CrashTestDummy::new(id))); + let map = BTreeMap::from_iter(dummies.iter().map(|dummy| { + let panic = if dummy.id == i { Panic::InClone } else { Panic::Never }; + (dummy.spawn(panic), ()) + })); + + catch_unwind(|| map.clone()).unwrap_err(); + for d in &dummies { + assert_eq!(d.cloned(), if d.id <= i { 1 } else { 0 }, "id={}/{}", d.id, i); + assert_eq!(d.dropped(), if d.id < i { 1 } else { 0 }, "id={}/{}", d.id, i); + } + assert_eq!(map.len(), size); + + drop(map); + for d in &dummies { + assert_eq!(d.cloned(), if d.id <= i { 1 } else { 0 }, "id={}/{}", d.id, i); + assert_eq!(d.dropped(), if d.id < i { 2 } else { 1 }, "id={}/{}", d.id, i); + } + } +} + +#[test] +fn test_clone_panic_leak_height_0() { + test_clone_panic_leak(3) +} + +#[test] +fn test_clone_panic_leak_height_1() { + test_clone_panic_leak(MIN_INSERTS_HEIGHT_1) +} + +#[test] +fn test_clone_from() { + let mut map1 = BTreeMap::new(); + let max_size = MIN_INSERTS_HEIGHT_1; + + // Range to max_size inclusive, because i is the size of map1 being tested. + for i in 0..=max_size { + let mut map2 = BTreeMap::new(); + for j in 0..i { + let mut map1_copy = map2.clone(); + map1_copy.clone_from(&map1); // small cloned from large + assert_eq!(map1_copy, map1); + let mut map2_copy = map1.clone(); + map2_copy.clone_from(&map2); // large cloned from small + assert_eq!(map2_copy, map2); + map2.insert(100 * j + 1, 2 * j + 1); + } + map2.clone_from(&map1); // same length + map2.check(); + assert_eq!(map2, map1); + map1.insert(i, 10 * i); + map1.check(); + } +} + +#[allow(dead_code)] +fn assert_covariance() { + fn map_key<'new>(v: BTreeMap<&'static str, ()>) -> BTreeMap<&'new str, ()> { + v + } + fn map_val<'new>(v: BTreeMap<(), &'static str>) -> BTreeMap<(), &'new str> { + v + } + + fn iter_key<'a, 'new>(v: Iter<'a, &'static str, ()>) -> Iter<'a, &'new str, ()> { + v + } + fn iter_val<'a, 'new>(v: Iter<'a, (), &'static str>) -> Iter<'a, (), &'new str> { + v + } + + fn into_iter_key<'new>(v: IntoIter<&'static str, ()>) -> IntoIter<&'new str, ()> { + v + } + fn into_iter_val<'new>(v: IntoIter<(), &'static str>) -> IntoIter<(), &'new str> { + v + } + + fn into_keys_key<'new>(v: IntoKeys<&'static str, ()>) -> IntoKeys<&'new str, ()> { + v + } + fn into_keys_val<'new>(v: IntoKeys<(), &'static str>) -> IntoKeys<(), &'new str> { + v + } + + fn into_values_key<'new>(v: IntoValues<&'static str, ()>) -> IntoValues<&'new str, ()> { + v + } + fn into_values_val<'new>(v: IntoValues<(), &'static str>) -> IntoValues<(), &'new str> { + v + } + + fn range_key<'a, 'new>(v: Range<'a, &'static str, ()>) -> Range<'a, &'new str, ()> { + v + } + fn range_val<'a, 'new>(v: Range<'a, (), &'static str>) -> Range<'a, (), &'new str> { + v + } + + fn keys_key<'a, 'new>(v: Keys<'a, &'static str, ()>) -> Keys<'a, &'new str, ()> { + v + } + fn keys_val<'a, 'new>(v: Keys<'a, (), &'static str>) -> Keys<'a, (), &'new str> { + v + } + + fn values_key<'a, 'new>(v: Values<'a, &'static str, ()>) -> Values<'a, &'new str, ()> { + v + } + fn values_val<'a, 'new>(v: Values<'a, (), &'static str>) -> Values<'a, (), &'new str> { + v + } +} + +#[allow(dead_code)] +fn assert_sync() { + fn map<T: Sync>(v: &BTreeMap<T, T>) -> impl Sync + '_ { + v + } + + fn into_iter<T: Sync>(v: BTreeMap<T, T>) -> impl Sync { + v.into_iter() + } + + fn into_keys<T: Sync + Ord>(v: BTreeMap<T, T>) -> impl Sync { + v.into_keys() + } + + fn into_values<T: Sync + Ord>(v: BTreeMap<T, T>) -> impl Sync { + v.into_values() + } + + fn drain_filter<T: Sync + Ord>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + v.drain_filter(|_, _| false) + } + + fn iter<T: Sync>(v: &BTreeMap<T, T>) -> impl Sync + '_ { + v.iter() + } + + fn iter_mut<T: Sync>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + v.iter_mut() + } + + fn keys<T: Sync>(v: &BTreeMap<T, T>) -> impl Sync + '_ { + v.keys() + } + + fn values<T: Sync>(v: &BTreeMap<T, T>) -> impl Sync + '_ { + v.values() + } + + fn values_mut<T: Sync>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + v.values_mut() + } + + fn range<T: Sync + Ord>(v: &BTreeMap<T, T>) -> impl Sync + '_ { + v.range(..) + } + + fn range_mut<T: Sync + Ord>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + v.range_mut(..) + } + + fn entry<T: Sync + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + v.entry(Default::default()) + } + + fn occupied_entry<T: Sync + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + match v.entry(Default::default()) { + Occupied(entry) => entry, + _ => unreachable!(), + } + } + + fn vacant_entry<T: Sync + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Sync + '_ { + match v.entry(Default::default()) { + Vacant(entry) => entry, + _ => unreachable!(), + } + } +} + +#[allow(dead_code)] +fn assert_send() { + fn map<T: Send>(v: BTreeMap<T, T>) -> impl Send { + v + } + + fn into_iter<T: Send>(v: BTreeMap<T, T>) -> impl Send { + v.into_iter() + } + + fn into_keys<T: Send + Ord>(v: BTreeMap<T, T>) -> impl Send { + v.into_keys() + } + + fn into_values<T: Send + Ord>(v: BTreeMap<T, T>) -> impl Send { + v.into_values() + } + + fn drain_filter<T: Send + Ord>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + v.drain_filter(|_, _| false) + } + + fn iter<T: Send + Sync>(v: &BTreeMap<T, T>) -> impl Send + '_ { + v.iter() + } + + fn iter_mut<T: Send>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + v.iter_mut() + } + + fn keys<T: Send + Sync>(v: &BTreeMap<T, T>) -> impl Send + '_ { + v.keys() + } + + fn values<T: Send + Sync>(v: &BTreeMap<T, T>) -> impl Send + '_ { + v.values() + } + + fn values_mut<T: Send>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + v.values_mut() + } + + fn range<T: Send + Sync + Ord>(v: &BTreeMap<T, T>) -> impl Send + '_ { + v.range(..) + } + + fn range_mut<T: Send + Ord>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + v.range_mut(..) + } + + fn entry<T: Send + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + v.entry(Default::default()) + } + + fn occupied_entry<T: Send + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + match v.entry(Default::default()) { + Occupied(entry) => entry, + _ => unreachable!(), + } + } + + fn vacant_entry<T: Send + Ord + Default>(v: &mut BTreeMap<T, T>) -> impl Send + '_ { + match v.entry(Default::default()) { + Vacant(entry) => entry, + _ => unreachable!(), + } + } +} + +#[test] +fn test_ord_absence() { + fn map<K>(mut map: BTreeMap<K, ()>) { + let _ = map.is_empty(); + let _ = map.len(); + map.clear(); + let _ = map.iter(); + let _ = map.iter_mut(); + let _ = map.keys(); + let _ = map.values(); + let _ = map.values_mut(); + if true { + let _ = map.into_values(); + } else if true { + let _ = map.into_iter(); + } else { + let _ = map.into_keys(); + } + } + + fn map_debug<K: Debug>(mut map: BTreeMap<K, ()>) { + format!("{map:?}"); + format!("{:?}", map.iter()); + format!("{:?}", map.iter_mut()); + format!("{:?}", map.keys()); + format!("{:?}", map.values()); + format!("{:?}", map.values_mut()); + if true { + format!("{:?}", map.into_iter()); + } else if true { + format!("{:?}", map.into_keys()); + } else { + format!("{:?}", map.into_values()); + } + } + + fn map_clone<K: Clone>(mut map: BTreeMap<K, ()>) { + map.clone_from(&map.clone()); + } + + #[derive(Debug, Clone)] + struct NonOrd; + map(BTreeMap::<NonOrd, _>::new()); + map_debug(BTreeMap::<NonOrd, _>::new()); + map_clone(BTreeMap::<NonOrd, _>::default()); +} + +#[test] +fn test_occupied_entry_key() { + let mut a = BTreeMap::new(); + let key = "hello there"; + let value = "value goes here"; + assert_eq!(a.height(), None); + a.insert(key, value); + assert_eq!(a.len(), 1); + assert_eq!(a[key], value); + + match a.entry(key) { + Vacant(_) => panic!(), + Occupied(e) => assert_eq!(key, *e.key()), + } + assert_eq!(a.len(), 1); + assert_eq!(a[key], value); + a.check(); +} + +#[test] +fn test_vacant_entry_key() { + let mut a = BTreeMap::new(); + let key = "hello there"; + let value = "value goes here"; + + assert_eq!(a.height(), None); + match a.entry(key) { + Occupied(_) => unreachable!(), + Vacant(e) => { + assert_eq!(key, *e.key()); + e.insert(value); + } + } + assert_eq!(a.len(), 1); + assert_eq!(a[key], value); + a.check(); +} + +#[test] +fn test_vacant_entry_no_insert() { + let mut a = BTreeMap::<&str, ()>::new(); + let key = "hello there"; + + // Non-allocated + assert_eq!(a.height(), None); + match a.entry(key) { + Occupied(_) => unreachable!(), + Vacant(e) => assert_eq!(key, *e.key()), + } + // Ensures the tree has no root. + assert_eq!(a.height(), None); + a.check(); + + // Allocated but still empty + a.insert(key, ()); + a.remove(&key); + assert_eq!(a.height(), Some(0)); + assert!(a.is_empty()); + match a.entry(key) { + Occupied(_) => unreachable!(), + Vacant(e) => assert_eq!(key, *e.key()), + } + // Ensures the allocated root is not changed. + assert_eq!(a.height(), Some(0)); + assert!(a.is_empty()); + a.check(); +} + +#[test] +fn test_first_last_entry() { + let mut a = BTreeMap::new(); + assert!(a.first_entry().is_none()); + assert!(a.last_entry().is_none()); + a.insert(1, 42); + assert_eq!(a.first_entry().unwrap().key(), &1); + assert_eq!(a.last_entry().unwrap().key(), &1); + a.insert(2, 24); + assert_eq!(a.first_entry().unwrap().key(), &1); + assert_eq!(a.last_entry().unwrap().key(), &2); + a.insert(0, 6); + assert_eq!(a.first_entry().unwrap().key(), &0); + assert_eq!(a.last_entry().unwrap().key(), &2); + let (k1, v1) = a.first_entry().unwrap().remove_entry(); + assert_eq!(k1, 0); + assert_eq!(v1, 6); + let (k2, v2) = a.last_entry().unwrap().remove_entry(); + assert_eq!(k2, 2); + assert_eq!(v2, 24); + assert_eq!(a.first_entry().unwrap().key(), &1); + assert_eq!(a.last_entry().unwrap().key(), &1); + a.check(); +} + +#[test] +fn test_pop_first_last() { + let mut map = BTreeMap::new(); + assert_eq!(map.pop_first(), None); + assert_eq!(map.pop_last(), None); + + map.insert(1, 10); + map.insert(2, 20); + map.insert(3, 30); + map.insert(4, 40); + + assert_eq!(map.len(), 4); + + let (key, val) = map.pop_first().unwrap(); + assert_eq!(key, 1); + assert_eq!(val, 10); + assert_eq!(map.len(), 3); + + let (key, val) = map.pop_first().unwrap(); + assert_eq!(key, 2); + assert_eq!(val, 20); + assert_eq!(map.len(), 2); + let (key, val) = map.pop_last().unwrap(); + assert_eq!(key, 4); + assert_eq!(val, 40); + assert_eq!(map.len(), 1); + + map.insert(5, 50); + map.insert(6, 60); + assert_eq!(map.len(), 3); + + let (key, val) = map.pop_first().unwrap(); + assert_eq!(key, 3); + assert_eq!(val, 30); + assert_eq!(map.len(), 2); + + let (key, val) = map.pop_last().unwrap(); + assert_eq!(key, 6); + assert_eq!(val, 60); + assert_eq!(map.len(), 1); + + let (key, val) = map.pop_last().unwrap(); + assert_eq!(key, 5); + assert_eq!(val, 50); + assert_eq!(map.len(), 0); + + assert_eq!(map.pop_first(), None); + assert_eq!(map.pop_last(), None); + + map.insert(7, 70); + map.insert(8, 80); + + let (key, val) = map.pop_last().unwrap(); + assert_eq!(key, 8); + assert_eq!(val, 80); + assert_eq!(map.len(), 1); + + let (key, val) = map.pop_last().unwrap(); + assert_eq!(key, 7); + assert_eq!(val, 70); + assert_eq!(map.len(), 0); + + assert_eq!(map.pop_first(), None); + assert_eq!(map.pop_last(), None); +} + +#[test] +fn test_get_key_value() { + let mut map = BTreeMap::new(); + + assert!(map.is_empty()); + assert_eq!(map.get_key_value(&1), None); + assert_eq!(map.get_key_value(&2), None); + + map.insert(1, 10); + map.insert(2, 20); + map.insert(3, 30); + + assert_eq!(map.len(), 3); + assert_eq!(map.get_key_value(&1), Some((&1, &10))); + assert_eq!(map.get_key_value(&3), Some((&3, &30))); + assert_eq!(map.get_key_value(&4), None); + + map.remove(&3); + + assert_eq!(map.len(), 2); + assert_eq!(map.get_key_value(&3), None); + assert_eq!(map.get_key_value(&2), Some((&2, &20))); +} + +#[test] +fn test_insert_into_full_height_0() { + let size = node::CAPACITY; + for pos in 0..=size { + let mut map = BTreeMap::from_iter((0..size).map(|i| (i * 2 + 1, ()))); + assert!(map.insert(pos * 2, ()).is_none()); + map.check(); + } +} + +#[test] +fn test_insert_into_full_height_1() { + let size = node::CAPACITY + 1 + node::CAPACITY; + for pos in 0..=size { + let mut map = BTreeMap::from_iter((0..size).map(|i| (i * 2 + 1, ()))); + map.compact(); + let root_node = map.root.as_ref().unwrap().reborrow(); + assert_eq!(root_node.len(), 1); + assert_eq!(root_node.first_leaf_edge().into_node().len(), node::CAPACITY); + assert_eq!(root_node.last_leaf_edge().into_node().len(), node::CAPACITY); + + assert!(map.insert(pos * 2, ()).is_none()); + map.check(); + } +} + +#[test] +fn test_try_insert() { + let mut map = BTreeMap::new(); + + assert!(map.is_empty()); + + assert_eq!(map.try_insert(1, 10).unwrap(), &10); + assert_eq!(map.try_insert(2, 20).unwrap(), &20); + + let err = map.try_insert(2, 200).unwrap_err(); + assert_eq!(err.entry.key(), &2); + assert_eq!(err.entry.get(), &20); + assert_eq!(err.value, 200); +} + +macro_rules! create_append_test { + ($name:ident, $len:expr) => { + #[test] + fn $name() { + let mut a = BTreeMap::new(); + for i in 0..8 { + a.insert(i, i); + } + + let mut b = BTreeMap::new(); + for i in 5..$len { + b.insert(i, 2 * i); + } + + a.append(&mut b); + + assert_eq!(a.len(), $len); + assert_eq!(b.len(), 0); + + for i in 0..$len { + if i < 5 { + assert_eq!(a[&i], i); + } else { + assert_eq!(a[&i], 2 * i); + } + } + + a.check(); + assert_eq!(a.remove(&($len - 1)), Some(2 * ($len - 1))); + assert_eq!(a.insert($len - 1, 20), None); + a.check(); + } + }; +} + +// These are mostly for testing the algorithm that "fixes" the right edge after insertion. +// Single node. +create_append_test!(test_append_9, 9); +// Two leafs that don't need fixing. +create_append_test!(test_append_17, 17); +// Two leafs where the second one ends up underfull and needs stealing at the end. +create_append_test!(test_append_14, 14); +// Two leafs where the second one ends up empty because the insertion finished at the root. +create_append_test!(test_append_12, 12); +// Three levels; insertion finished at the root. +create_append_test!(test_append_144, 144); +// Three levels; insertion finished at leaf while there is an empty node on the second level. +create_append_test!(test_append_145, 145); +// Tests for several randomly chosen sizes. +create_append_test!(test_append_170, 170); +create_append_test!(test_append_181, 181); +#[cfg(not(miri))] // Miri is too slow +create_append_test!(test_append_239, 239); +#[cfg(not(miri))] // Miri is too slow +create_append_test!(test_append_1700, 1700); + +#[test] +fn test_append_drop_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut left = BTreeMap::new(); + let mut right = BTreeMap::new(); + left.insert(a.spawn(Panic::Never), ()); + left.insert(b.spawn(Panic::InDrop), ()); // first duplicate key, dropped during append + left.insert(c.spawn(Panic::Never), ()); + right.insert(b.spawn(Panic::Never), ()); + right.insert(c.spawn(Panic::Never), ()); + + catch_unwind(move || left.append(&mut right)).unwrap_err(); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); // should be 2 were it not for Rust issue #47949 + assert_eq!(c.dropped(), 2); +} + +#[test] +fn test_append_ord_chaos() { + let mut map1 = BTreeMap::new(); + map1.insert(Cyclic3::A, ()); + map1.insert(Cyclic3::B, ()); + let mut map2 = BTreeMap::new(); + map2.insert(Cyclic3::A, ()); + map2.insert(Cyclic3::B, ()); + map2.insert(Cyclic3::C, ()); // lands first, before A + map2.insert(Cyclic3::B, ()); // lands first, before C + map1.check(); + map2.check(); // keys are not unique but still strictly ascending + assert_eq!(map1.len(), 2); + assert_eq!(map2.len(), 4); + map1.append(&mut map2); + assert_eq!(map1.len(), 5); + assert_eq!(map2.len(), 0); + map1.check(); + map2.check(); +} + +fn rand_data(len: usize) -> Vec<(u32, u32)> { + let mut rng = DeterministicRng::new(); + Vec::from_iter((0..len).map(|_| (rng.next(), rng.next()))) +} + +#[test] +fn test_split_off_empty_right() { + let mut data = rand_data(173); + + let mut map = BTreeMap::from_iter(data.clone()); + let right = map.split_off(&(data.iter().max().unwrap().0 + 1)); + map.check(); + right.check(); + + data.sort(); + assert!(map.into_iter().eq(data)); + assert!(right.into_iter().eq(None)); +} + +#[test] +fn test_split_off_empty_left() { + let mut data = rand_data(314); + + let mut map = BTreeMap::from_iter(data.clone()); + let right = map.split_off(&data.iter().min().unwrap().0); + map.check(); + right.check(); + + data.sort(); + assert!(map.into_iter().eq(None)); + assert!(right.into_iter().eq(data)); +} + +// In a tree with 3 levels, if all but a part of the first leaf node is split off, +// make sure fix_top eliminates both top levels. +#[test] +fn test_split_off_tiny_left_height_2() { + let pairs = (0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i)); + let mut left = BTreeMap::from_iter(pairs.clone()); + let right = left.split_off(&1); + left.check(); + right.check(); + assert_eq!(left.len(), 1); + assert_eq!(right.len(), MIN_INSERTS_HEIGHT_2 - 1); + assert_eq!(*left.first_key_value().unwrap().0, 0); + assert_eq!(*right.first_key_value().unwrap().0, 1); +} + +// In a tree with 3 levels, if only part of the last leaf node is split off, +// make sure fix_top eliminates both top levels. +#[test] +fn test_split_off_tiny_right_height_2() { + let pairs = (0..MIN_INSERTS_HEIGHT_2).map(|i| (i, i)); + let last = MIN_INSERTS_HEIGHT_2 - 1; + let mut left = BTreeMap::from_iter(pairs.clone()); + assert_eq!(*left.last_key_value().unwrap().0, last); + let right = left.split_off(&last); + left.check(); + right.check(); + assert_eq!(left.len(), MIN_INSERTS_HEIGHT_2 - 1); + assert_eq!(right.len(), 1); + assert_eq!(*left.last_key_value().unwrap().0, last - 1); + assert_eq!(*right.last_key_value().unwrap().0, last); +} + +#[test] +fn test_split_off_halfway() { + let mut rng = DeterministicRng::new(); + for &len in &[node::CAPACITY, 25, 50, 75, 100] { + let mut data = Vec::from_iter((0..len).map(|_| (rng.next(), ()))); + // Insertion in non-ascending order creates some variation in node length. + let mut map = BTreeMap::from_iter(data.iter().copied()); + data.sort(); + let small_keys = data.iter().take(len / 2).map(|kv| kv.0); + let large_keys = data.iter().skip(len / 2).map(|kv| kv.0); + let split_key = large_keys.clone().next().unwrap(); + let right = map.split_off(&split_key); + map.check(); + right.check(); + assert!(map.keys().copied().eq(small_keys)); + assert!(right.keys().copied().eq(large_keys)); + } +} + +#[test] +fn test_split_off_large_random_sorted() { + // Miri is too slow + let mut data = if cfg!(miri) { rand_data(529) } else { rand_data(1529) }; + // special case with maximum height. + data.sort(); + + let mut map = BTreeMap::from_iter(data.clone()); + let key = data[data.len() / 2].0; + let right = map.split_off(&key); + map.check(); + right.check(); + + assert!(map.into_iter().eq(data.clone().into_iter().filter(|x| x.0 < key))); + assert!(right.into_iter().eq(data.into_iter().filter(|x| x.0 >= key))); +} + +#[test] +fn test_into_iter_drop_leak_height_0() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let d = CrashTestDummy::new(3); + let e = CrashTestDummy::new(4); + let mut map = BTreeMap::new(); + map.insert("a", a.spawn(Panic::Never)); + map.insert("b", b.spawn(Panic::Never)); + map.insert("c", c.spawn(Panic::Never)); + map.insert("d", d.spawn(Panic::InDrop)); + map.insert("e", e.spawn(Panic::Never)); + + catch_unwind(move || drop(map.into_iter())).unwrap_err(); + + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); + assert_eq!(c.dropped(), 1); + assert_eq!(d.dropped(), 1); + assert_eq!(e.dropped(), 1); +} + +#[test] +fn test_into_iter_drop_leak_height_1() { + let size = MIN_INSERTS_HEIGHT_1; + for panic_point in vec![0, 1, size - 2, size - 1] { + let dummies = Vec::from_iter((0..size).map(|i| CrashTestDummy::new(i))); + let map = BTreeMap::from_iter((0..size).map(|i| { + let panic = if i == panic_point { Panic::InDrop } else { Panic::Never }; + (dummies[i].spawn(Panic::Never), dummies[i].spawn(panic)) + })); + catch_unwind(move || drop(map.into_iter())).unwrap_err(); + for i in 0..size { + assert_eq!(dummies[i].dropped(), 2); + } + } +} + +#[test] +fn test_into_keys() { + let map = BTreeMap::from([(1, 'a'), (2, 'b'), (3, 'c')]); + let keys = Vec::from_iter(map.into_keys()); + + assert_eq!(keys.len(), 3); + assert!(keys.contains(&1)); + assert!(keys.contains(&2)); + assert!(keys.contains(&3)); +} + +#[test] +fn test_into_values() { + let map = BTreeMap::from([(1, 'a'), (2, 'b'), (3, 'c')]); + let values = Vec::from_iter(map.into_values()); + + assert_eq!(values.len(), 3); + assert!(values.contains(&'a')); + assert!(values.contains(&'b')); + assert!(values.contains(&'c')); +} + +#[test] +fn test_insert_remove_intertwined() { + let loops = if cfg!(miri) { 100 } else { 1_000_000 }; + let mut map = BTreeMap::new(); + let mut i = 1; + let offset = 165; // somewhat arbitrarily chosen to cover some code paths + for _ in 0..loops { + i = (i + offset) & 0xFF; + map.insert(i, i); + map.remove(&(0xFF - i)); + } + map.check(); +} + +#[test] +fn test_insert_remove_intertwined_ord_chaos() { + let loops = if cfg!(miri) { 100 } else { 1_000_000 }; + let gov = Governor::new(); + let mut map = BTreeMap::new(); + let mut i = 1; + let offset = 165; // more arbitrarily copied from above + for _ in 0..loops { + i = (i + offset) & 0xFF; + map.insert(Governed(i, &gov), ()); + map.remove(&Governed(0xFF - i, &gov)); + gov.flip(); + } + map.check_invariants(); +} + +#[test] +fn from_array() { + let map = BTreeMap::from([(1, 2), (3, 4)]); + let unordered_duplicates = BTreeMap::from([(3, 4), (1, 2), (1, 2)]); + assert_eq!(map, unordered_duplicates); +} diff --git a/library/alloc/src/collections/btree/mem.rs b/library/alloc/src/collections/btree/mem.rs new file mode 100644 index 000000000..e1363d1ae --- /dev/null +++ b/library/alloc/src/collections/btree/mem.rs @@ -0,0 +1,35 @@ +use core::intrinsics; +use core::mem; +use core::ptr; + +/// This replaces the value behind the `v` unique reference by calling the +/// relevant function. +/// +/// If a panic occurs in the `change` closure, the entire process will be aborted. +#[allow(dead_code)] // keep as illustration and for future use +#[inline] +pub fn take_mut<T>(v: &mut T, change: impl FnOnce(T) -> T) { + replace(v, |value| (change(value), ())) +} + +/// This replaces the value behind the `v` unique reference by calling the +/// relevant function, and returns a result obtained along the way. +/// +/// If a panic occurs in the `change` closure, the entire process will be aborted. +#[inline] +pub fn replace<T, R>(v: &mut T, change: impl FnOnce(T) -> (T, R)) -> R { + struct PanicGuard; + impl Drop for PanicGuard { + fn drop(&mut self) { + intrinsics::abort() + } + } + let guard = PanicGuard; + let value = unsafe { ptr::read(v) }; + let (new_value, ret) = change(value); + unsafe { + ptr::write(v, new_value); + } + mem::forget(guard); + ret +} diff --git a/library/alloc/src/collections/btree/merge_iter.rs b/library/alloc/src/collections/btree/merge_iter.rs new file mode 100644 index 000000000..7f23d93b9 --- /dev/null +++ b/library/alloc/src/collections/btree/merge_iter.rs @@ -0,0 +1,98 @@ +use core::cmp::Ordering; +use core::fmt::{self, Debug}; +use core::iter::FusedIterator; + +/// Core of an iterator that merges the output of two strictly ascending iterators, +/// for instance a union or a symmetric difference. +pub struct MergeIterInner<I: Iterator> { + a: I, + b: I, + peeked: Option<Peeked<I>>, +} + +/// Benchmarks faster than wrapping both iterators in a Peekable, +/// probably because we can afford to impose a FusedIterator bound. +#[derive(Clone, Debug)] +enum Peeked<I: Iterator> { + A(I::Item), + B(I::Item), +} + +impl<I: Iterator> Clone for MergeIterInner<I> +where + I: Clone, + I::Item: Clone, +{ + fn clone(&self) -> Self { + Self { a: self.a.clone(), b: self.b.clone(), peeked: self.peeked.clone() } + } +} + +impl<I: Iterator> Debug for MergeIterInner<I> +where + I: Debug, + I::Item: Debug, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("MergeIterInner").field(&self.a).field(&self.b).field(&self.peeked).finish() + } +} + +impl<I: Iterator> MergeIterInner<I> { + /// Creates a new core for an iterator merging a pair of sources. + pub fn new(a: I, b: I) -> Self { + MergeIterInner { a, b, peeked: None } + } + + /// Returns the next pair of items stemming from the pair of sources + /// being merged. If both returned options contain a value, that value + /// is equal and occurs in both sources. If one of the returned options + /// contains a value, that value doesn't occur in the other source (or + /// the sources are not strictly ascending). If neither returned option + /// contains a value, iteration has finished and subsequent calls will + /// return the same empty pair. + pub fn nexts<Cmp: Fn(&I::Item, &I::Item) -> Ordering>( + &mut self, + cmp: Cmp, + ) -> (Option<I::Item>, Option<I::Item>) + where + I: FusedIterator, + { + let mut a_next; + let mut b_next; + match self.peeked.take() { + Some(Peeked::A(next)) => { + a_next = Some(next); + b_next = self.b.next(); + } + Some(Peeked::B(next)) => { + b_next = Some(next); + a_next = self.a.next(); + } + None => { + a_next = self.a.next(); + b_next = self.b.next(); + } + } + if let (Some(ref a1), Some(ref b1)) = (&a_next, &b_next) { + match cmp(a1, b1) { + Ordering::Less => self.peeked = b_next.take().map(Peeked::B), + Ordering::Greater => self.peeked = a_next.take().map(Peeked::A), + Ordering::Equal => (), + } + } + (a_next, b_next) + } + + /// Returns a pair of upper bounds for the `size_hint` of the final iterator. + pub fn lens(&self) -> (usize, usize) + where + I: ExactSizeIterator, + { + match self.peeked { + Some(Peeked::A(_)) => (1 + self.a.len(), self.b.len()), + Some(Peeked::B(_)) => (self.a.len(), 1 + self.b.len()), + _ => (self.a.len(), self.b.len()), + } + } +} diff --git a/library/alloc/src/collections/btree/mod.rs b/library/alloc/src/collections/btree/mod.rs new file mode 100644 index 000000000..9d43ac5c5 --- /dev/null +++ b/library/alloc/src/collections/btree/mod.rs @@ -0,0 +1,26 @@ +mod append; +mod borrow; +mod dedup_sorted_iter; +mod fix; +pub mod map; +mod mem; +mod merge_iter; +mod navigate; +mod node; +mod remove; +mod search; +pub mod set; +mod set_val; +mod split; + +#[doc(hidden)] +trait Recover<Q: ?Sized> { + type Key; + + fn get(&self, key: &Q) -> Option<&Self::Key>; + fn take(&mut self, key: &Q) -> Option<Self::Key>; + fn replace(&mut self, key: Self::Key) -> Option<Self::Key>; +} + +#[cfg(test)] +mod testing; diff --git a/library/alloc/src/collections/btree/navigate.rs b/library/alloc/src/collections/btree/navigate.rs new file mode 100644 index 000000000..1e33c1e64 --- /dev/null +++ b/library/alloc/src/collections/btree/navigate.rs @@ -0,0 +1,719 @@ +use core::borrow::Borrow; +use core::hint; +use core::ops::RangeBounds; +use core::ptr; + +use super::node::{marker, ForceResult::*, Handle, NodeRef}; + +use crate::alloc::Allocator; +// `front` and `back` are always both `None` or both `Some`. +pub struct LeafRange<BorrowType, K, V> { + front: Option<Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>>, + back: Option<Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>>, +} + +impl<'a, K: 'a, V: 'a> Clone for LeafRange<marker::Immut<'a>, K, V> { + fn clone(&self) -> Self { + LeafRange { front: self.front.clone(), back: self.back.clone() } + } +} + +impl<BorrowType, K, V> LeafRange<BorrowType, K, V> { + pub fn none() -> Self { + LeafRange { front: None, back: None } + } + + fn is_empty(&self) -> bool { + self.front == self.back + } + + /// Temporarily takes out another, immutable equivalent of the same range. + pub fn reborrow(&self) -> LeafRange<marker::Immut<'_>, K, V> { + LeafRange { + front: self.front.as_ref().map(|f| f.reborrow()), + back: self.back.as_ref().map(|b| b.reborrow()), + } + } +} + +impl<'a, K, V> LeafRange<marker::Immut<'a>, K, V> { + #[inline] + pub fn next_checked(&mut self) -> Option<(&'a K, &'a V)> { + self.perform_next_checked(|kv| kv.into_kv()) + } + + #[inline] + pub fn next_back_checked(&mut self) -> Option<(&'a K, &'a V)> { + self.perform_next_back_checked(|kv| kv.into_kv()) + } +} + +impl<'a, K, V> LeafRange<marker::ValMut<'a>, K, V> { + #[inline] + pub fn next_checked(&mut self) -> Option<(&'a K, &'a mut V)> { + self.perform_next_checked(|kv| unsafe { ptr::read(kv) }.into_kv_valmut()) + } + + #[inline] + pub fn next_back_checked(&mut self) -> Option<(&'a K, &'a mut V)> { + self.perform_next_back_checked(|kv| unsafe { ptr::read(kv) }.into_kv_valmut()) + } +} + +impl<BorrowType: marker::BorrowType, K, V> LeafRange<BorrowType, K, V> { + /// If possible, extract some result from the following KV and move to the edge beyond it. + fn perform_next_checked<F, R>(&mut self, f: F) -> Option<R> + where + F: Fn(&Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV>) -> R, + { + if self.is_empty() { + None + } else { + super::mem::replace(self.front.as_mut().unwrap(), |front| { + let kv = front.next_kv().ok().unwrap(); + let result = f(&kv); + (kv.next_leaf_edge(), Some(result)) + }) + } + } + + /// If possible, extract some result from the preceding KV and move to the edge beyond it. + fn perform_next_back_checked<F, R>(&mut self, f: F) -> Option<R> + where + F: Fn(&Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV>) -> R, + { + if self.is_empty() { + None + } else { + super::mem::replace(self.back.as_mut().unwrap(), |back| { + let kv = back.next_back_kv().ok().unwrap(); + let result = f(&kv); + (kv.next_back_leaf_edge(), Some(result)) + }) + } + } +} + +enum LazyLeafHandle<BorrowType, K, V> { + Root(NodeRef<BorrowType, K, V, marker::LeafOrInternal>), // not yet descended + Edge(Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>), +} + +impl<'a, K: 'a, V: 'a> Clone for LazyLeafHandle<marker::Immut<'a>, K, V> { + fn clone(&self) -> Self { + match self { + LazyLeafHandle::Root(root) => LazyLeafHandle::Root(*root), + LazyLeafHandle::Edge(edge) => LazyLeafHandle::Edge(*edge), + } + } +} + +impl<BorrowType, K, V> LazyLeafHandle<BorrowType, K, V> { + fn reborrow(&self) -> LazyLeafHandle<marker::Immut<'_>, K, V> { + match self { + LazyLeafHandle::Root(root) => LazyLeafHandle::Root(root.reborrow()), + LazyLeafHandle::Edge(edge) => LazyLeafHandle::Edge(edge.reborrow()), + } + } +} + +// `front` and `back` are always both `None` or both `Some`. +pub struct LazyLeafRange<BorrowType, K, V> { + front: Option<LazyLeafHandle<BorrowType, K, V>>, + back: Option<LazyLeafHandle<BorrowType, K, V>>, +} + +impl<'a, K: 'a, V: 'a> Clone for LazyLeafRange<marker::Immut<'a>, K, V> { + fn clone(&self) -> Self { + LazyLeafRange { front: self.front.clone(), back: self.back.clone() } + } +} + +impl<BorrowType, K, V> LazyLeafRange<BorrowType, K, V> { + pub fn none() -> Self { + LazyLeafRange { front: None, back: None } + } + + /// Temporarily takes out another, immutable equivalent of the same range. + pub fn reborrow(&self) -> LazyLeafRange<marker::Immut<'_>, K, V> { + LazyLeafRange { + front: self.front.as_ref().map(|f| f.reborrow()), + back: self.back.as_ref().map(|b| b.reborrow()), + } + } +} + +impl<'a, K, V> LazyLeafRange<marker::Immut<'a>, K, V> { + #[inline] + pub unsafe fn next_unchecked(&mut self) -> (&'a K, &'a V) { + unsafe { self.init_front().unwrap().next_unchecked() } + } + + #[inline] + pub unsafe fn next_back_unchecked(&mut self) -> (&'a K, &'a V) { + unsafe { self.init_back().unwrap().next_back_unchecked() } + } +} + +impl<'a, K, V> LazyLeafRange<marker::ValMut<'a>, K, V> { + #[inline] + pub unsafe fn next_unchecked(&mut self) -> (&'a K, &'a mut V) { + unsafe { self.init_front().unwrap().next_unchecked() } + } + + #[inline] + pub unsafe fn next_back_unchecked(&mut self) -> (&'a K, &'a mut V) { + unsafe { self.init_back().unwrap().next_back_unchecked() } + } +} + +impl<K, V> LazyLeafRange<marker::Dying, K, V> { + fn take_front( + &mut self, + ) -> Option<Handle<NodeRef<marker::Dying, K, V, marker::Leaf>, marker::Edge>> { + match self.front.take()? { + LazyLeafHandle::Root(root) => Some(root.first_leaf_edge()), + LazyLeafHandle::Edge(edge) => Some(edge), + } + } + + #[inline] + pub unsafe fn deallocating_next_unchecked<A: Allocator + Clone>( + &mut self, + alloc: A, + ) -> Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV> { + debug_assert!(self.front.is_some()); + let front = self.init_front().unwrap(); + unsafe { front.deallocating_next_unchecked(alloc) } + } + + #[inline] + pub unsafe fn deallocating_next_back_unchecked<A: Allocator + Clone>( + &mut self, + alloc: A, + ) -> Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV> { + debug_assert!(self.back.is_some()); + let back = self.init_back().unwrap(); + unsafe { back.deallocating_next_back_unchecked(alloc) } + } + + #[inline] + pub fn deallocating_end<A: Allocator + Clone>(&mut self, alloc: A) { + if let Some(front) = self.take_front() { + front.deallocating_end(alloc) + } + } +} + +impl<BorrowType: marker::BorrowType, K, V> LazyLeafRange<BorrowType, K, V> { + fn init_front( + &mut self, + ) -> Option<&mut Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>> { + if let Some(LazyLeafHandle::Root(root)) = &self.front { + self.front = Some(LazyLeafHandle::Edge(unsafe { ptr::read(root) }.first_leaf_edge())); + } + match &mut self.front { + None => None, + Some(LazyLeafHandle::Edge(edge)) => Some(edge), + // SAFETY: the code above would have replaced it. + Some(LazyLeafHandle::Root(_)) => unsafe { hint::unreachable_unchecked() }, + } + } + + fn init_back( + &mut self, + ) -> Option<&mut Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>> { + if let Some(LazyLeafHandle::Root(root)) = &self.back { + self.back = Some(LazyLeafHandle::Edge(unsafe { ptr::read(root) }.last_leaf_edge())); + } + match &mut self.back { + None => None, + Some(LazyLeafHandle::Edge(edge)) => Some(edge), + // SAFETY: the code above would have replaced it. + Some(LazyLeafHandle::Root(_)) => unsafe { hint::unreachable_unchecked() }, + } + } +} + +impl<BorrowType: marker::BorrowType, K, V> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + /// Finds the distinct leaf edges delimiting a specified range in a tree. + /// + /// If such distinct edges exist, returns them in ascending order, meaning + /// that a non-zero number of calls to `next_unchecked` on the `front` of + /// the result and/or calls to `next_back_unchecked` on the `back` of the + /// result will eventually reach the same edge. + /// + /// If there are no such edges, i.e., if the tree contains no key within + /// the range, returns an empty `front` and `back`. + /// + /// # Safety + /// Unless `BorrowType` is `Immut`, do not use the handles to visit the same + /// KV twice. + unsafe fn find_leaf_edges_spanning_range<Q: ?Sized, R>( + self, + range: R, + ) -> LeafRange<BorrowType, K, V> + where + Q: Ord, + K: Borrow<Q>, + R: RangeBounds<Q>, + { + match self.search_tree_for_bifurcation(&range) { + Err(_) => LeafRange::none(), + Ok(( + node, + lower_edge_idx, + upper_edge_idx, + mut lower_child_bound, + mut upper_child_bound, + )) => { + let mut lower_edge = unsafe { Handle::new_edge(ptr::read(&node), lower_edge_idx) }; + let mut upper_edge = unsafe { Handle::new_edge(node, upper_edge_idx) }; + loop { + match (lower_edge.force(), upper_edge.force()) { + (Leaf(f), Leaf(b)) => return LeafRange { front: Some(f), back: Some(b) }, + (Internal(f), Internal(b)) => { + (lower_edge, lower_child_bound) = + f.descend().find_lower_bound_edge(lower_child_bound); + (upper_edge, upper_child_bound) = + b.descend().find_upper_bound_edge(upper_child_bound); + } + _ => unreachable!("BTreeMap has different depths"), + } + } + } + } + } +} + +fn full_range<BorrowType: marker::BorrowType, K, V>( + root1: NodeRef<BorrowType, K, V, marker::LeafOrInternal>, + root2: NodeRef<BorrowType, K, V, marker::LeafOrInternal>, +) -> LazyLeafRange<BorrowType, K, V> { + LazyLeafRange { + front: Some(LazyLeafHandle::Root(root1)), + back: Some(LazyLeafHandle::Root(root2)), + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Immut<'a>, K, V, marker::LeafOrInternal> { + /// Finds the pair of leaf edges delimiting a specific range in a tree. + /// + /// The result is meaningful only if the tree is ordered by key, like the tree + /// in a `BTreeMap` is. + pub fn range_search<Q, R>(self, range: R) -> LeafRange<marker::Immut<'a>, K, V> + where + Q: ?Sized + Ord, + K: Borrow<Q>, + R: RangeBounds<Q>, + { + // SAFETY: our borrow type is immutable. + unsafe { self.find_leaf_edges_spanning_range(range) } + } + + /// Finds the pair of leaf edges delimiting an entire tree. + pub fn full_range(self) -> LazyLeafRange<marker::Immut<'a>, K, V> { + full_range(self, self) + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::ValMut<'a>, K, V, marker::LeafOrInternal> { + /// Splits a unique reference into a pair of leaf edges delimiting a specified range. + /// The result are non-unique references allowing (some) mutation, which must be used + /// carefully. + /// + /// The result is meaningful only if the tree is ordered by key, like the tree + /// in a `BTreeMap` is. + /// + /// # Safety + /// Do not use the duplicate handles to visit the same KV twice. + pub fn range_search<Q, R>(self, range: R) -> LeafRange<marker::ValMut<'a>, K, V> + where + Q: ?Sized + Ord, + K: Borrow<Q>, + R: RangeBounds<Q>, + { + unsafe { self.find_leaf_edges_spanning_range(range) } + } + + /// Splits a unique reference into a pair of leaf edges delimiting the full range of the tree. + /// The results are non-unique references allowing mutation (of values only), so must be used + /// with care. + pub fn full_range(self) -> LazyLeafRange<marker::ValMut<'a>, K, V> { + // We duplicate the root NodeRef here -- we will never visit the same KV + // twice, and never end up with overlapping value references. + let self2 = unsafe { ptr::read(&self) }; + full_range(self, self2) + } +} + +impl<K, V> NodeRef<marker::Dying, K, V, marker::LeafOrInternal> { + /// Splits a unique reference into a pair of leaf edges delimiting the full range of the tree. + /// The results are non-unique references allowing massively destructive mutation, so must be + /// used with the utmost care. + pub fn full_range(self) -> LazyLeafRange<marker::Dying, K, V> { + // We duplicate the root NodeRef here -- we will never access it in a way + // that overlaps references obtained from the root. + let self2 = unsafe { ptr::read(&self) }; + full_range(self, self2) + } +} + +impl<BorrowType: marker::BorrowType, K, V> + Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> +{ + /// Given a leaf edge handle, returns [`Result::Ok`] with a handle to the neighboring KV + /// on the right side, which is either in the same leaf node or in an ancestor node. + /// If the leaf edge is the last one in the tree, returns [`Result::Err`] with the root node. + pub fn next_kv( + self, + ) -> Result< + Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV>, + NodeRef<BorrowType, K, V, marker::LeafOrInternal>, + > { + let mut edge = self.forget_node_type(); + loop { + edge = match edge.right_kv() { + Ok(kv) => return Ok(kv), + Err(last_edge) => match last_edge.into_node().ascend() { + Ok(parent_edge) => parent_edge.forget_node_type(), + Err(root) => return Err(root), + }, + } + } + } + + /// Given a leaf edge handle, returns [`Result::Ok`] with a handle to the neighboring KV + /// on the left side, which is either in the same leaf node or in an ancestor node. + /// If the leaf edge is the first one in the tree, returns [`Result::Err`] with the root node. + fn next_back_kv( + self, + ) -> Result< + Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV>, + NodeRef<BorrowType, K, V, marker::LeafOrInternal>, + > { + let mut edge = self.forget_node_type(); + loop { + edge = match edge.left_kv() { + Ok(kv) => return Ok(kv), + Err(last_edge) => match last_edge.into_node().ascend() { + Ok(parent_edge) => parent_edge.forget_node_type(), + Err(root) => return Err(root), + }, + } + } + } +} + +impl<BorrowType: marker::BorrowType, K, V> + Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::Edge> +{ + /// Given an internal edge handle, returns [`Result::Ok`] with a handle to the neighboring KV + /// on the right side, which is either in the same internal node or in an ancestor node. + /// If the internal edge is the last one in the tree, returns [`Result::Err`] with the root node. + fn next_kv( + self, + ) -> Result< + Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::KV>, + NodeRef<BorrowType, K, V, marker::Internal>, + > { + let mut edge = self; + loop { + edge = match edge.right_kv() { + Ok(internal_kv) => return Ok(internal_kv), + Err(last_edge) => match last_edge.into_node().ascend() { + Ok(parent_edge) => parent_edge, + Err(root) => return Err(root), + }, + } + } + } +} + +impl<K, V> Handle<NodeRef<marker::Dying, K, V, marker::Leaf>, marker::Edge> { + /// Given a leaf edge handle into a dying tree, returns the next leaf edge + /// on the right side, and the key-value pair in between, if they exist. + /// + /// If the given edge is the last one in a leaf, this method deallocates + /// the leaf, as well as any ancestor nodes whose last edge was reached. + /// This implies that if no more key-value pair follows, the entire tree + /// will have been deallocated and there is nothing left to return. + /// + /// # Safety + /// - The given edge must not have been previously returned by counterpart + /// `deallocating_next_back`. + /// - The returned KV handle is only valid to access the key and value, + /// and only valid until the next call to a `deallocating_` method. + unsafe fn deallocating_next<A: Allocator + Clone>( + self, + alloc: A, + ) -> Option<(Self, Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV>)> + { + let mut edge = self.forget_node_type(); + loop { + edge = match edge.right_kv() { + Ok(kv) => return Some((unsafe { ptr::read(&kv) }.next_leaf_edge(), kv)), + Err(last_edge) => { + match unsafe { last_edge.into_node().deallocate_and_ascend(alloc.clone()) } { + Some(parent_edge) => parent_edge.forget_node_type(), + None => return None, + } + } + } + } + } + + /// Given a leaf edge handle into a dying tree, returns the next leaf edge + /// on the left side, and the key-value pair in between, if they exist. + /// + /// If the given edge is the first one in a leaf, this method deallocates + /// the leaf, as well as any ancestor nodes whose first edge was reached. + /// This implies that if no more key-value pair follows, the entire tree + /// will have been deallocated and there is nothing left to return. + /// + /// # Safety + /// - The given edge must not have been previously returned by counterpart + /// `deallocating_next`. + /// - The returned KV handle is only valid to access the key and value, + /// and only valid until the next call to a `deallocating_` method. + unsafe fn deallocating_next_back<A: Allocator + Clone>( + self, + alloc: A, + ) -> Option<(Self, Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV>)> + { + let mut edge = self.forget_node_type(); + loop { + edge = match edge.left_kv() { + Ok(kv) => return Some((unsafe { ptr::read(&kv) }.next_back_leaf_edge(), kv)), + Err(last_edge) => { + match unsafe { last_edge.into_node().deallocate_and_ascend(alloc.clone()) } { + Some(parent_edge) => parent_edge.forget_node_type(), + None => return None, + } + } + } + } + } + + /// Deallocates a pile of nodes from the leaf up to the root. + /// This is the only way to deallocate the remainder of a tree after + /// `deallocating_next` and `deallocating_next_back` have been nibbling at + /// both sides of the tree, and have hit the same edge. As it is intended + /// only to be called when all keys and values have been returned, + /// no cleanup is done on any of the keys or values. + fn deallocating_end<A: Allocator + Clone>(self, alloc: A) { + let mut edge = self.forget_node_type(); + while let Some(parent_edge) = + unsafe { edge.into_node().deallocate_and_ascend(alloc.clone()) } + { + edge = parent_edge.forget_node_type(); + } + } +} + +impl<'a, K, V> Handle<NodeRef<marker::Immut<'a>, K, V, marker::Leaf>, marker::Edge> { + /// Moves the leaf edge handle to the next leaf edge and returns references to the + /// key and value in between. + /// + /// # Safety + /// There must be another KV in the direction travelled. + unsafe fn next_unchecked(&mut self) -> (&'a K, &'a V) { + super::mem::replace(self, |leaf_edge| { + let kv = leaf_edge.next_kv().ok().unwrap(); + (kv.next_leaf_edge(), kv.into_kv()) + }) + } + + /// Moves the leaf edge handle to the previous leaf edge and returns references to the + /// key and value in between. + /// + /// # Safety + /// There must be another KV in the direction travelled. + unsafe fn next_back_unchecked(&mut self) -> (&'a K, &'a V) { + super::mem::replace(self, |leaf_edge| { + let kv = leaf_edge.next_back_kv().ok().unwrap(); + (kv.next_back_leaf_edge(), kv.into_kv()) + }) + } +} + +impl<'a, K, V> Handle<NodeRef<marker::ValMut<'a>, K, V, marker::Leaf>, marker::Edge> { + /// Moves the leaf edge handle to the next leaf edge and returns references to the + /// key and value in between. + /// + /// # Safety + /// There must be another KV in the direction travelled. + unsafe fn next_unchecked(&mut self) -> (&'a K, &'a mut V) { + let kv = super::mem::replace(self, |leaf_edge| { + let kv = leaf_edge.next_kv().ok().unwrap(); + (unsafe { ptr::read(&kv) }.next_leaf_edge(), kv) + }); + // Doing this last is faster, according to benchmarks. + kv.into_kv_valmut() + } + + /// Moves the leaf edge handle to the previous leaf and returns references to the + /// key and value in between. + /// + /// # Safety + /// There must be another KV in the direction travelled. + unsafe fn next_back_unchecked(&mut self) -> (&'a K, &'a mut V) { + let kv = super::mem::replace(self, |leaf_edge| { + let kv = leaf_edge.next_back_kv().ok().unwrap(); + (unsafe { ptr::read(&kv) }.next_back_leaf_edge(), kv) + }); + // Doing this last is faster, according to benchmarks. + kv.into_kv_valmut() + } +} + +impl<K, V> Handle<NodeRef<marker::Dying, K, V, marker::Leaf>, marker::Edge> { + /// Moves the leaf edge handle to the next leaf edge and returns the key and value + /// in between, deallocating any node left behind while leaving the corresponding + /// edge in its parent node dangling. + /// + /// # Safety + /// - There must be another KV in the direction travelled. + /// - That KV was not previously returned by counterpart + /// `deallocating_next_back_unchecked` on any copy of the handles + /// being used to traverse the tree. + /// + /// The only safe way to proceed with the updated handle is to compare it, drop it, + /// or call this method or counterpart `deallocating_next_back_unchecked` again. + unsafe fn deallocating_next_unchecked<A: Allocator + Clone>( + &mut self, + alloc: A, + ) -> Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV> { + super::mem::replace(self, |leaf_edge| unsafe { + leaf_edge.deallocating_next(alloc).unwrap() + }) + } + + /// Moves the leaf edge handle to the previous leaf edge and returns the key and value + /// in between, deallocating any node left behind while leaving the corresponding + /// edge in its parent node dangling. + /// + /// # Safety + /// - There must be another KV in the direction travelled. + /// - That leaf edge was not previously returned by counterpart + /// `deallocating_next_unchecked` on any copy of the handles + /// being used to traverse the tree. + /// + /// The only safe way to proceed with the updated handle is to compare it, drop it, + /// or call this method or counterpart `deallocating_next_unchecked` again. + unsafe fn deallocating_next_back_unchecked<A: Allocator + Clone>( + &mut self, + alloc: A, + ) -> Handle<NodeRef<marker::Dying, K, V, marker::LeafOrInternal>, marker::KV> { + super::mem::replace(self, |leaf_edge| unsafe { + leaf_edge.deallocating_next_back(alloc).unwrap() + }) + } +} + +impl<BorrowType: marker::BorrowType, K, V> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + /// Returns the leftmost leaf edge in or underneath a node - in other words, the edge + /// you need first when navigating forward (or last when navigating backward). + #[inline] + pub fn first_leaf_edge(self) -> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> { + let mut node = self; + loop { + match node.force() { + Leaf(leaf) => return leaf.first_edge(), + Internal(internal) => node = internal.first_edge().descend(), + } + } + } + + /// Returns the rightmost leaf edge in or underneath a node - in other words, the edge + /// you need last when navigating forward (or first when navigating backward). + #[inline] + pub fn last_leaf_edge(self) -> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> { + let mut node = self; + loop { + match node.force() { + Leaf(leaf) => return leaf.last_edge(), + Internal(internal) => node = internal.last_edge().descend(), + } + } + } +} + +pub enum Position<BorrowType, K, V> { + Leaf(NodeRef<BorrowType, K, V, marker::Leaf>), + Internal(NodeRef<BorrowType, K, V, marker::Internal>), + InternalKV(Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::KV>), +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Immut<'a>, K, V, marker::LeafOrInternal> { + /// Visits leaf nodes and internal KVs in order of ascending keys, and also + /// visits internal nodes as a whole in a depth first order, meaning that + /// internal nodes precede their individual KVs and their child nodes. + pub fn visit_nodes_in_order<F>(self, mut visit: F) + where + F: FnMut(Position<marker::Immut<'a>, K, V>), + { + match self.force() { + Leaf(leaf) => visit(Position::Leaf(leaf)), + Internal(internal) => { + visit(Position::Internal(internal)); + let mut edge = internal.first_edge(); + loop { + edge = match edge.descend().force() { + Leaf(leaf) => { + visit(Position::Leaf(leaf)); + match edge.next_kv() { + Ok(kv) => { + visit(Position::InternalKV(kv)); + kv.right_edge() + } + Err(_) => return, + } + } + Internal(internal) => { + visit(Position::Internal(internal)); + internal.first_edge() + } + } + } + } + } + } + + /// Calculates the number of elements in a (sub)tree. + pub fn calc_length(self) -> usize { + let mut result = 0; + self.visit_nodes_in_order(|pos| match pos { + Position::Leaf(node) => result += node.len(), + Position::Internal(node) => result += node.len(), + Position::InternalKV(_) => (), + }); + result + } +} + +impl<BorrowType: marker::BorrowType, K, V> + Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV> +{ + /// Returns the leaf edge closest to a KV for forward navigation. + pub fn next_leaf_edge(self) -> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> { + match self.force() { + Leaf(leaf_kv) => leaf_kv.right_edge(), + Internal(internal_kv) => { + let next_internal_edge = internal_kv.right_edge(); + next_internal_edge.descend().first_leaf_edge() + } + } + } + + /// Returns the leaf edge closest to a KV for backward navigation. + fn next_back_leaf_edge(self) -> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> { + match self.force() { + Leaf(leaf_kv) => leaf_kv.left_edge(), + Internal(internal_kv) => { + let next_internal_edge = internal_kv.left_edge(); + next_internal_edge.descend().last_leaf_edge() + } + } + } +} diff --git a/library/alloc/src/collections/btree/node.rs b/library/alloc/src/collections/btree/node.rs new file mode 100644 index 000000000..d831161bc --- /dev/null +++ b/library/alloc/src/collections/btree/node.rs @@ -0,0 +1,1753 @@ +// This is an attempt at an implementation following the ideal +// +// ``` +// struct BTreeMap<K, V> { +// height: usize, +// root: Option<Box<Node<K, V, height>>> +// } +// +// struct Node<K, V, height: usize> { +// keys: [K; 2 * B - 1], +// vals: [V; 2 * B - 1], +// edges: [if height > 0 { Box<Node<K, V, height - 1>> } else { () }; 2 * B], +// parent: Option<(NonNull<Node<K, V, height + 1>>, u16)>, +// len: u16, +// } +// ``` +// +// Since Rust doesn't actually have dependent types and polymorphic recursion, +// we make do with lots of unsafety. + +// A major goal of this module is to avoid complexity by treating the tree as a generic (if +// weirdly shaped) container and avoiding dealing with most of the B-Tree invariants. As such, +// this module doesn't care whether the entries are sorted, which nodes can be underfull, or +// even what underfull means. However, we do rely on a few invariants: +// +// - Trees must have uniform depth/height. This means that every path down to a leaf from a +// given node has exactly the same length. +// - A node of length `n` has `n` keys, `n` values, and `n + 1` edges. +// This implies that even an empty node has at least one edge. +// For a leaf node, "having an edge" only means we can identify a position in the node, +// since leaf edges are empty and need no data representation. In an internal node, +// an edge both identifies a position and contains a pointer to a child node. + +use core::marker::PhantomData; +use core::mem::{self, MaybeUninit}; +use core::ptr::{self, NonNull}; +use core::slice::SliceIndex; + +use crate::alloc::{Allocator, Layout}; +use crate::boxed::Box; + +const B: usize = 6; +pub const CAPACITY: usize = 2 * B - 1; +pub const MIN_LEN_AFTER_SPLIT: usize = B - 1; +const KV_IDX_CENTER: usize = B - 1; +const EDGE_IDX_LEFT_OF_CENTER: usize = B - 1; +const EDGE_IDX_RIGHT_OF_CENTER: usize = B; + +/// The underlying representation of leaf nodes and part of the representation of internal nodes. +struct LeafNode<K, V> { + /// We want to be covariant in `K` and `V`. + parent: Option<NonNull<InternalNode<K, V>>>, + + /// This node's index into the parent node's `edges` array. + /// `*node.parent.edges[node.parent_idx]` should be the same thing as `node`. + /// This is only guaranteed to be initialized when `parent` is non-null. + parent_idx: MaybeUninit<u16>, + + /// The number of keys and values this node stores. + len: u16, + + /// The arrays storing the actual data of the node. Only the first `len` elements of each + /// array are initialized and valid. + keys: [MaybeUninit<K>; CAPACITY], + vals: [MaybeUninit<V>; CAPACITY], +} + +impl<K, V> LeafNode<K, V> { + /// Initializes a new `LeafNode` in-place. + unsafe fn init(this: *mut Self) { + // As a general policy, we leave fields uninitialized if they can be, as this should + // be both slightly faster and easier to track in Valgrind. + unsafe { + // parent_idx, keys, and vals are all MaybeUninit + ptr::addr_of_mut!((*this).parent).write(None); + ptr::addr_of_mut!((*this).len).write(0); + } + } + + /// Creates a new boxed `LeafNode`. + fn new<A: Allocator + Clone>(alloc: A) -> Box<Self, A> { + unsafe { + let mut leaf = Box::new_uninit_in(alloc); + LeafNode::init(leaf.as_mut_ptr()); + leaf.assume_init() + } + } +} + +/// The underlying representation of internal nodes. As with `LeafNode`s, these should be hidden +/// behind `BoxedNode`s to prevent dropping uninitialized keys and values. Any pointer to an +/// `InternalNode` can be directly cast to a pointer to the underlying `LeafNode` portion of the +/// node, allowing code to act on leaf and internal nodes generically without having to even check +/// which of the two a pointer is pointing at. This property is enabled by the use of `repr(C)`. +#[repr(C)] +// gdb_providers.py uses this type name for introspection. +struct InternalNode<K, V> { + data: LeafNode<K, V>, + + /// The pointers to the children of this node. `len + 1` of these are considered + /// initialized and valid, except that near the end, while the tree is held + /// through borrow type `Dying`, some of these pointers are dangling. + edges: [MaybeUninit<BoxedNode<K, V>>; 2 * B], +} + +impl<K, V> InternalNode<K, V> { + /// Creates a new boxed `InternalNode`. + /// + /// # Safety + /// An invariant of internal nodes is that they have at least one + /// initialized and valid edge. This function does not set up + /// such an edge. + unsafe fn new<A: Allocator + Clone>(alloc: A) -> Box<Self, A> { + unsafe { + let mut node = Box::<Self, _>::new_uninit_in(alloc); + // We only need to initialize the data; the edges are MaybeUninit. + LeafNode::init(ptr::addr_of_mut!((*node.as_mut_ptr()).data)); + node.assume_init() + } + } +} + +/// A managed, non-null pointer to a node. This is either an owned pointer to +/// `LeafNode<K, V>` or an owned pointer to `InternalNode<K, V>`. +/// +/// However, `BoxedNode` contains no information as to which of the two types +/// of nodes it actually contains, and, partially due to this lack of information, +/// is not a separate type and has no destructor. +type BoxedNode<K, V> = NonNull<LeafNode<K, V>>; + +// N.B. `NodeRef` is always covariant in `K` and `V`, even when the `BorrowType` +// is `Mut`. This is technically wrong, but cannot result in any unsafety due to +// internal use of `NodeRef` because we stay completely generic over `K` and `V`. +// However, whenever a public type wraps `NodeRef`, make sure that it has the +// correct variance. +/// +/// A reference to a node. +/// +/// This type has a number of parameters that controls how it acts: +/// - `BorrowType`: A dummy type that describes the kind of borrow and carries a lifetime. +/// - When this is `Immut<'a>`, the `NodeRef` acts roughly like `&'a Node`. +/// - When this is `ValMut<'a>`, the `NodeRef` acts roughly like `&'a Node` +/// with respect to keys and tree structure, but also allows many +/// mutable references to values throughout the tree to coexist. +/// - When this is `Mut<'a>`, the `NodeRef` acts roughly like `&'a mut Node`, +/// although insert methods allow a mutable pointer to a value to coexist. +/// - When this is `Owned`, the `NodeRef` acts roughly like `Box<Node>`, +/// but does not have a destructor, and must be cleaned up manually. +/// - When this is `Dying`, the `NodeRef` still acts roughly like `Box<Node>`, +/// but has methods to destroy the tree bit by bit, and ordinary methods, +/// while not marked as unsafe to call, can invoke UB if called incorrectly. +/// Since any `NodeRef` allows navigating through the tree, `BorrowType` +/// effectively applies to the entire tree, not just to the node itself. +/// - `K` and `V`: These are the types of keys and values stored in the nodes. +/// - `Type`: This can be `Leaf`, `Internal`, or `LeafOrInternal`. When this is +/// `Leaf`, the `NodeRef` points to a leaf node, when this is `Internal` the +/// `NodeRef` points to an internal node, and when this is `LeafOrInternal` the +/// `NodeRef` could be pointing to either type of node. +/// `Type` is named `NodeType` when used outside `NodeRef`. +/// +/// Both `BorrowType` and `NodeType` restrict what methods we implement, to +/// exploit static type safety. There are limitations in the way we can apply +/// such restrictions: +/// - For each type parameter, we can only define a method either generically +/// or for one particular type. For example, we cannot define a method like +/// `into_kv` generically for all `BorrowType`, or once for all types that +/// carry a lifetime, because we want it to return `&'a` references. +/// Therefore, we define it only for the least powerful type `Immut<'a>`. +/// - We cannot get implicit coercion from say `Mut<'a>` to `Immut<'a>`. +/// Therefore, we have to explicitly call `reborrow` on a more powerful +/// `NodeRef` in order to reach a method like `into_kv`. +/// +/// All methods on `NodeRef` that return some kind of reference, either: +/// - Take `self` by value, and return the lifetime carried by `BorrowType`. +/// Sometimes, to invoke such a method, we need to call `reborrow_mut`. +/// - Take `self` by reference, and (implicitly) return that reference's +/// lifetime, instead of the lifetime carried by `BorrowType`. That way, +/// the borrow checker guarantees that the `NodeRef` remains borrowed as long +/// as the returned reference is used. +/// The methods supporting insert bend this rule by returning a raw pointer, +/// i.e., a reference without any lifetime. +pub struct NodeRef<BorrowType, K, V, Type> { + /// The number of levels that the node and the level of leaves are apart, a + /// constant of the node that cannot be entirely described by `Type`, and that + /// the node itself does not store. We only need to store the height of the root + /// node, and derive every other node's height from it. + /// Must be zero if `Type` is `Leaf` and non-zero if `Type` is `Internal`. + height: usize, + /// The pointer to the leaf or internal node. The definition of `InternalNode` + /// ensures that the pointer is valid either way. + node: NonNull<LeafNode<K, V>>, + _marker: PhantomData<(BorrowType, Type)>, +} + +/// The root node of an owned tree. +/// +/// Note that this does not have a destructor, and must be cleaned up manually. +pub type Root<K, V> = NodeRef<marker::Owned, K, V, marker::LeafOrInternal>; + +impl<'a, K: 'a, V: 'a, Type> Copy for NodeRef<marker::Immut<'a>, K, V, Type> {} +impl<'a, K: 'a, V: 'a, Type> Clone for NodeRef<marker::Immut<'a>, K, V, Type> { + fn clone(&self) -> Self { + *self + } +} + +unsafe impl<BorrowType, K: Sync, V: Sync, Type> Sync for NodeRef<BorrowType, K, V, Type> {} + +unsafe impl<'a, K: Sync + 'a, V: Sync + 'a, Type> Send for NodeRef<marker::Immut<'a>, K, V, Type> {} +unsafe impl<'a, K: Send + 'a, V: Send + 'a, Type> Send for NodeRef<marker::Mut<'a>, K, V, Type> {} +unsafe impl<'a, K: Send + 'a, V: Send + 'a, Type> Send for NodeRef<marker::ValMut<'a>, K, V, Type> {} +unsafe impl<K: Send, V: Send, Type> Send for NodeRef<marker::Owned, K, V, Type> {} +unsafe impl<K: Send, V: Send, Type> Send for NodeRef<marker::Dying, K, V, Type> {} + +impl<K, V> NodeRef<marker::Owned, K, V, marker::Leaf> { + pub fn new_leaf<A: Allocator + Clone>(alloc: A) -> Self { + Self::from_new_leaf(LeafNode::new(alloc)) + } + + fn from_new_leaf<A: Allocator + Clone>(leaf: Box<LeafNode<K, V>, A>) -> Self { + NodeRef { height: 0, node: NonNull::from(Box::leak(leaf)), _marker: PhantomData } + } +} + +impl<K, V> NodeRef<marker::Owned, K, V, marker::Internal> { + fn new_internal<A: Allocator + Clone>(child: Root<K, V>, alloc: A) -> Self { + let mut new_node = unsafe { InternalNode::new(alloc) }; + new_node.edges[0].write(child.node); + unsafe { NodeRef::from_new_internal(new_node, child.height + 1) } + } + + /// # Safety + /// `height` must not be zero. + unsafe fn from_new_internal<A: Allocator + Clone>( + internal: Box<InternalNode<K, V>, A>, + height: usize, + ) -> Self { + debug_assert!(height > 0); + let node = NonNull::from(Box::leak(internal)).cast(); + let mut this = NodeRef { height, node, _marker: PhantomData }; + this.borrow_mut().correct_all_childrens_parent_links(); + this + } +} + +impl<BorrowType, K, V> NodeRef<BorrowType, K, V, marker::Internal> { + /// Unpack a node reference that was packed as `NodeRef::parent`. + fn from_internal(node: NonNull<InternalNode<K, V>>, height: usize) -> Self { + debug_assert!(height > 0); + NodeRef { height, node: node.cast(), _marker: PhantomData } + } +} + +impl<BorrowType, K, V> NodeRef<BorrowType, K, V, marker::Internal> { + /// Exposes the data of an internal node. + /// + /// Returns a raw ptr to avoid invalidating other references to this node. + fn as_internal_ptr(this: &Self) -> *mut InternalNode<K, V> { + // SAFETY: the static node type is `Internal`. + this.node.as_ptr() as *mut InternalNode<K, V> + } +} + +impl<'a, K, V> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + /// Borrows exclusive access to the data of an internal node. + fn as_internal_mut(&mut self) -> &mut InternalNode<K, V> { + let ptr = Self::as_internal_ptr(self); + unsafe { &mut *ptr } + } +} + +impl<BorrowType, K, V, Type> NodeRef<BorrowType, K, V, Type> { + /// Finds the length of the node. This is the number of keys or values. + /// The number of edges is `len() + 1`. + /// Note that, despite being safe, calling this function can have the side effect + /// of invalidating mutable references that unsafe code has created. + pub fn len(&self) -> usize { + // Crucially, we only access the `len` field here. If BorrowType is marker::ValMut, + // there might be outstanding mutable references to values that we must not invalidate. + unsafe { usize::from((*Self::as_leaf_ptr(self)).len) } + } + + /// Returns the number of levels that the node and leaves are apart. Zero + /// height means the node is a leaf itself. If you picture trees with the + /// root on top, the number says at which elevation the node appears. + /// If you picture trees with leaves on top, the number says how high + /// the tree extends above the node. + pub fn height(&self) -> usize { + self.height + } + + /// Temporarily takes out another, immutable reference to the same node. + pub fn reborrow(&self) -> NodeRef<marker::Immut<'_>, K, V, Type> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Exposes the leaf portion of any leaf or internal node. + /// + /// Returns a raw ptr to avoid invalidating other references to this node. + fn as_leaf_ptr(this: &Self) -> *mut LeafNode<K, V> { + // The node must be valid for at least the LeafNode portion. + // This is not a reference in the NodeRef type because we don't know if + // it should be unique or shared. + this.node.as_ptr() + } +} + +impl<BorrowType: marker::BorrowType, K, V, Type> NodeRef<BorrowType, K, V, Type> { + /// Finds the parent of the current node. Returns `Ok(handle)` if the current + /// node actually has a parent, where `handle` points to the edge of the parent + /// that points to the current node. Returns `Err(self)` if the current node has + /// no parent, giving back the original `NodeRef`. + /// + /// The method name assumes you picture trees with the root node on top. + /// + /// `edge.descend().ascend().unwrap()` and `node.ascend().unwrap().descend()` should + /// both, upon success, do nothing. + pub fn ascend( + self, + ) -> Result<Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::Edge>, Self> { + assert!(BorrowType::PERMITS_TRAVERSAL); + // We need to use raw pointers to nodes because, if BorrowType is marker::ValMut, + // there might be outstanding mutable references to values that we must not invalidate. + let leaf_ptr: *const _ = Self::as_leaf_ptr(&self); + unsafe { (*leaf_ptr).parent } + .as_ref() + .map(|parent| Handle { + node: NodeRef::from_internal(*parent, self.height + 1), + idx: unsafe { usize::from((*leaf_ptr).parent_idx.assume_init()) }, + _marker: PhantomData, + }) + .ok_or(self) + } + + pub fn first_edge(self) -> Handle<Self, marker::Edge> { + unsafe { Handle::new_edge(self, 0) } + } + + pub fn last_edge(self) -> Handle<Self, marker::Edge> { + let len = self.len(); + unsafe { Handle::new_edge(self, len) } + } + + /// Note that `self` must be nonempty. + pub fn first_kv(self) -> Handle<Self, marker::KV> { + let len = self.len(); + assert!(len > 0); + unsafe { Handle::new_kv(self, 0) } + } + + /// Note that `self` must be nonempty. + pub fn last_kv(self) -> Handle<Self, marker::KV> { + let len = self.len(); + assert!(len > 0); + unsafe { Handle::new_kv(self, len - 1) } + } +} + +impl<BorrowType, K, V, Type> NodeRef<BorrowType, K, V, Type> { + /// Could be a public implementation of PartialEq, but only used in this module. + fn eq(&self, other: &Self) -> bool { + let Self { node, height, _marker } = self; + if node.eq(&other.node) { + debug_assert_eq!(*height, other.height); + true + } else { + false + } + } +} + +impl<'a, K: 'a, V: 'a, Type> NodeRef<marker::Immut<'a>, K, V, Type> { + /// Exposes the leaf portion of any leaf or internal node in an immutable tree. + fn into_leaf(self) -> &'a LeafNode<K, V> { + let ptr = Self::as_leaf_ptr(&self); + // SAFETY: there can be no mutable references into this tree borrowed as `Immut`. + unsafe { &*ptr } + } + + /// Borrows a view into the keys stored in the node. + pub fn keys(&self) -> &[K] { + let leaf = self.into_leaf(); + unsafe { + MaybeUninit::slice_assume_init_ref(leaf.keys.get_unchecked(..usize::from(leaf.len))) + } + } +} + +impl<K, V> NodeRef<marker::Dying, K, V, marker::LeafOrInternal> { + /// Similar to `ascend`, gets a reference to a node's parent node, but also + /// deallocates the current node in the process. This is unsafe because the + /// current node will still be accessible despite being deallocated. + pub unsafe fn deallocate_and_ascend<A: Allocator + Clone>( + self, + alloc: A, + ) -> Option<Handle<NodeRef<marker::Dying, K, V, marker::Internal>, marker::Edge>> { + let height = self.height; + let node = self.node; + let ret = self.ascend().ok(); + unsafe { + alloc.deallocate( + node.cast(), + if height > 0 { + Layout::new::<InternalNode<K, V>>() + } else { + Layout::new::<LeafNode<K, V>>() + }, + ); + } + ret + } +} + +impl<'a, K, V, Type> NodeRef<marker::Mut<'a>, K, V, Type> { + /// Temporarily takes out another mutable reference to the same node. Beware, as + /// this method is very dangerous, doubly so since it might not immediately appear + /// dangerous. + /// + /// Because mutable pointers can roam anywhere around the tree, the returned + /// pointer can easily be used to make the original pointer dangling, out of + /// bounds, or invalid under stacked borrow rules. + // FIXME(@gereeter) consider adding yet another type parameter to `NodeRef` + // that restricts the use of navigation methods on reborrowed pointers, + // preventing this unsafety. + unsafe fn reborrow_mut(&mut self) -> NodeRef<marker::Mut<'_>, K, V, Type> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Borrows exclusive access to the leaf portion of a leaf or internal node. + fn as_leaf_mut(&mut self) -> &mut LeafNode<K, V> { + let ptr = Self::as_leaf_ptr(self); + // SAFETY: we have exclusive access to the entire node. + unsafe { &mut *ptr } + } + + /// Offers exclusive access to the leaf portion of a leaf or internal node. + fn into_leaf_mut(mut self) -> &'a mut LeafNode<K, V> { + let ptr = Self::as_leaf_ptr(&mut self); + // SAFETY: we have exclusive access to the entire node. + unsafe { &mut *ptr } + } +} + +impl<K, V, Type> NodeRef<marker::Dying, K, V, Type> { + /// Borrows exclusive access to the leaf portion of a dying leaf or internal node. + fn as_leaf_dying(&mut self) -> &mut LeafNode<K, V> { + let ptr = Self::as_leaf_ptr(self); + // SAFETY: we have exclusive access to the entire node. + unsafe { &mut *ptr } + } +} + +impl<'a, K: 'a, V: 'a, Type> NodeRef<marker::Mut<'a>, K, V, Type> { + /// Borrows exclusive access to an element of the key storage area. + /// + /// # Safety + /// `index` is in bounds of 0..CAPACITY + unsafe fn key_area_mut<I, Output: ?Sized>(&mut self, index: I) -> &mut Output + where + I: SliceIndex<[MaybeUninit<K>], Output = Output>, + { + // SAFETY: the caller will not be able to call further methods on self + // until the key slice reference is dropped, as we have unique access + // for the lifetime of the borrow. + unsafe { self.as_leaf_mut().keys.as_mut_slice().get_unchecked_mut(index) } + } + + /// Borrows exclusive access to an element or slice of the node's value storage area. + /// + /// # Safety + /// `index` is in bounds of 0..CAPACITY + unsafe fn val_area_mut<I, Output: ?Sized>(&mut self, index: I) -> &mut Output + where + I: SliceIndex<[MaybeUninit<V>], Output = Output>, + { + // SAFETY: the caller will not be able to call further methods on self + // until the value slice reference is dropped, as we have unique access + // for the lifetime of the borrow. + unsafe { self.as_leaf_mut().vals.as_mut_slice().get_unchecked_mut(index) } + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + /// Borrows exclusive access to an element or slice of the node's storage area for edge contents. + /// + /// # Safety + /// `index` is in bounds of 0..CAPACITY + 1 + unsafe fn edge_area_mut<I, Output: ?Sized>(&mut self, index: I) -> &mut Output + where + I: SliceIndex<[MaybeUninit<BoxedNode<K, V>>], Output = Output>, + { + // SAFETY: the caller will not be able to call further methods on self + // until the edge slice reference is dropped, as we have unique access + // for the lifetime of the borrow. + unsafe { self.as_internal_mut().edges.as_mut_slice().get_unchecked_mut(index) } + } +} + +impl<'a, K, V, Type> NodeRef<marker::ValMut<'a>, K, V, Type> { + /// # Safety + /// - The node has more than `idx` initialized elements. + unsafe fn into_key_val_mut_at(mut self, idx: usize) -> (&'a K, &'a mut V) { + // We only create a reference to the one element we are interested in, + // to avoid aliasing with outstanding references to other elements, + // in particular, those returned to the caller in earlier iterations. + let leaf = Self::as_leaf_ptr(&mut self); + let keys = unsafe { ptr::addr_of!((*leaf).keys) }; + let vals = unsafe { ptr::addr_of_mut!((*leaf).vals) }; + // We must coerce to unsized array pointers because of Rust issue #74679. + let keys: *const [_] = keys; + let vals: *mut [_] = vals; + let key = unsafe { (&*keys.get_unchecked(idx)).assume_init_ref() }; + let val = unsafe { (&mut *vals.get_unchecked_mut(idx)).assume_init_mut() }; + (key, val) + } +} + +impl<'a, K: 'a, V: 'a, Type> NodeRef<marker::Mut<'a>, K, V, Type> { + /// Borrows exclusive access to the length of the node. + pub fn len_mut(&mut self) -> &mut u16 { + &mut self.as_leaf_mut().len + } +} + +impl<'a, K, V> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + /// # Safety + /// Every item returned by `range` is a valid edge index for the node. + unsafe fn correct_childrens_parent_links<R: Iterator<Item = usize>>(&mut self, range: R) { + for i in range { + debug_assert!(i <= self.len()); + unsafe { Handle::new_edge(self.reborrow_mut(), i) }.correct_parent_link(); + } + } + + fn correct_all_childrens_parent_links(&mut self) { + let len = self.len(); + unsafe { self.correct_childrens_parent_links(0..=len) }; + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + /// Sets the node's link to its parent edge, + /// without invalidating other references to the node. + fn set_parent_link(&mut self, parent: NonNull<InternalNode<K, V>>, parent_idx: usize) { + let leaf = Self::as_leaf_ptr(self); + unsafe { (*leaf).parent = Some(parent) }; + unsafe { (*leaf).parent_idx.write(parent_idx as u16) }; + } +} + +impl<K, V> NodeRef<marker::Owned, K, V, marker::LeafOrInternal> { + /// Clears the root's link to its parent edge. + fn clear_parent_link(&mut self) { + let mut root_node = self.borrow_mut(); + let leaf = root_node.as_leaf_mut(); + leaf.parent = None; + } +} + +impl<K, V> NodeRef<marker::Owned, K, V, marker::LeafOrInternal> { + /// Returns a new owned tree, with its own root node that is initially empty. + pub fn new<A: Allocator + Clone>(alloc: A) -> Self { + NodeRef::new_leaf(alloc).forget_type() + } + + /// Adds a new internal node with a single edge pointing to the previous root node, + /// make that new node the root node, and return it. This increases the height by 1 + /// and is the opposite of `pop_internal_level`. + pub fn push_internal_level<A: Allocator + Clone>( + &mut self, + alloc: A, + ) -> NodeRef<marker::Mut<'_>, K, V, marker::Internal> { + super::mem::take_mut(self, |old_root| NodeRef::new_internal(old_root, alloc).forget_type()); + + // `self.borrow_mut()`, except that we just forgot we're internal now: + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Removes the internal root node, using its first child as the new root node. + /// As it is intended only to be called when the root node has only one child, + /// no cleanup is done on any of the keys, values and other children. + /// This decreases the height by 1 and is the opposite of `push_internal_level`. + /// + /// Requires exclusive access to the `NodeRef` object but not to the root node; + /// it will not invalidate other handles or references to the root node. + /// + /// Panics if there is no internal level, i.e., if the root node is a leaf. + pub fn pop_internal_level<A: Allocator + Clone>(&mut self, alloc: A) { + assert!(self.height > 0); + + let top = self.node; + + // SAFETY: we asserted to be internal. + let internal_self = unsafe { self.borrow_mut().cast_to_internal_unchecked() }; + // SAFETY: we borrowed `self` exclusively and its borrow type is exclusive. + let internal_node = unsafe { &mut *NodeRef::as_internal_ptr(&internal_self) }; + // SAFETY: the first edge is always initialized. + self.node = unsafe { internal_node.edges[0].assume_init_read() }; + self.height -= 1; + self.clear_parent_link(); + + unsafe { + alloc.deallocate(top.cast(), Layout::new::<InternalNode<K, V>>()); + } + } +} + +impl<K, V, Type> NodeRef<marker::Owned, K, V, Type> { + /// Mutably borrows the owned root node. Unlike `reborrow_mut`, this is safe + /// because the return value cannot be used to destroy the root, and there + /// cannot be other references to the tree. + pub fn borrow_mut(&mut self) -> NodeRef<marker::Mut<'_>, K, V, Type> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Slightly mutably borrows the owned root node. + pub fn borrow_valmut(&mut self) -> NodeRef<marker::ValMut<'_>, K, V, Type> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Irreversibly transitions to a reference that permits traversal and offers + /// destructive methods and little else. + pub fn into_dying(self) -> NodeRef<marker::Dying, K, V, Type> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::Leaf> { + /// Adds a key-value pair to the end of the node, and returns + /// the mutable reference of the inserted value. + pub fn push(&mut self, key: K, val: V) -> &mut V { + let len = self.len_mut(); + let idx = usize::from(*len); + assert!(idx < CAPACITY); + *len += 1; + unsafe { + self.key_area_mut(idx).write(key); + self.val_area_mut(idx).write(val) + } + } +} + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + /// Adds a key-value pair, and an edge to go to the right of that pair, + /// to the end of the node. + pub fn push(&mut self, key: K, val: V, edge: Root<K, V>) { + assert!(edge.height == self.height - 1); + + let len = self.len_mut(); + let idx = usize::from(*len); + assert!(idx < CAPACITY); + *len += 1; + unsafe { + self.key_area_mut(idx).write(key); + self.val_area_mut(idx).write(val); + self.edge_area_mut(idx + 1).write(edge.node); + Handle::new_edge(self.reborrow_mut(), idx + 1).correct_parent_link(); + } + } +} + +impl<BorrowType, K, V> NodeRef<BorrowType, K, V, marker::Leaf> { + /// Removes any static information asserting that this node is a `Leaf` node. + pub fn forget_type(self) -> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } +} + +impl<BorrowType, K, V> NodeRef<BorrowType, K, V, marker::Internal> { + /// Removes any static information asserting that this node is an `Internal` node. + pub fn forget_type(self) -> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } +} + +impl<BorrowType, K, V> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + /// Checks whether a node is an `Internal` node or a `Leaf` node. + pub fn force( + self, + ) -> ForceResult< + NodeRef<BorrowType, K, V, marker::Leaf>, + NodeRef<BorrowType, K, V, marker::Internal>, + > { + if self.height == 0 { + ForceResult::Leaf(NodeRef { + height: self.height, + node: self.node, + _marker: PhantomData, + }) + } else { + ForceResult::Internal(NodeRef { + height: self.height, + node: self.node, + _marker: PhantomData, + }) + } + } +} + +impl<'a, K, V> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + /// Unsafely asserts to the compiler the static information that this node is a `Leaf`. + unsafe fn cast_to_leaf_unchecked(self) -> NodeRef<marker::Mut<'a>, K, V, marker::Leaf> { + debug_assert!(self.height == 0); + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } + + /// Unsafely asserts to the compiler the static information that this node is an `Internal`. + unsafe fn cast_to_internal_unchecked(self) -> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + debug_assert!(self.height > 0); + NodeRef { height: self.height, node: self.node, _marker: PhantomData } + } +} + +/// A reference to a specific key-value pair or edge within a node. The `Node` parameter +/// must be a `NodeRef`, while the `Type` can either be `KV` (signifying a handle on a key-value +/// pair) or `Edge` (signifying a handle on an edge). +/// +/// Note that even `Leaf` nodes can have `Edge` handles. Instead of representing a pointer to +/// a child node, these represent the spaces where child pointers would go between the key-value +/// pairs. For example, in a node with length 2, there would be 3 possible edge locations - one +/// to the left of the node, one between the two pairs, and one at the right of the node. +pub struct Handle<Node, Type> { + node: Node, + idx: usize, + _marker: PhantomData<Type>, +} + +impl<Node: Copy, Type> Copy for Handle<Node, Type> {} +// We don't need the full generality of `#[derive(Clone)]`, as the only time `Node` will be +// `Clone`able is when it is an immutable reference and therefore `Copy`. +impl<Node: Copy, Type> Clone for Handle<Node, Type> { + fn clone(&self) -> Self { + *self + } +} + +impl<Node, Type> Handle<Node, Type> { + /// Retrieves the node that contains the edge or key-value pair this handle points to. + pub fn into_node(self) -> Node { + self.node + } + + /// Returns the position of this handle in the node. + pub fn idx(&self) -> usize { + self.idx + } +} + +impl<BorrowType, K, V, NodeType> Handle<NodeRef<BorrowType, K, V, NodeType>, marker::KV> { + /// Creates a new handle to a key-value pair in `node`. + /// Unsafe because the caller must ensure that `idx < node.len()`. + pub unsafe fn new_kv(node: NodeRef<BorrowType, K, V, NodeType>, idx: usize) -> Self { + debug_assert!(idx < node.len()); + + Handle { node, idx, _marker: PhantomData } + } + + pub fn left_edge(self) -> Handle<NodeRef<BorrowType, K, V, NodeType>, marker::Edge> { + unsafe { Handle::new_edge(self.node, self.idx) } + } + + pub fn right_edge(self) -> Handle<NodeRef<BorrowType, K, V, NodeType>, marker::Edge> { + unsafe { Handle::new_edge(self.node, self.idx + 1) } + } +} + +impl<BorrowType, K, V, NodeType, HandleType> PartialEq + for Handle<NodeRef<BorrowType, K, V, NodeType>, HandleType> +{ + fn eq(&self, other: &Self) -> bool { + let Self { node, idx, _marker } = self; + node.eq(&other.node) && *idx == other.idx + } +} + +impl<BorrowType, K, V, NodeType, HandleType> + Handle<NodeRef<BorrowType, K, V, NodeType>, HandleType> +{ + /// Temporarily takes out another immutable handle on the same location. + pub fn reborrow(&self) -> Handle<NodeRef<marker::Immut<'_>, K, V, NodeType>, HandleType> { + // We can't use Handle::new_kv or Handle::new_edge because we don't know our type + Handle { node: self.node.reborrow(), idx: self.idx, _marker: PhantomData } + } +} + +impl<'a, K, V, NodeType, HandleType> Handle<NodeRef<marker::Mut<'a>, K, V, NodeType>, HandleType> { + /// Temporarily takes out another mutable handle on the same location. Beware, as + /// this method is very dangerous, doubly so since it might not immediately appear + /// dangerous. + /// + /// For details, see `NodeRef::reborrow_mut`. + pub unsafe fn reborrow_mut( + &mut self, + ) -> Handle<NodeRef<marker::Mut<'_>, K, V, NodeType>, HandleType> { + // We can't use Handle::new_kv or Handle::new_edge because we don't know our type + Handle { node: unsafe { self.node.reborrow_mut() }, idx: self.idx, _marker: PhantomData } + } +} + +impl<BorrowType, K, V, NodeType> Handle<NodeRef<BorrowType, K, V, NodeType>, marker::Edge> { + /// Creates a new handle to an edge in `node`. + /// Unsafe because the caller must ensure that `idx <= node.len()`. + pub unsafe fn new_edge(node: NodeRef<BorrowType, K, V, NodeType>, idx: usize) -> Self { + debug_assert!(idx <= node.len()); + + Handle { node, idx, _marker: PhantomData } + } + + pub fn left_kv(self) -> Result<Handle<NodeRef<BorrowType, K, V, NodeType>, marker::KV>, Self> { + if self.idx > 0 { + Ok(unsafe { Handle::new_kv(self.node, self.idx - 1) }) + } else { + Err(self) + } + } + + pub fn right_kv(self) -> Result<Handle<NodeRef<BorrowType, K, V, NodeType>, marker::KV>, Self> { + if self.idx < self.node.len() { + Ok(unsafe { Handle::new_kv(self.node, self.idx) }) + } else { + Err(self) + } + } +} + +pub enum LeftOrRight<T> { + Left(T), + Right(T), +} + +/// Given an edge index where we want to insert into a node filled to capacity, +/// computes a sensible KV index of a split point and where to perform the insertion. +/// The goal of the split point is for its key and value to end up in a parent node; +/// the keys, values and edges to the left of the split point become the left child; +/// the keys, values and edges to the right of the split point become the right child. +fn splitpoint(edge_idx: usize) -> (usize, LeftOrRight<usize>) { + debug_assert!(edge_idx <= CAPACITY); + // Rust issue #74834 tries to explain these symmetric rules. + match edge_idx { + 0..EDGE_IDX_LEFT_OF_CENTER => (KV_IDX_CENTER - 1, LeftOrRight::Left(edge_idx)), + EDGE_IDX_LEFT_OF_CENTER => (KV_IDX_CENTER, LeftOrRight::Left(edge_idx)), + EDGE_IDX_RIGHT_OF_CENTER => (KV_IDX_CENTER, LeftOrRight::Right(0)), + _ => (KV_IDX_CENTER + 1, LeftOrRight::Right(edge_idx - (KV_IDX_CENTER + 1 + 1))), + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge> { + /// Inserts a new key-value pair between the key-value pairs to the right and left of + /// this edge. This method assumes that there is enough space in the node for the new + /// pair to fit. + /// + /// The returned pointer points to the inserted value. + fn insert_fit(&mut self, key: K, val: V) -> *mut V { + debug_assert!(self.node.len() < CAPACITY); + let new_len = self.node.len() + 1; + + unsafe { + slice_insert(self.node.key_area_mut(..new_len), self.idx, key); + slice_insert(self.node.val_area_mut(..new_len), self.idx, val); + *self.node.len_mut() = new_len as u16; + + self.node.val_area_mut(self.idx).assume_init_mut() + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge> { + /// Inserts a new key-value pair between the key-value pairs to the right and left of + /// this edge. This method splits the node if there isn't enough room. + /// + /// The returned pointer points to the inserted value. + fn insert<A: Allocator + Clone>( + mut self, + key: K, + val: V, + alloc: A, + ) -> (Option<SplitResult<'a, K, V, marker::Leaf>>, *mut V) { + if self.node.len() < CAPACITY { + let val_ptr = self.insert_fit(key, val); + (None, val_ptr) + } else { + let (middle_kv_idx, insertion) = splitpoint(self.idx); + let middle = unsafe { Handle::new_kv(self.node, middle_kv_idx) }; + let mut result = middle.split(alloc); + let mut insertion_edge = match insertion { + LeftOrRight::Left(insert_idx) => unsafe { + Handle::new_edge(result.left.reborrow_mut(), insert_idx) + }, + LeftOrRight::Right(insert_idx) => unsafe { + Handle::new_edge(result.right.borrow_mut(), insert_idx) + }, + }; + let val_ptr = insertion_edge.insert_fit(key, val); + (Some(result), val_ptr) + } + } +} + +impl<'a, K, V> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::Edge> { + /// Fixes the parent pointer and index in the child node that this edge + /// links to. This is useful when the ordering of edges has been changed, + fn correct_parent_link(self) { + // Create backpointer without invalidating other references to the node. + let ptr = unsafe { NonNull::new_unchecked(NodeRef::as_internal_ptr(&self.node)) }; + let idx = self.idx; + let mut child = self.descend(); + child.set_parent_link(ptr, idx); + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::Edge> { + /// Inserts a new key-value pair and an edge that will go to the right of that new pair + /// between this edge and the key-value pair to the right of this edge. This method assumes + /// that there is enough space in the node for the new pair to fit. + fn insert_fit(&mut self, key: K, val: V, edge: Root<K, V>) { + debug_assert!(self.node.len() < CAPACITY); + debug_assert!(edge.height == self.node.height - 1); + let new_len = self.node.len() + 1; + + unsafe { + slice_insert(self.node.key_area_mut(..new_len), self.idx, key); + slice_insert(self.node.val_area_mut(..new_len), self.idx, val); + slice_insert(self.node.edge_area_mut(..new_len + 1), self.idx + 1, edge.node); + *self.node.len_mut() = new_len as u16; + + self.node.correct_childrens_parent_links(self.idx + 1..new_len + 1); + } + } + + /// Inserts a new key-value pair and an edge that will go to the right of that new pair + /// between this edge and the key-value pair to the right of this edge. This method splits + /// the node if there isn't enough room. + fn insert<A: Allocator + Clone>( + mut self, + key: K, + val: V, + edge: Root<K, V>, + alloc: A, + ) -> Option<SplitResult<'a, K, V, marker::Internal>> { + assert!(edge.height == self.node.height - 1); + + if self.node.len() < CAPACITY { + self.insert_fit(key, val, edge); + None + } else { + let (middle_kv_idx, insertion) = splitpoint(self.idx); + let middle = unsafe { Handle::new_kv(self.node, middle_kv_idx) }; + let mut result = middle.split(alloc); + let mut insertion_edge = match insertion { + LeftOrRight::Left(insert_idx) => unsafe { + Handle::new_edge(result.left.reborrow_mut(), insert_idx) + }, + LeftOrRight::Right(insert_idx) => unsafe { + Handle::new_edge(result.right.borrow_mut(), insert_idx) + }, + }; + insertion_edge.insert_fit(key, val, edge); + Some(result) + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge> { + /// Inserts a new key-value pair between the key-value pairs to the right and left of + /// this edge. This method splits the node if there isn't enough room, and tries to + /// insert the split off portion into the parent node recursively, until the root is reached. + /// + /// If the returned result is some `SplitResult`, the `left` field will be the root node. + /// The returned pointer points to the inserted value, which in the case of `SplitResult` + /// is in the `left` or `right` tree. + pub fn insert_recursing<A: Allocator + Clone>( + self, + key: K, + value: V, + alloc: A, + ) -> (Option<SplitResult<'a, K, V, marker::LeafOrInternal>>, *mut V) { + let (mut split, val_ptr) = match self.insert(key, value, alloc.clone()) { + (None, val_ptr) => return (None, val_ptr), + (Some(split), val_ptr) => (split.forget_node_type(), val_ptr), + }; + + loop { + split = match split.left.ascend() { + Ok(parent) => { + match parent.insert(split.kv.0, split.kv.1, split.right, alloc.clone()) { + None => return (None, val_ptr), + Some(split) => split.forget_node_type(), + } + } + Err(root) => return (Some(SplitResult { left: root, ..split }), val_ptr), + }; + } + } +} + +impl<BorrowType: marker::BorrowType, K, V> + Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::Edge> +{ + /// Finds the node pointed to by this edge. + /// + /// The method name assumes you picture trees with the root node on top. + /// + /// `edge.descend().ascend().unwrap()` and `node.ascend().unwrap().descend()` should + /// both, upon success, do nothing. + pub fn descend(self) -> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + assert!(BorrowType::PERMITS_TRAVERSAL); + // We need to use raw pointers to nodes because, if BorrowType is + // marker::ValMut, there might be outstanding mutable references to + // values that we must not invalidate. There's no worry accessing the + // height field because that value is copied. Beware that, once the + // node pointer is dereferenced, we access the edges array with a + // reference (Rust issue #73987) and invalidate any other references + // to or inside the array, should any be around. + let parent_ptr = NodeRef::as_internal_ptr(&self.node); + let node = unsafe { (*parent_ptr).edges.get_unchecked(self.idx).assume_init_read() }; + NodeRef { node, height: self.node.height - 1, _marker: PhantomData } + } +} + +impl<'a, K: 'a, V: 'a, NodeType> Handle<NodeRef<marker::Immut<'a>, K, V, NodeType>, marker::KV> { + pub fn into_kv(self) -> (&'a K, &'a V) { + debug_assert!(self.idx < self.node.len()); + let leaf = self.node.into_leaf(); + let k = unsafe { leaf.keys.get_unchecked(self.idx).assume_init_ref() }; + let v = unsafe { leaf.vals.get_unchecked(self.idx).assume_init_ref() }; + (k, v) + } +} + +impl<'a, K: 'a, V: 'a, NodeType> Handle<NodeRef<marker::Mut<'a>, K, V, NodeType>, marker::KV> { + pub fn key_mut(&mut self) -> &mut K { + unsafe { self.node.key_area_mut(self.idx).assume_init_mut() } + } + + pub fn into_val_mut(self) -> &'a mut V { + debug_assert!(self.idx < self.node.len()); + let leaf = self.node.into_leaf_mut(); + unsafe { leaf.vals.get_unchecked_mut(self.idx).assume_init_mut() } + } +} + +impl<'a, K, V, NodeType> Handle<NodeRef<marker::ValMut<'a>, K, V, NodeType>, marker::KV> { + pub fn into_kv_valmut(self) -> (&'a K, &'a mut V) { + unsafe { self.node.into_key_val_mut_at(self.idx) } + } +} + +impl<'a, K: 'a, V: 'a, NodeType> Handle<NodeRef<marker::Mut<'a>, K, V, NodeType>, marker::KV> { + pub fn kv_mut(&mut self) -> (&mut K, &mut V) { + debug_assert!(self.idx < self.node.len()); + // We cannot call separate key and value methods, because calling the second one + // invalidates the reference returned by the first. + unsafe { + let leaf = self.node.as_leaf_mut(); + let key = leaf.keys.get_unchecked_mut(self.idx).assume_init_mut(); + let val = leaf.vals.get_unchecked_mut(self.idx).assume_init_mut(); + (key, val) + } + } + + /// Replaces the key and value that the KV handle refers to. + pub fn replace_kv(&mut self, k: K, v: V) -> (K, V) { + let (key, val) = self.kv_mut(); + (mem::replace(key, k), mem::replace(val, v)) + } +} + +impl<K, V, NodeType> Handle<NodeRef<marker::Dying, K, V, NodeType>, marker::KV> { + /// Extracts the key and value that the KV handle refers to. + /// # Safety + /// The node that the handle refers to must not yet have been deallocated. + pub unsafe fn into_key_val(mut self) -> (K, V) { + debug_assert!(self.idx < self.node.len()); + let leaf = self.node.as_leaf_dying(); + unsafe { + let key = leaf.keys.get_unchecked_mut(self.idx).assume_init_read(); + let val = leaf.vals.get_unchecked_mut(self.idx).assume_init_read(); + (key, val) + } + } + + /// Drops the key and value that the KV handle refers to. + /// # Safety + /// The node that the handle refers to must not yet have been deallocated. + #[inline] + pub unsafe fn drop_key_val(mut self) { + debug_assert!(self.idx < self.node.len()); + let leaf = self.node.as_leaf_dying(); + unsafe { + leaf.keys.get_unchecked_mut(self.idx).assume_init_drop(); + leaf.vals.get_unchecked_mut(self.idx).assume_init_drop(); + } + } +} + +impl<'a, K: 'a, V: 'a, NodeType> Handle<NodeRef<marker::Mut<'a>, K, V, NodeType>, marker::KV> { + /// Helps implementations of `split` for a particular `NodeType`, + /// by taking care of leaf data. + fn split_leaf_data(&mut self, new_node: &mut LeafNode<K, V>) -> (K, V) { + debug_assert!(self.idx < self.node.len()); + let old_len = self.node.len(); + let new_len = old_len - self.idx - 1; + new_node.len = new_len as u16; + unsafe { + let k = self.node.key_area_mut(self.idx).assume_init_read(); + let v = self.node.val_area_mut(self.idx).assume_init_read(); + + move_to_slice( + self.node.key_area_mut(self.idx + 1..old_len), + &mut new_node.keys[..new_len], + ); + move_to_slice( + self.node.val_area_mut(self.idx + 1..old_len), + &mut new_node.vals[..new_len], + ); + + *self.node.len_mut() = self.idx as u16; + (k, v) + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::KV> { + /// Splits the underlying node into three parts: + /// + /// - The node is truncated to only contain the key-value pairs to the left of + /// this handle. + /// - The key and value pointed to by this handle are extracted. + /// - All the key-value pairs to the right of this handle are put into a newly + /// allocated node. + pub fn split<A: Allocator + Clone>(mut self, alloc: A) -> SplitResult<'a, K, V, marker::Leaf> { + let mut new_node = LeafNode::new(alloc); + + let kv = self.split_leaf_data(&mut new_node); + + let right = NodeRef::from_new_leaf(new_node); + SplitResult { left: self.node, kv, right } + } + + /// Removes the key-value pair pointed to by this handle and returns it, along with the edge + /// that the key-value pair collapsed into. + pub fn remove( + mut self, + ) -> ((K, V), Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>) { + let old_len = self.node.len(); + unsafe { + let k = slice_remove(self.node.key_area_mut(..old_len), self.idx); + let v = slice_remove(self.node.val_area_mut(..old_len), self.idx); + *self.node.len_mut() = (old_len - 1) as u16; + ((k, v), self.left_edge()) + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::KV> { + /// Splits the underlying node into three parts: + /// + /// - The node is truncated to only contain the edges and key-value pairs to the + /// left of this handle. + /// - The key and value pointed to by this handle are extracted. + /// - All the edges and key-value pairs to the right of this handle are put into + /// a newly allocated node. + pub fn split<A: Allocator + Clone>( + mut self, + alloc: A, + ) -> SplitResult<'a, K, V, marker::Internal> { + let old_len = self.node.len(); + unsafe { + let mut new_node = InternalNode::new(alloc); + let kv = self.split_leaf_data(&mut new_node.data); + let new_len = usize::from(new_node.data.len); + move_to_slice( + self.node.edge_area_mut(self.idx + 1..old_len + 1), + &mut new_node.edges[..new_len + 1], + ); + + let height = self.node.height; + let right = NodeRef::from_new_internal(new_node, height); + + SplitResult { left: self.node, kv, right } + } + } +} + +/// Represents a session for evaluating and performing a balancing operation +/// around an internal key-value pair. +pub struct BalancingContext<'a, K, V> { + parent: Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::KV>, + left_child: NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, + right_child: NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, +} + +impl<'a, K, V> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::KV> { + pub fn consider_for_balancing(self) -> BalancingContext<'a, K, V> { + let self1 = unsafe { ptr::read(&self) }; + let self2 = unsafe { ptr::read(&self) }; + BalancingContext { + parent: self, + left_child: self1.left_edge().descend(), + right_child: self2.right_edge().descend(), + } + } +} + +impl<'a, K, V> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + /// Chooses a balancing context involving the node as a child, thus between + /// the KV immediately to the left or to the right in the parent node. + /// Returns an `Err` if there is no parent. + /// Panics if the parent is empty. + /// + /// Prefers the left side, to be optimal if the given node is somehow + /// underfull, meaning here only that it has fewer elements than its left + /// sibling and than its right sibling, if they exist. In that case, + /// merging with the left sibling is faster, since we only need to move + /// the node's N elements, instead of shifting them to the right and moving + /// more than N elements in front. Stealing from the left sibling is also + /// typically faster, since we only need to shift the node's N elements to + /// the right, instead of shifting at least N of the sibling's elements to + /// the left. + pub fn choose_parent_kv(self) -> Result<LeftOrRight<BalancingContext<'a, K, V>>, Self> { + match unsafe { ptr::read(&self) }.ascend() { + Ok(parent_edge) => match parent_edge.left_kv() { + Ok(left_parent_kv) => Ok(LeftOrRight::Left(BalancingContext { + parent: unsafe { ptr::read(&left_parent_kv) }, + left_child: left_parent_kv.left_edge().descend(), + right_child: self, + })), + Err(parent_edge) => match parent_edge.right_kv() { + Ok(right_parent_kv) => Ok(LeftOrRight::Right(BalancingContext { + parent: unsafe { ptr::read(&right_parent_kv) }, + left_child: self, + right_child: right_parent_kv.right_edge().descend(), + })), + Err(_) => unreachable!("empty internal node"), + }, + }, + Err(root) => Err(root), + } + } +} + +impl<'a, K, V> BalancingContext<'a, K, V> { + pub fn left_child_len(&self) -> usize { + self.left_child.len() + } + + pub fn right_child_len(&self) -> usize { + self.right_child.len() + } + + pub fn into_left_child(self) -> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + self.left_child + } + + pub fn into_right_child(self) -> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + self.right_child + } + + /// Returns whether merging is possible, i.e., whether there is enough room + /// in a node to combine the central KV with both adjacent child nodes. + pub fn can_merge(&self) -> bool { + self.left_child.len() + 1 + self.right_child.len() <= CAPACITY + } +} + +impl<'a, K: 'a, V: 'a> BalancingContext<'a, K, V> { + /// Performs a merge and lets a closure decide what to return. + fn do_merge< + F: FnOnce( + NodeRef<marker::Mut<'a>, K, V, marker::Internal>, + NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, + ) -> R, + R, + A: Allocator, + >( + self, + result: F, + alloc: A, + ) -> R { + let Handle { node: mut parent_node, idx: parent_idx, _marker } = self.parent; + let old_parent_len = parent_node.len(); + let mut left_node = self.left_child; + let old_left_len = left_node.len(); + let mut right_node = self.right_child; + let right_len = right_node.len(); + let new_left_len = old_left_len + 1 + right_len; + + assert!(new_left_len <= CAPACITY); + + unsafe { + *left_node.len_mut() = new_left_len as u16; + + let parent_key = slice_remove(parent_node.key_area_mut(..old_parent_len), parent_idx); + left_node.key_area_mut(old_left_len).write(parent_key); + move_to_slice( + right_node.key_area_mut(..right_len), + left_node.key_area_mut(old_left_len + 1..new_left_len), + ); + + let parent_val = slice_remove(parent_node.val_area_mut(..old_parent_len), parent_idx); + left_node.val_area_mut(old_left_len).write(parent_val); + move_to_slice( + right_node.val_area_mut(..right_len), + left_node.val_area_mut(old_left_len + 1..new_left_len), + ); + + slice_remove(&mut parent_node.edge_area_mut(..old_parent_len + 1), parent_idx + 1); + parent_node.correct_childrens_parent_links(parent_idx + 1..old_parent_len); + *parent_node.len_mut() -= 1; + + if parent_node.height > 1 { + // SAFETY: the height of the nodes being merged is one below the height + // of the node of this edge, thus above zero, so they are internal. + let mut left_node = left_node.reborrow_mut().cast_to_internal_unchecked(); + let mut right_node = right_node.cast_to_internal_unchecked(); + move_to_slice( + right_node.edge_area_mut(..right_len + 1), + left_node.edge_area_mut(old_left_len + 1..new_left_len + 1), + ); + + left_node.correct_childrens_parent_links(old_left_len + 1..new_left_len + 1); + + alloc.deallocate(right_node.node.cast(), Layout::new::<InternalNode<K, V>>()); + } else { + alloc.deallocate(right_node.node.cast(), Layout::new::<LeafNode<K, V>>()); + } + } + result(parent_node, left_node) + } + + /// Merges the parent's key-value pair and both adjacent child nodes into + /// the left child node and returns the shrunk parent node. + /// + /// Panics unless we `.can_merge()`. + pub fn merge_tracking_parent<A: Allocator + Clone>( + self, + alloc: A, + ) -> NodeRef<marker::Mut<'a>, K, V, marker::Internal> { + self.do_merge(|parent, _child| parent, alloc) + } + + /// Merges the parent's key-value pair and both adjacent child nodes into + /// the left child node and returns that child node. + /// + /// Panics unless we `.can_merge()`. + pub fn merge_tracking_child<A: Allocator + Clone>( + self, + alloc: A, + ) -> NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal> { + self.do_merge(|_parent, child| child, alloc) + } + + /// Merges the parent's key-value pair and both adjacent child nodes into + /// the left child node and returns the edge handle in that child node + /// where the tracked child edge ended up, + /// + /// Panics unless we `.can_merge()`. + pub fn merge_tracking_child_edge<A: Allocator + Clone>( + self, + track_edge_idx: LeftOrRight<usize>, + alloc: A, + ) -> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::Edge> { + let old_left_len = self.left_child.len(); + let right_len = self.right_child.len(); + assert!(match track_edge_idx { + LeftOrRight::Left(idx) => idx <= old_left_len, + LeftOrRight::Right(idx) => idx <= right_len, + }); + let child = self.merge_tracking_child(alloc); + let new_idx = match track_edge_idx { + LeftOrRight::Left(idx) => idx, + LeftOrRight::Right(idx) => old_left_len + 1 + idx, + }; + unsafe { Handle::new_edge(child, new_idx) } + } + + /// Removes a key-value pair from the left child and places it in the key-value storage + /// of the parent, while pushing the old parent key-value pair into the right child. + /// Returns a handle to the edge in the right child corresponding to where the original + /// edge specified by `track_right_edge_idx` ended up. + pub fn steal_left( + mut self, + track_right_edge_idx: usize, + ) -> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::Edge> { + self.bulk_steal_left(1); + unsafe { Handle::new_edge(self.right_child, 1 + track_right_edge_idx) } + } + + /// Removes a key-value pair from the right child and places it in the key-value storage + /// of the parent, while pushing the old parent key-value pair onto the left child. + /// Returns a handle to the edge in the left child specified by `track_left_edge_idx`, + /// which didn't move. + pub fn steal_right( + mut self, + track_left_edge_idx: usize, + ) -> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::Edge> { + self.bulk_steal_right(1); + unsafe { Handle::new_edge(self.left_child, track_left_edge_idx) } + } + + /// This does stealing similar to `steal_left` but steals multiple elements at once. + pub fn bulk_steal_left(&mut self, count: usize) { + assert!(count > 0); + unsafe { + let left_node = &mut self.left_child; + let old_left_len = left_node.len(); + let right_node = &mut self.right_child; + let old_right_len = right_node.len(); + + // Make sure that we may steal safely. + assert!(old_right_len + count <= CAPACITY); + assert!(old_left_len >= count); + + let new_left_len = old_left_len - count; + let new_right_len = old_right_len + count; + *left_node.len_mut() = new_left_len as u16; + *right_node.len_mut() = new_right_len as u16; + + // Move leaf data. + { + // Make room for stolen elements in the right child. + slice_shr(right_node.key_area_mut(..new_right_len), count); + slice_shr(right_node.val_area_mut(..new_right_len), count); + + // Move elements from the left child to the right one. + move_to_slice( + left_node.key_area_mut(new_left_len + 1..old_left_len), + right_node.key_area_mut(..count - 1), + ); + move_to_slice( + left_node.val_area_mut(new_left_len + 1..old_left_len), + right_node.val_area_mut(..count - 1), + ); + + // Move the left-most stolen pair to the parent. + let k = left_node.key_area_mut(new_left_len).assume_init_read(); + let v = left_node.val_area_mut(new_left_len).assume_init_read(); + let (k, v) = self.parent.replace_kv(k, v); + + // Move parent's key-value pair to the right child. + right_node.key_area_mut(count - 1).write(k); + right_node.val_area_mut(count - 1).write(v); + } + + match (left_node.reborrow_mut().force(), right_node.reborrow_mut().force()) { + (ForceResult::Internal(mut left), ForceResult::Internal(mut right)) => { + // Make room for stolen edges. + slice_shr(right.edge_area_mut(..new_right_len + 1), count); + + // Steal edges. + move_to_slice( + left.edge_area_mut(new_left_len + 1..old_left_len + 1), + right.edge_area_mut(..count), + ); + + right.correct_childrens_parent_links(0..new_right_len + 1); + } + (ForceResult::Leaf(_), ForceResult::Leaf(_)) => {} + _ => unreachable!(), + } + } + } + + /// The symmetric clone of `bulk_steal_left`. + pub fn bulk_steal_right(&mut self, count: usize) { + assert!(count > 0); + unsafe { + let left_node = &mut self.left_child; + let old_left_len = left_node.len(); + let right_node = &mut self.right_child; + let old_right_len = right_node.len(); + + // Make sure that we may steal safely. + assert!(old_left_len + count <= CAPACITY); + assert!(old_right_len >= count); + + let new_left_len = old_left_len + count; + let new_right_len = old_right_len - count; + *left_node.len_mut() = new_left_len as u16; + *right_node.len_mut() = new_right_len as u16; + + // Move leaf data. + { + // Move the right-most stolen pair to the parent. + let k = right_node.key_area_mut(count - 1).assume_init_read(); + let v = right_node.val_area_mut(count - 1).assume_init_read(); + let (k, v) = self.parent.replace_kv(k, v); + + // Move parent's key-value pair to the left child. + left_node.key_area_mut(old_left_len).write(k); + left_node.val_area_mut(old_left_len).write(v); + + // Move elements from the right child to the left one. + move_to_slice( + right_node.key_area_mut(..count - 1), + left_node.key_area_mut(old_left_len + 1..new_left_len), + ); + move_to_slice( + right_node.val_area_mut(..count - 1), + left_node.val_area_mut(old_left_len + 1..new_left_len), + ); + + // Fill gap where stolen elements used to be. + slice_shl(right_node.key_area_mut(..old_right_len), count); + slice_shl(right_node.val_area_mut(..old_right_len), count); + } + + match (left_node.reborrow_mut().force(), right_node.reborrow_mut().force()) { + (ForceResult::Internal(mut left), ForceResult::Internal(mut right)) => { + // Steal edges. + move_to_slice( + right.edge_area_mut(..count), + left.edge_area_mut(old_left_len + 1..new_left_len + 1), + ); + + // Fill gap where stolen edges used to be. + slice_shl(right.edge_area_mut(..old_right_len + 1), count); + + left.correct_childrens_parent_links(old_left_len + 1..new_left_len + 1); + right.correct_childrens_parent_links(0..new_right_len + 1); + } + (ForceResult::Leaf(_), ForceResult::Leaf(_)) => {} + _ => unreachable!(), + } + } + } +} + +impl<BorrowType, K, V> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge> { + pub fn forget_node_type( + self, + ) -> Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::Edge> { + unsafe { Handle::new_edge(self.node.forget_type(), self.idx) } + } +} + +impl<BorrowType, K, V> Handle<NodeRef<BorrowType, K, V, marker::Internal>, marker::Edge> { + pub fn forget_node_type( + self, + ) -> Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::Edge> { + unsafe { Handle::new_edge(self.node.forget_type(), self.idx) } + } +} + +impl<BorrowType, K, V> Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::KV> { + pub fn forget_node_type( + self, + ) -> Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, marker::KV> { + unsafe { Handle::new_kv(self.node.forget_type(), self.idx) } + } +} + +impl<BorrowType, K, V, Type> Handle<NodeRef<BorrowType, K, V, marker::LeafOrInternal>, Type> { + /// Checks whether the underlying node is an `Internal` node or a `Leaf` node. + pub fn force( + self, + ) -> ForceResult< + Handle<NodeRef<BorrowType, K, V, marker::Leaf>, Type>, + Handle<NodeRef<BorrowType, K, V, marker::Internal>, Type>, + > { + match self.node.force() { + ForceResult::Leaf(node) => { + ForceResult::Leaf(Handle { node, idx: self.idx, _marker: PhantomData }) + } + ForceResult::Internal(node) => { + ForceResult::Internal(Handle { node, idx: self.idx, _marker: PhantomData }) + } + } + } +} + +impl<'a, K, V, Type> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, Type> { + /// Unsafely asserts to the compiler the static information that the handle's node is a `Leaf`. + pub unsafe fn cast_to_leaf_unchecked( + self, + ) -> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, Type> { + let node = unsafe { self.node.cast_to_leaf_unchecked() }; + Handle { node, idx: self.idx, _marker: PhantomData } + } +} + +impl<'a, K, V> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::Edge> { + /// Move the suffix after `self` from one node to another one. `right` must be empty. + /// The first edge of `right` remains unchanged. + pub fn move_suffix( + &mut self, + right: &mut NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, + ) { + unsafe { + let new_left_len = self.idx; + let mut left_node = self.reborrow_mut().into_node(); + let old_left_len = left_node.len(); + + let new_right_len = old_left_len - new_left_len; + let mut right_node = right.reborrow_mut(); + + assert!(right_node.len() == 0); + assert!(left_node.height == right_node.height); + + if new_right_len > 0 { + *left_node.len_mut() = new_left_len as u16; + *right_node.len_mut() = new_right_len as u16; + + move_to_slice( + left_node.key_area_mut(new_left_len..old_left_len), + right_node.key_area_mut(..new_right_len), + ); + move_to_slice( + left_node.val_area_mut(new_left_len..old_left_len), + right_node.val_area_mut(..new_right_len), + ); + match (left_node.force(), right_node.force()) { + (ForceResult::Internal(mut left), ForceResult::Internal(mut right)) => { + move_to_slice( + left.edge_area_mut(new_left_len + 1..old_left_len + 1), + right.edge_area_mut(1..new_right_len + 1), + ); + right.correct_childrens_parent_links(1..new_right_len + 1); + } + (ForceResult::Leaf(_), ForceResult::Leaf(_)) => {} + _ => unreachable!(), + } + } + } + } +} + +pub enum ForceResult<Leaf, Internal> { + Leaf(Leaf), + Internal(Internal), +} + +/// Result of insertion, when a node needed to expand beyond its capacity. +pub struct SplitResult<'a, K, V, NodeType> { + // Altered node in existing tree with elements and edges that belong to the left of `kv`. + pub left: NodeRef<marker::Mut<'a>, K, V, NodeType>, + // Some key and value that existed before and were split off, to be inserted elsewhere. + pub kv: (K, V), + // Owned, unattached, new node with elements and edges that belong to the right of `kv`. + pub right: NodeRef<marker::Owned, K, V, NodeType>, +} + +impl<'a, K, V> SplitResult<'a, K, V, marker::Leaf> { + pub fn forget_node_type(self) -> SplitResult<'a, K, V, marker::LeafOrInternal> { + SplitResult { left: self.left.forget_type(), kv: self.kv, right: self.right.forget_type() } + } +} + +impl<'a, K, V> SplitResult<'a, K, V, marker::Internal> { + pub fn forget_node_type(self) -> SplitResult<'a, K, V, marker::LeafOrInternal> { + SplitResult { left: self.left.forget_type(), kv: self.kv, right: self.right.forget_type() } + } +} + +pub mod marker { + use core::marker::PhantomData; + + pub enum Leaf {} + pub enum Internal {} + pub enum LeafOrInternal {} + + pub enum Owned {} + pub enum Dying {} + pub struct Immut<'a>(PhantomData<&'a ()>); + pub struct Mut<'a>(PhantomData<&'a mut ()>); + pub struct ValMut<'a>(PhantomData<&'a mut ()>); + + pub trait BorrowType { + // Whether node references of this borrow type allow traversing + // to other nodes in the tree. + const PERMITS_TRAVERSAL: bool = true; + } + impl BorrowType for Owned { + // Traversal isn't needed, it happens using the result of `borrow_mut`. + // By disabling traversal, and only creating new references to roots, + // we know that every reference of the `Owned` type is to a root node. + const PERMITS_TRAVERSAL: bool = false; + } + impl BorrowType for Dying {} + impl<'a> BorrowType for Immut<'a> {} + impl<'a> BorrowType for Mut<'a> {} + impl<'a> BorrowType for ValMut<'a> {} + + pub enum KV {} + pub enum Edge {} +} + +/// Inserts a value into a slice of initialized elements followed by one uninitialized element. +/// +/// # Safety +/// The slice has more than `idx` elements. +unsafe fn slice_insert<T>(slice: &mut [MaybeUninit<T>], idx: usize, val: T) { + unsafe { + let len = slice.len(); + debug_assert!(len > idx); + let slice_ptr = slice.as_mut_ptr(); + if len > idx + 1 { + ptr::copy(slice_ptr.add(idx), slice_ptr.add(idx + 1), len - idx - 1); + } + (*slice_ptr.add(idx)).write(val); + } +} + +/// Removes and returns a value from a slice of all initialized elements, leaving behind one +/// trailing uninitialized element. +/// +/// # Safety +/// The slice has more than `idx` elements. +unsafe fn slice_remove<T>(slice: &mut [MaybeUninit<T>], idx: usize) -> T { + unsafe { + let len = slice.len(); + debug_assert!(idx < len); + let slice_ptr = slice.as_mut_ptr(); + let ret = (*slice_ptr.add(idx)).assume_init_read(); + ptr::copy(slice_ptr.add(idx + 1), slice_ptr.add(idx), len - idx - 1); + ret + } +} + +/// Shifts the elements in a slice `distance` positions to the left. +/// +/// # Safety +/// The slice has at least `distance` elements. +unsafe fn slice_shl<T>(slice: &mut [MaybeUninit<T>], distance: usize) { + unsafe { + let slice_ptr = slice.as_mut_ptr(); + ptr::copy(slice_ptr.add(distance), slice_ptr, slice.len() - distance); + } +} + +/// Shifts the elements in a slice `distance` positions to the right. +/// +/// # Safety +/// The slice has at least `distance` elements. +unsafe fn slice_shr<T>(slice: &mut [MaybeUninit<T>], distance: usize) { + unsafe { + let slice_ptr = slice.as_mut_ptr(); + ptr::copy(slice_ptr, slice_ptr.add(distance), slice.len() - distance); + } +} + +/// Moves all values from a slice of initialized elements to a slice +/// of uninitialized elements, leaving behind `src` as all uninitialized. +/// Works like `dst.copy_from_slice(src)` but does not require `T` to be `Copy`. +fn move_to_slice<T>(src: &mut [MaybeUninit<T>], dst: &mut [MaybeUninit<T>]) { + assert!(src.len() == dst.len()); + unsafe { + ptr::copy_nonoverlapping(src.as_ptr(), dst.as_mut_ptr(), src.len()); + } +} + +#[cfg(test)] +mod tests; diff --git a/library/alloc/src/collections/btree/node/tests.rs b/library/alloc/src/collections/btree/node/tests.rs new file mode 100644 index 000000000..aadb0dc9c --- /dev/null +++ b/library/alloc/src/collections/btree/node/tests.rs @@ -0,0 +1,102 @@ +use super::super::navigate; +use super::*; +use crate::alloc::Global; +use crate::fmt::Debug; +use crate::string::String; + +impl<'a, K: 'a, V: 'a> NodeRef<marker::Immut<'a>, K, V, marker::LeafOrInternal> { + // Asserts that the back pointer in each reachable node points to its parent. + pub fn assert_back_pointers(self) { + if let ForceResult::Internal(node) = self.force() { + for idx in 0..=node.len() { + let edge = unsafe { Handle::new_edge(node, idx) }; + let child = edge.descend(); + assert!(child.ascend().ok() == Some(edge)); + child.assert_back_pointers(); + } + } + } + + // Renders a multi-line display of the keys in order and in tree hierarchy, + // picturing the tree growing sideways from its root on the left to its + // leaves on the right. + pub fn dump_keys(self) -> String + where + K: Debug, + { + let mut result = String::new(); + self.visit_nodes_in_order(|pos| match pos { + navigate::Position::Leaf(leaf) => { + let depth = self.height(); + let indent = " ".repeat(depth); + result += &format!("\n{}{:?}", indent, leaf.keys()); + } + navigate::Position::Internal(_) => {} + navigate::Position::InternalKV(kv) => { + let depth = self.height() - kv.into_node().height(); + let indent = " ".repeat(depth); + result += &format!("\n{}{:?}", indent, kv.into_kv().0); + } + }); + result + } +} + +#[test] +fn test_splitpoint() { + for idx in 0..=CAPACITY { + let (middle_kv_idx, insertion) = splitpoint(idx); + + // Simulate performing the split: + let mut left_len = middle_kv_idx; + let mut right_len = CAPACITY - middle_kv_idx - 1; + match insertion { + LeftOrRight::Left(edge_idx) => { + assert!(edge_idx <= left_len); + left_len += 1; + } + LeftOrRight::Right(edge_idx) => { + assert!(edge_idx <= right_len); + right_len += 1; + } + } + assert!(left_len >= MIN_LEN_AFTER_SPLIT); + assert!(right_len >= MIN_LEN_AFTER_SPLIT); + assert!(left_len + right_len == CAPACITY); + } +} + +#[test] +fn test_partial_eq() { + let mut root1 = NodeRef::new_leaf(Global); + root1.borrow_mut().push(1, ()); + let mut root1 = NodeRef::new_internal(root1.forget_type(), Global).forget_type(); + let root2 = Root::new(Global); + root1.reborrow().assert_back_pointers(); + root2.reborrow().assert_back_pointers(); + + let leaf_edge_1a = root1.reborrow().first_leaf_edge().forget_node_type(); + let leaf_edge_1b = root1.reborrow().last_leaf_edge().forget_node_type(); + let top_edge_1 = root1.reborrow().first_edge(); + let top_edge_2 = root2.reborrow().first_edge(); + + assert!(leaf_edge_1a == leaf_edge_1a); + assert!(leaf_edge_1a != leaf_edge_1b); + assert!(leaf_edge_1a != top_edge_1); + assert!(leaf_edge_1a != top_edge_2); + assert!(top_edge_1 == top_edge_1); + assert!(top_edge_1 != top_edge_2); + + root1.pop_internal_level(Global); + unsafe { root1.into_dying().deallocate_and_ascend(Global) }; + unsafe { root2.into_dying().deallocate_and_ascend(Global) }; +} + +#[test] +#[cfg(target_arch = "x86_64")] +fn test_sizes() { + assert_eq!(core::mem::size_of::<LeafNode<(), ()>>(), 16); + assert_eq!(core::mem::size_of::<LeafNode<i64, i64>>(), 16 + CAPACITY * 2 * 8); + assert_eq!(core::mem::size_of::<InternalNode<(), ()>>(), 16 + (CAPACITY + 1) * 8); + assert_eq!(core::mem::size_of::<InternalNode<i64, i64>>(), 16 + (CAPACITY * 3 + 1) * 8); +} diff --git a/library/alloc/src/collections/btree/remove.rs b/library/alloc/src/collections/btree/remove.rs new file mode 100644 index 000000000..090429925 --- /dev/null +++ b/library/alloc/src/collections/btree/remove.rs @@ -0,0 +1,95 @@ +use super::map::MIN_LEN; +use super::node::{marker, ForceResult::*, Handle, LeftOrRight::*, NodeRef}; +use core::alloc::Allocator; + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::LeafOrInternal>, marker::KV> { + /// Removes a key-value pair from the tree, and returns that pair, as well as + /// the leaf edge corresponding to that former pair. It's possible this empties + /// a root node that is internal, which the caller should pop from the map + /// holding the tree. The caller should also decrement the map's length. + pub fn remove_kv_tracking<F: FnOnce(), A: Allocator + Clone>( + self, + handle_emptied_internal_root: F, + alloc: A, + ) -> ((K, V), Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>) { + match self.force() { + Leaf(node) => node.remove_leaf_kv(handle_emptied_internal_root, alloc), + Internal(node) => node.remove_internal_kv(handle_emptied_internal_root, alloc), + } + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::KV> { + fn remove_leaf_kv<F: FnOnce(), A: Allocator + Clone>( + self, + handle_emptied_internal_root: F, + alloc: A, + ) -> ((K, V), Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>) { + let (old_kv, mut pos) = self.remove(); + let len = pos.reborrow().into_node().len(); + if len < MIN_LEN { + let idx = pos.idx(); + // We have to temporarily forget the child type, because there is no + // distinct node type for the immediate parents of a leaf. + let new_pos = match pos.into_node().forget_type().choose_parent_kv() { + Ok(Left(left_parent_kv)) => { + debug_assert!(left_parent_kv.right_child_len() == MIN_LEN - 1); + if left_parent_kv.can_merge() { + left_parent_kv.merge_tracking_child_edge(Right(idx), alloc.clone()) + } else { + debug_assert!(left_parent_kv.left_child_len() > MIN_LEN); + left_parent_kv.steal_left(idx) + } + } + Ok(Right(right_parent_kv)) => { + debug_assert!(right_parent_kv.left_child_len() == MIN_LEN - 1); + if right_parent_kv.can_merge() { + right_parent_kv.merge_tracking_child_edge(Left(idx), alloc.clone()) + } else { + debug_assert!(right_parent_kv.right_child_len() > MIN_LEN); + right_parent_kv.steal_right(idx) + } + } + Err(pos) => unsafe { Handle::new_edge(pos, idx) }, + }; + // SAFETY: `new_pos` is the leaf we started from or a sibling. + pos = unsafe { new_pos.cast_to_leaf_unchecked() }; + + // Only if we merged, the parent (if any) has shrunk, but skipping + // the following step otherwise does not pay off in benchmarks. + // + // SAFETY: We won't destroy or rearrange the leaf where `pos` is at + // by handling its parent recursively; at worst we will destroy or + // rearrange the parent through the grandparent, thus change the + // link to the parent inside the leaf. + if let Ok(parent) = unsafe { pos.reborrow_mut() }.into_node().ascend() { + if !parent.into_node().forget_type().fix_node_and_affected_ancestors(alloc) { + handle_emptied_internal_root(); + } + } + } + (old_kv, pos) + } +} + +impl<'a, K: 'a, V: 'a> Handle<NodeRef<marker::Mut<'a>, K, V, marker::Internal>, marker::KV> { + fn remove_internal_kv<F: FnOnce(), A: Allocator + Clone>( + self, + handle_emptied_internal_root: F, + alloc: A, + ) -> ((K, V), Handle<NodeRef<marker::Mut<'a>, K, V, marker::Leaf>, marker::Edge>) { + // Remove an adjacent KV from its leaf and then put it back in place of + // the element we were asked to remove. Prefer the left adjacent KV, + // for the reasons listed in `choose_parent_kv`. + let left_leaf_kv = self.left_edge().descend().last_leaf_edge().left_kv(); + let left_leaf_kv = unsafe { left_leaf_kv.ok().unwrap_unchecked() }; + let (left_kv, left_hole) = left_leaf_kv.remove_leaf_kv(handle_emptied_internal_root, alloc); + + // The internal node may have been stolen from or merged. Go back right + // to find where the original KV ended up. + let mut internal = unsafe { left_hole.next_kv().ok().unwrap_unchecked() }; + let old_kv = internal.replace_kv(left_kv.0, left_kv.1); + let pos = internal.next_leaf_edge(); + (old_kv, pos) + } +} diff --git a/library/alloc/src/collections/btree/search.rs b/library/alloc/src/collections/btree/search.rs new file mode 100644 index 000000000..ad3522b4e --- /dev/null +++ b/library/alloc/src/collections/btree/search.rs @@ -0,0 +1,285 @@ +use core::borrow::Borrow; +use core::cmp::Ordering; +use core::ops::{Bound, RangeBounds}; + +use super::node::{marker, ForceResult::*, Handle, NodeRef}; + +use SearchBound::*; +use SearchResult::*; + +pub enum SearchBound<T> { + /// An inclusive bound to look for, just like `Bound::Included(T)`. + Included(T), + /// An exclusive bound to look for, just like `Bound::Excluded(T)`. + Excluded(T), + /// An unconditional inclusive bound, just like `Bound::Unbounded`. + AllIncluded, + /// An unconditional exclusive bound. + AllExcluded, +} + +impl<T> SearchBound<T> { + pub fn from_range(range_bound: Bound<T>) -> Self { + match range_bound { + Bound::Included(t) => Included(t), + Bound::Excluded(t) => Excluded(t), + Bound::Unbounded => AllIncluded, + } + } +} + +pub enum SearchResult<BorrowType, K, V, FoundType, GoDownType> { + Found(Handle<NodeRef<BorrowType, K, V, FoundType>, marker::KV>), + GoDown(Handle<NodeRef<BorrowType, K, V, GoDownType>, marker::Edge>), +} + +pub enum IndexResult { + KV(usize), + Edge(usize), +} + +impl<BorrowType: marker::BorrowType, K, V> NodeRef<BorrowType, K, V, marker::LeafOrInternal> { + /// Looks up a given key in a (sub)tree headed by the node, recursively. + /// Returns a `Found` with the handle of the matching KV, if any. Otherwise, + /// returns a `GoDown` with the handle of the leaf edge where the key belongs. + /// + /// The result is meaningful only if the tree is ordered by key, like the tree + /// in a `BTreeMap` is. + pub fn search_tree<Q: ?Sized>( + mut self, + key: &Q, + ) -> SearchResult<BorrowType, K, V, marker::LeafOrInternal, marker::Leaf> + where + Q: Ord, + K: Borrow<Q>, + { + loop { + self = match self.search_node(key) { + Found(handle) => return Found(handle), + GoDown(handle) => match handle.force() { + Leaf(leaf) => return GoDown(leaf), + Internal(internal) => internal.descend(), + }, + } + } + } + + /// Descends to the nearest node where the edge matching the lower bound + /// of the range is different from the edge matching the upper bound, i.e., + /// the nearest node that has at least one key contained in the range. + /// + /// If found, returns an `Ok` with that node, the strictly ascending pair of + /// edge indices in the node delimiting the range, and the corresponding + /// pair of bounds for continuing the search in the child nodes, in case + /// the node is internal. + /// + /// If not found, returns an `Err` with the leaf edge matching the entire + /// range. + /// + /// As a diagnostic service, panics if the range specifies impossible bounds. + /// + /// The result is meaningful only if the tree is ordered by key. + pub fn search_tree_for_bifurcation<'r, Q: ?Sized, R>( + mut self, + range: &'r R, + ) -> Result< + ( + NodeRef<BorrowType, K, V, marker::LeafOrInternal>, + usize, + usize, + SearchBound<&'r Q>, + SearchBound<&'r Q>, + ), + Handle<NodeRef<BorrowType, K, V, marker::Leaf>, marker::Edge>, + > + where + Q: Ord, + K: Borrow<Q>, + R: RangeBounds<Q>, + { + // Determine if map or set is being searched + let is_set = <V as super::set_val::IsSetVal>::is_set_val(); + + // Inlining these variables should be avoided. We assume the bounds reported by `range` + // remain the same, but an adversarial implementation could change between calls (#81138). + let (start, end) = (range.start_bound(), range.end_bound()); + match (start, end) { + (Bound::Excluded(s), Bound::Excluded(e)) if s == e => { + if is_set { + panic!("range start and end are equal and excluded in BTreeSet") + } else { + panic!("range start and end are equal and excluded in BTreeMap") + } + } + (Bound::Included(s) | Bound::Excluded(s), Bound::Included(e) | Bound::Excluded(e)) + if s > e => + { + if is_set { + panic!("range start is greater than range end in BTreeSet") + } else { + panic!("range start is greater than range end in BTreeMap") + } + } + _ => {} + } + let mut lower_bound = SearchBound::from_range(start); + let mut upper_bound = SearchBound::from_range(end); + loop { + let (lower_edge_idx, lower_child_bound) = self.find_lower_bound_index(lower_bound); + let (upper_edge_idx, upper_child_bound) = + unsafe { self.find_upper_bound_index(upper_bound, lower_edge_idx) }; + if lower_edge_idx < upper_edge_idx { + return Ok(( + self, + lower_edge_idx, + upper_edge_idx, + lower_child_bound, + upper_child_bound, + )); + } + debug_assert_eq!(lower_edge_idx, upper_edge_idx); + let common_edge = unsafe { Handle::new_edge(self, lower_edge_idx) }; + match common_edge.force() { + Leaf(common_edge) => return Err(common_edge), + Internal(common_edge) => { + self = common_edge.descend(); + lower_bound = lower_child_bound; + upper_bound = upper_child_bound; + } + } + } + } + + /// Finds an edge in the node delimiting the lower bound of a range. + /// Also returns the lower bound to be used for continuing the search in + /// the matching child node, if `self` is an internal node. + /// + /// The result is meaningful only if the tree is ordered by key. + pub fn find_lower_bound_edge<'r, Q>( + self, + bound: SearchBound<&'r Q>, + ) -> (Handle<Self, marker::Edge>, SearchBound<&'r Q>) + where + Q: ?Sized + Ord, + K: Borrow<Q>, + { + let (edge_idx, bound) = self.find_lower_bound_index(bound); + let edge = unsafe { Handle::new_edge(self, edge_idx) }; + (edge, bound) + } + + /// Clone of `find_lower_bound_edge` for the upper bound. + pub fn find_upper_bound_edge<'r, Q>( + self, + bound: SearchBound<&'r Q>, + ) -> (Handle<Self, marker::Edge>, SearchBound<&'r Q>) + where + Q: ?Sized + Ord, + K: Borrow<Q>, + { + let (edge_idx, bound) = unsafe { self.find_upper_bound_index(bound, 0) }; + let edge = unsafe { Handle::new_edge(self, edge_idx) }; + (edge, bound) + } +} + +impl<BorrowType, K, V, Type> NodeRef<BorrowType, K, V, Type> { + /// Looks up a given key in the node, without recursion. + /// Returns a `Found` with the handle of the matching KV, if any. Otherwise, + /// returns a `GoDown` with the handle of the edge where the key might be found + /// (if the node is internal) or where the key can be inserted. + /// + /// The result is meaningful only if the tree is ordered by key, like the tree + /// in a `BTreeMap` is. + pub fn search_node<Q: ?Sized>(self, key: &Q) -> SearchResult<BorrowType, K, V, Type, Type> + where + Q: Ord, + K: Borrow<Q>, + { + match unsafe { self.find_key_index(key, 0) } { + IndexResult::KV(idx) => Found(unsafe { Handle::new_kv(self, idx) }), + IndexResult::Edge(idx) => GoDown(unsafe { Handle::new_edge(self, idx) }), + } + } + + /// Returns either the KV index in the node at which the key (or an equivalent) + /// exists, or the edge index where the key belongs, starting from a particular index. + /// + /// The result is meaningful only if the tree is ordered by key, like the tree + /// in a `BTreeMap` is. + /// + /// # Safety + /// `start_index` must be a valid edge index for the node. + unsafe fn find_key_index<Q: ?Sized>(&self, key: &Q, start_index: usize) -> IndexResult + where + Q: Ord, + K: Borrow<Q>, + { + let node = self.reborrow(); + let keys = node.keys(); + debug_assert!(start_index <= keys.len()); + for (offset, k) in unsafe { keys.get_unchecked(start_index..) }.iter().enumerate() { + match key.cmp(k.borrow()) { + Ordering::Greater => {} + Ordering::Equal => return IndexResult::KV(start_index + offset), + Ordering::Less => return IndexResult::Edge(start_index + offset), + } + } + IndexResult::Edge(keys.len()) + } + + /// Finds an edge index in the node delimiting the lower bound of a range. + /// Also returns the lower bound to be used for continuing the search in + /// the matching child node, if `self` is an internal node. + /// + /// The result is meaningful only if the tree is ordered by key. + fn find_lower_bound_index<'r, Q>( + &self, + bound: SearchBound<&'r Q>, + ) -> (usize, SearchBound<&'r Q>) + where + Q: ?Sized + Ord, + K: Borrow<Q>, + { + match bound { + Included(key) => match unsafe { self.find_key_index(key, 0) } { + IndexResult::KV(idx) => (idx, AllExcluded), + IndexResult::Edge(idx) => (idx, bound), + }, + Excluded(key) => match unsafe { self.find_key_index(key, 0) } { + IndexResult::KV(idx) => (idx + 1, AllIncluded), + IndexResult::Edge(idx) => (idx, bound), + }, + AllIncluded => (0, AllIncluded), + AllExcluded => (self.len(), AllExcluded), + } + } + + /// Mirror image of `find_lower_bound_index` for the upper bound, + /// with an additional parameter to skip part of the key array. + /// + /// # Safety + /// `start_index` must be a valid edge index for the node. + unsafe fn find_upper_bound_index<'r, Q>( + &self, + bound: SearchBound<&'r Q>, + start_index: usize, + ) -> (usize, SearchBound<&'r Q>) + where + Q: ?Sized + Ord, + K: Borrow<Q>, + { + match bound { + Included(key) => match unsafe { self.find_key_index(key, start_index) } { + IndexResult::KV(idx) => (idx + 1, AllExcluded), + IndexResult::Edge(idx) => (idx, bound), + }, + Excluded(key) => match unsafe { self.find_key_index(key, start_index) } { + IndexResult::KV(idx) => (idx, AllIncluded), + IndexResult::Edge(idx) => (idx, bound), + }, + AllIncluded => (self.len(), AllIncluded), + AllExcluded => (start_index, AllExcluded), + } + } +} diff --git a/library/alloc/src/collections/btree/set.rs b/library/alloc/src/collections/btree/set.rs new file mode 100644 index 000000000..2cfc08074 --- /dev/null +++ b/library/alloc/src/collections/btree/set.rs @@ -0,0 +1,1789 @@ +// This is pretty much entirely stolen from TreeSet, since BTreeMap has an identical interface +// to TreeMap + +use crate::vec::Vec; +use core::borrow::Borrow; +use core::cmp::Ordering::{self, Equal, Greater, Less}; +use core::cmp::{max, min}; +use core::fmt::{self, Debug}; +use core::hash::{Hash, Hasher}; +use core::iter::{FromIterator, FusedIterator, Peekable}; +use core::mem::ManuallyDrop; +use core::ops::{BitAnd, BitOr, BitXor, RangeBounds, Sub}; + +use super::map::{BTreeMap, Keys}; +use super::merge_iter::MergeIterInner; +use super::set_val::SetValZST; +use super::Recover; + +use crate::alloc::{Allocator, Global}; + +// FIXME(conventions): implement bounded iterators + +/// An ordered set based on a B-Tree. +/// +/// See [`BTreeMap`]'s documentation for a detailed discussion of this collection's performance +/// benefits and drawbacks. +/// +/// It is a logic error for an item to be modified in such a way that the item's ordering relative +/// to any other item, as determined by the [`Ord`] trait, changes while it is in the set. This is +/// normally only possible through [`Cell`], [`RefCell`], global state, I/O, or unsafe code. +/// The behavior resulting from such a logic error is not specified, but will be encapsulated to the +/// `BTreeSet` that observed the logic error and not result in undefined behavior. This could +/// include panics, incorrect results, aborts, memory leaks, and non-termination. +/// +/// Iterators returned by [`BTreeSet::iter`] produce their items in order, and take worst-case +/// logarithmic and amortized constant time per item returned. +/// +/// [`Ord`]: core::cmp::Ord +/// [`Cell`]: core::cell::Cell +/// [`RefCell`]: core::cell::RefCell +/// +/// # Examples +/// +/// ``` +/// use std::collections::BTreeSet; +/// +/// // Type inference lets us omit an explicit type signature (which +/// // would be `BTreeSet<&str>` in this example). +/// let mut books = BTreeSet::new(); +/// +/// // Add some books. +/// books.insert("A Dance With Dragons"); +/// books.insert("To Kill a Mockingbird"); +/// books.insert("The Odyssey"); +/// books.insert("The Great Gatsby"); +/// +/// // Check for a specific one. +/// if !books.contains("The Winds of Winter") { +/// println!("We have {} books, but The Winds of Winter ain't one.", +/// books.len()); +/// } +/// +/// // Remove a book. +/// books.remove("The Odyssey"); +/// +/// // Iterate over everything. +/// for book in &books { +/// println!("{book}"); +/// } +/// ``` +/// +/// A `BTreeSet` with a known list of items can be initialized from an array: +/// +/// ``` +/// use std::collections::BTreeSet; +/// +/// let set = BTreeSet::from([1, 2, 3]); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "BTreeSet")] +pub struct BTreeSet< + T, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + map: BTreeMap<T, SetValZST, A>, +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Hash, A: Allocator + Clone> Hash for BTreeSet<T, A> { + fn hash<H: Hasher>(&self, state: &mut H) { + self.map.hash(state) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialEq, A: Allocator + Clone> PartialEq for BTreeSet<T, A> { + fn eq(&self, other: &BTreeSet<T, A>) -> bool { + self.map.eq(&other.map) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Eq, A: Allocator + Clone> Eq for BTreeSet<T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialOrd, A: Allocator + Clone> PartialOrd for BTreeSet<T, A> { + fn partial_cmp(&self, other: &BTreeSet<T, A>) -> Option<Ordering> { + self.map.partial_cmp(&other.map) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord, A: Allocator + Clone> Ord for BTreeSet<T, A> { + fn cmp(&self, other: &BTreeSet<T, A>) -> Ordering { + self.map.cmp(&other.map) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone, A: Allocator + Clone> Clone for BTreeSet<T, A> { + fn clone(&self) -> Self { + BTreeSet { map: self.map.clone() } + } + + fn clone_from(&mut self, other: &Self) { + self.map.clone_from(&other.map); + } +} + +/// An iterator over the items of a `BTreeSet`. +/// +/// This `struct` is created by the [`iter`] method on [`BTreeSet`]. +/// See its documentation for more. +/// +/// [`iter`]: BTreeSet::iter +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Iter<'a, T: 'a> { + iter: Keys<'a, T, SetValZST>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for Iter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Iter").field(&self.iter.clone()).finish() + } +} + +/// An owning iterator over the items of a `BTreeSet`. +/// +/// This `struct` is created by the [`into_iter`] method on [`BTreeSet`] +/// (provided by the [`IntoIterator`] trait). See its documentation for more. +/// +/// [`into_iter`]: BTreeSet#method.into_iter +/// [`IntoIterator`]: core::iter::IntoIterator +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Debug)] +pub struct IntoIter< + T, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + iter: super::map::IntoIter<T, SetValZST, A>, +} + +/// An iterator over a sub-range of items in a `BTreeSet`. +/// +/// This `struct` is created by the [`range`] method on [`BTreeSet`]. +/// See its documentation for more. +/// +/// [`range`]: BTreeSet::range +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[derive(Debug)] +#[stable(feature = "btree_range", since = "1.17.0")] +pub struct Range<'a, T: 'a> { + iter: super::map::Range<'a, T, SetValZST>, +} + +/// A lazy iterator producing elements in the difference of `BTreeSet`s. +/// +/// This `struct` is created by the [`difference`] method on [`BTreeSet`]. +/// See its documentation for more. +/// +/// [`difference`]: BTreeSet::difference +#[must_use = "this returns the difference as an iterator, \ + without modifying either input set"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Difference< + 'a, + T: 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + inner: DifferenceInner<'a, T, A>, +} +enum DifferenceInner<'a, T: 'a, A: Allocator + Clone> { + Stitch { + // iterate all of `self` and some of `other`, spotting matches along the way + self_iter: Iter<'a, T>, + other_iter: Peekable<Iter<'a, T>>, + }, + Search { + // iterate `self`, look up in `other` + self_iter: Iter<'a, T>, + other_set: &'a BTreeSet<T, A>, + }, + Iterate(Iter<'a, T>), // simply produce all elements in `self` +} + +// Explicit Debug impl necessary because of issue #26925 +impl<T: Debug, A: Allocator + Clone> Debug for DifferenceInner<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match self { + DifferenceInner::Stitch { self_iter, other_iter } => f + .debug_struct("Stitch") + .field("self_iter", self_iter) + .field("other_iter", other_iter) + .finish(), + DifferenceInner::Search { self_iter, other_set } => f + .debug_struct("Search") + .field("self_iter", self_iter) + .field("other_iter", other_set) + .finish(), + DifferenceInner::Iterate(x) => f.debug_tuple("Iterate").field(x).finish(), + } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug, A: Allocator + Clone> fmt::Debug for Difference<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Difference").field(&self.inner).finish() + } +} + +/// A lazy iterator producing elements in the symmetric difference of `BTreeSet`s. +/// +/// This `struct` is created by the [`symmetric_difference`] method on +/// [`BTreeSet`]. See its documentation for more. +/// +/// [`symmetric_difference`]: BTreeSet::symmetric_difference +#[must_use = "this returns the difference as an iterator, \ + without modifying either input set"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct SymmetricDifference<'a, T: 'a>(MergeIterInner<Iter<'a, T>>); + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for SymmetricDifference<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("SymmetricDifference").field(&self.0).finish() + } +} + +/// A lazy iterator producing elements in the intersection of `BTreeSet`s. +/// +/// This `struct` is created by the [`intersection`] method on [`BTreeSet`]. +/// See its documentation for more. +/// +/// [`intersection`]: BTreeSet::intersection +#[must_use = "this returns the intersection as an iterator, \ + without modifying either input set"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Intersection< + 'a, + T: 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> { + inner: IntersectionInner<'a, T, A>, +} +enum IntersectionInner<'a, T: 'a, A: Allocator + Clone> { + Stitch { + // iterate similarly sized sets jointly, spotting matches along the way + a: Iter<'a, T>, + b: Iter<'a, T>, + }, + Search { + // iterate a small set, look up in the large set + small_iter: Iter<'a, T>, + large_set: &'a BTreeSet<T, A>, + }, + Answer(Option<&'a T>), // return a specific element or emptiness +} + +// Explicit Debug impl necessary because of issue #26925 +impl<T: Debug, A: Allocator + Clone> Debug for IntersectionInner<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match self { + IntersectionInner::Stitch { a, b } => { + f.debug_struct("Stitch").field("a", a).field("b", b).finish() + } + IntersectionInner::Search { small_iter, large_set } => f + .debug_struct("Search") + .field("small_iter", small_iter) + .field("large_set", large_set) + .finish(), + IntersectionInner::Answer(x) => f.debug_tuple("Answer").field(x).finish(), + } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: Debug, A: Allocator + Clone> Debug for Intersection<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Intersection").field(&self.inner).finish() + } +} + +/// A lazy iterator producing elements in the union of `BTreeSet`s. +/// +/// This `struct` is created by the [`union`] method on [`BTreeSet`]. +/// See its documentation for more. +/// +/// [`union`]: BTreeSet::union +#[must_use = "this returns the union as an iterator, \ + without modifying either input set"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Union<'a, T: 'a>(MergeIterInner<Iter<'a, T>>); + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for Union<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Union").field(&self.0).finish() + } +} + +// This constant is used by functions that compare two sets. +// It estimates the relative size at which searching performs better +// than iterating, based on the benchmarks in +// https://github.com/ssomers/rust_bench_btreeset_intersection. +// It's used to divide rather than multiply sizes, to rule out overflow, +// and it's a power of two to make that division cheap. +const ITER_PERFORMANCE_TIPPING_SIZE_DIFF: usize = 16; + +impl<T> BTreeSet<T> { + /// Makes a new, empty `BTreeSet`. + /// + /// Does not allocate anything on its own. + /// + /// # Examples + /// + /// ``` + /// # #![allow(unused_mut)] + /// use std::collections::BTreeSet; + /// + /// let mut set: BTreeSet<i32> = BTreeSet::new(); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + #[must_use] + pub const fn new() -> BTreeSet<T> { + BTreeSet { map: BTreeMap::new() } + } +} + +impl<T, A: Allocator + Clone> BTreeSet<T, A> { + /// Makes a new `BTreeSet` with a reasonable choice of B. + /// + /// # Examples + /// + /// ``` + /// # #![allow(unused_mut)] + /// # #![feature(allocator_api)] + /// # #![feature(btreemap_alloc)] + /// use std::collections::BTreeSet; + /// use std::alloc::Global; + /// + /// let mut set: BTreeSet<i32> = BTreeSet::new_in(Global); + /// ``` + #[unstable(feature = "btreemap_alloc", issue = "32838")] + pub fn new_in(alloc: A) -> BTreeSet<T, A> { + BTreeSet { map: BTreeMap::new_in(alloc) } + } + + /// Constructs a double-ended iterator over a sub-range of elements in the set. + /// The simplest way is to use the range syntax `min..max`, thus `range(min..max)` will + /// yield elements from min (inclusive) to max (exclusive). + /// The range may also be entered as `(Bound<T>, Bound<T>)`, so for example + /// `range((Excluded(4), Included(10)))` will yield a left-exclusive, right-inclusive + /// range from 4 to 10. + /// + /// # Panics + /// + /// Panics if range `start > end`. + /// Panics if range `start == end` and both bounds are `Excluded`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// use std::ops::Bound::Included; + /// + /// let mut set = BTreeSet::new(); + /// set.insert(3); + /// set.insert(5); + /// set.insert(8); + /// for &elem in set.range((Included(&4), Included(&8))) { + /// println!("{elem}"); + /// } + /// assert_eq!(Some(&5), set.range(4..).next()); + /// ``` + #[stable(feature = "btree_range", since = "1.17.0")] + pub fn range<K: ?Sized, R>(&self, range: R) -> Range<'_, T> + where + K: Ord, + T: Borrow<K> + Ord, + R: RangeBounds<K>, + { + Range { iter: self.map.range(range) } + } + + /// Visits the elements representing the difference, + /// i.e., the elements that are in `self` but not in `other`, + /// in ascending order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// a.insert(2); + /// + /// let mut b = BTreeSet::new(); + /// b.insert(2); + /// b.insert(3); + /// + /// let diff: Vec<_> = a.difference(&b).cloned().collect(); + /// assert_eq!(diff, [1]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn difference<'a>(&'a self, other: &'a BTreeSet<T, A>) -> Difference<'a, T, A> + where + T: Ord, + { + let (self_min, self_max) = + if let (Some(self_min), Some(self_max)) = (self.first(), self.last()) { + (self_min, self_max) + } else { + return Difference { inner: DifferenceInner::Iterate(self.iter()) }; + }; + let (other_min, other_max) = + if let (Some(other_min), Some(other_max)) = (other.first(), other.last()) { + (other_min, other_max) + } else { + return Difference { inner: DifferenceInner::Iterate(self.iter()) }; + }; + Difference { + inner: match (self_min.cmp(other_max), self_max.cmp(other_min)) { + (Greater, _) | (_, Less) => DifferenceInner::Iterate(self.iter()), + (Equal, _) => { + let mut self_iter = self.iter(); + self_iter.next(); + DifferenceInner::Iterate(self_iter) + } + (_, Equal) => { + let mut self_iter = self.iter(); + self_iter.next_back(); + DifferenceInner::Iterate(self_iter) + } + _ if self.len() <= other.len() / ITER_PERFORMANCE_TIPPING_SIZE_DIFF => { + DifferenceInner::Search { self_iter: self.iter(), other_set: other } + } + _ => DifferenceInner::Stitch { + self_iter: self.iter(), + other_iter: other.iter().peekable(), + }, + }, + } + } + + /// Visits the elements representing the symmetric difference, + /// i.e., the elements that are in `self` or in `other` but not in both, + /// in ascending order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// a.insert(2); + /// + /// let mut b = BTreeSet::new(); + /// b.insert(2); + /// b.insert(3); + /// + /// let sym_diff: Vec<_> = a.symmetric_difference(&b).cloned().collect(); + /// assert_eq!(sym_diff, [1, 3]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn symmetric_difference<'a>( + &'a self, + other: &'a BTreeSet<T, A>, + ) -> SymmetricDifference<'a, T> + where + T: Ord, + { + SymmetricDifference(MergeIterInner::new(self.iter(), other.iter())) + } + + /// Visits the elements representing the intersection, + /// i.e., the elements that are both in `self` and `other`, + /// in ascending order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// a.insert(2); + /// + /// let mut b = BTreeSet::new(); + /// b.insert(2); + /// b.insert(3); + /// + /// let intersection: Vec<_> = a.intersection(&b).cloned().collect(); + /// assert_eq!(intersection, [2]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn intersection<'a>(&'a self, other: &'a BTreeSet<T, A>) -> Intersection<'a, T, A> + where + T: Ord, + { + let (self_min, self_max) = + if let (Some(self_min), Some(self_max)) = (self.first(), self.last()) { + (self_min, self_max) + } else { + return Intersection { inner: IntersectionInner::Answer(None) }; + }; + let (other_min, other_max) = + if let (Some(other_min), Some(other_max)) = (other.first(), other.last()) { + (other_min, other_max) + } else { + return Intersection { inner: IntersectionInner::Answer(None) }; + }; + Intersection { + inner: match (self_min.cmp(other_max), self_max.cmp(other_min)) { + (Greater, _) | (_, Less) => IntersectionInner::Answer(None), + (Equal, _) => IntersectionInner::Answer(Some(self_min)), + (_, Equal) => IntersectionInner::Answer(Some(self_max)), + _ if self.len() <= other.len() / ITER_PERFORMANCE_TIPPING_SIZE_DIFF => { + IntersectionInner::Search { small_iter: self.iter(), large_set: other } + } + _ if other.len() <= self.len() / ITER_PERFORMANCE_TIPPING_SIZE_DIFF => { + IntersectionInner::Search { small_iter: other.iter(), large_set: self } + } + _ => IntersectionInner::Stitch { a: self.iter(), b: other.iter() }, + }, + } + } + + /// Visits the elements representing the union, + /// i.e., all the elements in `self` or `other`, without duplicates, + /// in ascending order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// + /// let mut b = BTreeSet::new(); + /// b.insert(2); + /// + /// let union: Vec<_> = a.union(&b).cloned().collect(); + /// assert_eq!(union, [1, 2]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn union<'a>(&'a self, other: &'a BTreeSet<T, A>) -> Union<'a, T> + where + T: Ord, + { + Union(MergeIterInner::new(self.iter(), other.iter())) + } + + /// Clears the set, removing all elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut v = BTreeSet::new(); + /// v.insert(1); + /// v.clear(); + /// assert!(v.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) + where + A: Clone, + { + self.map.clear() + } + + /// Returns `true` if the set contains an element equal to the value. + /// + /// The value may be any borrowed form of the set's element type, + /// but the ordering on the borrowed form *must* match the + /// ordering on the element type. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set = BTreeSet::from([1, 2, 3]); + /// assert_eq!(set.contains(&1), true); + /// assert_eq!(set.contains(&4), false); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn contains<Q: ?Sized>(&self, value: &Q) -> bool + where + T: Borrow<Q> + Ord, + Q: Ord, + { + self.map.contains_key(value) + } + + /// Returns a reference to the element in the set, if any, that is equal to + /// the value. + /// + /// The value may be any borrowed form of the set's element type, + /// but the ordering on the borrowed form *must* match the + /// ordering on the element type. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set = BTreeSet::from([1, 2, 3]); + /// assert_eq!(set.get(&2), Some(&2)); + /// assert_eq!(set.get(&4), None); + /// ``` + #[stable(feature = "set_recovery", since = "1.9.0")] + pub fn get<Q: ?Sized>(&self, value: &Q) -> Option<&T> + where + T: Borrow<Q> + Ord, + Q: Ord, + { + Recover::get(&self.map, value) + } + + /// Returns `true` if `self` has no elements in common with `other`. + /// This is equivalent to checking for an empty intersection. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let a = BTreeSet::from([1, 2, 3]); + /// let mut b = BTreeSet::new(); + /// + /// assert_eq!(a.is_disjoint(&b), true); + /// b.insert(4); + /// assert_eq!(a.is_disjoint(&b), true); + /// b.insert(1); + /// assert_eq!(a.is_disjoint(&b), false); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_disjoint(&self, other: &BTreeSet<T, A>) -> bool + where + T: Ord, + { + self.intersection(other).next().is_none() + } + + /// Returns `true` if the set is a subset of another, + /// i.e., `other` contains at least all the elements in `self`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let sup = BTreeSet::from([1, 2, 3]); + /// let mut set = BTreeSet::new(); + /// + /// assert_eq!(set.is_subset(&sup), true); + /// set.insert(2); + /// assert_eq!(set.is_subset(&sup), true); + /// set.insert(4); + /// assert_eq!(set.is_subset(&sup), false); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_subset(&self, other: &BTreeSet<T, A>) -> bool + where + T: Ord, + { + // Same result as self.difference(other).next().is_none() + // but the code below is faster (hugely in some cases). + if self.len() > other.len() { + return false; + } + let (self_min, self_max) = + if let (Some(self_min), Some(self_max)) = (self.first(), self.last()) { + (self_min, self_max) + } else { + return true; // self is empty + }; + let (other_min, other_max) = + if let (Some(other_min), Some(other_max)) = (other.first(), other.last()) { + (other_min, other_max) + } else { + return false; // other is empty + }; + let mut self_iter = self.iter(); + match self_min.cmp(other_min) { + Less => return false, + Equal => { + self_iter.next(); + } + Greater => (), + } + match self_max.cmp(other_max) { + Greater => return false, + Equal => { + self_iter.next_back(); + } + Less => (), + } + if self_iter.len() <= other.len() / ITER_PERFORMANCE_TIPPING_SIZE_DIFF { + for next in self_iter { + if !other.contains(next) { + return false; + } + } + } else { + let mut other_iter = other.iter(); + other_iter.next(); + other_iter.next_back(); + let mut self_next = self_iter.next(); + while let Some(self1) = self_next { + match other_iter.next().map_or(Less, |other1| self1.cmp(other1)) { + Less => return false, + Equal => self_next = self_iter.next(), + Greater => (), + } + } + } + true + } + + /// Returns `true` if the set is a superset of another, + /// i.e., `self` contains at least all the elements in `other`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let sub = BTreeSet::from([1, 2]); + /// let mut set = BTreeSet::new(); + /// + /// assert_eq!(set.is_superset(&sub), false); + /// + /// set.insert(0); + /// set.insert(1); + /// assert_eq!(set.is_superset(&sub), false); + /// + /// set.insert(2); + /// assert_eq!(set.is_superset(&sub), true); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_superset(&self, other: &BTreeSet<T, A>) -> bool + where + T: Ord, + { + other.is_subset(self) + } + + /// Returns a reference to the first element in the set, if any. + /// This element is always the minimum of all elements in the set. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// assert_eq!(set.first(), None); + /// set.insert(1); + /// assert_eq!(set.first(), Some(&1)); + /// set.insert(2); + /// assert_eq!(set.first(), Some(&1)); + /// ``` + #[must_use] + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn first(&self) -> Option<&T> + where + T: Ord, + { + self.map.first_key_value().map(|(k, _)| k) + } + + /// Returns a reference to the last element in the set, if any. + /// This element is always the maximum of all elements in the set. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// assert_eq!(set.last(), None); + /// set.insert(1); + /// assert_eq!(set.last(), Some(&1)); + /// set.insert(2); + /// assert_eq!(set.last(), Some(&2)); + /// ``` + #[must_use] + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn last(&self) -> Option<&T> + where + T: Ord, + { + self.map.last_key_value().map(|(k, _)| k) + } + + /// Removes the first element from the set and returns it, if any. + /// The first element is always the minimum element in the set. + /// + /// # Examples + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// + /// set.insert(1); + /// while let Some(n) = set.pop_first() { + /// assert_eq!(n, 1); + /// } + /// assert!(set.is_empty()); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn pop_first(&mut self) -> Option<T> + where + T: Ord, + { + self.map.pop_first().map(|kv| kv.0) + } + + /// Removes the last element from the set and returns it, if any. + /// The last element is always the maximum element in the set. + /// + /// # Examples + /// + /// ``` + /// #![feature(map_first_last)] + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// + /// set.insert(1); + /// while let Some(n) = set.pop_last() { + /// assert_eq!(n, 1); + /// } + /// assert!(set.is_empty()); + /// ``` + #[unstable(feature = "map_first_last", issue = "62924")] + pub fn pop_last(&mut self) -> Option<T> + where + T: Ord, + { + self.map.pop_last().map(|kv| kv.0) + } + + /// Adds a value to the set. + /// + /// Returns whether the value was newly inserted. That is: + /// + /// - If the set did not previously contain an equal value, `true` is + /// returned. + /// - If the set already contained an equal value, `false` is returned, and + /// the entry is not updated. + /// + /// See the [module-level documentation] for more. + /// + /// [module-level documentation]: index.html#insert-and-complex-keys + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// + /// assert_eq!(set.insert(2), true); + /// assert_eq!(set.insert(2), false); + /// assert_eq!(set.len(), 1); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(&mut self, value: T) -> bool + where + T: Ord, + { + self.map.insert(value, SetValZST::default()).is_none() + } + + /// Adds a value to the set, replacing the existing element, if any, that is + /// equal to the value. Returns the replaced element. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// set.insert(Vec::<i32>::new()); + /// + /// assert_eq!(set.get(&[][..]).unwrap().capacity(), 0); + /// set.replace(Vec::with_capacity(10)); + /// assert_eq!(set.get(&[][..]).unwrap().capacity(), 10); + /// ``` + #[stable(feature = "set_recovery", since = "1.9.0")] + pub fn replace(&mut self, value: T) -> Option<T> + where + T: Ord, + { + Recover::replace(&mut self.map, value) + } + + /// If the set contains an element equal to the value, removes it from the + /// set and drops it. Returns whether such an element was present. + /// + /// The value may be any borrowed form of the set's element type, + /// but the ordering on the borrowed form *must* match the + /// ordering on the element type. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::new(); + /// + /// set.insert(2); + /// assert_eq!(set.remove(&2), true); + /// assert_eq!(set.remove(&2), false); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn remove<Q: ?Sized>(&mut self, value: &Q) -> bool + where + T: Borrow<Q> + Ord, + Q: Ord, + { + self.map.remove(value).is_some() + } + + /// Removes and returns the element in the set, if any, that is equal to + /// the value. + /// + /// The value may be any borrowed form of the set's element type, + /// but the ordering on the borrowed form *must* match the + /// ordering on the element type. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::from([1, 2, 3]); + /// assert_eq!(set.take(&2), Some(2)); + /// assert_eq!(set.take(&2), None); + /// ``` + #[stable(feature = "set_recovery", since = "1.9.0")] + pub fn take<Q: ?Sized>(&mut self, value: &Q) -> Option<T> + where + T: Borrow<Q> + Ord, + Q: Ord, + { + Recover::take(&mut self.map, value) + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all elements `e` for which `f(&e)` returns `false`. + /// The elements are visited in ascending order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut set = BTreeSet::from([1, 2, 3, 4, 5, 6]); + /// // Keep only the even numbers. + /// set.retain(|&k| k % 2 == 0); + /// assert!(set.iter().eq([2, 4, 6].iter())); + /// ``` + #[stable(feature = "btree_retain", since = "1.53.0")] + pub fn retain<F>(&mut self, mut f: F) + where + T: Ord, + F: FnMut(&T) -> bool, + { + self.drain_filter(|v| !f(v)); + } + + /// Moves all elements from `other` into `self`, leaving `other` empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// a.insert(2); + /// a.insert(3); + /// + /// let mut b = BTreeSet::new(); + /// b.insert(3); + /// b.insert(4); + /// b.insert(5); + /// + /// a.append(&mut b); + /// + /// assert_eq!(a.len(), 5); + /// assert_eq!(b.len(), 0); + /// + /// assert!(a.contains(&1)); + /// assert!(a.contains(&2)); + /// assert!(a.contains(&3)); + /// assert!(a.contains(&4)); + /// assert!(a.contains(&5)); + /// ``` + #[stable(feature = "btree_append", since = "1.11.0")] + pub fn append(&mut self, other: &mut Self) + where + T: Ord, + A: Clone, + { + self.map.append(&mut other.map); + } + + /// Splits the collection into two at the value. Returns a new collection + /// with all elements greater than or equal to the value. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut a = BTreeSet::new(); + /// a.insert(1); + /// a.insert(2); + /// a.insert(3); + /// a.insert(17); + /// a.insert(41); + /// + /// let b = a.split_off(&3); + /// + /// assert_eq!(a.len(), 2); + /// assert_eq!(b.len(), 3); + /// + /// assert!(a.contains(&1)); + /// assert!(a.contains(&2)); + /// + /// assert!(b.contains(&3)); + /// assert!(b.contains(&17)); + /// assert!(b.contains(&41)); + /// ``` + #[stable(feature = "btree_split_off", since = "1.11.0")] + pub fn split_off<Q: ?Sized + Ord>(&mut self, value: &Q) -> Self + where + T: Borrow<Q> + Ord, + A: Clone, + { + BTreeSet { map: self.map.split_off(value) } + } + + /// Creates an iterator that visits all elements in ascending order and + /// uses a closure to determine if an element should be removed. + /// + /// If the closure returns `true`, the element is removed from the set and + /// yielded. If the closure returns `false`, or panics, the element remains + /// in the set and will not be yielded. + /// + /// If the iterator is only partially consumed or not consumed at all, each + /// of the remaining elements is still subjected to the closure and removed + /// and dropped if it returns `true`. + /// + /// It is unspecified how many more elements will be subjected to the + /// closure if a panic occurs in the closure, or if a panic occurs while + /// dropping an element, or if the `DrainFilter` itself is leaked. + /// + /// # Examples + /// + /// Splitting a set into even and odd values, reusing the original set: + /// + /// ``` + /// #![feature(btree_drain_filter)] + /// use std::collections::BTreeSet; + /// + /// let mut set: BTreeSet<i32> = (0..8).collect(); + /// let evens: BTreeSet<_> = set.drain_filter(|v| v % 2 == 0).collect(); + /// let odds = set; + /// assert_eq!(evens.into_iter().collect::<Vec<_>>(), vec![0, 2, 4, 6]); + /// assert_eq!(odds.into_iter().collect::<Vec<_>>(), vec![1, 3, 5, 7]); + /// ``` + #[unstable(feature = "btree_drain_filter", issue = "70530")] + pub fn drain_filter<'a, F>(&'a mut self, pred: F) -> DrainFilter<'a, T, F, A> + where + T: Ord, + F: 'a + FnMut(&T) -> bool, + { + let (inner, alloc) = self.map.drain_filter_inner(); + DrainFilter { pred, inner, alloc } + } + + /// Gets an iterator that visits the elements in the `BTreeSet` in ascending + /// order. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set = BTreeSet::from([1, 2, 3]); + /// let mut set_iter = set.iter(); + /// assert_eq!(set_iter.next(), Some(&1)); + /// assert_eq!(set_iter.next(), Some(&2)); + /// assert_eq!(set_iter.next(), Some(&3)); + /// assert_eq!(set_iter.next(), None); + /// ``` + /// + /// Values returned by the iterator are returned in ascending order: + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set = BTreeSet::from([3, 1, 2]); + /// let mut set_iter = set.iter(); + /// assert_eq!(set_iter.next(), Some(&1)); + /// assert_eq!(set_iter.next(), Some(&2)); + /// assert_eq!(set_iter.next(), Some(&3)); + /// assert_eq!(set_iter.next(), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter(&self) -> Iter<'_, T> { + Iter { iter: self.map.keys() } + } + + /// Returns the number of elements in the set. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut v = BTreeSet::new(); + /// assert_eq!(v.len(), 0); + /// v.insert(1); + /// assert_eq!(v.len(), 1); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + pub const fn len(&self) -> usize { + self.map.len() + } + + /// Returns `true` if the set contains no elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let mut v = BTreeSet::new(); + /// assert!(v.is_empty()); + /// v.insert(1); + /// assert!(!v.is_empty()); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + #[rustc_const_unstable(feature = "const_btree_new", issue = "71835")] + pub const fn is_empty(&self) -> bool { + self.len() == 0 + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord> FromIterator<T> for BTreeSet<T> { + fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> BTreeSet<T> { + let mut inputs: Vec<_> = iter.into_iter().collect(); + + if inputs.is_empty() { + return BTreeSet::new(); + } + + // use stable sort to preserve the insertion order. + inputs.sort(); + BTreeSet::from_sorted_iter(inputs.into_iter(), Global) + } +} + +impl<T: Ord, A: Allocator + Clone> BTreeSet<T, A> { + fn from_sorted_iter<I: Iterator<Item = T>>(iter: I, alloc: A) -> BTreeSet<T, A> { + let iter = iter.map(|k| (k, SetValZST::default())); + let map = BTreeMap::bulk_build_from_sorted_iter(iter, alloc); + BTreeSet { map } + } +} + +#[stable(feature = "std_collections_from_array", since = "1.56.0")] +impl<T: Ord, const N: usize> From<[T; N]> for BTreeSet<T> { + /// Converts a `[T; N]` into a `BTreeSet<T>`. + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set1 = BTreeSet::from([1, 2, 3, 4]); + /// let set2: BTreeSet<_> = [1, 2, 3, 4].into(); + /// assert_eq!(set1, set2); + /// ``` + fn from(mut arr: [T; N]) -> Self { + if N == 0 { + return BTreeSet::new(); + } + + // use stable sort to preserve the insertion order. + arr.sort(); + let iter = IntoIterator::into_iter(arr).map(|k| (k, SetValZST::default())); + let map = BTreeMap::bulk_build_from_sorted_iter(iter, Global); + BTreeSet { map } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> IntoIterator for BTreeSet<T, A> { + type Item = T; + type IntoIter = IntoIter<T, A>; + + /// Gets an iterator for moving out the `BTreeSet`'s contents. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let set = BTreeSet::from([1, 2, 3, 4]); + /// + /// let v: Vec<_> = set.into_iter().collect(); + /// assert_eq!(v, [1, 2, 3, 4]); + /// ``` + fn into_iter(self) -> IntoIter<T, A> { + IntoIter { iter: self.map.into_iter() } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T, A: Allocator + Clone> IntoIterator for &'a BTreeSet<T, A> { + type Item = &'a T; + type IntoIter = Iter<'a, T>; + + fn into_iter(self) -> Iter<'a, T> { + self.iter() + } +} + +/// An iterator produced by calling `drain_filter` on BTreeSet. +#[unstable(feature = "btree_drain_filter", issue = "70530")] +pub struct DrainFilter< + 'a, + T, + F, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + Clone = Global, +> where + T: 'a, + F: 'a + FnMut(&T) -> bool, +{ + pred: F, + inner: super::map::DrainFilterInner<'a, T, SetValZST>, + /// The BTreeMap will outlive this IntoIter so we don't care about drop order for `alloc`. + alloc: A, +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<T, F, A: Allocator + Clone> Drop for DrainFilter<'_, T, F, A> +where + F: FnMut(&T) -> bool, +{ + fn drop(&mut self) { + self.for_each(drop); + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<T, F, A: Allocator + Clone> fmt::Debug for DrainFilter<'_, T, F, A> +where + T: fmt::Debug, + F: FnMut(&T) -> bool, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("DrainFilter").field(&self.inner.peek().map(|(k, _)| k)).finish() + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<'a, T, F, A: Allocator + Clone> Iterator for DrainFilter<'_, T, F, A> +where + F: 'a + FnMut(&T) -> bool, +{ + type Item = T; + + fn next(&mut self) -> Option<T> { + let pred = &mut self.pred; + let mut mapped_pred = |k: &T, _v: &mut SetValZST| pred(k); + self.inner.next(&mut mapped_pred, self.alloc.clone()).map(|(k, _)| k) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.inner.size_hint() + } +} + +#[unstable(feature = "btree_drain_filter", issue = "70530")] +impl<T, F, A: Allocator + Clone> FusedIterator for DrainFilter<'_, T, F, A> where + F: FnMut(&T) -> bool +{ +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord, A: Allocator + Clone> Extend<T> for BTreeSet<T, A> { + #[inline] + fn extend<Iter: IntoIterator<Item = T>>(&mut self, iter: Iter) { + iter.into_iter().for_each(move |elem| { + self.insert(elem); + }); + } + + #[inline] + fn extend_one(&mut self, elem: T) { + self.insert(elem); + } +} + +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, T: 'a + Ord + Copy, A: Allocator + Clone> Extend<&'a T> for BTreeSet<T, A> { + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iter: I) { + self.extend(iter.into_iter().cloned()); + } + + #[inline] + fn extend_one(&mut self, &elem: &'a T) { + self.insert(elem); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Default for BTreeSet<T> { + /// Creates an empty `BTreeSet`. + fn default() -> BTreeSet<T> { + BTreeSet::new() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord + Clone, A: Allocator + Clone> Sub<&BTreeSet<T, A>> for &BTreeSet<T, A> { + type Output = BTreeSet<T, A>; + + /// Returns the difference of `self` and `rhs` as a new `BTreeSet<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let a = BTreeSet::from([1, 2, 3]); + /// let b = BTreeSet::from([3, 4, 5]); + /// + /// let result = &a - &b; + /// assert_eq!(result, BTreeSet::from([1, 2])); + /// ``` + fn sub(self, rhs: &BTreeSet<T, A>) -> BTreeSet<T, A> { + BTreeSet::from_sorted_iter( + self.difference(rhs).cloned(), + ManuallyDrop::into_inner(self.map.alloc.clone()), + ) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord + Clone, A: Allocator + Clone> BitXor<&BTreeSet<T, A>> for &BTreeSet<T, A> { + type Output = BTreeSet<T, A>; + + /// Returns the symmetric difference of `self` and `rhs` as a new `BTreeSet<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let a = BTreeSet::from([1, 2, 3]); + /// let b = BTreeSet::from([2, 3, 4]); + /// + /// let result = &a ^ &b; + /// assert_eq!(result, BTreeSet::from([1, 4])); + /// ``` + fn bitxor(self, rhs: &BTreeSet<T, A>) -> BTreeSet<T, A> { + BTreeSet::from_sorted_iter( + self.symmetric_difference(rhs).cloned(), + ManuallyDrop::into_inner(self.map.alloc.clone()), + ) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord + Clone, A: Allocator + Clone> BitAnd<&BTreeSet<T, A>> for &BTreeSet<T, A> { + type Output = BTreeSet<T, A>; + + /// Returns the intersection of `self` and `rhs` as a new `BTreeSet<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let a = BTreeSet::from([1, 2, 3]); + /// let b = BTreeSet::from([2, 3, 4]); + /// + /// let result = &a & &b; + /// assert_eq!(result, BTreeSet::from([2, 3])); + /// ``` + fn bitand(self, rhs: &BTreeSet<T, A>) -> BTreeSet<T, A> { + BTreeSet::from_sorted_iter( + self.intersection(rhs).cloned(), + ManuallyDrop::into_inner(self.map.alloc.clone()), + ) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord + Clone, A: Allocator + Clone> BitOr<&BTreeSet<T, A>> for &BTreeSet<T, A> { + type Output = BTreeSet<T, A>; + + /// Returns the union of `self` and `rhs` as a new `BTreeSet<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::BTreeSet; + /// + /// let a = BTreeSet::from([1, 2, 3]); + /// let b = BTreeSet::from([3, 4, 5]); + /// + /// let result = &a | &b; + /// assert_eq!(result, BTreeSet::from([1, 2, 3, 4, 5])); + /// ``` + fn bitor(self, rhs: &BTreeSet<T, A>) -> BTreeSet<T, A> { + BTreeSet::from_sorted_iter( + self.union(rhs).cloned(), + ManuallyDrop::into_inner(self.map.alloc.clone()), + ) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Debug, A: Allocator + Clone> Debug for BTreeSet<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_set().entries(self.iter()).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for Iter<'_, T> { + fn clone(&self) -> Self { + Iter { iter: self.iter.clone() } + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for Iter<'a, T> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + self.iter.next() + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + fn last(mut self) -> Option<&'a T> { + self.next_back() + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } + + fn max(mut self) -> Option<&'a T> { + self.next_back() + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for Iter<'a, T> { + fn next_back(&mut self) -> Option<&'a T> { + self.iter.next_back() + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for Iter<'_, T> { + fn len(&self) -> usize { + self.iter.len() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Iter<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> Iterator for IntoIter<T, A> { + type Item = T; + + fn next(&mut self) -> Option<T> { + self.iter.next().map(|(k, _)| k) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> DoubleEndedIterator for IntoIter<T, A> { + fn next_back(&mut self) -> Option<T> { + self.iter.next_back().map(|(k, _)| k) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> ExactSizeIterator for IntoIter<T, A> { + fn len(&self) -> usize { + self.iter.len() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T, A: Allocator + Clone> FusedIterator for IntoIter<T, A> {} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<T> Clone for Range<'_, T> { + fn clone(&self) -> Self { + Range { iter: self.iter.clone() } + } +} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, T> Iterator for Range<'a, T> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + self.iter.next().map(|(k, _)| k) + } + + fn last(mut self) -> Option<&'a T> { + self.next_back() + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } + + fn max(mut self) -> Option<&'a T> { + self.next_back() + } +} + +#[stable(feature = "btree_range", since = "1.17.0")] +impl<'a, T> DoubleEndedIterator for Range<'a, T> { + fn next_back(&mut self) -> Option<&'a T> { + self.iter.next_back().map(|(k, _)| k) + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Range<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> Clone for Difference<'_, T, A> { + fn clone(&self) -> Self { + Difference { + inner: match &self.inner { + DifferenceInner::Stitch { self_iter, other_iter } => DifferenceInner::Stitch { + self_iter: self_iter.clone(), + other_iter: other_iter.clone(), + }, + DifferenceInner::Search { self_iter, other_set } => { + DifferenceInner::Search { self_iter: self_iter.clone(), other_set } + } + DifferenceInner::Iterate(iter) => DifferenceInner::Iterate(iter.clone()), + }, + } + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T: Ord, A: Allocator + Clone> Iterator for Difference<'a, T, A> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + match &mut self.inner { + DifferenceInner::Stitch { self_iter, other_iter } => { + let mut self_next = self_iter.next()?; + loop { + match other_iter.peek().map_or(Less, |other_next| self_next.cmp(other_next)) { + Less => return Some(self_next), + Equal => { + self_next = self_iter.next()?; + other_iter.next(); + } + Greater => { + other_iter.next(); + } + } + } + } + DifferenceInner::Search { self_iter, other_set } => loop { + let self_next = self_iter.next()?; + if !other_set.contains(&self_next) { + return Some(self_next); + } + }, + DifferenceInner::Iterate(iter) => iter.next(), + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + let (self_len, other_len) = match &self.inner { + DifferenceInner::Stitch { self_iter, other_iter } => { + (self_iter.len(), other_iter.len()) + } + DifferenceInner::Search { self_iter, other_set } => (self_iter.len(), other_set.len()), + DifferenceInner::Iterate(iter) => (iter.len(), 0), + }; + (self_len.saturating_sub(other_len), Some(self_len)) + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T: Ord, A: Allocator + Clone> FusedIterator for Difference<'_, T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for SymmetricDifference<'_, T> { + fn clone(&self) -> Self { + SymmetricDifference(self.0.clone()) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T: Ord> Iterator for SymmetricDifference<'a, T> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + loop { + let (a_next, b_next) = self.0.nexts(Self::Item::cmp); + if a_next.and(b_next).is_none() { + return a_next.or(b_next); + } + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + let (a_len, b_len) = self.0.lens(); + // No checked_add, because even if a and b refer to the same set, + // and T is a zero-sized type, the storage overhead of sets limits + // the number of elements to less than half the range of usize. + (0, Some(a_len + b_len)) + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T: Ord> FusedIterator for SymmetricDifference<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator + Clone> Clone for Intersection<'_, T, A> { + fn clone(&self) -> Self { + Intersection { + inner: match &self.inner { + IntersectionInner::Stitch { a, b } => { + IntersectionInner::Stitch { a: a.clone(), b: b.clone() } + } + IntersectionInner::Search { small_iter, large_set } => { + IntersectionInner::Search { small_iter: small_iter.clone(), large_set } + } + IntersectionInner::Answer(answer) => IntersectionInner::Answer(*answer), + }, + } + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T: Ord, A: Allocator + Clone> Iterator for Intersection<'a, T, A> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + match &mut self.inner { + IntersectionInner::Stitch { a, b } => { + let mut a_next = a.next()?; + let mut b_next = b.next()?; + loop { + match a_next.cmp(b_next) { + Less => a_next = a.next()?, + Greater => b_next = b.next()?, + Equal => return Some(a_next), + } + } + } + IntersectionInner::Search { small_iter, large_set } => loop { + let small_next = small_iter.next()?; + if large_set.contains(&small_next) { + return Some(small_next); + } + }, + IntersectionInner::Answer(answer) => answer.take(), + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + match &self.inner { + IntersectionInner::Stitch { a, b } => (0, Some(min(a.len(), b.len()))), + IntersectionInner::Search { small_iter, .. } => (0, Some(small_iter.len())), + IntersectionInner::Answer(None) => (0, Some(0)), + IntersectionInner::Answer(Some(_)) => (1, Some(1)), + } + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T: Ord, A: Allocator + Clone> FusedIterator for Intersection<'_, T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for Union<'_, T> { + fn clone(&self) -> Self { + Union(self.0.clone()) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T: Ord> Iterator for Union<'a, T> { + type Item = &'a T; + + fn next(&mut self) -> Option<&'a T> { + let (a_next, b_next) = self.0.nexts(Self::Item::cmp); + a_next.or(b_next) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + let (a_len, b_len) = self.0.lens(); + // No checked_add - see SymmetricDifference::size_hint. + (max(a_len, b_len), Some(a_len + b_len)) + } + + fn min(mut self) -> Option<&'a T> { + self.next() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T: Ord> FusedIterator for Union<'_, T> {} + +#[cfg(test)] +mod tests; diff --git a/library/alloc/src/collections/btree/set/tests.rs b/library/alloc/src/collections/btree/set/tests.rs new file mode 100644 index 000000000..502d3e1d1 --- /dev/null +++ b/library/alloc/src/collections/btree/set/tests.rs @@ -0,0 +1,856 @@ +use super::super::testing::crash_test::{CrashTestDummy, Panic}; +use super::super::testing::rng::DeterministicRng; +use super::*; +use crate::vec::Vec; +use std::cmp::Ordering; +use std::hash::{Hash, Hasher}; +use std::iter::FromIterator; +use std::ops::Bound::{Excluded, Included}; +use std::panic::{catch_unwind, AssertUnwindSafe}; + +#[test] +fn test_clone_eq() { + let mut m = BTreeSet::new(); + + m.insert(1); + m.insert(2); + + assert_eq!(m.clone(), m); +} + +#[test] +fn test_iter_min_max() { + let mut a = BTreeSet::new(); + assert_eq!(a.iter().min(), None); + assert_eq!(a.iter().max(), None); + assert_eq!(a.range(..).min(), None); + assert_eq!(a.range(..).max(), None); + assert_eq!(a.difference(&BTreeSet::new()).min(), None); + assert_eq!(a.difference(&BTreeSet::new()).max(), None); + assert_eq!(a.intersection(&a).min(), None); + assert_eq!(a.intersection(&a).max(), None); + assert_eq!(a.symmetric_difference(&BTreeSet::new()).min(), None); + assert_eq!(a.symmetric_difference(&BTreeSet::new()).max(), None); + assert_eq!(a.union(&a).min(), None); + assert_eq!(a.union(&a).max(), None); + a.insert(1); + a.insert(2); + assert_eq!(a.iter().min(), Some(&1)); + assert_eq!(a.iter().max(), Some(&2)); + assert_eq!(a.range(..).min(), Some(&1)); + assert_eq!(a.range(..).max(), Some(&2)); + assert_eq!(a.difference(&BTreeSet::new()).min(), Some(&1)); + assert_eq!(a.difference(&BTreeSet::new()).max(), Some(&2)); + assert_eq!(a.intersection(&a).min(), Some(&1)); + assert_eq!(a.intersection(&a).max(), Some(&2)); + assert_eq!(a.symmetric_difference(&BTreeSet::new()).min(), Some(&1)); + assert_eq!(a.symmetric_difference(&BTreeSet::new()).max(), Some(&2)); + assert_eq!(a.union(&a).min(), Some(&1)); + assert_eq!(a.union(&a).max(), Some(&2)); +} + +fn check<F>(a: &[i32], b: &[i32], expected: &[i32], f: F) +where + F: FnOnce(&BTreeSet<i32>, &BTreeSet<i32>, &mut dyn FnMut(&i32) -> bool) -> bool, +{ + let mut set_a = BTreeSet::new(); + let mut set_b = BTreeSet::new(); + + for x in a { + assert!(set_a.insert(*x)) + } + for y in b { + assert!(set_b.insert(*y)) + } + + let mut i = 0; + f(&set_a, &set_b, &mut |&x| { + if i < expected.len() { + assert_eq!(x, expected[i]); + } + i += 1; + true + }); + assert_eq!(i, expected.len()); +} + +#[test] +fn test_intersection() { + fn check_intersection(a: &[i32], b: &[i32], expected: &[i32]) { + check(a, b, expected, |x, y, f| x.intersection(y).all(f)) + } + + check_intersection(&[], &[], &[]); + check_intersection(&[1, 2, 3], &[], &[]); + check_intersection(&[], &[1, 2, 3], &[]); + check_intersection(&[2], &[1, 2, 3], &[2]); + check_intersection(&[1, 2, 3], &[2], &[2]); + check_intersection(&[11, 1, 3, 77, 103, 5, -5], &[2, 11, 77, -9, -42, 5, 3], &[3, 5, 11, 77]); + + if cfg!(miri) { + // Miri is too slow + return; + } + + let large = Vec::from_iter(0..100); + check_intersection(&[], &large, &[]); + check_intersection(&large, &[], &[]); + check_intersection(&[-1], &large, &[]); + check_intersection(&large, &[-1], &[]); + check_intersection(&[0], &large, &[0]); + check_intersection(&large, &[0], &[0]); + check_intersection(&[99], &large, &[99]); + check_intersection(&large, &[99], &[99]); + check_intersection(&[100], &large, &[]); + check_intersection(&large, &[100], &[]); + check_intersection(&[11, 5000, 1, 3, 77, 8924], &large, &[1, 3, 11, 77]); +} + +#[test] +fn test_intersection_size_hint() { + let x = BTreeSet::from([3, 4]); + let y = BTreeSet::from([1, 2, 3]); + let mut iter = x.intersection(&y); + assert_eq!(iter.size_hint(), (1, Some(1))); + assert_eq!(iter.next(), Some(&3)); + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + + iter = y.intersection(&y); + assert_eq!(iter.size_hint(), (0, Some(3))); + assert_eq!(iter.next(), Some(&1)); + assert_eq!(iter.size_hint(), (0, Some(2))); +} + +#[test] +fn test_difference() { + fn check_difference(a: &[i32], b: &[i32], expected: &[i32]) { + check(a, b, expected, |x, y, f| x.difference(y).all(f)) + } + + check_difference(&[], &[], &[]); + check_difference(&[1, 12], &[], &[1, 12]); + check_difference(&[], &[1, 2, 3, 9], &[]); + check_difference(&[1, 3, 5, 9, 11], &[3, 9], &[1, 5, 11]); + check_difference(&[1, 3, 5, 9, 11], &[3, 6, 9], &[1, 5, 11]); + check_difference(&[1, 3, 5, 9, 11], &[0, 1], &[3, 5, 9, 11]); + check_difference(&[1, 3, 5, 9, 11], &[11, 12], &[1, 3, 5, 9]); + check_difference( + &[-5, 11, 22, 33, 40, 42], + &[-12, -5, 14, 23, 34, 38, 39, 50], + &[11, 22, 33, 40, 42], + ); + + if cfg!(miri) { + // Miri is too slow + return; + } + + let large = Vec::from_iter(0..100); + check_difference(&[], &large, &[]); + check_difference(&[-1], &large, &[-1]); + check_difference(&[0], &large, &[]); + check_difference(&[99], &large, &[]); + check_difference(&[100], &large, &[100]); + check_difference(&[11, 5000, 1, 3, 77, 8924], &large, &[5000, 8924]); + check_difference(&large, &[], &large); + check_difference(&large, &[-1], &large); + check_difference(&large, &[100], &large); +} + +#[test] +fn test_difference_size_hint() { + let s246 = BTreeSet::from([2, 4, 6]); + let s23456 = BTreeSet::from_iter(2..=6); + let mut iter = s246.difference(&s23456); + assert_eq!(iter.size_hint(), (0, Some(3))); + assert_eq!(iter.next(), None); + + let s12345 = BTreeSet::from_iter(1..=5); + iter = s246.difference(&s12345); + assert_eq!(iter.size_hint(), (0, Some(3))); + assert_eq!(iter.next(), Some(&6)); + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + + let s34567 = BTreeSet::from_iter(3..=7); + iter = s246.difference(&s34567); + assert_eq!(iter.size_hint(), (0, Some(3))); + assert_eq!(iter.next(), Some(&2)); + assert_eq!(iter.size_hint(), (0, Some(2))); + assert_eq!(iter.next(), None); + + let s1 = BTreeSet::from_iter(-9..=1); + iter = s246.difference(&s1); + assert_eq!(iter.size_hint(), (3, Some(3))); + + let s2 = BTreeSet::from_iter(-9..=2); + iter = s246.difference(&s2); + assert_eq!(iter.size_hint(), (2, Some(2))); + assert_eq!(iter.next(), Some(&4)); + assert_eq!(iter.size_hint(), (1, Some(1))); + + let s23 = BTreeSet::from([2, 3]); + iter = s246.difference(&s23); + assert_eq!(iter.size_hint(), (1, Some(3))); + assert_eq!(iter.next(), Some(&4)); + assert_eq!(iter.size_hint(), (1, Some(1))); + + let s4 = BTreeSet::from([4]); + iter = s246.difference(&s4); + assert_eq!(iter.size_hint(), (2, Some(3))); + assert_eq!(iter.next(), Some(&2)); + assert_eq!(iter.size_hint(), (1, Some(2))); + assert_eq!(iter.next(), Some(&6)); + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + + let s56 = BTreeSet::from([5, 6]); + iter = s246.difference(&s56); + assert_eq!(iter.size_hint(), (1, Some(3))); + assert_eq!(iter.next(), Some(&2)); + assert_eq!(iter.size_hint(), (0, Some(2))); + + let s6 = BTreeSet::from_iter(6..=19); + iter = s246.difference(&s6); + assert_eq!(iter.size_hint(), (2, Some(2))); + assert_eq!(iter.next(), Some(&2)); + assert_eq!(iter.size_hint(), (1, Some(1))); + + let s7 = BTreeSet::from_iter(7..=19); + iter = s246.difference(&s7); + assert_eq!(iter.size_hint(), (3, Some(3))); +} + +#[test] +fn test_symmetric_difference() { + fn check_symmetric_difference(a: &[i32], b: &[i32], expected: &[i32]) { + check(a, b, expected, |x, y, f| x.symmetric_difference(y).all(f)) + } + + check_symmetric_difference(&[], &[], &[]); + check_symmetric_difference(&[1, 2, 3], &[2], &[1, 3]); + check_symmetric_difference(&[2], &[1, 2, 3], &[1, 3]); + check_symmetric_difference(&[1, 3, 5, 9, 11], &[-2, 3, 9, 14, 22], &[-2, 1, 5, 11, 14, 22]); +} + +#[test] +fn test_symmetric_difference_size_hint() { + let x = BTreeSet::from([2, 4]); + let y = BTreeSet::from([1, 2, 3]); + let mut iter = x.symmetric_difference(&y); + assert_eq!(iter.size_hint(), (0, Some(5))); + assert_eq!(iter.next(), Some(&1)); + assert_eq!(iter.size_hint(), (0, Some(4))); + assert_eq!(iter.next(), Some(&3)); + assert_eq!(iter.size_hint(), (0, Some(1))); +} + +#[test] +fn test_union() { + fn check_union(a: &[i32], b: &[i32], expected: &[i32]) { + check(a, b, expected, |x, y, f| x.union(y).all(f)) + } + + check_union(&[], &[], &[]); + check_union(&[1, 2, 3], &[2], &[1, 2, 3]); + check_union(&[2], &[1, 2, 3], &[1, 2, 3]); + check_union( + &[1, 3, 5, 9, 11, 16, 19, 24], + &[-2, 1, 5, 9, 13, 19], + &[-2, 1, 3, 5, 9, 11, 13, 16, 19, 24], + ); +} + +#[test] +fn test_union_size_hint() { + let x = BTreeSet::from([2, 4]); + let y = BTreeSet::from([1, 2, 3]); + let mut iter = x.union(&y); + assert_eq!(iter.size_hint(), (3, Some(5))); + assert_eq!(iter.next(), Some(&1)); + assert_eq!(iter.size_hint(), (2, Some(4))); + assert_eq!(iter.next(), Some(&2)); + assert_eq!(iter.size_hint(), (1, Some(2))); +} + +#[test] +// Only tests the simple function definition with respect to intersection +fn test_is_disjoint() { + let one = BTreeSet::from([1]); + let two = BTreeSet::from([2]); + assert!(one.is_disjoint(&two)); +} + +#[test] +// Also implicitly tests the trivial function definition of is_superset +fn test_is_subset() { + fn is_subset(a: &[i32], b: &[i32]) -> bool { + let set_a = BTreeSet::from_iter(a.iter()); + let set_b = BTreeSet::from_iter(b.iter()); + set_a.is_subset(&set_b) + } + + assert_eq!(is_subset(&[], &[]), true); + assert_eq!(is_subset(&[], &[1, 2]), true); + assert_eq!(is_subset(&[0], &[1, 2]), false); + assert_eq!(is_subset(&[1], &[1, 2]), true); + assert_eq!(is_subset(&[2], &[1, 2]), true); + assert_eq!(is_subset(&[3], &[1, 2]), false); + assert_eq!(is_subset(&[1, 2], &[1]), false); + assert_eq!(is_subset(&[1, 2], &[1, 2]), true); + assert_eq!(is_subset(&[1, 2], &[2, 3]), false); + assert_eq!( + is_subset(&[-5, 11, 22, 33, 40, 42], &[-12, -5, 11, 14, 22, 23, 33, 34, 38, 39, 40, 42]), + true + ); + assert_eq!(is_subset(&[-5, 11, 22, 33, 40, 42], &[-12, -5, 11, 14, 22, 23, 34, 38]), false); + + if cfg!(miri) { + // Miri is too slow + return; + } + + let large = Vec::from_iter(0..100); + assert_eq!(is_subset(&[], &large), true); + assert_eq!(is_subset(&large, &[]), false); + assert_eq!(is_subset(&[-1], &large), false); + assert_eq!(is_subset(&[0], &large), true); + assert_eq!(is_subset(&[1, 2], &large), true); + assert_eq!(is_subset(&[99, 100], &large), false); +} + +#[test] +fn test_is_superset() { + fn is_superset(a: &[i32], b: &[i32]) -> bool { + let set_a = BTreeSet::from_iter(a.iter()); + let set_b = BTreeSet::from_iter(b.iter()); + set_a.is_superset(&set_b) + } + + assert_eq!(is_superset(&[], &[]), true); + assert_eq!(is_superset(&[], &[1, 2]), false); + assert_eq!(is_superset(&[0], &[1, 2]), false); + assert_eq!(is_superset(&[1], &[1, 2]), false); + assert_eq!(is_superset(&[4], &[1, 2]), false); + assert_eq!(is_superset(&[1, 4], &[1, 2]), false); + assert_eq!(is_superset(&[1, 2], &[1, 2]), true); + assert_eq!(is_superset(&[1, 2, 3], &[1, 3]), true); + assert_eq!(is_superset(&[1, 2, 3], &[]), true); + assert_eq!(is_superset(&[-1, 1, 2, 3], &[-1, 3]), true); + + if cfg!(miri) { + // Miri is too slow + return; + } + + let large = Vec::from_iter(0..100); + assert_eq!(is_superset(&[], &large), false); + assert_eq!(is_superset(&large, &[]), true); + assert_eq!(is_superset(&large, &[1]), true); + assert_eq!(is_superset(&large, &[50, 99]), true); + assert_eq!(is_superset(&large, &[100]), false); + assert_eq!(is_superset(&large, &[0, 99]), true); + assert_eq!(is_superset(&[-1], &large), false); + assert_eq!(is_superset(&[0], &large), false); + assert_eq!(is_superset(&[99, 100], &large), false); +} + +#[test] +fn test_retain() { + let mut set = BTreeSet::from([1, 2, 3, 4, 5, 6]); + set.retain(|&k| k % 2 == 0); + assert_eq!(set.len(), 3); + assert!(set.contains(&2)); + assert!(set.contains(&4)); + assert!(set.contains(&6)); +} + +#[test] +fn test_drain_filter() { + let mut x = BTreeSet::from([1]); + let mut y = BTreeSet::from([1]); + + x.drain_filter(|_| true); + y.drain_filter(|_| false); + assert_eq!(x.len(), 0); + assert_eq!(y.len(), 1); +} + +#[test] +fn test_drain_filter_drop_panic_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut set = BTreeSet::new(); + set.insert(a.spawn(Panic::Never)); + set.insert(b.spawn(Panic::InDrop)); + set.insert(c.spawn(Panic::Never)); + + catch_unwind(move || drop(set.drain_filter(|dummy| dummy.query(true)))).ok(); + + assert_eq!(a.queried(), 1); + assert_eq!(b.queried(), 1); + assert_eq!(c.queried(), 0); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 1); + assert_eq!(c.dropped(), 1); +} + +#[test] +fn test_drain_filter_pred_panic_leak() { + let a = CrashTestDummy::new(0); + let b = CrashTestDummy::new(1); + let c = CrashTestDummy::new(2); + let mut set = BTreeSet::new(); + set.insert(a.spawn(Panic::Never)); + set.insert(b.spawn(Panic::InQuery)); + set.insert(c.spawn(Panic::InQuery)); + + catch_unwind(AssertUnwindSafe(|| drop(set.drain_filter(|dummy| dummy.query(true))))).ok(); + + assert_eq!(a.queried(), 1); + assert_eq!(b.queried(), 1); + assert_eq!(c.queried(), 0); + assert_eq!(a.dropped(), 1); + assert_eq!(b.dropped(), 0); + assert_eq!(c.dropped(), 0); + assert_eq!(set.len(), 2); + assert_eq!(set.first().unwrap().id(), 1); + assert_eq!(set.last().unwrap().id(), 2); +} + +#[test] +fn test_clear() { + let mut x = BTreeSet::new(); + x.insert(1); + + x.clear(); + assert!(x.is_empty()); +} +#[test] +fn test_remove() { + let mut x = BTreeSet::new(); + assert!(x.is_empty()); + + x.insert(1); + x.insert(2); + x.insert(3); + x.insert(4); + + assert_eq!(x.remove(&2), true); + assert_eq!(x.remove(&0), false); + assert_eq!(x.remove(&5), false); + assert_eq!(x.remove(&1), true); + assert_eq!(x.remove(&2), false); + assert_eq!(x.remove(&3), true); + assert_eq!(x.remove(&4), true); + assert_eq!(x.remove(&4), false); + assert!(x.is_empty()); +} + +#[test] +fn test_zip() { + let mut x = BTreeSet::new(); + x.insert(5); + x.insert(12); + x.insert(11); + + let mut y = BTreeSet::new(); + y.insert("foo"); + y.insert("bar"); + + let x = x; + let y = y; + let mut z = x.iter().zip(&y); + + assert_eq!(z.next().unwrap(), (&5, &("bar"))); + assert_eq!(z.next().unwrap(), (&11, &("foo"))); + assert!(z.next().is_none()); +} + +#[test] +fn test_from_iter() { + let xs = [1, 2, 3, 4, 5, 6, 7, 8, 9]; + + let set = BTreeSet::from_iter(xs.iter()); + + for x in &xs { + assert!(set.contains(x)); + } +} + +#[test] +fn test_show() { + let mut set = BTreeSet::new(); + let empty = BTreeSet::<i32>::new(); + + set.insert(1); + set.insert(2); + + let set_str = format!("{set:?}"); + + assert_eq!(set_str, "{1, 2}"); + assert_eq!(format!("{empty:?}"), "{}"); +} + +#[test] +fn test_extend_ref() { + let mut a = BTreeSet::new(); + a.insert(1); + + a.extend(&[2, 3, 4]); + + assert_eq!(a.len(), 4); + assert!(a.contains(&1)); + assert!(a.contains(&2)); + assert!(a.contains(&3)); + assert!(a.contains(&4)); + + let mut b = BTreeSet::new(); + b.insert(5); + b.insert(6); + + a.extend(&b); + + assert_eq!(a.len(), 6); + assert!(a.contains(&1)); + assert!(a.contains(&2)); + assert!(a.contains(&3)); + assert!(a.contains(&4)); + assert!(a.contains(&5)); + assert!(a.contains(&6)); +} + +#[test] +fn test_recovery() { + #[derive(Debug)] + struct Foo(&'static str, i32); + + impl PartialEq for Foo { + fn eq(&self, other: &Self) -> bool { + self.0 == other.0 + } + } + + impl Eq for Foo {} + + impl PartialOrd for Foo { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + self.0.partial_cmp(&other.0) + } + } + + impl Ord for Foo { + fn cmp(&self, other: &Self) -> Ordering { + self.0.cmp(&other.0) + } + } + + let mut s = BTreeSet::new(); + assert_eq!(s.replace(Foo("a", 1)), None); + assert_eq!(s.len(), 1); + assert_eq!(s.replace(Foo("a", 2)), Some(Foo("a", 1))); + assert_eq!(s.len(), 1); + + { + let mut it = s.iter(); + assert_eq!(it.next(), Some(&Foo("a", 2))); + assert_eq!(it.next(), None); + } + + assert_eq!(s.get(&Foo("a", 1)), Some(&Foo("a", 2))); + assert_eq!(s.take(&Foo("a", 1)), Some(Foo("a", 2))); + assert_eq!(s.len(), 0); + + assert_eq!(s.get(&Foo("a", 1)), None); + assert_eq!(s.take(&Foo("a", 1)), None); + + assert_eq!(s.iter().next(), None); +} + +#[allow(dead_code)] +fn assert_covariance() { + fn set<'new>(v: BTreeSet<&'static str>) -> BTreeSet<&'new str> { + v + } + fn iter<'a, 'new>(v: Iter<'a, &'static str>) -> Iter<'a, &'new str> { + v + } + fn into_iter<'new>(v: IntoIter<&'static str>) -> IntoIter<&'new str> { + v + } + fn range<'a, 'new>(v: Range<'a, &'static str>) -> Range<'a, &'new str> { + v + } + // not applied to Difference, Intersection, SymmetricDifference, Union +} + +#[allow(dead_code)] +fn assert_sync() { + fn set<T: Sync>(v: &BTreeSet<T>) -> impl Sync + '_ { + v + } + + fn iter<T: Sync>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.iter() + } + + fn into_iter<T: Sync>(v: BTreeSet<T>) -> impl Sync { + v.into_iter() + } + + fn range<T: Sync + Ord>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.range(..) + } + + fn drain_filter<T: Sync + Ord>(v: &mut BTreeSet<T>) -> impl Sync + '_ { + v.drain_filter(|_| false) + } + + fn difference<T: Sync + Ord>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.difference(&v) + } + + fn intersection<T: Sync + Ord>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.intersection(&v) + } + + fn symmetric_difference<T: Sync + Ord>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.symmetric_difference(&v) + } + + fn union<T: Sync + Ord>(v: &BTreeSet<T>) -> impl Sync + '_ { + v.union(&v) + } +} + +#[allow(dead_code)] +fn assert_send() { + fn set<T: Send>(v: BTreeSet<T>) -> impl Send { + v + } + + fn iter<T: Send + Sync>(v: &BTreeSet<T>) -> impl Send + '_ { + v.iter() + } + + fn into_iter<T: Send>(v: BTreeSet<T>) -> impl Send { + v.into_iter() + } + + fn range<T: Send + Sync + Ord>(v: &BTreeSet<T>) -> impl Send + '_ { + v.range(..) + } + + fn drain_filter<T: Send + Ord>(v: &mut BTreeSet<T>) -> impl Send + '_ { + v.drain_filter(|_| false) + } + + fn difference<T: Send + Sync + Ord>(v: &BTreeSet<T>) -> impl Send + '_ { + v.difference(&v) + } + + fn intersection<T: Send + Sync + Ord>(v: &BTreeSet<T>) -> impl Send + '_ { + v.intersection(&v) + } + + fn symmetric_difference<T: Send + Sync + Ord>(v: &BTreeSet<T>) -> impl Send + '_ { + v.symmetric_difference(&v) + } + + fn union<T: Send + Sync + Ord>(v: &BTreeSet<T>) -> impl Send + '_ { + v.union(&v) + } +} + +#[allow(dead_code)] +// Check that the member-like functions conditionally provided by #[derive()] +// are not overridden by genuine member functions with a different signature. +fn assert_derives() { + fn hash<T: Hash, H: Hasher>(v: BTreeSet<T>, state: &mut H) { + v.hash(state); + // Tested much more thoroughly outside the crate in btree_set_hash.rs + } + fn eq<T: PartialEq>(v: BTreeSet<T>) { + let _ = v.eq(&v); + } + fn ne<T: PartialEq>(v: BTreeSet<T>) { + let _ = v.ne(&v); + } + fn cmp<T: Ord>(v: BTreeSet<T>) { + let _ = v.cmp(&v); + } + fn min<T: Ord>(v: BTreeSet<T>, w: BTreeSet<T>) { + let _ = v.min(w); + } + fn max<T: Ord>(v: BTreeSet<T>, w: BTreeSet<T>) { + let _ = v.max(w); + } + fn clamp<T: Ord>(v: BTreeSet<T>, w: BTreeSet<T>, x: BTreeSet<T>) { + let _ = v.clamp(w, x); + } + fn partial_cmp<T: PartialOrd>(v: &BTreeSet<T>) { + let _ = v.partial_cmp(&v); + } +} + +#[test] +fn test_ord_absence() { + fn set<K>(mut set: BTreeSet<K>) { + let _ = set.is_empty(); + let _ = set.len(); + set.clear(); + let _ = set.iter(); + let _ = set.into_iter(); + } + + fn set_debug<K: Debug>(set: BTreeSet<K>) { + format!("{set:?}"); + format!("{:?}", set.iter()); + format!("{:?}", set.into_iter()); + } + + fn set_clone<K: Clone>(mut set: BTreeSet<K>) { + set.clone_from(&set.clone()); + } + + #[derive(Debug, Clone)] + struct NonOrd; + set(BTreeSet::<NonOrd>::new()); + set_debug(BTreeSet::<NonOrd>::new()); + set_clone(BTreeSet::<NonOrd>::default()); +} + +#[test] +fn test_append() { + let mut a = BTreeSet::new(); + a.insert(1); + a.insert(2); + a.insert(3); + + let mut b = BTreeSet::new(); + b.insert(3); + b.insert(4); + b.insert(5); + + a.append(&mut b); + + assert_eq!(a.len(), 5); + assert_eq!(b.len(), 0); + + assert_eq!(a.contains(&1), true); + assert_eq!(a.contains(&2), true); + assert_eq!(a.contains(&3), true); + assert_eq!(a.contains(&4), true); + assert_eq!(a.contains(&5), true); +} + +#[test] +fn test_first_last() { + let mut a = BTreeSet::new(); + assert_eq!(a.first(), None); + assert_eq!(a.last(), None); + a.insert(1); + assert_eq!(a.first(), Some(&1)); + assert_eq!(a.last(), Some(&1)); + a.insert(2); + assert_eq!(a.first(), Some(&1)); + assert_eq!(a.last(), Some(&2)); + for i in 3..=12 { + a.insert(i); + } + assert_eq!(a.first(), Some(&1)); + assert_eq!(a.last(), Some(&12)); + assert_eq!(a.pop_first(), Some(1)); + assert_eq!(a.pop_last(), Some(12)); + assert_eq!(a.pop_first(), Some(2)); + assert_eq!(a.pop_last(), Some(11)); + assert_eq!(a.pop_first(), Some(3)); + assert_eq!(a.pop_last(), Some(10)); + assert_eq!(a.pop_first(), Some(4)); + assert_eq!(a.pop_first(), Some(5)); + assert_eq!(a.pop_first(), Some(6)); + assert_eq!(a.pop_first(), Some(7)); + assert_eq!(a.pop_first(), Some(8)); + assert_eq!(a.clone().pop_last(), Some(9)); + assert_eq!(a.pop_first(), Some(9)); + assert_eq!(a.pop_first(), None); + assert_eq!(a.pop_last(), None); +} + +// Unlike the function with the same name in map/tests, returns no values. +// Which also means it returns different predetermined pseudo-random keys, +// and the test cases using this function explore slightly different trees. +fn rand_data(len: usize) -> Vec<u32> { + let mut rng = DeterministicRng::new(); + Vec::from_iter((0..len).map(|_| rng.next())) +} + +#[test] +fn test_split_off_empty_right() { + let mut data = rand_data(173); + + let mut set = BTreeSet::from_iter(data.clone()); + let right = set.split_off(&(data.iter().max().unwrap() + 1)); + + data.sort(); + assert!(set.into_iter().eq(data)); + assert!(right.into_iter().eq(None)); +} + +#[test] +fn test_split_off_empty_left() { + let mut data = rand_data(314); + + let mut set = BTreeSet::from_iter(data.clone()); + let right = set.split_off(data.iter().min().unwrap()); + + data.sort(); + assert!(set.into_iter().eq(None)); + assert!(right.into_iter().eq(data)); +} + +#[test] +fn test_split_off_large_random_sorted() { + // Miri is too slow + let mut data = if cfg!(miri) { rand_data(529) } else { rand_data(1529) }; + // special case with maximum height. + data.sort(); + + let mut set = BTreeSet::from_iter(data.clone()); + let key = data[data.len() / 2]; + let right = set.split_off(&key); + + assert!(set.into_iter().eq(data.clone().into_iter().filter(|x| *x < key))); + assert!(right.into_iter().eq(data.into_iter().filter(|x| *x >= key))); +} + +#[test] +fn from_array() { + let set = BTreeSet::from([1, 2, 3, 4]); + let unordered_duplicates = BTreeSet::from([4, 1, 4, 3, 2]); + assert_eq!(set, unordered_duplicates); +} + +#[should_panic(expected = "range start is greater than range end in BTreeSet")] +#[test] +fn test_range_panic_1() { + let mut set = BTreeSet::new(); + set.insert(3); + set.insert(5); + set.insert(8); + + let _invalid_range = set.range((Included(&8), Included(&3))); +} + +#[should_panic(expected = "range start and end are equal and excluded in BTreeSet")] +#[test] +fn test_range_panic_2() { + let mut set = BTreeSet::new(); + set.insert(3); + set.insert(5); + set.insert(8); + + let _invalid_range = set.range((Excluded(&5), Excluded(&5))); +} diff --git a/library/alloc/src/collections/btree/set_val.rs b/library/alloc/src/collections/btree/set_val.rs new file mode 100644 index 000000000..80c459bcf --- /dev/null +++ b/library/alloc/src/collections/btree/set_val.rs @@ -0,0 +1,29 @@ +/// Zero-Sized Type (ZST) for internal `BTreeSet` values. +/// Used instead of `()` to differentiate between: +/// * `BTreeMap<T, ()>` (possible user-defined map) +/// * `BTreeMap<T, SetValZST>` (internal set representation) +#[derive(Debug, Eq, PartialEq, Ord, PartialOrd, Hash, Clone, Default)] +pub struct SetValZST; + +/// A trait to differentiate between `BTreeMap` and `BTreeSet` values. +/// Returns `true` only for type `SetValZST`, `false` for all other types (blanket implementation). +/// `TypeId` requires a `'static` lifetime, use of this trait avoids that restriction. +/// +/// [`TypeId`]: std::any::TypeId +pub trait IsSetVal { + fn is_set_val() -> bool; +} + +// Blanket implementation +impl<V> IsSetVal for V { + default fn is_set_val() -> bool { + false + } +} + +// Specialization +impl IsSetVal for SetValZST { + fn is_set_val() -> bool { + true + } +} diff --git a/library/alloc/src/collections/btree/split.rs b/library/alloc/src/collections/btree/split.rs new file mode 100644 index 000000000..638dc98fc --- /dev/null +++ b/library/alloc/src/collections/btree/split.rs @@ -0,0 +1,73 @@ +use super::node::{ForceResult::*, Root}; +use super::search::SearchResult::*; +use core::alloc::Allocator; +use core::borrow::Borrow; + +impl<K, V> Root<K, V> { + /// Calculates the length of both trees that result from splitting up + /// a given number of distinct key-value pairs. + pub fn calc_split_length( + total_num: usize, + root_a: &Root<K, V>, + root_b: &Root<K, V>, + ) -> (usize, usize) { + let (length_a, length_b); + if root_a.height() < root_b.height() { + length_a = root_a.reborrow().calc_length(); + length_b = total_num - length_a; + debug_assert_eq!(length_b, root_b.reborrow().calc_length()); + } else { + length_b = root_b.reborrow().calc_length(); + length_a = total_num - length_b; + debug_assert_eq!(length_a, root_a.reborrow().calc_length()); + } + (length_a, length_b) + } + + /// Split off a tree with key-value pairs at and after the given key. + /// The result is meaningful only if the tree is ordered by key, + /// and if the ordering of `Q` corresponds to that of `K`. + /// If `self` respects all `BTreeMap` tree invariants, then both + /// `self` and the returned tree will respect those invariants. + pub fn split_off<Q: ?Sized + Ord, A: Allocator + Clone>(&mut self, key: &Q, alloc: A) -> Self + where + K: Borrow<Q>, + { + let left_root = self; + let mut right_root = Root::new_pillar(left_root.height(), alloc.clone()); + let mut left_node = left_root.borrow_mut(); + let mut right_node = right_root.borrow_mut(); + + loop { + let mut split_edge = match left_node.search_node(key) { + // key is going to the right tree + Found(kv) => kv.left_edge(), + GoDown(edge) => edge, + }; + + split_edge.move_suffix(&mut right_node); + + match (split_edge.force(), right_node.force()) { + (Internal(edge), Internal(node)) => { + left_node = edge.descend(); + right_node = node.first_edge().descend(); + } + (Leaf(_), Leaf(_)) => break, + _ => unreachable!(), + } + } + + left_root.fix_right_border(alloc.clone()); + right_root.fix_left_border(alloc); + right_root + } + + /// Creates a tree consisting of empty nodes. + fn new_pillar<A: Allocator + Clone>(height: usize, alloc: A) -> Self { + let mut root = Root::new(alloc.clone()); + for _ in 0..height { + root.push_internal_level(alloc.clone()); + } + root + } +} diff --git a/library/alloc/src/collections/btree/testing/crash_test.rs b/library/alloc/src/collections/btree/testing/crash_test.rs new file mode 100644 index 000000000..bcf5f5f72 --- /dev/null +++ b/library/alloc/src/collections/btree/testing/crash_test.rs @@ -0,0 +1,119 @@ +// We avoid relying on anything else in the crate, apart from the `Debug` trait. +use crate::fmt::Debug; +use std::cmp::Ordering; +use std::sync::atomic::{AtomicUsize, Ordering::SeqCst}; + +/// A blueprint for crash test dummy instances that monitor particular events. +/// Some instances may be configured to panic at some point. +/// Events are `clone`, `drop` or some anonymous `query`. +/// +/// Crash test dummies are identified and ordered by an id, so they can be used +/// as keys in a BTreeMap. +#[derive(Debug)] +pub struct CrashTestDummy { + pub id: usize, + cloned: AtomicUsize, + dropped: AtomicUsize, + queried: AtomicUsize, +} + +impl CrashTestDummy { + /// Creates a crash test dummy design. The `id` determines order and equality of instances. + pub fn new(id: usize) -> CrashTestDummy { + CrashTestDummy { + id, + cloned: AtomicUsize::new(0), + dropped: AtomicUsize::new(0), + queried: AtomicUsize::new(0), + } + } + + /// Creates an instance of a crash test dummy that records what events it experiences + /// and optionally panics. + pub fn spawn(&self, panic: Panic) -> Instance<'_> { + Instance { origin: self, panic } + } + + /// Returns how many times instances of the dummy have been cloned. + pub fn cloned(&self) -> usize { + self.cloned.load(SeqCst) + } + + /// Returns how many times instances of the dummy have been dropped. + pub fn dropped(&self) -> usize { + self.dropped.load(SeqCst) + } + + /// Returns how many times instances of the dummy have had their `query` member invoked. + pub fn queried(&self) -> usize { + self.queried.load(SeqCst) + } +} + +#[derive(Debug)] +pub struct Instance<'a> { + origin: &'a CrashTestDummy, + panic: Panic, +} + +#[derive(Copy, Clone, Debug, PartialEq, Eq)] +pub enum Panic { + Never, + InClone, + InDrop, + InQuery, +} + +impl Instance<'_> { + pub fn id(&self) -> usize { + self.origin.id + } + + /// Some anonymous query, the result of which is already given. + pub fn query<R>(&self, result: R) -> R { + self.origin.queried.fetch_add(1, SeqCst); + if self.panic == Panic::InQuery { + panic!("panic in `query`"); + } + result + } +} + +impl Clone for Instance<'_> { + fn clone(&self) -> Self { + self.origin.cloned.fetch_add(1, SeqCst); + if self.panic == Panic::InClone { + panic!("panic in `clone`"); + } + Self { origin: self.origin, panic: Panic::Never } + } +} + +impl Drop for Instance<'_> { + fn drop(&mut self) { + self.origin.dropped.fetch_add(1, SeqCst); + if self.panic == Panic::InDrop { + panic!("panic in `drop`"); + } + } +} + +impl PartialOrd for Instance<'_> { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + self.id().partial_cmp(&other.id()) + } +} + +impl Ord for Instance<'_> { + fn cmp(&self, other: &Self) -> Ordering { + self.id().cmp(&other.id()) + } +} + +impl PartialEq for Instance<'_> { + fn eq(&self, other: &Self) -> bool { + self.id().eq(&other.id()) + } +} + +impl Eq for Instance<'_> {} diff --git a/library/alloc/src/collections/btree/testing/mod.rs b/library/alloc/src/collections/btree/testing/mod.rs new file mode 100644 index 000000000..7a094f8a5 --- /dev/null +++ b/library/alloc/src/collections/btree/testing/mod.rs @@ -0,0 +1,3 @@ +pub mod crash_test; +pub mod ord_chaos; +pub mod rng; diff --git a/library/alloc/src/collections/btree/testing/ord_chaos.rs b/library/alloc/src/collections/btree/testing/ord_chaos.rs new file mode 100644 index 000000000..96ce7c157 --- /dev/null +++ b/library/alloc/src/collections/btree/testing/ord_chaos.rs @@ -0,0 +1,81 @@ +use std::cell::Cell; +use std::cmp::Ordering::{self, *}; +use std::ptr; + +// Minimal type with an `Ord` implementation violating transitivity. +#[derive(Debug)] +pub enum Cyclic3 { + A, + B, + C, +} +use Cyclic3::*; + +impl PartialOrd for Cyclic3 { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + Some(self.cmp(other)) + } +} + +impl Ord for Cyclic3 { + fn cmp(&self, other: &Self) -> Ordering { + match (self, other) { + (A, A) | (B, B) | (C, C) => Equal, + (A, B) | (B, C) | (C, A) => Less, + (A, C) | (B, A) | (C, B) => Greater, + } + } +} + +impl PartialEq for Cyclic3 { + fn eq(&self, other: &Self) -> bool { + self.cmp(&other) == Equal + } +} + +impl Eq for Cyclic3 {} + +// Controls the ordering of values wrapped by `Governed`. +#[derive(Debug)] +pub struct Governor { + flipped: Cell<bool>, +} + +impl Governor { + pub fn new() -> Self { + Governor { flipped: Cell::new(false) } + } + + pub fn flip(&self) { + self.flipped.set(!self.flipped.get()); + } +} + +// Type with an `Ord` implementation that forms a total order at any moment +// (assuming that `T` respects total order), but can suddenly be made to invert +// that total order. +#[derive(Debug)] +pub struct Governed<'a, T>(pub T, pub &'a Governor); + +impl<T: Ord> PartialOrd for Governed<'_, T> { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + Some(self.cmp(other)) + } +} + +impl<T: Ord> Ord for Governed<'_, T> { + fn cmp(&self, other: &Self) -> Ordering { + assert!(ptr::eq(self.1, other.1)); + let ord = self.0.cmp(&other.0); + if self.1.flipped.get() { ord.reverse() } else { ord } + } +} + +impl<T: PartialEq> PartialEq for Governed<'_, T> { + fn eq(&self, other: &Self) -> bool { + assert!(ptr::eq(self.1, other.1)); + self.0.eq(&other.0) + } +} + +impl<T: Eq> Eq for Governed<'_, T> {} diff --git a/library/alloc/src/collections/btree/testing/rng.rs b/library/alloc/src/collections/btree/testing/rng.rs new file mode 100644 index 000000000..ecf543bee --- /dev/null +++ b/library/alloc/src/collections/btree/testing/rng.rs @@ -0,0 +1,28 @@ +/// XorShiftRng +pub struct DeterministicRng { + count: usize, + x: u32, + y: u32, + z: u32, + w: u32, +} + +impl DeterministicRng { + pub fn new() -> Self { + DeterministicRng { count: 0, x: 0x193a6754, y: 0xa8a7d469, z: 0x97830e05, w: 0x113ba7bb } + } + + /// Guarantees that each returned number is unique. + pub fn next(&mut self) -> u32 { + self.count += 1; + assert!(self.count <= 70029); + let x = self.x; + let t = x ^ (x << 11); + self.x = self.y; + self.y = self.z; + self.z = self.w; + let w_ = self.w; + self.w = w_ ^ (w_ >> 19) ^ (t ^ (t >> 8)); + self.w + } +} diff --git a/library/alloc/src/collections/linked_list.rs b/library/alloc/src/collections/linked_list.rs new file mode 100644 index 000000000..e21c8aa3b --- /dev/null +++ b/library/alloc/src/collections/linked_list.rs @@ -0,0 +1,2012 @@ +//! A doubly-linked list with owned nodes. +//! +//! The `LinkedList` allows pushing and popping elements at either end +//! in constant time. +//! +//! NOTE: It is almost always better to use [`Vec`] or [`VecDeque`] because +//! array-based containers are generally faster, +//! more memory efficient, and make better use of CPU cache. +//! +//! [`Vec`]: crate::vec::Vec +//! [`VecDeque`]: super::vec_deque::VecDeque + +#![stable(feature = "rust1", since = "1.0.0")] + +use core::cmp::Ordering; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::iter::{FromIterator, FusedIterator}; +use core::marker::PhantomData; +use core::mem; +use core::ptr::NonNull; + +use super::SpecExtend; +use crate::boxed::Box; + +#[cfg(test)] +mod tests; + +/// A doubly-linked list with owned nodes. +/// +/// The `LinkedList` allows pushing and popping elements at either end +/// in constant time. +/// +/// A `LinkedList` with a known list of items can be initialized from an array: +/// ``` +/// use std::collections::LinkedList; +/// +/// let list = LinkedList::from([1, 2, 3]); +/// ``` +/// +/// NOTE: It is almost always better to use [`Vec`] or [`VecDeque`] because +/// array-based containers are generally faster, +/// more memory efficient, and make better use of CPU cache. +/// +/// [`Vec`]: crate::vec::Vec +/// [`VecDeque`]: super::vec_deque::VecDeque +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "LinkedList")] +#[rustc_insignificant_dtor] +pub struct LinkedList<T> { + head: Option<NonNull<Node<T>>>, + tail: Option<NonNull<Node<T>>>, + len: usize, + marker: PhantomData<Box<Node<T>>>, +} + +struct Node<T> { + next: Option<NonNull<Node<T>>>, + prev: Option<NonNull<Node<T>>>, + element: T, +} + +/// An iterator over the elements of a `LinkedList`. +/// +/// This `struct` is created by [`LinkedList::iter()`]. See its +/// documentation for more. +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Iter<'a, T: 'a> { + head: Option<NonNull<Node<T>>>, + tail: Option<NonNull<Node<T>>>, + len: usize, + marker: PhantomData<&'a Node<T>>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for Iter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Iter") + .field(&*mem::ManuallyDrop::new(LinkedList { + head: self.head, + tail: self.tail, + len: self.len, + marker: PhantomData, + })) + .field(&self.len) + .finish() + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for Iter<'_, T> { + fn clone(&self) -> Self { + Iter { ..*self } + } +} + +/// A mutable iterator over the elements of a `LinkedList`. +/// +/// This `struct` is created by [`LinkedList::iter_mut()`]. See its +/// documentation for more. +#[must_use = "iterators are lazy and do nothing unless consumed"] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct IterMut<'a, T: 'a> { + head: Option<NonNull<Node<T>>>, + tail: Option<NonNull<Node<T>>>, + len: usize, + marker: PhantomData<&'a mut Node<T>>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for IterMut<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("IterMut") + .field(&*mem::ManuallyDrop::new(LinkedList { + head: self.head, + tail: self.tail, + len: self.len, + marker: PhantomData, + })) + .field(&self.len) + .finish() + } +} + +/// An owning iterator over the elements of a `LinkedList`. +/// +/// This `struct` is created by the [`into_iter`] method on [`LinkedList`] +/// (provided by the [`IntoIterator`] trait). See its documentation for more. +/// +/// [`into_iter`]: LinkedList::into_iter +/// [`IntoIterator`]: core::iter::IntoIterator +#[derive(Clone)] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct IntoIter<T> { + list: LinkedList<T>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for IntoIter<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("IntoIter").field(&self.list).finish() + } +} + +impl<T> Node<T> { + fn new(element: T) -> Self { + Node { next: None, prev: None, element } + } + + fn into_element(self: Box<Self>) -> T { + self.element + } +} + +// private methods +impl<T> LinkedList<T> { + /// Adds the given node to the front of the list. + #[inline] + fn push_front_node(&mut self, mut node: Box<Node<T>>) { + // This method takes care not to create mutable references to whole nodes, + // to maintain validity of aliasing pointers into `element`. + unsafe { + node.next = self.head; + node.prev = None; + let node = Some(Box::leak(node).into()); + + match self.head { + None => self.tail = node, + // Not creating new mutable (unique!) references overlapping `element`. + Some(head) => (*head.as_ptr()).prev = node, + } + + self.head = node; + self.len += 1; + } + } + + /// Removes and returns the node at the front of the list. + #[inline] + fn pop_front_node(&mut self) -> Option<Box<Node<T>>> { + // This method takes care not to create mutable references to whole nodes, + // to maintain validity of aliasing pointers into `element`. + self.head.map(|node| unsafe { + let node = Box::from_raw(node.as_ptr()); + self.head = node.next; + + match self.head { + None => self.tail = None, + // Not creating new mutable (unique!) references overlapping `element`. + Some(head) => (*head.as_ptr()).prev = None, + } + + self.len -= 1; + node + }) + } + + /// Adds the given node to the back of the list. + #[inline] + fn push_back_node(&mut self, mut node: Box<Node<T>>) { + // This method takes care not to create mutable references to whole nodes, + // to maintain validity of aliasing pointers into `element`. + unsafe { + node.next = None; + node.prev = self.tail; + let node = Some(Box::leak(node).into()); + + match self.tail { + None => self.head = node, + // Not creating new mutable (unique!) references overlapping `element`. + Some(tail) => (*tail.as_ptr()).next = node, + } + + self.tail = node; + self.len += 1; + } + } + + /// Removes and returns the node at the back of the list. + #[inline] + fn pop_back_node(&mut self) -> Option<Box<Node<T>>> { + // This method takes care not to create mutable references to whole nodes, + // to maintain validity of aliasing pointers into `element`. + self.tail.map(|node| unsafe { + let node = Box::from_raw(node.as_ptr()); + self.tail = node.prev; + + match self.tail { + None => self.head = None, + // Not creating new mutable (unique!) references overlapping `element`. + Some(tail) => (*tail.as_ptr()).next = None, + } + + self.len -= 1; + node + }) + } + + /// Unlinks the specified node from the current list. + /// + /// Warning: this will not check that the provided node belongs to the current list. + /// + /// This method takes care not to create mutable references to `element`, to + /// maintain validity of aliasing pointers. + #[inline] + unsafe fn unlink_node(&mut self, mut node: NonNull<Node<T>>) { + let node = unsafe { node.as_mut() }; // this one is ours now, we can create an &mut. + + // Not creating new mutable (unique!) references overlapping `element`. + match node.prev { + Some(prev) => unsafe { (*prev.as_ptr()).next = node.next }, + // this node is the head node + None => self.head = node.next, + }; + + match node.next { + Some(next) => unsafe { (*next.as_ptr()).prev = node.prev }, + // this node is the tail node + None => self.tail = node.prev, + }; + + self.len -= 1; + } + + /// Splices a series of nodes between two existing nodes. + /// + /// Warning: this will not check that the provided node belongs to the two existing lists. + #[inline] + unsafe fn splice_nodes( + &mut self, + existing_prev: Option<NonNull<Node<T>>>, + existing_next: Option<NonNull<Node<T>>>, + mut splice_start: NonNull<Node<T>>, + mut splice_end: NonNull<Node<T>>, + splice_length: usize, + ) { + // This method takes care not to create multiple mutable references to whole nodes at the same time, + // to maintain validity of aliasing pointers into `element`. + if let Some(mut existing_prev) = existing_prev { + unsafe { + existing_prev.as_mut().next = Some(splice_start); + } + } else { + self.head = Some(splice_start); + } + if let Some(mut existing_next) = existing_next { + unsafe { + existing_next.as_mut().prev = Some(splice_end); + } + } else { + self.tail = Some(splice_end); + } + unsafe { + splice_start.as_mut().prev = existing_prev; + splice_end.as_mut().next = existing_next; + } + + self.len += splice_length; + } + + /// Detaches all nodes from a linked list as a series of nodes. + #[inline] + fn detach_all_nodes(mut self) -> Option<(NonNull<Node<T>>, NonNull<Node<T>>, usize)> { + let head = self.head.take(); + let tail = self.tail.take(); + let len = mem::replace(&mut self.len, 0); + if let Some(head) = head { + // SAFETY: In a LinkedList, either both the head and tail are None because + // the list is empty, or both head and tail are Some because the list is populated. + // Since we have verified the head is Some, we are sure the tail is Some too. + let tail = unsafe { tail.unwrap_unchecked() }; + Some((head, tail, len)) + } else { + None + } + } + + #[inline] + unsafe fn split_off_before_node( + &mut self, + split_node: Option<NonNull<Node<T>>>, + at: usize, + ) -> Self { + // The split node is the new head node of the second part + if let Some(mut split_node) = split_node { + let first_part_head; + let first_part_tail; + unsafe { + first_part_tail = split_node.as_mut().prev.take(); + } + if let Some(mut tail) = first_part_tail { + unsafe { + tail.as_mut().next = None; + } + first_part_head = self.head; + } else { + first_part_head = None; + } + + let first_part = LinkedList { + head: first_part_head, + tail: first_part_tail, + len: at, + marker: PhantomData, + }; + + // Fix the head ptr of the second part + self.head = Some(split_node); + self.len = self.len - at; + + first_part + } else { + mem::replace(self, LinkedList::new()) + } + } + + #[inline] + unsafe fn split_off_after_node( + &mut self, + split_node: Option<NonNull<Node<T>>>, + at: usize, + ) -> Self { + // The split node is the new tail node of the first part and owns + // the head of the second part. + if let Some(mut split_node) = split_node { + let second_part_head; + let second_part_tail; + unsafe { + second_part_head = split_node.as_mut().next.take(); + } + if let Some(mut head) = second_part_head { + unsafe { + head.as_mut().prev = None; + } + second_part_tail = self.tail; + } else { + second_part_tail = None; + } + + let second_part = LinkedList { + head: second_part_head, + tail: second_part_tail, + len: self.len - at, + marker: PhantomData, + }; + + // Fix the tail ptr of the first part + self.tail = Some(split_node); + self.len = at; + + second_part + } else { + mem::replace(self, LinkedList::new()) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Default for LinkedList<T> { + /// Creates an empty `LinkedList<T>`. + #[inline] + fn default() -> Self { + Self::new() + } +} + +impl<T> LinkedList<T> { + /// Creates an empty `LinkedList`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let list: LinkedList<u32> = LinkedList::new(); + /// ``` + #[inline] + #[rustc_const_stable(feature = "const_linked_list_new", since = "1.39.0")] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub const fn new() -> Self { + LinkedList { head: None, tail: None, len: 0, marker: PhantomData } + } + + /// Moves all elements from `other` to the end of the list. + /// + /// This reuses all the nodes from `other` and moves them into `self`. After + /// this operation, `other` becomes empty. + /// + /// This operation should compute in *O*(1) time and *O*(1) memory. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut list1 = LinkedList::new(); + /// list1.push_back('a'); + /// + /// let mut list2 = LinkedList::new(); + /// list2.push_back('b'); + /// list2.push_back('c'); + /// + /// list1.append(&mut list2); + /// + /// let mut iter = list1.iter(); + /// assert_eq!(iter.next(), Some(&'a')); + /// assert_eq!(iter.next(), Some(&'b')); + /// assert_eq!(iter.next(), Some(&'c')); + /// assert!(iter.next().is_none()); + /// + /// assert!(list2.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn append(&mut self, other: &mut Self) { + match self.tail { + None => mem::swap(self, other), + Some(mut tail) => { + // `as_mut` is okay here because we have exclusive access to the entirety + // of both lists. + if let Some(mut other_head) = other.head.take() { + unsafe { + tail.as_mut().next = Some(other_head); + other_head.as_mut().prev = Some(tail); + } + + self.tail = other.tail.take(); + self.len += mem::replace(&mut other.len, 0); + } + } + } + } + + /// Provides a forward iterator. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut list: LinkedList<u32> = LinkedList::new(); + /// + /// list.push_back(0); + /// list.push_back(1); + /// list.push_back(2); + /// + /// let mut iter = list.iter(); + /// assert_eq!(iter.next(), Some(&0)); + /// assert_eq!(iter.next(), Some(&1)); + /// assert_eq!(iter.next(), Some(&2)); + /// assert_eq!(iter.next(), None); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter(&self) -> Iter<'_, T> { + Iter { head: self.head, tail: self.tail, len: self.len, marker: PhantomData } + } + + /// Provides a forward iterator with mutable references. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut list: LinkedList<u32> = LinkedList::new(); + /// + /// list.push_back(0); + /// list.push_back(1); + /// list.push_back(2); + /// + /// for element in list.iter_mut() { + /// *element += 10; + /// } + /// + /// let mut iter = list.iter(); + /// assert_eq!(iter.next(), Some(&10)); + /// assert_eq!(iter.next(), Some(&11)); + /// assert_eq!(iter.next(), Some(&12)); + /// assert_eq!(iter.next(), None); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter_mut(&mut self) -> IterMut<'_, T> { + IterMut { head: self.head, tail: self.tail, len: self.len, marker: PhantomData } + } + + /// Provides a cursor at the front element. + /// + /// The cursor is pointing to the "ghost" non-element if the list is empty. + #[inline] + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn cursor_front(&self) -> Cursor<'_, T> { + Cursor { index: 0, current: self.head, list: self } + } + + /// Provides a cursor with editing operations at the front element. + /// + /// The cursor is pointing to the "ghost" non-element if the list is empty. + #[inline] + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn cursor_front_mut(&mut self) -> CursorMut<'_, T> { + CursorMut { index: 0, current: self.head, list: self } + } + + /// Provides a cursor at the back element. + /// + /// The cursor is pointing to the "ghost" non-element if the list is empty. + #[inline] + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn cursor_back(&self) -> Cursor<'_, T> { + Cursor { index: self.len.checked_sub(1).unwrap_or(0), current: self.tail, list: self } + } + + /// Provides a cursor with editing operations at the back element. + /// + /// The cursor is pointing to the "ghost" non-element if the list is empty. + #[inline] + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn cursor_back_mut(&mut self) -> CursorMut<'_, T> { + CursorMut { index: self.len.checked_sub(1).unwrap_or(0), current: self.tail, list: self } + } + + /// Returns `true` if the `LinkedList` is empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// assert!(dl.is_empty()); + /// + /// dl.push_front("foo"); + /// assert!(!dl.is_empty()); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_empty(&self) -> bool { + self.head.is_none() + } + + /// Returns the length of the `LinkedList`. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// + /// dl.push_front(2); + /// assert_eq!(dl.len(), 1); + /// + /// dl.push_front(1); + /// assert_eq!(dl.len(), 2); + /// + /// dl.push_back(3); + /// assert_eq!(dl.len(), 3); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn len(&self) -> usize { + self.len + } + + /// Removes all elements from the `LinkedList`. + /// + /// This operation should compute in *O*(*n*) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// + /// dl.push_front(2); + /// dl.push_front(1); + /// assert_eq!(dl.len(), 2); + /// assert_eq!(dl.front(), Some(&1)); + /// + /// dl.clear(); + /// assert_eq!(dl.len(), 0); + /// assert_eq!(dl.front(), None); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) { + *self = Self::new(); + } + + /// Returns `true` if the `LinkedList` contains an element equal to the + /// given value. + /// + /// This operation should compute linearly in *O*(*n*) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut list: LinkedList<u32> = LinkedList::new(); + /// + /// list.push_back(0); + /// list.push_back(1); + /// list.push_back(2); + /// + /// assert_eq!(list.contains(&0), true); + /// assert_eq!(list.contains(&10), false); + /// ``` + #[stable(feature = "linked_list_contains", since = "1.12.0")] + pub fn contains(&self, x: &T) -> bool + where + T: PartialEq<T>, + { + self.iter().any(|e| e == x) + } + + /// Provides a reference to the front element, or `None` if the list is + /// empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// assert_eq!(dl.front(), None); + /// + /// dl.push_front(1); + /// assert_eq!(dl.front(), Some(&1)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn front(&self) -> Option<&T> { + unsafe { self.head.as_ref().map(|node| &node.as_ref().element) } + } + + /// Provides a mutable reference to the front element, or `None` if the list + /// is empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// assert_eq!(dl.front(), None); + /// + /// dl.push_front(1); + /// assert_eq!(dl.front(), Some(&1)); + /// + /// match dl.front_mut() { + /// None => {}, + /// Some(x) => *x = 5, + /// } + /// assert_eq!(dl.front(), Some(&5)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn front_mut(&mut self) -> Option<&mut T> { + unsafe { self.head.as_mut().map(|node| &mut node.as_mut().element) } + } + + /// Provides a reference to the back element, or `None` if the list is + /// empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// assert_eq!(dl.back(), None); + /// + /// dl.push_back(1); + /// assert_eq!(dl.back(), Some(&1)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn back(&self) -> Option<&T> { + unsafe { self.tail.as_ref().map(|node| &node.as_ref().element) } + } + + /// Provides a mutable reference to the back element, or `None` if the list + /// is empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// assert_eq!(dl.back(), None); + /// + /// dl.push_back(1); + /// assert_eq!(dl.back(), Some(&1)); + /// + /// match dl.back_mut() { + /// None => {}, + /// Some(x) => *x = 5, + /// } + /// assert_eq!(dl.back(), Some(&5)); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn back_mut(&mut self) -> Option<&mut T> { + unsafe { self.tail.as_mut().map(|node| &mut node.as_mut().element) } + } + + /// Adds an element first in the list. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut dl = LinkedList::new(); + /// + /// dl.push_front(2); + /// assert_eq!(dl.front().unwrap(), &2); + /// + /// dl.push_front(1); + /// assert_eq!(dl.front().unwrap(), &1); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push_front(&mut self, elt: T) { + self.push_front_node(Box::new(Node::new(elt))); + } + + /// Removes the first element and returns it, or `None` if the list is + /// empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut d = LinkedList::new(); + /// assert_eq!(d.pop_front(), None); + /// + /// d.push_front(1); + /// d.push_front(3); + /// assert_eq!(d.pop_front(), Some(3)); + /// assert_eq!(d.pop_front(), Some(1)); + /// assert_eq!(d.pop_front(), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop_front(&mut self) -> Option<T> { + self.pop_front_node().map(Node::into_element) + } + + /// Appends an element to the back of a list. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut d = LinkedList::new(); + /// d.push_back(1); + /// d.push_back(3); + /// assert_eq!(3, *d.back().unwrap()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push_back(&mut self, elt: T) { + self.push_back_node(Box::new(Node::new(elt))); + } + + /// Removes the last element from a list and returns it, or `None` if + /// it is empty. + /// + /// This operation should compute in *O*(1) time. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut d = LinkedList::new(); + /// assert_eq!(d.pop_back(), None); + /// d.push_back(1); + /// d.push_back(3); + /// assert_eq!(d.pop_back(), Some(3)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop_back(&mut self) -> Option<T> { + self.pop_back_node().map(Node::into_element) + } + + /// Splits the list into two at the given index. Returns everything after the given index, + /// including the index. + /// + /// This operation should compute in *O*(*n*) time. + /// + /// # Panics + /// + /// Panics if `at > len`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let mut d = LinkedList::new(); + /// + /// d.push_front(1); + /// d.push_front(2); + /// d.push_front(3); + /// + /// let mut split = d.split_off(2); + /// + /// assert_eq!(split.pop_front(), Some(1)); + /// assert_eq!(split.pop_front(), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn split_off(&mut self, at: usize) -> LinkedList<T> { + let len = self.len(); + assert!(at <= len, "Cannot split off at a nonexistent index"); + if at == 0 { + return mem::take(self); + } else if at == len { + return Self::new(); + } + + // Below, we iterate towards the `i-1`th node, either from the start or the end, + // depending on which would be faster. + let split_node = if at - 1 <= len - 1 - (at - 1) { + let mut iter = self.iter_mut(); + // instead of skipping using .skip() (which creates a new struct), + // we skip manually so we can access the head field without + // depending on implementation details of Skip + for _ in 0..at - 1 { + iter.next(); + } + iter.head + } else { + // better off starting from the end + let mut iter = self.iter_mut(); + for _ in 0..len - 1 - (at - 1) { + iter.next_back(); + } + iter.tail + }; + unsafe { self.split_off_after_node(split_node, at) } + } + + /// Removes the element at the given index and returns it. + /// + /// This operation should compute in *O*(*n*) time. + /// + /// # Panics + /// Panics if at >= len + /// + /// # Examples + /// + /// ``` + /// #![feature(linked_list_remove)] + /// use std::collections::LinkedList; + /// + /// let mut d = LinkedList::new(); + /// + /// d.push_front(1); + /// d.push_front(2); + /// d.push_front(3); + /// + /// assert_eq!(d.remove(1), 2); + /// assert_eq!(d.remove(0), 3); + /// assert_eq!(d.remove(0), 1); + /// ``` + #[unstable(feature = "linked_list_remove", issue = "69210")] + pub fn remove(&mut self, at: usize) -> T { + let len = self.len(); + assert!(at < len, "Cannot remove at an index outside of the list bounds"); + + // Below, we iterate towards the node at the given index, either from + // the start or the end, depending on which would be faster. + let offset_from_end = len - at - 1; + if at <= offset_from_end { + let mut cursor = self.cursor_front_mut(); + for _ in 0..at { + cursor.move_next(); + } + cursor.remove_current().unwrap() + } else { + let mut cursor = self.cursor_back_mut(); + for _ in 0..offset_from_end { + cursor.move_prev(); + } + cursor.remove_current().unwrap() + } + } + + /// Creates an iterator which uses a closure to determine if an element should be removed. + /// + /// If the closure returns true, then the element is removed and yielded. + /// If the closure returns false, the element will remain in the list and will not be yielded + /// by the iterator. + /// + /// Note that `drain_filter` lets you mutate every element in the filter closure, regardless of + /// whether you choose to keep or remove it. + /// + /// # Examples + /// + /// Splitting a list into evens and odds, reusing the original list: + /// + /// ``` + /// #![feature(drain_filter)] + /// use std::collections::LinkedList; + /// + /// let mut numbers: LinkedList<u32> = LinkedList::new(); + /// numbers.extend(&[1, 2, 3, 4, 5, 6, 8, 9, 11, 13, 14, 15]); + /// + /// let evens = numbers.drain_filter(|x| *x % 2 == 0).collect::<LinkedList<_>>(); + /// let odds = numbers; + /// + /// assert_eq!(evens.into_iter().collect::<Vec<_>>(), vec![2, 4, 6, 8, 14]); + /// assert_eq!(odds.into_iter().collect::<Vec<_>>(), vec![1, 3, 5, 9, 11, 13, 15]); + /// ``` + #[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] + pub fn drain_filter<F>(&mut self, filter: F) -> DrainFilter<'_, T, F> + where + F: FnMut(&mut T) -> bool, + { + // avoid borrow issues. + let it = self.head; + let old_len = self.len; + + DrainFilter { list: self, it, pred: filter, idx: 0, old_len } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T> Drop for LinkedList<T> { + fn drop(&mut self) { + struct DropGuard<'a, T>(&'a mut LinkedList<T>); + + impl<'a, T> Drop for DropGuard<'a, T> { + fn drop(&mut self) { + // Continue the same loop we do below. This only runs when a destructor has + // panicked. If another one panics this will abort. + while self.0.pop_front_node().is_some() {} + } + } + + while let Some(node) = self.pop_front_node() { + let guard = DropGuard(self); + drop(node); + mem::forget(guard); + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for Iter<'a, T> { + type Item = &'a T; + + #[inline] + fn next(&mut self) -> Option<&'a T> { + if self.len == 0 { + None + } else { + self.head.map(|node| unsafe { + // Need an unbound lifetime to get 'a + let node = &*node.as_ptr(); + self.len -= 1; + self.head = node.next; + &node.element + }) + } + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + (self.len, Some(self.len)) + } + + #[inline] + fn last(mut self) -> Option<&'a T> { + self.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for Iter<'a, T> { + #[inline] + fn next_back(&mut self) -> Option<&'a T> { + if self.len == 0 { + None + } else { + self.tail.map(|node| unsafe { + // Need an unbound lifetime to get 'a + let node = &*node.as_ptr(); + self.len -= 1; + self.tail = node.prev; + &node.element + }) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for Iter<'_, T> {} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Iter<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for IterMut<'a, T> { + type Item = &'a mut T; + + #[inline] + fn next(&mut self) -> Option<&'a mut T> { + if self.len == 0 { + None + } else { + self.head.map(|node| unsafe { + // Need an unbound lifetime to get 'a + let node = &mut *node.as_ptr(); + self.len -= 1; + self.head = node.next; + &mut node.element + }) + } + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + (self.len, Some(self.len)) + } + + #[inline] + fn last(mut self) -> Option<&'a mut T> { + self.next_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for IterMut<'a, T> { + #[inline] + fn next_back(&mut self) -> Option<&'a mut T> { + if self.len == 0 { + None + } else { + self.tail.map(|node| unsafe { + // Need an unbound lifetime to get 'a + let node = &mut *node.as_ptr(); + self.len -= 1; + self.tail = node.prev; + &mut node.element + }) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for IterMut<'_, T> {} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for IterMut<'_, T> {} + +/// A cursor over a `LinkedList`. +/// +/// A `Cursor` is like an iterator, except that it can freely seek back-and-forth. +/// +/// Cursors always rest between two elements in the list, and index in a logically circular way. +/// To accommodate this, there is a "ghost" non-element that yields `None` between the head and +/// tail of the list. +/// +/// When created, cursors start at the front of the list, or the "ghost" non-element if the list is empty. +#[unstable(feature = "linked_list_cursors", issue = "58533")] +pub struct Cursor<'a, T: 'a> { + index: usize, + current: Option<NonNull<Node<T>>>, + list: &'a LinkedList<T>, +} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +impl<T> Clone for Cursor<'_, T> { + fn clone(&self) -> Self { + let Cursor { index, current, list } = *self; + Cursor { index, current, list } + } +} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +impl<T: fmt::Debug> fmt::Debug for Cursor<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Cursor").field(&self.list).field(&self.index()).finish() + } +} + +/// A cursor over a `LinkedList` with editing operations. +/// +/// A `Cursor` is like an iterator, except that it can freely seek back-and-forth, and can +/// safely mutate the list during iteration. This is because the lifetime of its yielded +/// references is tied to its own lifetime, instead of just the underlying list. This means +/// cursors cannot yield multiple elements at once. +/// +/// Cursors always rest between two elements in the list, and index in a logically circular way. +/// To accommodate this, there is a "ghost" non-element that yields `None` between the head and +/// tail of the list. +#[unstable(feature = "linked_list_cursors", issue = "58533")] +pub struct CursorMut<'a, T: 'a> { + index: usize, + current: Option<NonNull<Node<T>>>, + list: &'a mut LinkedList<T>, +} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +impl<T: fmt::Debug> fmt::Debug for CursorMut<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("CursorMut").field(&self.list).field(&self.index()).finish() + } +} + +impl<'a, T> Cursor<'a, T> { + /// Returns the cursor position index within the `LinkedList`. + /// + /// This returns `None` if the cursor is currently pointing to the + /// "ghost" non-element. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn index(&self) -> Option<usize> { + let _ = self.current?; + Some(self.index) + } + + /// Moves the cursor to the next element of the `LinkedList`. + /// + /// If the cursor is pointing to the "ghost" non-element then this will move it to + /// the first element of the `LinkedList`. If it is pointing to the last + /// element of the `LinkedList` then this will move it to the "ghost" non-element. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn move_next(&mut self) { + match self.current.take() { + // We had no current element; the cursor was sitting at the start position + // Next element should be the head of the list + None => { + self.current = self.list.head; + self.index = 0; + } + // We had a previous element, so let's go to its next + Some(current) => unsafe { + self.current = current.as_ref().next; + self.index += 1; + }, + } + } + + /// Moves the cursor to the previous element of the `LinkedList`. + /// + /// If the cursor is pointing to the "ghost" non-element then this will move it to + /// the last element of the `LinkedList`. If it is pointing to the first + /// element of the `LinkedList` then this will move it to the "ghost" non-element. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn move_prev(&mut self) { + match self.current.take() { + // No current. We're at the start of the list. Yield None and jump to the end. + None => { + self.current = self.list.tail; + self.index = self.list.len().checked_sub(1).unwrap_or(0); + } + // Have a prev. Yield it and go to the previous element. + Some(current) => unsafe { + self.current = current.as_ref().prev; + self.index = self.index.checked_sub(1).unwrap_or_else(|| self.list.len()); + }, + } + } + + /// Returns a reference to the element that the cursor is currently + /// pointing to. + /// + /// This returns `None` if the cursor is currently pointing to the + /// "ghost" non-element. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn current(&self) -> Option<&'a T> { + unsafe { self.current.map(|current| &(*current.as_ptr()).element) } + } + + /// Returns a reference to the next element. + /// + /// If the cursor is pointing to the "ghost" non-element then this returns + /// the first element of the `LinkedList`. If it is pointing to the last + /// element of the `LinkedList` then this returns `None`. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn peek_next(&self) -> Option<&'a T> { + unsafe { + let next = match self.current { + None => self.list.head, + Some(current) => current.as_ref().next, + }; + next.map(|next| &(*next.as_ptr()).element) + } + } + + /// Returns a reference to the previous element. + /// + /// If the cursor is pointing to the "ghost" non-element then this returns + /// the last element of the `LinkedList`. If it is pointing to the first + /// element of the `LinkedList` then this returns `None`. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn peek_prev(&self) -> Option<&'a T> { + unsafe { + let prev = match self.current { + None => self.list.tail, + Some(current) => current.as_ref().prev, + }; + prev.map(|prev| &(*prev.as_ptr()).element) + } + } + + /// Provides a reference to the front element of the cursor's parent list, + /// or None if the list is empty. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn front(&self) -> Option<&'a T> { + self.list.front() + } + + /// Provides a reference to the back element of the cursor's parent list, + /// or None if the list is empty. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn back(&self) -> Option<&'a T> { + self.list.back() + } +} + +impl<'a, T> CursorMut<'a, T> { + /// Returns the cursor position index within the `LinkedList`. + /// + /// This returns `None` if the cursor is currently pointing to the + /// "ghost" non-element. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn index(&self) -> Option<usize> { + let _ = self.current?; + Some(self.index) + } + + /// Moves the cursor to the next element of the `LinkedList`. + /// + /// If the cursor is pointing to the "ghost" non-element then this will move it to + /// the first element of the `LinkedList`. If it is pointing to the last + /// element of the `LinkedList` then this will move it to the "ghost" non-element. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn move_next(&mut self) { + match self.current.take() { + // We had no current element; the cursor was sitting at the start position + // Next element should be the head of the list + None => { + self.current = self.list.head; + self.index = 0; + } + // We had a previous element, so let's go to its next + Some(current) => unsafe { + self.current = current.as_ref().next; + self.index += 1; + }, + } + } + + /// Moves the cursor to the previous element of the `LinkedList`. + /// + /// If the cursor is pointing to the "ghost" non-element then this will move it to + /// the last element of the `LinkedList`. If it is pointing to the first + /// element of the `LinkedList` then this will move it to the "ghost" non-element. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn move_prev(&mut self) { + match self.current.take() { + // No current. We're at the start of the list. Yield None and jump to the end. + None => { + self.current = self.list.tail; + self.index = self.list.len().checked_sub(1).unwrap_or(0); + } + // Have a prev. Yield it and go to the previous element. + Some(current) => unsafe { + self.current = current.as_ref().prev; + self.index = self.index.checked_sub(1).unwrap_or_else(|| self.list.len()); + }, + } + } + + /// Returns a reference to the element that the cursor is currently + /// pointing to. + /// + /// This returns `None` if the cursor is currently pointing to the + /// "ghost" non-element. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn current(&mut self) -> Option<&mut T> { + unsafe { self.current.map(|current| &mut (*current.as_ptr()).element) } + } + + /// Returns a reference to the next element. + /// + /// If the cursor is pointing to the "ghost" non-element then this returns + /// the first element of the `LinkedList`. If it is pointing to the last + /// element of the `LinkedList` then this returns `None`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn peek_next(&mut self) -> Option<&mut T> { + unsafe { + let next = match self.current { + None => self.list.head, + Some(current) => current.as_ref().next, + }; + next.map(|next| &mut (*next.as_ptr()).element) + } + } + + /// Returns a reference to the previous element. + /// + /// If the cursor is pointing to the "ghost" non-element then this returns + /// the last element of the `LinkedList`. If it is pointing to the first + /// element of the `LinkedList` then this returns `None`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn peek_prev(&mut self) -> Option<&mut T> { + unsafe { + let prev = match self.current { + None => self.list.tail, + Some(current) => current.as_ref().prev, + }; + prev.map(|prev| &mut (*prev.as_ptr()).element) + } + } + + /// Returns a read-only cursor pointing to the current element. + /// + /// The lifetime of the returned `Cursor` is bound to that of the + /// `CursorMut`, which means it cannot outlive the `CursorMut` and that the + /// `CursorMut` is frozen for the lifetime of the `Cursor`. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn as_cursor(&self) -> Cursor<'_, T> { + Cursor { list: self.list, current: self.current, index: self.index } + } +} + +// Now the list editing operations + +impl<'a, T> CursorMut<'a, T> { + /// Inserts a new element into the `LinkedList` after the current one. + /// + /// If the cursor is pointing at the "ghost" non-element then the new element is + /// inserted at the front of the `LinkedList`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn insert_after(&mut self, item: T) { + unsafe { + let spliced_node = Box::leak(Box::new(Node::new(item))).into(); + let node_next = match self.current { + None => self.list.head, + Some(node) => node.as_ref().next, + }; + self.list.splice_nodes(self.current, node_next, spliced_node, spliced_node, 1); + if self.current.is_none() { + // The "ghost" non-element's index has changed. + self.index = self.list.len; + } + } + } + + /// Inserts a new element into the `LinkedList` before the current one. + /// + /// If the cursor is pointing at the "ghost" non-element then the new element is + /// inserted at the end of the `LinkedList`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn insert_before(&mut self, item: T) { + unsafe { + let spliced_node = Box::leak(Box::new(Node::new(item))).into(); + let node_prev = match self.current { + None => self.list.tail, + Some(node) => node.as_ref().prev, + }; + self.list.splice_nodes(node_prev, self.current, spliced_node, spliced_node, 1); + self.index += 1; + } + } + + /// Removes the current element from the `LinkedList`. + /// + /// The element that was removed is returned, and the cursor is + /// moved to point to the next element in the `LinkedList`. + /// + /// If the cursor is currently pointing to the "ghost" non-element then no element + /// is removed and `None` is returned. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn remove_current(&mut self) -> Option<T> { + let unlinked_node = self.current?; + unsafe { + self.current = unlinked_node.as_ref().next; + self.list.unlink_node(unlinked_node); + let unlinked_node = Box::from_raw(unlinked_node.as_ptr()); + Some(unlinked_node.element) + } + } + + /// Removes the current element from the `LinkedList` without deallocating the list node. + /// + /// The node that was removed is returned as a new `LinkedList` containing only this node. + /// The cursor is moved to point to the next element in the current `LinkedList`. + /// + /// If the cursor is currently pointing to the "ghost" non-element then no element + /// is removed and `None` is returned. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn remove_current_as_list(&mut self) -> Option<LinkedList<T>> { + let mut unlinked_node = self.current?; + unsafe { + self.current = unlinked_node.as_ref().next; + self.list.unlink_node(unlinked_node); + + unlinked_node.as_mut().prev = None; + unlinked_node.as_mut().next = None; + Some(LinkedList { + head: Some(unlinked_node), + tail: Some(unlinked_node), + len: 1, + marker: PhantomData, + }) + } + } + + /// Inserts the elements from the given `LinkedList` after the current one. + /// + /// If the cursor is pointing at the "ghost" non-element then the new elements are + /// inserted at the start of the `LinkedList`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn splice_after(&mut self, list: LinkedList<T>) { + unsafe { + let (splice_head, splice_tail, splice_len) = match list.detach_all_nodes() { + Some(parts) => parts, + _ => return, + }; + let node_next = match self.current { + None => self.list.head, + Some(node) => node.as_ref().next, + }; + self.list.splice_nodes(self.current, node_next, splice_head, splice_tail, splice_len); + if self.current.is_none() { + // The "ghost" non-element's index has changed. + self.index = self.list.len; + } + } + } + + /// Inserts the elements from the given `LinkedList` before the current one. + /// + /// If the cursor is pointing at the "ghost" non-element then the new elements are + /// inserted at the end of the `LinkedList`. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn splice_before(&mut self, list: LinkedList<T>) { + unsafe { + let (splice_head, splice_tail, splice_len) = match list.detach_all_nodes() { + Some(parts) => parts, + _ => return, + }; + let node_prev = match self.current { + None => self.list.tail, + Some(node) => node.as_ref().prev, + }; + self.list.splice_nodes(node_prev, self.current, splice_head, splice_tail, splice_len); + self.index += splice_len; + } + } + + /// Splits the list into two after the current element. This will return a + /// new list consisting of everything after the cursor, with the original + /// list retaining everything before. + /// + /// If the cursor is pointing at the "ghost" non-element then the entire contents + /// of the `LinkedList` are moved. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn split_after(&mut self) -> LinkedList<T> { + let split_off_idx = if self.index == self.list.len { 0 } else { self.index + 1 }; + if self.index == self.list.len { + // The "ghost" non-element's index has changed to 0. + self.index = 0; + } + unsafe { self.list.split_off_after_node(self.current, split_off_idx) } + } + + /// Splits the list into two before the current element. This will return a + /// new list consisting of everything before the cursor, with the original + /// list retaining everything after. + /// + /// If the cursor is pointing at the "ghost" non-element then the entire contents + /// of the `LinkedList` are moved. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn split_before(&mut self) -> LinkedList<T> { + let split_off_idx = self.index; + self.index = 0; + unsafe { self.list.split_off_before_node(self.current, split_off_idx) } + } + + /// Appends an element to the front of the cursor's parent list. The node + /// that the cursor points to is unchanged, even if it is the "ghost" node. + /// + /// This operation should compute in *O*(1) time. + // `push_front` continues to point to "ghost" when it addes a node to mimic + // the behavior of `insert_before` on an empty list. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn push_front(&mut self, elt: T) { + // Safety: We know that `push_front` does not change the position in + // memory of other nodes. This ensures that `self.current` remains + // valid. + self.list.push_front(elt); + self.index += 1; + } + + /// Appends an element to the back of the cursor's parent list. The node + /// that the cursor points to is unchanged, even if it is the "ghost" node. + /// + /// This operation should compute in *O*(1) time. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn push_back(&mut self, elt: T) { + // Safety: We know that `push_back` does not change the position in + // memory of other nodes. This ensures that `self.current` remains + // valid. + self.list.push_back(elt); + if self.current().is_none() { + // The index of "ghost" is the length of the list, so we just need + // to increment self.index to reflect the new length of the list. + self.index += 1; + } + } + + /// Removes the first element from the cursor's parent list and returns it, + /// or None if the list is empty. The element the cursor points to remains + /// unchanged, unless it was pointing to the front element. In that case, it + /// points to the new front element. + /// + /// This operation should compute in *O*(1) time. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn pop_front(&mut self) -> Option<T> { + // We can't check if current is empty, we must check the list directly. + // It is possible for `self.current == None` and the list to be + // non-empty. + if self.list.is_empty() { + None + } else { + // We can't point to the node that we pop. Copying the behavior of + // `remove_current`, we move on the the next node in the sequence. + // If the list is of length 1 then we end pointing to the "ghost" + // node at index 0, which is expected. + if self.list.head == self.current { + self.move_next(); + } else { + self.index -= 1; + } + self.list.pop_front() + } + } + + /// Removes the last element from the cursor's parent list and returns it, + /// or None if the list is empty. The element the cursor points to remains + /// unchanged, unless it was pointing to the back element. In that case, it + /// points to the "ghost" element. + /// + /// This operation should compute in *O*(1) time. + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn pop_back(&mut self) -> Option<T> { + if self.list.is_empty() { + None + } else { + if self.list.tail == self.current { + // The index now reflects the length of the list. It was the + // length of the list minus 1, but now the list is 1 smaller. No + // change is needed for `index`. + self.current = None; + } else if self.current.is_none() { + self.index = self.list.len - 1; + } + self.list.pop_back() + } + } + + /// Provides a reference to the front element of the cursor's parent list, + /// or None if the list is empty. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn front(&self) -> Option<&T> { + self.list.front() + } + + /// Provides a mutable reference to the front element of the cursor's + /// parent list, or None if the list is empty. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn front_mut(&mut self) -> Option<&mut T> { + self.list.front_mut() + } + + /// Provides a reference to the back element of the cursor's parent list, + /// or None if the list is empty. + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn back(&self) -> Option<&T> { + self.list.back() + } + + /// Provides a mutable reference to back element of the cursor's parent + /// list, or `None` if the list is empty. + /// + /// # Examples + /// Building and mutating a list with a cursor, then getting the back element: + /// ``` + /// #![feature(linked_list_cursors)] + /// use std::collections::LinkedList; + /// let mut dl = LinkedList::new(); + /// dl.push_front(3); + /// dl.push_front(2); + /// dl.push_front(1); + /// let mut cursor = dl.cursor_front_mut(); + /// *cursor.current().unwrap() = 99; + /// *cursor.back_mut().unwrap() = 0; + /// let mut contents = dl.into_iter(); + /// assert_eq!(contents.next(), Some(99)); + /// assert_eq!(contents.next(), Some(2)); + /// assert_eq!(contents.next(), Some(0)); + /// assert_eq!(contents.next(), None); + /// ``` + #[must_use] + #[unstable(feature = "linked_list_cursors", issue = "58533")] + pub fn back_mut(&mut self) -> Option<&mut T> { + self.list.back_mut() + } +} + +/// An iterator produced by calling `drain_filter` on LinkedList. +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +pub struct DrainFilter<'a, T: 'a, F: 'a> +where + F: FnMut(&mut T) -> bool, +{ + list: &'a mut LinkedList<T>, + it: Option<NonNull<Node<T>>>, + pred: F, + idx: usize, + old_len: usize, +} + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +impl<T, F> Iterator for DrainFilter<'_, T, F> +where + F: FnMut(&mut T) -> bool, +{ + type Item = T; + + fn next(&mut self) -> Option<T> { + while let Some(mut node) = self.it { + unsafe { + self.it = node.as_ref().next; + self.idx += 1; + + if (self.pred)(&mut node.as_mut().element) { + // `unlink_node` is okay with aliasing `element` references. + self.list.unlink_node(node); + return Some(Box::from_raw(node.as_ptr()).element); + } + } + } + + None + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (0, Some(self.old_len - self.idx)) + } +} + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +impl<T, F> Drop for DrainFilter<'_, T, F> +where + F: FnMut(&mut T) -> bool, +{ + fn drop(&mut self) { + struct DropGuard<'r, 'a, T, F>(&'r mut DrainFilter<'a, T, F>) + where + F: FnMut(&mut T) -> bool; + + impl<'r, 'a, T, F> Drop for DropGuard<'r, 'a, T, F> + where + F: FnMut(&mut T) -> bool, + { + fn drop(&mut self) { + self.0.for_each(drop); + } + } + + while let Some(item) = self.next() { + let guard = DropGuard(self); + drop(item); + mem::forget(guard); + } + } +} + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +impl<T: fmt::Debug, F> fmt::Debug for DrainFilter<'_, T, F> +where + F: FnMut(&mut T) -> bool, +{ + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("DrainFilter").field(&self.list).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Iterator for IntoIter<T> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.list.pop_front() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + (self.list.len, Some(self.list.len)) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> DoubleEndedIterator for IntoIter<T> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.list.pop_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for IntoIter<T> {} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for IntoIter<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> FromIterator<T> for LinkedList<T> { + fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> Self { + let mut list = Self::new(); + list.extend(iter); + list + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> IntoIterator for LinkedList<T> { + type Item = T; + type IntoIter = IntoIter<T>; + + /// Consumes the list into an iterator yielding elements by value. + #[inline] + fn into_iter(self) -> IntoIter<T> { + IntoIter { list: self } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> IntoIterator for &'a LinkedList<T> { + type Item = &'a T; + type IntoIter = Iter<'a, T>; + + fn into_iter(self) -> Iter<'a, T> { + self.iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> IntoIterator for &'a mut LinkedList<T> { + type Item = &'a mut T; + type IntoIter = IterMut<'a, T>; + + fn into_iter(self) -> IterMut<'a, T> { + self.iter_mut() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Extend<T> for LinkedList<T> { + fn extend<I: IntoIterator<Item = T>>(&mut self, iter: I) { + <Self as SpecExtend<I>>::spec_extend(self, iter); + } + + #[inline] + fn extend_one(&mut self, elem: T) { + self.push_back(elem); + } +} + +impl<I: IntoIterator> SpecExtend<I> for LinkedList<I::Item> { + default fn spec_extend(&mut self, iter: I) { + iter.into_iter().for_each(move |elt| self.push_back(elt)); + } +} + +impl<T> SpecExtend<LinkedList<T>> for LinkedList<T> { + fn spec_extend(&mut self, ref mut other: LinkedList<T>) { + self.append(other); + } +} + +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, T: 'a + Copy> Extend<&'a T> for LinkedList<T> { + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iter: I) { + self.extend(iter.into_iter().cloned()); + } + + #[inline] + fn extend_one(&mut self, &elem: &'a T) { + self.push_back(elem); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialEq> PartialEq for LinkedList<T> { + fn eq(&self, other: &Self) -> bool { + self.len() == other.len() && self.iter().eq(other) + } + + fn ne(&self, other: &Self) -> bool { + self.len() != other.len() || self.iter().ne(other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Eq> Eq for LinkedList<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialOrd> PartialOrd for LinkedList<T> { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + self.iter().partial_cmp(other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord> Ord for LinkedList<T> { + #[inline] + fn cmp(&self, other: &Self) -> Ordering { + self.iter().cmp(other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone> Clone for LinkedList<T> { + fn clone(&self) -> Self { + self.iter().cloned().collect() + } + + fn clone_from(&mut self, other: &Self) { + let mut iter_other = other.iter(); + if self.len() > other.len() { + self.split_off(other.len()); + } + for (elem, elem_other) in self.iter_mut().zip(&mut iter_other) { + elem.clone_from(elem_other); + } + if !iter_other.is_empty() { + self.extend(iter_other.cloned()); + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Debug> fmt::Debug for LinkedList<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Hash> Hash for LinkedList<T> { + fn hash<H: Hasher>(&self, state: &mut H) { + state.write_length_prefix(self.len()); + for elt in self { + elt.hash(state); + } + } +} + +#[stable(feature = "std_collections_from_array", since = "1.56.0")] +impl<T, const N: usize> From<[T; N]> for LinkedList<T> { + /// Converts a `[T; N]` into a `LinkedList<T>`. + /// + /// ``` + /// use std::collections::LinkedList; + /// + /// let list1 = LinkedList::from([1, 2, 3, 4]); + /// let list2: LinkedList<_> = [1, 2, 3, 4].into(); + /// assert_eq!(list1, list2); + /// ``` + fn from(arr: [T; N]) -> Self { + Self::from_iter(arr) + } +} + +// Ensure that `LinkedList` and its read-only iterators are covariant in their type parameters. +#[allow(dead_code)] +fn assert_covariance() { + fn a<'a>(x: LinkedList<&'static str>) -> LinkedList<&'a str> { + x + } + fn b<'i, 'a>(x: Iter<'i, &'static str>) -> Iter<'i, &'a str> { + x + } + fn c<'a>(x: IntoIter<&'static str>) -> IntoIter<&'a str> { + x + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Send> Send for LinkedList<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync> Sync for LinkedList<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync> Send for Iter<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync> Sync for Iter<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Send> Send for IterMut<'_, T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync> Sync for IterMut<'_, T> {} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +unsafe impl<T: Sync> Send for Cursor<'_, T> {} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +unsafe impl<T: Sync> Sync for Cursor<'_, T> {} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +unsafe impl<T: Send> Send for CursorMut<'_, T> {} + +#[unstable(feature = "linked_list_cursors", issue = "58533")] +unsafe impl<T: Sync> Sync for CursorMut<'_, T> {} diff --git a/library/alloc/src/collections/linked_list/tests.rs b/library/alloc/src/collections/linked_list/tests.rs new file mode 100644 index 000000000..f8fbfa1bf --- /dev/null +++ b/library/alloc/src/collections/linked_list/tests.rs @@ -0,0 +1,1156 @@ +use super::*; +use crate::vec::Vec; + +use std::panic::{catch_unwind, AssertUnwindSafe}; +use std::thread; + +use rand::{thread_rng, RngCore}; + +#[test] +fn test_basic() { + let mut m = LinkedList::<Box<_>>::new(); + assert_eq!(m.pop_front(), None); + assert_eq!(m.pop_back(), None); + assert_eq!(m.pop_front(), None); + m.push_front(Box::new(1)); + assert_eq!(m.pop_front(), Some(Box::new(1))); + m.push_back(Box::new(2)); + m.push_back(Box::new(3)); + assert_eq!(m.len(), 2); + assert_eq!(m.pop_front(), Some(Box::new(2))); + assert_eq!(m.pop_front(), Some(Box::new(3))); + assert_eq!(m.len(), 0); + assert_eq!(m.pop_front(), None); + m.push_back(Box::new(1)); + m.push_back(Box::new(3)); + m.push_back(Box::new(5)); + m.push_back(Box::new(7)); + assert_eq!(m.pop_front(), Some(Box::new(1))); + + let mut n = LinkedList::new(); + n.push_front(2); + n.push_front(3); + { + assert_eq!(n.front().unwrap(), &3); + let x = n.front_mut().unwrap(); + assert_eq!(*x, 3); + *x = 0; + } + { + assert_eq!(n.back().unwrap(), &2); + let y = n.back_mut().unwrap(); + assert_eq!(*y, 2); + *y = 1; + } + assert_eq!(n.pop_front(), Some(0)); + assert_eq!(n.pop_front(), Some(1)); +} + +fn generate_test() -> LinkedList<i32> { + list_from(&[0, 1, 2, 3, 4, 5, 6]) +} + +fn list_from<T: Clone>(v: &[T]) -> LinkedList<T> { + v.iter().cloned().collect() +} + +pub fn check_links<T>(list: &LinkedList<T>) { + unsafe { + let mut len = 0; + let mut last_ptr: Option<&Node<T>> = None; + let mut node_ptr: &Node<T>; + match list.head { + None => { + // tail node should also be None. + assert!(list.tail.is_none()); + assert_eq!(0, list.len); + return; + } + Some(node) => node_ptr = &*node.as_ptr(), + } + loop { + match (last_ptr, node_ptr.prev) { + (None, None) => {} + (None, _) => panic!("prev link for head"), + (Some(p), Some(pptr)) => { + assert_eq!(p as *const Node<T>, pptr.as_ptr() as *const Node<T>); + } + _ => panic!("prev link is none, not good"), + } + match node_ptr.next { + Some(next) => { + last_ptr = Some(node_ptr); + node_ptr = &*next.as_ptr(); + len += 1; + } + None => { + len += 1; + break; + } + } + } + + // verify that the tail node points to the last node. + let tail = list.tail.as_ref().expect("some tail node").as_ref(); + assert_eq!(tail as *const Node<T>, node_ptr as *const Node<T>); + // check that len matches interior links. + assert_eq!(len, list.len); + } +} + +#[test] +fn test_append() { + // Empty to empty + { + let mut m = LinkedList::<i32>::new(); + let mut n = LinkedList::new(); + m.append(&mut n); + check_links(&m); + assert_eq!(m.len(), 0); + assert_eq!(n.len(), 0); + } + // Non-empty to empty + { + let mut m = LinkedList::new(); + let mut n = LinkedList::new(); + n.push_back(2); + m.append(&mut n); + check_links(&m); + assert_eq!(m.len(), 1); + assert_eq!(m.pop_back(), Some(2)); + assert_eq!(n.len(), 0); + check_links(&m); + } + // Empty to non-empty + { + let mut m = LinkedList::new(); + let mut n = LinkedList::new(); + m.push_back(2); + m.append(&mut n); + check_links(&m); + assert_eq!(m.len(), 1); + assert_eq!(m.pop_back(), Some(2)); + check_links(&m); + } + + // Non-empty to non-empty + let v = vec![1, 2, 3, 4, 5]; + let u = vec![9, 8, 1, 2, 3, 4, 5]; + let mut m = list_from(&v); + let mut n = list_from(&u); + m.append(&mut n); + check_links(&m); + let mut sum = v; + sum.extend_from_slice(&u); + assert_eq!(sum.len(), m.len()); + for elt in sum { + assert_eq!(m.pop_front(), Some(elt)) + } + assert_eq!(n.len(), 0); + // Let's make sure it's working properly, since we + // did some direct changes to private members. + n.push_back(3); + assert_eq!(n.len(), 1); + assert_eq!(n.pop_front(), Some(3)); + check_links(&n); +} + +#[test] +fn test_iterator() { + let m = generate_test(); + for (i, elt) in m.iter().enumerate() { + assert_eq!(i as i32, *elt); + } + let mut n = LinkedList::new(); + assert_eq!(n.iter().next(), None); + n.push_front(4); + let mut it = n.iter(); + assert_eq!(it.size_hint(), (1, Some(1))); + assert_eq!(it.next().unwrap(), &4); + assert_eq!(it.size_hint(), (0, Some(0))); + assert_eq!(it.next(), None); +} + +#[test] +fn test_iterator_clone() { + let mut n = LinkedList::new(); + n.push_back(2); + n.push_back(3); + n.push_back(4); + let mut it = n.iter(); + it.next(); + let mut jt = it.clone(); + assert_eq!(it.next(), jt.next()); + assert_eq!(it.next_back(), jt.next_back()); + assert_eq!(it.next(), jt.next()); +} + +#[test] +fn test_iterator_double_end() { + let mut n = LinkedList::new(); + assert_eq!(n.iter().next(), None); + n.push_front(4); + n.push_front(5); + n.push_front(6); + let mut it = n.iter(); + assert_eq!(it.size_hint(), (3, Some(3))); + assert_eq!(it.next().unwrap(), &6); + assert_eq!(it.size_hint(), (2, Some(2))); + assert_eq!(it.next_back().unwrap(), &4); + assert_eq!(it.size_hint(), (1, Some(1))); + assert_eq!(it.next_back().unwrap(), &5); + assert_eq!(it.next_back(), None); + assert_eq!(it.next(), None); +} + +#[test] +fn test_rev_iter() { + let m = generate_test(); + for (i, elt) in m.iter().rev().enumerate() { + assert_eq!((6 - i) as i32, *elt); + } + let mut n = LinkedList::new(); + assert_eq!(n.iter().rev().next(), None); + n.push_front(4); + let mut it = n.iter().rev(); + assert_eq!(it.size_hint(), (1, Some(1))); + assert_eq!(it.next().unwrap(), &4); + assert_eq!(it.size_hint(), (0, Some(0))); + assert_eq!(it.next(), None); +} + +#[test] +fn test_mut_iter() { + let mut m = generate_test(); + let mut len = m.len(); + for (i, elt) in m.iter_mut().enumerate() { + assert_eq!(i as i32, *elt); + len -= 1; + } + assert_eq!(len, 0); + let mut n = LinkedList::new(); + assert!(n.iter_mut().next().is_none()); + n.push_front(4); + n.push_back(5); + let mut it = n.iter_mut(); + assert_eq!(it.size_hint(), (2, Some(2))); + assert!(it.next().is_some()); + assert!(it.next().is_some()); + assert_eq!(it.size_hint(), (0, Some(0))); + assert!(it.next().is_none()); +} + +#[test] +fn test_iterator_mut_double_end() { + let mut n = LinkedList::new(); + assert!(n.iter_mut().next_back().is_none()); + n.push_front(4); + n.push_front(5); + n.push_front(6); + let mut it = n.iter_mut(); + assert_eq!(it.size_hint(), (3, Some(3))); + assert_eq!(*it.next().unwrap(), 6); + assert_eq!(it.size_hint(), (2, Some(2))); + assert_eq!(*it.next_back().unwrap(), 4); + assert_eq!(it.size_hint(), (1, Some(1))); + assert_eq!(*it.next_back().unwrap(), 5); + assert!(it.next_back().is_none()); + assert!(it.next().is_none()); +} + +#[test] +fn test_mut_rev_iter() { + let mut m = generate_test(); + for (i, elt) in m.iter_mut().rev().enumerate() { + assert_eq!((6 - i) as i32, *elt); + } + let mut n = LinkedList::new(); + assert!(n.iter_mut().rev().next().is_none()); + n.push_front(4); + let mut it = n.iter_mut().rev(); + assert!(it.next().is_some()); + assert!(it.next().is_none()); +} + +#[test] +fn test_clone_from() { + // Short cloned from long + { + let v = vec![1, 2, 3, 4, 5]; + let u = vec![8, 7, 6, 2, 3, 4, 5]; + let mut m = list_from(&v); + let n = list_from(&u); + m.clone_from(&n); + check_links(&m); + assert_eq!(m, n); + for elt in u { + assert_eq!(m.pop_front(), Some(elt)) + } + } + // Long cloned from short + { + let v = vec![1, 2, 3, 4, 5]; + let u = vec![6, 7, 8]; + let mut m = list_from(&v); + let n = list_from(&u); + m.clone_from(&n); + check_links(&m); + assert_eq!(m, n); + for elt in u { + assert_eq!(m.pop_front(), Some(elt)) + } + } + // Two equal length lists + { + let v = vec![1, 2, 3, 4, 5]; + let u = vec![9, 8, 1, 2, 3]; + let mut m = list_from(&v); + let n = list_from(&u); + m.clone_from(&n); + check_links(&m); + assert_eq!(m, n); + for elt in u { + assert_eq!(m.pop_front(), Some(elt)) + } + } +} + +#[test] +#[cfg_attr(target_os = "emscripten", ignore)] +fn test_send() { + let n = list_from(&[1, 2, 3]); + thread::spawn(move || { + check_links(&n); + let a: &[_] = &[&1, &2, &3]; + assert_eq!(a, &*n.iter().collect::<Vec<_>>()); + }) + .join() + .ok() + .unwrap(); +} + +#[test] +fn test_eq() { + let mut n = list_from(&[]); + let mut m = list_from(&[]); + assert!(n == m); + n.push_front(1); + assert!(n != m); + m.push_back(1); + assert!(n == m); + + let n = list_from(&[2, 3, 4]); + let m = list_from(&[1, 2, 3]); + assert!(n != m); +} + +#[test] +fn test_ord() { + let n = list_from(&[]); + let m = list_from(&[1, 2, 3]); + assert!(n < m); + assert!(m > n); + assert!(n <= n); + assert!(n >= n); +} + +#[test] +fn test_ord_nan() { + let nan = 0.0f64 / 0.0; + let n = list_from(&[nan]); + let m = list_from(&[nan]); + assert!(!(n < m)); + assert!(!(n > m)); + assert!(!(n <= m)); + assert!(!(n >= m)); + + let n = list_from(&[nan]); + let one = list_from(&[1.0f64]); + assert!(!(n < one)); + assert!(!(n > one)); + assert!(!(n <= one)); + assert!(!(n >= one)); + + let u = list_from(&[1.0f64, 2.0, nan]); + let v = list_from(&[1.0f64, 2.0, 3.0]); + assert!(!(u < v)); + assert!(!(u > v)); + assert!(!(u <= v)); + assert!(!(u >= v)); + + let s = list_from(&[1.0f64, 2.0, 4.0, 2.0]); + let t = list_from(&[1.0f64, 2.0, 3.0, 2.0]); + assert!(!(s < t)); + assert!(s > one); + assert!(!(s <= one)); + assert!(s >= one); +} + +#[test] +fn test_26021() { + // There was a bug in split_off that failed to null out the RHS's head's prev ptr. + // This caused the RHS's dtor to walk up into the LHS at drop and delete all of + // its nodes. + // + // https://github.com/rust-lang/rust/issues/26021 + let mut v1 = LinkedList::new(); + v1.push_front(1); + v1.push_front(1); + v1.push_front(1); + v1.push_front(1); + let _ = v1.split_off(3); // Dropping this now should not cause laundry consumption + assert_eq!(v1.len(), 3); + + assert_eq!(v1.iter().len(), 3); + assert_eq!(v1.iter().collect::<Vec<_>>().len(), 3); +} + +#[test] +fn test_split_off() { + let mut v1 = LinkedList::new(); + v1.push_front(1); + v1.push_front(1); + v1.push_front(1); + v1.push_front(1); + + // test all splits + for ix in 0..1 + v1.len() { + let mut a = v1.clone(); + let b = a.split_off(ix); + check_links(&a); + check_links(&b); + a.extend(b); + assert_eq!(v1, a); + } +} + +#[test] +fn test_split_off_2() { + // singleton + { + let mut m = LinkedList::new(); + m.push_back(1); + + let p = m.split_off(0); + assert_eq!(m.len(), 0); + assert_eq!(p.len(), 1); + assert_eq!(p.back(), Some(&1)); + assert_eq!(p.front(), Some(&1)); + } + + // not singleton, forwards + { + let u = vec![1, 2, 3, 4, 5]; + let mut m = list_from(&u); + let mut n = m.split_off(2); + assert_eq!(m.len(), 2); + assert_eq!(n.len(), 3); + for elt in 1..3 { + assert_eq!(m.pop_front(), Some(elt)); + } + for elt in 3..6 { + assert_eq!(n.pop_front(), Some(elt)); + } + } + // not singleton, backwards + { + let u = vec![1, 2, 3, 4, 5]; + let mut m = list_from(&u); + let mut n = m.split_off(4); + assert_eq!(m.len(), 4); + assert_eq!(n.len(), 1); + for elt in 1..5 { + assert_eq!(m.pop_front(), Some(elt)); + } + for elt in 5..6 { + assert_eq!(n.pop_front(), Some(elt)); + } + } + + // no-op on the last index + { + let mut m = LinkedList::new(); + m.push_back(1); + + let p = m.split_off(1); + assert_eq!(m.len(), 1); + assert_eq!(p.len(), 0); + assert_eq!(m.back(), Some(&1)); + assert_eq!(m.front(), Some(&1)); + } +} + +fn fuzz_test(sz: i32) { + let mut m: LinkedList<_> = LinkedList::new(); + let mut v = vec![]; + for i in 0..sz { + check_links(&m); + let r: u8 = thread_rng().next_u32() as u8; + match r % 6 { + 0 => { + m.pop_back(); + v.pop(); + } + 1 => { + if !v.is_empty() { + m.pop_front(); + v.remove(0); + } + } + 2 | 4 => { + m.push_front(-i); + v.insert(0, -i); + } + 3 | 5 | _ => { + m.push_back(i); + v.push(i); + } + } + } + + check_links(&m); + + let mut i = 0; + for (a, &b) in m.into_iter().zip(&v) { + i += 1; + assert_eq!(a, b); + } + assert_eq!(i, v.len()); +} + +#[test] +fn test_fuzz() { + for _ in 0..25 { + fuzz_test(3); + fuzz_test(16); + #[cfg(not(miri))] // Miri is too slow + fuzz_test(189); + } +} + +#[test] +fn test_show() { + let list: LinkedList<_> = (0..10).collect(); + assert_eq!(format!("{list:?}"), "[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]"); + + let list: LinkedList<_> = ["just", "one", "test", "more"].into_iter().collect(); + assert_eq!(format!("{list:?}"), "[\"just\", \"one\", \"test\", \"more\"]"); +} + +#[test] +fn drain_filter_test() { + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let deleted = m.drain_filter(|v| *v < 4).collect::<Vec<_>>(); + + check_links(&m); + + assert_eq!(deleted, &[1, 2, 3]); + assert_eq!(m.into_iter().collect::<Vec<_>>(), &[4, 5, 6]); +} + +#[test] +fn drain_to_empty_test() { + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let deleted = m.drain_filter(|_| true).collect::<Vec<_>>(); + + check_links(&m); + + assert_eq!(deleted, &[1, 2, 3, 4, 5, 6]); + assert_eq!(m.into_iter().collect::<Vec<_>>(), &[]); +} + +#[test] +fn test_cursor_move_peek() { + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_front(); + assert_eq!(cursor.current(), Some(&1)); + assert_eq!(cursor.peek_next(), Some(&2)); + assert_eq!(cursor.peek_prev(), None); + assert_eq!(cursor.index(), Some(0)); + cursor.move_prev(); + assert_eq!(cursor.current(), None); + assert_eq!(cursor.peek_next(), Some(&1)); + assert_eq!(cursor.peek_prev(), Some(&6)); + assert_eq!(cursor.index(), None); + cursor.move_next(); + cursor.move_next(); + assert_eq!(cursor.current(), Some(&2)); + assert_eq!(cursor.peek_next(), Some(&3)); + assert_eq!(cursor.peek_prev(), Some(&1)); + assert_eq!(cursor.index(), Some(1)); + + let mut cursor = m.cursor_back(); + assert_eq!(cursor.current(), Some(&6)); + assert_eq!(cursor.peek_next(), None); + assert_eq!(cursor.peek_prev(), Some(&5)); + assert_eq!(cursor.index(), Some(5)); + cursor.move_next(); + assert_eq!(cursor.current(), None); + assert_eq!(cursor.peek_next(), Some(&1)); + assert_eq!(cursor.peek_prev(), Some(&6)); + assert_eq!(cursor.index(), None); + cursor.move_prev(); + cursor.move_prev(); + assert_eq!(cursor.current(), Some(&5)); + assert_eq!(cursor.peek_next(), Some(&6)); + assert_eq!(cursor.peek_prev(), Some(&4)); + assert_eq!(cursor.index(), Some(4)); + + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_front_mut(); + assert_eq!(cursor.current(), Some(&mut 1)); + assert_eq!(cursor.peek_next(), Some(&mut 2)); + assert_eq!(cursor.peek_prev(), None); + assert_eq!(cursor.index(), Some(0)); + cursor.move_prev(); + assert_eq!(cursor.current(), None); + assert_eq!(cursor.peek_next(), Some(&mut 1)); + assert_eq!(cursor.peek_prev(), Some(&mut 6)); + assert_eq!(cursor.index(), None); + cursor.move_next(); + cursor.move_next(); + assert_eq!(cursor.current(), Some(&mut 2)); + assert_eq!(cursor.peek_next(), Some(&mut 3)); + assert_eq!(cursor.peek_prev(), Some(&mut 1)); + assert_eq!(cursor.index(), Some(1)); + let mut cursor2 = cursor.as_cursor(); + assert_eq!(cursor2.current(), Some(&2)); + assert_eq!(cursor2.index(), Some(1)); + cursor2.move_next(); + assert_eq!(cursor2.current(), Some(&3)); + assert_eq!(cursor2.index(), Some(2)); + assert_eq!(cursor.current(), Some(&mut 2)); + assert_eq!(cursor.index(), Some(1)); + + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_back_mut(); + assert_eq!(cursor.current(), Some(&mut 6)); + assert_eq!(cursor.peek_next(), None); + assert_eq!(cursor.peek_prev(), Some(&mut 5)); + assert_eq!(cursor.index(), Some(5)); + cursor.move_next(); + assert_eq!(cursor.current(), None); + assert_eq!(cursor.peek_next(), Some(&mut 1)); + assert_eq!(cursor.peek_prev(), Some(&mut 6)); + assert_eq!(cursor.index(), None); + cursor.move_prev(); + cursor.move_prev(); + assert_eq!(cursor.current(), Some(&mut 5)); + assert_eq!(cursor.peek_next(), Some(&mut 6)); + assert_eq!(cursor.peek_prev(), Some(&mut 4)); + assert_eq!(cursor.index(), Some(4)); + let mut cursor2 = cursor.as_cursor(); + assert_eq!(cursor2.current(), Some(&5)); + assert_eq!(cursor2.index(), Some(4)); + cursor2.move_prev(); + assert_eq!(cursor2.current(), Some(&4)); + assert_eq!(cursor2.index(), Some(3)); + assert_eq!(cursor.current(), Some(&mut 5)); + assert_eq!(cursor.index(), Some(4)); +} + +#[test] +fn test_cursor_mut_insert() { + let mut m: LinkedList<u32> = LinkedList::new(); + m.extend(&[1, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_front_mut(); + cursor.insert_before(7); + cursor.insert_after(8); + check_links(&m); + assert_eq!(m.iter().cloned().collect::<Vec<_>>(), &[7, 1, 8, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_front_mut(); + cursor.move_prev(); + cursor.insert_before(9); + cursor.insert_after(10); + check_links(&m); + assert_eq!(m.iter().cloned().collect::<Vec<_>>(), &[10, 7, 1, 8, 2, 3, 4, 5, 6, 9]); + let mut cursor = m.cursor_front_mut(); + cursor.move_prev(); + assert_eq!(cursor.remove_current(), None); + cursor.move_next(); + cursor.move_next(); + assert_eq!(cursor.remove_current(), Some(7)); + cursor.move_prev(); + cursor.move_prev(); + cursor.move_prev(); + assert_eq!(cursor.remove_current(), Some(9)); + cursor.move_next(); + assert_eq!(cursor.remove_current(), Some(10)); + check_links(&m); + assert_eq!(m.iter().cloned().collect::<Vec<_>>(), &[1, 8, 2, 3, 4, 5, 6]); + let mut cursor = m.cursor_front_mut(); + let mut p: LinkedList<u32> = LinkedList::new(); + p.extend(&[100, 101, 102, 103]); + let mut q: LinkedList<u32> = LinkedList::new(); + q.extend(&[200, 201, 202, 203]); + cursor.splice_after(p); + cursor.splice_before(q); + check_links(&m); + assert_eq!( + m.iter().cloned().collect::<Vec<_>>(), + &[200, 201, 202, 203, 1, 100, 101, 102, 103, 8, 2, 3, 4, 5, 6] + ); + let mut cursor = m.cursor_front_mut(); + cursor.move_prev(); + let tmp = cursor.split_before(); + assert_eq!(m.into_iter().collect::<Vec<_>>(), &[]); + m = tmp; + let mut cursor = m.cursor_front_mut(); + cursor.move_next(); + cursor.move_next(); + cursor.move_next(); + cursor.move_next(); + cursor.move_next(); + cursor.move_next(); + let tmp = cursor.split_after(); + assert_eq!(tmp.into_iter().collect::<Vec<_>>(), &[102, 103, 8, 2, 3, 4, 5, 6]); + check_links(&m); + assert_eq!(m.iter().cloned().collect::<Vec<_>>(), &[200, 201, 202, 203, 1, 100, 101]); +} + +#[test] +fn test_cursor_push_front_back() { + let mut ll: LinkedList<u32> = LinkedList::new(); + ll.extend(&[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]); + let mut c = ll.cursor_front_mut(); + assert_eq!(c.current(), Some(&mut 1)); + assert_eq!(c.index(), Some(0)); + c.push_front(0); + assert_eq!(c.current(), Some(&mut 1)); + assert_eq!(c.peek_prev(), Some(&mut 0)); + assert_eq!(c.index(), Some(1)); + c.push_back(11); + drop(c); + let p = ll.cursor_back().front().unwrap(); + assert_eq!(p, &0); + assert_eq!(ll, (0..12).collect()); + check_links(&ll); +} + +#[test] +fn test_cursor_pop_front_back() { + let mut ll: LinkedList<u32> = LinkedList::new(); + ll.extend(&[1, 2, 3, 4, 5, 6]); + let mut c = ll.cursor_back_mut(); + assert_eq!(c.pop_front(), Some(1)); + c.move_prev(); + c.move_prev(); + c.move_prev(); + assert_eq!(c.pop_back(), Some(6)); + let c = c.as_cursor(); + assert_eq!(c.front(), Some(&2)); + assert_eq!(c.back(), Some(&5)); + assert_eq!(c.index(), Some(1)); + drop(c); + assert_eq!(ll, (2..6).collect()); + check_links(&ll); + let mut c = ll.cursor_back_mut(); + assert_eq!(c.current(), Some(&mut 5)); + assert_eq!(c.index, 3); + assert_eq!(c.pop_back(), Some(5)); + assert_eq!(c.current(), None); + assert_eq!(c.index, 3); + assert_eq!(c.pop_back(), Some(4)); + assert_eq!(c.current(), None); + assert_eq!(c.index, 2); +} + +#[test] +fn test_extend_ref() { + let mut a = LinkedList::new(); + a.push_back(1); + + a.extend(&[2, 3, 4]); + + assert_eq!(a.len(), 4); + assert_eq!(a, list_from(&[1, 2, 3, 4])); + + let mut b = LinkedList::new(); + b.push_back(5); + b.push_back(6); + a.extend(&b); + + assert_eq!(a.len(), 6); + assert_eq!(a, list_from(&[1, 2, 3, 4, 5, 6])); +} + +#[test] +fn test_extend() { + let mut a = LinkedList::new(); + a.push_back(1); + a.extend(vec![2, 3, 4]); // uses iterator + + assert_eq!(a.len(), 4); + assert!(a.iter().eq(&[1, 2, 3, 4])); + + let b: LinkedList<_> = [5, 6, 7].into_iter().collect(); + a.extend(b); // specializes to `append` + + assert_eq!(a.len(), 7); + assert!(a.iter().eq(&[1, 2, 3, 4, 5, 6, 7])); +} + +#[test] +fn test_contains() { + let mut l = LinkedList::new(); + l.extend(&[2, 3, 4]); + + assert!(l.contains(&3)); + assert!(!l.contains(&1)); + + l.clear(); + + assert!(!l.contains(&3)); +} + +#[test] +fn drain_filter_empty() { + let mut list: LinkedList<i32> = LinkedList::new(); + + { + let mut iter = list.drain_filter(|_| true); + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + assert_eq!(iter.size_hint(), (0, Some(0))); + } + + assert_eq!(list.len(), 0); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![]); +} + +#[test] +fn drain_filter_zst() { + let mut list: LinkedList<_> = [(), (), (), (), ()].into_iter().collect(); + let initial_len = list.len(); + let mut count = 0; + + { + let mut iter = list.drain_filter(|_| true); + assert_eq!(iter.size_hint(), (0, Some(initial_len))); + while let Some(_) = iter.next() { + count += 1; + assert_eq!(iter.size_hint(), (0, Some(initial_len - count))); + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + assert_eq!(iter.size_hint(), (0, Some(0))); + } + + assert_eq!(count, initial_len); + assert_eq!(list.len(), 0); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![]); +} + +#[test] +fn drain_filter_false() { + let mut list: LinkedList<_> = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10].into_iter().collect(); + + let initial_len = list.len(); + let mut count = 0; + + { + let mut iter = list.drain_filter(|_| false); + assert_eq!(iter.size_hint(), (0, Some(initial_len))); + for _ in iter.by_ref() { + count += 1; + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + assert_eq!(iter.size_hint(), (0, Some(0))); + } + + assert_eq!(count, 0); + assert_eq!(list.len(), initial_len); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![1, 2, 3, 4, 5, 6, 7, 8, 9, 10]); +} + +#[test] +fn drain_filter_true() { + let mut list: LinkedList<_> = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10].into_iter().collect(); + + let initial_len = list.len(); + let mut count = 0; + + { + let mut iter = list.drain_filter(|_| true); + assert_eq!(iter.size_hint(), (0, Some(initial_len))); + while let Some(_) = iter.next() { + count += 1; + assert_eq!(iter.size_hint(), (0, Some(initial_len - count))); + } + assert_eq!(iter.size_hint(), (0, Some(0))); + assert_eq!(iter.next(), None); + assert_eq!(iter.size_hint(), (0, Some(0))); + } + + assert_eq!(count, initial_len); + assert_eq!(list.len(), 0); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![]); +} + +#[test] +fn drain_filter_complex() { + { + // [+xxx++++++xxxxx++++x+x++] + let mut list = [ + 1, 2, 4, 6, 7, 9, 11, 13, 15, 17, 18, 20, 22, 24, 26, 27, 29, 31, 33, 34, 35, 36, 37, + 39, + ] + .into_iter() + .collect::<LinkedList<_>>(); + + let removed = list.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + assert_eq!(removed.len(), 10); + assert_eq!(removed, vec![2, 4, 6, 18, 20, 22, 24, 26, 34, 36]); + + assert_eq!(list.len(), 14); + assert_eq!( + list.into_iter().collect::<Vec<_>>(), + vec![1, 7, 9, 11, 13, 15, 17, 27, 29, 31, 33, 35, 37, 39] + ); + } + + { + // [xxx++++++xxxxx++++x+x++] + let mut list = + [2, 4, 6, 7, 9, 11, 13, 15, 17, 18, 20, 22, 24, 26, 27, 29, 31, 33, 34, 35, 36, 37, 39] + .into_iter() + .collect::<LinkedList<_>>(); + + let removed = list.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + assert_eq!(removed.len(), 10); + assert_eq!(removed, vec![2, 4, 6, 18, 20, 22, 24, 26, 34, 36]); + + assert_eq!(list.len(), 13); + assert_eq!( + list.into_iter().collect::<Vec<_>>(), + vec![7, 9, 11, 13, 15, 17, 27, 29, 31, 33, 35, 37, 39] + ); + } + + { + // [xxx++++++xxxxx++++x+x] + let mut list = + [2, 4, 6, 7, 9, 11, 13, 15, 17, 18, 20, 22, 24, 26, 27, 29, 31, 33, 34, 35, 36] + .into_iter() + .collect::<LinkedList<_>>(); + + let removed = list.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + assert_eq!(removed.len(), 10); + assert_eq!(removed, vec![2, 4, 6, 18, 20, 22, 24, 26, 34, 36]); + + assert_eq!(list.len(), 11); + assert_eq!( + list.into_iter().collect::<Vec<_>>(), + vec![7, 9, 11, 13, 15, 17, 27, 29, 31, 33, 35] + ); + } + + { + // [xxxxxxxxxx+++++++++++] + let mut list = [2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19] + .into_iter() + .collect::<LinkedList<_>>(); + + let removed = list.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + assert_eq!(removed.len(), 10); + assert_eq!(removed, vec![2, 4, 6, 8, 10, 12, 14, 16, 18, 20]); + + assert_eq!(list.len(), 10); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![1, 3, 5, 7, 9, 11, 13, 15, 17, 19]); + } + + { + // [+++++++++++xxxxxxxxxx] + let mut list = [1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20] + .into_iter() + .collect::<LinkedList<_>>(); + + let removed = list.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + assert_eq!(removed.len(), 10); + assert_eq!(removed, vec![2, 4, 6, 8, 10, 12, 14, 16, 18, 20]); + + assert_eq!(list.len(), 10); + assert_eq!(list.into_iter().collect::<Vec<_>>(), vec![1, 3, 5, 7, 9, 11, 13, 15, 17, 19]); + } +} + +#[test] +fn drain_filter_drop_panic_leak() { + static mut DROPS: i32 = 0; + + struct D(bool); + + impl Drop for D { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + + if self.0 { + panic!("panic in `drop`"); + } + } + } + + let mut q = LinkedList::new(); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_front(D(false)); + q.push_front(D(true)); + q.push_front(D(false)); + + catch_unwind(AssertUnwindSafe(|| drop(q.drain_filter(|_| true)))).ok(); + + assert_eq!(unsafe { DROPS }, 8); + assert!(q.is_empty()); +} + +#[test] +fn drain_filter_pred_panic_leak() { + static mut DROPS: i32 = 0; + + #[derive(Debug)] + struct D(u32); + + impl Drop for D { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + } + } + + let mut q = LinkedList::new(); + q.push_back(D(3)); + q.push_back(D(4)); + q.push_back(D(5)); + q.push_back(D(6)); + q.push_back(D(7)); + q.push_front(D(2)); + q.push_front(D(1)); + q.push_front(D(0)); + + catch_unwind(AssertUnwindSafe(|| { + drop(q.drain_filter(|item| if item.0 >= 2 { panic!() } else { true })) + })) + .ok(); + + assert_eq!(unsafe { DROPS }, 2); // 0 and 1 + assert_eq!(q.len(), 6); +} + +#[test] +fn test_drop() { + static mut DROPS: i32 = 0; + struct Elem; + impl Drop for Elem { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + } + } + + let mut ring = LinkedList::new(); + ring.push_back(Elem); + ring.push_front(Elem); + ring.push_back(Elem); + ring.push_front(Elem); + drop(ring); + + assert_eq!(unsafe { DROPS }, 4); +} + +#[test] +fn test_drop_with_pop() { + static mut DROPS: i32 = 0; + struct Elem; + impl Drop for Elem { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + } + } + + let mut ring = LinkedList::new(); + ring.push_back(Elem); + ring.push_front(Elem); + ring.push_back(Elem); + ring.push_front(Elem); + + drop(ring.pop_back()); + drop(ring.pop_front()); + assert_eq!(unsafe { DROPS }, 2); + + drop(ring); + assert_eq!(unsafe { DROPS }, 4); +} + +#[test] +fn test_drop_clear() { + static mut DROPS: i32 = 0; + struct Elem; + impl Drop for Elem { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + } + } + + let mut ring = LinkedList::new(); + ring.push_back(Elem); + ring.push_front(Elem); + ring.push_back(Elem); + ring.push_front(Elem); + ring.clear(); + assert_eq!(unsafe { DROPS }, 4); + + drop(ring); + assert_eq!(unsafe { DROPS }, 4); +} + +#[test] +fn test_drop_panic() { + static mut DROPS: i32 = 0; + + struct D(bool); + + impl Drop for D { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + + if self.0 { + panic!("panic in `drop`"); + } + } + } + + let mut q = LinkedList::new(); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_back(D(false)); + q.push_front(D(false)); + q.push_front(D(false)); + q.push_front(D(true)); + + catch_unwind(move || drop(q)).ok(); + + assert_eq!(unsafe { DROPS }, 8); +} diff --git a/library/alloc/src/collections/mod.rs b/library/alloc/src/collections/mod.rs new file mode 100644 index 000000000..628a5b155 --- /dev/null +++ b/library/alloc/src/collections/mod.rs @@ -0,0 +1,154 @@ +//! Collection types. + +#![stable(feature = "rust1", since = "1.0.0")] + +#[cfg(not(no_global_oom_handling))] +pub mod binary_heap; +#[cfg(not(no_global_oom_handling))] +mod btree; +#[cfg(not(no_global_oom_handling))] +pub mod linked_list; +#[cfg(not(no_global_oom_handling))] +pub mod vec_deque; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +pub mod btree_map { + //! An ordered map based on a B-Tree. + #[stable(feature = "rust1", since = "1.0.0")] + pub use super::btree::map::*; +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +pub mod btree_set { + //! An ordered set based on a B-Tree. + #[stable(feature = "rust1", since = "1.0.0")] + pub use super::btree::set::*; +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[doc(no_inline)] +pub use binary_heap::BinaryHeap; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[doc(no_inline)] +pub use btree_map::BTreeMap; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[doc(no_inline)] +pub use btree_set::BTreeSet; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[doc(no_inline)] +pub use linked_list::LinkedList; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +#[doc(no_inline)] +pub use vec_deque::VecDeque; + +use crate::alloc::{Layout, LayoutError}; +use core::fmt::Display; + +/// The error type for `try_reserve` methods. +#[derive(Clone, PartialEq, Eq, Debug)] +#[stable(feature = "try_reserve", since = "1.57.0")] +pub struct TryReserveError { + kind: TryReserveErrorKind, +} + +impl TryReserveError { + /// Details about the allocation that caused the error + #[inline] + #[must_use] + #[unstable( + feature = "try_reserve_kind", + reason = "Uncertain how much info should be exposed", + issue = "48043" + )] + pub fn kind(&self) -> TryReserveErrorKind { + self.kind.clone() + } +} + +/// Details of the allocation that caused a `TryReserveError` +#[derive(Clone, PartialEq, Eq, Debug)] +#[unstable( + feature = "try_reserve_kind", + reason = "Uncertain how much info should be exposed", + issue = "48043" +)] +pub enum TryReserveErrorKind { + /// Error due to the computed capacity exceeding the collection's maximum + /// (usually `isize::MAX` bytes). + CapacityOverflow, + + /// The memory allocator returned an error + AllocError { + /// The layout of allocation request that failed + layout: Layout, + + #[doc(hidden)] + #[unstable( + feature = "container_error_extra", + issue = "none", + reason = "\ + Enable exposing the allocator’s custom error value \ + if an associated type is added in the future: \ + https://github.com/rust-lang/wg-allocators/issues/23" + )] + non_exhaustive: (), + }, +} + +#[unstable( + feature = "try_reserve_kind", + reason = "Uncertain how much info should be exposed", + issue = "48043" +)] +impl From<TryReserveErrorKind> for TryReserveError { + #[inline] + fn from(kind: TryReserveErrorKind) -> Self { + Self { kind } + } +} + +#[unstable(feature = "try_reserve_kind", reason = "new API", issue = "48043")] +impl From<LayoutError> for TryReserveErrorKind { + /// Always evaluates to [`TryReserveErrorKind::CapacityOverflow`]. + #[inline] + fn from(_: LayoutError) -> Self { + TryReserveErrorKind::CapacityOverflow + } +} + +#[stable(feature = "try_reserve", since = "1.57.0")] +impl Display for TryReserveError { + fn fmt( + &self, + fmt: &mut core::fmt::Formatter<'_>, + ) -> core::result::Result<(), core::fmt::Error> { + fmt.write_str("memory allocation failed")?; + let reason = match self.kind { + TryReserveErrorKind::CapacityOverflow => { + " because the computed capacity exceeded the collection's maximum" + } + TryReserveErrorKind::AllocError { .. } => { + " because the memory allocator returned a error" + } + }; + fmt.write_str(reason) + } +} + +/// An intermediate trait for specialization of `Extend`. +#[doc(hidden)] +trait SpecExtend<I: IntoIterator> { + /// Extends `self` with the contents of the given iterator. + fn spec_extend(&mut self, iter: I); +} diff --git a/library/alloc/src/collections/vec_deque/drain.rs b/library/alloc/src/collections/vec_deque/drain.rs new file mode 100644 index 000000000..05f94da6d --- /dev/null +++ b/library/alloc/src/collections/vec_deque/drain.rs @@ -0,0 +1,142 @@ +use core::iter::FusedIterator; +use core::ptr::{self, NonNull}; +use core::{fmt, mem}; + +use crate::alloc::{Allocator, Global}; + +use super::{count, Iter, VecDeque}; + +/// A draining iterator over the elements of a `VecDeque`. +/// +/// This `struct` is created by the [`drain`] method on [`VecDeque`]. See its +/// documentation for more. +/// +/// [`drain`]: VecDeque::drain +#[stable(feature = "drain", since = "1.6.0")] +pub struct Drain< + 'a, + T: 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +> { + after_tail: usize, + after_head: usize, + iter: Iter<'a, T>, + deque: NonNull<VecDeque<T, A>>, +} + +impl<'a, T, A: Allocator> Drain<'a, T, A> { + pub(super) unsafe fn new( + after_tail: usize, + after_head: usize, + iter: Iter<'a, T>, + deque: NonNull<VecDeque<T, A>>, + ) -> Self { + Drain { after_tail, after_head, iter, deque } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for Drain<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Drain") + .field(&self.after_tail) + .field(&self.after_head) + .field(&self.iter) + .finish() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl<T: Sync, A: Allocator + Sync> Sync for Drain<'_, T, A> {} +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl<T: Send, A: Allocator + Send> Send for Drain<'_, T, A> {} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> Drop for Drain<'_, T, A> { + fn drop(&mut self) { + struct DropGuard<'r, 'a, T, A: Allocator>(&'r mut Drain<'a, T, A>); + + impl<'r, 'a, T, A: Allocator> Drop for DropGuard<'r, 'a, T, A> { + fn drop(&mut self) { + self.0.for_each(drop); + + let source_deque = unsafe { self.0.deque.as_mut() }; + + // T = source_deque_tail; H = source_deque_head; t = drain_tail; h = drain_head + // + // T t h H + // [. . . o o x x o o . . .] + // + let orig_tail = source_deque.tail; + let drain_tail = source_deque.head; + let drain_head = self.0.after_tail; + let orig_head = self.0.after_head; + + let tail_len = count(orig_tail, drain_tail, source_deque.cap()); + let head_len = count(drain_head, orig_head, source_deque.cap()); + + // Restore the original head value + source_deque.head = orig_head; + + match (tail_len, head_len) { + (0, 0) => { + source_deque.head = 0; + source_deque.tail = 0; + } + (0, _) => { + source_deque.tail = drain_head; + } + (_, 0) => { + source_deque.head = drain_tail; + } + _ => unsafe { + if tail_len <= head_len { + source_deque.tail = source_deque.wrap_sub(drain_head, tail_len); + source_deque.wrap_copy(source_deque.tail, orig_tail, tail_len); + } else { + source_deque.head = source_deque.wrap_add(drain_tail, head_len); + source_deque.wrap_copy(drain_tail, drain_head, head_len); + } + }, + } + } + } + + while let Some(item) = self.next() { + let guard = DropGuard(self); + drop(item); + mem::forget(guard); + } + + DropGuard(self); + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> Iterator for Drain<'_, T, A> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.iter.next().map(|elt| unsafe { ptr::read(elt) }) + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> DoubleEndedIterator for Drain<'_, T, A> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.iter.next_back().map(|elt| unsafe { ptr::read(elt) }) + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> ExactSizeIterator for Drain<'_, T, A> {} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T, A: Allocator> FusedIterator for Drain<'_, T, A> {} diff --git a/library/alloc/src/collections/vec_deque/into_iter.rs b/library/alloc/src/collections/vec_deque/into_iter.rs new file mode 100644 index 000000000..55f6138cd --- /dev/null +++ b/library/alloc/src/collections/vec_deque/into_iter.rs @@ -0,0 +1,72 @@ +use core::fmt; +use core::iter::{FusedIterator, TrustedLen}; + +use crate::alloc::{Allocator, Global}; + +use super::VecDeque; + +/// An owning iterator over the elements of a `VecDeque`. +/// +/// This `struct` is created by the [`into_iter`] method on [`VecDeque`] +/// (provided by the [`IntoIterator`] trait). See its documentation for more. +/// +/// [`into_iter`]: VecDeque::into_iter +/// [`IntoIterator`]: core::iter::IntoIterator +#[derive(Clone)] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct IntoIter< + T, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +> { + inner: VecDeque<T, A>, +} + +impl<T, A: Allocator> IntoIter<T, A> { + pub(super) fn new(inner: VecDeque<T, A>) -> Self { + IntoIter { inner } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for IntoIter<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("IntoIter").field(&self.inner).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Iterator for IntoIter<T, A> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.inner.pop_front() + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let len = self.inner.len(); + (len, Some(len)) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> DoubleEndedIterator for IntoIter<T, A> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.inner.pop_back() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> ExactSizeIterator for IntoIter<T, A> { + fn is_empty(&self) -> bool { + self.inner.is_empty() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T, A: Allocator> FusedIterator for IntoIter<T, A> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T, A: Allocator> TrustedLen for IntoIter<T, A> {} diff --git a/library/alloc/src/collections/vec_deque/iter.rs b/library/alloc/src/collections/vec_deque/iter.rs new file mode 100644 index 000000000..e696d7ed6 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/iter.rs @@ -0,0 +1,219 @@ +use core::fmt; +use core::iter::{FusedIterator, TrustedLen, TrustedRandomAccess, TrustedRandomAccessNoCoerce}; +use core::mem::MaybeUninit; +use core::ops::Try; + +use super::{count, wrap_index, RingSlices}; + +/// An iterator over the elements of a `VecDeque`. +/// +/// This `struct` is created by the [`iter`] method on [`super::VecDeque`]. See its +/// documentation for more. +/// +/// [`iter`]: super::VecDeque::iter +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Iter<'a, T: 'a> { + ring: &'a [MaybeUninit<T>], + tail: usize, + head: usize, +} + +impl<'a, T> Iter<'a, T> { + pub(super) fn new(ring: &'a [MaybeUninit<T>], tail: usize, head: usize) -> Self { + Iter { ring, tail, head } + } +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for Iter<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + f.debug_tuple("Iter") + .field(&MaybeUninit::slice_assume_init_ref(front)) + .field(&MaybeUninit::slice_assume_init_ref(back)) + .finish() + } + } +} + +// FIXME(#26925) Remove in favor of `#[derive(Clone)]` +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Clone for Iter<'_, T> { + fn clone(&self) -> Self { + Iter { ring: self.ring, tail: self.tail, head: self.head } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for Iter<'a, T> { + type Item = &'a T; + + #[inline] + fn next(&mut self) -> Option<&'a T> { + if self.tail == self.head { + return None; + } + let tail = self.tail; + self.tail = wrap_index(self.tail.wrapping_add(1), self.ring.len()); + // Safety: + // - `self.tail` in a ring buffer is always a valid index. + // - `self.head` and `self.tail` equality is checked above. + unsafe { Some(self.ring.get_unchecked(tail).assume_init_ref()) } + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let len = count(self.tail, self.head, self.ring.len()); + (len, Some(len)) + } + + fn fold<Acc, F>(self, mut accum: Acc, mut f: F) -> Acc + where + F: FnMut(Acc, Self::Item) -> Acc, + { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + accum = MaybeUninit::slice_assume_init_ref(front).iter().fold(accum, &mut f); + MaybeUninit::slice_assume_init_ref(back).iter().fold(accum, &mut f) + } + } + + fn try_fold<B, F, R>(&mut self, init: B, mut f: F) -> R + where + Self: Sized, + F: FnMut(B, Self::Item) -> R, + R: Try<Output = B>, + { + let (mut iter, final_res); + if self.tail <= self.head { + // Safety: single slice self.ring[self.tail..self.head] is initialized. + iter = unsafe { MaybeUninit::slice_assume_init_ref(&self.ring[self.tail..self.head]) } + .iter(); + final_res = iter.try_fold(init, &mut f); + } else { + // Safety: two slices: self.ring[self.tail..], self.ring[..self.head] both are initialized. + let (front, back) = self.ring.split_at(self.tail); + + let mut back_iter = unsafe { MaybeUninit::slice_assume_init_ref(back).iter() }; + let res = back_iter.try_fold(init, &mut f); + let len = self.ring.len(); + self.tail = (self.ring.len() - back_iter.len()) & (len - 1); + iter = unsafe { MaybeUninit::slice_assume_init_ref(&front[..self.head]).iter() }; + final_res = iter.try_fold(res?, &mut f); + } + self.tail = self.head - iter.len(); + final_res + } + + fn nth(&mut self, n: usize) -> Option<Self::Item> { + if n >= count(self.tail, self.head, self.ring.len()) { + self.tail = self.head; + None + } else { + self.tail = wrap_index(self.tail.wrapping_add(n), self.ring.len()); + self.next() + } + } + + #[inline] + fn last(mut self) -> Option<&'a T> { + self.next_back() + } + + #[inline] + unsafe fn __iterator_get_unchecked(&mut self, idx: usize) -> Self::Item { + // Safety: The TrustedRandomAccess contract requires that callers only pass an index + // that is in bounds. + unsafe { + let idx = wrap_index(self.tail.wrapping_add(idx), self.ring.len()); + self.ring.get_unchecked(idx).assume_init_ref() + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for Iter<'a, T> { + #[inline] + fn next_back(&mut self) -> Option<&'a T> { + if self.tail == self.head { + return None; + } + self.head = wrap_index(self.head.wrapping_sub(1), self.ring.len()); + // Safety: + // - `self.head` in a ring buffer is always a valid index. + // - `self.head` and `self.tail` equality is checked above. + unsafe { Some(self.ring.get_unchecked(self.head).assume_init_ref()) } + } + + fn rfold<Acc, F>(self, mut accum: Acc, mut f: F) -> Acc + where + F: FnMut(Acc, Self::Item) -> Acc, + { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + accum = MaybeUninit::slice_assume_init_ref(back).iter().rfold(accum, &mut f); + MaybeUninit::slice_assume_init_ref(front).iter().rfold(accum, &mut f) + } + } + + fn try_rfold<B, F, R>(&mut self, init: B, mut f: F) -> R + where + Self: Sized, + F: FnMut(B, Self::Item) -> R, + R: Try<Output = B>, + { + let (mut iter, final_res); + if self.tail <= self.head { + // Safety: single slice self.ring[self.tail..self.head] is initialized. + iter = unsafe { + MaybeUninit::slice_assume_init_ref(&self.ring[self.tail..self.head]).iter() + }; + final_res = iter.try_rfold(init, &mut f); + } else { + // Safety: two slices: self.ring[self.tail..], self.ring[..self.head] both are initialized. + let (front, back) = self.ring.split_at(self.tail); + + let mut front_iter = + unsafe { MaybeUninit::slice_assume_init_ref(&front[..self.head]).iter() }; + let res = front_iter.try_rfold(init, &mut f); + self.head = front_iter.len(); + iter = unsafe { MaybeUninit::slice_assume_init_ref(back).iter() }; + final_res = iter.try_rfold(res?, &mut f); + } + self.head = self.tail + iter.len(); + final_res + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for Iter<'_, T> { + fn is_empty(&self) -> bool { + self.head == self.tail + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for Iter<'_, T> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T> TrustedLen for Iter<'_, T> {} + +#[doc(hidden)] +#[unstable(feature = "trusted_random_access", issue = "none")] +unsafe impl<T> TrustedRandomAccess for Iter<'_, T> {} + +#[doc(hidden)] +#[unstable(feature = "trusted_random_access", issue = "none")] +unsafe impl<T> TrustedRandomAccessNoCoerce for Iter<'_, T> { + const MAY_HAVE_SIDE_EFFECT: bool = false; +} diff --git a/library/alloc/src/collections/vec_deque/iter_mut.rs b/library/alloc/src/collections/vec_deque/iter_mut.rs new file mode 100644 index 000000000..b78c0d5e1 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/iter_mut.rs @@ -0,0 +1,162 @@ +use core::fmt; +use core::iter::{FusedIterator, TrustedLen, TrustedRandomAccess, TrustedRandomAccessNoCoerce}; +use core::marker::PhantomData; + +use super::{count, wrap_index, RingSlices}; + +/// A mutable iterator over the elements of a `VecDeque`. +/// +/// This `struct` is created by the [`iter_mut`] method on [`super::VecDeque`]. See its +/// documentation for more. +/// +/// [`iter_mut`]: super::VecDeque::iter_mut +#[stable(feature = "rust1", since = "1.0.0")] +pub struct IterMut<'a, T: 'a> { + // Internal safety invariant: the entire slice is dereferenceable. + ring: *mut [T], + tail: usize, + head: usize, + phantom: PhantomData<&'a mut [T]>, +} + +impl<'a, T> IterMut<'a, T> { + pub(super) unsafe fn new( + ring: *mut [T], + tail: usize, + head: usize, + phantom: PhantomData<&'a mut [T]>, + ) -> Self { + IterMut { ring, tail, head, phantom } + } +} + +// SAFETY: we do nothing thread-local and there is no interior mutability, +// so the usual structural `Send`/`Sync` apply. +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Send> Send for IterMut<'_, T> {} +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync> Sync for IterMut<'_, T> {} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug> fmt::Debug for IterMut<'_, T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // SAFETY: these are the elements we have not handed out yet, so aliasing is fine. + // The `IterMut` invariant also ensures everything is dereferenceable. + let (front, back) = unsafe { (&*front, &*back) }; + f.debug_tuple("IterMut").field(&front).field(&back).finish() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> Iterator for IterMut<'a, T> { + type Item = &'a mut T; + + #[inline] + fn next(&mut self) -> Option<&'a mut T> { + if self.tail == self.head { + return None; + } + let tail = self.tail; + self.tail = wrap_index(self.tail.wrapping_add(1), self.ring.len()); + + unsafe { + let elem = self.ring.get_unchecked_mut(tail); + Some(&mut *elem) + } + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let len = count(self.tail, self.head, self.ring.len()); + (len, Some(len)) + } + + fn fold<Acc, F>(self, mut accum: Acc, mut f: F) -> Acc + where + F: FnMut(Acc, Self::Item) -> Acc, + { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // SAFETY: these are the elements we have not handed out yet, so aliasing is fine. + // The `IterMut` invariant also ensures everything is dereferenceable. + let (front, back) = unsafe { (&mut *front, &mut *back) }; + accum = front.iter_mut().fold(accum, &mut f); + back.iter_mut().fold(accum, &mut f) + } + + fn nth(&mut self, n: usize) -> Option<Self::Item> { + if n >= count(self.tail, self.head, self.ring.len()) { + self.tail = self.head; + None + } else { + self.tail = wrap_index(self.tail.wrapping_add(n), self.ring.len()); + self.next() + } + } + + #[inline] + fn last(mut self) -> Option<&'a mut T> { + self.next_back() + } + + #[inline] + unsafe fn __iterator_get_unchecked(&mut self, idx: usize) -> Self::Item { + // Safety: The TrustedRandomAccess contract requires that callers only pass an index + // that is in bounds. + unsafe { + let idx = wrap_index(self.tail.wrapping_add(idx), self.ring.len()); + &mut *self.ring.get_unchecked_mut(idx) + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> DoubleEndedIterator for IterMut<'a, T> { + #[inline] + fn next_back(&mut self) -> Option<&'a mut T> { + if self.tail == self.head { + return None; + } + self.head = wrap_index(self.head.wrapping_sub(1), self.ring.len()); + + unsafe { + let elem = self.ring.get_unchecked_mut(self.head); + Some(&mut *elem) + } + } + + fn rfold<Acc, F>(self, mut accum: Acc, mut f: F) -> Acc + where + F: FnMut(Acc, Self::Item) -> Acc, + { + let (front, back) = RingSlices::ring_slices(self.ring, self.head, self.tail); + // SAFETY: these are the elements we have not handed out yet, so aliasing is fine. + // The `IterMut` invariant also ensures everything is dereferenceable. + let (front, back) = unsafe { (&mut *front, &mut *back) }; + accum = back.iter_mut().rfold(accum, &mut f); + front.iter_mut().rfold(accum, &mut f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> ExactSizeIterator for IterMut<'_, T> { + fn is_empty(&self) -> bool { + self.head == self.tail + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T> FusedIterator for IterMut<'_, T> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T> TrustedLen for IterMut<'_, T> {} + +#[doc(hidden)] +#[unstable(feature = "trusted_random_access", issue = "none")] +unsafe impl<T> TrustedRandomAccess for IterMut<'_, T> {} + +#[doc(hidden)] +#[unstable(feature = "trusted_random_access", issue = "none")] +unsafe impl<T> TrustedRandomAccessNoCoerce for IterMut<'_, T> { + const MAY_HAVE_SIDE_EFFECT: bool = false; +} diff --git a/library/alloc/src/collections/vec_deque/macros.rs b/library/alloc/src/collections/vec_deque/macros.rs new file mode 100644 index 000000000..5c7913073 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/macros.rs @@ -0,0 +1,19 @@ +macro_rules! __impl_slice_eq1 { + ([$($vars:tt)*] $lhs:ty, $rhs:ty, $($constraints:tt)*) => { + #[stable(feature = "vec_deque_partial_eq_slice", since = "1.17.0")] + impl<T, U, A: Allocator, $($vars)*> PartialEq<$rhs> for $lhs + where + T: PartialEq<U>, + $($constraints)* + { + fn eq(&self, other: &$rhs) -> bool { + if self.len() != other.len() { + return false; + } + let (sa, sb) = self.as_slices(); + let (oa, ob) = other[..].split_at(sa.len()); + sa == oa && sb == ob + } + } + } +} diff --git a/library/alloc/src/collections/vec_deque/mod.rs b/library/alloc/src/collections/vec_deque/mod.rs new file mode 100644 index 000000000..4d895d837 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/mod.rs @@ -0,0 +1,3137 @@ +//! A double-ended queue (deque) implemented with a growable ring buffer. +//! +//! This queue has *O*(1) amortized inserts and removals from both ends of the +//! container. It also has *O*(1) indexing like a vector. The contained elements +//! are not required to be copyable, and the queue will be sendable if the +//! contained type is sendable. + +#![stable(feature = "rust1", since = "1.0.0")] + +use core::cmp::{self, Ordering}; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::iter::{repeat_with, FromIterator}; +use core::marker::PhantomData; +use core::mem::{self, ManuallyDrop, MaybeUninit}; +use core::ops::{Index, IndexMut, Range, RangeBounds}; +use core::ptr::{self, NonNull}; +use core::slice; + +use crate::alloc::{Allocator, Global}; +use crate::collections::TryReserveError; +use crate::collections::TryReserveErrorKind; +use crate::raw_vec::RawVec; +use crate::vec::Vec; + +#[macro_use] +mod macros; + +#[stable(feature = "drain", since = "1.6.0")] +pub use self::drain::Drain; + +mod drain; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use self::iter_mut::IterMut; + +mod iter_mut; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use self::into_iter::IntoIter; + +mod into_iter; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use self::iter::Iter; + +mod iter; + +use self::pair_slices::PairSlices; + +mod pair_slices; + +use self::ring_slices::RingSlices; + +mod ring_slices; + +use self::spec_extend::SpecExtend; + +mod spec_extend; + +#[cfg(test)] +mod tests; + +const INITIAL_CAPACITY: usize = 7; // 2^3 - 1 +const MINIMUM_CAPACITY: usize = 1; // 2 - 1 + +const MAXIMUM_ZST_CAPACITY: usize = 1 << (usize::BITS - 1); // Largest possible power of two + +/// A double-ended queue implemented with a growable ring buffer. +/// +/// The "default" usage of this type as a queue is to use [`push_back`] to add to +/// the queue, and [`pop_front`] to remove from the queue. [`extend`] and [`append`] +/// push onto the back in this manner, and iterating over `VecDeque` goes front +/// to back. +/// +/// A `VecDeque` with a known list of items can be initialized from an array: +/// +/// ``` +/// use std::collections::VecDeque; +/// +/// let deq = VecDeque::from([-1, 0, 1]); +/// ``` +/// +/// Since `VecDeque` is a ring buffer, its elements are not necessarily contiguous +/// in memory. If you want to access the elements as a single slice, such as for +/// efficient sorting, you can use [`make_contiguous`]. It rotates the `VecDeque` +/// so that its elements do not wrap, and returns a mutable slice to the +/// now-contiguous element sequence. +/// +/// [`push_back`]: VecDeque::push_back +/// [`pop_front`]: VecDeque::pop_front +/// [`extend`]: VecDeque::extend +/// [`append`]: VecDeque::append +/// [`make_contiguous`]: VecDeque::make_contiguous +#[cfg_attr(not(test), rustc_diagnostic_item = "VecDeque")] +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_insignificant_dtor] +pub struct VecDeque< + T, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +> { + // tail and head are pointers into the buffer. Tail always points + // to the first element that could be read, Head always points + // to where data should be written. + // If tail == head the buffer is empty. The length of the ringbuffer + // is defined as the distance between the two. + tail: usize, + head: usize, + buf: RawVec<T, A>, +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone, A: Allocator + Clone> Clone for VecDeque<T, A> { + fn clone(&self) -> Self { + let mut deq = Self::with_capacity_in(self.len(), self.allocator().clone()); + deq.extend(self.iter().cloned()); + deq + } + + fn clone_from(&mut self, other: &Self) { + self.truncate(other.len()); + + let mut iter = PairSlices::from(self, other); + while let Some((dst, src)) = iter.next() { + dst.clone_from_slice(&src); + } + + if iter.has_remainder() { + for remainder in iter.remainder() { + self.extend(remainder.iter().cloned()); + } + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T, A: Allocator> Drop for VecDeque<T, A> { + fn drop(&mut self) { + /// Runs the destructor for all items in the slice when it gets dropped (normally or + /// during unwinding). + struct Dropper<'a, T>(&'a mut [T]); + + impl<'a, T> Drop for Dropper<'a, T> { + fn drop(&mut self) { + unsafe { + ptr::drop_in_place(self.0); + } + } + } + + let (front, back) = self.as_mut_slices(); + unsafe { + let _back_dropper = Dropper(back); + // use drop for [T] + ptr::drop_in_place(front); + } + // RawVec handles deallocation + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> Default for VecDeque<T> { + /// Creates an empty deque. + #[inline] + fn default() -> VecDeque<T> { + VecDeque::new() + } +} + +impl<T, A: Allocator> VecDeque<T, A> { + /// Marginally more convenient + #[inline] + fn ptr(&self) -> *mut T { + self.buf.ptr() + } + + /// Marginally more convenient + #[inline] + fn cap(&self) -> usize { + if mem::size_of::<T>() == 0 { + // For zero sized types, we are always at maximum capacity + MAXIMUM_ZST_CAPACITY + } else { + self.buf.capacity() + } + } + + /// Turn ptr into a slice, since the elements of the backing buffer may be uninitialized, + /// we will return a slice of [`MaybeUninit<T>`]. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[inline] + unsafe fn buffer_as_slice(&self) -> &[MaybeUninit<T>] { + unsafe { slice::from_raw_parts(self.ptr() as *mut MaybeUninit<T>, self.cap()) } + } + + /// Turn ptr into a mut slice, since the elements of the backing buffer may be uninitialized, + /// we will return a slice of [`MaybeUninit<T>`]. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[inline] + unsafe fn buffer_as_mut_slice(&mut self) -> &mut [MaybeUninit<T>] { + unsafe { slice::from_raw_parts_mut(self.ptr() as *mut MaybeUninit<T>, self.cap()) } + } + + /// Moves an element out of the buffer + #[inline] + unsafe fn buffer_read(&mut self, off: usize) -> T { + unsafe { ptr::read(self.ptr().add(off)) } + } + + /// Writes an element into the buffer, moving it. + #[inline] + unsafe fn buffer_write(&mut self, off: usize, value: T) { + unsafe { + ptr::write(self.ptr().add(off), value); + } + } + + /// Returns `true` if the buffer is at full capacity. + #[inline] + fn is_full(&self) -> bool { + self.cap() - self.len() == 1 + } + + /// Returns the index in the underlying buffer for a given logical element + /// index. + #[inline] + fn wrap_index(&self, idx: usize) -> usize { + wrap_index(idx, self.cap()) + } + + /// Returns the index in the underlying buffer for a given logical element + /// index + addend. + #[inline] + fn wrap_add(&self, idx: usize, addend: usize) -> usize { + wrap_index(idx.wrapping_add(addend), self.cap()) + } + + /// Returns the index in the underlying buffer for a given logical element + /// index - subtrahend. + #[inline] + fn wrap_sub(&self, idx: usize, subtrahend: usize) -> usize { + wrap_index(idx.wrapping_sub(subtrahend), self.cap()) + } + + /// Copies a contiguous block of memory len long from src to dst + #[inline] + unsafe fn copy(&self, dst: usize, src: usize, len: usize) { + debug_assert!( + dst + len <= self.cap(), + "cpy dst={} src={} len={} cap={}", + dst, + src, + len, + self.cap() + ); + debug_assert!( + src + len <= self.cap(), + "cpy dst={} src={} len={} cap={}", + dst, + src, + len, + self.cap() + ); + unsafe { + ptr::copy(self.ptr().add(src), self.ptr().add(dst), len); + } + } + + /// Copies a contiguous block of memory len long from src to dst + #[inline] + unsafe fn copy_nonoverlapping(&self, dst: usize, src: usize, len: usize) { + debug_assert!( + dst + len <= self.cap(), + "cno dst={} src={} len={} cap={}", + dst, + src, + len, + self.cap() + ); + debug_assert!( + src + len <= self.cap(), + "cno dst={} src={} len={} cap={}", + dst, + src, + len, + self.cap() + ); + unsafe { + ptr::copy_nonoverlapping(self.ptr().add(src), self.ptr().add(dst), len); + } + } + + /// Copies a potentially wrapping block of memory len long from src to dest. + /// (abs(dst - src) + len) must be no larger than cap() (There must be at + /// most one continuous overlapping region between src and dest). + unsafe fn wrap_copy(&self, dst: usize, src: usize, len: usize) { + #[allow(dead_code)] + fn diff(a: usize, b: usize) -> usize { + if a <= b { b - a } else { a - b } + } + debug_assert!( + cmp::min(diff(dst, src), self.cap() - diff(dst, src)) + len <= self.cap(), + "wrc dst={} src={} len={} cap={}", + dst, + src, + len, + self.cap() + ); + + if src == dst || len == 0 { + return; + } + + let dst_after_src = self.wrap_sub(dst, src) < len; + + let src_pre_wrap_len = self.cap() - src; + let dst_pre_wrap_len = self.cap() - dst; + let src_wraps = src_pre_wrap_len < len; + let dst_wraps = dst_pre_wrap_len < len; + + match (dst_after_src, src_wraps, dst_wraps) { + (_, false, false) => { + // src doesn't wrap, dst doesn't wrap + // + // S . . . + // 1 [_ _ A A B B C C _] + // 2 [_ _ A A A A B B _] + // D . . . + // + unsafe { + self.copy(dst, src, len); + } + } + (false, false, true) => { + // dst before src, src doesn't wrap, dst wraps + // + // S . . . + // 1 [A A B B _ _ _ C C] + // 2 [A A B B _ _ _ A A] + // 3 [B B B B _ _ _ A A] + // . . D . + // + unsafe { + self.copy(dst, src, dst_pre_wrap_len); + self.copy(0, src + dst_pre_wrap_len, len - dst_pre_wrap_len); + } + } + (true, false, true) => { + // src before dst, src doesn't wrap, dst wraps + // + // S . . . + // 1 [C C _ _ _ A A B B] + // 2 [B B _ _ _ A A B B] + // 3 [B B _ _ _ A A A A] + // . . D . + // + unsafe { + self.copy(0, src + dst_pre_wrap_len, len - dst_pre_wrap_len); + self.copy(dst, src, dst_pre_wrap_len); + } + } + (false, true, false) => { + // dst before src, src wraps, dst doesn't wrap + // + // . . S . + // 1 [C C _ _ _ A A B B] + // 2 [C C _ _ _ B B B B] + // 3 [C C _ _ _ B B C C] + // D . . . + // + unsafe { + self.copy(dst, src, src_pre_wrap_len); + self.copy(dst + src_pre_wrap_len, 0, len - src_pre_wrap_len); + } + } + (true, true, false) => { + // src before dst, src wraps, dst doesn't wrap + // + // . . S . + // 1 [A A B B _ _ _ C C] + // 2 [A A A A _ _ _ C C] + // 3 [C C A A _ _ _ C C] + // D . . . + // + unsafe { + self.copy(dst + src_pre_wrap_len, 0, len - src_pre_wrap_len); + self.copy(dst, src, src_pre_wrap_len); + } + } + (false, true, true) => { + // dst before src, src wraps, dst wraps + // + // . . . S . + // 1 [A B C D _ E F G H] + // 2 [A B C D _ E G H H] + // 3 [A B C D _ E G H A] + // 4 [B C C D _ E G H A] + // . . D . . + // + debug_assert!(dst_pre_wrap_len > src_pre_wrap_len); + let delta = dst_pre_wrap_len - src_pre_wrap_len; + unsafe { + self.copy(dst, src, src_pre_wrap_len); + self.copy(dst + src_pre_wrap_len, 0, delta); + self.copy(0, delta, len - dst_pre_wrap_len); + } + } + (true, true, true) => { + // src before dst, src wraps, dst wraps + // + // . . S . . + // 1 [A B C D _ E F G H] + // 2 [A A B D _ E F G H] + // 3 [H A B D _ E F G H] + // 4 [H A B D _ E F F G] + // . . . D . + // + debug_assert!(src_pre_wrap_len > dst_pre_wrap_len); + let delta = src_pre_wrap_len - dst_pre_wrap_len; + unsafe { + self.copy(delta, 0, len - src_pre_wrap_len); + self.copy(0, self.cap() - delta, delta); + self.copy(dst, src, dst_pre_wrap_len); + } + } + } + } + + /// Copies all values from `src` to `dst`, wrapping around if needed. + /// Assumes capacity is sufficient. + #[inline] + unsafe fn copy_slice(&mut self, dst: usize, src: &[T]) { + debug_assert!(src.len() <= self.cap()); + let head_room = self.cap() - dst; + if src.len() <= head_room { + unsafe { + ptr::copy_nonoverlapping(src.as_ptr(), self.ptr().add(dst), src.len()); + } + } else { + let (left, right) = src.split_at(head_room); + unsafe { + ptr::copy_nonoverlapping(left.as_ptr(), self.ptr().add(dst), left.len()); + ptr::copy_nonoverlapping(right.as_ptr(), self.ptr(), right.len()); + } + } + } + + /// Writes all values from `iter` to `dst`. + /// + /// # Safety + /// + /// Assumes no wrapping around happens. + /// Assumes capacity is sufficient. + #[inline] + unsafe fn write_iter( + &mut self, + dst: usize, + iter: impl Iterator<Item = T>, + written: &mut usize, + ) { + iter.enumerate().for_each(|(i, element)| unsafe { + self.buffer_write(dst + i, element); + *written += 1; + }); + } + + /// Frobs the head and tail sections around to handle the fact that we + /// just reallocated. Unsafe because it trusts old_capacity. + #[inline] + unsafe fn handle_capacity_increase(&mut self, old_capacity: usize) { + let new_capacity = self.cap(); + + // Move the shortest contiguous section of the ring buffer + // T H + // [o o o o o o o . ] + // T H + // A [o o o o o o o . . . . . . . . . ] + // H T + // [o o . o o o o o ] + // T H + // B [. . . o o o o o o o . . . . . . ] + // H T + // [o o o o o . o o ] + // H T + // C [o o o o o . . . . . . . . . o o ] + + if self.tail <= self.head { + // A + // Nop + } else if self.head < old_capacity - self.tail { + // B + unsafe { + self.copy_nonoverlapping(old_capacity, 0, self.head); + } + self.head += old_capacity; + debug_assert!(self.head > self.tail); + } else { + // C + let new_tail = new_capacity - (old_capacity - self.tail); + unsafe { + self.copy_nonoverlapping(new_tail, self.tail, old_capacity - self.tail); + } + self.tail = new_tail; + debug_assert!(self.head < self.tail); + } + debug_assert!(self.head < self.cap()); + debug_assert!(self.tail < self.cap()); + debug_assert!(self.cap().count_ones() == 1); + } +} + +impl<T> VecDeque<T> { + /// Creates an empty deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<u32> = VecDeque::new(); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn new() -> VecDeque<T> { + VecDeque::new_in(Global) + } + + /// Creates an empty deque with space for at least `capacity` elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<u32> = VecDeque::with_capacity(10); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn with_capacity(capacity: usize) -> VecDeque<T> { + Self::with_capacity_in(capacity, Global) + } +} + +impl<T, A: Allocator> VecDeque<T, A> { + /// Creates an empty deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<u32> = VecDeque::new(); + /// ``` + #[inline] + #[unstable(feature = "allocator_api", issue = "32838")] + pub fn new_in(alloc: A) -> VecDeque<T, A> { + VecDeque::with_capacity_in(INITIAL_CAPACITY, alloc) + } + + /// Creates an empty deque with space for at least `capacity` elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<u32> = VecDeque::with_capacity(10); + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + pub fn with_capacity_in(capacity: usize, alloc: A) -> VecDeque<T, A> { + assert!(capacity < 1_usize << usize::BITS - 1, "capacity overflow"); + // +1 since the ringbuffer always leaves one space empty + let cap = cmp::max(capacity + 1, MINIMUM_CAPACITY + 1).next_power_of_two(); + + VecDeque { tail: 0, head: 0, buf: RawVec::with_capacity_in(cap, alloc) } + } + + /// Provides a reference to the element at the given index. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(3); + /// buf.push_back(4); + /// buf.push_back(5); + /// assert_eq!(buf.get(1), Some(&4)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get(&self, index: usize) -> Option<&T> { + if index < self.len() { + let idx = self.wrap_add(self.tail, index); + unsafe { Some(&*self.ptr().add(idx)) } + } else { + None + } + } + + /// Provides a mutable reference to the element at the given index. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(3); + /// buf.push_back(4); + /// buf.push_back(5); + /// if let Some(elem) = buf.get_mut(1) { + /// *elem = 7; + /// } + /// + /// assert_eq!(buf[1], 7); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn get_mut(&mut self, index: usize) -> Option<&mut T> { + if index < self.len() { + let idx = self.wrap_add(self.tail, index); + unsafe { Some(&mut *self.ptr().add(idx)) } + } else { + None + } + } + + /// Swaps elements at indices `i` and `j`. + /// + /// `i` and `j` may be equal. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Panics + /// + /// Panics if either index is out of bounds. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(3); + /// buf.push_back(4); + /// buf.push_back(5); + /// assert_eq!(buf, [3, 4, 5]); + /// buf.swap(0, 2); + /// assert_eq!(buf, [5, 4, 3]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn swap(&mut self, i: usize, j: usize) { + assert!(i < self.len()); + assert!(j < self.len()); + let ri = self.wrap_add(self.tail, i); + let rj = self.wrap_add(self.tail, j); + unsafe { ptr::swap(self.ptr().add(ri), self.ptr().add(rj)) } + } + + /// Returns the number of elements the deque can hold without + /// reallocating. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let buf: VecDeque<i32> = VecDeque::with_capacity(10); + /// assert!(buf.capacity() >= 10); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn capacity(&self) -> usize { + self.cap() - 1 + } + + /// Reserves the minimum capacity for at least `additional` more elements to be inserted in the + /// given deque. Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the collection more space than it requests. Therefore + /// capacity can not be relied upon to be precisely minimal. Prefer [`reserve`] if future + /// insertions are expected. + /// + /// # Panics + /// + /// Panics if the new capacity overflows `usize`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<i32> = [1].into(); + /// buf.reserve_exact(10); + /// assert!(buf.capacity() >= 11); + /// ``` + /// + /// [`reserve`]: VecDeque::reserve + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve_exact(&mut self, additional: usize) { + self.reserve(additional); + } + + /// Reserves capacity for at least `additional` more elements to be inserted in the given + /// deque. The collection may reserve more space to speculatively avoid frequent reallocations. + /// + /// # Panics + /// + /// Panics if the new capacity overflows `usize`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<i32> = [1].into(); + /// buf.reserve(10); + /// assert!(buf.capacity() >= 11); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve(&mut self, additional: usize) { + let old_cap = self.cap(); + let used_cap = self.len() + 1; + let new_cap = used_cap + .checked_add(additional) + .and_then(|needed_cap| needed_cap.checked_next_power_of_two()) + .expect("capacity overflow"); + + if new_cap > old_cap { + self.buf.reserve_exact(used_cap, new_cap - used_cap); + unsafe { + self.handle_capacity_increase(old_cap); + } + } + } + + /// Tries to reserve the minimum capacity for at least `additional` more elements to + /// be inserted in the given deque. After calling `try_reserve_exact`, + /// capacity will be greater than or equal to `self.len() + additional` if + /// it returns `Ok(())`. Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the collection more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: VecDeque::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows `usize`, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// use std::collections::VecDeque; + /// + /// fn process_data(data: &[u32]) -> Result<VecDeque<u32>, TryReserveError> { + /// let mut output = VecDeque::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve_exact(data.len())?; + /// + /// // Now we know this can't OOM(Out-Of-Memory) in the middle of our complex work + /// output.extend(data.iter().map(|&val| { + /// val * 2 + 5 // very complicated + /// })); + /// + /// Ok(output) + /// } + /// # process_data(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.try_reserve(additional) + } + + /// Tries to reserve capacity for at least `additional` more elements to be inserted + /// in the given deque. The collection may reserve more space to speculatively avoid + /// frequent reallocations. After calling `try_reserve`, capacity will be + /// greater than or equal to `self.len() + additional` if it returns + /// `Ok(())`. Does nothing if capacity is already sufficient. + /// + /// # Errors + /// + /// If the capacity overflows `usize`, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// use std::collections::VecDeque; + /// + /// fn process_data(data: &[u32]) -> Result<VecDeque<u32>, TryReserveError> { + /// let mut output = VecDeque::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// output.extend(data.iter().map(|&val| { + /// val * 2 + 5 // very complicated + /// })); + /// + /// Ok(output) + /// } + /// # process_data(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + let old_cap = self.cap(); + let used_cap = self.len() + 1; + let new_cap = used_cap + .checked_add(additional) + .and_then(|needed_cap| needed_cap.checked_next_power_of_two()) + .ok_or(TryReserveErrorKind::CapacityOverflow)?; + + if new_cap > old_cap { + self.buf.try_reserve_exact(used_cap, new_cap - used_cap)?; + unsafe { + self.handle_capacity_increase(old_cap); + } + } + Ok(()) + } + + /// Shrinks the capacity of the deque as much as possible. + /// + /// It will drop down as close as possible to the length but the allocator may still inform the + /// deque that there is space for a few more elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::with_capacity(15); + /// buf.extend(0..4); + /// assert_eq!(buf.capacity(), 15); + /// buf.shrink_to_fit(); + /// assert!(buf.capacity() >= 4); + /// ``` + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn shrink_to_fit(&mut self) { + self.shrink_to(0); + } + + /// Shrinks the capacity of the deque with a lower bound. + /// + /// The capacity will remain at least as large as both the length + /// and the supplied value. + /// + /// If the current capacity is less than the lower limit, this is a no-op. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::with_capacity(15); + /// buf.extend(0..4); + /// assert_eq!(buf.capacity(), 15); + /// buf.shrink_to(6); + /// assert!(buf.capacity() >= 6); + /// buf.shrink_to(0); + /// assert!(buf.capacity() >= 4); + /// ``` + #[stable(feature = "shrink_to", since = "1.56.0")] + pub fn shrink_to(&mut self, min_capacity: usize) { + let min_capacity = cmp::min(min_capacity, self.capacity()); + // We don't have to worry about an overflow as neither `self.len()` nor `self.capacity()` + // can ever be `usize::MAX`. +1 as the ringbuffer always leaves one space empty. + let target_cap = cmp::max(cmp::max(min_capacity, self.len()) + 1, MINIMUM_CAPACITY + 1) + .next_power_of_two(); + + if target_cap < self.cap() { + // There are three cases of interest: + // All elements are out of desired bounds + // Elements are contiguous, and head is out of desired bounds + // Elements are discontiguous, and tail is out of desired bounds + // + // At all other times, element positions are unaffected. + // + // Indicates that elements at the head should be moved. + let head_outside = self.head == 0 || self.head >= target_cap; + // Move elements from out of desired bounds (positions after target_cap) + if self.tail >= target_cap && head_outside { + // T H + // [. . . . . . . . o o o o o o o . ] + // T H + // [o o o o o o o . ] + unsafe { + self.copy_nonoverlapping(0, self.tail, self.len()); + } + self.head = self.len(); + self.tail = 0; + } else if self.tail != 0 && self.tail < target_cap && head_outside { + // T H + // [. . . o o o o o o o . . . . . . ] + // H T + // [o o . o o o o o ] + let len = self.wrap_sub(self.head, target_cap); + unsafe { + self.copy_nonoverlapping(0, target_cap, len); + } + self.head = len; + debug_assert!(self.head < self.tail); + } else if self.tail >= target_cap { + // H T + // [o o o o o . . . . . . . . . o o ] + // H T + // [o o o o o . o o ] + debug_assert!(self.wrap_sub(self.head, 1) < target_cap); + let len = self.cap() - self.tail; + let new_tail = target_cap - len; + unsafe { + self.copy_nonoverlapping(new_tail, self.tail, len); + } + self.tail = new_tail; + debug_assert!(self.head < self.tail); + } + + self.buf.shrink_to_fit(target_cap); + + debug_assert!(self.head < self.cap()); + debug_assert!(self.tail < self.cap()); + debug_assert!(self.cap().count_ones() == 1); + } + } + + /// Shortens the deque, keeping the first `len` elements and dropping + /// the rest. + /// + /// If `len` is greater than the deque's current length, this has no + /// effect. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(5); + /// buf.push_back(10); + /// buf.push_back(15); + /// assert_eq!(buf, [5, 10, 15]); + /// buf.truncate(1); + /// assert_eq!(buf, [5]); + /// ``` + #[stable(feature = "deque_extras", since = "1.16.0")] + pub fn truncate(&mut self, len: usize) { + /// Runs the destructor for all items in the slice when it gets dropped (normally or + /// during unwinding). + struct Dropper<'a, T>(&'a mut [T]); + + impl<'a, T> Drop for Dropper<'a, T> { + fn drop(&mut self) { + unsafe { + ptr::drop_in_place(self.0); + } + } + } + + // Safe because: + // + // * Any slice passed to `drop_in_place` is valid; the second case has + // `len <= front.len()` and returning on `len > self.len()` ensures + // `begin <= back.len()` in the first case + // * The head of the VecDeque is moved before calling `drop_in_place`, + // so no value is dropped twice if `drop_in_place` panics + unsafe { + if len > self.len() { + return; + } + let num_dropped = self.len() - len; + let (front, back) = self.as_mut_slices(); + if len > front.len() { + let begin = len - front.len(); + let drop_back = back.get_unchecked_mut(begin..) as *mut _; + self.head = self.wrap_sub(self.head, num_dropped); + ptr::drop_in_place(drop_back); + } else { + let drop_back = back as *mut _; + let drop_front = front.get_unchecked_mut(len..) as *mut _; + self.head = self.wrap_sub(self.head, num_dropped); + + // Make sure the second half is dropped even when a destructor + // in the first one panics. + let _back_dropper = Dropper(&mut *drop_back); + ptr::drop_in_place(drop_front); + } + } + } + + /// Returns a reference to the underlying allocator. + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn allocator(&self) -> &A { + self.buf.allocator() + } + + /// Returns a front-to-back iterator. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(5); + /// buf.push_back(3); + /// buf.push_back(4); + /// let b: &[_] = &[&5, &3, &4]; + /// let c: Vec<&i32> = buf.iter().collect(); + /// assert_eq!(&c[..], b); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter(&self) -> Iter<'_, T> { + Iter::new(unsafe { self.buffer_as_slice() }, self.tail, self.head) + } + + /// Returns a front-to-back iterator that returns mutable references. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(5); + /// buf.push_back(3); + /// buf.push_back(4); + /// for num in buf.iter_mut() { + /// *num = *num - 2; + /// } + /// let b: &[_] = &[&mut 3, &mut 1, &mut 2]; + /// assert_eq!(&buf.iter_mut().collect::<Vec<&mut i32>>()[..], b); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn iter_mut(&mut self) -> IterMut<'_, T> { + // SAFETY: The internal `IterMut` safety invariant is established because the + // `ring` we create is a dereferenceable slice for lifetime '_. + let ring = ptr::slice_from_raw_parts_mut(self.ptr(), self.cap()); + + unsafe { IterMut::new(ring, self.tail, self.head, PhantomData) } + } + + /// Returns a pair of slices which contain, in order, the contents of the + /// deque. + /// + /// If [`make_contiguous`] was previously called, all elements of the + /// deque will be in the first slice and the second slice will be empty. + /// + /// [`make_contiguous`]: VecDeque::make_contiguous + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque = VecDeque::new(); + /// + /// deque.push_back(0); + /// deque.push_back(1); + /// deque.push_back(2); + /// + /// assert_eq!(deque.as_slices(), (&[0, 1, 2][..], &[][..])); + /// + /// deque.push_front(10); + /// deque.push_front(9); + /// + /// assert_eq!(deque.as_slices(), (&[9, 10][..], &[0, 1, 2][..])); + /// ``` + #[inline] + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn as_slices(&self) -> (&[T], &[T]) { + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + let buf = self.buffer_as_slice(); + let (front, back) = RingSlices::ring_slices(buf, self.head, self.tail); + (MaybeUninit::slice_assume_init_ref(front), MaybeUninit::slice_assume_init_ref(back)) + } + } + + /// Returns a pair of slices which contain, in order, the contents of the + /// deque. + /// + /// If [`make_contiguous`] was previously called, all elements of the + /// deque will be in the first slice and the second slice will be empty. + /// + /// [`make_contiguous`]: VecDeque::make_contiguous + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque = VecDeque::new(); + /// + /// deque.push_back(0); + /// deque.push_back(1); + /// + /// deque.push_front(10); + /// deque.push_front(9); + /// + /// deque.as_mut_slices().0[0] = 42; + /// deque.as_mut_slices().1[0] = 24; + /// assert_eq!(deque.as_slices(), (&[42, 10][..], &[24, 1][..])); + /// ``` + #[inline] + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn as_mut_slices(&mut self) -> (&mut [T], &mut [T]) { + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + let head = self.head; + let tail = self.tail; + let buf = self.buffer_as_mut_slice(); + let (front, back) = RingSlices::ring_slices(buf, head, tail); + (MaybeUninit::slice_assume_init_mut(front), MaybeUninit::slice_assume_init_mut(back)) + } + } + + /// Returns the number of elements in the deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque = VecDeque::new(); + /// assert_eq!(deque.len(), 0); + /// deque.push_back(1); + /// assert_eq!(deque.len(), 1); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn len(&self) -> usize { + count(self.tail, self.head, self.cap()) + } + + /// Returns `true` if the deque is empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque = VecDeque::new(); + /// assert!(deque.is_empty()); + /// deque.push_front(1); + /// assert!(!deque.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_empty(&self) -> bool { + self.tail == self.head + } + + fn range_tail_head<R>(&self, range: R) -> (usize, usize) + where + R: RangeBounds<usize>, + { + let Range { start, end } = slice::range(range, ..self.len()); + let tail = self.wrap_add(self.tail, start); + let head = self.wrap_add(self.tail, end); + (tail, head) + } + + /// Creates an iterator that covers the specified range in the deque. + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<_> = [1, 2, 3].into(); + /// let range = deque.range(2..).copied().collect::<VecDeque<_>>(); + /// assert_eq!(range, [3]); + /// + /// // A full range covers all contents + /// let all = deque.range(..); + /// assert_eq!(all.len(), 3); + /// ``` + #[inline] + #[stable(feature = "deque_range", since = "1.51.0")] + pub fn range<R>(&self, range: R) -> Iter<'_, T> + where + R: RangeBounds<usize>, + { + let (tail, head) = self.range_tail_head(range); + // The shared reference we have in &self is maintained in the '_ of Iter. + Iter::new(unsafe { self.buffer_as_slice() }, tail, head) + } + + /// Creates an iterator that covers the specified mutable range in the deque. + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque: VecDeque<_> = [1, 2, 3].into(); + /// for v in deque.range_mut(2..) { + /// *v *= 2; + /// } + /// assert_eq!(deque, [1, 2, 6]); + /// + /// // A full range covers all contents + /// for v in deque.range_mut(..) { + /// *v *= 2; + /// } + /// assert_eq!(deque, [2, 4, 12]); + /// ``` + #[inline] + #[stable(feature = "deque_range", since = "1.51.0")] + pub fn range_mut<R>(&mut self, range: R) -> IterMut<'_, T> + where + R: RangeBounds<usize>, + { + let (tail, head) = self.range_tail_head(range); + + // SAFETY: The internal `IterMut` safety invariant is established because the + // `ring` we create is a dereferenceable slice for lifetime '_. + let ring = ptr::slice_from_raw_parts_mut(self.ptr(), self.cap()); + + unsafe { IterMut::new(ring, tail, head, PhantomData) } + } + + /// Removes the specified range from the deque in bulk, returning all + /// removed elements as an iterator. If the iterator is dropped before + /// being fully consumed, it drops the remaining removed elements. + /// + /// The returned iterator keeps a mutable borrow on the queue to optimize + /// its implementation. + /// + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the deque. + /// + /// # Leaking + /// + /// If the returned iterator goes out of scope without being dropped (due to + /// [`mem::forget`], for example), the deque may have lost and leaked + /// elements arbitrarily, including elements outside the range. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque: VecDeque<_> = [1, 2, 3].into(); + /// let drained = deque.drain(2..).collect::<VecDeque<_>>(); + /// assert_eq!(drained, [3]); + /// assert_eq!(deque, [1, 2]); + /// + /// // A full range clears all contents, like `clear()` does + /// deque.drain(..); + /// assert!(deque.is_empty()); + /// ``` + #[inline] + #[stable(feature = "drain", since = "1.6.0")] + pub fn drain<R>(&mut self, range: R) -> Drain<'_, T, A> + where + R: RangeBounds<usize>, + { + // Memory safety + // + // When the Drain is first created, the source deque is shortened to + // make sure no uninitialized or moved-from elements are accessible at + // all if the Drain's destructor never gets to run. + // + // Drain will ptr::read out the values to remove. + // When finished, the remaining data will be copied back to cover the hole, + // and the head/tail values will be restored correctly. + // + let (drain_tail, drain_head) = self.range_tail_head(range); + + // The deque's elements are parted into three segments: + // * self.tail -> drain_tail + // * drain_tail -> drain_head + // * drain_head -> self.head + // + // T = self.tail; H = self.head; t = drain_tail; h = drain_head + // + // We store drain_tail as self.head, and drain_head and self.head as + // after_tail and after_head respectively on the Drain. This also + // truncates the effective array such that if the Drain is leaked, we + // have forgotten about the potentially moved values after the start of + // the drain. + // + // T t h H + // [. . . o o x x o o . . .] + // + let head = self.head; + + // "forget" about the values after the start of the drain until after + // the drain is complete and the Drain destructor is run. + self.head = drain_tail; + + let deque = NonNull::from(&mut *self); + unsafe { + // Crucially, we only create shared references from `self` here and read from + // it. We do not write to `self` nor reborrow to a mutable reference. + // Hence the raw pointer we created above, for `deque`, remains valid. + let ring = self.buffer_as_slice(); + let iter = Iter::new(ring, drain_tail, drain_head); + + Drain::new(drain_head, head, iter, deque) + } + } + + /// Clears the deque, removing all values. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque = VecDeque::new(); + /// deque.push_back(1); + /// deque.clear(); + /// assert!(deque.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn clear(&mut self) { + self.truncate(0); + } + + /// Returns `true` if the deque contains an element equal to the + /// given value. + /// + /// This operation is *O*(*n*). + /// + /// Note that if you have a sorted `VecDeque`, [`binary_search`] may be faster. + /// + /// [`binary_search`]: VecDeque::binary_search + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque: VecDeque<u32> = VecDeque::new(); + /// + /// deque.push_back(0); + /// deque.push_back(1); + /// + /// assert_eq!(deque.contains(&1), true); + /// assert_eq!(deque.contains(&10), false); + /// ``` + #[stable(feature = "vec_deque_contains", since = "1.12.0")] + pub fn contains(&self, x: &T) -> bool + where + T: PartialEq<T>, + { + let (a, b) = self.as_slices(); + a.contains(x) || b.contains(x) + } + + /// Provides a reference to the front element, or `None` if the deque is + /// empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// assert_eq!(d.front(), None); + /// + /// d.push_back(1); + /// d.push_back(2); + /// assert_eq!(d.front(), Some(&1)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn front(&self) -> Option<&T> { + self.get(0) + } + + /// Provides a mutable reference to the front element, or `None` if the + /// deque is empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// assert_eq!(d.front_mut(), None); + /// + /// d.push_back(1); + /// d.push_back(2); + /// match d.front_mut() { + /// Some(x) => *x = 9, + /// None => (), + /// } + /// assert_eq!(d.front(), Some(&9)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn front_mut(&mut self) -> Option<&mut T> { + self.get_mut(0) + } + + /// Provides a reference to the back element, or `None` if the deque is + /// empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// assert_eq!(d.back(), None); + /// + /// d.push_back(1); + /// d.push_back(2); + /// assert_eq!(d.back(), Some(&2)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn back(&self) -> Option<&T> { + self.get(self.len().wrapping_sub(1)) + } + + /// Provides a mutable reference to the back element, or `None` if the + /// deque is empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// assert_eq!(d.back(), None); + /// + /// d.push_back(1); + /// d.push_back(2); + /// match d.back_mut() { + /// Some(x) => *x = 9, + /// None => (), + /// } + /// assert_eq!(d.back(), Some(&9)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn back_mut(&mut self) -> Option<&mut T> { + self.get_mut(self.len().wrapping_sub(1)) + } + + /// Removes the first element and returns it, or `None` if the deque is + /// empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// d.push_back(1); + /// d.push_back(2); + /// + /// assert_eq!(d.pop_front(), Some(1)); + /// assert_eq!(d.pop_front(), Some(2)); + /// assert_eq!(d.pop_front(), None); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop_front(&mut self) -> Option<T> { + if self.is_empty() { + None + } else { + let tail = self.tail; + self.tail = self.wrap_add(self.tail, 1); + unsafe { Some(self.buffer_read(tail)) } + } + } + + /// Removes the last element from the deque and returns it, or `None` if + /// it is empty. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// assert_eq!(buf.pop_back(), None); + /// buf.push_back(1); + /// buf.push_back(3); + /// assert_eq!(buf.pop_back(), Some(3)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop_back(&mut self) -> Option<T> { + if self.is_empty() { + None + } else { + self.head = self.wrap_sub(self.head, 1); + let head = self.head; + unsafe { Some(self.buffer_read(head)) } + } + } + + /// Prepends an element to the deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut d = VecDeque::new(); + /// d.push_front(1); + /// d.push_front(2); + /// assert_eq!(d.front(), Some(&2)); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push_front(&mut self, value: T) { + if self.is_full() { + self.grow(); + } + + self.tail = self.wrap_sub(self.tail, 1); + let tail = self.tail; + unsafe { + self.buffer_write(tail, value); + } + } + + /// Appends an element to the back of the deque. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(1); + /// buf.push_back(3); + /// assert_eq!(3, *buf.back().unwrap()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push_back(&mut self, value: T) { + if self.is_full() { + self.grow(); + } + + let head = self.head; + self.head = self.wrap_add(self.head, 1); + unsafe { self.buffer_write(head, value) } + } + + #[inline] + fn is_contiguous(&self) -> bool { + // FIXME: Should we consider `head == 0` to mean + // that `self` is contiguous? + self.tail <= self.head + } + + /// Removes an element from anywhere in the deque and returns it, + /// replacing it with the first element. + /// + /// This does not preserve ordering, but is *O*(1). + /// + /// Returns `None` if `index` is out of bounds. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// assert_eq!(buf.swap_remove_front(0), None); + /// buf.push_back(1); + /// buf.push_back(2); + /// buf.push_back(3); + /// assert_eq!(buf, [1, 2, 3]); + /// + /// assert_eq!(buf.swap_remove_front(2), Some(3)); + /// assert_eq!(buf, [2, 1]); + /// ``` + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn swap_remove_front(&mut self, index: usize) -> Option<T> { + let length = self.len(); + if length > 0 && index < length && index != 0 { + self.swap(index, 0); + } else if index >= length { + return None; + } + self.pop_front() + } + + /// Removes an element from anywhere in the deque and returns it, + /// replacing it with the last element. + /// + /// This does not preserve ordering, but is *O*(1). + /// + /// Returns `None` if `index` is out of bounds. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// assert_eq!(buf.swap_remove_back(0), None); + /// buf.push_back(1); + /// buf.push_back(2); + /// buf.push_back(3); + /// assert_eq!(buf, [1, 2, 3]); + /// + /// assert_eq!(buf.swap_remove_back(0), Some(1)); + /// assert_eq!(buf, [3, 2]); + /// ``` + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn swap_remove_back(&mut self, index: usize) -> Option<T> { + let length = self.len(); + if length > 0 && index < length - 1 { + self.swap(index, length - 1); + } else if index >= length { + return None; + } + self.pop_back() + } + + /// Inserts an element at `index` within the deque, shifting all elements + /// with indices greater than or equal to `index` towards the back. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Panics + /// + /// Panics if `index` is greater than deque's length + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut vec_deque = VecDeque::new(); + /// vec_deque.push_back('a'); + /// vec_deque.push_back('b'); + /// vec_deque.push_back('c'); + /// assert_eq!(vec_deque, &['a', 'b', 'c']); + /// + /// vec_deque.insert(1, 'd'); + /// assert_eq!(vec_deque, &['a', 'd', 'b', 'c']); + /// ``` + #[stable(feature = "deque_extras_15", since = "1.5.0")] + pub fn insert(&mut self, index: usize, value: T) { + assert!(index <= self.len(), "index out of bounds"); + if self.is_full() { + self.grow(); + } + + // Move the least number of elements in the ring buffer and insert + // the given object + // + // At most len/2 - 1 elements will be moved. O(min(n, n-i)) + // + // There are three main cases: + // Elements are contiguous + // - special case when tail is 0 + // Elements are discontiguous and the insert is in the tail section + // Elements are discontiguous and the insert is in the head section + // + // For each of those there are two more cases: + // Insert is closer to tail + // Insert is closer to head + // + // Key: H - self.head + // T - self.tail + // o - Valid element + // I - Insertion element + // A - The element that should be after the insertion point + // M - Indicates element was moved + + let idx = self.wrap_add(self.tail, index); + + let distance_to_tail = index; + let distance_to_head = self.len() - index; + + let contiguous = self.is_contiguous(); + + match (contiguous, distance_to_tail <= distance_to_head, idx >= self.tail) { + (true, true, _) if index == 0 => { + // push_front + // + // T + // I H + // [A o o o o o o . . . . . . . . .] + // + // H T + // [A o o o o o o o . . . . . I] + // + + self.tail = self.wrap_sub(self.tail, 1); + } + (true, true, _) => { + unsafe { + // contiguous, insert closer to tail: + // + // T I H + // [. . . o o A o o o o . . . . . .] + // + // T H + // [. . o o I A o o o o . . . . . .] + // M M + // + // contiguous, insert closer to tail and tail is 0: + // + // + // T I H + // [o o A o o o o . . . . . . . . .] + // + // H T + // [o I A o o o o o . . . . . . . o] + // M M + + let new_tail = self.wrap_sub(self.tail, 1); + + self.copy(new_tail, self.tail, 1); + // Already moved the tail, so we only copy `index - 1` elements. + self.copy(self.tail, self.tail + 1, index - 1); + + self.tail = new_tail; + } + } + (true, false, _) => { + unsafe { + // contiguous, insert closer to head: + // + // T I H + // [. . . o o o o A o o . . . . . .] + // + // T H + // [. . . o o o o I A o o . . . . .] + // M M M + + self.copy(idx + 1, idx, self.head - idx); + self.head = self.wrap_add(self.head, 1); + } + } + (false, true, true) => { + unsafe { + // discontiguous, insert closer to tail, tail section: + // + // H T I + // [o o o o o o . . . . . o o A o o] + // + // H T + // [o o o o o o . . . . o o I A o o] + // M M + + self.copy(self.tail - 1, self.tail, index); + self.tail -= 1; + } + } + (false, false, true) => { + unsafe { + // discontiguous, insert closer to head, tail section: + // + // H T I + // [o o . . . . . . . o o o o o A o] + // + // H T + // [o o o . . . . . . o o o o o I A] + // M M M M + + // copy elements up to new head + self.copy(1, 0, self.head); + + // copy last element into empty spot at bottom of buffer + self.copy(0, self.cap() - 1, 1); + + // move elements from idx to end forward not including ^ element + self.copy(idx + 1, idx, self.cap() - 1 - idx); + + self.head += 1; + } + } + (false, true, false) if idx == 0 => { + unsafe { + // discontiguous, insert is closer to tail, head section, + // and is at index zero in the internal buffer: + // + // I H T + // [A o o o o o o o o o . . . o o o] + // + // H T + // [A o o o o o o o o o . . o o o I] + // M M M + + // copy elements up to new tail + self.copy(self.tail - 1, self.tail, self.cap() - self.tail); + + // copy last element into empty spot at bottom of buffer + self.copy(self.cap() - 1, 0, 1); + + self.tail -= 1; + } + } + (false, true, false) => { + unsafe { + // discontiguous, insert closer to tail, head section: + // + // I H T + // [o o o A o o o o o o . . . o o o] + // + // H T + // [o o I A o o o o o o . . o o o o] + // M M M M M M + + // copy elements up to new tail + self.copy(self.tail - 1, self.tail, self.cap() - self.tail); + + // copy last element into empty spot at bottom of buffer + self.copy(self.cap() - 1, 0, 1); + + // move elements from idx-1 to end forward not including ^ element + self.copy(0, 1, idx - 1); + + self.tail -= 1; + } + } + (false, false, false) => { + unsafe { + // discontiguous, insert closer to head, head section: + // + // I H T + // [o o o o A o o . . . . . . o o o] + // + // H T + // [o o o o I A o o . . . . . o o o] + // M M M + + self.copy(idx + 1, idx, self.head - idx); + self.head += 1; + } + } + } + + // tail might've been changed so we need to recalculate + let new_idx = self.wrap_add(self.tail, index); + unsafe { + self.buffer_write(new_idx, value); + } + } + + /// Removes and returns the element at `index` from the deque. + /// Whichever end is closer to the removal point will be moved to make + /// room, and all the affected elements will be moved to new positions. + /// Returns `None` if `index` is out of bounds. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(1); + /// buf.push_back(2); + /// buf.push_back(3); + /// assert_eq!(buf, [1, 2, 3]); + /// + /// assert_eq!(buf.remove(1), Some(2)); + /// assert_eq!(buf, [1, 3]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn remove(&mut self, index: usize) -> Option<T> { + if self.is_empty() || self.len() <= index { + return None; + } + + // There are three main cases: + // Elements are contiguous + // Elements are discontiguous and the removal is in the tail section + // Elements are discontiguous and the removal is in the head section + // - special case when elements are technically contiguous, + // but self.head = 0 + // + // For each of those there are two more cases: + // Insert is closer to tail + // Insert is closer to head + // + // Key: H - self.head + // T - self.tail + // o - Valid element + // x - Element marked for removal + // R - Indicates element that is being removed + // M - Indicates element was moved + + let idx = self.wrap_add(self.tail, index); + + let elem = unsafe { Some(self.buffer_read(idx)) }; + + let distance_to_tail = index; + let distance_to_head = self.len() - index; + + let contiguous = self.is_contiguous(); + + match (contiguous, distance_to_tail <= distance_to_head, idx >= self.tail) { + (true, true, _) => { + unsafe { + // contiguous, remove closer to tail: + // + // T R H + // [. . . o o x o o o o . . . . . .] + // + // T H + // [. . . . o o o o o o . . . . . .] + // M M + + self.copy(self.tail + 1, self.tail, index); + self.tail += 1; + } + } + (true, false, _) => { + unsafe { + // contiguous, remove closer to head: + // + // T R H + // [. . . o o o o x o o . . . . . .] + // + // T H + // [. . . o o o o o o . . . . . . .] + // M M + + self.copy(idx, idx + 1, self.head - idx - 1); + self.head -= 1; + } + } + (false, true, true) => { + unsafe { + // discontiguous, remove closer to tail, tail section: + // + // H T R + // [o o o o o o . . . . . o o x o o] + // + // H T + // [o o o o o o . . . . . . o o o o] + // M M + + self.copy(self.tail + 1, self.tail, index); + self.tail = self.wrap_add(self.tail, 1); + } + } + (false, false, false) => { + unsafe { + // discontiguous, remove closer to head, head section: + // + // R H T + // [o o o o x o o . . . . . . o o o] + // + // H T + // [o o o o o o . . . . . . . o o o] + // M M + + self.copy(idx, idx + 1, self.head - idx - 1); + self.head -= 1; + } + } + (false, false, true) => { + unsafe { + // discontiguous, remove closer to head, tail section: + // + // H T R + // [o o o . . . . . . o o o o o x o] + // + // H T + // [o o . . . . . . . o o o o o o o] + // M M M M + // + // or quasi-discontiguous, remove next to head, tail section: + // + // H T R + // [. . . . . . . . . o o o o o x o] + // + // T H + // [. . . . . . . . . o o o o o o .] + // M + + // draw in elements in the tail section + self.copy(idx, idx + 1, self.cap() - idx - 1); + + // Prevents underflow. + if self.head != 0 { + // copy first element into empty spot + self.copy(self.cap() - 1, 0, 1); + + // move elements in the head section backwards + self.copy(0, 1, self.head - 1); + } + + self.head = self.wrap_sub(self.head, 1); + } + } + (false, true, false) => { + unsafe { + // discontiguous, remove closer to tail, head section: + // + // R H T + // [o o x o o o o o o o . . . o o o] + // + // H T + // [o o o o o o o o o o . . . . o o] + // M M M M M + + // draw in elements up to idx + self.copy(1, 0, idx); + + // copy last element into empty spot + self.copy(0, self.cap() - 1, 1); + + // move elements from tail to end forward, excluding the last one + self.copy(self.tail + 1, self.tail, self.cap() - self.tail - 1); + + self.tail = self.wrap_add(self.tail, 1); + } + } + } + + elem + } + + /// Splits the deque into two at the given index. + /// + /// Returns a newly allocated `VecDeque`. `self` contains elements `[0, at)`, + /// and the returned deque contains elements `[at, len)`. + /// + /// Note that the capacity of `self` does not change. + /// + /// Element at index 0 is the front of the queue. + /// + /// # Panics + /// + /// Panics if `at > len`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<_> = [1, 2, 3].into(); + /// let buf2 = buf.split_off(1); + /// assert_eq!(buf, [1]); + /// assert_eq!(buf2, [2, 3]); + /// ``` + #[inline] + #[must_use = "use `.truncate()` if you don't need the other half"] + #[stable(feature = "split_off", since = "1.4.0")] + pub fn split_off(&mut self, at: usize) -> Self + where + A: Clone, + { + let len = self.len(); + assert!(at <= len, "`at` out of bounds"); + + let other_len = len - at; + let mut other = VecDeque::with_capacity_in(other_len, self.allocator().clone()); + + unsafe { + let (first_half, second_half) = self.as_slices(); + + let first_len = first_half.len(); + let second_len = second_half.len(); + if at < first_len { + // `at` lies in the first half. + let amount_in_first = first_len - at; + + ptr::copy_nonoverlapping(first_half.as_ptr().add(at), other.ptr(), amount_in_first); + + // just take all of the second half. + ptr::copy_nonoverlapping( + second_half.as_ptr(), + other.ptr().add(amount_in_first), + second_len, + ); + } else { + // `at` lies in the second half, need to factor in the elements we skipped + // in the first half. + let offset = at - first_len; + let amount_in_second = second_len - offset; + ptr::copy_nonoverlapping( + second_half.as_ptr().add(offset), + other.ptr(), + amount_in_second, + ); + } + } + + // Cleanup where the ends of the buffers are + self.head = self.wrap_sub(self.head, other_len); + other.head = other.wrap_index(other_len); + + other + } + + /// Moves all the elements of `other` into `self`, leaving `other` empty. + /// + /// # Panics + /// + /// Panics if the new number of elements in self overflows a `usize`. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<_> = [1, 2].into(); + /// let mut buf2: VecDeque<_> = [3, 4].into(); + /// buf.append(&mut buf2); + /// assert_eq!(buf, [1, 2, 3, 4]); + /// assert_eq!(buf2, []); + /// ``` + #[inline] + #[stable(feature = "append", since = "1.4.0")] + pub fn append(&mut self, other: &mut Self) { + self.reserve(other.len()); + unsafe { + let (left, right) = other.as_slices(); + self.copy_slice(self.head, left); + self.copy_slice(self.wrap_add(self.head, left.len()), right); + } + // SAFETY: Update pointers after copying to avoid leaving doppelganger + // in case of panics. + self.head = self.wrap_add(self.head, other.len()); + // Silently drop values in `other`. + other.tail = other.head; + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all elements `e` for which `f(&e)` returns false. + /// This method operates in place, visiting each element exactly once in the + /// original order, and preserves the order of the retained elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.extend(1..5); + /// buf.retain(|&x| x % 2 == 0); + /// assert_eq!(buf, [2, 4]); + /// ``` + /// + /// Because the elements are visited exactly once in the original order, + /// external state may be used to decide which elements to keep. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.extend(1..6); + /// + /// let keep = [false, true, true, false, true]; + /// let mut iter = keep.iter(); + /// buf.retain(|_| *iter.next().unwrap()); + /// assert_eq!(buf, [2, 3, 5]); + /// ``` + #[stable(feature = "vec_deque_retain", since = "1.4.0")] + pub fn retain<F>(&mut self, mut f: F) + where + F: FnMut(&T) -> bool, + { + self.retain_mut(|elem| f(elem)); + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all elements `e` for which `f(&e)` returns false. + /// This method operates in place, visiting each element exactly once in the + /// original order, and preserves the order of the retained elements. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.extend(1..5); + /// buf.retain_mut(|x| if *x % 2 == 0 { + /// *x += 1; + /// true + /// } else { + /// false + /// }); + /// assert_eq!(buf, [3, 5]); + /// ``` + #[stable(feature = "vec_retain_mut", since = "1.61.0")] + pub fn retain_mut<F>(&mut self, mut f: F) + where + F: FnMut(&mut T) -> bool, + { + let len = self.len(); + let mut idx = 0; + let mut cur = 0; + + // Stage 1: All values are retained. + while cur < len { + if !f(&mut self[cur]) { + cur += 1; + break; + } + cur += 1; + idx += 1; + } + // Stage 2: Swap retained value into current idx. + while cur < len { + if !f(&mut self[cur]) { + cur += 1; + continue; + } + + self.swap(idx, cur); + cur += 1; + idx += 1; + } + // Stage 3: Truncate all values after idx. + if cur != idx { + self.truncate(idx); + } + } + + // Double the buffer size. This method is inline(never), so we expect it to only + // be called in cold paths. + // This may panic or abort + #[inline(never)] + fn grow(&mut self) { + // Extend or possibly remove this assertion when valid use-cases for growing the + // buffer without it being full emerge + debug_assert!(self.is_full()); + let old_cap = self.cap(); + self.buf.reserve_exact(old_cap, old_cap); + assert!(self.cap() == old_cap * 2); + unsafe { + self.handle_capacity_increase(old_cap); + } + debug_assert!(!self.is_full()); + } + + /// Modifies the deque in-place so that `len()` is equal to `new_len`, + /// either by removing excess elements from the back or by appending + /// elements generated by calling `generator` to the back. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(5); + /// buf.push_back(10); + /// buf.push_back(15); + /// assert_eq!(buf, [5, 10, 15]); + /// + /// buf.resize_with(5, Default::default); + /// assert_eq!(buf, [5, 10, 15, 0, 0]); + /// + /// buf.resize_with(2, || unreachable!()); + /// assert_eq!(buf, [5, 10]); + /// + /// let mut state = 100; + /// buf.resize_with(5, || { state += 1; state }); + /// assert_eq!(buf, [5, 10, 101, 102, 103]); + /// ``` + #[stable(feature = "vec_resize_with", since = "1.33.0")] + pub fn resize_with(&mut self, new_len: usize, generator: impl FnMut() -> T) { + let len = self.len(); + + if new_len > len { + self.extend(repeat_with(generator).take(new_len - len)) + } else { + self.truncate(new_len); + } + } + + /// Rearranges the internal storage of this deque so it is one contiguous + /// slice, which is then returned. + /// + /// This method does not allocate and does not change the order of the + /// inserted elements. As it returns a mutable slice, this can be used to + /// sort a deque. + /// + /// Once the internal storage is contiguous, the [`as_slices`] and + /// [`as_mut_slices`] methods will return the entire contents of the + /// deque in a single slice. + /// + /// [`as_slices`]: VecDeque::as_slices + /// [`as_mut_slices`]: VecDeque::as_mut_slices + /// + /// # Examples + /// + /// Sorting the content of a deque. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::with_capacity(15); + /// + /// buf.push_back(2); + /// buf.push_back(1); + /// buf.push_front(3); + /// + /// // sorting the deque + /// buf.make_contiguous().sort(); + /// assert_eq!(buf.as_slices(), (&[1, 2, 3] as &[_], &[] as &[_])); + /// + /// // sorting it in reverse order + /// buf.make_contiguous().sort_by(|a, b| b.cmp(a)); + /// assert_eq!(buf.as_slices(), (&[3, 2, 1] as &[_], &[] as &[_])); + /// ``` + /// + /// Getting immutable access to the contiguous slice. + /// + /// ```rust + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// + /// buf.push_back(2); + /// buf.push_back(1); + /// buf.push_front(3); + /// + /// buf.make_contiguous(); + /// if let (slice, &[]) = buf.as_slices() { + /// // we can now be sure that `slice` contains all elements of the deque, + /// // while still having immutable access to `buf`. + /// assert_eq!(buf.len(), slice.len()); + /// assert_eq!(slice, &[3, 2, 1] as &[_]); + /// } + /// ``` + #[stable(feature = "deque_make_contiguous", since = "1.48.0")] + pub fn make_contiguous(&mut self) -> &mut [T] { + if self.is_contiguous() { + let tail = self.tail; + let head = self.head; + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + return unsafe { + MaybeUninit::slice_assume_init_mut( + RingSlices::ring_slices(self.buffer_as_mut_slice(), head, tail).0, + ) + }; + } + + let buf = self.buf.ptr(); + let cap = self.cap(); + let len = self.len(); + + let free = self.tail - self.head; + let tail_len = cap - self.tail; + + if free >= tail_len { + // there is enough free space to copy the tail in one go, + // this means that we first shift the head backwards, and then + // copy the tail to the correct position. + // + // from: DEFGH....ABC + // to: ABCDEFGH.... + unsafe { + ptr::copy(buf, buf.add(tail_len), self.head); + // ...DEFGH.ABC + ptr::copy_nonoverlapping(buf.add(self.tail), buf, tail_len); + // ABCDEFGH.... + + self.tail = 0; + self.head = len; + } + } else if free > self.head { + // FIXME: We currently do not consider ....ABCDEFGH + // to be contiguous because `head` would be `0` in this + // case. While we probably want to change this it + // isn't trivial as a few places expect `is_contiguous` + // to mean that we can just slice using `buf[tail..head]`. + + // there is enough free space to copy the head in one go, + // this means that we first shift the tail forwards, and then + // copy the head to the correct position. + // + // from: FGH....ABCDE + // to: ...ABCDEFGH. + unsafe { + ptr::copy(buf.add(self.tail), buf.add(self.head), tail_len); + // FGHABCDE.... + ptr::copy_nonoverlapping(buf, buf.add(self.head + tail_len), self.head); + // ...ABCDEFGH. + + self.tail = self.head; + self.head = self.wrap_add(self.tail, len); + } + } else { + // free is smaller than both head and tail, + // this means we have to slowly "swap" the tail and the head. + // + // from: EFGHI...ABCD or HIJK.ABCDEFG + // to: ABCDEFGHI... or ABCDEFGHIJK. + let mut left_edge: usize = 0; + let mut right_edge: usize = self.tail; + unsafe { + // The general problem looks like this + // GHIJKLM...ABCDEF - before any swaps + // ABCDEFM...GHIJKL - after 1 pass of swaps + // ABCDEFGHIJM...KL - swap until the left edge reaches the temp store + // - then restart the algorithm with a new (smaller) store + // Sometimes the temp store is reached when the right edge is at the end + // of the buffer - this means we've hit the right order with fewer swaps! + // E.g + // EF..ABCD + // ABCDEF.. - after four only swaps we've finished + while left_edge < len && right_edge != cap { + let mut right_offset = 0; + for i in left_edge..right_edge { + right_offset = (i - left_edge) % (cap - right_edge); + let src: isize = (right_edge + right_offset) as isize; + ptr::swap(buf.add(i), buf.offset(src)); + } + let n_ops = right_edge - left_edge; + left_edge += n_ops; + right_edge += right_offset + 1; + } + + self.tail = 0; + self.head = len; + } + } + + let tail = self.tail; + let head = self.head; + // Safety: + // - `self.head` and `self.tail` in a ring buffer are always valid indices. + // - `RingSlices::ring_slices` guarantees that the slices split according to `self.head` and `self.tail` are initialized. + unsafe { + MaybeUninit::slice_assume_init_mut( + RingSlices::ring_slices(self.buffer_as_mut_slice(), head, tail).0, + ) + } + } + + /// Rotates the double-ended queue `mid` places to the left. + /// + /// Equivalently, + /// - Rotates item `mid` into the first position. + /// - Pops the first `mid` items and pushes them to the end. + /// - Rotates `len() - mid` places to the right. + /// + /// # Panics + /// + /// If `mid` is greater than `len()`. Note that `mid == len()` + /// does _not_ panic and is a no-op rotation. + /// + /// # Complexity + /// + /// Takes `*O*(min(mid, len() - mid))` time and no extra space. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<_> = (0..10).collect(); + /// + /// buf.rotate_left(3); + /// assert_eq!(buf, [3, 4, 5, 6, 7, 8, 9, 0, 1, 2]); + /// + /// for i in 1..10 { + /// assert_eq!(i * 3 % 10, buf[0]); + /// buf.rotate_left(3); + /// } + /// assert_eq!(buf, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]); + /// ``` + #[stable(feature = "vecdeque_rotate", since = "1.36.0")] + pub fn rotate_left(&mut self, mid: usize) { + assert!(mid <= self.len()); + let k = self.len() - mid; + if mid <= k { + unsafe { self.rotate_left_inner(mid) } + } else { + unsafe { self.rotate_right_inner(k) } + } + } + + /// Rotates the double-ended queue `k` places to the right. + /// + /// Equivalently, + /// - Rotates the first item into position `k`. + /// - Pops the last `k` items and pushes them to the front. + /// - Rotates `len() - k` places to the left. + /// + /// # Panics + /// + /// If `k` is greater than `len()`. Note that `k == len()` + /// does _not_ panic and is a no-op rotation. + /// + /// # Complexity + /// + /// Takes `*O*(min(k, len() - k))` time and no extra space. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf: VecDeque<_> = (0..10).collect(); + /// + /// buf.rotate_right(3); + /// assert_eq!(buf, [7, 8, 9, 0, 1, 2, 3, 4, 5, 6]); + /// + /// for i in 1..10 { + /// assert_eq!(0, buf[i * 3 % 10]); + /// buf.rotate_right(3); + /// } + /// assert_eq!(buf, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]); + /// ``` + #[stable(feature = "vecdeque_rotate", since = "1.36.0")] + pub fn rotate_right(&mut self, k: usize) { + assert!(k <= self.len()); + let mid = self.len() - k; + if k <= mid { + unsafe { self.rotate_right_inner(k) } + } else { + unsafe { self.rotate_left_inner(mid) } + } + } + + // SAFETY: the following two methods require that the rotation amount + // be less than half the length of the deque. + // + // `wrap_copy` requires that `min(x, cap() - x) + copy_len <= cap()`, + // but than `min` is never more than half the capacity, regardless of x, + // so it's sound to call here because we're calling with something + // less than half the length, which is never above half the capacity. + + unsafe fn rotate_left_inner(&mut self, mid: usize) { + debug_assert!(mid * 2 <= self.len()); + unsafe { + self.wrap_copy(self.head, self.tail, mid); + } + self.head = self.wrap_add(self.head, mid); + self.tail = self.wrap_add(self.tail, mid); + } + + unsafe fn rotate_right_inner(&mut self, k: usize) { + debug_assert!(k * 2 <= self.len()); + self.head = self.wrap_sub(self.head, k); + self.tail = self.wrap_sub(self.tail, k); + unsafe { + self.wrap_copy(self.tail, self.head, k); + } + } + + /// Binary searches this `VecDeque` for a given element. + /// This behaves similarly to [`contains`] if this `VecDeque` is sorted. + /// + /// If the value is found then [`Result::Ok`] is returned, containing the + /// index of the matching element. If there are multiple matches, then any + /// one of the matches could be returned. If the value is not found then + /// [`Result::Err`] is returned, containing the index where a matching + /// element could be inserted while maintaining sorted order. + /// + /// See also [`binary_search_by`], [`binary_search_by_key`], and [`partition_point`]. + /// + /// [`contains`]: VecDeque::contains + /// [`binary_search_by`]: VecDeque::binary_search_by + /// [`binary_search_by_key`]: VecDeque::binary_search_by_key + /// [`partition_point`]: VecDeque::partition_point + /// + /// # Examples + /// + /// Looks up a series of four elements. The first is found, with a + /// uniquely determined position; the second and third are not + /// found; the fourth could match any position in `[1, 4]`. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<_> = [0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + /// + /// assert_eq!(deque.binary_search(&13), Ok(9)); + /// assert_eq!(deque.binary_search(&4), Err(7)); + /// assert_eq!(deque.binary_search(&100), Err(13)); + /// let r = deque.binary_search(&1); + /// assert!(matches!(r, Ok(1..=4))); + /// ``` + /// + /// If you want to insert an item to a sorted deque, while maintaining + /// sort order, consider using [`partition_point`]: + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque: VecDeque<_> = [0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + /// let num = 42; + /// let idx = deque.partition_point(|&x| x < num); + /// // The above is equivalent to `let idx = deque.binary_search(&num).unwrap_or_else(|x| x);` + /// deque.insert(idx, num); + /// assert_eq!(deque, &[0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 42, 55]); + /// ``` + #[stable(feature = "vecdeque_binary_search", since = "1.54.0")] + #[inline] + pub fn binary_search(&self, x: &T) -> Result<usize, usize> + where + T: Ord, + { + self.binary_search_by(|e| e.cmp(x)) + } + + /// Binary searches this `VecDeque` with a comparator function. + /// This behaves similarly to [`contains`] if this `VecDeque` is sorted. + /// + /// The comparator function should implement an order consistent + /// with the sort order of the deque, returning an order code that + /// indicates whether its argument is `Less`, `Equal` or `Greater` + /// than the desired target. + /// + /// If the value is found then [`Result::Ok`] is returned, containing the + /// index of the matching element. If there are multiple matches, then any + /// one of the matches could be returned. If the value is not found then + /// [`Result::Err`] is returned, containing the index where a matching + /// element could be inserted while maintaining sorted order. + /// + /// See also [`binary_search`], [`binary_search_by_key`], and [`partition_point`]. + /// + /// [`contains`]: VecDeque::contains + /// [`binary_search`]: VecDeque::binary_search + /// [`binary_search_by_key`]: VecDeque::binary_search_by_key + /// [`partition_point`]: VecDeque::partition_point + /// + /// # Examples + /// + /// Looks up a series of four elements. The first is found, with a + /// uniquely determined position; the second and third are not + /// found; the fourth could match any position in `[1, 4]`. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<_> = [0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + /// + /// assert_eq!(deque.binary_search_by(|x| x.cmp(&13)), Ok(9)); + /// assert_eq!(deque.binary_search_by(|x| x.cmp(&4)), Err(7)); + /// assert_eq!(deque.binary_search_by(|x| x.cmp(&100)), Err(13)); + /// let r = deque.binary_search_by(|x| x.cmp(&1)); + /// assert!(matches!(r, Ok(1..=4))); + /// ``` + #[stable(feature = "vecdeque_binary_search", since = "1.54.0")] + pub fn binary_search_by<'a, F>(&'a self, mut f: F) -> Result<usize, usize> + where + F: FnMut(&'a T) -> Ordering, + { + let (front, back) = self.as_slices(); + let cmp_back = back.first().map(|elem| f(elem)); + + if let Some(Ordering::Equal) = cmp_back { + Ok(front.len()) + } else if let Some(Ordering::Less) = cmp_back { + back.binary_search_by(f).map(|idx| idx + front.len()).map_err(|idx| idx + front.len()) + } else { + front.binary_search_by(f) + } + } + + /// Binary searches this `VecDeque` with a key extraction function. + /// This behaves similarly to [`contains`] if this `VecDeque` is sorted. + /// + /// Assumes that the deque is sorted by the key, for instance with + /// [`make_contiguous().sort_by_key()`] using the same key extraction function. + /// + /// If the value is found then [`Result::Ok`] is returned, containing the + /// index of the matching element. If there are multiple matches, then any + /// one of the matches could be returned. If the value is not found then + /// [`Result::Err`] is returned, containing the index where a matching + /// element could be inserted while maintaining sorted order. + /// + /// See also [`binary_search`], [`binary_search_by`], and [`partition_point`]. + /// + /// [`contains`]: VecDeque::contains + /// [`make_contiguous().sort_by_key()`]: VecDeque::make_contiguous + /// [`binary_search`]: VecDeque::binary_search + /// [`binary_search_by`]: VecDeque::binary_search_by + /// [`partition_point`]: VecDeque::partition_point + /// + /// # Examples + /// + /// Looks up a series of four elements in a slice of pairs sorted by + /// their second elements. The first is found, with a uniquely + /// determined position; the second and third are not found; the + /// fourth could match any position in `[1, 4]`. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<_> = [(0, 0), (2, 1), (4, 1), (5, 1), + /// (3, 1), (1, 2), (2, 3), (4, 5), (5, 8), (3, 13), + /// (1, 21), (2, 34), (4, 55)].into(); + /// + /// assert_eq!(deque.binary_search_by_key(&13, |&(a, b)| b), Ok(9)); + /// assert_eq!(deque.binary_search_by_key(&4, |&(a, b)| b), Err(7)); + /// assert_eq!(deque.binary_search_by_key(&100, |&(a, b)| b), Err(13)); + /// let r = deque.binary_search_by_key(&1, |&(a, b)| b); + /// assert!(matches!(r, Ok(1..=4))); + /// ``` + #[stable(feature = "vecdeque_binary_search", since = "1.54.0")] + #[inline] + pub fn binary_search_by_key<'a, B, F>(&'a self, b: &B, mut f: F) -> Result<usize, usize> + where + F: FnMut(&'a T) -> B, + B: Ord, + { + self.binary_search_by(|k| f(k).cmp(b)) + } + + /// Returns the index of the partition point according to the given predicate + /// (the index of the first element of the second partition). + /// + /// The deque is assumed to be partitioned according to the given predicate. + /// This means that all elements for which the predicate returns true are at the start of the deque + /// and all elements for which the predicate returns false are at the end. + /// For example, [7, 15, 3, 5, 4, 12, 6] is a partitioned under the predicate x % 2 != 0 + /// (all odd numbers are at the start, all even at the end). + /// + /// If the deque is not partitioned, the returned result is unspecified and meaningless, + /// as this method performs a kind of binary search. + /// + /// See also [`binary_search`], [`binary_search_by`], and [`binary_search_by_key`]. + /// + /// [`binary_search`]: VecDeque::binary_search + /// [`binary_search_by`]: VecDeque::binary_search_by + /// [`binary_search_by_key`]: VecDeque::binary_search_by_key + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deque: VecDeque<_> = [1, 2, 3, 3, 5, 6, 7].into(); + /// let i = deque.partition_point(|&x| x < 5); + /// + /// assert_eq!(i, 4); + /// assert!(deque.iter().take(i).all(|&x| x < 5)); + /// assert!(deque.iter().skip(i).all(|&x| !(x < 5))); + /// ``` + /// + /// If you want to insert an item to a sorted deque, while maintaining + /// sort order: + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut deque: VecDeque<_> = [0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + /// let num = 42; + /// let idx = deque.partition_point(|&x| x < num); + /// deque.insert(idx, num); + /// assert_eq!(deque, &[0, 1, 1, 1, 1, 2, 3, 5, 8, 13, 21, 34, 42, 55]); + /// ``` + #[stable(feature = "vecdeque_binary_search", since = "1.54.0")] + pub fn partition_point<P>(&self, mut pred: P) -> usize + where + P: FnMut(&T) -> bool, + { + let (front, back) = self.as_slices(); + + if let Some(true) = back.first().map(|v| pred(v)) { + back.partition_point(pred) + front.len() + } else { + front.partition_point(pred) + } + } +} + +impl<T: Clone, A: Allocator> VecDeque<T, A> { + /// Modifies the deque in-place so that `len()` is equal to new_len, + /// either by removing excess elements from the back or by appending clones of `value` + /// to the back. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let mut buf = VecDeque::new(); + /// buf.push_back(5); + /// buf.push_back(10); + /// buf.push_back(15); + /// assert_eq!(buf, [5, 10, 15]); + /// + /// buf.resize(2, 0); + /// assert_eq!(buf, [5, 10]); + /// + /// buf.resize(5, 20); + /// assert_eq!(buf, [5, 10, 20, 20, 20]); + /// ``` + #[stable(feature = "deque_extras", since = "1.16.0")] + pub fn resize(&mut self, new_len: usize, value: T) { + self.resize_with(new_len, || value.clone()); + } +} + +/// Returns the index in the underlying buffer for a given logical element index. +#[inline] +fn wrap_index(index: usize, size: usize) -> usize { + // size is always a power of 2 + debug_assert!(size.is_power_of_two()); + index & (size - 1) +} + +/// Calculate the number of elements left to be read in the buffer +#[inline] +fn count(tail: usize, head: usize, size: usize) -> usize { + // size is always a power of 2 + (head.wrapping_sub(tail)) & (size - 1) +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialEq, A: Allocator> PartialEq for VecDeque<T, A> { + fn eq(&self, other: &Self) -> bool { + if self.len() != other.len() { + return false; + } + let (sa, sb) = self.as_slices(); + let (oa, ob) = other.as_slices(); + if sa.len() == oa.len() { + sa == oa && sb == ob + } else if sa.len() < oa.len() { + // Always divisible in three sections, for example: + // self: [a b c|d e f] + // other: [0 1 2 3|4 5] + // front = 3, mid = 1, + // [a b c] == [0 1 2] && [d] == [3] && [e f] == [4 5] + let front = sa.len(); + let mid = oa.len() - front; + + let (oa_front, oa_mid) = oa.split_at(front); + let (sb_mid, sb_back) = sb.split_at(mid); + debug_assert_eq!(sa.len(), oa_front.len()); + debug_assert_eq!(sb_mid.len(), oa_mid.len()); + debug_assert_eq!(sb_back.len(), ob.len()); + sa == oa_front && sb_mid == oa_mid && sb_back == ob + } else { + let front = oa.len(); + let mid = sa.len() - front; + + let (sa_front, sa_mid) = sa.split_at(front); + let (ob_mid, ob_back) = ob.split_at(mid); + debug_assert_eq!(sa_front.len(), oa.len()); + debug_assert_eq!(sa_mid.len(), ob_mid.len()); + debug_assert_eq!(sb.len(), ob_back.len()); + sa_front == oa && sa_mid == ob_mid && sb == ob_back + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Eq, A: Allocator> Eq for VecDeque<T, A> {} + +__impl_slice_eq1! { [] VecDeque<T, A>, Vec<U, A>, } +__impl_slice_eq1! { [] VecDeque<T, A>, &[U], } +__impl_slice_eq1! { [] VecDeque<T, A>, &mut [U], } +__impl_slice_eq1! { [const N: usize] VecDeque<T, A>, [U; N], } +__impl_slice_eq1! { [const N: usize] VecDeque<T, A>, &[U; N], } +__impl_slice_eq1! { [const N: usize] VecDeque<T, A>, &mut [U; N], } + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialOrd, A: Allocator> PartialOrd for VecDeque<T, A> { + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + self.iter().partial_cmp(other.iter()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord, A: Allocator> Ord for VecDeque<T, A> { + #[inline] + fn cmp(&self, other: &Self) -> Ordering { + self.iter().cmp(other.iter()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Hash, A: Allocator> Hash for VecDeque<T, A> { + fn hash<H: Hasher>(&self, state: &mut H) { + state.write_length_prefix(self.len()); + // It's not possible to use Hash::hash_slice on slices + // returned by as_slices method as their length can vary + // in otherwise identical deques. + // + // Hasher only guarantees equivalence for the exact same + // set of calls to its methods. + self.iter().for_each(|elem| elem.hash(state)); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Index<usize> for VecDeque<T, A> { + type Output = T; + + #[inline] + fn index(&self, index: usize) -> &T { + self.get(index).expect("Out of bounds access") + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> IndexMut<usize> for VecDeque<T, A> { + #[inline] + fn index_mut(&mut self, index: usize) -> &mut T { + self.get_mut(index).expect("Out of bounds access") + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> FromIterator<T> for VecDeque<T> { + fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> VecDeque<T> { + let iterator = iter.into_iter(); + let (lower, _) = iterator.size_hint(); + let mut deq = VecDeque::with_capacity(lower); + deq.extend(iterator); + deq + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> IntoIterator for VecDeque<T, A> { + type Item = T; + type IntoIter = IntoIter<T, A>; + + /// Consumes the deque into a front-to-back iterator yielding elements by + /// value. + fn into_iter(self) -> IntoIter<T, A> { + IntoIter::new(self) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T, A: Allocator> IntoIterator for &'a VecDeque<T, A> { + type Item = &'a T; + type IntoIter = Iter<'a, T>; + + fn into_iter(self) -> Iter<'a, T> { + self.iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T, A: Allocator> IntoIterator for &'a mut VecDeque<T, A> { + type Item = &'a mut T; + type IntoIter = IterMut<'a, T>; + + fn into_iter(self) -> IterMut<'a, T> { + self.iter_mut() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Extend<T> for VecDeque<T, A> { + fn extend<I: IntoIterator<Item = T>>(&mut self, iter: I) { + <Self as SpecExtend<T, I::IntoIter>>::spec_extend(self, iter.into_iter()); + } + + #[inline] + fn extend_one(&mut self, elem: T) { + self.push_back(elem); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, T: 'a + Copy, A: Allocator> Extend<&'a T> for VecDeque<T, A> { + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iter: I) { + self.spec_extend(iter.into_iter()); + } + + #[inline] + fn extend_one(&mut self, &elem: &T) { + self.push_back(elem); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for VecDeque<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_list().entries(self).finish() + } +} + +#[stable(feature = "vecdeque_vec_conversions", since = "1.10.0")] +impl<T, A: Allocator> From<Vec<T, A>> for VecDeque<T, A> { + /// Turn a [`Vec<T>`] into a [`VecDeque<T>`]. + /// + /// [`Vec<T>`]: crate::vec::Vec + /// [`VecDeque<T>`]: crate::collections::VecDeque + /// + /// This avoids reallocating where possible, but the conditions for that are + /// strict, and subject to change, and so shouldn't be relied upon unless the + /// `Vec<T>` came from `From<VecDeque<T>>` and hasn't been reallocated. + fn from(mut other: Vec<T, A>) -> Self { + let len = other.len(); + if mem::size_of::<T>() == 0 { + // There's no actual allocation for ZSTs to worry about capacity, + // but `VecDeque` can't handle as much length as `Vec`. + assert!(len < MAXIMUM_ZST_CAPACITY, "capacity overflow"); + } else { + // We need to resize if the capacity is not a power of two, too small or + // doesn't have at least one free space. We do this while it's still in + // the `Vec` so the items will drop on panic. + let min_cap = cmp::max(MINIMUM_CAPACITY, len) + 1; + let cap = cmp::max(min_cap, other.capacity()).next_power_of_two(); + if other.capacity() != cap { + other.reserve_exact(cap - len); + } + } + + unsafe { + let (other_buf, len, capacity, alloc) = other.into_raw_parts_with_alloc(); + let buf = RawVec::from_raw_parts_in(other_buf, capacity, alloc); + VecDeque { tail: 0, head: len, buf } + } + } +} + +#[stable(feature = "vecdeque_vec_conversions", since = "1.10.0")] +impl<T, A: Allocator> From<VecDeque<T, A>> for Vec<T, A> { + /// Turn a [`VecDeque<T>`] into a [`Vec<T>`]. + /// + /// [`Vec<T>`]: crate::vec::Vec + /// [`VecDeque<T>`]: crate::collections::VecDeque + /// + /// This never needs to re-allocate, but does need to do *O*(*n*) data movement if + /// the circular buffer doesn't happen to be at the beginning of the allocation. + /// + /// # Examples + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// // This one is *O*(1). + /// let deque: VecDeque<_> = (1..5).collect(); + /// let ptr = deque.as_slices().0.as_ptr(); + /// let vec = Vec::from(deque); + /// assert_eq!(vec, [1, 2, 3, 4]); + /// assert_eq!(vec.as_ptr(), ptr); + /// + /// // This one needs data rearranging. + /// let mut deque: VecDeque<_> = (1..5).collect(); + /// deque.push_front(9); + /// deque.push_front(8); + /// let ptr = deque.as_slices().1.as_ptr(); + /// let vec = Vec::from(deque); + /// assert_eq!(vec, [8, 9, 1, 2, 3, 4]); + /// assert_eq!(vec.as_ptr(), ptr); + /// ``` + fn from(mut other: VecDeque<T, A>) -> Self { + other.make_contiguous(); + + unsafe { + let other = ManuallyDrop::new(other); + let buf = other.buf.ptr(); + let len = other.len(); + let cap = other.cap(); + let alloc = ptr::read(other.allocator()); + + if other.tail != 0 { + ptr::copy(buf.add(other.tail), buf, len); + } + Vec::from_raw_parts_in(buf, len, cap, alloc) + } + } +} + +#[stable(feature = "std_collections_from_array", since = "1.56.0")] +impl<T, const N: usize> From<[T; N]> for VecDeque<T> { + /// Converts a `[T; N]` into a `VecDeque<T>`. + /// + /// ``` + /// use std::collections::VecDeque; + /// + /// let deq1 = VecDeque::from([1, 2, 3, 4]); + /// let deq2: VecDeque<_> = [1, 2, 3, 4].into(); + /// assert_eq!(deq1, deq2); + /// ``` + fn from(arr: [T; N]) -> Self { + let mut deq = VecDeque::with_capacity(N); + let arr = ManuallyDrop::new(arr); + if mem::size_of::<T>() != 0 { + // SAFETY: VecDeque::with_capacity ensures that there is enough capacity. + unsafe { + ptr::copy_nonoverlapping(arr.as_ptr(), deq.ptr(), N); + } + } + deq.tail = 0; + deq.head = N; + deq + } +} diff --git a/library/alloc/src/collections/vec_deque/pair_slices.rs b/library/alloc/src/collections/vec_deque/pair_slices.rs new file mode 100644 index 000000000..6735424a3 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/pair_slices.rs @@ -0,0 +1,67 @@ +use core::cmp::{self}; +use core::mem::replace; + +use crate::alloc::Allocator; + +use super::VecDeque; + +/// PairSlices pairs up equal length slice parts of two deques +/// +/// For example, given deques "A" and "B" with the following division into slices: +/// +/// A: [0 1 2] [3 4 5] +/// B: [a b] [c d e] +/// +/// It produces the following sequence of matching slices: +/// +/// ([0 1], [a b]) +/// (\[2\], \[c\]) +/// ([3 4], [d e]) +/// +/// and the uneven remainder of either A or B is skipped. +pub struct PairSlices<'a, 'b, T> { + a0: &'a mut [T], + a1: &'a mut [T], + b0: &'b [T], + b1: &'b [T], +} + +impl<'a, 'b, T> PairSlices<'a, 'b, T> { + pub fn from<A: Allocator>(to: &'a mut VecDeque<T, A>, from: &'b VecDeque<T, A>) -> Self { + let (a0, a1) = to.as_mut_slices(); + let (b0, b1) = from.as_slices(); + PairSlices { a0, a1, b0, b1 } + } + + pub fn has_remainder(&self) -> bool { + !self.b0.is_empty() + } + + pub fn remainder(self) -> impl Iterator<Item = &'b [T]> { + IntoIterator::into_iter([self.b0, self.b1]) + } +} + +impl<'a, 'b, T> Iterator for PairSlices<'a, 'b, T> { + type Item = (&'a mut [T], &'b [T]); + fn next(&mut self) -> Option<Self::Item> { + // Get next part length + let part = cmp::min(self.a0.len(), self.b0.len()); + if part == 0 { + return None; + } + let (p0, p1) = replace(&mut self.a0, &mut []).split_at_mut(part); + let (q0, q1) = self.b0.split_at(part); + + // Move a1 into a0, if it's empty (and b1, b0 the same way). + self.a0 = p1; + self.b0 = q1; + if self.a0.is_empty() { + self.a0 = replace(&mut self.a1, &mut []); + } + if self.b0.is_empty() { + self.b0 = replace(&mut self.b1, &[]); + } + Some((p0, q0)) + } +} diff --git a/library/alloc/src/collections/vec_deque/ring_slices.rs b/library/alloc/src/collections/vec_deque/ring_slices.rs new file mode 100644 index 000000000..dd0fa7d60 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/ring_slices.rs @@ -0,0 +1,56 @@ +use core::ptr::{self}; + +/// Returns the two slices that cover the `VecDeque`'s valid range +pub trait RingSlices: Sized { + fn slice(self, from: usize, to: usize) -> Self; + fn split_at(self, i: usize) -> (Self, Self); + + fn ring_slices(buf: Self, head: usize, tail: usize) -> (Self, Self) { + let contiguous = tail <= head; + if contiguous { + let (empty, buf) = buf.split_at(0); + (buf.slice(tail, head), empty) + } else { + let (mid, right) = buf.split_at(tail); + let (left, _) = mid.split_at(head); + (right, left) + } + } +} + +impl<T> RingSlices for &[T] { + fn slice(self, from: usize, to: usize) -> Self { + &self[from..to] + } + fn split_at(self, i: usize) -> (Self, Self) { + (*self).split_at(i) + } +} + +impl<T> RingSlices for &mut [T] { + fn slice(self, from: usize, to: usize) -> Self { + &mut self[from..to] + } + fn split_at(self, i: usize) -> (Self, Self) { + (*self).split_at_mut(i) + } +} + +impl<T> RingSlices for *mut [T] { + fn slice(self, from: usize, to: usize) -> Self { + assert!(from <= to && to < self.len()); + // Not using `get_unchecked_mut` to keep this a safe operation. + let len = to - from; + ptr::slice_from_raw_parts_mut(self.as_mut_ptr().wrapping_add(from), len) + } + + fn split_at(self, mid: usize) -> (Self, Self) { + let len = self.len(); + let ptr = self.as_mut_ptr(); + assert!(mid <= len); + ( + ptr::slice_from_raw_parts_mut(ptr, mid), + ptr::slice_from_raw_parts_mut(ptr.wrapping_add(mid), len - mid), + ) + } +} diff --git a/library/alloc/src/collections/vec_deque/spec_extend.rs b/library/alloc/src/collections/vec_deque/spec_extend.rs new file mode 100644 index 000000000..97ff8b765 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/spec_extend.rs @@ -0,0 +1,132 @@ +use crate::alloc::Allocator; +use crate::vec; +use core::iter::{ByRefSized, TrustedLen}; +use core::slice; + +use super::VecDeque; + +// Specialization trait used for VecDeque::extend +pub(super) trait SpecExtend<T, I> { + fn spec_extend(&mut self, iter: I); +} + +impl<T, I, A: Allocator> SpecExtend<T, I> for VecDeque<T, A> +where + I: Iterator<Item = T>, +{ + default fn spec_extend(&mut self, mut iter: I) { + // This function should be the moral equivalent of: + // + // for item in iter { + // self.push_back(item); + // } + while let Some(element) = iter.next() { + if self.len() == self.capacity() { + let (lower, _) = iter.size_hint(); + self.reserve(lower.saturating_add(1)); + } + + let head = self.head; + self.head = self.wrap_add(self.head, 1); + unsafe { + self.buffer_write(head, element); + } + } + } +} + +impl<T, I, A: Allocator> SpecExtend<T, I> for VecDeque<T, A> +where + I: TrustedLen<Item = T>, +{ + default fn spec_extend(&mut self, mut iter: I) { + // This is the case for a TrustedLen iterator. + let (low, high) = iter.size_hint(); + if let Some(additional) = high { + debug_assert_eq!( + low, + additional, + "TrustedLen iterator's size hint is not exact: {:?}", + (low, high) + ); + self.reserve(additional); + + struct WrapAddOnDrop<'a, T, A: Allocator> { + vec_deque: &'a mut VecDeque<T, A>, + written: usize, + } + + impl<'a, T, A: Allocator> Drop for WrapAddOnDrop<'a, T, A> { + fn drop(&mut self) { + self.vec_deque.head = + self.vec_deque.wrap_add(self.vec_deque.head, self.written); + } + } + + let mut wrapper = WrapAddOnDrop { vec_deque: self, written: 0 }; + + let head_room = wrapper.vec_deque.cap() - wrapper.vec_deque.head; + unsafe { + wrapper.vec_deque.write_iter( + wrapper.vec_deque.head, + ByRefSized(&mut iter).take(head_room), + &mut wrapper.written, + ); + + if additional > head_room { + wrapper.vec_deque.write_iter(0, iter, &mut wrapper.written); + } + } + + debug_assert_eq!( + additional, wrapper.written, + "The number of items written to VecDeque doesn't match the TrustedLen size hint" + ); + } else { + // Per TrustedLen contract a `None` upper bound means that the iterator length + // truly exceeds usize::MAX, which would eventually lead to a capacity overflow anyway. + // Since the other branch already panics eagerly (via `reserve()`) we do the same here. + // This avoids additional codegen for a fallback code path which would eventually + // panic anyway. + panic!("capacity overflow"); + } + } +} + +impl<T, A: Allocator> SpecExtend<T, vec::IntoIter<T>> for VecDeque<T, A> { + fn spec_extend(&mut self, mut iterator: vec::IntoIter<T>) { + let slice = iterator.as_slice(); + self.reserve(slice.len()); + + unsafe { + self.copy_slice(self.head, slice); + self.head = self.wrap_add(self.head, slice.len()); + } + iterator.forget_remaining_elements(); + } +} + +impl<'a, T: 'a, I, A: Allocator> SpecExtend<&'a T, I> for VecDeque<T, A> +where + I: Iterator<Item = &'a T>, + T: Copy, +{ + default fn spec_extend(&mut self, iterator: I) { + self.spec_extend(iterator.copied()) + } +} + +impl<'a, T: 'a, A: Allocator> SpecExtend<&'a T, slice::Iter<'a, T>> for VecDeque<T, A> +where + T: Copy, +{ + fn spec_extend(&mut self, iterator: slice::Iter<'a, T>) { + let slice = iterator.as_slice(); + self.reserve(slice.len()); + + unsafe { + self.copy_slice(self.head, slice); + self.head = self.wrap_add(self.head, slice.len()); + } + } +} diff --git a/library/alloc/src/collections/vec_deque/tests.rs b/library/alloc/src/collections/vec_deque/tests.rs new file mode 100644 index 000000000..1f2daef21 --- /dev/null +++ b/library/alloc/src/collections/vec_deque/tests.rs @@ -0,0 +1,1110 @@ +use core::iter::TrustedLen; + +use super::*; + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_push_back_100(b: &mut test::Bencher) { + let mut deq = VecDeque::with_capacity(101); + b.iter(|| { + for i in 0..100 { + deq.push_back(i); + } + deq.head = 0; + deq.tail = 0; + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_push_front_100(b: &mut test::Bencher) { + let mut deq = VecDeque::with_capacity(101); + b.iter(|| { + for i in 0..100 { + deq.push_front(i); + } + deq.head = 0; + deq.tail = 0; + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_pop_back_100(b: &mut test::Bencher) { + let mut deq = VecDeque::<i32>::with_capacity(101); + + b.iter(|| { + deq.head = 100; + deq.tail = 0; + while !deq.is_empty() { + test::black_box(deq.pop_back()); + } + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_retain_whole_10000(b: &mut test::Bencher) { + let v = (1..100000).collect::<VecDeque<u32>>(); + + b.iter(|| { + let mut v = v.clone(); + v.retain(|x| *x > 0) + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_retain_odd_10000(b: &mut test::Bencher) { + let v = (1..100000).collect::<VecDeque<u32>>(); + + b.iter(|| { + let mut v = v.clone(); + v.retain(|x| x & 1 == 0) + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_retain_half_10000(b: &mut test::Bencher) { + let v = (1..100000).collect::<VecDeque<u32>>(); + + b.iter(|| { + let mut v = v.clone(); + v.retain(|x| *x > 50000) + }) +} + +#[bench] +#[cfg_attr(miri, ignore)] // isolated Miri does not support benchmarks +fn bench_pop_front_100(b: &mut test::Bencher) { + let mut deq = VecDeque::<i32>::with_capacity(101); + + b.iter(|| { + deq.head = 100; + deq.tail = 0; + while !deq.is_empty() { + test::black_box(deq.pop_front()); + } + }) +} + +#[test] +fn test_swap_front_back_remove() { + fn test(back: bool) { + // This test checks that every single combination of tail position and length is tested. + // Capacity 15 should be large enough to cover every case. + let mut tester = VecDeque::with_capacity(15); + let usable_cap = tester.capacity(); + let final_len = usable_cap / 2; + + for len in 0..final_len { + let expected: VecDeque<_> = + if back { (0..len).collect() } else { (0..len).rev().collect() }; + for tail_pos in 0..usable_cap { + tester.tail = tail_pos; + tester.head = tail_pos; + if back { + for i in 0..len * 2 { + tester.push_front(i); + } + for i in 0..len { + assert_eq!(tester.swap_remove_back(i), Some(len * 2 - 1 - i)); + } + } else { + for i in 0..len * 2 { + tester.push_back(i); + } + for i in 0..len { + let idx = tester.len() - 1 - i; + assert_eq!(tester.swap_remove_front(idx), Some(len * 2 - 1 - i)); + } + } + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + assert_eq!(tester, expected); + } + } + } + test(true); + test(false); +} + +#[test] +fn test_insert() { + // This test checks that every single combination of tail position, length, and + // insertion position is tested. Capacity 15 should be large enough to cover every case. + + let mut tester = VecDeque::with_capacity(15); + // can't guarantee we got 15, so have to get what we got. + // 15 would be great, but we will definitely get 2^k - 1, for k >= 4, or else + // this test isn't covering what it wants to + let cap = tester.capacity(); + + // len is the length *after* insertion + let minlen = if cfg!(miri) { cap - 1 } else { 1 }; // Miri is too slow + for len in minlen..cap { + // 0, 1, 2, .., len - 1 + let expected = (0..).take(len).collect::<VecDeque<_>>(); + for tail_pos in 0..cap { + for to_insert in 0..len { + tester.tail = tail_pos; + tester.head = tail_pos; + for i in 0..len { + if i != to_insert { + tester.push_back(i); + } + } + tester.insert(to_insert, to_insert); + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + assert_eq!(tester, expected); + } + } + } +} + +#[test] +fn test_get() { + let mut tester = VecDeque::new(); + tester.push_back(1); + tester.push_back(2); + tester.push_back(3); + + assert_eq!(tester.len(), 3); + + assert_eq!(tester.get(1), Some(&2)); + assert_eq!(tester.get(2), Some(&3)); + assert_eq!(tester.get(0), Some(&1)); + assert_eq!(tester.get(3), None); + + tester.remove(0); + + assert_eq!(tester.len(), 2); + assert_eq!(tester.get(0), Some(&2)); + assert_eq!(tester.get(1), Some(&3)); + assert_eq!(tester.get(2), None); +} + +#[test] +fn test_get_mut() { + let mut tester = VecDeque::new(); + tester.push_back(1); + tester.push_back(2); + tester.push_back(3); + + assert_eq!(tester.len(), 3); + + if let Some(elem) = tester.get_mut(0) { + assert_eq!(*elem, 1); + *elem = 10; + } + + if let Some(elem) = tester.get_mut(2) { + assert_eq!(*elem, 3); + *elem = 30; + } + + assert_eq!(tester.get(0), Some(&10)); + assert_eq!(tester.get(2), Some(&30)); + assert_eq!(tester.get_mut(3), None); + + tester.remove(2); + + assert_eq!(tester.len(), 2); + assert_eq!(tester.get(0), Some(&10)); + assert_eq!(tester.get(1), Some(&2)); + assert_eq!(tester.get(2), None); +} + +#[test] +fn test_swap() { + let mut tester = VecDeque::new(); + tester.push_back(1); + tester.push_back(2); + tester.push_back(3); + + assert_eq!(tester, [1, 2, 3]); + + tester.swap(0, 0); + assert_eq!(tester, [1, 2, 3]); + tester.swap(0, 1); + assert_eq!(tester, [2, 1, 3]); + tester.swap(2, 1); + assert_eq!(tester, [2, 3, 1]); + tester.swap(1, 2); + assert_eq!(tester, [2, 1, 3]); + tester.swap(0, 2); + assert_eq!(tester, [3, 1, 2]); + tester.swap(2, 2); + assert_eq!(tester, [3, 1, 2]); +} + +#[test] +#[should_panic = "assertion failed: j < self.len()"] +fn test_swap_panic() { + let mut tester = VecDeque::new(); + tester.push_back(1); + tester.push_back(2); + tester.push_back(3); + tester.swap(2, 3); +} + +#[test] +fn test_reserve_exact() { + let mut tester: VecDeque<i32> = VecDeque::with_capacity(1); + assert!(tester.capacity() == 1); + tester.reserve_exact(50); + assert!(tester.capacity() >= 51); + tester.reserve_exact(40); + assert!(tester.capacity() >= 51); + tester.reserve_exact(200); + assert!(tester.capacity() >= 200); +} + +#[test] +#[should_panic = "capacity overflow"] +fn test_reserve_exact_panic() { + let mut tester: VecDeque<i32> = VecDeque::new(); + tester.reserve_exact(usize::MAX); +} + +#[test] +fn test_try_reserve_exact() { + let mut tester: VecDeque<i32> = VecDeque::with_capacity(1); + assert!(tester.capacity() == 1); + assert_eq!(tester.try_reserve_exact(100), Ok(())); + assert!(tester.capacity() >= 100); + assert_eq!(tester.try_reserve_exact(50), Ok(())); + assert!(tester.capacity() >= 100); + assert_eq!(tester.try_reserve_exact(200), Ok(())); + assert!(tester.capacity() >= 200); + assert_eq!(tester.try_reserve_exact(0), Ok(())); + assert!(tester.capacity() >= 200); + assert!(tester.try_reserve_exact(usize::MAX).is_err()); +} + +#[test] +fn test_try_reserve() { + let mut tester: VecDeque<i32> = VecDeque::with_capacity(1); + assert!(tester.capacity() == 1); + assert_eq!(tester.try_reserve(100), Ok(())); + assert!(tester.capacity() >= 100); + assert_eq!(tester.try_reserve(50), Ok(())); + assert!(tester.capacity() >= 100); + assert_eq!(tester.try_reserve(200), Ok(())); + assert!(tester.capacity() >= 200); + assert_eq!(tester.try_reserve(0), Ok(())); + assert!(tester.capacity() >= 200); + assert!(tester.try_reserve(usize::MAX).is_err()); +} + +#[test] +fn test_contains() { + let mut tester = VecDeque::new(); + tester.push_back(1); + tester.push_back(2); + tester.push_back(3); + + assert!(tester.contains(&1)); + assert!(tester.contains(&3)); + assert!(!tester.contains(&0)); + assert!(!tester.contains(&4)); + tester.remove(0); + assert!(!tester.contains(&1)); + assert!(tester.contains(&2)); + assert!(tester.contains(&3)); +} + +#[test] +fn test_rotate_left_right() { + let mut tester: VecDeque<_> = (1..=10).collect(); + + assert_eq!(tester.len(), 10); + + tester.rotate_left(0); + assert_eq!(tester, [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]); + + tester.rotate_right(0); + assert_eq!(tester, [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]); + + tester.rotate_left(3); + assert_eq!(tester, [4, 5, 6, 7, 8, 9, 10, 1, 2, 3]); + + tester.rotate_right(5); + assert_eq!(tester, [9, 10, 1, 2, 3, 4, 5, 6, 7, 8]); + + tester.rotate_left(tester.len()); + assert_eq!(tester, [9, 10, 1, 2, 3, 4, 5, 6, 7, 8]); + + tester.rotate_right(tester.len()); + assert_eq!(tester, [9, 10, 1, 2, 3, 4, 5, 6, 7, 8]); + + tester.rotate_left(1); + assert_eq!(tester, [10, 1, 2, 3, 4, 5, 6, 7, 8, 9]); +} + +#[test] +#[should_panic = "assertion failed: mid <= self.len()"] +fn test_rotate_left_panic() { + let mut tester: VecDeque<_> = (1..=10).collect(); + tester.rotate_left(tester.len() + 1); +} + +#[test] +#[should_panic = "assertion failed: k <= self.len()"] +fn test_rotate_right_panic() { + let mut tester: VecDeque<_> = (1..=10).collect(); + tester.rotate_right(tester.len() + 1); +} + +#[test] +fn test_binary_search() { + // If the givin VecDeque is not sorted, the returned result is unspecified and meaningless, + // as this method performs a binary search. + + let tester: VecDeque<_> = [0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + + assert_eq!(tester.binary_search(&0), Ok(0)); + assert_eq!(tester.binary_search(&5), Ok(5)); + assert_eq!(tester.binary_search(&55), Ok(10)); + assert_eq!(tester.binary_search(&4), Err(5)); + assert_eq!(tester.binary_search(&-1), Err(0)); + assert!(matches!(tester.binary_search(&1), Ok(1..=2))); + + let tester: VecDeque<_> = [1, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3].into(); + assert_eq!(tester.binary_search(&1), Ok(0)); + assert!(matches!(tester.binary_search(&2), Ok(1..=4))); + assert!(matches!(tester.binary_search(&3), Ok(5..=13))); + assert_eq!(tester.binary_search(&-2), Err(0)); + assert_eq!(tester.binary_search(&0), Err(0)); + assert_eq!(tester.binary_search(&4), Err(14)); + assert_eq!(tester.binary_search(&5), Err(14)); +} + +#[test] +fn test_binary_search_by() { + // If the givin VecDeque is not sorted, the returned result is unspecified and meaningless, + // as this method performs a binary search. + + let tester: VecDeque<_> = [0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55].into(); + + assert_eq!(tester.binary_search_by(|x| x.cmp(&0)), Ok(0)); + assert_eq!(tester.binary_search_by(|x| x.cmp(&5)), Ok(5)); + assert_eq!(tester.binary_search_by(|x| x.cmp(&55)), Ok(10)); + assert_eq!(tester.binary_search_by(|x| x.cmp(&4)), Err(5)); + assert_eq!(tester.binary_search_by(|x| x.cmp(&-1)), Err(0)); + assert!(matches!(tester.binary_search_by(|x| x.cmp(&1)), Ok(1..=2))); +} + +#[test] +fn test_binary_search_key() { + // If the givin VecDeque is not sorted, the returned result is unspecified and meaningless, + // as this method performs a binary search. + + let tester: VecDeque<_> = [ + (-1, 0), + (2, 10), + (6, 5), + (7, 1), + (8, 10), + (10, 2), + (20, 3), + (24, 5), + (25, 18), + (28, 13), + (31, 21), + (32, 4), + (54, 25), + ] + .into(); + + assert_eq!(tester.binary_search_by_key(&-1, |&(a, _b)| a), Ok(0)); + assert_eq!(tester.binary_search_by_key(&8, |&(a, _b)| a), Ok(4)); + assert_eq!(tester.binary_search_by_key(&25, |&(a, _b)| a), Ok(8)); + assert_eq!(tester.binary_search_by_key(&54, |&(a, _b)| a), Ok(12)); + assert_eq!(tester.binary_search_by_key(&-2, |&(a, _b)| a), Err(0)); + assert_eq!(tester.binary_search_by_key(&1, |&(a, _b)| a), Err(1)); + assert_eq!(tester.binary_search_by_key(&4, |&(a, _b)| a), Err(2)); + assert_eq!(tester.binary_search_by_key(&13, |&(a, _b)| a), Err(6)); + assert_eq!(tester.binary_search_by_key(&55, |&(a, _b)| a), Err(13)); + assert_eq!(tester.binary_search_by_key(&100, |&(a, _b)| a), Err(13)); + + let tester: VecDeque<_> = [ + (0, 0), + (2, 1), + (6, 1), + (5, 1), + (3, 1), + (1, 2), + (2, 3), + (4, 5), + (5, 8), + (8, 13), + (1, 21), + (2, 34), + (4, 55), + ] + .into(); + + assert_eq!(tester.binary_search_by_key(&0, |&(_a, b)| b), Ok(0)); + assert!(matches!(tester.binary_search_by_key(&1, |&(_a, b)| b), Ok(1..=4))); + assert_eq!(tester.binary_search_by_key(&8, |&(_a, b)| b), Ok(8)); + assert_eq!(tester.binary_search_by_key(&13, |&(_a, b)| b), Ok(9)); + assert_eq!(tester.binary_search_by_key(&55, |&(_a, b)| b), Ok(12)); + assert_eq!(tester.binary_search_by_key(&-1, |&(_a, b)| b), Err(0)); + assert_eq!(tester.binary_search_by_key(&4, |&(_a, b)| b), Err(7)); + assert_eq!(tester.binary_search_by_key(&56, |&(_a, b)| b), Err(13)); + assert_eq!(tester.binary_search_by_key(&100, |&(_a, b)| b), Err(13)); +} + +#[test] +fn make_contiguous_big_tail() { + let mut tester = VecDeque::with_capacity(15); + + for i in 0..3 { + tester.push_back(i); + } + + for i in 3..10 { + tester.push_front(i); + } + + // 012......9876543 + assert_eq!(tester.capacity(), 15); + assert_eq!((&[9, 8, 7, 6, 5, 4, 3] as &[_], &[0, 1, 2] as &[_]), tester.as_slices()); + + let expected_start = tester.head; + tester.make_contiguous(); + assert_eq!(tester.tail, expected_start); + assert_eq!((&[9, 8, 7, 6, 5, 4, 3, 0, 1, 2] as &[_], &[] as &[_]), tester.as_slices()); +} + +#[test] +fn make_contiguous_big_head() { + let mut tester = VecDeque::with_capacity(15); + + for i in 0..8 { + tester.push_back(i); + } + + for i in 8..10 { + tester.push_front(i); + } + + // 01234567......98 + let expected_start = 0; + tester.make_contiguous(); + assert_eq!(tester.tail, expected_start); + assert_eq!((&[9, 8, 0, 1, 2, 3, 4, 5, 6, 7] as &[_], &[] as &[_]), tester.as_slices()); +} + +#[test] +fn make_contiguous_small_free() { + let mut tester = VecDeque::with_capacity(15); + + for i in 'A' as u8..'I' as u8 { + tester.push_back(i as char); + } + + for i in 'I' as u8..'N' as u8 { + tester.push_front(i as char); + } + + // ABCDEFGH...MLKJI + let expected_start = 0; + tester.make_contiguous(); + assert_eq!(tester.tail, expected_start); + assert_eq!( + (&['M', 'L', 'K', 'J', 'I', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H'] as &[_], &[] as &[_]), + tester.as_slices() + ); + + tester.clear(); + for i in 'I' as u8..'N' as u8 { + tester.push_back(i as char); + } + + for i in 'A' as u8..'I' as u8 { + tester.push_front(i as char); + } + + // IJKLM...HGFEDCBA + let expected_start = 0; + tester.make_contiguous(); + assert_eq!(tester.tail, expected_start); + assert_eq!( + (&['H', 'G', 'F', 'E', 'D', 'C', 'B', 'A', 'I', 'J', 'K', 'L', 'M'] as &[_], &[] as &[_]), + tester.as_slices() + ); +} + +#[test] +fn make_contiguous_head_to_end() { + let mut dq = VecDeque::with_capacity(3); + dq.push_front('B'); + dq.push_front('A'); + dq.push_back('C'); + dq.make_contiguous(); + let expected_tail = 0; + let expected_head = 3; + assert_eq!(expected_tail, dq.tail); + assert_eq!(expected_head, dq.head); + assert_eq!((&['A', 'B', 'C'] as &[_], &[] as &[_]), dq.as_slices()); +} + +#[test] +fn make_contiguous_head_to_end_2() { + // Another test case for #79808, taken from #80293. + + let mut dq = VecDeque::from_iter(0..6); + dq.pop_front(); + dq.pop_front(); + dq.push_back(6); + dq.push_back(7); + dq.push_back(8); + dq.make_contiguous(); + let collected: Vec<_> = dq.iter().copied().collect(); + assert_eq!(dq.as_slices(), (&collected[..], &[] as &[_])); +} + +#[test] +fn test_remove() { + // This test checks that every single combination of tail position, length, and + // removal position is tested. Capacity 15 should be large enough to cover every case. + + let mut tester = VecDeque::with_capacity(15); + // can't guarantee we got 15, so have to get what we got. + // 15 would be great, but we will definitely get 2^k - 1, for k >= 4, or else + // this test isn't covering what it wants to + let cap = tester.capacity(); + + // len is the length *after* removal + let minlen = if cfg!(miri) { cap - 2 } else { 0 }; // Miri is too slow + for len in minlen..cap - 1 { + // 0, 1, 2, .., len - 1 + let expected = (0..).take(len).collect::<VecDeque<_>>(); + for tail_pos in 0..cap { + for to_remove in 0..=len { + tester.tail = tail_pos; + tester.head = tail_pos; + for i in 0..len { + if i == to_remove { + tester.push_back(1234); + } + tester.push_back(i); + } + if to_remove == len { + tester.push_back(1234); + } + tester.remove(to_remove); + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + assert_eq!(tester, expected); + } + } + } +} + +#[test] +fn test_range() { + let mut tester: VecDeque<usize> = VecDeque::with_capacity(7); + + let cap = tester.capacity(); + let minlen = if cfg!(miri) { cap - 1 } else { 0 }; // Miri is too slow + for len in minlen..=cap { + for tail in 0..=cap { + for start in 0..=len { + for end in start..=len { + tester.tail = tail; + tester.head = tail; + for i in 0..len { + tester.push_back(i); + } + + // Check that we iterate over the correct values + let range: VecDeque<_> = tester.range(start..end).copied().collect(); + let expected: VecDeque<_> = (start..end).collect(); + assert_eq!(range, expected); + } + } + } + } +} + +#[test] +fn test_range_mut() { + let mut tester: VecDeque<usize> = VecDeque::with_capacity(7); + + let cap = tester.capacity(); + for len in 0..=cap { + for tail in 0..=cap { + for start in 0..=len { + for end in start..=len { + tester.tail = tail; + tester.head = tail; + for i in 0..len { + tester.push_back(i); + } + + let head_was = tester.head; + let tail_was = tester.tail; + + // Check that we iterate over the correct values + let range: VecDeque<_> = tester.range_mut(start..end).map(|v| *v).collect(); + let expected: VecDeque<_> = (start..end).collect(); + assert_eq!(range, expected); + + // We shouldn't have changed the capacity or made the + // head or tail out of bounds + assert_eq!(tester.capacity(), cap); + assert_eq!(tester.tail, tail_was); + assert_eq!(tester.head, head_was); + } + } + } + } +} + +#[test] +fn test_drain() { + let mut tester: VecDeque<usize> = VecDeque::with_capacity(7); + + let cap = tester.capacity(); + for len in 0..=cap { + for tail in 0..=cap { + for drain_start in 0..=len { + for drain_end in drain_start..=len { + tester.tail = tail; + tester.head = tail; + for i in 0..len { + tester.push_back(i); + } + + // Check that we drain the correct values + let drained: VecDeque<_> = tester.drain(drain_start..drain_end).collect(); + let drained_expected: VecDeque<_> = (drain_start..drain_end).collect(); + assert_eq!(drained, drained_expected); + + // We shouldn't have changed the capacity or made the + // head or tail out of bounds + assert_eq!(tester.capacity(), cap); + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + + // We should see the correct values in the VecDeque + let expected: VecDeque<_> = (0..drain_start).chain(drain_end..len).collect(); + assert_eq!(expected, tester); + } + } + } + } +} + +#[test] +fn test_shrink_to_fit() { + // This test checks that every single combination of head and tail position, + // is tested. Capacity 15 should be large enough to cover every case. + + let mut tester = VecDeque::with_capacity(15); + // can't guarantee we got 15, so have to get what we got. + // 15 would be great, but we will definitely get 2^k - 1, for k >= 4, or else + // this test isn't covering what it wants to + let cap = tester.capacity(); + tester.reserve(63); + let max_cap = tester.capacity(); + + for len in 0..=cap { + // 0, 1, 2, .., len - 1 + let expected = (0..).take(len).collect::<VecDeque<_>>(); + for tail_pos in 0..=max_cap { + tester.tail = tail_pos; + tester.head = tail_pos; + tester.reserve(63); + for i in 0..len { + tester.push_back(i); + } + tester.shrink_to_fit(); + assert!(tester.capacity() <= cap); + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + assert_eq!(tester, expected); + } + } +} + +#[test] +fn test_split_off() { + // This test checks that every single combination of tail position, length, and + // split position is tested. Capacity 15 should be large enough to cover every case. + + let mut tester = VecDeque::with_capacity(15); + // can't guarantee we got 15, so have to get what we got. + // 15 would be great, but we will definitely get 2^k - 1, for k >= 4, or else + // this test isn't covering what it wants to + let cap = tester.capacity(); + + // len is the length *before* splitting + let minlen = if cfg!(miri) { cap - 1 } else { 0 }; // Miri is too slow + for len in minlen..cap { + // index to split at + for at in 0..=len { + // 0, 1, 2, .., at - 1 (may be empty) + let expected_self = (0..).take(at).collect::<VecDeque<_>>(); + // at, at + 1, .., len - 1 (may be empty) + let expected_other = (at..).take(len - at).collect::<VecDeque<_>>(); + + for tail_pos in 0..cap { + tester.tail = tail_pos; + tester.head = tail_pos; + for i in 0..len { + tester.push_back(i); + } + let result = tester.split_off(at); + assert!(tester.tail < tester.cap()); + assert!(tester.head < tester.cap()); + assert!(result.tail < result.cap()); + assert!(result.head < result.cap()); + assert_eq!(tester, expected_self); + assert_eq!(result, expected_other); + } + } + } +} + +#[test] +fn test_from_vec() { + use crate::vec::Vec; + for cap in 0..35 { + for len in 0..=cap { + let mut vec = Vec::with_capacity(cap); + vec.extend(0..len); + + let vd = VecDeque::from(vec.clone()); + assert!(vd.cap().is_power_of_two()); + assert_eq!(vd.len(), vec.len()); + assert!(vd.into_iter().eq(vec)); + } + } + + let vec = Vec::from([(); MAXIMUM_ZST_CAPACITY - 1]); + let vd = VecDeque::from(vec.clone()); + assert!(vd.cap().is_power_of_two()); + assert_eq!(vd.len(), vec.len()); +} + +#[test] +fn test_extend_basic() { + test_extend_impl(false); +} + +#[test] +fn test_extend_trusted_len() { + test_extend_impl(true); +} + +fn test_extend_impl(trusted_len: bool) { + struct VecDequeTester { + test: VecDeque<usize>, + expected: VecDeque<usize>, + trusted_len: bool, + } + + impl VecDequeTester { + fn new(trusted_len: bool) -> Self { + Self { test: VecDeque::new(), expected: VecDeque::new(), trusted_len } + } + + fn test_extend<I>(&mut self, iter: I) + where + I: Iterator<Item = usize> + TrustedLen + Clone, + { + struct BasicIterator<I>(I); + impl<I> Iterator for BasicIterator<I> + where + I: Iterator<Item = usize>, + { + type Item = usize; + + fn next(&mut self) -> Option<Self::Item> { + self.0.next() + } + } + + if self.trusted_len { + self.test.extend(iter.clone()); + } else { + self.test.extend(BasicIterator(iter.clone())); + } + + for item in iter { + self.expected.push_back(item) + } + + assert_eq!(self.test, self.expected); + let (a1, b1) = self.test.as_slices(); + let (a2, b2) = self.expected.as_slices(); + assert_eq!(a1, a2); + assert_eq!(b1, b2); + } + + fn drain<R: RangeBounds<usize> + Clone>(&mut self, range: R) { + self.test.drain(range.clone()); + self.expected.drain(range); + + assert_eq!(self.test, self.expected); + } + + fn clear(&mut self) { + self.test.clear(); + self.expected.clear(); + } + + fn remaining_capacity(&self) -> usize { + self.test.capacity() - self.test.len() + } + } + + let mut tester = VecDequeTester::new(trusted_len); + + // Initial capacity + tester.test_extend(0..tester.remaining_capacity() - 1); + + // Grow + tester.test_extend(1024..2048); + + // Wrap around + tester.drain(..128); + + tester.test_extend(0..tester.remaining_capacity() - 1); + + // Continue + tester.drain(256..); + tester.test_extend(4096..8196); + + tester.clear(); + + // Start again + tester.test_extend(0..32); +} + +#[test] +#[should_panic = "capacity overflow"] +fn test_from_vec_zst_overflow() { + use crate::vec::Vec; + let vec = Vec::from([(); MAXIMUM_ZST_CAPACITY]); + let vd = VecDeque::from(vec.clone()); // no room for +1 + assert!(vd.cap().is_power_of_two()); + assert_eq!(vd.len(), vec.len()); +} + +#[test] +fn test_from_array() { + fn test<const N: usize>() { + let mut array: [usize; N] = [0; N]; + + for i in 0..N { + array[i] = i; + } + + let deq: VecDeque<_> = array.into(); + + for i in 0..N { + assert_eq!(deq[i], i); + } + + assert!(deq.cap().is_power_of_two()); + assert_eq!(deq.len(), N); + } + test::<0>(); + test::<1>(); + test::<2>(); + test::<32>(); + test::<35>(); + + let array = [(); MAXIMUM_ZST_CAPACITY - 1]; + let deq = VecDeque::from(array); + assert!(deq.cap().is_power_of_two()); + assert_eq!(deq.len(), MAXIMUM_ZST_CAPACITY - 1); +} + +#[test] +fn test_vec_from_vecdeque() { + use crate::vec::Vec; + + fn create_vec_and_test_convert(capacity: usize, offset: usize, len: usize) { + let mut vd = VecDeque::with_capacity(capacity); + for _ in 0..offset { + vd.push_back(0); + vd.pop_front(); + } + vd.extend(0..len); + + let vec: Vec<_> = Vec::from(vd.clone()); + assert_eq!(vec.len(), vd.len()); + assert!(vec.into_iter().eq(vd)); + } + + // Miri is too slow + let max_pwr = if cfg!(miri) { 5 } else { 7 }; + + for cap_pwr in 0..max_pwr { + // Make capacity as a (2^x)-1, so that the ring size is 2^x + let cap = (2i32.pow(cap_pwr) - 1) as usize; + + // In these cases there is enough free space to solve it with copies + for len in 0..((cap + 1) / 2) { + // Test contiguous cases + for offset in 0..(cap - len) { + create_vec_and_test_convert(cap, offset, len) + } + + // Test cases where block at end of buffer is bigger than block at start + for offset in (cap - len)..(cap - (len / 2)) { + create_vec_and_test_convert(cap, offset, len) + } + + // Test cases where block at start of buffer is bigger than block at end + for offset in (cap - (len / 2))..cap { + create_vec_and_test_convert(cap, offset, len) + } + } + + // Now there's not (necessarily) space to straighten the ring with simple copies, + // the ring will use swapping when: + // (cap + 1 - offset) > (cap + 1 - len) && (len - (cap + 1 - offset)) > (cap + 1 - len)) + // right block size > free space && left block size > free space + for len in ((cap + 1) / 2)..cap { + // Test contiguous cases + for offset in 0..(cap - len) { + create_vec_and_test_convert(cap, offset, len) + } + + // Test cases where block at end of buffer is bigger than block at start + for offset in (cap - len)..(cap - (len / 2)) { + create_vec_and_test_convert(cap, offset, len) + } + + // Test cases where block at start of buffer is bigger than block at end + for offset in (cap - (len / 2))..cap { + create_vec_and_test_convert(cap, offset, len) + } + } + } +} + +#[test] +fn test_clone_from() { + let m = vec![1; 8]; + let n = vec![2; 12]; + let limit = if cfg!(miri) { 4 } else { 8 }; // Miri is too slow + for pfv in 0..limit { + for pfu in 0..limit { + for longer in 0..2 { + let (vr, ur) = if longer == 0 { (&m, &n) } else { (&n, &m) }; + let mut v = VecDeque::from(vr.clone()); + for _ in 0..pfv { + v.push_front(1); + } + let mut u = VecDeque::from(ur.clone()); + for _ in 0..pfu { + u.push_front(2); + } + v.clone_from(&u); + assert_eq!(&v, &u); + } + } + } +} + +#[test] +fn test_vec_deque_truncate_drop() { + static mut DROPS: u32 = 0; + #[derive(Clone)] + struct Elem(i32); + impl Drop for Elem { + fn drop(&mut self) { + unsafe { + DROPS += 1; + } + } + } + + let v = vec![Elem(1), Elem(2), Elem(3), Elem(4), Elem(5)]; + for push_front in 0..=v.len() { + let v = v.clone(); + let mut tester = VecDeque::with_capacity(5); + for (index, elem) in v.into_iter().enumerate() { + if index < push_front { + tester.push_front(elem); + } else { + tester.push_back(elem); + } + } + assert_eq!(unsafe { DROPS }, 0); + tester.truncate(3); + assert_eq!(unsafe { DROPS }, 2); + tester.truncate(0); + assert_eq!(unsafe { DROPS }, 5); + unsafe { + DROPS = 0; + } + } +} + +#[test] +fn issue_53529() { + use crate::boxed::Box; + + let mut dst = VecDeque::new(); + dst.push_front(Box::new(1)); + dst.push_front(Box::new(2)); + assert_eq!(*dst.pop_back().unwrap(), 1); + + let mut src = VecDeque::new(); + src.push_front(Box::new(2)); + dst.append(&mut src); + for a in dst { + assert_eq!(*a, 2); + } +} + +#[test] +fn issue_80303() { + use core::iter; + use core::num::Wrapping; + + // This is a valid, albeit rather bad hash function implementation. + struct SimpleHasher(Wrapping<u64>); + + impl Hasher for SimpleHasher { + fn finish(&self) -> u64 { + self.0.0 + } + + fn write(&mut self, bytes: &[u8]) { + // This particular implementation hashes value 24 in addition to bytes. + // Such an implementation is valid as Hasher only guarantees equivalence + // for the exact same set of calls to its methods. + for &v in iter::once(&24).chain(bytes) { + self.0 = Wrapping(31) * self.0 + Wrapping(u64::from(v)); + } + } + } + + fn hash_code(value: impl Hash) -> u64 { + let mut hasher = SimpleHasher(Wrapping(1)); + value.hash(&mut hasher); + hasher.finish() + } + + // This creates two deques for which values returned by as_slices + // method differ. + let vda: VecDeque<u8> = (0..10).collect(); + let mut vdb = VecDeque::with_capacity(10); + vdb.extend(5..10); + (0..5).rev().for_each(|elem| vdb.push_front(elem)); + assert_ne!(vda.as_slices(), vdb.as_slices()); + assert_eq!(vda, vdb); + assert_eq!(hash_code(vda), hash_code(vdb)); +} diff --git a/library/alloc/src/ffi/c_str.rs b/library/alloc/src/ffi/c_str.rs new file mode 100644 index 000000000..ae61b1f1e --- /dev/null +++ b/library/alloc/src/ffi/c_str.rs @@ -0,0 +1,1123 @@ +#[cfg(test)] +mod tests; + +use crate::borrow::{Cow, ToOwned}; +use crate::boxed::Box; +use crate::rc::Rc; +use crate::slice::hack::into_vec; +use crate::string::String; +use crate::vec::Vec; +use core::borrow::Borrow; +use core::ffi::{c_char, CStr}; +use core::fmt; +use core::mem; +use core::num::NonZeroU8; +use core::ops; +use core::ptr; +use core::slice; +use core::slice::memchr; +use core::str::{self, Utf8Error}; + +#[cfg(target_has_atomic = "ptr")] +use crate::sync::Arc; + +/// A type representing an owned, C-compatible, nul-terminated string with no nul bytes in the +/// middle. +/// +/// This type serves the purpose of being able to safely generate a +/// C-compatible string from a Rust byte slice or vector. An instance of this +/// type is a static guarantee that the underlying bytes contain no interior 0 +/// bytes ("nul characters") and that the final byte is 0 ("nul terminator"). +/// +/// `CString` is to <code>&[CStr]</code> as [`String`] is to <code>&[str]</code>: the former +/// in each pair are owned strings; the latter are borrowed +/// references. +/// +/// # Creating a `CString` +/// +/// A `CString` is created from either a byte slice or a byte vector, +/// or anything that implements <code>[Into]<[Vec]<[u8]>></code> (for +/// example, you can build a `CString` straight out of a [`String`] or +/// a <code>&[str]</code>, since both implement that trait). +/// +/// The [`CString::new`] method will actually check that the provided <code>&[[u8]]</code> +/// does not have 0 bytes in the middle, and return an error if it +/// finds one. +/// +/// # Extracting a raw pointer to the whole C string +/// +/// `CString` implements an [`as_ptr`][`CStr::as_ptr`] method through the [`Deref`] +/// trait. This method will give you a `*const c_char` which you can +/// feed directly to extern functions that expect a nul-terminated +/// string, like C's `strdup()`. Notice that [`as_ptr`][`CStr::as_ptr`] returns a +/// read-only pointer; if the C code writes to it, that causes +/// undefined behavior. +/// +/// # Extracting a slice of the whole C string +/// +/// Alternatively, you can obtain a <code>&[[u8]]</code> slice from a +/// `CString` with the [`CString::as_bytes`] method. Slices produced in this +/// way do *not* contain the trailing nul terminator. This is useful +/// when you will be calling an extern function that takes a `*const +/// u8` argument which is not necessarily nul-terminated, plus another +/// argument with the length of the string — like C's `strndup()`. +/// You can of course get the slice's length with its +/// [`len`][slice::len] method. +/// +/// If you need a <code>&[[u8]]</code> slice *with* the nul terminator, you +/// can use [`CString::as_bytes_with_nul`] instead. +/// +/// Once you have the kind of slice you need (with or without a nul +/// terminator), you can call the slice's own +/// [`as_ptr`][slice::as_ptr] method to get a read-only raw pointer to pass to +/// extern functions. See the documentation for that function for a +/// discussion on ensuring the lifetime of the raw pointer. +/// +/// [str]: prim@str "str" +/// [`Deref`]: ops::Deref +/// +/// # Examples +/// +/// ```ignore (extern-declaration) +/// # fn main() { +/// use std::ffi::CString; +/// use std::os::raw::c_char; +/// +/// extern "C" { +/// fn my_printer(s: *const c_char); +/// } +/// +/// // We are certain that our string doesn't have 0 bytes in the middle, +/// // so we can .expect() +/// let c_to_print = CString::new("Hello, world!").expect("CString::new failed"); +/// unsafe { +/// my_printer(c_to_print.as_ptr()); +/// } +/// # } +/// ``` +/// +/// # Safety +/// +/// `CString` is intended for working with traditional C-style strings +/// (a sequence of non-nul bytes terminated by a single nul byte); the +/// primary use case for these kinds of strings is interoperating with C-like +/// code. Often you will need to transfer ownership to/from that external +/// code. It is strongly recommended that you thoroughly read through the +/// documentation of `CString` before use, as improper ownership management +/// of `CString` instances can lead to invalid memory accesses, memory leaks, +/// and other memory errors. +#[derive(PartialEq, PartialOrd, Eq, Ord, Hash, Clone)] +#[cfg_attr(not(test), rustc_diagnostic_item = "cstring_type")] +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub struct CString { + // Invariant 1: the slice ends with a zero byte and has a length of at least one. + // Invariant 2: the slice contains only one zero byte. + // Improper usage of unsafe function can break Invariant 2, but not Invariant 1. + inner: Box<[u8]>, +} + +/// An error indicating that an interior nul byte was found. +/// +/// While Rust strings may contain nul bytes in the middle, C strings +/// can't, as that byte would effectively truncate the string. +/// +/// This error is created by the [`new`][`CString::new`] method on +/// [`CString`]. See its documentation for more. +/// +/// # Examples +/// +/// ``` +/// use std::ffi::{CString, NulError}; +/// +/// let _: NulError = CString::new(b"f\0oo".to_vec()).unwrap_err(); +/// ``` +#[derive(Clone, PartialEq, Eq, Debug)] +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub struct NulError(usize, Vec<u8>); + +#[derive(Clone, PartialEq, Eq, Debug)] +enum FromBytesWithNulErrorKind { + InteriorNul(usize), + NotNulTerminated, +} + +/// An error indicating that a nul byte was not in the expected position. +/// +/// The vector used to create a [`CString`] must have one and only one nul byte, +/// positioned at the end. +/// +/// This error is created by the [`CString::from_vec_with_nul`] method. +/// See its documentation for more. +/// +/// # Examples +/// +/// ``` +/// use std::ffi::{CString, FromVecWithNulError}; +/// +/// let _: FromVecWithNulError = CString::from_vec_with_nul(b"f\0oo".to_vec()).unwrap_err(); +/// ``` +#[derive(Clone, PartialEq, Eq, Debug)] +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub struct FromVecWithNulError { + error_kind: FromBytesWithNulErrorKind, + bytes: Vec<u8>, +} + +#[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] +impl FromVecWithNulError { + /// Returns a slice of [`u8`]s bytes that were attempted to convert to a [`CString`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::ffi::CString; + /// + /// // Some invalid bytes in a vector + /// let bytes = b"f\0oo".to_vec(); + /// + /// let value = CString::from_vec_with_nul(bytes.clone()); + /// + /// assert_eq!(&bytes[..], value.unwrap_err().as_bytes()); + /// ``` + #[must_use] + #[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] + pub fn as_bytes(&self) -> &[u8] { + &self.bytes[..] + } + + /// Returns the bytes that were attempted to convert to a [`CString`]. + /// + /// This method is carefully constructed to avoid allocation. It will + /// consume the error, moving out the bytes, so that a copy of the bytes + /// does not need to be made. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::ffi::CString; + /// + /// // Some invalid bytes in a vector + /// let bytes = b"f\0oo".to_vec(); + /// + /// let value = CString::from_vec_with_nul(bytes.clone()); + /// + /// assert_eq!(bytes, value.unwrap_err().into_bytes()); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] + pub fn into_bytes(self) -> Vec<u8> { + self.bytes + } +} + +/// An error indicating invalid UTF-8 when converting a [`CString`] into a [`String`]. +/// +/// `CString` is just a wrapper over a buffer of bytes with a nul terminator; +/// [`CString::into_string`] performs UTF-8 validation on those bytes and may +/// return this error. +/// +/// This `struct` is created by [`CString::into_string()`]. See +/// its documentation for more. +#[derive(Clone, PartialEq, Eq, Debug)] +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub struct IntoStringError { + inner: CString, + error: Utf8Error, +} + +impl CString { + /// Creates a new C-compatible string from a container of bytes. + /// + /// This function will consume the provided data and use the + /// underlying bytes to construct a new string, ensuring that + /// there is a trailing 0 byte. This trailing 0 byte will be + /// appended by this function; the provided data should *not* + /// contain any 0 bytes in it. + /// + /// # Examples + /// + /// ```ignore (extern-declaration) + /// use std::ffi::CString; + /// use std::os::raw::c_char; + /// + /// extern "C" { fn puts(s: *const c_char); } + /// + /// let to_print = CString::new("Hello!").expect("CString::new failed"); + /// unsafe { + /// puts(to_print.as_ptr()); + /// } + /// ``` + /// + /// # Errors + /// + /// This function will return an error if the supplied bytes contain an + /// internal 0 byte. The [`NulError`] returned will contain the bytes as well as + /// the position of the nul byte. + #[stable(feature = "rust1", since = "1.0.0")] + pub fn new<T: Into<Vec<u8>>>(t: T) -> Result<CString, NulError> { + trait SpecNewImpl { + fn spec_new_impl(self) -> Result<CString, NulError>; + } + + impl<T: Into<Vec<u8>>> SpecNewImpl for T { + default fn spec_new_impl(self) -> Result<CString, NulError> { + let bytes: Vec<u8> = self.into(); + match memchr::memchr(0, &bytes) { + Some(i) => Err(NulError(i, bytes)), + None => Ok(unsafe { CString::_from_vec_unchecked(bytes) }), + } + } + } + + // Specialization for avoiding reallocation + #[inline(always)] // Without that it is not inlined into specializations + fn spec_new_impl_bytes(bytes: &[u8]) -> Result<CString, NulError> { + // We cannot have such large slice that we would overflow here + // but using `checked_add` allows LLVM to assume that capacity never overflows + // and generate twice shorter code. + // `saturating_add` doesn't help for some reason. + let capacity = bytes.len().checked_add(1).unwrap(); + + // Allocate before validation to avoid duplication of allocation code. + // We still need to allocate and copy memory even if we get an error. + let mut buffer = Vec::with_capacity(capacity); + buffer.extend(bytes); + + // Check memory of self instead of new buffer. + // This allows better optimizations if lto enabled. + match memchr::memchr(0, bytes) { + Some(i) => Err(NulError(i, buffer)), + None => Ok(unsafe { CString::_from_vec_unchecked(buffer) }), + } + } + + impl SpecNewImpl for &'_ [u8] { + fn spec_new_impl(self) -> Result<CString, NulError> { + spec_new_impl_bytes(self) + } + } + + impl SpecNewImpl for &'_ str { + fn spec_new_impl(self) -> Result<CString, NulError> { + spec_new_impl_bytes(self.as_bytes()) + } + } + + impl SpecNewImpl for &'_ mut [u8] { + fn spec_new_impl(self) -> Result<CString, NulError> { + spec_new_impl_bytes(self) + } + } + + t.spec_new_impl() + } + + /// Creates a C-compatible string by consuming a byte vector, + /// without checking for interior 0 bytes. + /// + /// Trailing 0 byte will be appended by this function. + /// + /// This method is equivalent to [`CString::new`] except that no runtime + /// assertion is made that `v` contains no 0 bytes, and it requires an + /// actual byte vector, not anything that can be converted to one with Into. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let raw = b"foo".to_vec(); + /// unsafe { + /// let c_string = CString::from_vec_unchecked(raw); + /// } + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn from_vec_unchecked(v: Vec<u8>) -> Self { + debug_assert!(memchr::memchr(0, &v).is_none()); + unsafe { Self::_from_vec_unchecked(v) } + } + + unsafe fn _from_vec_unchecked(mut v: Vec<u8>) -> Self { + v.reserve_exact(1); + v.push(0); + Self { inner: v.into_boxed_slice() } + } + + /// Retakes ownership of a `CString` that was transferred to C via + /// [`CString::into_raw`]. + /// + /// Additionally, the length of the string will be recalculated from the pointer. + /// + /// # Safety + /// + /// This should only ever be called with a pointer that was earlier + /// obtained by calling [`CString::into_raw`]. Other usage (e.g., trying to take + /// ownership of a string that was allocated by foreign code) is likely to lead + /// to undefined behavior or allocator corruption. + /// + /// It should be noted that the length isn't just "recomputed," but that + /// the recomputed length must match the original length from the + /// [`CString::into_raw`] call. This means the [`CString::into_raw`]/`from_raw` + /// methods should not be used when passing the string to C functions that can + /// modify the string's length. + /// + /// > **Note:** If you need to borrow a string that was allocated by + /// > foreign code, use [`CStr`]. If you need to take ownership of + /// > a string that was allocated by foreign code, you will need to + /// > make your own provisions for freeing it appropriately, likely + /// > with the foreign code's API to do that. + /// + /// # Examples + /// + /// Creates a `CString`, pass ownership to an `extern` function (via raw pointer), then retake + /// ownership with `from_raw`: + /// + /// ```ignore (extern-declaration) + /// use std::ffi::CString; + /// use std::os::raw::c_char; + /// + /// extern "C" { + /// fn some_extern_function(s: *mut c_char); + /// } + /// + /// let c_string = CString::new("Hello!").expect("CString::new failed"); + /// let raw = c_string.into_raw(); + /// unsafe { + /// some_extern_function(raw); + /// let c_string = CString::from_raw(raw); + /// } + /// ``` + #[must_use = "call `drop(from_raw(ptr))` if you intend to drop the `CString`"] + #[stable(feature = "cstr_memory", since = "1.4.0")] + pub unsafe fn from_raw(ptr: *mut c_char) -> CString { + // SAFETY: This is called with a pointer that was obtained from a call + // to `CString::into_raw` and the length has not been modified. As such, + // we know there is a NUL byte (and only one) at the end and that the + // information about the size of the allocation is correct on Rust's + // side. + unsafe { + extern "C" { + /// Provided by libc or compiler_builtins. + fn strlen(s: *const c_char) -> usize; + } + let len = strlen(ptr) + 1; // Including the NUL byte + let slice = slice::from_raw_parts_mut(ptr, len as usize); + CString { inner: Box::from_raw(slice as *mut [c_char] as *mut [u8]) } + } + } + + /// Consumes the `CString` and transfers ownership of the string to a C caller. + /// + /// The pointer which this function returns must be returned to Rust and reconstituted using + /// [`CString::from_raw`] to be properly deallocated. Specifically, one + /// should *not* use the standard C `free()` function to deallocate + /// this string. + /// + /// Failure to call [`CString::from_raw`] will lead to a memory leak. + /// + /// The C side must **not** modify the length of the string (by writing a + /// `null` somewhere inside the string or removing the final one) before + /// it makes it back into Rust using [`CString::from_raw`]. See the safety section + /// in [`CString::from_raw`]. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new("foo").expect("CString::new failed"); + /// + /// let ptr = c_string.into_raw(); + /// + /// unsafe { + /// assert_eq!(b'f', *ptr as u8); + /// assert_eq!(b'o', *ptr.offset(1) as u8); + /// assert_eq!(b'o', *ptr.offset(2) as u8); + /// assert_eq!(b'\0', *ptr.offset(3) as u8); + /// + /// // retake pointer to free memory + /// let _ = CString::from_raw(ptr); + /// } + /// ``` + #[inline] + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "cstr_memory", since = "1.4.0")] + pub fn into_raw(self) -> *mut c_char { + Box::into_raw(self.into_inner()) as *mut c_char + } + + /// Converts the `CString` into a [`String`] if it contains valid UTF-8 data. + /// + /// On failure, ownership of the original `CString` is returned. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let valid_utf8 = vec![b'f', b'o', b'o']; + /// let cstring = CString::new(valid_utf8).expect("CString::new failed"); + /// assert_eq!(cstring.into_string().expect("into_string() call failed"), "foo"); + /// + /// let invalid_utf8 = vec![b'f', 0xff, b'o', b'o']; + /// let cstring = CString::new(invalid_utf8).expect("CString::new failed"); + /// let err = cstring.into_string().err().expect("into_string().err() failed"); + /// assert_eq!(err.utf8_error().valid_up_to(), 1); + /// ``` + #[stable(feature = "cstring_into", since = "1.7.0")] + pub fn into_string(self) -> Result<String, IntoStringError> { + String::from_utf8(self.into_bytes()).map_err(|e| IntoStringError { + error: e.utf8_error(), + inner: unsafe { Self::_from_vec_unchecked(e.into_bytes()) }, + }) + } + + /// Consumes the `CString` and returns the underlying byte buffer. + /// + /// The returned buffer does **not** contain the trailing nul + /// terminator, and it is guaranteed to not have any interior nul + /// bytes. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new("foo").expect("CString::new failed"); + /// let bytes = c_string.into_bytes(); + /// assert_eq!(bytes, vec![b'f', b'o', b'o']); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "cstring_into", since = "1.7.0")] + pub fn into_bytes(self) -> Vec<u8> { + let mut vec = into_vec(self.into_inner()); + let _nul = vec.pop(); + debug_assert_eq!(_nul, Some(0u8)); + vec + } + + /// Equivalent to [`CString::into_bytes()`] except that the + /// returned vector includes the trailing nul terminator. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new("foo").expect("CString::new failed"); + /// let bytes = c_string.into_bytes_with_nul(); + /// assert_eq!(bytes, vec![b'f', b'o', b'o', b'\0']); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "cstring_into", since = "1.7.0")] + pub fn into_bytes_with_nul(self) -> Vec<u8> { + into_vec(self.into_inner()) + } + + /// Returns the contents of this `CString` as a slice of bytes. + /// + /// The returned slice does **not** contain the trailing nul + /// terminator, and it is guaranteed to not have any interior nul + /// bytes. If you need the nul terminator, use + /// [`CString::as_bytes_with_nul`] instead. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new("foo").expect("CString::new failed"); + /// let bytes = c_string.as_bytes(); + /// assert_eq!(bytes, &[b'f', b'o', b'o']); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn as_bytes(&self) -> &[u8] { + // SAFETY: CString has a length at least 1 + unsafe { self.inner.get_unchecked(..self.inner.len() - 1) } + } + + /// Equivalent to [`CString::as_bytes()`] except that the + /// returned slice includes the trailing nul terminator. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new("foo").expect("CString::new failed"); + /// let bytes = c_string.as_bytes_with_nul(); + /// assert_eq!(bytes, &[b'f', b'o', b'o', b'\0']); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn as_bytes_with_nul(&self) -> &[u8] { + &self.inner + } + + /// Extracts a [`CStr`] slice containing the entire string. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::{CString, CStr}; + /// + /// let c_string = CString::new(b"foo".to_vec()).expect("CString::new failed"); + /// let cstr = c_string.as_c_str(); + /// assert_eq!(cstr, + /// CStr::from_bytes_with_nul(b"foo\0").expect("CStr::from_bytes_with_nul failed")); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "as_c_str", since = "1.20.0")] + pub fn as_c_str(&self) -> &CStr { + &*self + } + + /// Converts this `CString` into a boxed [`CStr`]. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::{CString, CStr}; + /// + /// let c_string = CString::new(b"foo".to_vec()).expect("CString::new failed"); + /// let boxed = c_string.into_boxed_c_str(); + /// assert_eq!(&*boxed, + /// CStr::from_bytes_with_nul(b"foo\0").expect("CStr::from_bytes_with_nul failed")); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "into_boxed_c_str", since = "1.20.0")] + pub fn into_boxed_c_str(self) -> Box<CStr> { + unsafe { Box::from_raw(Box::into_raw(self.into_inner()) as *mut CStr) } + } + + /// Bypass "move out of struct which implements [`Drop`] trait" restriction. + #[inline] + fn into_inner(self) -> Box<[u8]> { + // Rationale: `mem::forget(self)` invalidates the previous call to `ptr::read(&self.inner)` + // so we use `ManuallyDrop` to ensure `self` is not dropped. + // Then we can return the box directly without invalidating it. + // See https://github.com/rust-lang/rust/issues/62553. + let this = mem::ManuallyDrop::new(self); + unsafe { ptr::read(&this.inner) } + } + + /// Converts a <code>[Vec]<[u8]></code> to a [`CString`] without checking the + /// invariants on the given [`Vec`]. + /// + /// # Safety + /// + /// The given [`Vec`] **must** have one nul byte as its last element. + /// This means it cannot be empty nor have any other nul byte anywhere else. + /// + /// # Example + /// + /// ``` + /// use std::ffi::CString; + /// assert_eq!( + /// unsafe { CString::from_vec_with_nul_unchecked(b"abc\0".to_vec()) }, + /// unsafe { CString::from_vec_unchecked(b"abc".to_vec()) } + /// ); + /// ``` + #[must_use] + #[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] + pub unsafe fn from_vec_with_nul_unchecked(v: Vec<u8>) -> Self { + debug_assert!(memchr::memchr(0, &v).unwrap() + 1 == v.len()); + unsafe { Self::_from_vec_with_nul_unchecked(v) } + } + + unsafe fn _from_vec_with_nul_unchecked(v: Vec<u8>) -> Self { + Self { inner: v.into_boxed_slice() } + } + + /// Attempts to converts a <code>[Vec]<[u8]></code> to a [`CString`]. + /// + /// Runtime checks are present to ensure there is only one nul byte in the + /// [`Vec`], its last element. + /// + /// # Errors + /// + /// If a nul byte is present and not the last element or no nul bytes + /// is present, an error will be returned. + /// + /// # Examples + /// + /// A successful conversion will produce the same result as [`CString::new`] + /// when called without the ending nul byte. + /// + /// ``` + /// use std::ffi::CString; + /// assert_eq!( + /// CString::from_vec_with_nul(b"abc\0".to_vec()) + /// .expect("CString::from_vec_with_nul failed"), + /// CString::new(b"abc".to_vec()).expect("CString::new failed") + /// ); + /// ``` + /// + /// An incorrectly formatted [`Vec`] will produce an error. + /// + /// ``` + /// use std::ffi::{CString, FromVecWithNulError}; + /// // Interior nul byte + /// let _: FromVecWithNulError = CString::from_vec_with_nul(b"a\0bc".to_vec()).unwrap_err(); + /// // No nul byte + /// let _: FromVecWithNulError = CString::from_vec_with_nul(b"abc".to_vec()).unwrap_err(); + /// ``` + #[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] + pub fn from_vec_with_nul(v: Vec<u8>) -> Result<Self, FromVecWithNulError> { + let nul_pos = memchr::memchr(0, &v); + match nul_pos { + Some(nul_pos) if nul_pos + 1 == v.len() => { + // SAFETY: We know there is only one nul byte, at the end + // of the vec. + Ok(unsafe { Self::_from_vec_with_nul_unchecked(v) }) + } + Some(nul_pos) => Err(FromVecWithNulError { + error_kind: FromBytesWithNulErrorKind::InteriorNul(nul_pos), + bytes: v, + }), + None => Err(FromVecWithNulError { + error_kind: FromBytesWithNulErrorKind::NotNulTerminated, + bytes: v, + }), + } + } +} + +// Turns this `CString` into an empty string to prevent +// memory-unsafe code from working by accident. Inline +// to prevent LLVM from optimizing it away in debug builds. +#[stable(feature = "cstring_drop", since = "1.13.0")] +impl Drop for CString { + #[inline] + fn drop(&mut self) { + unsafe { + *self.inner.get_unchecked_mut(0) = 0; + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Deref for CString { + type Target = CStr; + + #[inline] + fn deref(&self) -> &CStr { + unsafe { CStr::from_bytes_with_nul_unchecked(self.as_bytes_with_nul()) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Debug for CString { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "cstring_into", since = "1.7.0")] +impl From<CString> for Vec<u8> { + /// Converts a [`CString`] into a <code>[Vec]<[u8]></code>. + /// + /// The conversion consumes the [`CString`], and removes the terminating NUL byte. + #[inline] + fn from(s: CString) -> Vec<u8> { + s.into_bytes() + } +} + +#[stable(feature = "cstr_default", since = "1.10.0")] +impl Default for CString { + /// Creates an empty `CString`. + fn default() -> CString { + let a: &CStr = Default::default(); + a.to_owned() + } +} + +#[stable(feature = "cstr_borrow", since = "1.3.0")] +impl Borrow<CStr> for CString { + #[inline] + fn borrow(&self) -> &CStr { + self + } +} + +#[stable(feature = "cstring_from_cow_cstr", since = "1.28.0")] +impl<'a> From<Cow<'a, CStr>> for CString { + /// Converts a `Cow<'a, CStr>` into a `CString`, by copying the contents if they are + /// borrowed. + #[inline] + fn from(s: Cow<'a, CStr>) -> Self { + s.into_owned() + } +} + +#[cfg(not(test))] +#[stable(feature = "box_from_c_str", since = "1.17.0")] +impl From<&CStr> for Box<CStr> { + /// Converts a `&CStr` into a `Box<CStr>`, + /// by copying the contents into a newly allocated [`Box`]. + fn from(s: &CStr) -> Box<CStr> { + let boxed: Box<[u8]> = Box::from(s.to_bytes_with_nul()); + unsafe { Box::from_raw(Box::into_raw(boxed) as *mut CStr) } + } +} + +#[stable(feature = "box_from_cow", since = "1.45.0")] +impl From<Cow<'_, CStr>> for Box<CStr> { + /// Converts a `Cow<'a, CStr>` into a `Box<CStr>`, + /// by copying the contents if they are borrowed. + #[inline] + fn from(cow: Cow<'_, CStr>) -> Box<CStr> { + match cow { + Cow::Borrowed(s) => Box::from(s), + Cow::Owned(s) => Box::from(s), + } + } +} + +#[stable(feature = "c_string_from_box", since = "1.18.0")] +impl From<Box<CStr>> for CString { + /// Converts a <code>[Box]<[CStr]></code> into a [`CString`] without copying or allocating. + #[inline] + fn from(s: Box<CStr>) -> CString { + let raw = Box::into_raw(s) as *mut [u8]; + CString { inner: unsafe { Box::from_raw(raw) } } + } +} + +#[stable(feature = "cstring_from_vec_of_nonzerou8", since = "1.43.0")] +impl From<Vec<NonZeroU8>> for CString { + /// Converts a <code>[Vec]<[NonZeroU8]></code> into a [`CString`] without + /// copying nor checking for inner null bytes. + #[inline] + fn from(v: Vec<NonZeroU8>) -> CString { + unsafe { + // Transmute `Vec<NonZeroU8>` to `Vec<u8>`. + let v: Vec<u8> = { + // SAFETY: + // - transmuting between `NonZeroU8` and `u8` is sound; + // - `alloc::Layout<NonZeroU8> == alloc::Layout<u8>`. + let (ptr, len, cap): (*mut NonZeroU8, _, _) = Vec::into_raw_parts(v); + Vec::from_raw_parts(ptr.cast::<u8>(), len, cap) + }; + // SAFETY: `v` cannot contain null bytes, given the type-level + // invariant of `NonZeroU8`. + Self::_from_vec_unchecked(v) + } + } +} + +#[cfg(not(test))] +#[stable(feature = "more_box_slice_clone", since = "1.29.0")] +impl Clone for Box<CStr> { + #[inline] + fn clone(&self) -> Self { + (**self).into() + } +} + +#[stable(feature = "box_from_c_string", since = "1.20.0")] +impl From<CString> for Box<CStr> { + /// Converts a [`CString`] into a <code>[Box]<[CStr]></code> without copying or allocating. + #[inline] + fn from(s: CString) -> Box<CStr> { + s.into_boxed_c_str() + } +} + +#[stable(feature = "cow_from_cstr", since = "1.28.0")] +impl<'a> From<CString> for Cow<'a, CStr> { + /// Converts a [`CString`] into an owned [`Cow`] without copying or allocating. + #[inline] + fn from(s: CString) -> Cow<'a, CStr> { + Cow::Owned(s) + } +} + +#[stable(feature = "cow_from_cstr", since = "1.28.0")] +impl<'a> From<&'a CStr> for Cow<'a, CStr> { + /// Converts a [`CStr`] into a borrowed [`Cow`] without copying or allocating. + #[inline] + fn from(s: &'a CStr) -> Cow<'a, CStr> { + Cow::Borrowed(s) + } +} + +#[stable(feature = "cow_from_cstr", since = "1.28.0")] +impl<'a> From<&'a CString> for Cow<'a, CStr> { + /// Converts a `&`[`CString`] into a borrowed [`Cow`] without copying or allocating. + #[inline] + fn from(s: &'a CString) -> Cow<'a, CStr> { + Cow::Borrowed(s.as_c_str()) + } +} + +#[cfg(target_has_atomic = "ptr")] +#[stable(feature = "shared_from_slice2", since = "1.24.0")] +impl From<CString> for Arc<CStr> { + /// Converts a [`CString`] into an <code>[Arc]<[CStr]></code> by moving the [`CString`] + /// data into a new [`Arc`] buffer. + #[inline] + fn from(s: CString) -> Arc<CStr> { + let arc: Arc<[u8]> = Arc::from(s.into_inner()); + unsafe { Arc::from_raw(Arc::into_raw(arc) as *const CStr) } + } +} + +#[cfg(target_has_atomic = "ptr")] +#[stable(feature = "shared_from_slice2", since = "1.24.0")] +impl From<&CStr> for Arc<CStr> { + /// Converts a `&CStr` into a `Arc<CStr>`, + /// by copying the contents into a newly allocated [`Arc`]. + #[inline] + fn from(s: &CStr) -> Arc<CStr> { + let arc: Arc<[u8]> = Arc::from(s.to_bytes_with_nul()); + unsafe { Arc::from_raw(Arc::into_raw(arc) as *const CStr) } + } +} + +#[stable(feature = "shared_from_slice2", since = "1.24.0")] +impl From<CString> for Rc<CStr> { + /// Converts a [`CString`] into an <code>[Rc]<[CStr]></code> by moving the [`CString`] + /// data into a new [`Arc`] buffer. + #[inline] + fn from(s: CString) -> Rc<CStr> { + let rc: Rc<[u8]> = Rc::from(s.into_inner()); + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const CStr) } + } +} + +#[stable(feature = "shared_from_slice2", since = "1.24.0")] +impl From<&CStr> for Rc<CStr> { + /// Converts a `&CStr` into a `Rc<CStr>`, + /// by copying the contents into a newly allocated [`Rc`]. + #[inline] + fn from(s: &CStr) -> Rc<CStr> { + let rc: Rc<[u8]> = Rc::from(s.to_bytes_with_nul()); + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const CStr) } + } +} + +#[cfg(not(test))] +#[stable(feature = "default_box_extra", since = "1.17.0")] +impl Default for Box<CStr> { + fn default() -> Box<CStr> { + let boxed: Box<[u8]> = Box::from([0]); + unsafe { Box::from_raw(Box::into_raw(boxed) as *mut CStr) } + } +} + +impl NulError { + /// Returns the position of the nul byte in the slice that caused + /// [`CString::new`] to fail. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let nul_error = CString::new("foo\0bar").unwrap_err(); + /// assert_eq!(nul_error.nul_position(), 3); + /// + /// let nul_error = CString::new("foo bar\0").unwrap_err(); + /// assert_eq!(nul_error.nul_position(), 7); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn nul_position(&self) -> usize { + self.0 + } + + /// Consumes this error, returning the underlying vector of bytes which + /// generated the error in the first place. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let nul_error = CString::new("foo\0bar").unwrap_err(); + /// assert_eq!(nul_error.into_vec(), b"foo\0bar"); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_vec(self) -> Vec<u8> { + self.1 + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Display for NulError { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(f, "nul byte found in provided data at position: {}", self.0) + } +} + +#[stable(feature = "cstring_from_vec_with_nul", since = "1.58.0")] +impl fmt::Display for FromVecWithNulError { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + match self.error_kind { + FromBytesWithNulErrorKind::InteriorNul(pos) => { + write!(f, "data provided contains an interior nul byte at pos {pos}") + } + FromBytesWithNulErrorKind::NotNulTerminated => { + write!(f, "data provided is not nul terminated") + } + } + } +} + +impl IntoStringError { + /// Consumes this error, returning original [`CString`] which generated the + /// error. + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "cstring_into", since = "1.7.0")] + pub fn into_cstring(self) -> CString { + self.inner + } + + /// Access the underlying UTF-8 error that was the cause of this error. + #[must_use] + #[stable(feature = "cstring_into", since = "1.7.0")] + pub fn utf8_error(&self) -> Utf8Error { + self.error + } + + #[doc(hidden)] + #[unstable(feature = "cstr_internals", issue = "none")] + pub fn __source(&self) -> &Utf8Error { + &self.error + } +} + +impl IntoStringError { + fn description(&self) -> &str { + "C string contained non-utf8 bytes" + } +} + +#[stable(feature = "cstring_into", since = "1.7.0")] +impl fmt::Display for IntoStringError { + #[allow(deprecated, deprecated_in_future)] + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + self.description().fmt(f) + } +} + +#[stable(feature = "cstr_borrow", since = "1.3.0")] +impl ToOwned for CStr { + type Owned = CString; + + fn to_owned(&self) -> CString { + CString { inner: self.to_bytes_with_nul().into() } + } + + fn clone_into(&self, target: &mut CString) { + let mut b = into_vec(mem::take(&mut target.inner)); + self.to_bytes_with_nul().clone_into(&mut b); + target.inner = b.into_boxed_slice(); + } +} + +#[stable(feature = "cstring_asref", since = "1.7.0")] +impl From<&CStr> for CString { + fn from(s: &CStr) -> CString { + s.to_owned() + } +} + +#[stable(feature = "cstring_asref", since = "1.7.0")] +impl ops::Index<ops::RangeFull> for CString { + type Output = CStr; + + #[inline] + fn index(&self, _index: ops::RangeFull) -> &CStr { + self + } +} + +#[stable(feature = "cstring_asref", since = "1.7.0")] +impl AsRef<CStr> for CString { + #[inline] + fn as_ref(&self) -> &CStr { + self + } +} + +#[cfg(not(test))] +impl CStr { + /// Converts a `CStr` into a <code>[Cow]<[str]></code>. + /// + /// If the contents of the `CStr` are valid UTF-8 data, this + /// function will return a <code>[Cow]::[Borrowed]\(&[str])</code> + /// with the corresponding <code>&[str]</code> slice. Otherwise, it will + /// replace any invalid UTF-8 sequences with + /// [`U+FFFD REPLACEMENT CHARACTER`][U+FFFD] and return a + /// <code>[Cow]::[Owned]\(&[str])</code> with the result. + /// + /// [str]: prim@str "str" + /// [Borrowed]: Cow::Borrowed + /// [Owned]: Cow::Owned + /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER "std::char::REPLACEMENT_CHARACTER" + /// + /// # Examples + /// + /// Calling `to_string_lossy` on a `CStr` containing valid UTF-8: + /// + /// ``` + /// use std::borrow::Cow; + /// use std::ffi::CStr; + /// + /// let cstr = CStr::from_bytes_with_nul(b"Hello World\0") + /// .expect("CStr::from_bytes_with_nul failed"); + /// assert_eq!(cstr.to_string_lossy(), Cow::Borrowed("Hello World")); + /// ``` + /// + /// Calling `to_string_lossy` on a `CStr` containing invalid UTF-8: + /// + /// ``` + /// use std::borrow::Cow; + /// use std::ffi::CStr; + /// + /// let cstr = CStr::from_bytes_with_nul(b"Hello \xF0\x90\x80World\0") + /// .expect("CStr::from_bytes_with_nul failed"); + /// assert_eq!( + /// cstr.to_string_lossy(), + /// Cow::Owned(String::from("Hello �World")) as Cow<'_, str> + /// ); + /// ``` + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the result of the operation, \ + without modifying the original"] + #[stable(feature = "cstr_to_str", since = "1.4.0")] + pub fn to_string_lossy(&self) -> Cow<'_, str> { + String::from_utf8_lossy(self.to_bytes()) + } + + /// Converts a <code>[Box]<[CStr]></code> into a [`CString`] without copying or allocating. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::CString; + /// + /// let c_string = CString::new(b"foo".to_vec()).expect("CString::new failed"); + /// let boxed = c_string.into_boxed_c_str(); + /// assert_eq!(boxed.into_c_string(), CString::new("foo").expect("CString::new failed")); + /// ``` + #[rustc_allow_incoherent_impl] + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "into_boxed_c_str", since = "1.20.0")] + pub fn into_c_string(self: Box<Self>) -> CString { + CString::from(self) + } +} diff --git a/library/alloc/src/ffi/c_str/tests.rs b/library/alloc/src/ffi/c_str/tests.rs new file mode 100644 index 000000000..0b7476d5c --- /dev/null +++ b/library/alloc/src/ffi/c_str/tests.rs @@ -0,0 +1,228 @@ +use super::*; +use crate::rc::Rc; +use crate::sync::Arc; +use core::assert_matches::assert_matches; +use core::ffi::FromBytesUntilNulError; +use core::hash::{Hash, Hasher}; + +#[allow(deprecated)] +use core::hash::SipHasher13 as DefaultHasher; + +#[test] +fn c_to_rust() { + let data = b"123\0"; + let ptr = data.as_ptr() as *const c_char; + unsafe { + assert_eq!(CStr::from_ptr(ptr).to_bytes(), b"123"); + assert_eq!(CStr::from_ptr(ptr).to_bytes_with_nul(), b"123\0"); + } +} + +#[test] +fn simple() { + let s = CString::new("1234").unwrap(); + assert_eq!(s.as_bytes(), b"1234"); + assert_eq!(s.as_bytes_with_nul(), b"1234\0"); +} + +#[test] +fn build_with_zero1() { + assert!(CString::new(&b"\0"[..]).is_err()); +} +#[test] +fn build_with_zero2() { + assert!(CString::new(vec![0]).is_err()); +} + +#[test] +fn formatted() { + let s = CString::new(&b"abc\x01\x02\n\xE2\x80\xA6\xFF"[..]).unwrap(); + assert_eq!(format!("{s:?}"), r#""abc\x01\x02\n\xe2\x80\xa6\xff""#); +} + +#[test] +fn borrowed() { + unsafe { + let s = CStr::from_ptr(b"12\0".as_ptr() as *const _); + assert_eq!(s.to_bytes(), b"12"); + assert_eq!(s.to_bytes_with_nul(), b"12\0"); + } +} + +#[test] +fn to_owned() { + let data = b"123\0"; + let ptr = data.as_ptr() as *const c_char; + + let owned = unsafe { CStr::from_ptr(ptr).to_owned() }; + assert_eq!(owned.as_bytes_with_nul(), data); +} + +#[test] +fn equal_hash() { + let data = b"123\xE2\xFA\xA6\0"; + let ptr = data.as_ptr() as *const c_char; + let cstr: &'static CStr = unsafe { CStr::from_ptr(ptr) }; + + #[allow(deprecated)] + let mut s = DefaultHasher::new(); + cstr.hash(&mut s); + let cstr_hash = s.finish(); + #[allow(deprecated)] + let mut s = DefaultHasher::new(); + CString::new(&data[..data.len() - 1]).unwrap().hash(&mut s); + let cstring_hash = s.finish(); + + assert_eq!(cstr_hash, cstring_hash); +} + +#[test] +fn from_bytes_with_nul() { + let data = b"123\0"; + let cstr = CStr::from_bytes_with_nul(data); + assert_eq!(cstr.map(CStr::to_bytes), Ok(&b"123"[..])); + let cstr = CStr::from_bytes_with_nul(data); + assert_eq!(cstr.map(CStr::to_bytes_with_nul), Ok(&b"123\0"[..])); + + unsafe { + let cstr = CStr::from_bytes_with_nul(data); + let cstr_unchecked = CStr::from_bytes_with_nul_unchecked(data); + assert_eq!(cstr, Ok(cstr_unchecked)); + } +} + +#[test] +fn from_bytes_with_nul_unterminated() { + let data = b"123"; + let cstr = CStr::from_bytes_with_nul(data); + assert!(cstr.is_err()); +} + +#[test] +fn from_bytes_with_nul_interior() { + let data = b"1\023\0"; + let cstr = CStr::from_bytes_with_nul(data); + assert!(cstr.is_err()); +} + +#[test] +fn cstr_from_bytes_until_nul() { + // Test an empty slice. This should fail because it + // does not contain a nul byte. + let b = b""; + assert_matches!(CStr::from_bytes_until_nul(&b[..]), Err(FromBytesUntilNulError { .. })); + + // Test a non-empty slice, that does not contain a nul byte. + let b = b"hello"; + assert_matches!(CStr::from_bytes_until_nul(&b[..]), Err(FromBytesUntilNulError { .. })); + + // Test an empty nul-terminated string + let b = b"\0"; + let r = CStr::from_bytes_until_nul(&b[..]).unwrap(); + assert_eq!(r.to_bytes(), b""); + + // Test a slice with the nul byte in the middle + let b = b"hello\0world!"; + let r = CStr::from_bytes_until_nul(&b[..]).unwrap(); + assert_eq!(r.to_bytes(), b"hello"); + + // Test a slice with the nul byte at the end + let b = b"hello\0"; + let r = CStr::from_bytes_until_nul(&b[..]).unwrap(); + assert_eq!(r.to_bytes(), b"hello"); + + // Test a slice with two nul bytes at the end + let b = b"hello\0\0"; + let r = CStr::from_bytes_until_nul(&b[..]).unwrap(); + assert_eq!(r.to_bytes(), b"hello"); + + // Test a slice containing lots of nul bytes + let b = b"\0\0\0\0"; + let r = CStr::from_bytes_until_nul(&b[..]).unwrap(); + assert_eq!(r.to_bytes(), b""); +} + +#[test] +fn into_boxed() { + let orig: &[u8] = b"Hello, world!\0"; + let cstr = CStr::from_bytes_with_nul(orig).unwrap(); + let boxed: Box<CStr> = Box::from(cstr); + let cstring = cstr.to_owned().into_boxed_c_str().into_c_string(); + assert_eq!(cstr, &*boxed); + assert_eq!(&*boxed, &*cstring); + assert_eq!(&*cstring, cstr); +} + +#[test] +fn boxed_default() { + let boxed = <Box<CStr>>::default(); + assert_eq!(boxed.to_bytes_with_nul(), &[0]); +} + +#[test] +fn test_c_str_clone_into() { + let mut c_string = CString::new("lorem").unwrap(); + let c_ptr = c_string.as_ptr(); + let c_str = CStr::from_bytes_with_nul(b"ipsum\0").unwrap(); + c_str.clone_into(&mut c_string); + assert_eq!(c_str, c_string.as_c_str()); + // The exact same size shouldn't have needed to move its allocation + assert_eq!(c_ptr, c_string.as_ptr()); +} + +#[test] +fn into_rc() { + let orig: &[u8] = b"Hello, world!\0"; + let cstr = CStr::from_bytes_with_nul(orig).unwrap(); + let rc: Rc<CStr> = Rc::from(cstr); + let arc: Arc<CStr> = Arc::from(cstr); + + assert_eq!(&*rc, cstr); + assert_eq!(&*arc, cstr); + + let rc2: Rc<CStr> = Rc::from(cstr.to_owned()); + let arc2: Arc<CStr> = Arc::from(cstr.to_owned()); + + assert_eq!(&*rc2, cstr); + assert_eq!(&*arc2, cstr); +} + +#[test] +fn cstr_const_constructor() { + const CSTR: &CStr = unsafe { CStr::from_bytes_with_nul_unchecked(b"Hello, world!\0") }; + + assert_eq!(CSTR.to_str().unwrap(), "Hello, world!"); +} + +#[test] +fn cstr_index_from() { + let original = b"Hello, world!\0"; + let cstr = CStr::from_bytes_with_nul(original).unwrap(); + let result = CStr::from_bytes_with_nul(&original[7..]).unwrap(); + + assert_eq!(&cstr[7..], result); +} + +#[test] +#[should_panic] +fn cstr_index_from_empty() { + let original = b"Hello, world!\0"; + let cstr = CStr::from_bytes_with_nul(original).unwrap(); + let _ = &cstr[original.len()..]; +} + +#[test] +fn c_string_from_empty_string() { + let original = ""; + let cstring = CString::new(original).unwrap(); + assert_eq!(original.as_bytes(), cstring.as_bytes()); + assert_eq!([b'\0'], cstring.as_bytes_with_nul()); +} + +#[test] +fn c_str_from_empty_string() { + let original = b"\0"; + let cstr = CStr::from_bytes_with_nul(original).unwrap(); + assert_eq!([] as [u8; 0], cstr.to_bytes()); + assert_eq!([b'\0'], cstr.to_bytes_with_nul()); +} diff --git a/library/alloc/src/ffi/mod.rs b/library/alloc/src/ffi/mod.rs new file mode 100644 index 000000000..e8530fbc1 --- /dev/null +++ b/library/alloc/src/ffi/mod.rs @@ -0,0 +1,88 @@ +//! Utilities related to FFI bindings. +//! +//! This module provides utilities to handle data across non-Rust +//! interfaces, like other programming languages and the underlying +//! operating system. It is mainly of use for FFI (Foreign Function +//! Interface) bindings and code that needs to exchange C-like strings +//! with other languages. +//! +//! # Overview +//! +//! Rust represents owned strings with the [`String`] type, and +//! borrowed slices of strings with the [`str`] primitive. Both are +//! always in UTF-8 encoding, and may contain nul bytes in the middle, +//! i.e., if you look at the bytes that make up the string, there may +//! be a `\0` among them. Both `String` and `str` store their length +//! explicitly; there are no nul terminators at the end of strings +//! like in C. +//! +//! C strings are different from Rust strings: +//! +//! * **Encodings** - Rust strings are UTF-8, but C strings may use +//! other encodings. If you are using a string from C, you should +//! check its encoding explicitly, rather than just assuming that it +//! is UTF-8 like you can do in Rust. +//! +//! * **Character size** - C strings may use `char` or `wchar_t`-sized +//! characters; please **note** that C's `char` is different from Rust's. +//! The C standard leaves the actual sizes of those types open to +//! interpretation, but defines different APIs for strings made up of +//! each character type. Rust strings are always UTF-8, so different +//! Unicode characters will be encoded in a variable number of bytes +//! each. The Rust type [`char`] represents a '[Unicode scalar +//! value]', which is similar to, but not the same as, a '[Unicode +//! code point]'. +//! +//! * **Nul terminators and implicit string lengths** - Often, C +//! strings are nul-terminated, i.e., they have a `\0` character at the +//! end. The length of a string buffer is not stored, but has to be +//! calculated; to compute the length of a string, C code must +//! manually call a function like `strlen()` for `char`-based strings, +//! or `wcslen()` for `wchar_t`-based ones. Those functions return +//! the number of characters in the string excluding the nul +//! terminator, so the buffer length is really `len+1` characters. +//! Rust strings don't have a nul terminator; their length is always +//! stored and does not need to be calculated. While in Rust +//! accessing a string's length is an *O*(1) operation (because the +//! length is stored); in C it is an *O*(*n*) operation because the +//! length needs to be computed by scanning the string for the nul +//! terminator. +//! +//! * **Internal nul characters** - When C strings have a nul +//! terminator character, this usually means that they cannot have nul +//! characters in the middle — a nul character would essentially +//! truncate the string. Rust strings *can* have nul characters in +//! the middle, because nul does not have to mark the end of the +//! string in Rust. +//! +//! # Representations of non-Rust strings +//! +//! [`CString`] and [`CStr`] are useful when you need to transfer +//! UTF-8 strings to and from languages with a C ABI, like Python. +//! +//! * **From Rust to C:** [`CString`] represents an owned, C-friendly +//! string: it is nul-terminated, and has no internal nul characters. +//! Rust code can create a [`CString`] out of a normal string (provided +//! that the string doesn't have nul characters in the middle), and +//! then use a variety of methods to obtain a raw <code>\*mut [u8]</code> that can +//! then be passed as an argument to functions which use the C +//! conventions for strings. +//! +//! * **From C to Rust:** [`CStr`] represents a borrowed C string; it +//! is what you would use to wrap a raw <code>\*const [u8]</code> that you got from +//! a C function. A [`CStr`] is guaranteed to be a nul-terminated array +//! of bytes. Once you have a [`CStr`], you can convert it to a Rust +//! <code>&[str]</code> if it's valid UTF-8, or lossily convert it by adding +//! replacement characters. +//! +//! [`String`]: crate::string::String +//! [`CStr`]: core::ffi::CStr + +#![stable(feature = "alloc_ffi", since = "1.64.0")] + +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub use self::c_str::FromVecWithNulError; +#[stable(feature = "alloc_c_string", since = "1.64.0")] +pub use self::c_str::{CString, IntoStringError, NulError}; + +mod c_str; diff --git a/library/alloc/src/fmt.rs b/library/alloc/src/fmt.rs new file mode 100644 index 000000000..ed398b566 --- /dev/null +++ b/library/alloc/src/fmt.rs @@ -0,0 +1,617 @@ +//! Utilities for formatting and printing `String`s. +//! +//! This module contains the runtime support for the [`format!`] syntax extension. +//! This macro is implemented in the compiler to emit calls to this module in +//! order to format arguments at runtime into strings. +//! +//! # Usage +//! +//! The [`format!`] macro is intended to be familiar to those coming from C's +//! `printf`/`fprintf` functions or Python's `str.format` function. +//! +//! Some examples of the [`format!`] extension are: +//! +//! ``` +//! format!("Hello"); // => "Hello" +//! format!("Hello, {}!", "world"); // => "Hello, world!" +//! format!("The number is {}", 1); // => "The number is 1" +//! format!("{:?}", (3, 4)); // => "(3, 4)" +//! format!("{value}", value=4); // => "4" +//! let people = "Rustaceans"; +//! format!("Hello {people}!"); // => "Hello Rustaceans!" +//! format!("{} {}", 1, 2); // => "1 2" +//! format!("{:04}", 42); // => "0042" with leading zeros +//! format!("{:#?}", (100, 200)); // => "( +//! // 100, +//! // 200, +//! // )" +//! ``` +//! +//! From these, you can see that the first argument is a format string. It is +//! required by the compiler for this to be a string literal; it cannot be a +//! variable passed in (in order to perform validity checking). The compiler +//! will then parse the format string and determine if the list of arguments +//! provided is suitable to pass to this format string. +//! +//! To convert a single value to a string, use the [`to_string`] method. This +//! will use the [`Display`] formatting trait. +//! +//! ## Positional parameters +//! +//! Each formatting argument is allowed to specify which value argument it's +//! referencing, and if omitted it is assumed to be "the next argument". For +//! example, the format string `{} {} {}` would take three parameters, and they +//! would be formatted in the same order as they're given. The format string +//! `{2} {1} {0}`, however, would format arguments in reverse order. +//! +//! Things can get a little tricky once you start intermingling the two types of +//! positional specifiers. The "next argument" specifier can be thought of as an +//! iterator over the argument. Each time a "next argument" specifier is seen, +//! the iterator advances. This leads to behavior like this: +//! +//! ``` +//! format!("{1} {} {0} {}", 1, 2); // => "2 1 1 2" +//! ``` +//! +//! The internal iterator over the argument has not been advanced by the time +//! the first `{}` is seen, so it prints the first argument. Then upon reaching +//! the second `{}`, the iterator has advanced forward to the second argument. +//! Essentially, parameters that explicitly name their argument do not affect +//! parameters that do not name an argument in terms of positional specifiers. +//! +//! A format string is required to use all of its arguments, otherwise it is a +//! compile-time error. You may refer to the same argument more than once in the +//! format string. +//! +//! ## Named parameters +//! +//! Rust itself does not have a Python-like equivalent of named parameters to a +//! function, but the [`format!`] macro is a syntax extension that allows it to +//! leverage named parameters. Named parameters are listed at the end of the +//! argument list and have the syntax: +//! +//! ```text +//! identifier '=' expression +//! ``` +//! +//! For example, the following [`format!`] expressions all use named arguments: +//! +//! ``` +//! format!("{argument}", argument = "test"); // => "test" +//! format!("{name} {}", 1, name = 2); // => "2 1" +//! format!("{a} {c} {b}", a="a", b='b', c=3); // => "a 3 b" +//! ``` +//! +//! If a named parameter does not appear in the argument list, `format!` will +//! reference a variable with that name in the current scope. +//! +//! ``` +//! let argument = 2 + 2; +//! format!("{argument}"); // => "4" +//! +//! fn make_string(a: u32, b: &str) -> String { +//! format!("{b} {a}") +//! } +//! make_string(927, "label"); // => "label 927" +//! ``` +//! +//! It is not valid to put positional parameters (those without names) after +//! arguments that have names. Like with positional parameters, it is not +//! valid to provide named parameters that are unused by the format string. +//! +//! # Formatting Parameters +//! +//! Each argument being formatted can be transformed by a number of formatting +//! parameters (corresponding to `format_spec` in [the syntax](#syntax)). These +//! parameters affect the string representation of what's being formatted. +//! +//! ## Width +//! +//! ``` +//! // All of these print "Hello x !" +//! println!("Hello {:5}!", "x"); +//! println!("Hello {:1$}!", "x", 5); +//! println!("Hello {1:0$}!", 5, "x"); +//! println!("Hello {:width$}!", "x", width = 5); +//! let width = 5; +//! println!("Hello {:width$}!", "x"); +//! ``` +//! +//! This is a parameter for the "minimum width" that the format should take up. +//! If the value's string does not fill up this many characters, then the +//! padding specified by fill/alignment will be used to take up the required +//! space (see below). +//! +//! The value for the width can also be provided as a [`usize`] in the list of +//! parameters by adding a postfix `$`, indicating that the second argument is +//! a [`usize`] specifying the width. +//! +//! Referring to an argument with the dollar syntax does not affect the "next +//! argument" counter, so it's usually a good idea to refer to arguments by +//! position, or use named arguments. +//! +//! ## Fill/Alignment +//! +//! ``` +//! assert_eq!(format!("Hello {:<5}!", "x"), "Hello x !"); +//! assert_eq!(format!("Hello {:-<5}!", "x"), "Hello x----!"); +//! assert_eq!(format!("Hello {:^5}!", "x"), "Hello x !"); +//! assert_eq!(format!("Hello {:>5}!", "x"), "Hello x!"); +//! ``` +//! +//! The optional fill character and alignment is provided normally in conjunction with the +//! [`width`](#width) parameter. It must be defined before `width`, right after the `:`. +//! This indicates that if the value being formatted is smaller than +//! `width` some extra characters will be printed around it. +//! Filling comes in the following variants for different alignments: +//! +//! * `[fill]<` - the argument is left-aligned in `width` columns +//! * `[fill]^` - the argument is center-aligned in `width` columns +//! * `[fill]>` - the argument is right-aligned in `width` columns +//! +//! The default [fill/alignment](#fillalignment) for non-numerics is a space and +//! left-aligned. The +//! default for numeric formatters is also a space character but with right-alignment. If +//! the `0` flag (see below) is specified for numerics, then the implicit fill character is +//! `0`. +//! +//! Note that alignment might not be implemented by some types. In particular, it +//! is not generally implemented for the `Debug` trait. A good way to ensure +//! padding is applied is to format your input, then pad this resulting string +//! to obtain your output: +//! +//! ``` +//! println!("Hello {:^15}!", format!("{:?}", Some("hi"))); // => "Hello Some("hi") !" +//! ``` +//! +//! ## Sign/`#`/`0` +//! +//! ``` +//! assert_eq!(format!("Hello {:+}!", 5), "Hello +5!"); +//! assert_eq!(format!("{:#x}!", 27), "0x1b!"); +//! assert_eq!(format!("Hello {:05}!", 5), "Hello 00005!"); +//! assert_eq!(format!("Hello {:05}!", -5), "Hello -0005!"); +//! assert_eq!(format!("{:#010x}!", 27), "0x0000001b!"); +//! ``` +//! +//! These are all flags altering the behavior of the formatter. +//! +//! * `+` - This is intended for numeric types and indicates that the sign +//! should always be printed. Positive signs are never printed by +//! default, and the negative sign is only printed by default for signed values. +//! This flag indicates that the correct sign (`+` or `-`) should always be printed. +//! * `-` - Currently not used +//! * `#` - This flag indicates that the "alternate" form of printing should +//! be used. The alternate forms are: +//! * `#?` - pretty-print the [`Debug`] formatting (adds linebreaks and indentation) +//! * `#x` - precedes the argument with a `0x` +//! * `#X` - precedes the argument with a `0x` +//! * `#b` - precedes the argument with a `0b` +//! * `#o` - precedes the argument with a `0o` +//! * `0` - This is used to indicate for integer formats that the padding to `width` should +//! both be done with a `0` character as well as be sign-aware. A format +//! like `{:08}` would yield `00000001` for the integer `1`, while the +//! same format would yield `-0000001` for the integer `-1`. Notice that +//! the negative version has one fewer zero than the positive version. +//! Note that padding zeros are always placed after the sign (if any) +//! and before the digits. When used together with the `#` flag, a similar +//! rule applies: padding zeros are inserted after the prefix but before +//! the digits. The prefix is included in the total width. +//! +//! ## Precision +//! +//! For non-numeric types, this can be considered a "maximum width". If the resulting string is +//! longer than this width, then it is truncated down to this many characters and that truncated +//! value is emitted with proper `fill`, `alignment` and `width` if those parameters are set. +//! +//! For integral types, this is ignored. +//! +//! For floating-point types, this indicates how many digits after the decimal point should be +//! printed. +//! +//! There are three possible ways to specify the desired `precision`: +//! +//! 1. An integer `.N`: +//! +//! the integer `N` itself is the precision. +//! +//! 2. An integer or name followed by dollar sign `.N$`: +//! +//! use format *argument* `N` (which must be a `usize`) as the precision. +//! +//! 3. An asterisk `.*`: +//! +//! `.*` means that this `{...}` is associated with *two* format inputs rather than one: +//! - If a format string in the fashion of `{:<spec>.*}` is used, then the first input holds +//! the `usize` precision, and the second holds the value to print. +//! - If a format string in the fashion of `{<arg>:<spec>.*}` is used, then the `<arg>` part +//! refers to the value to print, and the `precision` is taken like it was specified with an +//! omitted positional parameter (`{}` instead of `{<arg>:}`). +//! +//! For example, the following calls all print the same thing `Hello x is 0.01000`: +//! +//! ``` +//! // Hello {arg 0 ("x")} is {arg 1 (0.01) with precision specified inline (5)} +//! println!("Hello {0} is {1:.5}", "x", 0.01); +//! +//! // Hello {arg 1 ("x")} is {arg 2 (0.01) with precision specified in arg 0 (5)} +//! println!("Hello {1} is {2:.0$}", 5, "x", 0.01); +//! +//! // Hello {arg 0 ("x")} is {arg 2 (0.01) with precision specified in arg 1 (5)} +//! println!("Hello {0} is {2:.1$}", "x", 5, 0.01); +//! +//! // Hello {next arg -> arg 0 ("x")} is {second of next two args -> arg 2 (0.01) with precision +//! // specified in first of next two args -> arg 1 (5)} +//! println!("Hello {} is {:.*}", "x", 5, 0.01); +//! +//! // Hello {arg 1 ("x")} is {arg 2 (0.01) with precision +//! // specified in next arg -> arg 0 (5)} +//! println!("Hello {1} is {2:.*}", 5, "x", 0.01); +//! +//! // Hello {next arg -> arg 0 ("x")} is {arg 2 (0.01) with precision +//! // specified in next arg -> arg 1 (5)} +//! println!("Hello {} is {2:.*}", "x", 5, 0.01); +//! +//! // Hello {next arg -> arg 0 ("x")} is {arg "number" (0.01) with precision specified +//! // in arg "prec" (5)} +//! println!("Hello {} is {number:.prec$}", "x", prec = 5, number = 0.01); +//! ``` +//! +//! While these: +//! +//! ``` +//! println!("{}, `{name:.*}` has 3 fractional digits", "Hello", 3, name=1234.56); +//! println!("{}, `{name:.*}` has 3 characters", "Hello", 3, name="1234.56"); +//! println!("{}, `{name:>8.*}` has 3 right-aligned characters", "Hello", 3, name="1234.56"); +//! ``` +//! +//! print three significantly different things: +//! +//! ```text +//! Hello, `1234.560` has 3 fractional digits +//! Hello, `123` has 3 characters +//! Hello, ` 123` has 3 right-aligned characters +//! ``` +//! +//! ## Localization +//! +//! In some programming languages, the behavior of string formatting functions +//! depends on the operating system's locale setting. The format functions +//! provided by Rust's standard library do not have any concept of locale and +//! will produce the same results on all systems regardless of user +//! configuration. +//! +//! For example, the following code will always print `1.5` even if the system +//! locale uses a decimal separator other than a dot. +//! +//! ``` +//! println!("The value is {}", 1.5); +//! ``` +//! +//! # Escaping +//! +//! The literal characters `{` and `}` may be included in a string by preceding +//! them with the same character. For example, the `{` character is escaped with +//! `{{` and the `}` character is escaped with `}}`. +//! +//! ``` +//! assert_eq!(format!("Hello {{}}"), "Hello {}"); +//! assert_eq!(format!("{{ Hello"), "{ Hello"); +//! ``` +//! +//! # Syntax +//! +//! To summarize, here you can find the full grammar of format strings. +//! The syntax for the formatting language used is drawn from other languages, +//! so it should not be too alien. Arguments are formatted with Python-like +//! syntax, meaning that arguments are surrounded by `{}` instead of the C-like +//! `%`. The actual grammar for the formatting syntax is: +//! +//! ```text +//! format_string := text [ maybe_format text ] * +//! maybe_format := '{' '{' | '}' '}' | format +//! format := '{' [ argument ] [ ':' format_spec ] [ ws ] * '}' +//! argument := integer | identifier +//! +//! format_spec := [[fill]align][sign]['#']['0'][width]['.' precision]type +//! fill := character +//! align := '<' | '^' | '>' +//! sign := '+' | '-' +//! width := count +//! precision := count | '*' +//! type := '' | '?' | 'x?' | 'X?' | identifier +//! count := parameter | integer +//! parameter := argument '$' +//! ``` +//! In the above grammar, +//! - `text` must not contain any `'{'` or `'}'` characters, +//! - `ws` is any character for which [`char::is_whitespace`] returns `true`, has no semantic +//! meaning and is completely optional, +//! - `integer` is a decimal integer that may contain leading zeroes and +//! - `identifier` is an `IDENTIFIER_OR_KEYWORD` (not an `IDENTIFIER`) as defined by the [Rust language reference](https://doc.rust-lang.org/reference/identifiers.html). +//! +//! # Formatting traits +//! +//! When requesting that an argument be formatted with a particular type, you +//! are actually requesting that an argument ascribes to a particular trait. +//! This allows multiple actual types to be formatted via `{:x}` (like [`i8`] as +//! well as [`isize`]). The current mapping of types to traits is: +//! +//! * *nothing* ⇒ [`Display`] +//! * `?` ⇒ [`Debug`] +//! * `x?` ⇒ [`Debug`] with lower-case hexadecimal integers +//! * `X?` ⇒ [`Debug`] with upper-case hexadecimal integers +//! * `o` ⇒ [`Octal`] +//! * `x` ⇒ [`LowerHex`] +//! * `X` ⇒ [`UpperHex`] +//! * `p` ⇒ [`Pointer`] +//! * `b` ⇒ [`Binary`] +//! * `e` ⇒ [`LowerExp`] +//! * `E` ⇒ [`UpperExp`] +//! +//! What this means is that any type of argument which implements the +//! [`fmt::Binary`][`Binary`] trait can then be formatted with `{:b}`. Implementations +//! are provided for these traits for a number of primitive types by the +//! standard library as well. If no format is specified (as in `{}` or `{:6}`), +//! then the format trait used is the [`Display`] trait. +//! +//! When implementing a format trait for your own type, you will have to +//! implement a method of the signature: +//! +//! ``` +//! # #![allow(dead_code)] +//! # use std::fmt; +//! # struct Foo; // our custom type +//! # impl fmt::Display for Foo { +//! fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { +//! # write!(f, "testing, testing") +//! # } } +//! ``` +//! +//! Your type will be passed as `self` by-reference, and then the function +//! should emit output into the Formatter `f` which implements `fmt::Write`. It is up to each +//! format trait implementation to correctly adhere to the requested formatting parameters. +//! The values of these parameters can be accessed with methods of the +//! [`Formatter`] struct. In order to help with this, the [`Formatter`] struct also +//! provides some helper methods. +//! +//! Additionally, the return value of this function is [`fmt::Result`] which is a +//! type alias of <code>[Result]<(), [std::fmt::Error]></code>. Formatting implementations +//! should ensure that they propagate errors from the [`Formatter`] (e.g., when +//! calling [`write!`]). However, they should never return errors spuriously. That +//! is, a formatting implementation must and may only return an error if the +//! passed-in [`Formatter`] returns an error. This is because, contrary to what +//! the function signature might suggest, string formatting is an infallible +//! operation. This function only returns a result because writing to the +//! underlying stream might fail and it must provide a way to propagate the fact +//! that an error has occurred back up the stack. +//! +//! An example of implementing the formatting traits would look +//! like: +//! +//! ``` +//! use std::fmt; +//! +//! #[derive(Debug)] +//! struct Vector2D { +//! x: isize, +//! y: isize, +//! } +//! +//! impl fmt::Display for Vector2D { +//! fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { +//! // The `f` value implements the `Write` trait, which is what the +//! // write! macro is expecting. Note that this formatting ignores the +//! // various flags provided to format strings. +//! write!(f, "({}, {})", self.x, self.y) +//! } +//! } +//! +//! // Different traits allow different forms of output of a type. The meaning +//! // of this format is to print the magnitude of a vector. +//! impl fmt::Binary for Vector2D { +//! fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { +//! let magnitude = (self.x * self.x + self.y * self.y) as f64; +//! let magnitude = magnitude.sqrt(); +//! +//! // Respect the formatting flags by using the helper method +//! // `pad_integral` on the Formatter object. See the method +//! // documentation for details, and the function `pad` can be used +//! // to pad strings. +//! let decimals = f.precision().unwrap_or(3); +//! let string = format!("{:.*}", decimals, magnitude); +//! f.pad_integral(true, "", &string) +//! } +//! } +//! +//! fn main() { +//! let myvector = Vector2D { x: 3, y: 4 }; +//! +//! println!("{myvector}"); // => "(3, 4)" +//! println!("{myvector:?}"); // => "Vector2D {x: 3, y:4}" +//! println!("{myvector:10.3b}"); // => " 5.000" +//! } +//! ``` +//! +//! ### `fmt::Display` vs `fmt::Debug` +//! +//! These two formatting traits have distinct purposes: +//! +//! - [`fmt::Display`][`Display`] implementations assert that the type can be faithfully +//! represented as a UTF-8 string at all times. It is **not** expected that +//! all types implement the [`Display`] trait. +//! - [`fmt::Debug`][`Debug`] implementations should be implemented for **all** public types. +//! Output will typically represent the internal state as faithfully as possible. +//! The purpose of the [`Debug`] trait is to facilitate debugging Rust code. In +//! most cases, using `#[derive(Debug)]` is sufficient and recommended. +//! +//! Some examples of the output from both traits: +//! +//! ``` +//! assert_eq!(format!("{} {:?}", 3, 4), "3 4"); +//! assert_eq!(format!("{} {:?}", 'a', 'b'), "a 'b'"); +//! assert_eq!(format!("{} {:?}", "foo\n", "bar\n"), "foo\n \"bar\\n\""); +//! ``` +//! +//! # Related macros +//! +//! There are a number of related macros in the [`format!`] family. The ones that +//! are currently implemented are: +//! +//! ```ignore (only-for-syntax-highlight) +//! format! // described above +//! write! // first argument is either a &mut io::Write or a &mut fmt::Write, the destination +//! writeln! // same as write but appends a newline +//! print! // the format string is printed to the standard output +//! println! // same as print but appends a newline +//! eprint! // the format string is printed to the standard error +//! eprintln! // same as eprint but appends a newline +//! format_args! // described below. +//! ``` +//! +//! ### `write!` +//! +//! [`write!`] and [`writeln!`] are two macros which are used to emit the format string +//! to a specified stream. This is used to prevent intermediate allocations of +//! format strings and instead directly write the output. Under the hood, this +//! function is actually invoking the [`write_fmt`] function defined on the +//! [`std::io::Write`] and the [`std::fmt::Write`] trait. Example usage is: +//! +//! ``` +//! # #![allow(unused_must_use)] +//! use std::io::Write; +//! let mut w = Vec::new(); +//! write!(&mut w, "Hello {}!", "world"); +//! ``` +//! +//! ### `print!` +//! +//! This and [`println!`] emit their output to stdout. Similarly to the [`write!`] +//! macro, the goal of these macros is to avoid intermediate allocations when +//! printing output. Example usage is: +//! +//! ``` +//! print!("Hello {}!", "world"); +//! println!("I have a newline {}", "character at the end"); +//! ``` +//! ### `eprint!` +//! +//! The [`eprint!`] and [`eprintln!`] macros are identical to +//! [`print!`] and [`println!`], respectively, except they emit their +//! output to stderr. +//! +//! ### `format_args!` +//! +//! [`format_args!`] is a curious macro used to safely pass around +//! an opaque object describing the format string. This object +//! does not require any heap allocations to create, and it only +//! references information on the stack. Under the hood, all of +//! the related macros are implemented in terms of this. First +//! off, some example usage is: +//! +//! ``` +//! # #![allow(unused_must_use)] +//! use std::fmt; +//! use std::io::{self, Write}; +//! +//! let mut some_writer = io::stdout(); +//! write!(&mut some_writer, "{}", format_args!("print with a {}", "macro")); +//! +//! fn my_fmt_fn(args: fmt::Arguments) { +//! write!(&mut io::stdout(), "{}", args); +//! } +//! my_fmt_fn(format_args!(", or a {} too", "function")); +//! ``` +//! +//! The result of the [`format_args!`] macro is a value of type [`fmt::Arguments`]. +//! This structure can then be passed to the [`write`] and [`format`] functions +//! inside this module in order to process the format string. +//! The goal of this macro is to even further prevent intermediate allocations +//! when dealing with formatting strings. +//! +//! For example, a logging library could use the standard formatting syntax, but +//! it would internally pass around this structure until it has been determined +//! where output should go to. +//! +//! [`fmt::Result`]: Result "fmt::Result" +//! [Result]: core::result::Result "std::result::Result" +//! [std::fmt::Error]: Error "fmt::Error" +//! [`write`]: write() "fmt::write" +//! [`to_string`]: crate::string::ToString::to_string "ToString::to_string" +//! [`write_fmt`]: ../../std/io/trait.Write.html#method.write_fmt +//! [`std::io::Write`]: ../../std/io/trait.Write.html +//! [`std::fmt::Write`]: ../../std/fmt/trait.Write.html +//! [`print!`]: ../../std/macro.print.html "print!" +//! [`println!`]: ../../std/macro.println.html "println!" +//! [`eprint!`]: ../../std/macro.eprint.html "eprint!" +//! [`eprintln!`]: ../../std/macro.eprintln.html "eprintln!" +//! [`format_args!`]: ../../std/macro.format_args.html "format_args!" +//! [`fmt::Arguments`]: Arguments "fmt::Arguments" +//! [`format`]: format() "fmt::format" + +#![stable(feature = "rust1", since = "1.0.0")] + +#[unstable(feature = "fmt_internals", issue = "none")] +pub use core::fmt::rt; +#[stable(feature = "fmt_flags_align", since = "1.28.0")] +pub use core::fmt::Alignment; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::Error; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{write, ArgumentV1, Arguments}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{Binary, Octal}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{Debug, Display}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{DebugList, DebugMap, DebugSet, DebugStruct, DebugTuple}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{Formatter, Result, Write}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{LowerExp, UpperExp}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::fmt::{LowerHex, Pointer, UpperHex}; + +#[cfg(not(no_global_oom_handling))] +use crate::string; + +/// The `format` function takes an [`Arguments`] struct and returns the resulting +/// formatted string. +/// +/// The [`Arguments`] instance can be created with the [`format_args!`] macro. +/// +/// # Examples +/// +/// Basic usage: +/// +/// ``` +/// use std::fmt; +/// +/// let s = fmt::format(format_args!("Hello, {}!", "world")); +/// assert_eq!(s, "Hello, world!"); +/// ``` +/// +/// Please note that using [`format!`] might be preferable. +/// Example: +/// +/// ``` +/// let s = format!("Hello, {}!", "world"); +/// assert_eq!(s, "Hello, world!"); +/// ``` +/// +/// [`format_args!`]: core::format_args +/// [`format!`]: crate::format +#[cfg(not(no_global_oom_handling))] +#[must_use] +#[stable(feature = "rust1", since = "1.0.0")] +#[inline] +pub fn format(args: Arguments<'_>) -> string::String { + fn format_inner(args: Arguments<'_>) -> string::String { + let capacity = args.estimated_capacity(); + let mut output = string::String::with_capacity(capacity); + output.write_fmt(args).expect("a formatting trait implementation returned an error"); + output + } + + args.as_str().map_or_else(|| format_inner(args), crate::borrow::ToOwned::to_owned) +} diff --git a/library/alloc/src/lib.rs b/library/alloc/src/lib.rs new file mode 100644 index 000000000..8b6f40548 --- /dev/null +++ b/library/alloc/src/lib.rs @@ -0,0 +1,240 @@ +//! # The Rust core allocation and collections library +//! +//! This library provides smart pointers and collections for managing +//! heap-allocated values. +//! +//! This library, like libcore, normally doesn’t need to be used directly +//! since its contents are re-exported in the [`std` crate](../std/index.html). +//! Crates that use the `#![no_std]` attribute however will typically +//! not depend on `std`, so they’d use this crate instead. +//! +//! ## Boxed values +//! +//! The [`Box`] type is a smart pointer type. There can only be one owner of a +//! [`Box`], and the owner can decide to mutate the contents, which live on the +//! heap. +//! +//! This type can be sent among threads efficiently as the size of a `Box` value +//! is the same as that of a pointer. Tree-like data structures are often built +//! with boxes because each node often has only one owner, the parent. +//! +//! ## Reference counted pointers +//! +//! The [`Rc`] type is a non-threadsafe reference-counted pointer type intended +//! for sharing memory within a thread. An [`Rc`] pointer wraps a type, `T`, and +//! only allows access to `&T`, a shared reference. +//! +//! This type is useful when inherited mutability (such as using [`Box`]) is too +//! constraining for an application, and is often paired with the [`Cell`] or +//! [`RefCell`] types in order to allow mutation. +//! +//! ## Atomically reference counted pointers +//! +//! The [`Arc`] type is the threadsafe equivalent of the [`Rc`] type. It +//! provides all the same functionality of [`Rc`], except it requires that the +//! contained type `T` is shareable. Additionally, [`Arc<T>`][`Arc`] is itself +//! sendable while [`Rc<T>`][`Rc`] is not. +//! +//! This type allows for shared access to the contained data, and is often +//! paired with synchronization primitives such as mutexes to allow mutation of +//! shared resources. +//! +//! ## Collections +//! +//! Implementations of the most common general purpose data structures are +//! defined in this library. They are re-exported through the +//! [standard collections library](../std/collections/index.html). +//! +//! ## Heap interfaces +//! +//! The [`alloc`](alloc/index.html) module defines the low-level interface to the +//! default global allocator. It is not compatible with the libc allocator API. +//! +//! [`Arc`]: sync +//! [`Box`]: boxed +//! [`Cell`]: core::cell +//! [`Rc`]: rc +//! [`RefCell`]: core::cell + +// To run liballoc tests without x.py without ending up with two copies of liballoc, Miri needs to be +// able to "empty" this crate. See <https://github.com/rust-lang/miri-test-libstd/issues/4>. +// rustc itself never sets the feature, so this line has no affect there. +#![cfg(any(not(feature = "miri-test-libstd"), test, doctest))] +#![allow(unused_attributes)] +#![stable(feature = "alloc", since = "1.36.0")] +#![doc( + html_playground_url = "https://play.rust-lang.org/", + issue_tracker_base_url = "https://github.com/rust-lang/rust/issues/", + test(no_crate_inject, attr(allow(unused_variables), deny(warnings))) +)] +#![doc(cfg_hide( + not(test), + not(any(test, bootstrap)), + any(not(feature = "miri-test-libstd"), test, doctest), + no_global_oom_handling, + not(no_global_oom_handling), + target_has_atomic = "ptr" +))] +#![no_std] +#![needs_allocator] +// +// Lints: +#![deny(unsafe_op_in_unsafe_fn)] +#![warn(deprecated_in_future)] +#![warn(missing_debug_implementations)] +#![warn(missing_docs)] +#![allow(explicit_outlives_requirements)] +// +// Library features: +#![feature(alloc_layout_extra)] +#![feature(allocator_api)] +#![feature(array_chunks)] +#![feature(array_into_iter_constructors)] +#![feature(array_methods)] +#![feature(array_windows)] +#![feature(assert_matches)] +#![feature(async_iterator)] +#![feature(coerce_unsized)] +#![cfg_attr(not(no_global_oom_handling), feature(const_alloc_error))] +#![feature(const_box)] +#![cfg_attr(not(no_global_oom_handling), feature(const_btree_new))] +#![feature(const_cow_is_borrowed)] +#![feature(const_convert)] +#![feature(const_size_of_val)] +#![feature(const_align_of_val)] +#![feature(const_ptr_read)] +#![feature(const_maybe_uninit_write)] +#![feature(const_maybe_uninit_as_mut_ptr)] +#![feature(const_refs_to_cell)] +#![feature(core_intrinsics)] +#![feature(const_eval_select)] +#![feature(const_pin)] +#![feature(cstr_from_bytes_until_nul)] +#![feature(dispatch_from_dyn)] +#![feature(exact_size_is_empty)] +#![feature(extend_one)] +#![feature(fmt_internals)] +#![feature(fn_traits)] +#![feature(hasher_prefixfree_extras)] +#![feature(inplace_iteration)] +#![feature(iter_advance_by)] +#![feature(iter_next_chunk)] +#![feature(layout_for_ptr)] +#![feature(maybe_uninit_array_assume_init)] +#![feature(maybe_uninit_slice)] +#![feature(maybe_uninit_uninit_array)] +#![cfg_attr(test, feature(new_uninit))] +#![feature(nonnull_slice_from_raw_parts)] +#![feature(pattern)] +#![feature(pointer_byte_offsets)] +#![feature(ptr_internals)] +#![feature(ptr_metadata)] +#![feature(ptr_sub_ptr)] +#![feature(receiver_trait)] +#![feature(set_ptr_value)] +#![feature(slice_from_ptr_range)] +#![feature(slice_group_by)] +#![feature(slice_ptr_get)] +#![feature(slice_ptr_len)] +#![feature(slice_range)] +#![feature(str_internals)] +#![feature(strict_provenance)] +#![feature(trusted_len)] +#![feature(trusted_random_access)] +#![feature(try_trait_v2)] +#![feature(unchecked_math)] +#![feature(unicode_internals)] +#![feature(unsize)] +#![feature(std_internals)] +// +// Language features: +#![feature(allocator_internals)] +#![feature(allow_internal_unstable)] +#![feature(associated_type_bounds)] +#![feature(cfg_sanitize)] +#![feature(const_deref)] +#![feature(const_mut_refs)] +#![feature(const_ptr_write)] +#![feature(const_precise_live_drops)] +#![feature(const_trait_impl)] +#![feature(const_try)] +#![feature(dropck_eyepatch)] +#![feature(exclusive_range_pattern)] +#![feature(fundamental)] +#![cfg_attr(not(test), feature(generator_trait))] +#![feature(hashmap_internals)] +#![feature(lang_items)] +#![feature(let_else)] +#![feature(min_specialization)] +#![feature(negative_impls)] +#![feature(never_type)] +#![feature(rustc_allow_const_fn_unstable)] +#![feature(rustc_attrs)] +#![feature(pointer_is_aligned)] +#![feature(slice_internals)] +#![feature(staged_api)] +#![feature(stmt_expr_attributes)] +#![cfg_attr(test, feature(test))] +#![feature(unboxed_closures)] +#![feature(unsized_fn_params)] +#![feature(c_unwind)] +// +// Rustdoc features: +#![feature(doc_cfg)] +#![feature(doc_cfg_hide)] +// Technically, this is a bug in rustdoc: rustdoc sees the documentation on `#[lang = slice_alloc]` +// blocks is for `&[T]`, which also has documentation using this feature in `core`, and gets mad +// that the feature-gate isn't enabled. Ideally, it wouldn't check for the feature gate for docs +// from other crates, but since this can only appear for lang items, it doesn't seem worth fixing. +#![feature(intra_doc_pointers)] + +// Allow testing this library +#[cfg(test)] +#[macro_use] +extern crate std; +#[cfg(test)] +extern crate test; + +// Module with internal macros used by other modules (needs to be included before other modules). +#[macro_use] +mod macros; + +mod raw_vec; + +// Heaps provided for low-level allocation strategies + +pub mod alloc; + +// Primitive types using the heaps above + +// Need to conditionally define the mod from `boxed.rs` to avoid +// duplicating the lang-items when building in test cfg; but also need +// to allow code to have `use boxed::Box;` declarations. +#[cfg(not(test))] +pub mod boxed; +#[cfg(test)] +mod boxed { + pub use std::boxed::Box; +} +pub mod borrow; +pub mod collections; +#[cfg(not(no_global_oom_handling))] +pub mod ffi; +pub mod fmt; +pub mod rc; +pub mod slice; +pub mod str; +pub mod string; +#[cfg(target_has_atomic = "ptr")] +pub mod sync; +#[cfg(all(not(no_global_oom_handling), target_has_atomic = "ptr"))] +pub mod task; +#[cfg(test)] +mod tests; +pub mod vec; + +#[doc(hidden)] +#[unstable(feature = "liballoc_internals", issue = "none", reason = "implementation detail")] +pub mod __export { + pub use core::format_args; +} diff --git a/library/alloc/src/macros.rs b/library/alloc/src/macros.rs new file mode 100644 index 000000000..88eb6aa7a --- /dev/null +++ b/library/alloc/src/macros.rs @@ -0,0 +1,129 @@ +/// Creates a [`Vec`] containing the arguments. +/// +/// `vec!` allows `Vec`s to be defined with the same syntax as array expressions. +/// There are two forms of this macro: +/// +/// - Create a [`Vec`] containing a given list of elements: +/// +/// ``` +/// let v = vec![1, 2, 3]; +/// assert_eq!(v[0], 1); +/// assert_eq!(v[1], 2); +/// assert_eq!(v[2], 3); +/// ``` +/// +/// - Create a [`Vec`] from a given element and size: +/// +/// ``` +/// let v = vec![1; 3]; +/// assert_eq!(v, [1, 1, 1]); +/// ``` +/// +/// Note that unlike array expressions this syntax supports all elements +/// which implement [`Clone`] and the number of elements doesn't have to be +/// a constant. +/// +/// This will use `clone` to duplicate an expression, so one should be careful +/// using this with types having a nonstandard `Clone` implementation. For +/// example, `vec![Rc::new(1); 5]` will create a vector of five references +/// to the same boxed integer value, not five references pointing to independently +/// boxed integers. +/// +/// Also, note that `vec![expr; 0]` is allowed, and produces an empty vector. +/// This will still evaluate `expr`, however, and immediately drop the resulting value, so +/// be mindful of side effects. +/// +/// [`Vec`]: crate::vec::Vec +#[cfg(all(not(no_global_oom_handling), not(test)))] +#[macro_export] +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_diagnostic_item = "vec_macro"] +#[allow_internal_unstable(rustc_attrs, liballoc_internals)] +macro_rules! vec { + () => ( + $crate::__rust_force_expr!($crate::vec::Vec::new()) + ); + ($elem:expr; $n:expr) => ( + $crate::__rust_force_expr!($crate::vec::from_elem($elem, $n)) + ); + ($($x:expr),+ $(,)?) => ( + $crate::__rust_force_expr!(<[_]>::into_vec( + #[rustc_box] + $crate::boxed::Box::new([$($x),+]) + )) + ); +} + +// HACK(japaric): with cfg(test) the inherent `[T]::into_vec` method, which is +// required for this macro definition, is not available. Instead use the +// `slice::into_vec` function which is only available with cfg(test) +// NB see the slice::hack module in slice.rs for more information +#[cfg(all(not(no_global_oom_handling), test))] +#[allow(unused_macro_rules)] +macro_rules! vec { + () => ( + $crate::vec::Vec::new() + ); + ($elem:expr; $n:expr) => ( + $crate::vec::from_elem($elem, $n) + ); + ($($x:expr),*) => ( + $crate::slice::into_vec($crate::boxed::Box::new([$($x),*])) + ); + ($($x:expr,)*) => (vec![$($x),*]) +} + +/// Creates a `String` using interpolation of runtime expressions. +/// +/// The first argument `format!` receives is a format string. This must be a string +/// literal. The power of the formatting string is in the `{}`s contained. +/// +/// Additional parameters passed to `format!` replace the `{}`s within the +/// formatting string in the order given unless named or positional parameters +/// are used; see [`std::fmt`] for more information. +/// +/// A common use for `format!` is concatenation and interpolation of strings. +/// The same convention is used with [`print!`] and [`write!`] macros, +/// depending on the intended destination of the string. +/// +/// To convert a single value to a string, use the [`to_string`] method. This +/// will use the [`Display`] formatting trait. +/// +/// [`std::fmt`]: ../std/fmt/index.html +/// [`print!`]: ../std/macro.print.html +/// [`write!`]: core::write +/// [`to_string`]: crate::string::ToString +/// [`Display`]: core::fmt::Display +/// +/// # Panics +/// +/// `format!` panics if a formatting trait implementation returns an error. +/// This indicates an incorrect implementation +/// since `fmt::Write for String` never returns an error itself. +/// +/// # Examples +/// +/// ``` +/// format!("test"); +/// format!("hello {}", "world!"); +/// format!("x = {}, y = {y}", 10, y = 30); +/// ``` +#[macro_export] +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "format_macro")] +macro_rules! format { + ($($arg:tt)*) => {{ + let res = $crate::fmt::format($crate::__export::format_args!($($arg)*)); + res + }} +} + +/// Force AST node to an expression to improve diagnostics in pattern position. +#[doc(hidden)] +#[macro_export] +#[unstable(feature = "liballoc_internals", issue = "none", reason = "implementation detail")] +macro_rules! __rust_force_expr { + ($e:expr) => { + $e + }; +} diff --git a/library/alloc/src/raw_vec.rs b/library/alloc/src/raw_vec.rs new file mode 100644 index 000000000..b0f4529ab --- /dev/null +++ b/library/alloc/src/raw_vec.rs @@ -0,0 +1,519 @@ +#![unstable(feature = "raw_vec_internals", reason = "unstable const warnings", issue = "none")] + +use core::alloc::LayoutError; +use core::cmp; +use core::intrinsics; +use core::mem::{self, ManuallyDrop, MaybeUninit}; +use core::ops::Drop; +use core::ptr::{self, NonNull, Unique}; +use core::slice; + +#[cfg(not(no_global_oom_handling))] +use crate::alloc::handle_alloc_error; +use crate::alloc::{Allocator, Global, Layout}; +use crate::boxed::Box; +use crate::collections::TryReserveError; +use crate::collections::TryReserveErrorKind::*; + +#[cfg(test)] +mod tests; + +#[cfg(not(no_global_oom_handling))] +enum AllocInit { + /// The contents of the new memory are uninitialized. + Uninitialized, + /// The new memory is guaranteed to be zeroed. + Zeroed, +} + +/// A low-level utility for more ergonomically allocating, reallocating, and deallocating +/// a buffer of memory on the heap without having to worry about all the corner cases +/// involved. This type is excellent for building your own data structures like Vec and VecDeque. +/// In particular: +/// +/// * Produces `Unique::dangling()` on zero-sized types. +/// * Produces `Unique::dangling()` on zero-length allocations. +/// * Avoids freeing `Unique::dangling()`. +/// * Catches all overflows in capacity computations (promotes them to "capacity overflow" panics). +/// * Guards against 32-bit systems allocating more than isize::MAX bytes. +/// * Guards against overflowing your length. +/// * Calls `handle_alloc_error` for fallible allocations. +/// * Contains a `ptr::Unique` and thus endows the user with all related benefits. +/// * Uses the excess returned from the allocator to use the largest available capacity. +/// +/// This type does not in anyway inspect the memory that it manages. When dropped it *will* +/// free its memory, but it *won't* try to drop its contents. It is up to the user of `RawVec` +/// to handle the actual things *stored* inside of a `RawVec`. +/// +/// Note that the excess of a zero-sized types is always infinite, so `capacity()` always returns +/// `usize::MAX`. This means that you need to be careful when round-tripping this type with a +/// `Box<[T]>`, since `capacity()` won't yield the length. +#[allow(missing_debug_implementations)] +pub(crate) struct RawVec<T, A: Allocator = Global> { + ptr: Unique<T>, + cap: usize, + alloc: A, +} + +impl<T> RawVec<T, Global> { + /// HACK(Centril): This exists because stable `const fn` can only call stable `const fn`, so + /// they cannot call `Self::new()`. + /// + /// If you change `RawVec<T>::new` or dependencies, please take care to not introduce anything + /// that would truly const-call something unstable. + pub const NEW: Self = Self::new(); + + /// Creates the biggest possible `RawVec` (on the system heap) + /// without allocating. If `T` has positive size, then this makes a + /// `RawVec` with capacity `0`. If `T` is zero-sized, then it makes a + /// `RawVec` with capacity `usize::MAX`. Useful for implementing + /// delayed allocation. + #[must_use] + pub const fn new() -> Self { + Self::new_in(Global) + } + + /// Creates a `RawVec` (on the system heap) with exactly the + /// capacity and alignment requirements for a `[T; capacity]`. This is + /// equivalent to calling `RawVec::new` when `capacity` is `0` or `T` is + /// zero-sized. Note that if `T` is zero-sized this means you will + /// *not* get a `RawVec` with the requested capacity. + /// + /// # Panics + /// + /// Panics if the requested capacity exceeds `isize::MAX` bytes. + /// + /// # Aborts + /// + /// Aborts on OOM. + #[cfg(not(any(no_global_oom_handling, test)))] + #[must_use] + #[inline] + pub fn with_capacity(capacity: usize) -> Self { + Self::with_capacity_in(capacity, Global) + } + + /// Like `with_capacity`, but guarantees the buffer is zeroed. + #[cfg(not(any(no_global_oom_handling, test)))] + #[must_use] + #[inline] + pub fn with_capacity_zeroed(capacity: usize) -> Self { + Self::with_capacity_zeroed_in(capacity, Global) + } +} + +impl<T, A: Allocator> RawVec<T, A> { + // Tiny Vecs are dumb. Skip to: + // - 8 if the element size is 1, because any heap allocators is likely + // to round up a request of less than 8 bytes to at least 8 bytes. + // - 4 if elements are moderate-sized (<= 1 KiB). + // - 1 otherwise, to avoid wasting too much space for very short Vecs. + pub(crate) const MIN_NON_ZERO_CAP: usize = if mem::size_of::<T>() == 1 { + 8 + } else if mem::size_of::<T>() <= 1024 { + 4 + } else { + 1 + }; + + /// Like `new`, but parameterized over the choice of allocator for + /// the returned `RawVec`. + pub const fn new_in(alloc: A) -> Self { + // `cap: 0` means "unallocated". zero-sized types are ignored. + Self { ptr: Unique::dangling(), cap: 0, alloc } + } + + /// Like `with_capacity`, but parameterized over the choice of + /// allocator for the returned `RawVec`. + #[cfg(not(no_global_oom_handling))] + #[inline] + pub fn with_capacity_in(capacity: usize, alloc: A) -> Self { + Self::allocate_in(capacity, AllocInit::Uninitialized, alloc) + } + + /// Like `with_capacity_zeroed`, but parameterized over the choice + /// of allocator for the returned `RawVec`. + #[cfg(not(no_global_oom_handling))] + #[inline] + pub fn with_capacity_zeroed_in(capacity: usize, alloc: A) -> Self { + Self::allocate_in(capacity, AllocInit::Zeroed, alloc) + } + + /// Converts the entire buffer into `Box<[MaybeUninit<T>]>` with the specified `len`. + /// + /// Note that this will correctly reconstitute any `cap` changes + /// that may have been performed. (See description of type for details.) + /// + /// # Safety + /// + /// * `len` must be greater than or equal to the most recently requested capacity, and + /// * `len` must be less than or equal to `self.capacity()`. + /// + /// Note, that the requested capacity and `self.capacity()` could differ, as + /// an allocator could overallocate and return a greater memory block than requested. + pub unsafe fn into_box(self, len: usize) -> Box<[MaybeUninit<T>], A> { + // Sanity-check one half of the safety requirement (we cannot check the other half). + debug_assert!( + len <= self.capacity(), + "`len` must be smaller than or equal to `self.capacity()`" + ); + + let me = ManuallyDrop::new(self); + unsafe { + let slice = slice::from_raw_parts_mut(me.ptr() as *mut MaybeUninit<T>, len); + Box::from_raw_in(slice, ptr::read(&me.alloc)) + } + } + + #[cfg(not(no_global_oom_handling))] + fn allocate_in(capacity: usize, init: AllocInit, alloc: A) -> Self { + // Don't allocate here because `Drop` will not deallocate when `capacity` is 0. + if mem::size_of::<T>() == 0 || capacity == 0 { + Self::new_in(alloc) + } else { + // We avoid `unwrap_or_else` here because it bloats the amount of + // LLVM IR generated. + let layout = match Layout::array::<T>(capacity) { + Ok(layout) => layout, + Err(_) => capacity_overflow(), + }; + match alloc_guard(layout.size()) { + Ok(_) => {} + Err(_) => capacity_overflow(), + } + let result = match init { + AllocInit::Uninitialized => alloc.allocate(layout), + AllocInit::Zeroed => alloc.allocate_zeroed(layout), + }; + let ptr = match result { + Ok(ptr) => ptr, + Err(_) => handle_alloc_error(layout), + }; + + // Allocators currently return a `NonNull<[u8]>` whose length + // matches the size requested. If that ever changes, the capacity + // here should change to `ptr.len() / mem::size_of::<T>()`. + Self { + ptr: unsafe { Unique::new_unchecked(ptr.cast().as_ptr()) }, + cap: capacity, + alloc, + } + } + } + + /// Reconstitutes a `RawVec` from a pointer, capacity, and allocator. + /// + /// # Safety + /// + /// The `ptr` must be allocated (via the given allocator `alloc`), and with the given + /// `capacity`. + /// The `capacity` cannot exceed `isize::MAX` for sized types. (only a concern on 32-bit + /// systems). ZST vectors may have a capacity up to `usize::MAX`. + /// If the `ptr` and `capacity` come from a `RawVec` created via `alloc`, then this is + /// guaranteed. + #[inline] + pub unsafe fn from_raw_parts_in(ptr: *mut T, capacity: usize, alloc: A) -> Self { + Self { ptr: unsafe { Unique::new_unchecked(ptr) }, cap: capacity, alloc } + } + + /// Gets a raw pointer to the start of the allocation. Note that this is + /// `Unique::dangling()` if `capacity == 0` or `T` is zero-sized. In the former case, you must + /// be careful. + #[inline] + pub fn ptr(&self) -> *mut T { + self.ptr.as_ptr() + } + + /// Gets the capacity of the allocation. + /// + /// This will always be `usize::MAX` if `T` is zero-sized. + #[inline(always)] + pub fn capacity(&self) -> usize { + if mem::size_of::<T>() == 0 { usize::MAX } else { self.cap } + } + + /// Returns a shared reference to the allocator backing this `RawVec`. + pub fn allocator(&self) -> &A { + &self.alloc + } + + fn current_memory(&self) -> Option<(NonNull<u8>, Layout)> { + if mem::size_of::<T>() == 0 || self.cap == 0 { + None + } else { + // We have an allocated chunk of memory, so we can bypass runtime + // checks to get our current layout. + unsafe { + let layout = Layout::array::<T>(self.cap).unwrap_unchecked(); + Some((self.ptr.cast().into(), layout)) + } + } + } + + /// Ensures that the buffer contains at least enough space to hold `len + + /// additional` elements. If it doesn't already have enough capacity, will + /// reallocate enough space plus comfortable slack space to get amortized + /// *O*(1) behavior. Will limit this behavior if it would needlessly cause + /// itself to panic. + /// + /// If `len` exceeds `self.capacity()`, this may fail to actually allocate + /// the requested space. This is not really unsafe, but the unsafe + /// code *you* write that relies on the behavior of this function may break. + /// + /// This is ideal for implementing a bulk-push operation like `extend`. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Aborts + /// + /// Aborts on OOM. + #[cfg(not(no_global_oom_handling))] + #[inline] + pub fn reserve(&mut self, len: usize, additional: usize) { + // Callers expect this function to be very cheap when there is already sufficient capacity. + // Therefore, we move all the resizing and error-handling logic from grow_amortized and + // handle_reserve behind a call, while making sure that this function is likely to be + // inlined as just a comparison and a call if the comparison fails. + #[cold] + fn do_reserve_and_handle<T, A: Allocator>( + slf: &mut RawVec<T, A>, + len: usize, + additional: usize, + ) { + handle_reserve(slf.grow_amortized(len, additional)); + } + + if self.needs_to_grow(len, additional) { + do_reserve_and_handle(self, len, additional); + } + } + + /// A specialized version of `reserve()` used only by the hot and + /// oft-instantiated `Vec::push()`, which does its own capacity check. + #[cfg(not(no_global_oom_handling))] + #[inline(never)] + pub fn reserve_for_push(&mut self, len: usize) { + handle_reserve(self.grow_amortized(len, 1)); + } + + /// The same as `reserve`, but returns on errors instead of panicking or aborting. + pub fn try_reserve(&mut self, len: usize, additional: usize) -> Result<(), TryReserveError> { + if self.needs_to_grow(len, additional) { + self.grow_amortized(len, additional) + } else { + Ok(()) + } + } + + /// Ensures that the buffer contains at least enough space to hold `len + + /// additional` elements. If it doesn't already, will reallocate the + /// minimum possible amount of memory necessary. Generally this will be + /// exactly the amount of memory necessary, but in principle the allocator + /// is free to give back more than we asked for. + /// + /// If `len` exceeds `self.capacity()`, this may fail to actually allocate + /// the requested space. This is not really unsafe, but the unsafe code + /// *you* write that relies on the behavior of this function may break. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Aborts + /// + /// Aborts on OOM. + #[cfg(not(no_global_oom_handling))] + pub fn reserve_exact(&mut self, len: usize, additional: usize) { + handle_reserve(self.try_reserve_exact(len, additional)); + } + + /// The same as `reserve_exact`, but returns on errors instead of panicking or aborting. + pub fn try_reserve_exact( + &mut self, + len: usize, + additional: usize, + ) -> Result<(), TryReserveError> { + if self.needs_to_grow(len, additional) { self.grow_exact(len, additional) } else { Ok(()) } + } + + /// Shrinks the buffer down to the specified capacity. If the given amount + /// is 0, actually completely deallocates. + /// + /// # Panics + /// + /// Panics if the given amount is *larger* than the current capacity. + /// + /// # Aborts + /// + /// Aborts on OOM. + #[cfg(not(no_global_oom_handling))] + pub fn shrink_to_fit(&mut self, cap: usize) { + handle_reserve(self.shrink(cap)); + } +} + +impl<T, A: Allocator> RawVec<T, A> { + /// Returns if the buffer needs to grow to fulfill the needed extra capacity. + /// Mainly used to make inlining reserve-calls possible without inlining `grow`. + fn needs_to_grow(&self, len: usize, additional: usize) -> bool { + additional > self.capacity().wrapping_sub(len) + } + + fn set_ptr_and_cap(&mut self, ptr: NonNull<[u8]>, cap: usize) { + // Allocators currently return a `NonNull<[u8]>` whose length matches + // the size requested. If that ever changes, the capacity here should + // change to `ptr.len() / mem::size_of::<T>()`. + self.ptr = unsafe { Unique::new_unchecked(ptr.cast().as_ptr()) }; + self.cap = cap; + } + + // This method is usually instantiated many times. So we want it to be as + // small as possible, to improve compile times. But we also want as much of + // its contents to be statically computable as possible, to make the + // generated code run faster. Therefore, this method is carefully written + // so that all of the code that depends on `T` is within it, while as much + // of the code that doesn't depend on `T` as possible is in functions that + // are non-generic over `T`. + fn grow_amortized(&mut self, len: usize, additional: usize) -> Result<(), TryReserveError> { + // This is ensured by the calling contexts. + debug_assert!(additional > 0); + + if mem::size_of::<T>() == 0 { + // Since we return a capacity of `usize::MAX` when `elem_size` is + // 0, getting to here necessarily means the `RawVec` is overfull. + return Err(CapacityOverflow.into()); + } + + // Nothing we can really do about these checks, sadly. + let required_cap = len.checked_add(additional).ok_or(CapacityOverflow)?; + + // This guarantees exponential growth. The doubling cannot overflow + // because `cap <= isize::MAX` and the type of `cap` is `usize`. + let cap = cmp::max(self.cap * 2, required_cap); + let cap = cmp::max(Self::MIN_NON_ZERO_CAP, cap); + + let new_layout = Layout::array::<T>(cap); + + // `finish_grow` is non-generic over `T`. + let ptr = finish_grow(new_layout, self.current_memory(), &mut self.alloc)?; + self.set_ptr_and_cap(ptr, cap); + Ok(()) + } + + // The constraints on this method are much the same as those on + // `grow_amortized`, but this method is usually instantiated less often so + // it's less critical. + fn grow_exact(&mut self, len: usize, additional: usize) -> Result<(), TryReserveError> { + if mem::size_of::<T>() == 0 { + // Since we return a capacity of `usize::MAX` when the type size is + // 0, getting to here necessarily means the `RawVec` is overfull. + return Err(CapacityOverflow.into()); + } + + let cap = len.checked_add(additional).ok_or(CapacityOverflow)?; + let new_layout = Layout::array::<T>(cap); + + // `finish_grow` is non-generic over `T`. + let ptr = finish_grow(new_layout, self.current_memory(), &mut self.alloc)?; + self.set_ptr_and_cap(ptr, cap); + Ok(()) + } + + #[cfg(not(no_global_oom_handling))] + fn shrink(&mut self, cap: usize) -> Result<(), TryReserveError> { + assert!(cap <= self.capacity(), "Tried to shrink to a larger capacity"); + + let (ptr, layout) = if let Some(mem) = self.current_memory() { mem } else { return Ok(()) }; + + let ptr = unsafe { + // `Layout::array` cannot overflow here because it would have + // overflowed earlier when capacity was larger. + let new_layout = Layout::array::<T>(cap).unwrap_unchecked(); + self.alloc + .shrink(ptr, layout, new_layout) + .map_err(|_| AllocError { layout: new_layout, non_exhaustive: () })? + }; + self.set_ptr_and_cap(ptr, cap); + Ok(()) + } +} + +// This function is outside `RawVec` to minimize compile times. See the comment +// above `RawVec::grow_amortized` for details. (The `A` parameter isn't +// significant, because the number of different `A` types seen in practice is +// much smaller than the number of `T` types.) +#[inline(never)] +fn finish_grow<A>( + new_layout: Result<Layout, LayoutError>, + current_memory: Option<(NonNull<u8>, Layout)>, + alloc: &mut A, +) -> Result<NonNull<[u8]>, TryReserveError> +where + A: Allocator, +{ + // Check for the error here to minimize the size of `RawVec::grow_*`. + let new_layout = new_layout.map_err(|_| CapacityOverflow)?; + + alloc_guard(new_layout.size())?; + + let memory = if let Some((ptr, old_layout)) = current_memory { + debug_assert_eq!(old_layout.align(), new_layout.align()); + unsafe { + // The allocator checks for alignment equality + intrinsics::assume(old_layout.align() == new_layout.align()); + alloc.grow(ptr, old_layout, new_layout) + } + } else { + alloc.allocate(new_layout) + }; + + memory.map_err(|_| AllocError { layout: new_layout, non_exhaustive: () }.into()) +} + +unsafe impl<#[may_dangle] T, A: Allocator> Drop for RawVec<T, A> { + /// Frees the memory owned by the `RawVec` *without* trying to drop its contents. + fn drop(&mut self) { + if let Some((ptr, layout)) = self.current_memory() { + unsafe { self.alloc.deallocate(ptr, layout) } + } + } +} + +// Central function for reserve error handling. +#[cfg(not(no_global_oom_handling))] +#[inline] +fn handle_reserve(result: Result<(), TryReserveError>) { + match result.map_err(|e| e.kind()) { + Err(CapacityOverflow) => capacity_overflow(), + Err(AllocError { layout, .. }) => handle_alloc_error(layout), + Ok(()) => { /* yay */ } + } +} + +// We need to guarantee the following: +// * We don't ever allocate `> isize::MAX` byte-size objects. +// * We don't overflow `usize::MAX` and actually allocate too little. +// +// On 64-bit we just need to check for overflow since trying to allocate +// `> isize::MAX` bytes will surely fail. On 32-bit and 16-bit we need to add +// an extra guard for this in case we're running on a platform which can use +// all 4GB in user-space, e.g., PAE or x32. + +#[inline] +fn alloc_guard(alloc_size: usize) -> Result<(), TryReserveError> { + if usize::BITS < 64 && alloc_size > isize::MAX as usize { + Err(CapacityOverflow.into()) + } else { + Ok(()) + } +} + +// One central function responsible for reporting capacity overflows. This'll +// ensure that the code generation related to these panics is minimal as there's +// only one location which panics rather than a bunch throughout the module. +#[cfg(not(no_global_oom_handling))] +fn capacity_overflow() -> ! { + panic!("capacity overflow"); +} diff --git a/library/alloc/src/raw_vec/tests.rs b/library/alloc/src/raw_vec/tests.rs new file mode 100644 index 000000000..ff322f0da --- /dev/null +++ b/library/alloc/src/raw_vec/tests.rs @@ -0,0 +1,163 @@ +use super::*; +use std::cell::Cell; + +#[test] +fn allocator_param() { + use crate::alloc::AllocError; + + // Writing a test of integration between third-party + // allocators and `RawVec` is a little tricky because the `RawVec` + // API does not expose fallible allocation methods, so we + // cannot check what happens when allocator is exhausted + // (beyond detecting a panic). + // + // Instead, this just checks that the `RawVec` methods do at + // least go through the Allocator API when it reserves + // storage. + + // A dumb allocator that consumes a fixed amount of fuel + // before allocation attempts start failing. + struct BoundedAlloc { + fuel: Cell<usize>, + } + unsafe impl Allocator for BoundedAlloc { + fn allocate(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError> { + let size = layout.size(); + if size > self.fuel.get() { + return Err(AllocError); + } + match Global.allocate(layout) { + ok @ Ok(_) => { + self.fuel.set(self.fuel.get() - size); + ok + } + err @ Err(_) => err, + } + } + unsafe fn deallocate(&self, ptr: NonNull<u8>, layout: Layout) { + unsafe { Global.deallocate(ptr, layout) } + } + } + + let a = BoundedAlloc { fuel: Cell::new(500) }; + let mut v: RawVec<u8, _> = RawVec::with_capacity_in(50, a); + assert_eq!(v.alloc.fuel.get(), 450); + v.reserve(50, 150); // (causes a realloc, thus using 50 + 150 = 200 units of fuel) + assert_eq!(v.alloc.fuel.get(), 250); +} + +#[test] +fn reserve_does_not_overallocate() { + { + let mut v: RawVec<u32> = RawVec::new(); + // First, `reserve` allocates like `reserve_exact`. + v.reserve(0, 9); + assert_eq!(9, v.capacity()); + } + + { + let mut v: RawVec<u32> = RawVec::new(); + v.reserve(0, 7); + assert_eq!(7, v.capacity()); + // 97 is more than double of 7, so `reserve` should work + // like `reserve_exact`. + v.reserve(7, 90); + assert_eq!(97, v.capacity()); + } + + { + let mut v: RawVec<u32> = RawVec::new(); + v.reserve(0, 12); + assert_eq!(12, v.capacity()); + v.reserve(12, 3); + // 3 is less than half of 12, so `reserve` must grow + // exponentially. At the time of writing this test grow + // factor is 2, so new capacity is 24, however, grow factor + // of 1.5 is OK too. Hence `>= 18` in assert. + assert!(v.capacity() >= 12 + 12 / 2); + } +} + +struct ZST; + +// A `RawVec` holding zero-sized elements should always look like this. +fn zst_sanity<T>(v: &RawVec<T>) { + assert_eq!(v.capacity(), usize::MAX); + assert_eq!(v.ptr(), core::ptr::Unique::<T>::dangling().as_ptr()); + assert_eq!(v.current_memory(), None); +} + +#[test] +fn zst() { + let cap_err = Err(crate::collections::TryReserveErrorKind::CapacityOverflow.into()); + + assert_eq!(std::mem::size_of::<ZST>(), 0); + + // All these different ways of creating the RawVec produce the same thing. + + let v: RawVec<ZST> = RawVec::new(); + zst_sanity(&v); + + let v: RawVec<ZST> = RawVec::with_capacity_in(100, Global); + zst_sanity(&v); + + let v: RawVec<ZST> = RawVec::with_capacity_in(100, Global); + zst_sanity(&v); + + let v: RawVec<ZST> = RawVec::allocate_in(0, AllocInit::Uninitialized, Global); + zst_sanity(&v); + + let v: RawVec<ZST> = RawVec::allocate_in(100, AllocInit::Uninitialized, Global); + zst_sanity(&v); + + let mut v: RawVec<ZST> = RawVec::allocate_in(usize::MAX, AllocInit::Uninitialized, Global); + zst_sanity(&v); + + // Check all these operations work as expected with zero-sized elements. + + assert!(!v.needs_to_grow(100, usize::MAX - 100)); + assert!(v.needs_to_grow(101, usize::MAX - 100)); + zst_sanity(&v); + + v.reserve(100, usize::MAX - 100); + //v.reserve(101, usize::MAX - 100); // panics, in `zst_reserve_panic` below + zst_sanity(&v); + + v.reserve_exact(100, usize::MAX - 100); + //v.reserve_exact(101, usize::MAX - 100); // panics, in `zst_reserve_exact_panic` below + zst_sanity(&v); + + assert_eq!(v.try_reserve(100, usize::MAX - 100), Ok(())); + assert_eq!(v.try_reserve(101, usize::MAX - 100), cap_err); + zst_sanity(&v); + + assert_eq!(v.try_reserve_exact(100, usize::MAX - 100), Ok(())); + assert_eq!(v.try_reserve_exact(101, usize::MAX - 100), cap_err); + zst_sanity(&v); + + assert_eq!(v.grow_amortized(100, usize::MAX - 100), cap_err); + assert_eq!(v.grow_amortized(101, usize::MAX - 100), cap_err); + zst_sanity(&v); + + assert_eq!(v.grow_exact(100, usize::MAX - 100), cap_err); + assert_eq!(v.grow_exact(101, usize::MAX - 100), cap_err); + zst_sanity(&v); +} + +#[test] +#[should_panic(expected = "capacity overflow")] +fn zst_reserve_panic() { + let mut v: RawVec<ZST> = RawVec::new(); + zst_sanity(&v); + + v.reserve(101, usize::MAX - 100); +} + +#[test] +#[should_panic(expected = "capacity overflow")] +fn zst_reserve_exact_panic() { + let mut v: RawVec<ZST> = RawVec::new(); + zst_sanity(&v); + + v.reserve_exact(101, usize::MAX - 100); +} diff --git a/library/alloc/src/rc.rs b/library/alloc/src/rc.rs new file mode 100644 index 000000000..b89b03683 --- /dev/null +++ b/library/alloc/src/rc.rs @@ -0,0 +1,2700 @@ +//! Single-threaded reference-counting pointers. 'Rc' stands for 'Reference +//! Counted'. +//! +//! The type [`Rc<T>`][`Rc`] provides shared ownership of a value of type `T`, +//! allocated in the heap. Invoking [`clone`][clone] on [`Rc`] produces a new +//! pointer to the same allocation in the heap. When the last [`Rc`] pointer to a +//! given allocation is destroyed, the value stored in that allocation (often +//! referred to as "inner value") is also dropped. +//! +//! Shared references in Rust disallow mutation by default, and [`Rc`] +//! is no exception: you cannot generally obtain a mutable reference to +//! something inside an [`Rc`]. If you need mutability, put a [`Cell`] +//! or [`RefCell`] inside the [`Rc`]; see [an example of mutability +//! inside an `Rc`][mutability]. +//! +//! [`Rc`] uses non-atomic reference counting. This means that overhead is very +//! low, but an [`Rc`] cannot be sent between threads, and consequently [`Rc`] +//! does not implement [`Send`][send]. As a result, the Rust compiler +//! will check *at compile time* that you are not sending [`Rc`]s between +//! threads. If you need multi-threaded, atomic reference counting, use +//! [`sync::Arc`][arc]. +//! +//! The [`downgrade`][downgrade] method can be used to create a non-owning +//! [`Weak`] pointer. A [`Weak`] pointer can be [`upgrade`][upgrade]d +//! to an [`Rc`], but this will return [`None`] if the value stored in the allocation has +//! already been dropped. In other words, `Weak` pointers do not keep the value +//! inside the allocation alive; however, they *do* keep the allocation +//! (the backing store for the inner value) alive. +//! +//! A cycle between [`Rc`] pointers will never be deallocated. For this reason, +//! [`Weak`] is used to break cycles. For example, a tree could have strong +//! [`Rc`] pointers from parent nodes to children, and [`Weak`] pointers from +//! children back to their parents. +//! +//! `Rc<T>` automatically dereferences to `T` (via the [`Deref`] trait), +//! so you can call `T`'s methods on a value of type [`Rc<T>`][`Rc`]. To avoid name +//! clashes with `T`'s methods, the methods of [`Rc<T>`][`Rc`] itself are associated +//! functions, called using [fully qualified syntax]: +//! +//! ``` +//! use std::rc::Rc; +//! +//! let my_rc = Rc::new(()); +//! let my_weak = Rc::downgrade(&my_rc); +//! ``` +//! +//! `Rc<T>`'s implementations of traits like `Clone` may also be called using +//! fully qualified syntax. Some people prefer to use fully qualified syntax, +//! while others prefer using method-call syntax. +//! +//! ``` +//! use std::rc::Rc; +//! +//! let rc = Rc::new(()); +//! // Method-call syntax +//! let rc2 = rc.clone(); +//! // Fully qualified syntax +//! let rc3 = Rc::clone(&rc); +//! ``` +//! +//! [`Weak<T>`][`Weak`] does not auto-dereference to `T`, because the inner value may have +//! already been dropped. +//! +//! # Cloning references +//! +//! Creating a new reference to the same allocation as an existing reference counted pointer +//! is done using the `Clone` trait implemented for [`Rc<T>`][`Rc`] and [`Weak<T>`][`Weak`]. +//! +//! ``` +//! use std::rc::Rc; +//! +//! let foo = Rc::new(vec![1.0, 2.0, 3.0]); +//! // The two syntaxes below are equivalent. +//! let a = foo.clone(); +//! let b = Rc::clone(&foo); +//! // a and b both point to the same memory location as foo. +//! ``` +//! +//! The `Rc::clone(&from)` syntax is the most idiomatic because it conveys more explicitly +//! the meaning of the code. In the example above, this syntax makes it easier to see that +//! this code is creating a new reference rather than copying the whole content of foo. +//! +//! # Examples +//! +//! Consider a scenario where a set of `Gadget`s are owned by a given `Owner`. +//! We want to have our `Gadget`s point to their `Owner`. We can't do this with +//! unique ownership, because more than one gadget may belong to the same +//! `Owner`. [`Rc`] allows us to share an `Owner` between multiple `Gadget`s, +//! and have the `Owner` remain allocated as long as any `Gadget` points at it. +//! +//! ``` +//! use std::rc::Rc; +//! +//! struct Owner { +//! name: String, +//! // ...other fields +//! } +//! +//! struct Gadget { +//! id: i32, +//! owner: Rc<Owner>, +//! // ...other fields +//! } +//! +//! fn main() { +//! // Create a reference-counted `Owner`. +//! let gadget_owner: Rc<Owner> = Rc::new( +//! Owner { +//! name: "Gadget Man".to_string(), +//! } +//! ); +//! +//! // Create `Gadget`s belonging to `gadget_owner`. Cloning the `Rc<Owner>` +//! // gives us a new pointer to the same `Owner` allocation, incrementing +//! // the reference count in the process. +//! let gadget1 = Gadget { +//! id: 1, +//! owner: Rc::clone(&gadget_owner), +//! }; +//! let gadget2 = Gadget { +//! id: 2, +//! owner: Rc::clone(&gadget_owner), +//! }; +//! +//! // Dispose of our local variable `gadget_owner`. +//! drop(gadget_owner); +//! +//! // Despite dropping `gadget_owner`, we're still able to print out the name +//! // of the `Owner` of the `Gadget`s. This is because we've only dropped a +//! // single `Rc<Owner>`, not the `Owner` it points to. As long as there are +//! // other `Rc<Owner>` pointing at the same `Owner` allocation, it will remain +//! // live. The field projection `gadget1.owner.name` works because +//! // `Rc<Owner>` automatically dereferences to `Owner`. +//! println!("Gadget {} owned by {}", gadget1.id, gadget1.owner.name); +//! println!("Gadget {} owned by {}", gadget2.id, gadget2.owner.name); +//! +//! // At the end of the function, `gadget1` and `gadget2` are destroyed, and +//! // with them the last counted references to our `Owner`. Gadget Man now +//! // gets destroyed as well. +//! } +//! ``` +//! +//! If our requirements change, and we also need to be able to traverse from +//! `Owner` to `Gadget`, we will run into problems. An [`Rc`] pointer from `Owner` +//! to `Gadget` introduces a cycle. This means that their +//! reference counts can never reach 0, and the allocation will never be destroyed: +//! a memory leak. In order to get around this, we can use [`Weak`] +//! pointers. +//! +//! Rust actually makes it somewhat difficult to produce this loop in the first +//! place. In order to end up with two values that point at each other, one of +//! them needs to be mutable. This is difficult because [`Rc`] enforces +//! memory safety by only giving out shared references to the value it wraps, +//! and these don't allow direct mutation. We need to wrap the part of the +//! value we wish to mutate in a [`RefCell`], which provides *interior +//! mutability*: a method to achieve mutability through a shared reference. +//! [`RefCell`] enforces Rust's borrowing rules at runtime. +//! +//! ``` +//! use std::rc::Rc; +//! use std::rc::Weak; +//! use std::cell::RefCell; +//! +//! struct Owner { +//! name: String, +//! gadgets: RefCell<Vec<Weak<Gadget>>>, +//! // ...other fields +//! } +//! +//! struct Gadget { +//! id: i32, +//! owner: Rc<Owner>, +//! // ...other fields +//! } +//! +//! fn main() { +//! // Create a reference-counted `Owner`. Note that we've put the `Owner`'s +//! // vector of `Gadget`s inside a `RefCell` so that we can mutate it through +//! // a shared reference. +//! let gadget_owner: Rc<Owner> = Rc::new( +//! Owner { +//! name: "Gadget Man".to_string(), +//! gadgets: RefCell::new(vec![]), +//! } +//! ); +//! +//! // Create `Gadget`s belonging to `gadget_owner`, as before. +//! let gadget1 = Rc::new( +//! Gadget { +//! id: 1, +//! owner: Rc::clone(&gadget_owner), +//! } +//! ); +//! let gadget2 = Rc::new( +//! Gadget { +//! id: 2, +//! owner: Rc::clone(&gadget_owner), +//! } +//! ); +//! +//! // Add the `Gadget`s to their `Owner`. +//! { +//! let mut gadgets = gadget_owner.gadgets.borrow_mut(); +//! gadgets.push(Rc::downgrade(&gadget1)); +//! gadgets.push(Rc::downgrade(&gadget2)); +//! +//! // `RefCell` dynamic borrow ends here. +//! } +//! +//! // Iterate over our `Gadget`s, printing their details out. +//! for gadget_weak in gadget_owner.gadgets.borrow().iter() { +//! +//! // `gadget_weak` is a `Weak<Gadget>`. Since `Weak` pointers can't +//! // guarantee the allocation still exists, we need to call +//! // `upgrade`, which returns an `Option<Rc<Gadget>>`. +//! // +//! // In this case we know the allocation still exists, so we simply +//! // `unwrap` the `Option`. In a more complicated program, you might +//! // need graceful error handling for a `None` result. +//! +//! let gadget = gadget_weak.upgrade().unwrap(); +//! println!("Gadget {} owned by {}", gadget.id, gadget.owner.name); +//! } +//! +//! // At the end of the function, `gadget_owner`, `gadget1`, and `gadget2` +//! // are destroyed. There are now no strong (`Rc`) pointers to the +//! // gadgets, so they are destroyed. This zeroes the reference count on +//! // Gadget Man, so he gets destroyed as well. +//! } +//! ``` +//! +//! [clone]: Clone::clone +//! [`Cell`]: core::cell::Cell +//! [`RefCell`]: core::cell::RefCell +//! [send]: core::marker::Send +//! [arc]: crate::sync::Arc +//! [`Deref`]: core::ops::Deref +//! [downgrade]: Rc::downgrade +//! [upgrade]: Weak::upgrade +//! [mutability]: core::cell#introducing-mutability-inside-of-something-immutable +//! [fully qualified syntax]: https://doc.rust-lang.org/book/ch19-03-advanced-traits.html#fully-qualified-syntax-for-disambiguation-calling-methods-with-the-same-name + +#![stable(feature = "rust1", since = "1.0.0")] + +#[cfg(not(test))] +use crate::boxed::Box; +#[cfg(test)] +use std::boxed::Box; + +use core::any::Any; +use core::borrow; +use core::cell::Cell; +use core::cmp::Ordering; +use core::convert::{From, TryFrom}; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::intrinsics::abort; +#[cfg(not(no_global_oom_handling))] +use core::iter; +use core::marker::{self, PhantomData, Unpin, Unsize}; +#[cfg(not(no_global_oom_handling))] +use core::mem::size_of_val; +use core::mem::{self, align_of_val_raw, forget}; +use core::ops::{CoerceUnsized, Deref, DispatchFromDyn, Receiver}; +use core::panic::{RefUnwindSafe, UnwindSafe}; +#[cfg(not(no_global_oom_handling))] +use core::pin::Pin; +use core::ptr::{self, NonNull}; +#[cfg(not(no_global_oom_handling))] +use core::slice::from_raw_parts_mut; + +#[cfg(not(no_global_oom_handling))] +use crate::alloc::handle_alloc_error; +#[cfg(not(no_global_oom_handling))] +use crate::alloc::{box_free, WriteCloneIntoRaw}; +use crate::alloc::{AllocError, Allocator, Global, Layout}; +use crate::borrow::{Cow, ToOwned}; +#[cfg(not(no_global_oom_handling))] +use crate::string::String; +#[cfg(not(no_global_oom_handling))] +use crate::vec::Vec; + +#[cfg(test)] +mod tests; + +// This is repr(C) to future-proof against possible field-reordering, which +// would interfere with otherwise safe [into|from]_raw() of transmutable +// inner types. +#[repr(C)] +struct RcBox<T: ?Sized> { + strong: Cell<usize>, + weak: Cell<usize>, + value: T, +} + +/// A single-threaded reference-counting pointer. 'Rc' stands for 'Reference +/// Counted'. +/// +/// See the [module-level documentation](./index.html) for more details. +/// +/// The inherent methods of `Rc` are all associated functions, which means +/// that you have to call them as e.g., [`Rc::get_mut(&mut value)`][get_mut] instead of +/// `value.get_mut()`. This avoids conflicts with methods of the inner type `T`. +/// +/// [get_mut]: Rc::get_mut +#[cfg_attr(not(test), rustc_diagnostic_item = "Rc")] +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_insignificant_dtor] +pub struct Rc<T: ?Sized> { + ptr: NonNull<RcBox<T>>, + phantom: PhantomData<RcBox<T>>, +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> !marker::Send for Rc<T> {} + +// Note that this negative impl isn't strictly necessary for correctness, +// as `Rc` transitively contains a `Cell`, which is itself `!Sync`. +// However, given how important `Rc`'s `!Sync`-ness is, +// having an explicit negative impl is nice for documentation purposes +// and results in nicer error messages. +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> !marker::Sync for Rc<T> {} + +#[stable(feature = "catch_unwind", since = "1.9.0")] +impl<T: RefUnwindSafe + ?Sized> UnwindSafe for Rc<T> {} +#[stable(feature = "rc_ref_unwind_safe", since = "1.58.0")] +impl<T: RefUnwindSafe + ?Sized> RefUnwindSafe for Rc<T> {} + +#[unstable(feature = "coerce_unsized", issue = "27732")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> CoerceUnsized<Rc<U>> for Rc<T> {} + +#[unstable(feature = "dispatch_from_dyn", issue = "none")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> DispatchFromDyn<Rc<U>> for Rc<T> {} + +impl<T: ?Sized> Rc<T> { + #[inline(always)] + fn inner(&self) -> &RcBox<T> { + // This unsafety is ok because while this Rc is alive we're guaranteed + // that the inner pointer is valid. + unsafe { self.ptr.as_ref() } + } + + unsafe fn from_inner(ptr: NonNull<RcBox<T>>) -> Self { + Self { ptr, phantom: PhantomData } + } + + unsafe fn from_ptr(ptr: *mut RcBox<T>) -> Self { + unsafe { Self::from_inner(NonNull::new_unchecked(ptr)) } + } +} + +impl<T> Rc<T> { + /// Constructs a new `Rc<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn new(value: T) -> Rc<T> { + // There is an implicit weak pointer owned by all the strong + // pointers, which ensures that the weak destructor never frees + // the allocation while the strong destructor is running, even + // if the weak pointer is stored inside the strong one. + unsafe { + Self::from_inner( + Box::leak(Box::new(RcBox { strong: Cell::new(1), weak: Cell::new(1), value })) + .into(), + ) + } + } + + /// Constructs a new `Rc<T>` while giving you a `Weak<T>` to the allocation, + /// to allow you to construct a `T` which holds a weak pointer to itself. + /// + /// Generally, a structure circularly referencing itself, either directly or + /// indirectly, should not hold a strong reference to itself to prevent a memory leak. + /// Using this function, you get access to the weak pointer during the + /// initialization of `T`, before the `Rc<T>` is created, such that you can + /// clone and store it inside the `T`. + /// + /// `new_cyclic` first allocates the managed allocation for the `Rc<T>`, + /// then calls your closure, giving it a `Weak<T>` to this allocation, + /// and only afterwards completes the construction of the `Rc<T>` by placing + /// the `T` returned from your closure into the allocation. + /// + /// Since the new `Rc<T>` is not fully-constructed until `Rc<T>::new_cyclic` + /// returns, calling [`upgrade`] on the weak reference inside your closure will + /// fail and result in a `None` value. + /// + /// # Panics + /// + /// If `data_fn` panics, the panic is propagated to the caller, and the + /// temporary [`Weak<T>`] is dropped normally. + /// + /// # Examples + /// + /// ``` + /// # #![allow(dead_code)] + /// use std::rc::{Rc, Weak}; + /// + /// struct Gadget { + /// me: Weak<Gadget>, + /// } + /// + /// impl Gadget { + /// /// Construct a reference counted Gadget. + /// fn new() -> Rc<Self> { + /// // `me` is a `Weak<Gadget>` pointing at the new allocation of the + /// // `Rc` we're constructing. + /// Rc::new_cyclic(|me| { + /// // Create the actual struct here. + /// Gadget { me: me.clone() } + /// }) + /// } + /// + /// /// Return a reference counted pointer to Self. + /// fn me(&self) -> Rc<Self> { + /// self.me.upgrade().unwrap() + /// } + /// } + /// ``` + /// [`upgrade`]: Weak::upgrade + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "arc_new_cyclic", since = "1.60.0")] + pub fn new_cyclic<F>(data_fn: F) -> Rc<T> + where + F: FnOnce(&Weak<T>) -> T, + { + // Construct the inner in the "uninitialized" state with a single + // weak reference. + let uninit_ptr: NonNull<_> = Box::leak(Box::new(RcBox { + strong: Cell::new(0), + weak: Cell::new(1), + value: mem::MaybeUninit::<T>::uninit(), + })) + .into(); + + let init_ptr: NonNull<RcBox<T>> = uninit_ptr.cast(); + + let weak = Weak { ptr: init_ptr }; + + // It's important we don't give up ownership of the weak pointer, or + // else the memory might be freed by the time `data_fn` returns. If + // we really wanted to pass ownership, we could create an additional + // weak pointer for ourselves, but this would result in additional + // updates to the weak reference count which might not be necessary + // otherwise. + let data = data_fn(&weak); + + let strong = unsafe { + let inner = init_ptr.as_ptr(); + ptr::write(ptr::addr_of_mut!((*inner).value), data); + + let prev_value = (*inner).strong.get(); + debug_assert_eq!(prev_value, 0, "No prior strong references should exist"); + (*inner).strong.set(1); + + Rc::from_inner(init_ptr) + }; + + // Strong references should collectively own a shared weak reference, + // so don't run the destructor for our old weak reference. + mem::forget(weak); + strong + } + + /// Constructs a new `Rc` with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut five = Rc::<u32>::new_uninit(); + /// + /// // Deferred initialization: + /// Rc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit() -> Rc<mem::MaybeUninit<T>> { + unsafe { + Rc::from_ptr(Rc::allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate(layout), + |mem| mem as *mut RcBox<mem::MaybeUninit<T>>, + )) + } + } + + /// Constructs a new `Rc` with uninitialized contents, with the memory + /// being filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// use std::rc::Rc; + /// + /// let zero = Rc::<u32>::new_zeroed(); + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed() -> Rc<mem::MaybeUninit<T>> { + unsafe { + Rc::from_ptr(Rc::allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate_zeroed(layout), + |mem| mem as *mut RcBox<mem::MaybeUninit<T>>, + )) + } + } + + /// Constructs a new `Rc<T>`, returning an error if the allocation fails + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// use std::rc::Rc; + /// + /// let five = Rc::try_new(5); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + pub fn try_new(value: T) -> Result<Rc<T>, AllocError> { + // There is an implicit weak pointer owned by all the strong + // pointers, which ensures that the weak destructor never frees + // the allocation while the strong destructor is running, even + // if the weak pointer is stored inside the strong one. + unsafe { + Ok(Self::from_inner( + Box::leak(Box::try_new(RcBox { strong: Cell::new(1), weak: Cell::new(1), value })?) + .into(), + )) + } + } + + /// Constructs a new `Rc` with uninitialized contents, returning an error if the allocation fails + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut five = Rc::<u32>::try_new_uninit()?; + /// + /// // Deferred initialization: + /// Rc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + pub fn try_new_uninit() -> Result<Rc<mem::MaybeUninit<T>>, AllocError> { + unsafe { + Ok(Rc::from_ptr(Rc::try_allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate(layout), + |mem| mem as *mut RcBox<mem::MaybeUninit<T>>, + )?)) + } + } + + /// Constructs a new `Rc` with uninitialized contents, with the memory + /// being filled with `0` bytes, returning an error if the allocation fails + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, new_uninit)] + /// + /// use std::rc::Rc; + /// + /// let zero = Rc::<u32>::try_new_zeroed()?; + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + //#[unstable(feature = "new_uninit", issue = "63291")] + pub fn try_new_zeroed() -> Result<Rc<mem::MaybeUninit<T>>, AllocError> { + unsafe { + Ok(Rc::from_ptr(Rc::try_allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate_zeroed(layout), + |mem| mem as *mut RcBox<mem::MaybeUninit<T>>, + )?)) + } + } + /// Constructs a new `Pin<Rc<T>>`. If `T` does not implement `Unpin`, then + /// `value` will be pinned in memory and unable to be moved. + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "pin", since = "1.33.0")] + #[must_use] + pub fn pin(value: T) -> Pin<Rc<T>> { + unsafe { Pin::new_unchecked(Rc::new(value)) } + } + + /// Returns the inner value, if the `Rc` has exactly one strong reference. + /// + /// Otherwise, an [`Err`] is returned with the same `Rc` that was + /// passed in. + /// + /// This will succeed even if there are outstanding weak references. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let x = Rc::new(3); + /// assert_eq!(Rc::try_unwrap(x), Ok(3)); + /// + /// let x = Rc::new(4); + /// let _y = Rc::clone(&x); + /// assert_eq!(*Rc::try_unwrap(x).unwrap_err(), 4); + /// ``` + #[inline] + #[stable(feature = "rc_unique", since = "1.4.0")] + pub fn try_unwrap(this: Self) -> Result<T, Self> { + if Rc::strong_count(&this) == 1 { + unsafe { + let val = ptr::read(&*this); // copy the contained object + + // Indicate to Weaks that they can't be promoted by decrementing + // the strong count, and then remove the implicit "strong weak" + // pointer while also handling drop logic by just crafting a + // fake Weak. + this.inner().dec_strong(); + let _weak = Weak { ptr: this.ptr }; + forget(this); + Ok(val) + } + } else { + Err(this) + } + } +} + +impl<T> Rc<[T]> { + /// Constructs a new reference-counted slice with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut values = Rc::<[u32]>::new_uninit_slice(3); + /// + /// // Deferred initialization: + /// let data = Rc::get_mut(&mut values).unwrap(); + /// data[0].write(1); + /// data[1].write(2); + /// data[2].write(3); + /// + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit_slice(len: usize) -> Rc<[mem::MaybeUninit<T>]> { + unsafe { Rc::from_ptr(Rc::allocate_for_slice(len)) } + } + + /// Constructs a new reference-counted slice with uninitialized contents, with the memory being + /// filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// use std::rc::Rc; + /// + /// let values = Rc::<[u32]>::new_zeroed_slice(3); + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [0, 0, 0]) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed_slice(len: usize) -> Rc<[mem::MaybeUninit<T>]> { + unsafe { + Rc::from_ptr(Rc::allocate_for_layout( + Layout::array::<T>(len).unwrap(), + |layout| Global.allocate_zeroed(layout), + |mem| { + ptr::slice_from_raw_parts_mut(mem as *mut T, len) + as *mut RcBox<[mem::MaybeUninit<T>]> + }, + )) + } + } +} + +impl<T> Rc<mem::MaybeUninit<T>> { + /// Converts to `Rc<T>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the inner value + /// really is in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut five = Rc::<u32>::new_uninit(); + /// + /// // Deferred initialization: + /// Rc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[inline] + pub unsafe fn assume_init(self) -> Rc<T> { + unsafe { Rc::from_inner(mem::ManuallyDrop::new(self).ptr.cast()) } + } +} + +impl<T> Rc<[mem::MaybeUninit<T>]> { + /// Converts to `Rc<[T]>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the inner value + /// really is in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut values = Rc::<[u32]>::new_uninit_slice(3); + /// + /// // Deferred initialization: + /// let data = Rc::get_mut(&mut values).unwrap(); + /// data[0].write(1); + /// data[1].write(2); + /// data[2].write(3); + /// + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[inline] + pub unsafe fn assume_init(self) -> Rc<[T]> { + unsafe { Rc::from_ptr(mem::ManuallyDrop::new(self).ptr.as_ptr() as _) } + } +} + +impl<T: ?Sized> Rc<T> { + /// Consumes the `Rc`, returning the wrapped pointer. + /// + /// To avoid a memory leak the pointer must be converted back to an `Rc` using + /// [`Rc::from_raw`]. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let x = Rc::new("hello".to_owned()); + /// let x_ptr = Rc::into_raw(x); + /// assert_eq!(unsafe { &*x_ptr }, "hello"); + /// ``` + #[stable(feature = "rc_raw", since = "1.17.0")] + pub fn into_raw(this: Self) -> *const T { + let ptr = Self::as_ptr(&this); + mem::forget(this); + ptr + } + + /// Provides a raw pointer to the data. + /// + /// The counts are not affected in any way and the `Rc` is not consumed. The pointer is valid + /// for as long there are strong counts in the `Rc`. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let x = Rc::new("hello".to_owned()); + /// let y = Rc::clone(&x); + /// let x_ptr = Rc::as_ptr(&x); + /// assert_eq!(x_ptr, Rc::as_ptr(&y)); + /// assert_eq!(unsafe { &*x_ptr }, "hello"); + /// ``` + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub fn as_ptr(this: &Self) -> *const T { + let ptr: *mut RcBox<T> = NonNull::as_ptr(this.ptr); + + // SAFETY: This cannot go through Deref::deref or Rc::inner because + // this is required to retain raw/mut provenance such that e.g. `get_mut` can + // write through the pointer after the Rc is recovered through `from_raw`. + unsafe { ptr::addr_of_mut!((*ptr).value) } + } + + /// Constructs an `Rc<T>` from a raw pointer. + /// + /// The raw pointer must have been previously returned by a call to + /// [`Rc<U>::into_raw`][into_raw] where `U` must have the same size + /// and alignment as `T`. This is trivially true if `U` is `T`. + /// Note that if `U` is not `T` but has the same size and alignment, this is + /// basically like transmuting references of different types. See + /// [`mem::transmute`] for more information on what + /// restrictions apply in this case. + /// + /// The user of `from_raw` has to make sure a specific value of `T` is only + /// dropped once. + /// + /// This function is unsafe because improper use may lead to memory unsafety, + /// even if the returned `Rc<T>` is never accessed. + /// + /// [into_raw]: Rc::into_raw + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let x = Rc::new("hello".to_owned()); + /// let x_ptr = Rc::into_raw(x); + /// + /// unsafe { + /// // Convert back to an `Rc` to prevent leak. + /// let x = Rc::from_raw(x_ptr); + /// assert_eq!(&*x, "hello"); + /// + /// // Further calls to `Rc::from_raw(x_ptr)` would be memory-unsafe. + /// } + /// + /// // The memory was freed when `x` went out of scope above, so `x_ptr` is now dangling! + /// ``` + #[stable(feature = "rc_raw", since = "1.17.0")] + pub unsafe fn from_raw(ptr: *const T) -> Self { + let offset = unsafe { data_offset(ptr) }; + + // Reverse the offset to find the original RcBox. + let rc_ptr = unsafe { ptr.byte_sub(offset) as *mut RcBox<T> }; + + unsafe { Self::from_ptr(rc_ptr) } + } + + /// Creates a new [`Weak`] pointer to this allocation. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// let weak_five = Rc::downgrade(&five); + /// ``` + #[must_use = "this returns a new `Weak` pointer, \ + without modifying the original `Rc`"] + #[stable(feature = "rc_weak", since = "1.4.0")] + pub fn downgrade(this: &Self) -> Weak<T> { + this.inner().inc_weak(); + // Make sure we do not create a dangling Weak + debug_assert!(!is_dangling(this.ptr.as_ptr())); + Weak { ptr: this.ptr } + } + + /// Gets the number of [`Weak`] pointers to this allocation. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// let _weak_five = Rc::downgrade(&five); + /// + /// assert_eq!(1, Rc::weak_count(&five)); + /// ``` + #[inline] + #[stable(feature = "rc_counts", since = "1.15.0")] + pub fn weak_count(this: &Self) -> usize { + this.inner().weak() - 1 + } + + /// Gets the number of strong (`Rc`) pointers to this allocation. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// let _also_five = Rc::clone(&five); + /// + /// assert_eq!(2, Rc::strong_count(&five)); + /// ``` + #[inline] + #[stable(feature = "rc_counts", since = "1.15.0")] + pub fn strong_count(this: &Self) -> usize { + this.inner().strong() + } + + /// Increments the strong reference count on the `Rc<T>` associated with the + /// provided pointer by one. + /// + /// # Safety + /// + /// The pointer must have been obtained through `Rc::into_raw`, and the + /// associated `Rc` instance must be valid (i.e. the strong count must be at + /// least 1) for the duration of this method. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// unsafe { + /// let ptr = Rc::into_raw(five); + /// Rc::increment_strong_count(ptr); + /// + /// let five = Rc::from_raw(ptr); + /// assert_eq!(2, Rc::strong_count(&five)); + /// } + /// ``` + #[inline] + #[stable(feature = "rc_mutate_strong_count", since = "1.53.0")] + pub unsafe fn increment_strong_count(ptr: *const T) { + // Retain Rc, but don't touch refcount by wrapping in ManuallyDrop + let rc = unsafe { mem::ManuallyDrop::new(Rc::<T>::from_raw(ptr)) }; + // Now increase refcount, but don't drop new refcount either + let _rc_clone: mem::ManuallyDrop<_> = rc.clone(); + } + + /// Decrements the strong reference count on the `Rc<T>` associated with the + /// provided pointer by one. + /// + /// # Safety + /// + /// The pointer must have been obtained through `Rc::into_raw`, and the + /// associated `Rc` instance must be valid (i.e. the strong count must be at + /// least 1) when invoking this method. This method can be used to release + /// the final `Rc` and backing storage, but **should not** be called after + /// the final `Rc` has been released. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// unsafe { + /// let ptr = Rc::into_raw(five); + /// Rc::increment_strong_count(ptr); + /// + /// let five = Rc::from_raw(ptr); + /// assert_eq!(2, Rc::strong_count(&five)); + /// Rc::decrement_strong_count(ptr); + /// assert_eq!(1, Rc::strong_count(&five)); + /// } + /// ``` + #[inline] + #[stable(feature = "rc_mutate_strong_count", since = "1.53.0")] + pub unsafe fn decrement_strong_count(ptr: *const T) { + unsafe { mem::drop(Rc::from_raw(ptr)) }; + } + + /// Returns `true` if there are no other `Rc` or [`Weak`] pointers to + /// this allocation. + #[inline] + fn is_unique(this: &Self) -> bool { + Rc::weak_count(this) == 0 && Rc::strong_count(this) == 1 + } + + /// Returns a mutable reference into the given `Rc`, if there are + /// no other `Rc` or [`Weak`] pointers to the same allocation. + /// + /// Returns [`None`] otherwise, because it is not safe to + /// mutate a shared value. + /// + /// See also [`make_mut`][make_mut], which will [`clone`][clone] + /// the inner value when there are other `Rc` pointers. + /// + /// [make_mut]: Rc::make_mut + /// [clone]: Clone::clone + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let mut x = Rc::new(3); + /// *Rc::get_mut(&mut x).unwrap() = 4; + /// assert_eq!(*x, 4); + /// + /// let _y = Rc::clone(&x); + /// assert!(Rc::get_mut(&mut x).is_none()); + /// ``` + #[inline] + #[stable(feature = "rc_unique", since = "1.4.0")] + pub fn get_mut(this: &mut Self) -> Option<&mut T> { + if Rc::is_unique(this) { unsafe { Some(Rc::get_mut_unchecked(this)) } } else { None } + } + + /// Returns a mutable reference into the given `Rc`, + /// without any check. + /// + /// See also [`get_mut`], which is safe and does appropriate checks. + /// + /// [`get_mut`]: Rc::get_mut + /// + /// # Safety + /// + /// Any other `Rc` or [`Weak`] pointers to the same allocation must not be dereferenced + /// for the duration of the returned borrow. + /// This is trivially the case if no such pointers exist, + /// for example immediately after `Rc::new`. + /// + /// # Examples + /// + /// ``` + /// #![feature(get_mut_unchecked)] + /// + /// use std::rc::Rc; + /// + /// let mut x = Rc::new(String::new()); + /// unsafe { + /// Rc::get_mut_unchecked(&mut x).push_str("foo") + /// } + /// assert_eq!(*x, "foo"); + /// ``` + #[inline] + #[unstable(feature = "get_mut_unchecked", issue = "63292")] + pub unsafe fn get_mut_unchecked(this: &mut Self) -> &mut T { + // We are careful to *not* create a reference covering the "count" fields, as + // this would conflict with accesses to the reference counts (e.g. by `Weak`). + unsafe { &mut (*this.ptr.as_ptr()).value } + } + + #[inline] + #[stable(feature = "ptr_eq", since = "1.17.0")] + /// Returns `true` if the two `Rc`s point to the same allocation + /// (in a vein similar to [`ptr::eq`]). + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// let same_five = Rc::clone(&five); + /// let other_five = Rc::new(5); + /// + /// assert!(Rc::ptr_eq(&five, &same_five)); + /// assert!(!Rc::ptr_eq(&five, &other_five)); + /// ``` + pub fn ptr_eq(this: &Self, other: &Self) -> bool { + this.ptr.as_ptr() == other.ptr.as_ptr() + } +} + +impl<T: Clone> Rc<T> { + /// Makes a mutable reference into the given `Rc`. + /// + /// If there are other `Rc` pointers to the same allocation, then `make_mut` will + /// [`clone`] the inner value to a new allocation to ensure unique ownership. This is also + /// referred to as clone-on-write. + /// + /// However, if there are no other `Rc` pointers to this allocation, but some [`Weak`] + /// pointers, then the [`Weak`] pointers will be disassociated and the inner value will not + /// be cloned. + /// + /// See also [`get_mut`], which will fail rather than cloning the inner value + /// or diassociating [`Weak`] pointers. + /// + /// [`clone`]: Clone::clone + /// [`get_mut`]: Rc::get_mut + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let mut data = Rc::new(5); + /// + /// *Rc::make_mut(&mut data) += 1; // Won't clone anything + /// let mut other_data = Rc::clone(&data); // Won't clone inner data + /// *Rc::make_mut(&mut data) += 1; // Clones inner data + /// *Rc::make_mut(&mut data) += 1; // Won't clone anything + /// *Rc::make_mut(&mut other_data) *= 2; // Won't clone anything + /// + /// // Now `data` and `other_data` point to different allocations. + /// assert_eq!(*data, 8); + /// assert_eq!(*other_data, 12); + /// ``` + /// + /// [`Weak`] pointers will be disassociated: + /// + /// ``` + /// use std::rc::Rc; + /// + /// let mut data = Rc::new(75); + /// let weak = Rc::downgrade(&data); + /// + /// assert!(75 == *data); + /// assert!(75 == *weak.upgrade().unwrap()); + /// + /// *Rc::make_mut(&mut data) += 1; + /// + /// assert!(76 == *data); + /// assert!(weak.upgrade().is_none()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rc_unique", since = "1.4.0")] + pub fn make_mut(this: &mut Self) -> &mut T { + if Rc::strong_count(this) != 1 { + // Gotta clone the data, there are other Rcs. + // Pre-allocate memory to allow writing the cloned value directly. + let mut rc = Self::new_uninit(); + unsafe { + let data = Rc::get_mut_unchecked(&mut rc); + (**this).write_clone_into_raw(data.as_mut_ptr()); + *this = rc.assume_init(); + } + } else if Rc::weak_count(this) != 0 { + // Can just steal the data, all that's left is Weaks + let mut rc = Self::new_uninit(); + unsafe { + let data = Rc::get_mut_unchecked(&mut rc); + data.as_mut_ptr().copy_from_nonoverlapping(&**this, 1); + + this.inner().dec_strong(); + // Remove implicit strong-weak ref (no need to craft a fake + // Weak here -- we know other Weaks can clean up for us) + this.inner().dec_weak(); + ptr::write(this, rc.assume_init()); + } + } + // This unsafety is ok because we're guaranteed that the pointer + // returned is the *only* pointer that will ever be returned to T. Our + // reference count is guaranteed to be 1 at this point, and we required + // the `Rc<T>` itself to be `mut`, so we're returning the only possible + // reference to the allocation. + unsafe { &mut this.ptr.as_mut().value } + } + + /// If we have the only reference to `T` then unwrap it. Otherwise, clone `T` and return the + /// clone. + /// + /// Assuming `rc_t` is of type `Rc<T>`, this function is functionally equivalent to + /// `(*rc_t).clone()`, but will avoid cloning the inner value where possible. + /// + /// # Examples + /// + /// ``` + /// #![feature(arc_unwrap_or_clone)] + /// # use std::{ptr, rc::Rc}; + /// let inner = String::from("test"); + /// let ptr = inner.as_ptr(); + /// + /// let rc = Rc::new(inner); + /// let inner = Rc::unwrap_or_clone(rc); + /// // The inner value was not cloned + /// assert!(ptr::eq(ptr, inner.as_ptr())); + /// + /// let rc = Rc::new(inner); + /// let rc2 = rc.clone(); + /// let inner = Rc::unwrap_or_clone(rc); + /// // Because there were 2 references, we had to clone the inner value. + /// assert!(!ptr::eq(ptr, inner.as_ptr())); + /// // `rc2` is the last reference, so when we unwrap it we get back + /// // the original `String`. + /// let inner = Rc::unwrap_or_clone(rc2); + /// assert!(ptr::eq(ptr, inner.as_ptr())); + /// ``` + #[inline] + #[unstable(feature = "arc_unwrap_or_clone", issue = "93610")] + pub fn unwrap_or_clone(this: Self) -> T { + Rc::try_unwrap(this).unwrap_or_else(|rc| (*rc).clone()) + } +} + +impl Rc<dyn Any> { + /// Attempt to downcast the `Rc<dyn Any>` to a concrete type. + /// + /// # Examples + /// + /// ``` + /// use std::any::Any; + /// use std::rc::Rc; + /// + /// fn print_if_string(value: Rc<dyn Any>) { + /// if let Ok(string) = value.downcast::<String>() { + /// println!("String ({}): {}", string.len(), string); + /// } + /// } + /// + /// let my_string = "Hello World".to_string(); + /// print_if_string(Rc::new(my_string)); + /// print_if_string(Rc::new(0i8)); + /// ``` + #[inline] + #[stable(feature = "rc_downcast", since = "1.29.0")] + pub fn downcast<T: Any>(self) -> Result<Rc<T>, Rc<dyn Any>> { + if (*self).is::<T>() { + unsafe { + let ptr = self.ptr.cast::<RcBox<T>>(); + forget(self); + Ok(Rc::from_inner(ptr)) + } + } else { + Err(self) + } + } + + /// Downcasts the `Rc<dyn Any>` to a concrete type. + /// + /// For a safe alternative see [`downcast`]. + /// + /// # Examples + /// + /// ``` + /// #![feature(downcast_unchecked)] + /// + /// use std::any::Any; + /// use std::rc::Rc; + /// + /// let x: Rc<dyn Any> = Rc::new(1_usize); + /// + /// unsafe { + /// assert_eq!(*x.downcast_unchecked::<usize>(), 1); + /// } + /// ``` + /// + /// # Safety + /// + /// The contained value must be of type `T`. Calling this method + /// with the incorrect type is *undefined behavior*. + /// + /// + /// [`downcast`]: Self::downcast + #[inline] + #[unstable(feature = "downcast_unchecked", issue = "90850")] + pub unsafe fn downcast_unchecked<T: Any>(self) -> Rc<T> { + unsafe { + let ptr = self.ptr.cast::<RcBox<T>>(); + mem::forget(self); + Rc::from_inner(ptr) + } + } +} + +impl<T: ?Sized> Rc<T> { + /// Allocates an `RcBox<T>` with sufficient space for + /// a possibly-unsized inner value where the value has the layout provided. + /// + /// The function `mem_to_rcbox` is called with the data pointer + /// and must return back a (potentially fat)-pointer for the `RcBox<T>`. + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_layout( + value_layout: Layout, + allocate: impl FnOnce(Layout) -> Result<NonNull<[u8]>, AllocError>, + mem_to_rcbox: impl FnOnce(*mut u8) -> *mut RcBox<T>, + ) -> *mut RcBox<T> { + // Calculate layout using the given value layout. + // Previously, layout was calculated on the expression + // `&*(ptr as *const RcBox<T>)`, but this created a misaligned + // reference (see #54908). + let layout = Layout::new::<RcBox<()>>().extend(value_layout).unwrap().0.pad_to_align(); + unsafe { + Rc::try_allocate_for_layout(value_layout, allocate, mem_to_rcbox) + .unwrap_or_else(|_| handle_alloc_error(layout)) + } + } + + /// Allocates an `RcBox<T>` with sufficient space for + /// a possibly-unsized inner value where the value has the layout provided, + /// returning an error if allocation fails. + /// + /// The function `mem_to_rcbox` is called with the data pointer + /// and must return back a (potentially fat)-pointer for the `RcBox<T>`. + #[inline] + unsafe fn try_allocate_for_layout( + value_layout: Layout, + allocate: impl FnOnce(Layout) -> Result<NonNull<[u8]>, AllocError>, + mem_to_rcbox: impl FnOnce(*mut u8) -> *mut RcBox<T>, + ) -> Result<*mut RcBox<T>, AllocError> { + // Calculate layout using the given value layout. + // Previously, layout was calculated on the expression + // `&*(ptr as *const RcBox<T>)`, but this created a misaligned + // reference (see #54908). + let layout = Layout::new::<RcBox<()>>().extend(value_layout).unwrap().0.pad_to_align(); + + // Allocate for the layout. + let ptr = allocate(layout)?; + + // Initialize the RcBox + let inner = mem_to_rcbox(ptr.as_non_null_ptr().as_ptr()); + unsafe { + debug_assert_eq!(Layout::for_value(&*inner), layout); + + ptr::write(&mut (*inner).strong, Cell::new(1)); + ptr::write(&mut (*inner).weak, Cell::new(1)); + } + + Ok(inner) + } + + /// Allocates an `RcBox<T>` with sufficient space for an unsized inner value + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_ptr(ptr: *const T) -> *mut RcBox<T> { + // Allocate for the `RcBox<T>` using the given value. + unsafe { + Self::allocate_for_layout( + Layout::for_value(&*ptr), + |layout| Global.allocate(layout), + |mem| mem.with_metadata_of(ptr as *mut RcBox<T>), + ) + } + } + + #[cfg(not(no_global_oom_handling))] + fn from_box(v: Box<T>) -> Rc<T> { + unsafe { + let (box_unique, alloc) = Box::into_unique(v); + let bptr = box_unique.as_ptr(); + + let value_size = size_of_val(&*bptr); + let ptr = Self::allocate_for_ptr(bptr); + + // Copy value as bytes + ptr::copy_nonoverlapping( + bptr as *const T as *const u8, + &mut (*ptr).value as *mut _ as *mut u8, + value_size, + ); + + // Free the allocation without dropping its contents + box_free(box_unique, alloc); + + Self::from_ptr(ptr) + } + } +} + +impl<T> Rc<[T]> { + /// Allocates an `RcBox<[T]>` with the given length. + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_slice(len: usize) -> *mut RcBox<[T]> { + unsafe { + Self::allocate_for_layout( + Layout::array::<T>(len).unwrap(), + |layout| Global.allocate(layout), + |mem| ptr::slice_from_raw_parts_mut(mem as *mut T, len) as *mut RcBox<[T]>, + ) + } + } + + /// Copy elements from slice into newly allocated Rc<\[T\]> + /// + /// Unsafe because the caller must either take ownership or bind `T: Copy` + #[cfg(not(no_global_oom_handling))] + unsafe fn copy_from_slice(v: &[T]) -> Rc<[T]> { + unsafe { + let ptr = Self::allocate_for_slice(v.len()); + ptr::copy_nonoverlapping(v.as_ptr(), &mut (*ptr).value as *mut [T] as *mut T, v.len()); + Self::from_ptr(ptr) + } + } + + /// Constructs an `Rc<[T]>` from an iterator known to be of a certain size. + /// + /// Behavior is undefined should the size be wrong. + #[cfg(not(no_global_oom_handling))] + unsafe fn from_iter_exact(iter: impl iter::Iterator<Item = T>, len: usize) -> Rc<[T]> { + // Panic guard while cloning T elements. + // In the event of a panic, elements that have been written + // into the new RcBox will be dropped, then the memory freed. + struct Guard<T> { + mem: NonNull<u8>, + elems: *mut T, + layout: Layout, + n_elems: usize, + } + + impl<T> Drop for Guard<T> { + fn drop(&mut self) { + unsafe { + let slice = from_raw_parts_mut(self.elems, self.n_elems); + ptr::drop_in_place(slice); + + Global.deallocate(self.mem, self.layout); + } + } + } + + unsafe { + let ptr = Self::allocate_for_slice(len); + + let mem = ptr as *mut _ as *mut u8; + let layout = Layout::for_value(&*ptr); + + // Pointer to first element + let elems = &mut (*ptr).value as *mut [T] as *mut T; + + let mut guard = Guard { mem: NonNull::new_unchecked(mem), elems, layout, n_elems: 0 }; + + for (i, item) in iter.enumerate() { + ptr::write(elems.add(i), item); + guard.n_elems += 1; + } + + // All clear. Forget the guard so it doesn't free the new RcBox. + forget(guard); + + Self::from_ptr(ptr) + } + } +} + +/// Specialization trait used for `From<&[T]>`. +trait RcFromSlice<T> { + fn from_slice(slice: &[T]) -> Self; +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Clone> RcFromSlice<T> for Rc<[T]> { + #[inline] + default fn from_slice(v: &[T]) -> Self { + unsafe { Self::from_iter_exact(v.iter().cloned(), v.len()) } + } +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Copy> RcFromSlice<T> for Rc<[T]> { + #[inline] + fn from_slice(v: &[T]) -> Self { + unsafe { Rc::copy_from_slice(v) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> Deref for Rc<T> { + type Target = T; + + #[inline(always)] + fn deref(&self) -> &T { + &self.inner().value + } +} + +#[unstable(feature = "receiver_trait", issue = "none")] +impl<T: ?Sized> Receiver for Rc<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T: ?Sized> Drop for Rc<T> { + /// Drops the `Rc`. + /// + /// This will decrement the strong reference count. If the strong reference + /// count reaches zero then the only other references (if any) are + /// [`Weak`], so we `drop` the inner value. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// struct Foo; + /// + /// impl Drop for Foo { + /// fn drop(&mut self) { + /// println!("dropped!"); + /// } + /// } + /// + /// let foo = Rc::new(Foo); + /// let foo2 = Rc::clone(&foo); + /// + /// drop(foo); // Doesn't print anything + /// drop(foo2); // Prints "dropped!" + /// ``` + fn drop(&mut self) { + unsafe { + self.inner().dec_strong(); + if self.inner().strong() == 0 { + // destroy the contained object + ptr::drop_in_place(Self::get_mut_unchecked(self)); + + // remove the implicit "strong weak" pointer now that we've + // destroyed the contents. + self.inner().dec_weak(); + + if self.inner().weak() == 0 { + Global.deallocate(self.ptr.cast(), Layout::for_value(self.ptr.as_ref())); + } + } + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> Clone for Rc<T> { + /// Makes a clone of the `Rc` pointer. + /// + /// This creates another pointer to the same allocation, increasing the + /// strong reference count. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// let _ = Rc::clone(&five); + /// ``` + #[inline] + fn clone(&self) -> Rc<T> { + unsafe { + self.inner().inc_strong(); + Self::from_inner(self.ptr) + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Default> Default for Rc<T> { + /// Creates a new `Rc<T>`, with the `Default` value for `T`. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let x: Rc<i32> = Default::default(); + /// assert_eq!(*x, 0); + /// ``` + #[inline] + fn default() -> Rc<T> { + Rc::new(Default::default()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +trait RcEqIdent<T: ?Sized + PartialEq> { + fn eq(&self, other: &Rc<T>) -> bool; + fn ne(&self, other: &Rc<T>) -> bool; +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialEq> RcEqIdent<T> for Rc<T> { + #[inline] + default fn eq(&self, other: &Rc<T>) -> bool { + **self == **other + } + + #[inline] + default fn ne(&self, other: &Rc<T>) -> bool { + **self != **other + } +} + +// Hack to allow specializing on `Eq` even though `Eq` has a method. +#[rustc_unsafe_specialization_marker] +pub(crate) trait MarkerEq: PartialEq<Self> {} + +impl<T: Eq> MarkerEq for T {} + +/// We're doing this specialization here, and not as a more general optimization on `&T`, because it +/// would otherwise add a cost to all equality checks on refs. We assume that `Rc`s are used to +/// store large values, that are slow to clone, but also heavy to check for equality, causing this +/// cost to pay off more easily. It's also more likely to have two `Rc` clones, that point to +/// the same value, than two `&T`s. +/// +/// We can only do this when `T: Eq` as a `PartialEq` might be deliberately irreflexive. +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + MarkerEq> RcEqIdent<T> for Rc<T> { + #[inline] + fn eq(&self, other: &Rc<T>) -> bool { + Rc::ptr_eq(self, other) || **self == **other + } + + #[inline] + fn ne(&self, other: &Rc<T>) -> bool { + !Rc::ptr_eq(self, other) && **self != **other + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialEq> PartialEq for Rc<T> { + /// Equality for two `Rc`s. + /// + /// Two `Rc`s are equal if their inner values are equal, even if they are + /// stored in different allocation. + /// + /// If `T` also implements `Eq` (implying reflexivity of equality), + /// two `Rc`s that point to the same allocation are + /// always equal. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five == Rc::new(5)); + /// ``` + #[inline] + fn eq(&self, other: &Rc<T>) -> bool { + RcEqIdent::eq(self, other) + } + + /// Inequality for two `Rc`s. + /// + /// Two `Rc`s are unequal if their inner values are unequal. + /// + /// If `T` also implements `Eq` (implying reflexivity of equality), + /// two `Rc`s that point to the same allocation are + /// never unequal. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five != Rc::new(6)); + /// ``` + #[inline] + fn ne(&self, other: &Rc<T>) -> bool { + RcEqIdent::ne(self, other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Eq> Eq for Rc<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialOrd> PartialOrd for Rc<T> { + /// Partial comparison for two `Rc`s. + /// + /// The two are compared by calling `partial_cmp()` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// use std::cmp::Ordering; + /// + /// let five = Rc::new(5); + /// + /// assert_eq!(Some(Ordering::Less), five.partial_cmp(&Rc::new(6))); + /// ``` + #[inline(always)] + fn partial_cmp(&self, other: &Rc<T>) -> Option<Ordering> { + (**self).partial_cmp(&**other) + } + + /// Less-than comparison for two `Rc`s. + /// + /// The two are compared by calling `<` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five < Rc::new(6)); + /// ``` + #[inline(always)] + fn lt(&self, other: &Rc<T>) -> bool { + **self < **other + } + + /// 'Less than or equal to' comparison for two `Rc`s. + /// + /// The two are compared by calling `<=` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five <= Rc::new(5)); + /// ``` + #[inline(always)] + fn le(&self, other: &Rc<T>) -> bool { + **self <= **other + } + + /// Greater-than comparison for two `Rc`s. + /// + /// The two are compared by calling `>` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five > Rc::new(4)); + /// ``` + #[inline(always)] + fn gt(&self, other: &Rc<T>) -> bool { + **self > **other + } + + /// 'Greater than or equal to' comparison for two `Rc`s. + /// + /// The two are compared by calling `>=` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// assert!(five >= Rc::new(5)); + /// ``` + #[inline(always)] + fn ge(&self, other: &Rc<T>) -> bool { + **self >= **other + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Ord> Ord for Rc<T> { + /// Comparison for two `Rc`s. + /// + /// The two are compared by calling `cmp()` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// use std::cmp::Ordering; + /// + /// let five = Rc::new(5); + /// + /// assert_eq!(Ordering::Less, five.cmp(&Rc::new(6))); + /// ``` + #[inline] + fn cmp(&self, other: &Rc<T>) -> Ordering { + (**self).cmp(&**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Hash> Hash for Rc<T> { + fn hash<H: Hasher>(&self, state: &mut H) { + (**self).hash(state); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + fmt::Display> fmt::Display for Rc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + fmt::Debug> fmt::Debug for Rc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> fmt::Pointer for Rc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Pointer::fmt(&(&**self as *const T), f) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_for_ptrs", since = "1.6.0")] +impl<T> From<T> for Rc<T> { + /// Converts a generic type `T` into an `Rc<T>` + /// + /// The conversion allocates on the heap and moves `t` + /// from the stack into it. + /// + /// # Example + /// ```rust + /// # use std::rc::Rc; + /// let x = 5; + /// let rc = Rc::new(5); + /// + /// assert_eq!(Rc::from(x), rc); + /// ``` + fn from(t: T) -> Self { + Rc::new(t) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T: Clone> From<&[T]> for Rc<[T]> { + /// Allocate a reference-counted slice and fill it by cloning `v`'s items. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let original: &[i32] = &[1, 2, 3]; + /// let shared: Rc<[i32]> = Rc::from(original); + /// assert_eq!(&[1, 2, 3], &shared[..]); + /// ``` + #[inline] + fn from(v: &[T]) -> Rc<[T]> { + <Self as RcFromSlice<T>>::from_slice(v) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl From<&str> for Rc<str> { + /// Allocate a reference-counted string slice and copy `v` into it. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let shared: Rc<str> = Rc::from("statue"); + /// assert_eq!("statue", &shared[..]); + /// ``` + #[inline] + fn from(v: &str) -> Rc<str> { + let rc = Rc::<[u8]>::from(v.as_bytes()); + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const str) } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl From<String> for Rc<str> { + /// Allocate a reference-counted string slice and copy `v` into it. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let original: String = "statue".to_owned(); + /// let shared: Rc<str> = Rc::from(original); + /// assert_eq!("statue", &shared[..]); + /// ``` + #[inline] + fn from(v: String) -> Rc<str> { + Rc::from(&v[..]) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T: ?Sized> From<Box<T>> for Rc<T> { + /// Move a boxed object to a new, reference counted, allocation. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let original: Box<i32> = Box::new(1); + /// let shared: Rc<i32> = Rc::from(original); + /// assert_eq!(1, *shared); + /// ``` + #[inline] + fn from(v: Box<T>) -> Rc<T> { + Rc::from_box(v) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T> From<Vec<T>> for Rc<[T]> { + /// Allocate a reference-counted slice and move `v`'s items into it. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let original: Box<Vec<i32>> = Box::new(vec![1, 2, 3]); + /// let shared: Rc<Vec<i32>> = Rc::from(original); + /// assert_eq!(vec![1, 2, 3], *shared); + /// ``` + #[inline] + fn from(mut v: Vec<T>) -> Rc<[T]> { + unsafe { + let rc = Rc::copy_from_slice(&v); + + // Allow the Vec to free its memory, but not destroy its contents + v.set_len(0); + + rc + } + } +} + +#[stable(feature = "shared_from_cow", since = "1.45.0")] +impl<'a, B> From<Cow<'a, B>> for Rc<B> +where + B: ToOwned + ?Sized, + Rc<B>: From<&'a B> + From<B::Owned>, +{ + /// Create a reference-counted pointer from + /// a clone-on-write pointer by copying its content. + /// + /// # Example + /// + /// ```rust + /// # use std::rc::Rc; + /// # use std::borrow::Cow; + /// let cow: Cow<str> = Cow::Borrowed("eggplant"); + /// let shared: Rc<str> = Rc::from(cow); + /// assert_eq!("eggplant", &shared[..]); + /// ``` + #[inline] + fn from(cow: Cow<'a, B>) -> Rc<B> { + match cow { + Cow::Borrowed(s) => Rc::from(s), + Cow::Owned(s) => Rc::from(s), + } + } +} + +#[stable(feature = "shared_from_str", since = "1.62.0")] +impl From<Rc<str>> for Rc<[u8]> { + /// Converts a reference-counted string slice into a byte slice. + /// + /// # Example + /// + /// ``` + /// # use std::rc::Rc; + /// let string: Rc<str> = Rc::from("eggplant"); + /// let bytes: Rc<[u8]> = Rc::from(string); + /// assert_eq!("eggplant".as_bytes(), bytes.as_ref()); + /// ``` + #[inline] + fn from(rc: Rc<str>) -> Self { + // SAFETY: `str` has the same layout as `[u8]`. + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const [u8]) } + } +} + +#[stable(feature = "boxed_slice_try_from", since = "1.43.0")] +impl<T, const N: usize> TryFrom<Rc<[T]>> for Rc<[T; N]> { + type Error = Rc<[T]>; + + fn try_from(boxed_slice: Rc<[T]>) -> Result<Self, Self::Error> { + if boxed_slice.len() == N { + Ok(unsafe { Rc::from_raw(Rc::into_raw(boxed_slice) as *mut [T; N]) }) + } else { + Err(boxed_slice) + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_iter", since = "1.37.0")] +impl<T> iter::FromIterator<T> for Rc<[T]> { + /// Takes each element in the `Iterator` and collects it into an `Rc<[T]>`. + /// + /// # Performance characteristics + /// + /// ## The general case + /// + /// In the general case, collecting into `Rc<[T]>` is done by first + /// collecting into a `Vec<T>`. That is, when writing the following: + /// + /// ```rust + /// # use std::rc::Rc; + /// let evens: Rc<[u8]> = (0..10).filter(|&x| x % 2 == 0).collect(); + /// # assert_eq!(&*evens, &[0, 2, 4, 6, 8]); + /// ``` + /// + /// this behaves as if we wrote: + /// + /// ```rust + /// # use std::rc::Rc; + /// let evens: Rc<[u8]> = (0..10).filter(|&x| x % 2 == 0) + /// .collect::<Vec<_>>() // The first set of allocations happens here. + /// .into(); // A second allocation for `Rc<[T]>` happens here. + /// # assert_eq!(&*evens, &[0, 2, 4, 6, 8]); + /// ``` + /// + /// This will allocate as many times as needed for constructing the `Vec<T>` + /// and then it will allocate once for turning the `Vec<T>` into the `Rc<[T]>`. + /// + /// ## Iterators of known length + /// + /// When your `Iterator` implements `TrustedLen` and is of an exact size, + /// a single allocation will be made for the `Rc<[T]>`. For example: + /// + /// ```rust + /// # use std::rc::Rc; + /// let evens: Rc<[u8]> = (0..10).collect(); // Just a single allocation happens here. + /// # assert_eq!(&*evens, &*(0..10).collect::<Vec<_>>()); + /// ``` + fn from_iter<I: iter::IntoIterator<Item = T>>(iter: I) -> Self { + ToRcSlice::to_rc_slice(iter.into_iter()) + } +} + +/// Specialization trait used for collecting into `Rc<[T]>`. +#[cfg(not(no_global_oom_handling))] +trait ToRcSlice<T>: Iterator<Item = T> + Sized { + fn to_rc_slice(self) -> Rc<[T]>; +} + +#[cfg(not(no_global_oom_handling))] +impl<T, I: Iterator<Item = T>> ToRcSlice<T> for I { + default fn to_rc_slice(self) -> Rc<[T]> { + self.collect::<Vec<T>>().into() + } +} + +#[cfg(not(no_global_oom_handling))] +impl<T, I: iter::TrustedLen<Item = T>> ToRcSlice<T> for I { + fn to_rc_slice(self) -> Rc<[T]> { + // This is the case for a `TrustedLen` iterator. + let (low, high) = self.size_hint(); + if let Some(high) = high { + debug_assert_eq!( + low, + high, + "TrustedLen iterator's size hint is not exact: {:?}", + (low, high) + ); + + unsafe { + // SAFETY: We need to ensure that the iterator has an exact length and we have. + Rc::from_iter_exact(self, low) + } + } else { + // TrustedLen contract guarantees that `upper_bound == `None` implies an iterator + // length exceeding `usize::MAX`. + // The default implementation would collect into a vec which would panic. + // Thus we panic here immediately without invoking `Vec` code. + panic!("capacity overflow"); + } + } +} + +/// `Weak` is a version of [`Rc`] that holds a non-owning reference to the +/// managed allocation. The allocation is accessed by calling [`upgrade`] on the `Weak` +/// pointer, which returns an <code>[Option]<[Rc]\<T>></code>. +/// +/// Since a `Weak` reference does not count towards ownership, it will not +/// prevent the value stored in the allocation from being dropped, and `Weak` itself makes no +/// guarantees about the value still being present. Thus it may return [`None`] +/// when [`upgrade`]d. Note however that a `Weak` reference *does* prevent the allocation +/// itself (the backing store) from being deallocated. +/// +/// A `Weak` pointer is useful for keeping a temporary reference to the allocation +/// managed by [`Rc`] without preventing its inner value from being dropped. It is also used to +/// prevent circular references between [`Rc`] pointers, since mutual owning references +/// would never allow either [`Rc`] to be dropped. For example, a tree could +/// have strong [`Rc`] pointers from parent nodes to children, and `Weak` +/// pointers from children back to their parents. +/// +/// The typical way to obtain a `Weak` pointer is to call [`Rc::downgrade`]. +/// +/// [`upgrade`]: Weak::upgrade +#[stable(feature = "rc_weak", since = "1.4.0")] +pub struct Weak<T: ?Sized> { + // This is a `NonNull` to allow optimizing the size of this type in enums, + // but it is not necessarily a valid pointer. + // `Weak::new` sets this to `usize::MAX` so that it doesn’t need + // to allocate space on the heap. That's not a value a real pointer + // will ever have because RcBox has alignment at least 2. + // This is only possible when `T: Sized`; unsized `T` never dangle. + ptr: NonNull<RcBox<T>>, +} + +#[stable(feature = "rc_weak", since = "1.4.0")] +impl<T: ?Sized> !marker::Send for Weak<T> {} +#[stable(feature = "rc_weak", since = "1.4.0")] +impl<T: ?Sized> !marker::Sync for Weak<T> {} + +#[unstable(feature = "coerce_unsized", issue = "27732")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> CoerceUnsized<Weak<U>> for Weak<T> {} + +#[unstable(feature = "dispatch_from_dyn", issue = "none")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> DispatchFromDyn<Weak<U>> for Weak<T> {} + +impl<T> Weak<T> { + /// Constructs a new `Weak<T>`, without allocating any memory. + /// Calling [`upgrade`] on the return value always gives [`None`]. + /// + /// [`upgrade`]: Weak::upgrade + /// + /// # Examples + /// + /// ``` + /// use std::rc::Weak; + /// + /// let empty: Weak<i64> = Weak::new(); + /// assert!(empty.upgrade().is_none()); + /// ``` + #[stable(feature = "downgraded_weak", since = "1.10.0")] + #[rustc_const_unstable(feature = "const_weak_new", issue = "95091", reason = "recently added")] + #[must_use] + pub const fn new() -> Weak<T> { + Weak { ptr: unsafe { NonNull::new_unchecked(ptr::invalid_mut::<RcBox<T>>(usize::MAX)) } } + } +} + +pub(crate) fn is_dangling<T: ?Sized>(ptr: *mut T) -> bool { + (ptr as *mut ()).addr() == usize::MAX +} + +/// Helper type to allow accessing the reference counts without +/// making any assertions about the data field. +struct WeakInner<'a> { + weak: &'a Cell<usize>, + strong: &'a Cell<usize>, +} + +impl<T: ?Sized> Weak<T> { + /// Returns a raw pointer to the object `T` pointed to by this `Weak<T>`. + /// + /// The pointer is valid only if there are some strong references. The pointer may be dangling, + /// unaligned or even [`null`] otherwise. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// use std::ptr; + /// + /// let strong = Rc::new("hello".to_owned()); + /// let weak = Rc::downgrade(&strong); + /// // Both point to the same object + /// assert!(ptr::eq(&*strong, weak.as_ptr())); + /// // The strong here keeps it alive, so we can still access the object. + /// assert_eq!("hello", unsafe { &*weak.as_ptr() }); + /// + /// drop(strong); + /// // But not any more. We can do weak.as_ptr(), but accessing the pointer would lead to + /// // undefined behaviour. + /// // assert_eq!("hello", unsafe { &*weak.as_ptr() }); + /// ``` + /// + /// [`null`]: ptr::null + #[must_use] + #[stable(feature = "rc_as_ptr", since = "1.45.0")] + pub fn as_ptr(&self) -> *const T { + let ptr: *mut RcBox<T> = NonNull::as_ptr(self.ptr); + + if is_dangling(ptr) { + // If the pointer is dangling, we return the sentinel directly. This cannot be + // a valid payload address, as the payload is at least as aligned as RcBox (usize). + ptr as *const T + } else { + // SAFETY: if is_dangling returns false, then the pointer is dereferenceable. + // The payload may be dropped at this point, and we have to maintain provenance, + // so use raw pointer manipulation. + unsafe { ptr::addr_of_mut!((*ptr).value) } + } + } + + /// Consumes the `Weak<T>` and turns it into a raw pointer. + /// + /// This converts the weak pointer into a raw pointer, while still preserving the ownership of + /// one weak reference (the weak count is not modified by this operation). It can be turned + /// back into the `Weak<T>` with [`from_raw`]. + /// + /// The same restrictions of accessing the target of the pointer as with + /// [`as_ptr`] apply. + /// + /// # Examples + /// + /// ``` + /// use std::rc::{Rc, Weak}; + /// + /// let strong = Rc::new("hello".to_owned()); + /// let weak = Rc::downgrade(&strong); + /// let raw = weak.into_raw(); + /// + /// assert_eq!(1, Rc::weak_count(&strong)); + /// assert_eq!("hello", unsafe { &*raw }); + /// + /// drop(unsafe { Weak::from_raw(raw) }); + /// assert_eq!(0, Rc::weak_count(&strong)); + /// ``` + /// + /// [`from_raw`]: Weak::from_raw + /// [`as_ptr`]: Weak::as_ptr + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub fn into_raw(self) -> *const T { + let result = self.as_ptr(); + mem::forget(self); + result + } + + /// Converts a raw pointer previously created by [`into_raw`] back into `Weak<T>`. + /// + /// This can be used to safely get a strong reference (by calling [`upgrade`] + /// later) or to deallocate the weak count by dropping the `Weak<T>`. + /// + /// It takes ownership of one weak reference (with the exception of pointers created by [`new`], + /// as these don't own anything; the method still works on them). + /// + /// # Safety + /// + /// The pointer must have originated from the [`into_raw`] and must still own its potential + /// weak reference. + /// + /// It is allowed for the strong count to be 0 at the time of calling this. Nevertheless, this + /// takes ownership of one weak reference currently represented as a raw pointer (the weak + /// count is not modified by this operation) and therefore it must be paired with a previous + /// call to [`into_raw`]. + /// + /// # Examples + /// + /// ``` + /// use std::rc::{Rc, Weak}; + /// + /// let strong = Rc::new("hello".to_owned()); + /// + /// let raw_1 = Rc::downgrade(&strong).into_raw(); + /// let raw_2 = Rc::downgrade(&strong).into_raw(); + /// + /// assert_eq!(2, Rc::weak_count(&strong)); + /// + /// assert_eq!("hello", &*unsafe { Weak::from_raw(raw_1) }.upgrade().unwrap()); + /// assert_eq!(1, Rc::weak_count(&strong)); + /// + /// drop(strong); + /// + /// // Decrement the last weak count. + /// assert!(unsafe { Weak::from_raw(raw_2) }.upgrade().is_none()); + /// ``` + /// + /// [`into_raw`]: Weak::into_raw + /// [`upgrade`]: Weak::upgrade + /// [`new`]: Weak::new + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub unsafe fn from_raw(ptr: *const T) -> Self { + // See Weak::as_ptr for context on how the input pointer is derived. + + let ptr = if is_dangling(ptr as *mut T) { + // This is a dangling Weak. + ptr as *mut RcBox<T> + } else { + // Otherwise, we're guaranteed the pointer came from a nondangling Weak. + // SAFETY: data_offset is safe to call, as ptr references a real (potentially dropped) T. + let offset = unsafe { data_offset(ptr) }; + // Thus, we reverse the offset to get the whole RcBox. + // SAFETY: the pointer originated from a Weak, so this offset is safe. + unsafe { ptr.byte_sub(offset) as *mut RcBox<T> } + }; + + // SAFETY: we now have recovered the original Weak pointer, so can create the Weak. + Weak { ptr: unsafe { NonNull::new_unchecked(ptr) } } + } + + /// Attempts to upgrade the `Weak` pointer to an [`Rc`], delaying + /// dropping of the inner value if successful. + /// + /// Returns [`None`] if the inner value has since been dropped. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let five = Rc::new(5); + /// + /// let weak_five = Rc::downgrade(&five); + /// + /// let strong_five: Option<Rc<_>> = weak_five.upgrade(); + /// assert!(strong_five.is_some()); + /// + /// // Destroy all strong pointers. + /// drop(strong_five); + /// drop(five); + /// + /// assert!(weak_five.upgrade().is_none()); + /// ``` + #[must_use = "this returns a new `Rc`, \ + without modifying the original weak pointer"] + #[stable(feature = "rc_weak", since = "1.4.0")] + pub fn upgrade(&self) -> Option<Rc<T>> { + let inner = self.inner()?; + + if inner.strong() == 0 { + None + } else { + unsafe { + inner.inc_strong(); + Some(Rc::from_inner(self.ptr)) + } + } + } + + /// Gets the number of strong (`Rc`) pointers pointing to this allocation. + /// + /// If `self` was created using [`Weak::new`], this will return 0. + #[must_use] + #[stable(feature = "weak_counts", since = "1.41.0")] + pub fn strong_count(&self) -> usize { + if let Some(inner) = self.inner() { inner.strong() } else { 0 } + } + + /// Gets the number of `Weak` pointers pointing to this allocation. + /// + /// If no strong pointers remain, this will return zero. + #[must_use] + #[stable(feature = "weak_counts", since = "1.41.0")] + pub fn weak_count(&self) -> usize { + self.inner() + .map(|inner| { + if inner.strong() > 0 { + inner.weak() - 1 // subtract the implicit weak ptr + } else { + 0 + } + }) + .unwrap_or(0) + } + + /// Returns `None` when the pointer is dangling and there is no allocated `RcBox`, + /// (i.e., when this `Weak` was created by `Weak::new`). + #[inline] + fn inner(&self) -> Option<WeakInner<'_>> { + if is_dangling(self.ptr.as_ptr()) { + None + } else { + // We are careful to *not* create a reference covering the "data" field, as + // the field may be mutated concurrently (for example, if the last `Rc` + // is dropped, the data field will be dropped in-place). + Some(unsafe { + let ptr = self.ptr.as_ptr(); + WeakInner { strong: &(*ptr).strong, weak: &(*ptr).weak } + }) + } + } + + /// Returns `true` if the two `Weak`s point to the same allocation (similar to + /// [`ptr::eq`]), or if both don't point to any allocation + /// (because they were created with `Weak::new()`). + /// + /// # Notes + /// + /// Since this compares pointers it means that `Weak::new()` will equal each + /// other, even though they don't point to any allocation. + /// + /// # Examples + /// + /// ``` + /// use std::rc::Rc; + /// + /// let first_rc = Rc::new(5); + /// let first = Rc::downgrade(&first_rc); + /// let second = Rc::downgrade(&first_rc); + /// + /// assert!(first.ptr_eq(&second)); + /// + /// let third_rc = Rc::new(5); + /// let third = Rc::downgrade(&third_rc); + /// + /// assert!(!first.ptr_eq(&third)); + /// ``` + /// + /// Comparing `Weak::new`. + /// + /// ``` + /// use std::rc::{Rc, Weak}; + /// + /// let first = Weak::new(); + /// let second = Weak::new(); + /// assert!(first.ptr_eq(&second)); + /// + /// let third_rc = Rc::new(()); + /// let third = Rc::downgrade(&third_rc); + /// assert!(!first.ptr_eq(&third)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "weak_ptr_eq", since = "1.39.0")] + pub fn ptr_eq(&self, other: &Self) -> bool { + self.ptr.as_ptr() == other.ptr.as_ptr() + } +} + +#[stable(feature = "rc_weak", since = "1.4.0")] +unsafe impl<#[may_dangle] T: ?Sized> Drop for Weak<T> { + /// Drops the `Weak` pointer. + /// + /// # Examples + /// + /// ``` + /// use std::rc::{Rc, Weak}; + /// + /// struct Foo; + /// + /// impl Drop for Foo { + /// fn drop(&mut self) { + /// println!("dropped!"); + /// } + /// } + /// + /// let foo = Rc::new(Foo); + /// let weak_foo = Rc::downgrade(&foo); + /// let other_weak_foo = Weak::clone(&weak_foo); + /// + /// drop(weak_foo); // Doesn't print anything + /// drop(foo); // Prints "dropped!" + /// + /// assert!(other_weak_foo.upgrade().is_none()); + /// ``` + fn drop(&mut self) { + let inner = if let Some(inner) = self.inner() { inner } else { return }; + + inner.dec_weak(); + // the weak count starts at 1, and will only go to zero if all + // the strong pointers have disappeared. + if inner.weak() == 0 { + unsafe { + Global.deallocate(self.ptr.cast(), Layout::for_value_raw(self.ptr.as_ptr())); + } + } + } +} + +#[stable(feature = "rc_weak", since = "1.4.0")] +impl<T: ?Sized> Clone for Weak<T> { + /// Makes a clone of the `Weak` pointer that points to the same allocation. + /// + /// # Examples + /// + /// ``` + /// use std::rc::{Rc, Weak}; + /// + /// let weak_five = Rc::downgrade(&Rc::new(5)); + /// + /// let _ = Weak::clone(&weak_five); + /// ``` + #[inline] + fn clone(&self) -> Weak<T> { + if let Some(inner) = self.inner() { + inner.inc_weak() + } + Weak { ptr: self.ptr } + } +} + +#[stable(feature = "rc_weak", since = "1.4.0")] +impl<T: ?Sized + fmt::Debug> fmt::Debug for Weak<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(f, "(Weak)") + } +} + +#[stable(feature = "downgraded_weak", since = "1.10.0")] +impl<T> Default for Weak<T> { + /// Constructs a new `Weak<T>`, without allocating any memory. + /// Calling [`upgrade`] on the return value always gives [`None`]. + /// + /// [`upgrade`]: Weak::upgrade + /// + /// # Examples + /// + /// ``` + /// use std::rc::Weak; + /// + /// let empty: Weak<i64> = Default::default(); + /// assert!(empty.upgrade().is_none()); + /// ``` + fn default() -> Weak<T> { + Weak::new() + } +} + +// NOTE: We checked_add here to deal with mem::forget safely. In particular +// if you mem::forget Rcs (or Weaks), the ref-count can overflow, and then +// you can free the allocation while outstanding Rcs (or Weaks) exist. +// We abort because this is such a degenerate scenario that we don't care about +// what happens -- no real program should ever experience this. +// +// This should have negligible overhead since you don't actually need to +// clone these much in Rust thanks to ownership and move-semantics. + +#[doc(hidden)] +trait RcInnerPtr { + fn weak_ref(&self) -> &Cell<usize>; + fn strong_ref(&self) -> &Cell<usize>; + + #[inline] + fn strong(&self) -> usize { + self.strong_ref().get() + } + + #[inline] + fn inc_strong(&self) { + let strong = self.strong(); + + // We insert an `assume` here to hint LLVM at an otherwise + // missed optimization. + // SAFETY: The reference count will never be zero when this is + // called. + unsafe { + core::intrinsics::assume(strong != 0); + } + + let strong = strong.wrapping_add(1); + self.strong_ref().set(strong); + + // We want to abort on overflow instead of dropping the value. + // Checking for overflow after the store instead of before + // allows for slightly better code generation. + if core::intrinsics::unlikely(strong == 0) { + abort(); + } + } + + #[inline] + fn dec_strong(&self) { + self.strong_ref().set(self.strong() - 1); + } + + #[inline] + fn weak(&self) -> usize { + self.weak_ref().get() + } + + #[inline] + fn inc_weak(&self) { + let weak = self.weak(); + + // We insert an `assume` here to hint LLVM at an otherwise + // missed optimization. + // SAFETY: The reference count will never be zero when this is + // called. + unsafe { + core::intrinsics::assume(weak != 0); + } + + let weak = weak.wrapping_add(1); + self.weak_ref().set(weak); + + // We want to abort on overflow instead of dropping the value. + // Checking for overflow after the store instead of before + // allows for slightly better code generation. + if core::intrinsics::unlikely(weak == 0) { + abort(); + } + } + + #[inline] + fn dec_weak(&self) { + self.weak_ref().set(self.weak() - 1); + } +} + +impl<T: ?Sized> RcInnerPtr for RcBox<T> { + #[inline(always)] + fn weak_ref(&self) -> &Cell<usize> { + &self.weak + } + + #[inline(always)] + fn strong_ref(&self) -> &Cell<usize> { + &self.strong + } +} + +impl<'a> RcInnerPtr for WeakInner<'a> { + #[inline(always)] + fn weak_ref(&self) -> &Cell<usize> { + self.weak + } + + #[inline(always)] + fn strong_ref(&self) -> &Cell<usize> { + self.strong + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> borrow::Borrow<T> for Rc<T> { + fn borrow(&self) -> &T { + &**self + } +} + +#[stable(since = "1.5.0", feature = "smart_ptr_as_ref")] +impl<T: ?Sized> AsRef<T> for Rc<T> { + fn as_ref(&self) -> &T { + &**self + } +} + +#[stable(feature = "pin", since = "1.33.0")] +impl<T: ?Sized> Unpin for Rc<T> {} + +/// Get the offset within an `RcBox` for the payload behind a pointer. +/// +/// # Safety +/// +/// The pointer must point to (and have valid metadata for) a previously +/// valid instance of T, but the T is allowed to be dropped. +unsafe fn data_offset<T: ?Sized>(ptr: *const T) -> usize { + // Align the unsized value to the end of the RcBox. + // Because RcBox is repr(C), it will always be the last field in memory. + // SAFETY: since the only unsized types possible are slices, trait objects, + // and extern types, the input safety requirement is currently enough to + // satisfy the requirements of align_of_val_raw; this is an implementation + // detail of the language that must not be relied upon outside of std. + unsafe { data_offset_align(align_of_val_raw(ptr)) } +} + +#[inline] +fn data_offset_align(align: usize) -> usize { + let layout = Layout::new::<RcBox<()>>(); + layout.size() + layout.padding_needed_for(align) +} diff --git a/library/alloc/src/rc/tests.rs b/library/alloc/src/rc/tests.rs new file mode 100644 index 000000000..32433cfbd --- /dev/null +++ b/library/alloc/src/rc/tests.rs @@ -0,0 +1,561 @@ +use super::*; + +use std::boxed::Box; +use std::cell::RefCell; +use std::clone::Clone; +use std::convert::{From, TryInto}; +use std::mem::drop; +use std::option::Option::{self, None, Some}; +use std::result::Result::{Err, Ok}; + +#[test] +fn test_clone() { + let x = Rc::new(RefCell::new(5)); + let y = x.clone(); + *x.borrow_mut() = 20; + assert_eq!(*y.borrow(), 20); +} + +#[test] +fn test_simple() { + let x = Rc::new(5); + assert_eq!(*x, 5); +} + +#[test] +fn test_simple_clone() { + let x = Rc::new(5); + let y = x.clone(); + assert_eq!(*x, 5); + assert_eq!(*y, 5); +} + +#[test] +fn test_destructor() { + let x: Rc<Box<_>> = Rc::new(Box::new(5)); + assert_eq!(**x, 5); +} + +#[test] +fn test_live() { + let x = Rc::new(5); + let y = Rc::downgrade(&x); + assert!(y.upgrade().is_some()); +} + +#[test] +fn test_dead() { + let x = Rc::new(5); + let y = Rc::downgrade(&x); + drop(x); + assert!(y.upgrade().is_none()); +} + +#[test] +fn weak_self_cyclic() { + struct Cycle { + x: RefCell<Option<Weak<Cycle>>>, + } + + let a = Rc::new(Cycle { x: RefCell::new(None) }); + let b = Rc::downgrade(&a.clone()); + *a.x.borrow_mut() = Some(b); + + // hopefully we don't double-free (or leak)... +} + +#[test] +fn is_unique() { + let x = Rc::new(3); + assert!(Rc::is_unique(&x)); + let y = x.clone(); + assert!(!Rc::is_unique(&x)); + drop(y); + assert!(Rc::is_unique(&x)); + let w = Rc::downgrade(&x); + assert!(!Rc::is_unique(&x)); + drop(w); + assert!(Rc::is_unique(&x)); +} + +#[test] +fn test_strong_count() { + let a = Rc::new(0); + assert!(Rc::strong_count(&a) == 1); + let w = Rc::downgrade(&a); + assert!(Rc::strong_count(&a) == 1); + let b = w.upgrade().expect("upgrade of live rc failed"); + assert!(Rc::strong_count(&b) == 2); + assert!(Rc::strong_count(&a) == 2); + drop(w); + drop(a); + assert!(Rc::strong_count(&b) == 1); + let c = b.clone(); + assert!(Rc::strong_count(&b) == 2); + assert!(Rc::strong_count(&c) == 2); +} + +#[test] +fn test_weak_count() { + let a = Rc::new(0); + assert!(Rc::strong_count(&a) == 1); + assert!(Rc::weak_count(&a) == 0); + let w = Rc::downgrade(&a); + assert!(Rc::strong_count(&a) == 1); + assert!(Rc::weak_count(&a) == 1); + drop(w); + assert!(Rc::strong_count(&a) == 1); + assert!(Rc::weak_count(&a) == 0); + let c = a.clone(); + assert!(Rc::strong_count(&a) == 2); + assert!(Rc::weak_count(&a) == 0); + drop(c); +} + +#[test] +fn weak_counts() { + assert_eq!(Weak::weak_count(&Weak::<u64>::new()), 0); + assert_eq!(Weak::strong_count(&Weak::<u64>::new()), 0); + + let a = Rc::new(0); + let w = Rc::downgrade(&a); + assert_eq!(Weak::strong_count(&w), 1); + assert_eq!(Weak::weak_count(&w), 1); + let w2 = w.clone(); + assert_eq!(Weak::strong_count(&w), 1); + assert_eq!(Weak::weak_count(&w), 2); + assert_eq!(Weak::strong_count(&w2), 1); + assert_eq!(Weak::weak_count(&w2), 2); + drop(w); + assert_eq!(Weak::strong_count(&w2), 1); + assert_eq!(Weak::weak_count(&w2), 1); + let a2 = a.clone(); + assert_eq!(Weak::strong_count(&w2), 2); + assert_eq!(Weak::weak_count(&w2), 1); + drop(a2); + drop(a); + assert_eq!(Weak::strong_count(&w2), 0); + assert_eq!(Weak::weak_count(&w2), 0); + drop(w2); +} + +#[test] +fn try_unwrap() { + let x = Rc::new(3); + assert_eq!(Rc::try_unwrap(x), Ok(3)); + let x = Rc::new(4); + let _y = x.clone(); + assert_eq!(Rc::try_unwrap(x), Err(Rc::new(4))); + let x = Rc::new(5); + let _w = Rc::downgrade(&x); + assert_eq!(Rc::try_unwrap(x), Ok(5)); +} + +#[test] +fn into_from_raw() { + let x = Rc::new(Box::new("hello")); + let y = x.clone(); + + let x_ptr = Rc::into_raw(x); + drop(y); + unsafe { + assert_eq!(**x_ptr, "hello"); + + let x = Rc::from_raw(x_ptr); + assert_eq!(**x, "hello"); + + assert_eq!(Rc::try_unwrap(x).map(|x| *x), Ok("hello")); + } +} + +#[test] +fn test_into_from_raw_unsized() { + use std::fmt::Display; + use std::string::ToString; + + let rc: Rc<str> = Rc::from("foo"); + + let ptr = Rc::into_raw(rc.clone()); + let rc2 = unsafe { Rc::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }, "foo"); + assert_eq!(rc, rc2); + + let rc: Rc<dyn Display> = Rc::new(123); + + let ptr = Rc::into_raw(rc.clone()); + let rc2 = unsafe { Rc::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }.to_string(), "123"); + assert_eq!(rc2.to_string(), "123"); +} + +#[test] +fn into_from_weak_raw() { + let x = Rc::new(Box::new("hello")); + let y = Rc::downgrade(&x); + + let y_ptr = Weak::into_raw(y); + unsafe { + assert_eq!(**y_ptr, "hello"); + + let y = Weak::from_raw(y_ptr); + let y_up = Weak::upgrade(&y).unwrap(); + assert_eq!(**y_up, "hello"); + drop(y_up); + + assert_eq!(Rc::try_unwrap(x).map(|x| *x), Ok("hello")); + } +} + +#[test] +fn test_into_from_weak_raw_unsized() { + use std::fmt::Display; + use std::string::ToString; + + let arc: Rc<str> = Rc::from("foo"); + let weak: Weak<str> = Rc::downgrade(&arc); + + let ptr = Weak::into_raw(weak.clone()); + let weak2 = unsafe { Weak::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }, "foo"); + assert!(weak.ptr_eq(&weak2)); + + let arc: Rc<dyn Display> = Rc::new(123); + let weak: Weak<dyn Display> = Rc::downgrade(&arc); + + let ptr = Weak::into_raw(weak.clone()); + let weak2 = unsafe { Weak::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }.to_string(), "123"); + assert!(weak.ptr_eq(&weak2)); +} + +#[test] +fn get_mut() { + let mut x = Rc::new(3); + *Rc::get_mut(&mut x).unwrap() = 4; + assert_eq!(*x, 4); + let y = x.clone(); + assert!(Rc::get_mut(&mut x).is_none()); + drop(y); + assert!(Rc::get_mut(&mut x).is_some()); + let _w = Rc::downgrade(&x); + assert!(Rc::get_mut(&mut x).is_none()); +} + +#[test] +fn test_cowrc_clone_make_unique() { + let mut cow0 = Rc::new(75); + let mut cow1 = cow0.clone(); + let mut cow2 = cow1.clone(); + + assert!(75 == *Rc::make_mut(&mut cow0)); + assert!(75 == *Rc::make_mut(&mut cow1)); + assert!(75 == *Rc::make_mut(&mut cow2)); + + *Rc::make_mut(&mut cow0) += 1; + *Rc::make_mut(&mut cow1) += 2; + *Rc::make_mut(&mut cow2) += 3; + + assert!(76 == *cow0); + assert!(77 == *cow1); + assert!(78 == *cow2); + + // none should point to the same backing memory + assert!(*cow0 != *cow1); + assert!(*cow0 != *cow2); + assert!(*cow1 != *cow2); +} + +#[test] +fn test_cowrc_clone_unique2() { + let mut cow0 = Rc::new(75); + let cow1 = cow0.clone(); + let cow2 = cow1.clone(); + + assert!(75 == *cow0); + assert!(75 == *cow1); + assert!(75 == *cow2); + + *Rc::make_mut(&mut cow0) += 1; + + assert!(76 == *cow0); + assert!(75 == *cow1); + assert!(75 == *cow2); + + // cow1 and cow2 should share the same contents + // cow0 should have a unique reference + assert!(*cow0 != *cow1); + assert!(*cow0 != *cow2); + assert!(*cow1 == *cow2); +} + +#[test] +fn test_cowrc_clone_weak() { + let mut cow0 = Rc::new(75); + let cow1_weak = Rc::downgrade(&cow0); + + assert!(75 == *cow0); + assert!(75 == *cow1_weak.upgrade().unwrap()); + + *Rc::make_mut(&mut cow0) += 1; + + assert!(76 == *cow0); + assert!(cow1_weak.upgrade().is_none()); +} + +#[test] +fn test_show() { + let foo = Rc::new(75); + assert_eq!(format!("{foo:?}"), "75"); +} + +#[test] +fn test_unsized() { + let foo: Rc<[i32]> = Rc::new([1, 2, 3]); + assert_eq!(foo, foo.clone()); +} + +#[test] +fn test_maybe_thin_unsized() { + // If/when custom thin DSTs exist, this test should be updated to use one + use std::ffi::{CStr, CString}; + + let x: Rc<CStr> = Rc::from(CString::new("swordfish").unwrap().into_boxed_c_str()); + assert_eq!(format!("{x:?}"), "\"swordfish\""); + let y: Weak<CStr> = Rc::downgrade(&x); + drop(x); + + // At this point, the weak points to a dropped DST + assert!(y.upgrade().is_none()); + // But we still need to be able to get the alloc layout to drop. + // CStr has no drop glue, but custom DSTs might, and need to work. + drop(y); +} + +#[test] +fn test_from_owned() { + let foo = 123; + let foo_rc = Rc::from(foo); + assert!(123 == *foo_rc); +} + +#[test] +fn test_new_weak() { + let foo: Weak<usize> = Weak::new(); + assert!(foo.upgrade().is_none()); +} + +#[test] +fn test_ptr_eq() { + let five = Rc::new(5); + let same_five = five.clone(); + let other_five = Rc::new(5); + + assert!(Rc::ptr_eq(&five, &same_five)); + assert!(!Rc::ptr_eq(&five, &other_five)); +} + +#[test] +fn test_from_str() { + let r: Rc<str> = Rc::from("foo"); + + assert_eq!(&r[..], "foo"); +} + +#[test] +fn test_copy_from_slice() { + let s: &[u32] = &[1, 2, 3]; + let r: Rc<[u32]> = Rc::from(s); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_clone_from_slice() { + #[derive(Clone, Debug, Eq, PartialEq)] + struct X(u32); + + let s: &[X] = &[X(1), X(2), X(3)]; + let r: Rc<[X]> = Rc::from(s); + + assert_eq!(&r[..], s); +} + +#[test] +#[should_panic] +fn test_clone_from_slice_panic() { + use std::string::{String, ToString}; + + struct Fail(u32, String); + + impl Clone for Fail { + fn clone(&self) -> Fail { + if self.0 == 2 { + panic!(); + } + Fail(self.0, self.1.clone()) + } + } + + let s: &[Fail] = + &[Fail(0, "foo".to_string()), Fail(1, "bar".to_string()), Fail(2, "baz".to_string())]; + + // Should panic, but not cause memory corruption + let _r: Rc<[Fail]> = Rc::from(s); +} + +#[test] +fn test_from_box() { + let b: Box<u32> = Box::new(123); + let r: Rc<u32> = Rc::from(b); + + assert_eq!(*r, 123); +} + +#[test] +fn test_from_box_str() { + use std::string::String; + + let s = String::from("foo").into_boxed_str(); + let r: Rc<str> = Rc::from(s); + + assert_eq!(&r[..], "foo"); +} + +#[test] +fn test_from_box_slice() { + let s = vec![1, 2, 3].into_boxed_slice(); + let r: Rc<[u32]> = Rc::from(s); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_from_box_trait() { + use std::fmt::Display; + use std::string::ToString; + + let b: Box<dyn Display> = Box::new(123); + let r: Rc<dyn Display> = Rc::from(b); + + assert_eq!(r.to_string(), "123"); +} + +#[test] +fn test_from_box_trait_zero_sized() { + use std::fmt::Debug; + + let b: Box<dyn Debug> = Box::new(()); + let r: Rc<dyn Debug> = Rc::from(b); + + assert_eq!(format!("{r:?}"), "()"); +} + +#[test] +fn test_from_vec() { + let v = vec![1, 2, 3]; + let r: Rc<[u32]> = Rc::from(v); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_downcast() { + use std::any::Any; + + let r1: Rc<dyn Any> = Rc::new(i32::MAX); + let r2: Rc<dyn Any> = Rc::new("abc"); + + assert!(r1.clone().downcast::<u32>().is_err()); + + let r1i32 = r1.downcast::<i32>(); + assert!(r1i32.is_ok()); + assert_eq!(r1i32.unwrap(), Rc::new(i32::MAX)); + + assert!(r2.clone().downcast::<i32>().is_err()); + + let r2str = r2.downcast::<&'static str>(); + assert!(r2str.is_ok()); + assert_eq!(r2str.unwrap(), Rc::new("abc")); +} + +#[test] +fn test_array_from_slice() { + let v = vec![1, 2, 3]; + let r: Rc<[u32]> = Rc::from(v); + + let a: Result<Rc<[u32; 3]>, _> = r.clone().try_into(); + assert!(a.is_ok()); + + let a: Result<Rc<[u32; 2]>, _> = r.clone().try_into(); + assert!(a.is_err()); +} + +#[test] +fn test_rc_cyclic_with_zero_refs() { + struct ZeroRefs { + inner: Weak<ZeroRefs>, + } + + let zero_refs = Rc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + ZeroRefs { inner: Weak::new() } + }); + + assert_eq!(Rc::strong_count(&zero_refs), 1); + assert_eq!(Rc::weak_count(&zero_refs), 0); + assert_eq!(zero_refs.inner.strong_count(), 0); + assert_eq!(zero_refs.inner.weak_count(), 0); +} + +#[test] +fn test_rc_cyclic_with_one_ref() { + struct OneRef { + inner: Weak<OneRef>, + } + + let one_ref = Rc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + OneRef { inner: inner.clone() } + }); + + assert_eq!(Rc::strong_count(&one_ref), 1); + assert_eq!(Rc::weak_count(&one_ref), 1); + + let one_ref2 = Weak::upgrade(&one_ref.inner).unwrap(); + assert!(Rc::ptr_eq(&one_ref, &one_ref2)); + + assert_eq!(one_ref.inner.strong_count(), 2); + assert_eq!(one_ref.inner.weak_count(), 1); +} + +#[test] +fn test_rc_cyclic_with_two_ref() { + struct TwoRefs { + inner: Weak<TwoRefs>, + inner1: Weak<TwoRefs>, + } + + let two_refs = Rc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + TwoRefs { inner: inner.clone(), inner1: inner.clone() } + }); + + assert_eq!(Rc::strong_count(&two_refs), 1); + assert_eq!(Rc::weak_count(&two_refs), 2); + + let two_ref3 = Weak::upgrade(&two_refs.inner).unwrap(); + assert!(Rc::ptr_eq(&two_refs, &two_ref3)); + + let two_ref2 = Weak::upgrade(&two_refs.inner1).unwrap(); + assert!(Rc::ptr_eq(&two_refs, &two_ref2)); + + assert_eq!(Rc::strong_count(&two_refs), 3); + assert_eq!(Rc::weak_count(&two_refs), 2); +} diff --git a/library/alloc/src/slice.rs b/library/alloc/src/slice.rs new file mode 100644 index 000000000..63d4d9452 --- /dev/null +++ b/library/alloc/src/slice.rs @@ -0,0 +1,1204 @@ +//! A dynamically-sized view into a contiguous sequence, `[T]`. +//! +//! *[See also the slice primitive type](slice).* +//! +//! Slices are a view into a block of memory represented as a pointer and a +//! length. +//! +//! ``` +//! // slicing a Vec +//! let vec = vec![1, 2, 3]; +//! let int_slice = &vec[..]; +//! // coercing an array to a slice +//! let str_slice: &[&str] = &["one", "two", "three"]; +//! ``` +//! +//! Slices are either mutable or shared. The shared slice type is `&[T]`, +//! while the mutable slice type is `&mut [T]`, where `T` represents the element +//! type. For example, you can mutate the block of memory that a mutable slice +//! points to: +//! +//! ``` +//! let x = &mut [1, 2, 3]; +//! x[1] = 7; +//! assert_eq!(x, &[1, 7, 3]); +//! ``` +//! +//! Here are some of the things this module contains: +//! +//! ## Structs +//! +//! There are several structs that are useful for slices, such as [`Iter`], which +//! represents iteration over a slice. +//! +//! ## Trait Implementations +//! +//! There are several implementations of common traits for slices. Some examples +//! include: +//! +//! * [`Clone`] +//! * [`Eq`], [`Ord`] - for slices whose element type are [`Eq`] or [`Ord`]. +//! * [`Hash`] - for slices whose element type is [`Hash`]. +//! +//! ## Iteration +//! +//! The slices implement `IntoIterator`. The iterator yields references to the +//! slice elements. +//! +//! ``` +//! let numbers = &[0, 1, 2]; +//! for n in numbers { +//! println!("{n} is a number!"); +//! } +//! ``` +//! +//! The mutable slice yields mutable references to the elements: +//! +//! ``` +//! let mut scores = [7, 8, 9]; +//! for score in &mut scores[..] { +//! *score += 1; +//! } +//! ``` +//! +//! This iterator yields mutable references to the slice's elements, so while +//! the element type of the slice is `i32`, the element type of the iterator is +//! `&mut i32`. +//! +//! * [`.iter`] and [`.iter_mut`] are the explicit methods to return the default +//! iterators. +//! * Further methods that return iterators are [`.split`], [`.splitn`], +//! [`.chunks`], [`.windows`] and more. +//! +//! [`Hash`]: core::hash::Hash +//! [`.iter`]: slice::iter +//! [`.iter_mut`]: slice::iter_mut +//! [`.split`]: slice::split +//! [`.splitn`]: slice::splitn +//! [`.chunks`]: slice::chunks +//! [`.windows`]: slice::windows +#![stable(feature = "rust1", since = "1.0.0")] +// Many of the usings in this module are only used in the test configuration. +// It's cleaner to just turn off the unused_imports warning than to fix them. +#![cfg_attr(test, allow(unused_imports, dead_code))] + +use core::borrow::{Borrow, BorrowMut}; +#[cfg(not(no_global_oom_handling))] +use core::cmp::Ordering::{self, Less}; +#[cfg(not(no_global_oom_handling))] +use core::mem; +#[cfg(not(no_global_oom_handling))] +use core::mem::size_of; +#[cfg(not(no_global_oom_handling))] +use core::ptr; + +use crate::alloc::Allocator; +#[cfg(not(no_global_oom_handling))] +use crate::alloc::Global; +#[cfg(not(no_global_oom_handling))] +use crate::borrow::ToOwned; +use crate::boxed::Box; +use crate::vec::Vec; + +#[unstable(feature = "slice_range", issue = "76393")] +pub use core::slice::range; +#[unstable(feature = "array_chunks", issue = "74985")] +pub use core::slice::ArrayChunks; +#[unstable(feature = "array_chunks", issue = "74985")] +pub use core::slice::ArrayChunksMut; +#[unstable(feature = "array_windows", issue = "75027")] +pub use core::slice::ArrayWindows; +#[stable(feature = "inherent_ascii_escape", since = "1.60.0")] +pub use core::slice::EscapeAscii; +#[stable(feature = "slice_get_slice", since = "1.28.0")] +pub use core::slice::SliceIndex; +#[stable(feature = "from_ref", since = "1.28.0")] +pub use core::slice::{from_mut, from_ref}; +#[unstable(feature = "slice_from_ptr_range", issue = "89792")] +pub use core::slice::{from_mut_ptr_range, from_ptr_range}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::slice::{from_raw_parts, from_raw_parts_mut}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::slice::{Chunks, Windows}; +#[stable(feature = "chunks_exact", since = "1.31.0")] +pub use core::slice::{ChunksExact, ChunksExactMut}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::slice::{ChunksMut, Split, SplitMut}; +#[unstable(feature = "slice_group_by", issue = "80552")] +pub use core::slice::{GroupBy, GroupByMut}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::slice::{Iter, IterMut}; +#[stable(feature = "rchunks", since = "1.31.0")] +pub use core::slice::{RChunks, RChunksExact, RChunksExactMut, RChunksMut}; +#[stable(feature = "slice_rsplit", since = "1.27.0")] +pub use core::slice::{RSplit, RSplitMut}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::slice::{RSplitN, RSplitNMut, SplitN, SplitNMut}; +#[stable(feature = "split_inclusive", since = "1.51.0")] +pub use core::slice::{SplitInclusive, SplitInclusiveMut}; + +//////////////////////////////////////////////////////////////////////////////// +// Basic slice extension methods +//////////////////////////////////////////////////////////////////////////////// + +// HACK(japaric) needed for the implementation of `vec!` macro during testing +// N.B., see the `hack` module in this file for more details. +#[cfg(test)] +pub use hack::into_vec; + +// HACK(japaric) needed for the implementation of `Vec::clone` during testing +// N.B., see the `hack` module in this file for more details. +#[cfg(test)] +pub use hack::to_vec; + +// HACK(japaric): With cfg(test) `impl [T]` is not available, these three +// functions are actually methods that are in `impl [T]` but not in +// `core::slice::SliceExt` - we need to supply these functions for the +// `test_permutations` test +pub(crate) mod hack { + use core::alloc::Allocator; + + use crate::boxed::Box; + use crate::vec::Vec; + + // We shouldn't add inline attribute to this since this is used in + // `vec!` macro mostly and causes perf regression. See #71204 for + // discussion and perf results. + pub fn into_vec<T, A: Allocator>(b: Box<[T], A>) -> Vec<T, A> { + unsafe { + let len = b.len(); + let (b, alloc) = Box::into_raw_with_allocator(b); + Vec::from_raw_parts_in(b as *mut T, len, len, alloc) + } + } + + #[cfg(not(no_global_oom_handling))] + #[inline] + pub fn to_vec<T: ConvertVec, A: Allocator>(s: &[T], alloc: A) -> Vec<T, A> { + T::to_vec(s, alloc) + } + + #[cfg(not(no_global_oom_handling))] + pub trait ConvertVec { + fn to_vec<A: Allocator>(s: &[Self], alloc: A) -> Vec<Self, A> + where + Self: Sized; + } + + #[cfg(not(no_global_oom_handling))] + impl<T: Clone> ConvertVec for T { + #[inline] + default fn to_vec<A: Allocator>(s: &[Self], alloc: A) -> Vec<Self, A> { + struct DropGuard<'a, T, A: Allocator> { + vec: &'a mut Vec<T, A>, + num_init: usize, + } + impl<'a, T, A: Allocator> Drop for DropGuard<'a, T, A> { + #[inline] + fn drop(&mut self) { + // SAFETY: + // items were marked initialized in the loop below + unsafe { + self.vec.set_len(self.num_init); + } + } + } + let mut vec = Vec::with_capacity_in(s.len(), alloc); + let mut guard = DropGuard { vec: &mut vec, num_init: 0 }; + let slots = guard.vec.spare_capacity_mut(); + // .take(slots.len()) is necessary for LLVM to remove bounds checks + // and has better codegen than zip. + for (i, b) in s.iter().enumerate().take(slots.len()) { + guard.num_init = i; + slots[i].write(b.clone()); + } + core::mem::forget(guard); + // SAFETY: + // the vec was allocated and initialized above to at least this length. + unsafe { + vec.set_len(s.len()); + } + vec + } + } + + #[cfg(not(no_global_oom_handling))] + impl<T: Copy> ConvertVec for T { + #[inline] + fn to_vec<A: Allocator>(s: &[Self], alloc: A) -> Vec<Self, A> { + let mut v = Vec::with_capacity_in(s.len(), alloc); + // SAFETY: + // allocated above with the capacity of `s`, and initialize to `s.len()` in + // ptr::copy_to_non_overlapping below. + unsafe { + s.as_ptr().copy_to_nonoverlapping(v.as_mut_ptr(), s.len()); + v.set_len(s.len()); + } + v + } + } +} + +#[cfg(not(test))] +impl<T> [T] { + /// Sorts the slice. + /// + /// This sort is stable (i.e., does not reorder equal elements) and *O*(*n* \* log(*n*)) worst-case. + /// + /// When applicable, unstable sorting is preferred because it is generally faster than stable + /// sorting and it doesn't allocate auxiliary memory. + /// See [`sort_unstable`](slice::sort_unstable). + /// + /// # Current implementation + /// + /// The current algorithm is an adaptive, iterative merge sort inspired by + /// [timsort](https://en.wikipedia.org/wiki/Timsort). + /// It is designed to be very fast in cases where the slice is nearly sorted, or consists of + /// two or more sorted sequences concatenated one after another. + /// + /// Also, it allocates temporary storage half the size of `self`, but for short slices a + /// non-allocating insertion sort is used instead. + /// + /// # Examples + /// + /// ``` + /// let mut v = [-5, 4, 1, -3, 2]; + /// + /// v.sort(); + /// assert!(v == [-5, -3, 1, 2, 4]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn sort(&mut self) + where + T: Ord, + { + merge_sort(self, |a, b| a.lt(b)); + } + + /// Sorts the slice with a comparator function. + /// + /// This sort is stable (i.e., does not reorder equal elements) and *O*(*n* \* log(*n*)) worst-case. + /// + /// The comparator function must define a total ordering for the elements in the slice. If + /// the ordering is not total, the order of the elements is unspecified. An order is a + /// total order if it is (for all `a`, `b` and `c`): + /// + /// * total and antisymmetric: exactly one of `a < b`, `a == b` or `a > b` is true, and + /// * transitive, `a < b` and `b < c` implies `a < c`. The same must hold for both `==` and `>`. + /// + /// For example, while [`f64`] doesn't implement [`Ord`] because `NaN != NaN`, we can use + /// `partial_cmp` as our sort function when we know the slice doesn't contain a `NaN`. + /// + /// ``` + /// let mut floats = [5f64, 4.0, 1.0, 3.0, 2.0]; + /// floats.sort_by(|a, b| a.partial_cmp(b).unwrap()); + /// assert_eq!(floats, [1.0, 2.0, 3.0, 4.0, 5.0]); + /// ``` + /// + /// When applicable, unstable sorting is preferred because it is generally faster than stable + /// sorting and it doesn't allocate auxiliary memory. + /// See [`sort_unstable_by`](slice::sort_unstable_by). + /// + /// # Current implementation + /// + /// The current algorithm is an adaptive, iterative merge sort inspired by + /// [timsort](https://en.wikipedia.org/wiki/Timsort). + /// It is designed to be very fast in cases where the slice is nearly sorted, or consists of + /// two or more sorted sequences concatenated one after another. + /// + /// Also, it allocates temporary storage half the size of `self`, but for short slices a + /// non-allocating insertion sort is used instead. + /// + /// # Examples + /// + /// ``` + /// let mut v = [5, 4, 1, 3, 2]; + /// v.sort_by(|a, b| a.cmp(b)); + /// assert!(v == [1, 2, 3, 4, 5]); + /// + /// // reverse sorting + /// v.sort_by(|a, b| b.cmp(a)); + /// assert!(v == [5, 4, 3, 2, 1]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn sort_by<F>(&mut self, mut compare: F) + where + F: FnMut(&T, &T) -> Ordering, + { + merge_sort(self, |a, b| compare(a, b) == Less); + } + + /// Sorts the slice with a key extraction function. + /// + /// This sort is stable (i.e., does not reorder equal elements) and *O*(*m* \* *n* \* log(*n*)) + /// worst-case, where the key function is *O*(*m*). + /// + /// For expensive key functions (e.g. functions that are not simple property accesses or + /// basic operations), [`sort_by_cached_key`](slice::sort_by_cached_key) is likely to be + /// significantly faster, as it does not recompute element keys. + /// + /// When applicable, unstable sorting is preferred because it is generally faster than stable + /// sorting and it doesn't allocate auxiliary memory. + /// See [`sort_unstable_by_key`](slice::sort_unstable_by_key). + /// + /// # Current implementation + /// + /// The current algorithm is an adaptive, iterative merge sort inspired by + /// [timsort](https://en.wikipedia.org/wiki/Timsort). + /// It is designed to be very fast in cases where the slice is nearly sorted, or consists of + /// two or more sorted sequences concatenated one after another. + /// + /// Also, it allocates temporary storage half the size of `self`, but for short slices a + /// non-allocating insertion sort is used instead. + /// + /// # Examples + /// + /// ``` + /// let mut v = [-5i32, 4, 1, -3, 2]; + /// + /// v.sort_by_key(|k| k.abs()); + /// assert!(v == [1, 2, -3, 4, -5]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[stable(feature = "slice_sort_by_key", since = "1.7.0")] + #[inline] + pub fn sort_by_key<K, F>(&mut self, mut f: F) + where + F: FnMut(&T) -> K, + K: Ord, + { + merge_sort(self, |a, b| f(a).lt(&f(b))); + } + + /// Sorts the slice with a key extraction function. + /// + /// During sorting, the key function is called at most once per element, by using + /// temporary storage to remember the results of key evaluation. + /// The order of calls to the key function is unspecified and may change in future versions + /// of the standard library. + /// + /// This sort is stable (i.e., does not reorder equal elements) and *O*(*m* \* *n* + *n* \* log(*n*)) + /// worst-case, where the key function is *O*(*m*). + /// + /// For simple key functions (e.g., functions that are property accesses or + /// basic operations), [`sort_by_key`](slice::sort_by_key) is likely to be + /// faster. + /// + /// # Current implementation + /// + /// The current algorithm is based on [pattern-defeating quicksort][pdqsort] by Orson Peters, + /// which combines the fast average case of randomized quicksort with the fast worst case of + /// heapsort, while achieving linear time on slices with certain patterns. It uses some + /// randomization to avoid degenerate cases, but with a fixed seed to always provide + /// deterministic behavior. + /// + /// In the worst case, the algorithm allocates temporary storage in a `Vec<(K, usize)>` the + /// length of the slice. + /// + /// # Examples + /// + /// ``` + /// let mut v = [-5i32, 4, 32, -3, 2]; + /// + /// v.sort_by_cached_key(|k| k.to_string()); + /// assert!(v == [-3, -5, 2, 32, 4]); + /// ``` + /// + /// [pdqsort]: https://github.com/orlp/pdqsort + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[stable(feature = "slice_sort_by_cached_key", since = "1.34.0")] + #[inline] + pub fn sort_by_cached_key<K, F>(&mut self, f: F) + where + F: FnMut(&T) -> K, + K: Ord, + { + // Helper macro for indexing our vector by the smallest possible type, to reduce allocation. + macro_rules! sort_by_key { + ($t:ty, $slice:ident, $f:ident) => {{ + let mut indices: Vec<_> = + $slice.iter().map($f).enumerate().map(|(i, k)| (k, i as $t)).collect(); + // The elements of `indices` are unique, as they are indexed, so any sort will be + // stable with respect to the original slice. We use `sort_unstable` here because + // it requires less memory allocation. + indices.sort_unstable(); + for i in 0..$slice.len() { + let mut index = indices[i].1; + while (index as usize) < i { + index = indices[index as usize].1; + } + indices[i].1 = index; + $slice.swap(i, index as usize); + } + }}; + } + + let sz_u8 = mem::size_of::<(K, u8)>(); + let sz_u16 = mem::size_of::<(K, u16)>(); + let sz_u32 = mem::size_of::<(K, u32)>(); + let sz_usize = mem::size_of::<(K, usize)>(); + + let len = self.len(); + if len < 2 { + return; + } + if sz_u8 < sz_u16 && len <= (u8::MAX as usize) { + return sort_by_key!(u8, self, f); + } + if sz_u16 < sz_u32 && len <= (u16::MAX as usize) { + return sort_by_key!(u16, self, f); + } + if sz_u32 < sz_usize && len <= (u32::MAX as usize) { + return sort_by_key!(u32, self, f); + } + sort_by_key!(usize, self, f) + } + + /// Copies `self` into a new `Vec`. + /// + /// # Examples + /// + /// ``` + /// let s = [10, 40, 30]; + /// let x = s.to_vec(); + /// // Here, `s` and `x` can be modified independently. + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[rustc_conversion_suggestion] + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn to_vec(&self) -> Vec<T> + where + T: Clone, + { + self.to_vec_in(Global) + } + + /// Copies `self` into a new `Vec` with an allocator. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let s = [10, 40, 30]; + /// let x = s.to_vec_in(System); + /// // Here, `s` and `x` can be modified independently. + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[inline] + #[unstable(feature = "allocator_api", issue = "32838")] + pub fn to_vec_in<A: Allocator>(&self, alloc: A) -> Vec<T, A> + where + T: Clone, + { + // N.B., see the `hack` module in this file for more details. + hack::to_vec(self, alloc) + } + + /// Converts `self` into a vector without clones or allocation. + /// + /// The resulting vector can be converted back into a box via + /// `Vec<T>`'s `into_boxed_slice` method. + /// + /// # Examples + /// + /// ``` + /// let s: Box<[i32]> = Box::new([10, 40, 30]); + /// let x = s.into_vec(); + /// // `s` cannot be used anymore because it has been converted into `x`. + /// + /// assert_eq!(x, vec![10, 40, 30]); + /// ``` + #[rustc_allow_incoherent_impl] + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn into_vec<A: Allocator>(self: Box<Self, A>) -> Vec<T, A> { + // N.B., see the `hack` module in this file for more details. + hack::into_vec(self) + } + + /// Creates a vector by repeating a slice `n` times. + /// + /// # Panics + /// + /// This function will panic if the capacity would overflow. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// assert_eq!([1, 2].repeat(3), vec![1, 2, 1, 2, 1, 2]); + /// ``` + /// + /// A panic upon overflow: + /// + /// ```should_panic + /// // this will panic at runtime + /// b"0123456789abcdef".repeat(usize::MAX); + /// ``` + #[rustc_allow_incoherent_impl] + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "repeat_generic_slice", since = "1.40.0")] + pub fn repeat(&self, n: usize) -> Vec<T> + where + T: Copy, + { + if n == 0 { + return Vec::new(); + } + + // If `n` is larger than zero, it can be split as + // `n = 2^expn + rem (2^expn > rem, expn >= 0, rem >= 0)`. + // `2^expn` is the number represented by the leftmost '1' bit of `n`, + // and `rem` is the remaining part of `n`. + + // Using `Vec` to access `set_len()`. + let capacity = self.len().checked_mul(n).expect("capacity overflow"); + let mut buf = Vec::with_capacity(capacity); + + // `2^expn` repetition is done by doubling `buf` `expn`-times. + buf.extend(self); + { + let mut m = n >> 1; + // If `m > 0`, there are remaining bits up to the leftmost '1'. + while m > 0 { + // `buf.extend(buf)`: + unsafe { + ptr::copy_nonoverlapping( + buf.as_ptr(), + (buf.as_mut_ptr() as *mut T).add(buf.len()), + buf.len(), + ); + // `buf` has capacity of `self.len() * n`. + let buf_len = buf.len(); + buf.set_len(buf_len * 2); + } + + m >>= 1; + } + } + + // `rem` (`= n - 2^expn`) repetition is done by copying + // first `rem` repetitions from `buf` itself. + let rem_len = capacity - buf.len(); // `self.len() * rem` + if rem_len > 0 { + // `buf.extend(buf[0 .. rem_len])`: + unsafe { + // This is non-overlapping since `2^expn > rem`. + ptr::copy_nonoverlapping( + buf.as_ptr(), + (buf.as_mut_ptr() as *mut T).add(buf.len()), + rem_len, + ); + // `buf.len() + rem_len` equals to `buf.capacity()` (`= self.len() * n`). + buf.set_len(capacity); + } + } + buf + } + + /// Flattens a slice of `T` into a single value `Self::Output`. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(["hello", "world"].concat(), "helloworld"); + /// assert_eq!([[1, 2], [3, 4]].concat(), [1, 2, 3, 4]); + /// ``` + #[rustc_allow_incoherent_impl] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn concat<Item: ?Sized>(&self) -> <Self as Concat<Item>>::Output + where + Self: Concat<Item>, + { + Concat::concat(self) + } + + /// Flattens a slice of `T` into a single value `Self::Output`, placing a + /// given separator between each. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(["hello", "world"].join(" "), "hello world"); + /// assert_eq!([[1, 2], [3, 4]].join(&0), [1, 2, 0, 3, 4]); + /// assert_eq!([[1, 2], [3, 4]].join(&[0, 0][..]), [1, 2, 0, 0, 3, 4]); + /// ``` + #[rustc_allow_incoherent_impl] + #[stable(feature = "rename_connect_to_join", since = "1.3.0")] + pub fn join<Separator>(&self, sep: Separator) -> <Self as Join<Separator>>::Output + where + Self: Join<Separator>, + { + Join::join(self, sep) + } + + /// Flattens a slice of `T` into a single value `Self::Output`, placing a + /// given separator between each. + /// + /// # Examples + /// + /// ``` + /// # #![allow(deprecated)] + /// assert_eq!(["hello", "world"].connect(" "), "hello world"); + /// assert_eq!([[1, 2], [3, 4]].connect(&0), [1, 2, 0, 3, 4]); + /// ``` + #[rustc_allow_incoherent_impl] + #[stable(feature = "rust1", since = "1.0.0")] + #[deprecated(since = "1.3.0", note = "renamed to join")] + pub fn connect<Separator>(&self, sep: Separator) -> <Self as Join<Separator>>::Output + where + Self: Join<Separator>, + { + Join::join(self, sep) + } +} + +#[cfg(not(test))] +impl [u8] { + /// Returns a vector containing a copy of this slice where each byte + /// is mapped to its ASCII upper case equivalent. + /// + /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z', + /// but non-ASCII letters are unchanged. + /// + /// To uppercase the value in-place, use [`make_ascii_uppercase`]. + /// + /// [`make_ascii_uppercase`]: slice::make_ascii_uppercase + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the uppercase bytes as a new Vec, \ + without modifying the original"] + #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")] + #[inline] + pub fn to_ascii_uppercase(&self) -> Vec<u8> { + let mut me = self.to_vec(); + me.make_ascii_uppercase(); + me + } + + /// Returns a vector containing a copy of this slice where each byte + /// is mapped to its ASCII lower case equivalent. + /// + /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z', + /// but non-ASCII letters are unchanged. + /// + /// To lowercase the value in-place, use [`make_ascii_lowercase`]. + /// + /// [`make_ascii_lowercase`]: slice::make_ascii_lowercase + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the lowercase bytes as a new Vec, \ + without modifying the original"] + #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")] + #[inline] + pub fn to_ascii_lowercase(&self) -> Vec<u8> { + let mut me = self.to_vec(); + me.make_ascii_lowercase(); + me + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Extension traits for slices over specific kinds of data +//////////////////////////////////////////////////////////////////////////////// + +/// Helper trait for [`[T]::concat`](slice::concat). +/// +/// Note: the `Item` type parameter is not used in this trait, +/// but it allows impls to be more generic. +/// Without it, we get this error: +/// +/// ```error +/// error[E0207]: the type parameter `T` is not constrained by the impl trait, self type, or predica +/// --> src/liballoc/slice.rs:608:6 +/// | +/// 608 | impl<T: Clone, V: Borrow<[T]>> Concat for [V] { +/// | ^ unconstrained type parameter +/// ``` +/// +/// This is because there could exist `V` types with multiple `Borrow<[_]>` impls, +/// such that multiple `T` types would apply: +/// +/// ``` +/// # #[allow(dead_code)] +/// pub struct Foo(Vec<u32>, Vec<String>); +/// +/// impl std::borrow::Borrow<[u32]> for Foo { +/// fn borrow(&self) -> &[u32] { &self.0 } +/// } +/// +/// impl std::borrow::Borrow<[String]> for Foo { +/// fn borrow(&self) -> &[String] { &self.1 } +/// } +/// ``` +#[unstable(feature = "slice_concat_trait", issue = "27747")] +pub trait Concat<Item: ?Sized> { + #[unstable(feature = "slice_concat_trait", issue = "27747")] + /// The resulting type after concatenation + type Output; + + /// Implementation of [`[T]::concat`](slice::concat) + #[unstable(feature = "slice_concat_trait", issue = "27747")] + fn concat(slice: &Self) -> Self::Output; +} + +/// Helper trait for [`[T]::join`](slice::join) +#[unstable(feature = "slice_concat_trait", issue = "27747")] +pub trait Join<Separator> { + #[unstable(feature = "slice_concat_trait", issue = "27747")] + /// The resulting type after concatenation + type Output; + + /// Implementation of [`[T]::join`](slice::join) + #[unstable(feature = "slice_concat_trait", issue = "27747")] + fn join(slice: &Self, sep: Separator) -> Self::Output; +} + +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "slice_concat_ext", issue = "27747")] +impl<T: Clone, V: Borrow<[T]>> Concat<T> for [V] { + type Output = Vec<T>; + + fn concat(slice: &Self) -> Vec<T> { + let size = slice.iter().map(|slice| slice.borrow().len()).sum(); + let mut result = Vec::with_capacity(size); + for v in slice { + result.extend_from_slice(v.borrow()) + } + result + } +} + +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "slice_concat_ext", issue = "27747")] +impl<T: Clone, V: Borrow<[T]>> Join<&T> for [V] { + type Output = Vec<T>; + + fn join(slice: &Self, sep: &T) -> Vec<T> { + let mut iter = slice.iter(); + let first = match iter.next() { + Some(first) => first, + None => return vec![], + }; + let size = slice.iter().map(|v| v.borrow().len()).sum::<usize>() + slice.len() - 1; + let mut result = Vec::with_capacity(size); + result.extend_from_slice(first.borrow()); + + for v in iter { + result.push(sep.clone()); + result.extend_from_slice(v.borrow()) + } + result + } +} + +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "slice_concat_ext", issue = "27747")] +impl<T: Clone, V: Borrow<[T]>> Join<&[T]> for [V] { + type Output = Vec<T>; + + fn join(slice: &Self, sep: &[T]) -> Vec<T> { + let mut iter = slice.iter(); + let first = match iter.next() { + Some(first) => first, + None => return vec![], + }; + let size = + slice.iter().map(|v| v.borrow().len()).sum::<usize>() + sep.len() * (slice.len() - 1); + let mut result = Vec::with_capacity(size); + result.extend_from_slice(first.borrow()); + + for v in iter { + result.extend_from_slice(sep); + result.extend_from_slice(v.borrow()) + } + result + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Standard trait implementations for slices +//////////////////////////////////////////////////////////////////////////////// + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Borrow<[T]> for Vec<T, A> { + fn borrow(&self) -> &[T] { + &self[..] + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> BorrowMut<[T]> for Vec<T, A> { + fn borrow_mut(&mut self) -> &mut [T] { + &mut self[..] + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone> ToOwned for [T] { + type Owned = Vec<T>; + #[cfg(not(test))] + fn to_owned(&self) -> Vec<T> { + self.to_vec() + } + + #[cfg(test)] + fn to_owned(&self) -> Vec<T> { + hack::to_vec(self, Global) + } + + fn clone_into(&self, target: &mut Vec<T>) { + // drop anything in target that will not be overwritten + target.truncate(self.len()); + + // target.len <= self.len due to the truncate above, so the + // slices here are always in-bounds. + let (init, tail) = self.split_at(target.len()); + + // reuse the contained values' allocations/resources. + target.clone_from_slice(init); + target.extend_from_slice(tail); + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Sorting +//////////////////////////////////////////////////////////////////////////////// + +/// Inserts `v[0]` into pre-sorted sequence `v[1..]` so that whole `v[..]` becomes sorted. +/// +/// This is the integral subroutine of insertion sort. +#[cfg(not(no_global_oom_handling))] +fn insert_head<T, F>(v: &mut [T], is_less: &mut F) +where + F: FnMut(&T, &T) -> bool, +{ + if v.len() >= 2 && is_less(&v[1], &v[0]) { + unsafe { + // There are three ways to implement insertion here: + // + // 1. Swap adjacent elements until the first one gets to its final destination. + // However, this way we copy data around more than is necessary. If elements are big + // structures (costly to copy), this method will be slow. + // + // 2. Iterate until the right place for the first element is found. Then shift the + // elements succeeding it to make room for it and finally place it into the + // remaining hole. This is a good method. + // + // 3. Copy the first element into a temporary variable. Iterate until the right place + // for it is found. As we go along, copy every traversed element into the slot + // preceding it. Finally, copy data from the temporary variable into the remaining + // hole. This method is very good. Benchmarks demonstrated slightly better + // performance than with the 2nd method. + // + // All methods were benchmarked, and the 3rd showed best results. So we chose that one. + let tmp = mem::ManuallyDrop::new(ptr::read(&v[0])); + + // Intermediate state of the insertion process is always tracked by `hole`, which + // serves two purposes: + // 1. Protects integrity of `v` from panics in `is_less`. + // 2. Fills the remaining hole in `v` in the end. + // + // Panic safety: + // + // If `is_less` panics at any point during the process, `hole` will get dropped and + // fill the hole in `v` with `tmp`, thus ensuring that `v` still holds every object it + // initially held exactly once. + let mut hole = InsertionHole { src: &*tmp, dest: &mut v[1] }; + ptr::copy_nonoverlapping(&v[1], &mut v[0], 1); + + for i in 2..v.len() { + if !is_less(&v[i], &*tmp) { + break; + } + ptr::copy_nonoverlapping(&v[i], &mut v[i - 1], 1); + hole.dest = &mut v[i]; + } + // `hole` gets dropped and thus copies `tmp` into the remaining hole in `v`. + } + } + + // When dropped, copies from `src` into `dest`. + struct InsertionHole<T> { + src: *const T, + dest: *mut T, + } + + impl<T> Drop for InsertionHole<T> { + fn drop(&mut self) { + unsafe { + ptr::copy_nonoverlapping(self.src, self.dest, 1); + } + } + } +} + +/// Merges non-decreasing runs `v[..mid]` and `v[mid..]` using `buf` as temporary storage, and +/// stores the result into `v[..]`. +/// +/// # Safety +/// +/// The two slices must be non-empty and `mid` must be in bounds. Buffer `buf` must be long enough +/// to hold a copy of the shorter slice. Also, `T` must not be a zero-sized type. +#[cfg(not(no_global_oom_handling))] +unsafe fn merge<T, F>(v: &mut [T], mid: usize, buf: *mut T, is_less: &mut F) +where + F: FnMut(&T, &T) -> bool, +{ + let len = v.len(); + let v = v.as_mut_ptr(); + let (v_mid, v_end) = unsafe { (v.add(mid), v.add(len)) }; + + // The merge process first copies the shorter run into `buf`. Then it traces the newly copied + // run and the longer run forwards (or backwards), comparing their next unconsumed elements and + // copying the lesser (or greater) one into `v`. + // + // As soon as the shorter run is fully consumed, the process is done. If the longer run gets + // consumed first, then we must copy whatever is left of the shorter run into the remaining + // hole in `v`. + // + // Intermediate state of the process is always tracked by `hole`, which serves two purposes: + // 1. Protects integrity of `v` from panics in `is_less`. + // 2. Fills the remaining hole in `v` if the longer run gets consumed first. + // + // Panic safety: + // + // If `is_less` panics at any point during the process, `hole` will get dropped and fill the + // hole in `v` with the unconsumed range in `buf`, thus ensuring that `v` still holds every + // object it initially held exactly once. + let mut hole; + + if mid <= len - mid { + // The left run is shorter. + unsafe { + ptr::copy_nonoverlapping(v, buf, mid); + hole = MergeHole { start: buf, end: buf.add(mid), dest: v }; + } + + // Initially, these pointers point to the beginnings of their arrays. + let left = &mut hole.start; + let mut right = v_mid; + let out = &mut hole.dest; + + while *left < hole.end && right < v_end { + // Consume the lesser side. + // If equal, prefer the left run to maintain stability. + unsafe { + let to_copy = if is_less(&*right, &**left) { + get_and_increment(&mut right) + } else { + get_and_increment(left) + }; + ptr::copy_nonoverlapping(to_copy, get_and_increment(out), 1); + } + } + } else { + // The right run is shorter. + unsafe { + ptr::copy_nonoverlapping(v_mid, buf, len - mid); + hole = MergeHole { start: buf, end: buf.add(len - mid), dest: v_mid }; + } + + // Initially, these pointers point past the ends of their arrays. + let left = &mut hole.dest; + let right = &mut hole.end; + let mut out = v_end; + + while v < *left && buf < *right { + // Consume the greater side. + // If equal, prefer the right run to maintain stability. + unsafe { + let to_copy = if is_less(&*right.offset(-1), &*left.offset(-1)) { + decrement_and_get(left) + } else { + decrement_and_get(right) + }; + ptr::copy_nonoverlapping(to_copy, decrement_and_get(&mut out), 1); + } + } + } + // Finally, `hole` gets dropped. If the shorter run was not fully consumed, whatever remains of + // it will now be copied into the hole in `v`. + + unsafe fn get_and_increment<T>(ptr: &mut *mut T) -> *mut T { + let old = *ptr; + *ptr = unsafe { ptr.offset(1) }; + old + } + + unsafe fn decrement_and_get<T>(ptr: &mut *mut T) -> *mut T { + *ptr = unsafe { ptr.offset(-1) }; + *ptr + } + + // When dropped, copies the range `start..end` into `dest..`. + struct MergeHole<T> { + start: *mut T, + end: *mut T, + dest: *mut T, + } + + impl<T> Drop for MergeHole<T> { + fn drop(&mut self) { + // `T` is not a zero-sized type, and these are pointers into a slice's elements. + unsafe { + let len = self.end.sub_ptr(self.start); + ptr::copy_nonoverlapping(self.start, self.dest, len); + } + } + } +} + +/// This merge sort borrows some (but not all) ideas from TimSort, which is described in detail +/// [here](https://github.com/python/cpython/blob/main/Objects/listsort.txt). +/// +/// The algorithm identifies strictly descending and non-descending subsequences, which are called +/// natural runs. There is a stack of pending runs yet to be merged. Each newly found run is pushed +/// onto the stack, and then some pairs of adjacent runs are merged until these two invariants are +/// satisfied: +/// +/// 1. for every `i` in `1..runs.len()`: `runs[i - 1].len > runs[i].len` +/// 2. for every `i` in `2..runs.len()`: `runs[i - 2].len > runs[i - 1].len + runs[i].len` +/// +/// The invariants ensure that the total running time is *O*(*n* \* log(*n*)) worst-case. +#[cfg(not(no_global_oom_handling))] +fn merge_sort<T, F>(v: &mut [T], mut is_less: F) +where + F: FnMut(&T, &T) -> bool, +{ + // Slices of up to this length get sorted using insertion sort. + const MAX_INSERTION: usize = 20; + // Very short runs are extended using insertion sort to span at least this many elements. + const MIN_RUN: usize = 10; + + // Sorting has no meaningful behavior on zero-sized types. + if size_of::<T>() == 0 { + return; + } + + let len = v.len(); + + // Short arrays get sorted in-place via insertion sort to avoid allocations. + if len <= MAX_INSERTION { + if len >= 2 { + for i in (0..len - 1).rev() { + insert_head(&mut v[i..], &mut is_less); + } + } + return; + } + + // Allocate a buffer to use as scratch memory. We keep the length 0 so we can keep in it + // shallow copies of the contents of `v` without risking the dtors running on copies if + // `is_less` panics. When merging two sorted runs, this buffer holds a copy of the shorter run, + // which will always have length at most `len / 2`. + let mut buf = Vec::with_capacity(len / 2); + + // In order to identify natural runs in `v`, we traverse it backwards. That might seem like a + // strange decision, but consider the fact that merges more often go in the opposite direction + // (forwards). According to benchmarks, merging forwards is slightly faster than merging + // backwards. To conclude, identifying runs by traversing backwards improves performance. + let mut runs = vec![]; + let mut end = len; + while end > 0 { + // Find the next natural run, and reverse it if it's strictly descending. + let mut start = end - 1; + if start > 0 { + start -= 1; + unsafe { + if is_less(v.get_unchecked(start + 1), v.get_unchecked(start)) { + while start > 0 && is_less(v.get_unchecked(start), v.get_unchecked(start - 1)) { + start -= 1; + } + v[start..end].reverse(); + } else { + while start > 0 && !is_less(v.get_unchecked(start), v.get_unchecked(start - 1)) + { + start -= 1; + } + } + } + } + + // Insert some more elements into the run if it's too short. Insertion sort is faster than + // merge sort on short sequences, so this significantly improves performance. + while start > 0 && end - start < MIN_RUN { + start -= 1; + insert_head(&mut v[start..end], &mut is_less); + } + + // Push this run onto the stack. + runs.push(Run { start, len: end - start }); + end = start; + + // Merge some pairs of adjacent runs to satisfy the invariants. + while let Some(r) = collapse(&runs) { + let left = runs[r + 1]; + let right = runs[r]; + unsafe { + merge( + &mut v[left.start..right.start + right.len], + left.len, + buf.as_mut_ptr(), + &mut is_less, + ); + } + runs[r] = Run { start: left.start, len: left.len + right.len }; + runs.remove(r + 1); + } + } + + // Finally, exactly one run must remain in the stack. + debug_assert!(runs.len() == 1 && runs[0].start == 0 && runs[0].len == len); + + // Examines the stack of runs and identifies the next pair of runs to merge. More specifically, + // if `Some(r)` is returned, that means `runs[r]` and `runs[r + 1]` must be merged next. If the + // algorithm should continue building a new run instead, `None` is returned. + // + // TimSort is infamous for its buggy implementations, as described here: + // http://envisage-project.eu/timsort-specification-and-verification/ + // + // The gist of the story is: we must enforce the invariants on the top four runs on the stack. + // Enforcing them on just top three is not sufficient to ensure that the invariants will still + // hold for *all* runs in the stack. + // + // This function correctly checks invariants for the top four runs. Additionally, if the top + // run starts at index 0, it will always demand a merge operation until the stack is fully + // collapsed, in order to complete the sort. + #[inline] + fn collapse(runs: &[Run]) -> Option<usize> { + let n = runs.len(); + if n >= 2 + && (runs[n - 1].start == 0 + || runs[n - 2].len <= runs[n - 1].len + || (n >= 3 && runs[n - 3].len <= runs[n - 2].len + runs[n - 1].len) + || (n >= 4 && runs[n - 4].len <= runs[n - 3].len + runs[n - 2].len)) + { + if n >= 3 && runs[n - 3].len < runs[n - 1].len { Some(n - 3) } else { Some(n - 2) } + } else { + None + } + } + + #[derive(Clone, Copy)] + struct Run { + start: usize, + len: usize, + } +} diff --git a/library/alloc/src/str.rs b/library/alloc/src/str.rs new file mode 100644 index 000000000..d5ed2c4ad --- /dev/null +++ b/library/alloc/src/str.rs @@ -0,0 +1,686 @@ +//! Unicode string slices. +//! +//! *[See also the `str` primitive type](str).* +//! +//! The `&str` type is one of the two main string types, the other being `String`. +//! Unlike its `String` counterpart, its contents are borrowed. +//! +//! # Basic Usage +//! +//! A basic string declaration of `&str` type: +//! +//! ``` +//! let hello_world = "Hello, World!"; +//! ``` +//! +//! Here we have declared a string literal, also known as a string slice. +//! String literals have a static lifetime, which means the string `hello_world` +//! is guaranteed to be valid for the duration of the entire program. +//! We can explicitly specify `hello_world`'s lifetime as well: +//! +//! ``` +//! let hello_world: &'static str = "Hello, world!"; +//! ``` + +#![stable(feature = "rust1", since = "1.0.0")] +// Many of the usings in this module are only used in the test configuration. +// It's cleaner to just turn off the unused_imports warning than to fix them. +#![allow(unused_imports)] + +use core::borrow::{Borrow, BorrowMut}; +use core::iter::FusedIterator; +use core::mem; +use core::ptr; +use core::str::pattern::{DoubleEndedSearcher, Pattern, ReverseSearcher, Searcher}; +use core::unicode::conversions; + +use crate::borrow::ToOwned; +use crate::boxed::Box; +use crate::slice::{Concat, Join, SliceIndex}; +use crate::string::String; +use crate::vec::Vec; + +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::pattern; +#[stable(feature = "encode_utf16", since = "1.8.0")] +pub use core::str::EncodeUtf16; +#[stable(feature = "split_ascii_whitespace", since = "1.34.0")] +pub use core::str::SplitAsciiWhitespace; +#[stable(feature = "split_inclusive", since = "1.51.0")] +pub use core::str::SplitInclusive; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::SplitWhitespace; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{from_utf8, from_utf8_mut, Bytes, CharIndices, Chars}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{from_utf8_unchecked, from_utf8_unchecked_mut, ParseBoolError}; +#[stable(feature = "str_escape", since = "1.34.0")] +pub use core::str::{EscapeDebug, EscapeDefault, EscapeUnicode}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{FromStr, Utf8Error}; +#[allow(deprecated)] +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{Lines, LinesAny}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{MatchIndices, RMatchIndices}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{Matches, RMatches}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{RSplit, Split}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{RSplitN, SplitN}; +#[stable(feature = "rust1", since = "1.0.0")] +pub use core::str::{RSplitTerminator, SplitTerminator}; + +/// Note: `str` in `Concat<str>` is not meaningful here. +/// This type parameter of the trait only exists to enable another impl. +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "slice_concat_ext", issue = "27747")] +impl<S: Borrow<str>> Concat<str> for [S] { + type Output = String; + + fn concat(slice: &Self) -> String { + Join::join(slice, "") + } +} + +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "slice_concat_ext", issue = "27747")] +impl<S: Borrow<str>> Join<&str> for [S] { + type Output = String; + + fn join(slice: &Self, sep: &str) -> String { + unsafe { String::from_utf8_unchecked(join_generic_copy(slice, sep.as_bytes())) } + } +} + +#[cfg(not(no_global_oom_handling))] +macro_rules! specialize_for_lengths { + ($separator:expr, $target:expr, $iter:expr; $($num:expr),*) => {{ + let mut target = $target; + let iter = $iter; + let sep_bytes = $separator; + match $separator.len() { + $( + // loops with hardcoded sizes run much faster + // specialize the cases with small separator lengths + $num => { + for s in iter { + copy_slice_and_advance!(target, sep_bytes); + let content_bytes = s.borrow().as_ref(); + copy_slice_and_advance!(target, content_bytes); + } + }, + )* + _ => { + // arbitrary non-zero size fallback + for s in iter { + copy_slice_and_advance!(target, sep_bytes); + let content_bytes = s.borrow().as_ref(); + copy_slice_and_advance!(target, content_bytes); + } + } + } + target + }} +} + +#[cfg(not(no_global_oom_handling))] +macro_rules! copy_slice_and_advance { + ($target:expr, $bytes:expr) => { + let len = $bytes.len(); + let (head, tail) = { $target }.split_at_mut(len); + head.copy_from_slice($bytes); + $target = tail; + }; +} + +// Optimized join implementation that works for both Vec<T> (T: Copy) and String's inner vec +// Currently (2018-05-13) there is a bug with type inference and specialization (see issue #36262) +// For this reason SliceConcat<T> is not specialized for T: Copy and SliceConcat<str> is the +// only user of this function. It is left in place for the time when that is fixed. +// +// the bounds for String-join are S: Borrow<str> and for Vec-join Borrow<[T]> +// [T] and str both impl AsRef<[T]> for some T +// => s.borrow().as_ref() and we always have slices +#[cfg(not(no_global_oom_handling))] +fn join_generic_copy<B, T, S>(slice: &[S], sep: &[T]) -> Vec<T> +where + T: Copy, + B: AsRef<[T]> + ?Sized, + S: Borrow<B>, +{ + let sep_len = sep.len(); + let mut iter = slice.iter(); + + // the first slice is the only one without a separator preceding it + let first = match iter.next() { + Some(first) => first, + None => return vec![], + }; + + // compute the exact total length of the joined Vec + // if the `len` calculation overflows, we'll panic + // we would have run out of memory anyway and the rest of the function requires + // the entire Vec pre-allocated for safety + let reserved_len = sep_len + .checked_mul(iter.len()) + .and_then(|n| { + slice.iter().map(|s| s.borrow().as_ref().len()).try_fold(n, usize::checked_add) + }) + .expect("attempt to join into collection with len > usize::MAX"); + + // prepare an uninitialized buffer + let mut result = Vec::with_capacity(reserved_len); + debug_assert!(result.capacity() >= reserved_len); + + result.extend_from_slice(first.borrow().as_ref()); + + unsafe { + let pos = result.len(); + let target = result.spare_capacity_mut().get_unchecked_mut(..reserved_len - pos); + + // Convert the separator and slices to slices of MaybeUninit + // to simplify implementation in specialize_for_lengths + let sep_uninit = core::slice::from_raw_parts(sep.as_ptr().cast(), sep.len()); + let iter_uninit = iter.map(|it| { + let it = it.borrow().as_ref(); + core::slice::from_raw_parts(it.as_ptr().cast(), it.len()) + }); + + // copy separator and slices over without bounds checks + // generate loops with hardcoded offsets for small separators + // massive improvements possible (~ x2) + let remain = specialize_for_lengths!(sep_uninit, target, iter_uninit; 0, 1, 2, 3, 4); + + // A weird borrow implementation may return different + // slices for the length calculation and the actual copy. + // Make sure we don't expose uninitialized bytes to the caller. + let result_len = reserved_len - remain.len(); + result.set_len(result_len); + } + result +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl Borrow<str> for String { + #[inline] + fn borrow(&self) -> &str { + &self[..] + } +} + +#[stable(feature = "string_borrow_mut", since = "1.36.0")] +impl BorrowMut<str> for String { + #[inline] + fn borrow_mut(&mut self) -> &mut str { + &mut self[..] + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl ToOwned for str { + type Owned = String; + #[inline] + fn to_owned(&self) -> String { + unsafe { String::from_utf8_unchecked(self.as_bytes().to_owned()) } + } + + fn clone_into(&self, target: &mut String) { + let mut b = mem::take(target).into_bytes(); + self.as_bytes().clone_into(&mut b); + *target = unsafe { String::from_utf8_unchecked(b) } + } +} + +/// Methods for string slices. +#[cfg(not(test))] +impl str { + /// Converts a `Box<str>` into a `Box<[u8]>` without copying or allocating. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = "this is a string"; + /// let boxed_str = s.to_owned().into_boxed_str(); + /// let boxed_bytes = boxed_str.into_boxed_bytes(); + /// assert_eq!(*boxed_bytes, *s.as_bytes()); + /// ``` + #[rustc_allow_incoherent_impl] + #[stable(feature = "str_box_extras", since = "1.20.0")] + #[must_use = "`self` will be dropped if the result is not used"] + #[inline] + pub fn into_boxed_bytes(self: Box<str>) -> Box<[u8]> { + self.into() + } + + /// Replaces all matches of a pattern with another string. + /// + /// `replace` creates a new [`String`], and copies the data from this string slice into it. + /// While doing so, it attempts to find matches of a pattern. If it finds any, it + /// replaces them with the replacement string slice. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = "this is old"; + /// + /// assert_eq!("this is new", s.replace("old", "new")); + /// assert_eq!("than an old", s.replace("is", "an")); + /// ``` + /// + /// When the pattern doesn't match: + /// + /// ``` + /// let s = "this is old"; + /// assert_eq!(s, s.replace("cookie monster", "little lamb")); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the replaced string as a new allocation, \ + without modifying the original"] + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn replace<'a, P: Pattern<'a>>(&'a self, from: P, to: &str) -> String { + let mut result = String::new(); + let mut last_end = 0; + for (start, part) in self.match_indices(from) { + result.push_str(unsafe { self.get_unchecked(last_end..start) }); + result.push_str(to); + last_end = start + part.len(); + } + result.push_str(unsafe { self.get_unchecked(last_end..self.len()) }); + result + } + + /// Replaces first N matches of a pattern with another string. + /// + /// `replacen` creates a new [`String`], and copies the data from this string slice into it. + /// While doing so, it attempts to find matches of a pattern. If it finds any, it + /// replaces them with the replacement string slice at most `count` times. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = "foo foo 123 foo"; + /// assert_eq!("new new 123 foo", s.replacen("foo", "new", 2)); + /// assert_eq!("faa fao 123 foo", s.replacen('o', "a", 3)); + /// assert_eq!("foo foo new23 foo", s.replacen(char::is_numeric, "new", 1)); + /// ``` + /// + /// When the pattern doesn't match: + /// + /// ``` + /// let s = "this is old"; + /// assert_eq!(s, s.replacen("cookie monster", "little lamb", 10)); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the replaced string as a new allocation, \ + without modifying the original"] + #[stable(feature = "str_replacen", since = "1.16.0")] + pub fn replacen<'a, P: Pattern<'a>>(&'a self, pat: P, to: &str, count: usize) -> String { + // Hope to reduce the times of re-allocation + let mut result = String::with_capacity(32); + let mut last_end = 0; + for (start, part) in self.match_indices(pat).take(count) { + result.push_str(unsafe { self.get_unchecked(last_end..start) }); + result.push_str(to); + last_end = start + part.len(); + } + result.push_str(unsafe { self.get_unchecked(last_end..self.len()) }); + result + } + + /// Returns the lowercase equivalent of this string slice, as a new [`String`]. + /// + /// 'Lowercase' is defined according to the terms of the Unicode Derived Core Property + /// `Lowercase`. + /// + /// Since some characters can expand into multiple characters when changing + /// the case, this function returns a [`String`] instead of modifying the + /// parameter in-place. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = "HELLO"; + /// + /// assert_eq!("hello", s.to_lowercase()); + /// ``` + /// + /// A tricky example, with sigma: + /// + /// ``` + /// let sigma = "Σ"; + /// + /// assert_eq!("σ", sigma.to_lowercase()); + /// + /// // but at the end of a word, it's ς, not σ: + /// let odysseus = "ὈΔΥΣΣΕΎΣ"; + /// + /// assert_eq!("ὀδυσσεύς", odysseus.to_lowercase()); + /// ``` + /// + /// Languages without case are not changed: + /// + /// ``` + /// let new_year = "农历新年"; + /// + /// assert_eq!(new_year, new_year.to_lowercase()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the lowercase string as a new String, \ + without modifying the original"] + #[stable(feature = "unicode_case_mapping", since = "1.2.0")] + pub fn to_lowercase(&self) -> String { + let out = convert_while_ascii(self.as_bytes(), u8::to_ascii_lowercase); + + // Safety: we know this is a valid char boundary since + // out.len() is only progressed if ascii bytes are found + let rest = unsafe { self.get_unchecked(out.len()..) }; + + // Safety: We have written only valid ASCII to our vec + let mut s = unsafe { String::from_utf8_unchecked(out) }; + + for (i, c) in rest[..].char_indices() { + if c == 'Σ' { + // Σ maps to σ, except at the end of a word where it maps to ς. + // This is the only conditional (contextual) but language-independent mapping + // in `SpecialCasing.txt`, + // so hard-code it rather than have a generic "condition" mechanism. + // See https://github.com/rust-lang/rust/issues/26035 + map_uppercase_sigma(rest, i, &mut s) + } else { + match conversions::to_lower(c) { + [a, '\0', _] => s.push(a), + [a, b, '\0'] => { + s.push(a); + s.push(b); + } + [a, b, c] => { + s.push(a); + s.push(b); + s.push(c); + } + } + } + } + return s; + + fn map_uppercase_sigma(from: &str, i: usize, to: &mut String) { + // See https://www.unicode.org/versions/Unicode7.0.0/ch03.pdf#G33992 + // for the definition of `Final_Sigma`. + debug_assert!('Σ'.len_utf8() == 2); + let is_word_final = case_ignoreable_then_cased(from[..i].chars().rev()) + && !case_ignoreable_then_cased(from[i + 2..].chars()); + to.push_str(if is_word_final { "ς" } else { "σ" }); + } + + fn case_ignoreable_then_cased<I: Iterator<Item = char>>(iter: I) -> bool { + use core::unicode::{Case_Ignorable, Cased}; + match iter.skip_while(|&c| Case_Ignorable(c)).next() { + Some(c) => Cased(c), + None => false, + } + } + } + + /// Returns the uppercase equivalent of this string slice, as a new [`String`]. + /// + /// 'Uppercase' is defined according to the terms of the Unicode Derived Core Property + /// `Uppercase`. + /// + /// Since some characters can expand into multiple characters when changing + /// the case, this function returns a [`String`] instead of modifying the + /// parameter in-place. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = "hello"; + /// + /// assert_eq!("HELLO", s.to_uppercase()); + /// ``` + /// + /// Scripts without case are not changed: + /// + /// ``` + /// let new_year = "农历新年"; + /// + /// assert_eq!(new_year, new_year.to_uppercase()); + /// ``` + /// + /// One character can become multiple: + /// ``` + /// let s = "tschüß"; + /// + /// assert_eq!("TSCHÜSS", s.to_uppercase()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "this returns the uppercase string as a new String, \ + without modifying the original"] + #[stable(feature = "unicode_case_mapping", since = "1.2.0")] + pub fn to_uppercase(&self) -> String { + let out = convert_while_ascii(self.as_bytes(), u8::to_ascii_uppercase); + + // Safety: we know this is a valid char boundary since + // out.len() is only progressed if ascii bytes are found + let rest = unsafe { self.get_unchecked(out.len()..) }; + + // Safety: We have written only valid ASCII to our vec + let mut s = unsafe { String::from_utf8_unchecked(out) }; + + for c in rest.chars() { + match conversions::to_upper(c) { + [a, '\0', _] => s.push(a), + [a, b, '\0'] => { + s.push(a); + s.push(b); + } + [a, b, c] => { + s.push(a); + s.push(b); + s.push(c); + } + } + } + s + } + + /// Converts a [`Box<str>`] into a [`String`] without copying or allocating. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let string = String::from("birthday gift"); + /// let boxed_str = string.clone().into_boxed_str(); + /// + /// assert_eq!(boxed_str.into_string(), string); + /// ``` + #[stable(feature = "box_str", since = "1.4.0")] + #[rustc_allow_incoherent_impl] + #[must_use = "`self` will be dropped if the result is not used"] + #[inline] + pub fn into_string(self: Box<str>) -> String { + let slice = Box::<[u8]>::from(self); + unsafe { String::from_utf8_unchecked(slice.into_vec()) } + } + + /// Creates a new [`String`] by repeating a string `n` times. + /// + /// # Panics + /// + /// This function will panic if the capacity would overflow. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// assert_eq!("abc".repeat(4), String::from("abcabcabcabc")); + /// ``` + /// + /// A panic upon overflow: + /// + /// ```should_panic + /// // this will panic at runtime + /// let huge = "0123456789abcdef".repeat(usize::MAX); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use] + #[stable(feature = "repeat_str", since = "1.16.0")] + pub fn repeat(&self, n: usize) -> String { + unsafe { String::from_utf8_unchecked(self.as_bytes().repeat(n)) } + } + + /// Returns a copy of this string where each character is mapped to its + /// ASCII upper case equivalent. + /// + /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z', + /// but non-ASCII letters are unchanged. + /// + /// To uppercase the value in-place, use [`make_ascii_uppercase`]. + /// + /// To uppercase ASCII characters in addition to non-ASCII characters, use + /// [`to_uppercase`]. + /// + /// # Examples + /// + /// ``` + /// let s = "Grüße, Jürgen ❤"; + /// + /// assert_eq!("GRüßE, JüRGEN ❤", s.to_ascii_uppercase()); + /// ``` + /// + /// [`make_ascii_uppercase`]: str::make_ascii_uppercase + /// [`to_uppercase`]: #method.to_uppercase + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "to uppercase the value in-place, use `make_ascii_uppercase()`"] + #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")] + #[inline] + pub fn to_ascii_uppercase(&self) -> String { + let mut bytes = self.as_bytes().to_vec(); + bytes.make_ascii_uppercase(); + // make_ascii_uppercase() preserves the UTF-8 invariant. + unsafe { String::from_utf8_unchecked(bytes) } + } + + /// Returns a copy of this string where each character is mapped to its + /// ASCII lower case equivalent. + /// + /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z', + /// but non-ASCII letters are unchanged. + /// + /// To lowercase the value in-place, use [`make_ascii_lowercase`]. + /// + /// To lowercase ASCII characters in addition to non-ASCII characters, use + /// [`to_lowercase`]. + /// + /// # Examples + /// + /// ``` + /// let s = "Grüße, Jürgen ❤"; + /// + /// assert_eq!("grüße, jürgen ❤", s.to_ascii_lowercase()); + /// ``` + /// + /// [`make_ascii_lowercase`]: str::make_ascii_lowercase + /// [`to_lowercase`]: #method.to_lowercase + #[cfg(not(no_global_oom_handling))] + #[rustc_allow_incoherent_impl] + #[must_use = "to lowercase the value in-place, use `make_ascii_lowercase()`"] + #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")] + #[inline] + pub fn to_ascii_lowercase(&self) -> String { + let mut bytes = self.as_bytes().to_vec(); + bytes.make_ascii_lowercase(); + // make_ascii_lowercase() preserves the UTF-8 invariant. + unsafe { String::from_utf8_unchecked(bytes) } + } +} + +/// Converts a boxed slice of bytes to a boxed string slice without checking +/// that the string contains valid UTF-8. +/// +/// # Examples +/// +/// Basic usage: +/// +/// ``` +/// let smile_utf8 = Box::new([226, 152, 186]); +/// let smile = unsafe { std::str::from_boxed_utf8_unchecked(smile_utf8) }; +/// +/// assert_eq!("☺", &*smile); +/// ``` +#[stable(feature = "str_box_extras", since = "1.20.0")] +#[must_use] +#[inline] +pub unsafe fn from_boxed_utf8_unchecked(v: Box<[u8]>) -> Box<str> { + unsafe { Box::from_raw(Box::into_raw(v) as *mut str) } +} + +/// Converts the bytes while the bytes are still ascii. +/// For better average performance, this is happens in chunks of `2*size_of::<usize>()`. +/// Returns a vec with the converted bytes. +#[inline] +#[cfg(not(test))] +#[cfg(not(no_global_oom_handling))] +fn convert_while_ascii(b: &[u8], convert: fn(&u8) -> u8) -> Vec<u8> { + let mut out = Vec::with_capacity(b.len()); + + const USIZE_SIZE: usize = mem::size_of::<usize>(); + const MAGIC_UNROLL: usize = 2; + const N: usize = USIZE_SIZE * MAGIC_UNROLL; + const NONASCII_MASK: usize = usize::from_ne_bytes([0x80; USIZE_SIZE]); + + let mut i = 0; + unsafe { + while i + N <= b.len() { + // Safety: we have checks the sizes `b` and `out` to know that our + let in_chunk = b.get_unchecked(i..i + N); + let out_chunk = out.spare_capacity_mut().get_unchecked_mut(i..i + N); + + let mut bits = 0; + for j in 0..MAGIC_UNROLL { + // read the bytes 1 usize at a time (unaligned since we haven't checked the alignment) + // safety: in_chunk is valid bytes in the range + bits |= in_chunk.as_ptr().cast::<usize>().add(j).read_unaligned(); + } + // if our chunks aren't ascii, then return only the prior bytes as init + if bits & NONASCII_MASK != 0 { + break; + } + + // perform the case conversions on N bytes (gets heavily autovec'd) + for j in 0..N { + // safety: in_chunk and out_chunk is valid bytes in the range + let out = out_chunk.get_unchecked_mut(j); + out.write(convert(in_chunk.get_unchecked(j))); + } + + // mark these bytes as initialised + i += N; + } + out.set_len(i); + } + + out +} diff --git a/library/alloc/src/string.rs b/library/alloc/src/string.rs new file mode 100644 index 000000000..a5118e533 --- /dev/null +++ b/library/alloc/src/string.rs @@ -0,0 +1,2951 @@ +//! A UTF-8–encoded, growable string. +//! +//! This module contains the [`String`] type, the [`ToString`] trait for +//! converting to strings, and several error types that may result from +//! working with [`String`]s. +//! +//! # Examples +//! +//! There are multiple ways to create a new [`String`] from a string literal: +//! +//! ``` +//! let s = "Hello".to_string(); +//! +//! let s = String::from("world"); +//! let s: String = "also this".into(); +//! ``` +//! +//! You can create a new [`String`] from an existing one by concatenating with +//! `+`: +//! +//! ``` +//! let s = "Hello".to_string(); +//! +//! let message = s + " world!"; +//! ``` +//! +//! If you have a vector of valid UTF-8 bytes, you can make a [`String`] out of +//! it. You can do the reverse too. +//! +//! ``` +//! let sparkle_heart = vec![240, 159, 146, 150]; +//! +//! // We know these bytes are valid, so we'll use `unwrap()`. +//! let sparkle_heart = String::from_utf8(sparkle_heart).unwrap(); +//! +//! assert_eq!("💖", sparkle_heart); +//! +//! let bytes = sparkle_heart.into_bytes(); +//! +//! assert_eq!(bytes, [240, 159, 146, 150]); +//! ``` + +#![stable(feature = "rust1", since = "1.0.0")] + +#[cfg(not(no_global_oom_handling))] +use core::char::{decode_utf16, REPLACEMENT_CHARACTER}; +use core::fmt; +use core::hash; +use core::iter::FusedIterator; +#[cfg(not(no_global_oom_handling))] +use core::iter::{from_fn, FromIterator}; +#[cfg(not(no_global_oom_handling))] +use core::ops::Add; +#[cfg(not(no_global_oom_handling))] +use core::ops::AddAssign; +#[cfg(not(no_global_oom_handling))] +use core::ops::Bound::{Excluded, Included, Unbounded}; +use core::ops::{self, Index, IndexMut, Range, RangeBounds}; +use core::ptr; +use core::slice; +#[cfg(not(no_global_oom_handling))] +use core::str::lossy; +use core::str::pattern::Pattern; + +#[cfg(not(no_global_oom_handling))] +use crate::borrow::{Cow, ToOwned}; +use crate::boxed::Box; +use crate::collections::TryReserveError; +use crate::str::{self, Chars, Utf8Error}; +#[cfg(not(no_global_oom_handling))] +use crate::str::{from_boxed_utf8_unchecked, FromStr}; +use crate::vec::Vec; + +/// A UTF-8–encoded, growable string. +/// +/// The `String` type is the most common string type that has ownership over the +/// contents of the string. It has a close relationship with its borrowed +/// counterpart, the primitive [`str`]. +/// +/// # Examples +/// +/// You can create a `String` from [a literal string][`&str`] with [`String::from`]: +/// +/// [`String::from`]: From::from +/// +/// ``` +/// let hello = String::from("Hello, world!"); +/// ``` +/// +/// You can append a [`char`] to a `String` with the [`push`] method, and +/// append a [`&str`] with the [`push_str`] method: +/// +/// ``` +/// let mut hello = String::from("Hello, "); +/// +/// hello.push('w'); +/// hello.push_str("orld!"); +/// ``` +/// +/// [`push`]: String::push +/// [`push_str`]: String::push_str +/// +/// If you have a vector of UTF-8 bytes, you can create a `String` from it with +/// the [`from_utf8`] method: +/// +/// ``` +/// // some bytes, in a vector +/// let sparkle_heart = vec![240, 159, 146, 150]; +/// +/// // We know these bytes are valid, so we'll use `unwrap()`. +/// let sparkle_heart = String::from_utf8(sparkle_heart).unwrap(); +/// +/// assert_eq!("💖", sparkle_heart); +/// ``` +/// +/// [`from_utf8`]: String::from_utf8 +/// +/// # UTF-8 +/// +/// `String`s are always valid UTF-8. If you need a non-UTF-8 string, consider +/// [`OsString`]. It is similar, but without the UTF-8 constraint. Because UTF-8 +/// is a variable width encoding, `String`s are typically smaller than an array of +/// the same `chars`: +/// +/// ``` +/// use std::mem; +/// +/// // `s` is ASCII which represents each `char` as one byte +/// let s = "hello"; +/// assert_eq!(s.len(), 5); +/// +/// // A `char` array with the same contents would be longer because +/// // every `char` is four bytes +/// let s = ['h', 'e', 'l', 'l', 'o']; +/// let size: usize = s.into_iter().map(|c| mem::size_of_val(&c)).sum(); +/// assert_eq!(size, 20); +/// +/// // However, for non-ASCII strings, the difference will be smaller +/// // and sometimes they are the same +/// let s = "💖💖💖💖💖"; +/// assert_eq!(s.len(), 20); +/// +/// let s = ['💖', '💖', '💖', '💖', '💖']; +/// let size: usize = s.into_iter().map(|c| mem::size_of_val(&c)).sum(); +/// assert_eq!(size, 20); +/// ``` +/// +/// This raises interesting questions as to how `s[i]` should work. +/// What should `i` be here? Several options include byte indices and +/// `char` indices but, because of UTF-8 encoding, only byte indices +/// would provide constant time indexing. Getting the `i`th `char`, for +/// example, is available using [`chars`]: +/// +/// ``` +/// let s = "hello"; +/// let third_character = s.chars().nth(2); +/// assert_eq!(third_character, Some('l')); +/// +/// let s = "💖💖💖💖💖"; +/// let third_character = s.chars().nth(2); +/// assert_eq!(third_character, Some('💖')); +/// ``` +/// +/// Next, what should `s[i]` return? Because indexing returns a reference +/// to underlying data it could be `&u8`, `&[u8]`, or something else similar. +/// Since we're only providing one index, `&u8` makes the most sense but that +/// might not be what the user expects and can be explicitly achieved with +/// [`as_bytes()`]: +/// +/// ``` +/// // The first byte is 104 - the byte value of `'h'` +/// let s = "hello"; +/// assert_eq!(s.as_bytes()[0], 104); +/// // or +/// assert_eq!(s.as_bytes()[0], b'h'); +/// +/// // The first byte is 240 which isn't obviously useful +/// let s = "💖💖💖💖💖"; +/// assert_eq!(s.as_bytes()[0], 240); +/// ``` +/// +/// Due to these ambiguities/restrictions, indexing with a `usize` is simply +/// forbidden: +/// +/// ```compile_fail,E0277 +/// let s = "hello"; +/// +/// // The following will not compile! +/// println!("The first letter of s is {}", s[0]); +/// ``` +/// +/// It is more clear, however, how `&s[i..j]` should work (that is, +/// indexing with a range). It should accept byte indices (to be constant-time) +/// and return a `&str` which is UTF-8 encoded. This is also called "string slicing". +/// Note this will panic if the byte indices provided are not character +/// boundaries - see [`is_char_boundary`] for more details. See the implementations +/// for [`SliceIndex<str>`] for more details on string slicing. For a non-panicking +/// version of string slicing, see [`get`]. +/// +/// [`OsString`]: ../../std/ffi/struct.OsString.html "ffi::OsString" +/// [`SliceIndex<str>`]: core::slice::SliceIndex +/// [`as_bytes()`]: str::as_bytes +/// [`get`]: str::get +/// [`is_char_boundary`]: str::is_char_boundary +/// +/// The [`bytes`] and [`chars`] methods return iterators over the bytes and +/// codepoints of the string, respectively. To iterate over codepoints along +/// with byte indices, use [`char_indices`]. +/// +/// [`bytes`]: str::bytes +/// [`chars`]: str::chars +/// [`char_indices`]: str::char_indices +/// +/// # Deref +/// +/// `String` implements <code>[Deref]<Target = [str]></code>, and so inherits all of [`str`]'s +/// methods. In addition, this means that you can pass a `String` to a +/// function which takes a [`&str`] by using an ampersand (`&`): +/// +/// ``` +/// fn takes_str(s: &str) { } +/// +/// let s = String::from("Hello"); +/// +/// takes_str(&s); +/// ``` +/// +/// This will create a [`&str`] from the `String` and pass it in. This +/// conversion is very inexpensive, and so generally, functions will accept +/// [`&str`]s as arguments unless they need a `String` for some specific +/// reason. +/// +/// In certain cases Rust doesn't have enough information to make this +/// conversion, known as [`Deref`] coercion. In the following example a string +/// slice [`&'a str`][`&str`] implements the trait `TraitExample`, and the function +/// `example_func` takes anything that implements the trait. In this case Rust +/// would need to make two implicit conversions, which Rust doesn't have the +/// means to do. For that reason, the following example will not compile. +/// +/// ```compile_fail,E0277 +/// trait TraitExample {} +/// +/// impl<'a> TraitExample for &'a str {} +/// +/// fn example_func<A: TraitExample>(example_arg: A) {} +/// +/// let example_string = String::from("example_string"); +/// example_func(&example_string); +/// ``` +/// +/// There are two options that would work instead. The first would be to +/// change the line `example_func(&example_string);` to +/// `example_func(example_string.as_str());`, using the method [`as_str()`] +/// to explicitly extract the string slice containing the string. The second +/// way changes `example_func(&example_string);` to +/// `example_func(&*example_string);`. In this case we are dereferencing a +/// `String` to a [`str`], then referencing the [`str`] back to +/// [`&str`]. The second way is more idiomatic, however both work to do the +/// conversion explicitly rather than relying on the implicit conversion. +/// +/// # Representation +/// +/// A `String` is made up of three components: a pointer to some bytes, a +/// length, and a capacity. The pointer points to an internal buffer `String` +/// uses to store its data. The length is the number of bytes currently stored +/// in the buffer, and the capacity is the size of the buffer in bytes. As such, +/// the length will always be less than or equal to the capacity. +/// +/// This buffer is always stored on the heap. +/// +/// You can look at these with the [`as_ptr`], [`len`], and [`capacity`] +/// methods: +/// +/// ``` +/// use std::mem; +/// +/// let story = String::from("Once upon a time..."); +/// +// FIXME Update this when vec_into_raw_parts is stabilized +/// // Prevent automatically dropping the String's data +/// let mut story = mem::ManuallyDrop::new(story); +/// +/// let ptr = story.as_mut_ptr(); +/// let len = story.len(); +/// let capacity = story.capacity(); +/// +/// // story has nineteen bytes +/// assert_eq!(19, len); +/// +/// // We can re-build a String out of ptr, len, and capacity. This is all +/// // unsafe because we are responsible for making sure the components are +/// // valid: +/// let s = unsafe { String::from_raw_parts(ptr, len, capacity) } ; +/// +/// assert_eq!(String::from("Once upon a time..."), s); +/// ``` +/// +/// [`as_ptr`]: str::as_ptr +/// [`len`]: String::len +/// [`capacity`]: String::capacity +/// +/// If a `String` has enough capacity, adding elements to it will not +/// re-allocate. For example, consider this program: +/// +/// ``` +/// let mut s = String::new(); +/// +/// println!("{}", s.capacity()); +/// +/// for _ in 0..5 { +/// s.push_str("hello"); +/// println!("{}", s.capacity()); +/// } +/// ``` +/// +/// This will output the following: +/// +/// ```text +/// 0 +/// 8 +/// 16 +/// 16 +/// 32 +/// 32 +/// ``` +/// +/// At first, we have no memory allocated at all, but as we append to the +/// string, it increases its capacity appropriately. If we instead use the +/// [`with_capacity`] method to allocate the correct capacity initially: +/// +/// ``` +/// let mut s = String::with_capacity(25); +/// +/// println!("{}", s.capacity()); +/// +/// for _ in 0..5 { +/// s.push_str("hello"); +/// println!("{}", s.capacity()); +/// } +/// ``` +/// +/// [`with_capacity`]: String::with_capacity +/// +/// We end up with a different output: +/// +/// ```text +/// 25 +/// 25 +/// 25 +/// 25 +/// 25 +/// 25 +/// ``` +/// +/// Here, there's no need to allocate more memory inside the loop. +/// +/// [str]: prim@str "str" +/// [`str`]: prim@str "str" +/// [`&str`]: prim@str "&str" +/// [Deref]: core::ops::Deref "ops::Deref" +/// [`Deref`]: core::ops::Deref "ops::Deref" +/// [`as_str()`]: String::as_str +#[derive(PartialOrd, Eq, Ord)] +#[cfg_attr(not(test), rustc_diagnostic_item = "String")] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct String { + vec: Vec<u8>, +} + +/// A possible error value when converting a `String` from a UTF-8 byte vector. +/// +/// This type is the error type for the [`from_utf8`] method on [`String`]. It +/// is designed in such a way to carefully avoid reallocations: the +/// [`into_bytes`] method will give back the byte vector that was used in the +/// conversion attempt. +/// +/// [`from_utf8`]: String::from_utf8 +/// [`into_bytes`]: FromUtf8Error::into_bytes +/// +/// The [`Utf8Error`] type provided by [`std::str`] represents an error that may +/// occur when converting a slice of [`u8`]s to a [`&str`]. In this sense, it's +/// an analogue to `FromUtf8Error`, and you can get one from a `FromUtf8Error` +/// through the [`utf8_error`] method. +/// +/// [`Utf8Error`]: str::Utf8Error "std::str::Utf8Error" +/// [`std::str`]: core::str "std::str" +/// [`&str`]: prim@str "&str" +/// [`utf8_error`]: FromUtf8Error::utf8_error +/// +/// # Examples +/// +/// Basic usage: +/// +/// ``` +/// // some invalid bytes, in a vector +/// let bytes = vec![0, 159]; +/// +/// let value = String::from_utf8(bytes); +/// +/// assert!(value.is_err()); +/// assert_eq!(vec![0, 159], value.unwrap_err().into_bytes()); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(no_global_oom_handling), derive(Clone))] +#[derive(Debug, PartialEq, Eq)] +pub struct FromUtf8Error { + bytes: Vec<u8>, + error: Utf8Error, +} + +/// A possible error value when converting a `String` from a UTF-16 byte slice. +/// +/// This type is the error type for the [`from_utf16`] method on [`String`]. +/// +/// [`from_utf16`]: String::from_utf16 +/// # Examples +/// +/// Basic usage: +/// +/// ``` +/// // 𝄞mu<invalid>ic +/// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075, +/// 0xD800, 0x0069, 0x0063]; +/// +/// assert!(String::from_utf16(v).is_err()); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Debug)] +pub struct FromUtf16Error(()); + +impl String { + /// Creates a new empty `String`. + /// + /// Given that the `String` is empty, this will not allocate any initial + /// buffer. While that means that this initial operation is very + /// inexpensive, it may cause excessive allocation later when you add + /// data. If you have an idea of how much data the `String` will hold, + /// consider the [`with_capacity`] method to prevent excessive + /// re-allocation. + /// + /// [`with_capacity`]: String::with_capacity + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::new(); + /// ``` + #[inline] + #[rustc_const_stable(feature = "const_string_new", since = "1.39.0")] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub const fn new() -> String { + String { vec: Vec::new() } + } + + /// Creates a new empty `String` with at least the specified capacity. + /// + /// `String`s have an internal buffer to hold their data. The capacity is + /// the length of that buffer, and can be queried with the [`capacity`] + /// method. This method creates an empty `String`, but one with an initial + /// buffer that can hold at least `capacity` bytes. This is useful when you + /// may be appending a bunch of data to the `String`, reducing the number of + /// reallocations it needs to do. + /// + /// [`capacity`]: String::capacity + /// + /// If the given capacity is `0`, no allocation will occur, and this method + /// is identical to the [`new`] method. + /// + /// [`new`]: String::new + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::with_capacity(10); + /// + /// // The String contains no chars, even though it has capacity for more + /// assert_eq!(s.len(), 0); + /// + /// // These are all done without reallocating... + /// let cap = s.capacity(); + /// for _ in 0..10 { + /// s.push('a'); + /// } + /// + /// assert_eq!(s.capacity(), cap); + /// + /// // ...but this may make the string reallocate + /// s.push('a'); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn with_capacity(capacity: usize) -> String { + String { vec: Vec::with_capacity(capacity) } + } + + // HACK(japaric): with cfg(test) the inherent `[T]::to_vec` method, which is + // required for this method definition, is not available. Since we don't + // require this method for testing purposes, I'll just stub it + // NB see the slice::hack module in slice.rs for more information + #[inline] + #[cfg(test)] + pub fn from_str(_: &str) -> String { + panic!("not available with cfg(test)"); + } + + /// Converts a vector of bytes to a `String`. + /// + /// A string ([`String`]) is made of bytes ([`u8`]), and a vector of bytes + /// ([`Vec<u8>`]) is made of bytes, so this function converts between the + /// two. Not all byte slices are valid `String`s, however: `String` + /// requires that it is valid UTF-8. `from_utf8()` checks to ensure that + /// the bytes are valid UTF-8, and then does the conversion. + /// + /// If you are sure that the byte slice is valid UTF-8, and you don't want + /// to incur the overhead of the validity check, there is an unsafe version + /// of this function, [`from_utf8_unchecked`], which has the same behavior + /// but skips the check. + /// + /// This method will take care to not copy the vector, for efficiency's + /// sake. + /// + /// If you need a [`&str`] instead of a `String`, consider + /// [`str::from_utf8`]. + /// + /// The inverse of this method is [`into_bytes`]. + /// + /// # Errors + /// + /// Returns [`Err`] if the slice is not UTF-8 with a description as to why the + /// provided bytes are not UTF-8. The vector you moved in is also included. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some bytes, in a vector + /// let sparkle_heart = vec![240, 159, 146, 150]; + /// + /// // We know these bytes are valid, so we'll use `unwrap()`. + /// let sparkle_heart = String::from_utf8(sparkle_heart).unwrap(); + /// + /// assert_eq!("💖", sparkle_heart); + /// ``` + /// + /// Incorrect bytes: + /// + /// ``` + /// // some invalid bytes, in a vector + /// let sparkle_heart = vec![0, 159, 146, 150]; + /// + /// assert!(String::from_utf8(sparkle_heart).is_err()); + /// ``` + /// + /// See the docs for [`FromUtf8Error`] for more details on what you can do + /// with this error. + /// + /// [`from_utf8_unchecked`]: String::from_utf8_unchecked + /// [`Vec<u8>`]: crate::vec::Vec "Vec" + /// [`&str`]: prim@str "&str" + /// [`into_bytes`]: String::into_bytes + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn from_utf8(vec: Vec<u8>) -> Result<String, FromUtf8Error> { + match str::from_utf8(&vec) { + Ok(..) => Ok(String { vec }), + Err(e) => Err(FromUtf8Error { bytes: vec, error: e }), + } + } + + /// Converts a slice of bytes to a string, including invalid characters. + /// + /// Strings are made of bytes ([`u8`]), and a slice of bytes + /// ([`&[u8]`][byteslice]) is made of bytes, so this function converts + /// between the two. Not all byte slices are valid strings, however: strings + /// are required to be valid UTF-8. During this conversion, + /// `from_utf8_lossy()` will replace any invalid UTF-8 sequences with + /// [`U+FFFD REPLACEMENT CHARACTER`][U+FFFD], which looks like this: � + /// + /// [byteslice]: prim@slice + /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER + /// + /// If you are sure that the byte slice is valid UTF-8, and you don't want + /// to incur the overhead of the conversion, there is an unsafe version + /// of this function, [`from_utf8_unchecked`], which has the same behavior + /// but skips the checks. + /// + /// [`from_utf8_unchecked`]: String::from_utf8_unchecked + /// + /// This function returns a [`Cow<'a, str>`]. If our byte slice is invalid + /// UTF-8, then we need to insert the replacement characters, which will + /// change the size of the string, and hence, require a `String`. But if + /// it's already valid UTF-8, we don't need a new allocation. This return + /// type allows us to handle both cases. + /// + /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow" + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some bytes, in a vector + /// let sparkle_heart = vec![240, 159, 146, 150]; + /// + /// let sparkle_heart = String::from_utf8_lossy(&sparkle_heart); + /// + /// assert_eq!("💖", sparkle_heart); + /// ``` + /// + /// Incorrect bytes: + /// + /// ``` + /// // some invalid bytes + /// let input = b"Hello \xF0\x90\x80World"; + /// let output = String::from_utf8_lossy(input); + /// + /// assert_eq!("Hello �World", output); + /// ``` + #[must_use] + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn from_utf8_lossy(v: &[u8]) -> Cow<'_, str> { + let mut iter = lossy::Utf8Lossy::from_bytes(v).chunks(); + + let first_valid = if let Some(chunk) = iter.next() { + let lossy::Utf8LossyChunk { valid, broken } = chunk; + if broken.is_empty() { + debug_assert_eq!(valid.len(), v.len()); + return Cow::Borrowed(valid); + } + valid + } else { + return Cow::Borrowed(""); + }; + + const REPLACEMENT: &str = "\u{FFFD}"; + + let mut res = String::with_capacity(v.len()); + res.push_str(first_valid); + res.push_str(REPLACEMENT); + + for lossy::Utf8LossyChunk { valid, broken } in iter { + res.push_str(valid); + if !broken.is_empty() { + res.push_str(REPLACEMENT); + } + } + + Cow::Owned(res) + } + + /// Decode a UTF-16–encoded vector `v` into a `String`, returning [`Err`] + /// if `v` contains any invalid data. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // 𝄞music + /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075, + /// 0x0073, 0x0069, 0x0063]; + /// assert_eq!(String::from("𝄞music"), + /// String::from_utf16(v).unwrap()); + /// + /// // 𝄞mu<invalid>ic + /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075, + /// 0xD800, 0x0069, 0x0063]; + /// assert!(String::from_utf16(v).is_err()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn from_utf16(v: &[u16]) -> Result<String, FromUtf16Error> { + // This isn't done via collect::<Result<_, _>>() for performance reasons. + // FIXME: the function can be simplified again when #48994 is closed. + let mut ret = String::with_capacity(v.len()); + for c in decode_utf16(v.iter().cloned()) { + if let Ok(c) = c { + ret.push(c); + } else { + return Err(FromUtf16Error(())); + } + } + Ok(ret) + } + + /// Decode a UTF-16–encoded slice `v` into a `String`, replacing + /// invalid data with [the replacement character (`U+FFFD`)][U+FFFD]. + /// + /// Unlike [`from_utf8_lossy`] which returns a [`Cow<'a, str>`], + /// `from_utf16_lossy` returns a `String` since the UTF-16 to UTF-8 + /// conversion requires a memory allocation. + /// + /// [`from_utf8_lossy`]: String::from_utf8_lossy + /// [`Cow<'a, str>`]: crate::borrow::Cow "borrow::Cow" + /// [U+FFFD]: core::char::REPLACEMENT_CHARACTER + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // 𝄞mus<invalid>ic<invalid> + /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075, + /// 0x0073, 0xDD1E, 0x0069, 0x0063, + /// 0xD834]; + /// + /// assert_eq!(String::from("𝄞mus\u{FFFD}ic\u{FFFD}"), + /// String::from_utf16_lossy(v)); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[must_use] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn from_utf16_lossy(v: &[u16]) -> String { + decode_utf16(v.iter().cloned()).map(|r| r.unwrap_or(REPLACEMENT_CHARACTER)).collect() + } + + /// Decomposes a `String` into its raw components. + /// + /// Returns the raw pointer to the underlying data, the length of + /// the string (in bytes), and the allocated capacity of the data + /// (in bytes). These are the same arguments in the same order as + /// the arguments to [`from_raw_parts`]. + /// + /// After calling this function, the caller is responsible for the + /// memory previously managed by the `String`. The only way to do + /// this is to convert the raw pointer, length, and capacity back + /// into a `String` with the [`from_raw_parts`] function, allowing + /// the destructor to perform the cleanup. + /// + /// [`from_raw_parts`]: String::from_raw_parts + /// + /// # Examples + /// + /// ``` + /// #![feature(vec_into_raw_parts)] + /// let s = String::from("hello"); + /// + /// let (ptr, len, cap) = s.into_raw_parts(); + /// + /// let rebuilt = unsafe { String::from_raw_parts(ptr, len, cap) }; + /// assert_eq!(rebuilt, "hello"); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[unstable(feature = "vec_into_raw_parts", reason = "new API", issue = "65816")] + pub fn into_raw_parts(self) -> (*mut u8, usize, usize) { + self.vec.into_raw_parts() + } + + /// Creates a new `String` from a length, capacity, and pointer. + /// + /// # Safety + /// + /// This is highly unsafe, due to the number of invariants that aren't + /// checked: + /// + /// * The memory at `buf` needs to have been previously allocated by the + /// same allocator the standard library uses, with a required alignment of exactly 1. + /// * `length` needs to be less than or equal to `capacity`. + /// * `capacity` needs to be the correct value. + /// * The first `length` bytes at `buf` need to be valid UTF-8. + /// + /// Violating these may cause problems like corrupting the allocator's + /// internal data structures. For example, it is normally **not** safe to + /// build a `String` from a pointer to a C `char` array containing UTF-8 + /// _unless_ you are certain that array was originally allocated by the + /// Rust standard library's allocator. + /// + /// The ownership of `buf` is effectively transferred to the + /// `String` which may then deallocate, reallocate or change the + /// contents of memory pointed to by the pointer at will. Ensure + /// that nothing else uses the pointer after calling this + /// function. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use std::mem; + /// + /// unsafe { + /// let s = String::from("hello"); + /// + // FIXME Update this when vec_into_raw_parts is stabilized + /// // Prevent automatically dropping the String's data + /// let mut s = mem::ManuallyDrop::new(s); + /// + /// let ptr = s.as_mut_ptr(); + /// let len = s.len(); + /// let capacity = s.capacity(); + /// + /// let s = String::from_raw_parts(ptr, len, capacity); + /// + /// assert_eq!(String::from("hello"), s); + /// } + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn from_raw_parts(buf: *mut u8, length: usize, capacity: usize) -> String { + unsafe { String { vec: Vec::from_raw_parts(buf, length, capacity) } } + } + + /// Converts a vector of bytes to a `String` without checking that the + /// string contains valid UTF-8. + /// + /// See the safe version, [`from_utf8`], for more details. + /// + /// [`from_utf8`]: String::from_utf8 + /// + /// # Safety + /// + /// This function is unsafe because it does not check that the bytes passed + /// to it are valid UTF-8. If this constraint is violated, it may cause + /// memory unsafety issues with future users of the `String`, as the rest of + /// the standard library assumes that `String`s are valid UTF-8. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some bytes, in a vector + /// let sparkle_heart = vec![240, 159, 146, 150]; + /// + /// let sparkle_heart = unsafe { + /// String::from_utf8_unchecked(sparkle_heart) + /// }; + /// + /// assert_eq!("💖", sparkle_heart); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn from_utf8_unchecked(bytes: Vec<u8>) -> String { + String { vec: bytes } + } + + /// Converts a `String` into a byte vector. + /// + /// This consumes the `String`, so we do not need to copy its contents. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::from("hello"); + /// let bytes = s.into_bytes(); + /// + /// assert_eq!(&[104, 101, 108, 108, 111][..], &bytes[..]); + /// ``` + #[inline] + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_bytes(self) -> Vec<u8> { + self.vec + } + + /// Extracts a string slice containing the entire `String`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::from("foo"); + /// + /// assert_eq!("foo", s.as_str()); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "string_as_str", since = "1.7.0")] + pub fn as_str(&self) -> &str { + self + } + + /// Converts a `String` into a mutable string slice. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foobar"); + /// let s_mut_str = s.as_mut_str(); + /// + /// s_mut_str.make_ascii_uppercase(); + /// + /// assert_eq!("FOOBAR", s_mut_str); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "string_as_str", since = "1.7.0")] + pub fn as_mut_str(&mut self) -> &mut str { + self + } + + /// Appends a given string slice onto the end of this `String`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// s.push_str("bar"); + /// + /// assert_eq!("foobar", s); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push_str(&mut self, string: &str) { + self.vec.extend_from_slice(string.as_bytes()) + } + + /// Copies elements from `src` range to the end of the string. + /// + /// ## Panics + /// + /// Panics if the starting point or end point do not lie on a [`char`] + /// boundary, or if they're out of bounds. + /// + /// ## Examples + /// + /// ``` + /// #![feature(string_extend_from_within)] + /// let mut string = String::from("abcde"); + /// + /// string.extend_from_within(2..); + /// assert_eq!(string, "abcdecde"); + /// + /// string.extend_from_within(..2); + /// assert_eq!(string, "abcdecdeab"); + /// + /// string.extend_from_within(4..8); + /// assert_eq!(string, "abcdecdeabecde"); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "string_extend_from_within", issue = "none")] + pub fn extend_from_within<R>(&mut self, src: R) + where + R: RangeBounds<usize>, + { + let src @ Range { start, end } = slice::range(src, ..self.len()); + + assert!(self.is_char_boundary(start)); + assert!(self.is_char_boundary(end)); + + self.vec.extend_from_within(src); + } + + /// Returns this `String`'s capacity, in bytes. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::with_capacity(10); + /// + /// assert!(s.capacity() >= 10); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn capacity(&self) -> usize { + self.vec.capacity() + } + + /// Reserves capacity for at least `additional` bytes more than the + /// current length. The allocator may reserve more space to speculatively + /// avoid frequent allocations. After calling `reserve`, + /// capacity will be greater than or equal to `self.len() + additional`. + /// Does nothing if capacity is already sufficient. + /// + /// # Panics + /// + /// Panics if the new capacity overflows [`usize`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::new(); + /// + /// s.reserve(10); + /// + /// assert!(s.capacity() >= 10); + /// ``` + /// + /// This might not actually increase the capacity: + /// + /// ``` + /// let mut s = String::with_capacity(10); + /// s.push('a'); + /// s.push('b'); + /// + /// // s now has a length of 2 and a capacity of at least 10 + /// let capacity = s.capacity(); + /// assert_eq!(2, s.len()); + /// assert!(capacity >= 10); + /// + /// // Since we already have at least an extra 8 capacity, calling this... + /// s.reserve(8); + /// + /// // ... doesn't actually increase. + /// assert_eq!(capacity, s.capacity()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve(&mut self, additional: usize) { + self.vec.reserve(additional) + } + + /// Reserves the minimum capacity for at least `additional` bytes more than + /// the current length. Unlike [`reserve`], this will not + /// deliberately over-allocate to speculatively avoid frequent allocations. + /// After calling `reserve_exact`, capacity will be greater than or equal to + /// `self.len() + additional`. Does nothing if the capacity is already + /// sufficient. + /// + /// [`reserve`]: String::reserve + /// + /// # Panics + /// + /// Panics if the new capacity overflows [`usize`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::new(); + /// + /// s.reserve_exact(10); + /// + /// assert!(s.capacity() >= 10); + /// ``` + /// + /// This might not actually increase the capacity: + /// + /// ``` + /// let mut s = String::with_capacity(10); + /// s.push('a'); + /// s.push('b'); + /// + /// // s now has a length of 2 and a capacity of at least 10 + /// let capacity = s.capacity(); + /// assert_eq!(2, s.len()); + /// assert!(capacity >= 10); + /// + /// // Since we already have at least an extra 8 capacity, calling this... + /// s.reserve_exact(8); + /// + /// // ... doesn't actually increase. + /// assert_eq!(capacity, s.capacity()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve_exact(&mut self, additional: usize) { + self.vec.reserve_exact(additional) + } + + /// Tries to reserve capacity for at least `additional` bytes more than the + /// current length. The allocator may reserve more space to speculatively + /// avoid frequent allocations. After calling `try_reserve`, capacity will be + /// greater than or equal to `self.len() + additional` if it returns + /// `Ok(())`. Does nothing if capacity is already sufficient. + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// + /// fn process_data(data: &str) -> Result<String, TryReserveError> { + /// let mut output = String::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// output.push_str(data); + /// + /// Ok(output) + /// } + /// # process_data("rust").expect("why is the test harness OOMing on 4 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.vec.try_reserve(additional) + } + + /// Tries to reserve the minimum capacity for at least `additional` bytes + /// more than the current length. Unlike [`try_reserve`], this will not + /// deliberately over-allocate to speculatively avoid frequent allocations. + /// After calling `try_reserve_exact`, capacity will be greater than or + /// equal to `self.len() + additional` if it returns `Ok(())`. + /// Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the collection more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: String::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// + /// fn process_data(data: &str) -> Result<String, TryReserveError> { + /// let mut output = String::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve_exact(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// output.push_str(data); + /// + /// Ok(output) + /// } + /// # process_data("rust").expect("why is the test harness OOMing on 4 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.vec.try_reserve_exact(additional) + } + + /// Shrinks the capacity of this `String` to match its length. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// s.reserve(100); + /// assert!(s.capacity() >= 100); + /// + /// s.shrink_to_fit(); + /// assert_eq!(3, s.capacity()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn shrink_to_fit(&mut self) { + self.vec.shrink_to_fit() + } + + /// Shrinks the capacity of this `String` with a lower bound. + /// + /// The capacity will remain at least as large as both the length + /// and the supplied value. + /// + /// If the current capacity is less than the lower limit, this is a no-op. + /// + /// # Examples + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// s.reserve(100); + /// assert!(s.capacity() >= 100); + /// + /// s.shrink_to(10); + /// assert!(s.capacity() >= 10); + /// s.shrink_to(0); + /// assert!(s.capacity() >= 3); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "shrink_to", since = "1.56.0")] + pub fn shrink_to(&mut self, min_capacity: usize) { + self.vec.shrink_to(min_capacity) + } + + /// Appends the given [`char`] to the end of this `String`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("abc"); + /// + /// s.push('1'); + /// s.push('2'); + /// s.push('3'); + /// + /// assert_eq!("abc123", s); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push(&mut self, ch: char) { + match ch.len_utf8() { + 1 => self.vec.push(ch as u8), + _ => self.vec.extend_from_slice(ch.encode_utf8(&mut [0; 4]).as_bytes()), + } + } + + /// Returns a byte slice of this `String`'s contents. + /// + /// The inverse of this method is [`from_utf8`]. + /// + /// [`from_utf8`]: String::from_utf8 + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::from("hello"); + /// + /// assert_eq!(&[104, 101, 108, 108, 111], s.as_bytes()); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn as_bytes(&self) -> &[u8] { + &self.vec + } + + /// Shortens this `String` to the specified length. + /// + /// If `new_len` is greater than the string's current length, this has no + /// effect. + /// + /// Note that this method has no effect on the allocated capacity + /// of the string + /// + /// # Panics + /// + /// Panics if `new_len` does not lie on a [`char`] boundary. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("hello"); + /// + /// s.truncate(2); + /// + /// assert_eq!("he", s); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn truncate(&mut self, new_len: usize) { + if new_len <= self.len() { + assert!(self.is_char_boundary(new_len)); + self.vec.truncate(new_len) + } + } + + /// Removes the last character from the string buffer and returns it. + /// + /// Returns [`None`] if this `String` is empty. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// assert_eq!(s.pop(), Some('o')); + /// assert_eq!(s.pop(), Some('o')); + /// assert_eq!(s.pop(), Some('f')); + /// + /// assert_eq!(s.pop(), None); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop(&mut self) -> Option<char> { + let ch = self.chars().rev().next()?; + let newlen = self.len() - ch.len_utf8(); + unsafe { + self.vec.set_len(newlen); + } + Some(ch) + } + + /// Removes a [`char`] from this `String` at a byte position and returns it. + /// + /// This is an *O*(*n*) operation, as it requires copying every element in the + /// buffer. + /// + /// # Panics + /// + /// Panics if `idx` is larger than or equal to the `String`'s length, + /// or if it does not lie on a [`char`] boundary. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// assert_eq!(s.remove(0), 'f'); + /// assert_eq!(s.remove(1), 'o'); + /// assert_eq!(s.remove(0), 'o'); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn remove(&mut self, idx: usize) -> char { + let ch = match self[idx..].chars().next() { + Some(ch) => ch, + None => panic!("cannot remove a char from the end of a string"), + }; + + let next = idx + ch.len_utf8(); + let len = self.len(); + unsafe { + ptr::copy(self.vec.as_ptr().add(next), self.vec.as_mut_ptr().add(idx), len - next); + self.vec.set_len(len - (next - idx)); + } + ch + } + + /// Remove all matches of pattern `pat` in the `String`. + /// + /// # Examples + /// + /// ``` + /// #![feature(string_remove_matches)] + /// let mut s = String::from("Trees are not green, the sky is not blue."); + /// s.remove_matches("not "); + /// assert_eq!("Trees are green, the sky is blue.", s); + /// ``` + /// + /// Matches will be detected and removed iteratively, so in cases where + /// patterns overlap, only the first pattern will be removed: + /// + /// ``` + /// #![feature(string_remove_matches)] + /// let mut s = String::from("banana"); + /// s.remove_matches("ana"); + /// assert_eq!("bna", s); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "string_remove_matches", reason = "new API", issue = "72826")] + pub fn remove_matches<'a, P>(&'a mut self, pat: P) + where + P: for<'x> Pattern<'x>, + { + use core::str::pattern::Searcher; + + let rejections = { + let mut searcher = pat.into_searcher(self); + // Per Searcher::next: + // + // A Match result needs to contain the whole matched pattern, + // however Reject results may be split up into arbitrary many + // adjacent fragments. Both ranges may have zero length. + // + // In practice the implementation of Searcher::next_match tends to + // be more efficient, so we use it here and do some work to invert + // matches into rejections since that's what we want to copy below. + let mut front = 0; + let rejections: Vec<_> = from_fn(|| { + let (start, end) = searcher.next_match()?; + let prev_front = front; + front = end; + Some((prev_front, start)) + }) + .collect(); + rejections.into_iter().chain(core::iter::once((front, self.len()))) + }; + + let mut len = 0; + let ptr = self.vec.as_mut_ptr(); + + for (start, end) in rejections { + let count = end - start; + if start != len { + // SAFETY: per Searcher::next: + // + // The stream of Match and Reject values up to a Done will + // contain index ranges that are adjacent, non-overlapping, + // covering the whole haystack, and laying on utf8 + // boundaries. + unsafe { + ptr::copy(ptr.add(start), ptr.add(len), count); + } + } + len += count; + } + + unsafe { + self.vec.set_len(len); + } + } + + /// Retains only the characters specified by the predicate. + /// + /// In other words, remove all characters `c` such that `f(c)` returns `false`. + /// This method operates in place, visiting each character exactly once in the + /// original order, and preserves the order of the retained characters. + /// + /// # Examples + /// + /// ``` + /// let mut s = String::from("f_o_ob_ar"); + /// + /// s.retain(|c| c != '_'); + /// + /// assert_eq!(s, "foobar"); + /// ``` + /// + /// Because the elements are visited exactly once in the original order, + /// external state may be used to decide which elements to keep. + /// + /// ``` + /// let mut s = String::from("abcde"); + /// let keep = [false, true, true, false, true]; + /// let mut iter = keep.iter(); + /// s.retain(|_| *iter.next().unwrap()); + /// assert_eq!(s, "bce"); + /// ``` + #[inline] + #[stable(feature = "string_retain", since = "1.26.0")] + pub fn retain<F>(&mut self, mut f: F) + where + F: FnMut(char) -> bool, + { + struct SetLenOnDrop<'a> { + s: &'a mut String, + idx: usize, + del_bytes: usize, + } + + impl<'a> Drop for SetLenOnDrop<'a> { + fn drop(&mut self) { + let new_len = self.idx - self.del_bytes; + debug_assert!(new_len <= self.s.len()); + unsafe { self.s.vec.set_len(new_len) }; + } + } + + let len = self.len(); + let mut guard = SetLenOnDrop { s: self, idx: 0, del_bytes: 0 }; + + while guard.idx < len { + let ch = + // SAFETY: `guard.idx` is positive-or-zero and less that len so the `get_unchecked` + // is in bound. `self` is valid UTF-8 like string and the returned slice starts at + // a unicode code point so the `Chars` always return one character. + unsafe { guard.s.get_unchecked(guard.idx..len).chars().next().unwrap_unchecked() }; + let ch_len = ch.len_utf8(); + + if !f(ch) { + guard.del_bytes += ch_len; + } else if guard.del_bytes > 0 { + // SAFETY: `guard.idx` is in bound and `guard.del_bytes` represent the number of + // bytes that are erased from the string so the resulting `guard.idx - + // guard.del_bytes` always represent a valid unicode code point. + // + // `guard.del_bytes` >= `ch.len_utf8()`, so taking a slice with `ch.len_utf8()` len + // is safe. + ch.encode_utf8(unsafe { + crate::slice::from_raw_parts_mut( + guard.s.as_mut_ptr().add(guard.idx - guard.del_bytes), + ch.len_utf8(), + ) + }); + } + + // Point idx to the next char + guard.idx += ch_len; + } + + drop(guard); + } + + /// Inserts a character into this `String` at a byte position. + /// + /// This is an *O*(*n*) operation as it requires copying every element in the + /// buffer. + /// + /// # Panics + /// + /// Panics if `idx` is larger than the `String`'s length, or if it does not + /// lie on a [`char`] boundary. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::with_capacity(3); + /// + /// s.insert(0, 'f'); + /// s.insert(1, 'o'); + /// s.insert(2, 'o'); + /// + /// assert_eq!("foo", s); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(&mut self, idx: usize, ch: char) { + assert!(self.is_char_boundary(idx)); + let mut bits = [0; 4]; + let bits = ch.encode_utf8(&mut bits).as_bytes(); + + unsafe { + self.insert_bytes(idx, bits); + } + } + + #[cfg(not(no_global_oom_handling))] + unsafe fn insert_bytes(&mut self, idx: usize, bytes: &[u8]) { + let len = self.len(); + let amt = bytes.len(); + self.vec.reserve(amt); + + unsafe { + ptr::copy(self.vec.as_ptr().add(idx), self.vec.as_mut_ptr().add(idx + amt), len - idx); + ptr::copy_nonoverlapping(bytes.as_ptr(), self.vec.as_mut_ptr().add(idx), amt); + self.vec.set_len(len + amt); + } + } + + /// Inserts a string slice into this `String` at a byte position. + /// + /// This is an *O*(*n*) operation as it requires copying every element in the + /// buffer. + /// + /// # Panics + /// + /// Panics if `idx` is larger than the `String`'s length, or if it does not + /// lie on a [`char`] boundary. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("bar"); + /// + /// s.insert_str(0, "foo"); + /// + /// assert_eq!("foobar", s); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "insert_str", since = "1.16.0")] + pub fn insert_str(&mut self, idx: usize, string: &str) { + assert!(self.is_char_boundary(idx)); + + unsafe { + self.insert_bytes(idx, string.as_bytes()); + } + } + + /// Returns a mutable reference to the contents of this `String`. + /// + /// # Safety + /// + /// This function is unsafe because the returned `&mut Vec` allows writing + /// bytes which are not valid UTF-8. If this constraint is violated, using + /// the original `String` after dropping the `&mut Vec` may violate memory + /// safety, as the rest of the standard library assumes that `String`s are + /// valid UTF-8. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("hello"); + /// + /// unsafe { + /// let vec = s.as_mut_vec(); + /// assert_eq!(&[104, 101, 108, 108, 111][..], &vec[..]); + /// + /// vec.reverse(); + /// } + /// assert_eq!(s, "olleh"); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn as_mut_vec(&mut self) -> &mut Vec<u8> { + &mut self.vec + } + + /// Returns the length of this `String`, in bytes, not [`char`]s or + /// graphemes. In other words, it might not be what a human considers the + /// length of the string. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let a = String::from("foo"); + /// assert_eq!(a.len(), 3); + /// + /// let fancy_f = String::from("ƒoo"); + /// assert_eq!(fancy_f.len(), 4); + /// assert_eq!(fancy_f.chars().count(), 3); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn len(&self) -> usize { + self.vec.len() + } + + /// Returns `true` if this `String` has a length of zero, and `false` otherwise. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut v = String::new(); + /// assert!(v.is_empty()); + /// + /// v.push('a'); + /// assert!(!v.is_empty()); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_empty(&self) -> bool { + self.len() == 0 + } + + /// Splits the string into two at the given byte index. + /// + /// Returns a newly allocated `String`. `self` contains bytes `[0, at)`, and + /// the returned `String` contains bytes `[at, len)`. `at` must be on the + /// boundary of a UTF-8 code point. + /// + /// Note that the capacity of `self` does not change. + /// + /// # Panics + /// + /// Panics if `at` is not on a `UTF-8` code point boundary, or if it is beyond the last + /// code point of the string. + /// + /// # Examples + /// + /// ``` + /// # fn main() { + /// let mut hello = String::from("Hello, World!"); + /// let world = hello.split_off(7); + /// assert_eq!(hello, "Hello, "); + /// assert_eq!(world, "World!"); + /// # } + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "string_split_off", since = "1.16.0")] + #[must_use = "use `.truncate()` if you don't need the other half"] + pub fn split_off(&mut self, at: usize) -> String { + assert!(self.is_char_boundary(at)); + let other = self.vec.split_off(at); + unsafe { String::from_utf8_unchecked(other) } + } + + /// Truncates this `String`, removing all contents. + /// + /// While this means the `String` will have a length of zero, it does not + /// touch its capacity. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("foo"); + /// + /// s.clear(); + /// + /// assert!(s.is_empty()); + /// assert_eq!(0, s.len()); + /// assert_eq!(3, s.capacity()); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) { + self.vec.clear() + } + + /// Removes the specified range from the string in bulk, returning all + /// removed characters as an iterator. + /// + /// The returned iterator keeps a mutable borrow on the string to optimize + /// its implementation. + /// + /// # Panics + /// + /// Panics if the starting point or end point do not lie on a [`char`] + /// boundary, or if they're out of bounds. + /// + /// # Leaking + /// + /// If the returned iterator goes out of scope without being dropped (due to + /// [`core::mem::forget`], for example), the string may still contain a copy + /// of any drained characters, or may have lost characters arbitrarily, + /// including characters outside the range. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("α is alpha, β is beta"); + /// let beta_offset = s.find('β').unwrap_or(s.len()); + /// + /// // Remove the range up until the β from the string + /// let t: String = s.drain(..beta_offset).collect(); + /// assert_eq!(t, "α is alpha, "); + /// assert_eq!(s, "β is beta"); + /// + /// // A full range clears the string, like `clear()` does + /// s.drain(..); + /// assert_eq!(s, ""); + /// ``` + #[stable(feature = "drain", since = "1.6.0")] + pub fn drain<R>(&mut self, range: R) -> Drain<'_> + where + R: RangeBounds<usize>, + { + // Memory safety + // + // The String version of Drain does not have the memory safety issues + // of the vector version. The data is just plain bytes. + // Because the range removal happens in Drop, if the Drain iterator is leaked, + // the removal will not happen. + let Range { start, end } = slice::range(range, ..self.len()); + assert!(self.is_char_boundary(start)); + assert!(self.is_char_boundary(end)); + + // Take out two simultaneous borrows. The &mut String won't be accessed + // until iteration is over, in Drop. + let self_ptr = self as *mut _; + // SAFETY: `slice::range` and `is_char_boundary` do the appropriate bounds checks. + let chars_iter = unsafe { self.get_unchecked(start..end) }.chars(); + + Drain { start, end, iter: chars_iter, string: self_ptr } + } + + /// Removes the specified range in the string, + /// and replaces it with the given string. + /// The given string doesn't need to be the same length as the range. + /// + /// # Panics + /// + /// Panics if the starting point or end point do not lie on a [`char`] + /// boundary, or if they're out of bounds. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let mut s = String::from("α is alpha, β is beta"); + /// let beta_offset = s.find('β').unwrap_or(s.len()); + /// + /// // Replace the range up until the β from the string + /// s.replace_range(..beta_offset, "Α is capital alpha; "); + /// assert_eq!(s, "Α is capital alpha; β is beta"); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "splice", since = "1.27.0")] + pub fn replace_range<R>(&mut self, range: R, replace_with: &str) + where + R: RangeBounds<usize>, + { + // Memory safety + // + // Replace_range does not have the memory safety issues of a vector Splice. + // of the vector version. The data is just plain bytes. + + // WARNING: Inlining this variable would be unsound (#81138) + let start = range.start_bound(); + match start { + Included(&n) => assert!(self.is_char_boundary(n)), + Excluded(&n) => assert!(self.is_char_boundary(n + 1)), + Unbounded => {} + }; + // WARNING: Inlining this variable would be unsound (#81138) + let end = range.end_bound(); + match end { + Included(&n) => assert!(self.is_char_boundary(n + 1)), + Excluded(&n) => assert!(self.is_char_boundary(n)), + Unbounded => {} + }; + + // Using `range` again would be unsound (#81138) + // We assume the bounds reported by `range` remain the same, but + // an adversarial implementation could change between calls + unsafe { self.as_mut_vec() }.splice((start, end), replace_with.bytes()); + } + + /// Converts this `String` into a <code>[Box]<[str]></code>. + /// + /// This will drop any excess capacity. + /// + /// [str]: prim@str "str" + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s = String::from("hello"); + /// + /// let b = s.into_boxed_str(); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "box_str", since = "1.4.0")] + #[must_use = "`self` will be dropped if the result is not used"] + #[inline] + pub fn into_boxed_str(self) -> Box<str> { + let slice = self.vec.into_boxed_slice(); + unsafe { from_boxed_utf8_unchecked(slice) } + } +} + +impl FromUtf8Error { + /// Returns a slice of [`u8`]s bytes that were attempted to convert to a `String`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some invalid bytes, in a vector + /// let bytes = vec![0, 159]; + /// + /// let value = String::from_utf8(bytes); + /// + /// assert_eq!(&[0, 159], value.unwrap_err().as_bytes()); + /// ``` + #[must_use] + #[stable(feature = "from_utf8_error_as_bytes", since = "1.26.0")] + pub fn as_bytes(&self) -> &[u8] { + &self.bytes[..] + } + + /// Returns the bytes that were attempted to convert to a `String`. + /// + /// This method is carefully constructed to avoid allocation. It will + /// consume the error, moving out the bytes, so that a copy of the bytes + /// does not need to be made. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some invalid bytes, in a vector + /// let bytes = vec![0, 159]; + /// + /// let value = String::from_utf8(bytes); + /// + /// assert_eq!(vec![0, 159], value.unwrap_err().into_bytes()); + /// ``` + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_bytes(self) -> Vec<u8> { + self.bytes + } + + /// Fetch a `Utf8Error` to get more details about the conversion failure. + /// + /// The [`Utf8Error`] type provided by [`std::str`] represents an error that may + /// occur when converting a slice of [`u8`]s to a [`&str`]. In this sense, it's + /// an analogue to `FromUtf8Error`. See its documentation for more details + /// on using it. + /// + /// [`std::str`]: core::str "std::str" + /// [`&str`]: prim@str "&str" + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// // some invalid bytes, in a vector + /// let bytes = vec![0, 159]; + /// + /// let error = String::from_utf8(bytes).unwrap_err().utf8_error(); + /// + /// // the first byte is invalid here + /// assert_eq!(1, error.valid_up_to()); + /// ``` + #[must_use] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn utf8_error(&self) -> Utf8Error { + self.error + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Display for FromUtf8Error { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt(&self.error, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Display for FromUtf16Error { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt("invalid utf-16: lone surrogate found", f) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl Clone for String { + fn clone(&self) -> Self { + String { vec: self.vec.clone() } + } + + fn clone_from(&mut self, source: &Self) { + self.vec.clone_from(&source.vec); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl FromIterator<char> for String { + fn from_iter<I: IntoIterator<Item = char>>(iter: I) -> String { + let mut buf = String::new(); + buf.extend(iter); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "string_from_iter_by_ref", since = "1.17.0")] +impl<'a> FromIterator<&'a char> for String { + fn from_iter<I: IntoIterator<Item = &'a char>>(iter: I) -> String { + let mut buf = String::new(); + buf.extend(iter); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a> FromIterator<&'a str> for String { + fn from_iter<I: IntoIterator<Item = &'a str>>(iter: I) -> String { + let mut buf = String::new(); + buf.extend(iter); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "extend_string", since = "1.4.0")] +impl FromIterator<String> for String { + fn from_iter<I: IntoIterator<Item = String>>(iter: I) -> String { + let mut iterator = iter.into_iter(); + + // Because we're iterating over `String`s, we can avoid at least + // one allocation by getting the first string from the iterator + // and appending to it all the subsequent strings. + match iterator.next() { + None => String::new(), + Some(mut buf) => { + buf.extend(iterator); + buf + } + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_str2", since = "1.45.0")] +impl FromIterator<Box<str>> for String { + fn from_iter<I: IntoIterator<Item = Box<str>>>(iter: I) -> String { + let mut buf = String::new(); + buf.extend(iter); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "herd_cows", since = "1.19.0")] +impl<'a> FromIterator<Cow<'a, str>> for String { + fn from_iter<I: IntoIterator<Item = Cow<'a, str>>>(iter: I) -> String { + let mut iterator = iter.into_iter(); + + // Because we're iterating over CoWs, we can (potentially) avoid at least + // one allocation by getting the first item and appending to it all the + // subsequent items. + match iterator.next() { + None => String::new(), + Some(cow) => { + let mut buf = cow.into_owned(); + buf.extend(iterator); + buf + } + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl Extend<char> for String { + fn extend<I: IntoIterator<Item = char>>(&mut self, iter: I) { + let iterator = iter.into_iter(); + let (lower_bound, _) = iterator.size_hint(); + self.reserve(lower_bound); + iterator.for_each(move |c| self.push(c)); + } + + #[inline] + fn extend_one(&mut self, c: char) { + self.push(c); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a> Extend<&'a char> for String { + fn extend<I: IntoIterator<Item = &'a char>>(&mut self, iter: I) { + self.extend(iter.into_iter().cloned()); + } + + #[inline] + fn extend_one(&mut self, &c: &'a char) { + self.push(c); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a> Extend<&'a str> for String { + fn extend<I: IntoIterator<Item = &'a str>>(&mut self, iter: I) { + iter.into_iter().for_each(move |s| self.push_str(s)); + } + + #[inline] + fn extend_one(&mut self, s: &'a str) { + self.push_str(s); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_str2", since = "1.45.0")] +impl Extend<Box<str>> for String { + fn extend<I: IntoIterator<Item = Box<str>>>(&mut self, iter: I) { + iter.into_iter().for_each(move |s| self.push_str(&s)); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "extend_string", since = "1.4.0")] +impl Extend<String> for String { + fn extend<I: IntoIterator<Item = String>>(&mut self, iter: I) { + iter.into_iter().for_each(move |s| self.push_str(&s)); + } + + #[inline] + fn extend_one(&mut self, s: String) { + self.push_str(&s); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "herd_cows", since = "1.19.0")] +impl<'a> Extend<Cow<'a, str>> for String { + fn extend<I: IntoIterator<Item = Cow<'a, str>>>(&mut self, iter: I) { + iter.into_iter().for_each(move |s| self.push_str(&s)); + } + + #[inline] + fn extend_one(&mut self, s: Cow<'a, str>) { + self.push_str(&s); + } +} + +/// A convenience impl that delegates to the impl for `&str`. +/// +/// # Examples +/// +/// ``` +/// assert_eq!(String::from("Hello world").find("world"), Some(6)); +/// ``` +#[unstable( + feature = "pattern", + reason = "API not fully fleshed out and ready to be stabilized", + issue = "27721" +)] +impl<'a, 'b> Pattern<'a> for &'b String { + type Searcher = <&'b str as Pattern<'a>>::Searcher; + + fn into_searcher(self, haystack: &'a str) -> <&'b str as Pattern<'a>>::Searcher { + self[..].into_searcher(haystack) + } + + #[inline] + fn is_contained_in(self, haystack: &'a str) -> bool { + self[..].is_contained_in(haystack) + } + + #[inline] + fn is_prefix_of(self, haystack: &'a str) -> bool { + self[..].is_prefix_of(haystack) + } + + #[inline] + fn strip_prefix_of(self, haystack: &'a str) -> Option<&'a str> { + self[..].strip_prefix_of(haystack) + } + + #[inline] + fn is_suffix_of(self, haystack: &'a str) -> bool { + self[..].is_suffix_of(haystack) + } + + #[inline] + fn strip_suffix_of(self, haystack: &'a str) -> Option<&'a str> { + self[..].strip_suffix_of(haystack) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl PartialEq for String { + #[inline] + fn eq(&self, other: &String) -> bool { + PartialEq::eq(&self[..], &other[..]) + } + #[inline] + fn ne(&self, other: &String) -> bool { + PartialEq::ne(&self[..], &other[..]) + } +} + +macro_rules! impl_eq { + ($lhs:ty, $rhs: ty) => { + #[stable(feature = "rust1", since = "1.0.0")] + #[allow(unused_lifetimes)] + impl<'a, 'b> PartialEq<$rhs> for $lhs { + #[inline] + fn eq(&self, other: &$rhs) -> bool { + PartialEq::eq(&self[..], &other[..]) + } + #[inline] + fn ne(&self, other: &$rhs) -> bool { + PartialEq::ne(&self[..], &other[..]) + } + } + + #[stable(feature = "rust1", since = "1.0.0")] + #[allow(unused_lifetimes)] + impl<'a, 'b> PartialEq<$lhs> for $rhs { + #[inline] + fn eq(&self, other: &$lhs) -> bool { + PartialEq::eq(&self[..], &other[..]) + } + #[inline] + fn ne(&self, other: &$lhs) -> bool { + PartialEq::ne(&self[..], &other[..]) + } + } + }; +} + +impl_eq! { String, str } +impl_eq! { String, &'a str } +#[cfg(not(no_global_oom_handling))] +impl_eq! { Cow<'a, str>, str } +#[cfg(not(no_global_oom_handling))] +impl_eq! { Cow<'a, str>, &'b str } +#[cfg(not(no_global_oom_handling))] +impl_eq! { Cow<'a, str>, String } + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_default_impls", issue = "87864")] +impl const Default for String { + /// Creates an empty `String`. + #[inline] + fn default() -> String { + String::new() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Display for String { + #[inline] + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Debug for String { + #[inline] + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl hash::Hash for String { + #[inline] + fn hash<H: hash::Hasher>(&self, hasher: &mut H) { + (**self).hash(hasher) + } +} + +/// Implements the `+` operator for concatenating two strings. +/// +/// This consumes the `String` on the left-hand side and re-uses its buffer (growing it if +/// necessary). This is done to avoid allocating a new `String` and copying the entire contents on +/// every operation, which would lead to *O*(*n*^2) running time when building an *n*-byte string by +/// repeated concatenation. +/// +/// The string on the right-hand side is only borrowed; its contents are copied into the returned +/// `String`. +/// +/// # Examples +/// +/// Concatenating two `String`s takes the first by value and borrows the second: +/// +/// ``` +/// let a = String::from("hello"); +/// let b = String::from(" world"); +/// let c = a + &b; +/// // `a` is moved and can no longer be used here. +/// ``` +/// +/// If you want to keep using the first `String`, you can clone it and append to the clone instead: +/// +/// ``` +/// let a = String::from("hello"); +/// let b = String::from(" world"); +/// let c = a.clone() + &b; +/// // `a` is still valid here. +/// ``` +/// +/// Concatenating `&str` slices can be done by converting the first to a `String`: +/// +/// ``` +/// let a = "hello"; +/// let b = " world"; +/// let c = a.to_string() + b; +/// ``` +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl Add<&str> for String { + type Output = String; + + #[inline] + fn add(mut self, other: &str) -> String { + self.push_str(other); + self + } +} + +/// Implements the `+=` operator for appending to a `String`. +/// +/// This has the same behavior as the [`push_str`][String::push_str] method. +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "stringaddassign", since = "1.12.0")] +impl AddAssign<&str> for String { + #[inline] + fn add_assign(&mut self, other: &str) { + self.push_str(other); + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Index<ops::Range<usize>> for String { + type Output = str; + + #[inline] + fn index(&self, index: ops::Range<usize>) -> &str { + &self[..][index] + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Index<ops::RangeTo<usize>> for String { + type Output = str; + + #[inline] + fn index(&self, index: ops::RangeTo<usize>) -> &str { + &self[..][index] + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Index<ops::RangeFrom<usize>> for String { + type Output = str; + + #[inline] + fn index(&self, index: ops::RangeFrom<usize>) -> &str { + &self[..][index] + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Index<ops::RangeFull> for String { + type Output = str; + + #[inline] + fn index(&self, _index: ops::RangeFull) -> &str { + unsafe { str::from_utf8_unchecked(&self.vec) } + } +} +#[stable(feature = "inclusive_range", since = "1.26.0")] +impl ops::Index<ops::RangeInclusive<usize>> for String { + type Output = str; + + #[inline] + fn index(&self, index: ops::RangeInclusive<usize>) -> &str { + Index::index(&**self, index) + } +} +#[stable(feature = "inclusive_range", since = "1.26.0")] +impl ops::Index<ops::RangeToInclusive<usize>> for String { + type Output = str; + + #[inline] + fn index(&self, index: ops::RangeToInclusive<usize>) -> &str { + Index::index(&**self, index) + } +} + +#[stable(feature = "derefmut_for_string", since = "1.3.0")] +impl ops::IndexMut<ops::Range<usize>> for String { + #[inline] + fn index_mut(&mut self, index: ops::Range<usize>) -> &mut str { + &mut self[..][index] + } +} +#[stable(feature = "derefmut_for_string", since = "1.3.0")] +impl ops::IndexMut<ops::RangeTo<usize>> for String { + #[inline] + fn index_mut(&mut self, index: ops::RangeTo<usize>) -> &mut str { + &mut self[..][index] + } +} +#[stable(feature = "derefmut_for_string", since = "1.3.0")] +impl ops::IndexMut<ops::RangeFrom<usize>> for String { + #[inline] + fn index_mut(&mut self, index: ops::RangeFrom<usize>) -> &mut str { + &mut self[..][index] + } +} +#[stable(feature = "derefmut_for_string", since = "1.3.0")] +impl ops::IndexMut<ops::RangeFull> for String { + #[inline] + fn index_mut(&mut self, _index: ops::RangeFull) -> &mut str { + unsafe { str::from_utf8_unchecked_mut(&mut *self.vec) } + } +} +#[stable(feature = "inclusive_range", since = "1.26.0")] +impl ops::IndexMut<ops::RangeInclusive<usize>> for String { + #[inline] + fn index_mut(&mut self, index: ops::RangeInclusive<usize>) -> &mut str { + IndexMut::index_mut(&mut **self, index) + } +} +#[stable(feature = "inclusive_range", since = "1.26.0")] +impl ops::IndexMut<ops::RangeToInclusive<usize>> for String { + #[inline] + fn index_mut(&mut self, index: ops::RangeToInclusive<usize>) -> &mut str { + IndexMut::index_mut(&mut **self, index) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl ops::Deref for String { + type Target = str; + + #[inline] + fn deref(&self) -> &str { + unsafe { str::from_utf8_unchecked(&self.vec) } + } +} + +#[stable(feature = "derefmut_for_string", since = "1.3.0")] +impl ops::DerefMut for String { + #[inline] + fn deref_mut(&mut self) -> &mut str { + unsafe { str::from_utf8_unchecked_mut(&mut *self.vec) } + } +} + +/// A type alias for [`Infallible`]. +/// +/// This alias exists for backwards compatibility, and may be eventually deprecated. +/// +/// [`Infallible`]: core::convert::Infallible "convert::Infallible" +#[stable(feature = "str_parse_error", since = "1.5.0")] +pub type ParseError = core::convert::Infallible; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl FromStr for String { + type Err = core::convert::Infallible; + #[inline] + fn from_str(s: &str) -> Result<String, Self::Err> { + Ok(String::from(s)) + } +} + +/// A trait for converting a value to a `String`. +/// +/// This trait is automatically implemented for any type which implements the +/// [`Display`] trait. As such, `ToString` shouldn't be implemented directly: +/// [`Display`] should be implemented instead, and you get the `ToString` +/// implementation for free. +/// +/// [`Display`]: fmt::Display +#[cfg_attr(not(test), rustc_diagnostic_item = "ToString")] +#[stable(feature = "rust1", since = "1.0.0")] +pub trait ToString { + /// Converts the given value to a `String`. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let i = 5; + /// let five = String::from("5"); + /// + /// assert_eq!(five, i.to_string()); + /// ``` + #[rustc_conversion_suggestion] + #[stable(feature = "rust1", since = "1.0.0")] + fn to_string(&self) -> String; +} + +/// # Panics +/// +/// In this implementation, the `to_string` method panics +/// if the `Display` implementation returns an error. +/// This indicates an incorrect `Display` implementation +/// since `fmt::Write for String` never returns an error itself. +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Display + ?Sized> ToString for T { + // A common guideline is to not inline generic functions. However, + // removing `#[inline]` from this method causes non-negligible regressions. + // See <https://github.com/rust-lang/rust/pull/74852>, the last attempt + // to try to remove it. + #[inline] + default fn to_string(&self) -> String { + let mut buf = String::new(); + let mut formatter = core::fmt::Formatter::new(&mut buf); + // Bypass format_args!() to avoid write_str with zero-length strs + fmt::Display::fmt(self, &mut formatter) + .expect("a Display implementation returned an error unexpectedly"); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "char_to_string_specialization", since = "1.46.0")] +impl ToString for char { + #[inline] + fn to_string(&self) -> String { + String::from(self.encode_utf8(&mut [0; 4])) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "u8_to_string_specialization", since = "1.54.0")] +impl ToString for u8 { + #[inline] + fn to_string(&self) -> String { + let mut buf = String::with_capacity(3); + let mut n = *self; + if n >= 10 { + if n >= 100 { + buf.push((b'0' + n / 100) as char); + n %= 100; + } + buf.push((b'0' + n / 10) as char); + n %= 10; + } + buf.push((b'0' + n) as char); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "i8_to_string_specialization", since = "1.54.0")] +impl ToString for i8 { + #[inline] + fn to_string(&self) -> String { + let mut buf = String::with_capacity(4); + if self.is_negative() { + buf.push('-'); + } + let mut n = self.unsigned_abs(); + if n >= 10 { + if n >= 100 { + buf.push('1'); + n -= 100; + } + buf.push((b'0' + n / 10) as char); + n %= 10; + } + buf.push((b'0' + n) as char); + buf + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "str_to_string_specialization", since = "1.9.0")] +impl ToString for str { + #[inline] + fn to_string(&self) -> String { + String::from(self) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_str_to_string_specialization", since = "1.17.0")] +impl ToString for Cow<'_, str> { + #[inline] + fn to_string(&self) -> String { + self[..].to_owned() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "string_to_string_specialization", since = "1.17.0")] +impl ToString for String { + #[inline] + fn to_string(&self) -> String { + self.to_owned() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl AsRef<str> for String { + #[inline] + fn as_ref(&self) -> &str { + self + } +} + +#[stable(feature = "string_as_mut", since = "1.43.0")] +impl AsMut<str> for String { + #[inline] + fn as_mut(&mut self) -> &mut str { + self + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl AsRef<[u8]> for String { + #[inline] + fn as_ref(&self) -> &[u8] { + self.as_bytes() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl From<&str> for String { + /// Converts a `&str` into a [`String`]. + /// + /// The result is allocated on the heap. + #[inline] + fn from(s: &str) -> String { + s.to_owned() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_mut_str_for_string", since = "1.44.0")] +impl From<&mut str> for String { + /// Converts a `&mut str` into a [`String`]. + /// + /// The result is allocated on the heap. + #[inline] + fn from(s: &mut str) -> String { + s.to_owned() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_ref_string", since = "1.35.0")] +impl From<&String> for String { + /// Converts a `&String` into a [`String`]. + /// + /// This clones `s` and returns the clone. + #[inline] + fn from(s: &String) -> String { + s.clone() + } +} + +// note: test pulls in libstd, which causes errors here +#[cfg(not(test))] +#[stable(feature = "string_from_box", since = "1.18.0")] +impl From<Box<str>> for String { + /// Converts the given boxed `str` slice to a [`String`]. + /// It is notable that the `str` slice is owned. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s1: String = String::from("hello world"); + /// let s2: Box<str> = s1.into_boxed_str(); + /// let s3: String = String::from(s2); + /// + /// assert_eq!("hello world", s3) + /// ``` + fn from(s: Box<str>) -> String { + s.into_string() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "box_from_str", since = "1.20.0")] +impl From<String> for Box<str> { + /// Converts the given [`String`] to a boxed `str` slice that is owned. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s1: String = String::from("hello world"); + /// let s2: Box<str> = Box::from(s1); + /// let s3: String = String::from(s2); + /// + /// assert_eq!("hello world", s3) + /// ``` + fn from(s: String) -> Box<str> { + s.into_boxed_str() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "string_from_cow_str", since = "1.14.0")] +impl<'a> From<Cow<'a, str>> for String { + /// Converts a clone-on-write string to an owned + /// instance of [`String`]. + /// + /// This extracts the owned string, + /// clones the string if it is not already owned. + /// + /// # Example + /// + /// ``` + /// # use std::borrow::Cow; + /// // If the string is not owned... + /// let cow: Cow<str> = Cow::Borrowed("eggplant"); + /// // It will allocate on the heap and copy the string. + /// let owned: String = String::from(cow); + /// assert_eq!(&owned[..], "eggplant"); + /// ``` + fn from(s: Cow<'a, str>) -> String { + s.into_owned() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a> From<&'a str> for Cow<'a, str> { + /// Converts a string slice into a [`Borrowed`] variant. + /// No heap allocation is performed, and the string + /// is not copied. + /// + /// # Example + /// + /// ``` + /// # use std::borrow::Cow; + /// assert_eq!(Cow::from("eggplant"), Cow::Borrowed("eggplant")); + /// ``` + /// + /// [`Borrowed`]: crate::borrow::Cow::Borrowed "borrow::Cow::Borrowed" + #[inline] + fn from(s: &'a str) -> Cow<'a, str> { + Cow::Borrowed(s) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a> From<String> for Cow<'a, str> { + /// Converts a [`String`] into an [`Owned`] variant. + /// No heap allocation is performed, and the string + /// is not copied. + /// + /// # Example + /// + /// ``` + /// # use std::borrow::Cow; + /// let s = "eggplant".to_string(); + /// let s2 = "eggplant".to_string(); + /// assert_eq!(Cow::from(s), Cow::<'static, str>::Owned(s2)); + /// ``` + /// + /// [`Owned`]: crate::borrow::Cow::Owned "borrow::Cow::Owned" + #[inline] + fn from(s: String) -> Cow<'a, str> { + Cow::Owned(s) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_from_string_ref", since = "1.28.0")] +impl<'a> From<&'a String> for Cow<'a, str> { + /// Converts a [`String`] reference into a [`Borrowed`] variant. + /// No heap allocation is performed, and the string + /// is not copied. + /// + /// # Example + /// + /// ``` + /// # use std::borrow::Cow; + /// let s = "eggplant".to_string(); + /// assert_eq!(Cow::from(&s), Cow::Borrowed("eggplant")); + /// ``` + /// + /// [`Borrowed`]: crate::borrow::Cow::Borrowed "borrow::Cow::Borrowed" + #[inline] + fn from(s: &'a String) -> Cow<'a, str> { + Cow::Borrowed(s.as_str()) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_str_from_iter", since = "1.12.0")] +impl<'a> FromIterator<char> for Cow<'a, str> { + fn from_iter<I: IntoIterator<Item = char>>(it: I) -> Cow<'a, str> { + Cow::Owned(FromIterator::from_iter(it)) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_str_from_iter", since = "1.12.0")] +impl<'a, 'b> FromIterator<&'b str> for Cow<'a, str> { + fn from_iter<I: IntoIterator<Item = &'b str>>(it: I) -> Cow<'a, str> { + Cow::Owned(FromIterator::from_iter(it)) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "cow_str_from_iter", since = "1.12.0")] +impl<'a> FromIterator<String> for Cow<'a, str> { + fn from_iter<I: IntoIterator<Item = String>>(it: I) -> Cow<'a, str> { + Cow::Owned(FromIterator::from_iter(it)) + } +} + +#[stable(feature = "from_string_for_vec_u8", since = "1.14.0")] +impl From<String> for Vec<u8> { + /// Converts the given [`String`] to a vector [`Vec`] that holds values of type [`u8`]. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// let s1 = String::from("hello world"); + /// let v1 = Vec::from(s1); + /// + /// for b in v1 { + /// println!("{b}"); + /// } + /// ``` + fn from(string: String) -> Vec<u8> { + string.into_bytes() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl fmt::Write for String { + #[inline] + fn write_str(&mut self, s: &str) -> fmt::Result { + self.push_str(s); + Ok(()) + } + + #[inline] + fn write_char(&mut self, c: char) -> fmt::Result { + self.push(c); + Ok(()) + } +} + +/// A draining iterator for `String`. +/// +/// This struct is created by the [`drain`] method on [`String`]. See its +/// documentation for more. +/// +/// [`drain`]: String::drain +#[stable(feature = "drain", since = "1.6.0")] +pub struct Drain<'a> { + /// Will be used as &'a mut String in the destructor + string: *mut String, + /// Start of part to remove + start: usize, + /// End of part to remove + end: usize, + /// Current remaining range to remove + iter: Chars<'a>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl fmt::Debug for Drain<'_> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Drain").field(&self.as_str()).finish() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl Sync for Drain<'_> {} +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl Send for Drain<'_> {} + +#[stable(feature = "drain", since = "1.6.0")] +impl Drop for Drain<'_> { + fn drop(&mut self) { + unsafe { + // Use Vec::drain. "Reaffirm" the bounds checks to avoid + // panic code being inserted again. + let self_vec = (*self.string).as_mut_vec(); + if self.start <= self.end && self.end <= self_vec.len() { + self_vec.drain(self.start..self.end); + } + } + } +} + +impl<'a> Drain<'a> { + /// Returns the remaining (sub)string of this iterator as a slice. + /// + /// # Examples + /// + /// ``` + /// let mut s = String::from("abc"); + /// let mut drain = s.drain(..); + /// assert_eq!(drain.as_str(), "abc"); + /// let _ = drain.next().unwrap(); + /// assert_eq!(drain.as_str(), "bc"); + /// ``` + #[must_use] + #[stable(feature = "string_drain_as_str", since = "1.55.0")] + pub fn as_str(&self) -> &str { + self.iter.as_str() + } +} + +#[stable(feature = "string_drain_as_str", since = "1.55.0")] +impl<'a> AsRef<str> for Drain<'a> { + fn as_ref(&self) -> &str { + self.as_str() + } +} + +#[stable(feature = "string_drain_as_str", since = "1.55.0")] +impl<'a> AsRef<[u8]> for Drain<'a> { + fn as_ref(&self) -> &[u8] { + self.as_str().as_bytes() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl Iterator for Drain<'_> { + type Item = char; + + #[inline] + fn next(&mut self) -> Option<char> { + self.iter.next() + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } + + #[inline] + fn last(mut self) -> Option<char> { + self.next_back() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl DoubleEndedIterator for Drain<'_> { + #[inline] + fn next_back(&mut self) -> Option<char> { + self.iter.next_back() + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl FusedIterator for Drain<'_> {} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_char_for_string", since = "1.46.0")] +impl From<char> for String { + /// Allocates an owned [`String`] from a single character. + /// + /// # Example + /// ```rust + /// let c: char = 'a'; + /// let s: String = String::from(c); + /// assert_eq!("a", &s[..]); + /// ``` + #[inline] + fn from(c: char) -> Self { + c.to_string() + } +} diff --git a/library/alloc/src/sync.rs b/library/alloc/src/sync.rs new file mode 100644 index 000000000..4c03cc3ed --- /dev/null +++ b/library/alloc/src/sync.rs @@ -0,0 +1,2765 @@ +#![stable(feature = "rust1", since = "1.0.0")] + +//! Thread-safe reference-counting pointers. +//! +//! See the [`Arc<T>`][Arc] documentation for more details. + +use core::any::Any; +use core::borrow; +use core::cmp::Ordering; +use core::convert::{From, TryFrom}; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::hint; +use core::intrinsics::abort; +#[cfg(not(no_global_oom_handling))] +use core::iter; +use core::marker::{PhantomData, Unpin, Unsize}; +#[cfg(not(no_global_oom_handling))] +use core::mem::size_of_val; +use core::mem::{self, align_of_val_raw}; +use core::ops::{CoerceUnsized, Deref, DispatchFromDyn, Receiver}; +use core::panic::{RefUnwindSafe, UnwindSafe}; +use core::pin::Pin; +use core::ptr::{self, NonNull}; +#[cfg(not(no_global_oom_handling))] +use core::slice::from_raw_parts_mut; +use core::sync::atomic; +use core::sync::atomic::Ordering::{Acquire, Relaxed, Release}; + +#[cfg(not(no_global_oom_handling))] +use crate::alloc::handle_alloc_error; +#[cfg(not(no_global_oom_handling))] +use crate::alloc::{box_free, WriteCloneIntoRaw}; +use crate::alloc::{AllocError, Allocator, Global, Layout}; +use crate::borrow::{Cow, ToOwned}; +use crate::boxed::Box; +use crate::rc::is_dangling; +#[cfg(not(no_global_oom_handling))] +use crate::string::String; +#[cfg(not(no_global_oom_handling))] +use crate::vec::Vec; + +#[cfg(test)] +mod tests; + +/// A soft limit on the amount of references that may be made to an `Arc`. +/// +/// Going above this limit will abort your program (although not +/// necessarily) at _exactly_ `MAX_REFCOUNT + 1` references. +const MAX_REFCOUNT: usize = (isize::MAX) as usize; + +#[cfg(not(sanitize = "thread"))] +macro_rules! acquire { + ($x:expr) => { + atomic::fence(Acquire) + }; +} + +// ThreadSanitizer does not support memory fences. To avoid false positive +// reports in Arc / Weak implementation use atomic loads for synchronization +// instead. +#[cfg(sanitize = "thread")] +macro_rules! acquire { + ($x:expr) => { + $x.load(Acquire) + }; +} + +/// A thread-safe reference-counting pointer. 'Arc' stands for 'Atomically +/// Reference Counted'. +/// +/// The type `Arc<T>` provides shared ownership of a value of type `T`, +/// allocated in the heap. Invoking [`clone`][clone] on `Arc` produces +/// a new `Arc` instance, which points to the same allocation on the heap as the +/// source `Arc`, while increasing a reference count. When the last `Arc` +/// pointer to a given allocation is destroyed, the value stored in that allocation (often +/// referred to as "inner value") is also dropped. +/// +/// Shared references in Rust disallow mutation by default, and `Arc` is no +/// exception: you cannot generally obtain a mutable reference to something +/// inside an `Arc`. If you need to mutate through an `Arc`, use +/// [`Mutex`][mutex], [`RwLock`][rwlock], or one of the [`Atomic`][atomic] +/// types. +/// +/// ## Thread Safety +/// +/// Unlike [`Rc<T>`], `Arc<T>` uses atomic operations for its reference +/// counting. This means that it is thread-safe. The disadvantage is that +/// atomic operations are more expensive than ordinary memory accesses. If you +/// are not sharing reference-counted allocations between threads, consider using +/// [`Rc<T>`] for lower overhead. [`Rc<T>`] is a safe default, because the +/// compiler will catch any attempt to send an [`Rc<T>`] between threads. +/// However, a library might choose `Arc<T>` in order to give library consumers +/// more flexibility. +/// +/// `Arc<T>` will implement [`Send`] and [`Sync`] as long as the `T` implements +/// [`Send`] and [`Sync`]. Why can't you put a non-thread-safe type `T` in an +/// `Arc<T>` to make it thread-safe? This may be a bit counter-intuitive at +/// first: after all, isn't the point of `Arc<T>` thread safety? The key is +/// this: `Arc<T>` makes it thread safe to have multiple ownership of the same +/// data, but it doesn't add thread safety to its data. Consider +/// <code>Arc<[RefCell\<T>]></code>. [`RefCell<T>`] isn't [`Sync`], and if `Arc<T>` was always +/// [`Send`], <code>Arc<[RefCell\<T>]></code> would be as well. But then we'd have a problem: +/// [`RefCell<T>`] is not thread safe; it keeps track of the borrowing count using +/// non-atomic operations. +/// +/// In the end, this means that you may need to pair `Arc<T>` with some sort of +/// [`std::sync`] type, usually [`Mutex<T>`][mutex]. +/// +/// ## Breaking cycles with `Weak` +/// +/// The [`downgrade`][downgrade] method can be used to create a non-owning +/// [`Weak`] pointer. A [`Weak`] pointer can be [`upgrade`][upgrade]d +/// to an `Arc`, but this will return [`None`] if the value stored in the allocation has +/// already been dropped. In other words, `Weak` pointers do not keep the value +/// inside the allocation alive; however, they *do* keep the allocation +/// (the backing store for the value) alive. +/// +/// A cycle between `Arc` pointers will never be deallocated. For this reason, +/// [`Weak`] is used to break cycles. For example, a tree could have +/// strong `Arc` pointers from parent nodes to children, and [`Weak`] +/// pointers from children back to their parents. +/// +/// # Cloning references +/// +/// Creating a new reference from an existing reference-counted pointer is done using the +/// `Clone` trait implemented for [`Arc<T>`][Arc] and [`Weak<T>`][Weak]. +/// +/// ``` +/// use std::sync::Arc; +/// let foo = Arc::new(vec![1.0, 2.0, 3.0]); +/// // The two syntaxes below are equivalent. +/// let a = foo.clone(); +/// let b = Arc::clone(&foo); +/// // a, b, and foo are all Arcs that point to the same memory location +/// ``` +/// +/// ## `Deref` behavior +/// +/// `Arc<T>` automatically dereferences to `T` (via the [`Deref`][deref] trait), +/// so you can call `T`'s methods on a value of type `Arc<T>`. To avoid name +/// clashes with `T`'s methods, the methods of `Arc<T>` itself are associated +/// functions, called using [fully qualified syntax]: +/// +/// ``` +/// use std::sync::Arc; +/// +/// let my_arc = Arc::new(()); +/// let my_weak = Arc::downgrade(&my_arc); +/// ``` +/// +/// `Arc<T>`'s implementations of traits like `Clone` may also be called using +/// fully qualified syntax. Some people prefer to use fully qualified syntax, +/// while others prefer using method-call syntax. +/// +/// ``` +/// use std::sync::Arc; +/// +/// let arc = Arc::new(()); +/// // Method-call syntax +/// let arc2 = arc.clone(); +/// // Fully qualified syntax +/// let arc3 = Arc::clone(&arc); +/// ``` +/// +/// [`Weak<T>`][Weak] does not auto-dereference to `T`, because the inner value may have +/// already been dropped. +/// +/// [`Rc<T>`]: crate::rc::Rc +/// [clone]: Clone::clone +/// [mutex]: ../../std/sync/struct.Mutex.html +/// [rwlock]: ../../std/sync/struct.RwLock.html +/// [atomic]: core::sync::atomic +/// [`Send`]: core::marker::Send +/// [`Sync`]: core::marker::Sync +/// [deref]: core::ops::Deref +/// [downgrade]: Arc::downgrade +/// [upgrade]: Weak::upgrade +/// [RefCell\<T>]: core::cell::RefCell +/// [`RefCell<T>`]: core::cell::RefCell +/// [`std::sync`]: ../../std/sync/index.html +/// [`Arc::clone(&from)`]: Arc::clone +/// [fully qualified syntax]: https://doc.rust-lang.org/book/ch19-03-advanced-traits.html#fully-qualified-syntax-for-disambiguation-calling-methods-with-the-same-name +/// +/// # Examples +/// +/// Sharing some immutable data between threads: +/// +// Note that we **do not** run these tests here. The windows builders get super +// unhappy if a thread outlives the main thread and then exits at the same time +// (something deadlocks) so we just avoid this entirely by not running these +// tests. +/// ```no_run +/// use std::sync::Arc; +/// use std::thread; +/// +/// let five = Arc::new(5); +/// +/// for _ in 0..10 { +/// let five = Arc::clone(&five); +/// +/// thread::spawn(move || { +/// println!("{five:?}"); +/// }); +/// } +/// ``` +/// +/// Sharing a mutable [`AtomicUsize`]: +/// +/// [`AtomicUsize`]: core::sync::atomic::AtomicUsize "sync::atomic::AtomicUsize" +/// +/// ```no_run +/// use std::sync::Arc; +/// use std::sync::atomic::{AtomicUsize, Ordering}; +/// use std::thread; +/// +/// let val = Arc::new(AtomicUsize::new(5)); +/// +/// for _ in 0..10 { +/// let val = Arc::clone(&val); +/// +/// thread::spawn(move || { +/// let v = val.fetch_add(1, Ordering::SeqCst); +/// println!("{v:?}"); +/// }); +/// } +/// ``` +/// +/// See the [`rc` documentation][rc_examples] for more examples of reference +/// counting in general. +/// +/// [rc_examples]: crate::rc#examples +#[cfg_attr(not(test), rustc_diagnostic_item = "Arc")] +#[stable(feature = "rust1", since = "1.0.0")] +pub struct Arc<T: ?Sized> { + ptr: NonNull<ArcInner<T>>, + phantom: PhantomData<ArcInner<T>>, +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: ?Sized + Sync + Send> Send for Arc<T> {} +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: ?Sized + Sync + Send> Sync for Arc<T> {} + +#[stable(feature = "catch_unwind", since = "1.9.0")] +impl<T: RefUnwindSafe + ?Sized> UnwindSafe for Arc<T> {} + +#[unstable(feature = "coerce_unsized", issue = "27732")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> CoerceUnsized<Arc<U>> for Arc<T> {} + +#[unstable(feature = "dispatch_from_dyn", issue = "none")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> DispatchFromDyn<Arc<U>> for Arc<T> {} + +impl<T: ?Sized> Arc<T> { + unsafe fn from_inner(ptr: NonNull<ArcInner<T>>) -> Self { + Self { ptr, phantom: PhantomData } + } + + unsafe fn from_ptr(ptr: *mut ArcInner<T>) -> Self { + unsafe { Self::from_inner(NonNull::new_unchecked(ptr)) } + } +} + +/// `Weak` is a version of [`Arc`] that holds a non-owning reference to the +/// managed allocation. The allocation is accessed by calling [`upgrade`] on the `Weak` +/// pointer, which returns an <code>[Option]<[Arc]\<T>></code>. +/// +/// Since a `Weak` reference does not count towards ownership, it will not +/// prevent the value stored in the allocation from being dropped, and `Weak` itself makes no +/// guarantees about the value still being present. Thus it may return [`None`] +/// when [`upgrade`]d. Note however that a `Weak` reference *does* prevent the allocation +/// itself (the backing store) from being deallocated. +/// +/// A `Weak` pointer is useful for keeping a temporary reference to the allocation +/// managed by [`Arc`] without preventing its inner value from being dropped. It is also used to +/// prevent circular references between [`Arc`] pointers, since mutual owning references +/// would never allow either [`Arc`] to be dropped. For example, a tree could +/// have strong [`Arc`] pointers from parent nodes to children, and `Weak` +/// pointers from children back to their parents. +/// +/// The typical way to obtain a `Weak` pointer is to call [`Arc::downgrade`]. +/// +/// [`upgrade`]: Weak::upgrade +#[stable(feature = "arc_weak", since = "1.4.0")] +pub struct Weak<T: ?Sized> { + // This is a `NonNull` to allow optimizing the size of this type in enums, + // but it is not necessarily a valid pointer. + // `Weak::new` sets this to `usize::MAX` so that it doesn’t need + // to allocate space on the heap. That's not a value a real pointer + // will ever have because RcBox has alignment at least 2. + // This is only possible when `T: Sized`; unsized `T` never dangle. + ptr: NonNull<ArcInner<T>>, +} + +#[stable(feature = "arc_weak", since = "1.4.0")] +unsafe impl<T: ?Sized + Sync + Send> Send for Weak<T> {} +#[stable(feature = "arc_weak", since = "1.4.0")] +unsafe impl<T: ?Sized + Sync + Send> Sync for Weak<T> {} + +#[unstable(feature = "coerce_unsized", issue = "27732")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> CoerceUnsized<Weak<U>> for Weak<T> {} +#[unstable(feature = "dispatch_from_dyn", issue = "none")] +impl<T: ?Sized + Unsize<U>, U: ?Sized> DispatchFromDyn<Weak<U>> for Weak<T> {} + +#[stable(feature = "arc_weak", since = "1.4.0")] +impl<T: ?Sized + fmt::Debug> fmt::Debug for Weak<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(f, "(Weak)") + } +} + +// This is repr(C) to future-proof against possible field-reordering, which +// would interfere with otherwise safe [into|from]_raw() of transmutable +// inner types. +#[repr(C)] +struct ArcInner<T: ?Sized> { + strong: atomic::AtomicUsize, + + // the value usize::MAX acts as a sentinel for temporarily "locking" the + // ability to upgrade weak pointers or downgrade strong ones; this is used + // to avoid races in `make_mut` and `get_mut`. + weak: atomic::AtomicUsize, + + data: T, +} + +unsafe impl<T: ?Sized + Sync + Send> Send for ArcInner<T> {} +unsafe impl<T: ?Sized + Sync + Send> Sync for ArcInner<T> {} + +impl<T> Arc<T> { + /// Constructs a new `Arc<T>`. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn new(data: T) -> Arc<T> { + // Start the weak pointer count as 1 which is the weak pointer that's + // held by all the strong pointers (kinda), see std/rc.rs for more info + let x: Box<_> = Box::new(ArcInner { + strong: atomic::AtomicUsize::new(1), + weak: atomic::AtomicUsize::new(1), + data, + }); + unsafe { Self::from_inner(Box::leak(x).into()) } + } + + /// Constructs a new `Arc<T>` while giving you a `Weak<T>` to the allocation, + /// to allow you to construct a `T` which holds a weak pointer to itself. + /// + /// Generally, a structure circularly referencing itself, either directly or + /// indirectly, should not hold a strong reference to itself to prevent a memory leak. + /// Using this function, you get access to the weak pointer during the + /// initialization of `T`, before the `Arc<T>` is created, such that you can + /// clone and store it inside the `T`. + /// + /// `new_cyclic` first allocates the managed allocation for the `Arc<T>`, + /// then calls your closure, giving it a `Weak<T>` to this allocation, + /// and only afterwards completes the construction of the `Arc<T>` by placing + /// the `T` returned from your closure into the allocation. + /// + /// Since the new `Arc<T>` is not fully-constructed until `Arc<T>::new_cyclic` + /// returns, calling [`upgrade`] on the weak reference inside your closure will + /// fail and result in a `None` value. + /// + /// # Panics + /// + /// If `data_fn` panics, the panic is propagated to the caller, and the + /// temporary [`Weak<T>`] is dropped normally. + /// + /// # Example + /// + /// ``` + /// # #![allow(dead_code)] + /// use std::sync::{Arc, Weak}; + /// + /// struct Gadget { + /// me: Weak<Gadget>, + /// } + /// + /// impl Gadget { + /// /// Construct a reference counted Gadget. + /// fn new() -> Arc<Self> { + /// // `me` is a `Weak<Gadget>` pointing at the new allocation of the + /// // `Arc` we're constructing. + /// Arc::new_cyclic(|me| { + /// // Create the actual struct here. + /// Gadget { me: me.clone() } + /// }) + /// } + /// + /// /// Return a reference counted pointer to Self. + /// fn me(&self) -> Arc<Self> { + /// self.me.upgrade().unwrap() + /// } + /// } + /// ``` + /// [`upgrade`]: Weak::upgrade + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "arc_new_cyclic", since = "1.60.0")] + pub fn new_cyclic<F>(data_fn: F) -> Arc<T> + where + F: FnOnce(&Weak<T>) -> T, + { + // Construct the inner in the "uninitialized" state with a single + // weak reference. + let uninit_ptr: NonNull<_> = Box::leak(Box::new(ArcInner { + strong: atomic::AtomicUsize::new(0), + weak: atomic::AtomicUsize::new(1), + data: mem::MaybeUninit::<T>::uninit(), + })) + .into(); + let init_ptr: NonNull<ArcInner<T>> = uninit_ptr.cast(); + + let weak = Weak { ptr: init_ptr }; + + // It's important we don't give up ownership of the weak pointer, or + // else the memory might be freed by the time `data_fn` returns. If + // we really wanted to pass ownership, we could create an additional + // weak pointer for ourselves, but this would result in additional + // updates to the weak reference count which might not be necessary + // otherwise. + let data = data_fn(&weak); + + // Now we can properly initialize the inner value and turn our weak + // reference into a strong reference. + let strong = unsafe { + let inner = init_ptr.as_ptr(); + ptr::write(ptr::addr_of_mut!((*inner).data), data); + + // The above write to the data field must be visible to any threads which + // observe a non-zero strong count. Therefore we need at least "Release" ordering + // in order to synchronize with the `compare_exchange_weak` in `Weak::upgrade`. + // + // "Acquire" ordering is not required. When considering the possible behaviours + // of `data_fn` we only need to look at what it could do with a reference to a + // non-upgradeable `Weak`: + // - It can *clone* the `Weak`, increasing the weak reference count. + // - It can drop those clones, decreasing the weak reference count (but never to zero). + // + // These side effects do not impact us in any way, and no other side effects are + // possible with safe code alone. + let prev_value = (*inner).strong.fetch_add(1, Release); + debug_assert_eq!(prev_value, 0, "No prior strong references should exist"); + + Arc::from_inner(init_ptr) + }; + + // Strong references should collectively own a shared weak reference, + // so don't run the destructor for our old weak reference. + mem::forget(weak); + strong + } + + /// Constructs a new `Arc` with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut five = Arc::<u32>::new_uninit(); + /// + /// // Deferred initialization: + /// Arc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit() -> Arc<mem::MaybeUninit<T>> { + unsafe { + Arc::from_ptr(Arc::allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate(layout), + |mem| mem as *mut ArcInner<mem::MaybeUninit<T>>, + )) + } + } + + /// Constructs a new `Arc` with uninitialized contents, with the memory + /// being filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// use std::sync::Arc; + /// + /// let zero = Arc::<u32>::new_zeroed(); + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed() -> Arc<mem::MaybeUninit<T>> { + unsafe { + Arc::from_ptr(Arc::allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate_zeroed(layout), + |mem| mem as *mut ArcInner<mem::MaybeUninit<T>>, + )) + } + } + + /// Constructs a new `Pin<Arc<T>>`. If `T` does not implement `Unpin`, then + /// `data` will be pinned in memory and unable to be moved. + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "pin", since = "1.33.0")] + #[must_use] + pub fn pin(data: T) -> Pin<Arc<T>> { + unsafe { Pin::new_unchecked(Arc::new(data)) } + } + + /// Constructs a new `Pin<Arc<T>>`, return an error if allocation fails. + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn try_pin(data: T) -> Result<Pin<Arc<T>>, AllocError> { + unsafe { Ok(Pin::new_unchecked(Arc::try_new(data)?)) } + } + + /// Constructs a new `Arc<T>`, returning an error if allocation fails. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// use std::sync::Arc; + /// + /// let five = Arc::try_new(5)?; + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn try_new(data: T) -> Result<Arc<T>, AllocError> { + // Start the weak pointer count as 1 which is the weak pointer that's + // held by all the strong pointers (kinda), see std/rc.rs for more info + let x: Box<_> = Box::try_new(ArcInner { + strong: atomic::AtomicUsize::new(1), + weak: atomic::AtomicUsize::new(1), + data, + })?; + unsafe { Ok(Self::from_inner(Box::leak(x).into())) } + } + + /// Constructs a new `Arc` with uninitialized contents, returning an error + /// if allocation fails. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit, allocator_api)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut five = Arc::<u32>::try_new_uninit()?; + /// + /// // Deferred initialization: + /// Arc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + pub fn try_new_uninit() -> Result<Arc<mem::MaybeUninit<T>>, AllocError> { + unsafe { + Ok(Arc::from_ptr(Arc::try_allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate(layout), + |mem| mem as *mut ArcInner<mem::MaybeUninit<T>>, + )?)) + } + } + + /// Constructs a new `Arc` with uninitialized contents, with the memory + /// being filled with `0` bytes, returning an error if allocation fails. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and incorrect usage + /// of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit, allocator_api)] + /// + /// use std::sync::Arc; + /// + /// let zero = Arc::<u32>::try_new_zeroed()?; + /// let zero = unsafe { zero.assume_init() }; + /// + /// assert_eq!(*zero, 0); + /// # Ok::<(), std::alloc::AllocError>(()) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "new_uninit", issue = "63291")] + pub fn try_new_zeroed() -> Result<Arc<mem::MaybeUninit<T>>, AllocError> { + unsafe { + Ok(Arc::from_ptr(Arc::try_allocate_for_layout( + Layout::new::<T>(), + |layout| Global.allocate_zeroed(layout), + |mem| mem as *mut ArcInner<mem::MaybeUninit<T>>, + )?)) + } + } + /// Returns the inner value, if the `Arc` has exactly one strong reference. + /// + /// Otherwise, an [`Err`] is returned with the same `Arc` that was + /// passed in. + /// + /// This will succeed even if there are outstanding weak references. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let x = Arc::new(3); + /// assert_eq!(Arc::try_unwrap(x), Ok(3)); + /// + /// let x = Arc::new(4); + /// let _y = Arc::clone(&x); + /// assert_eq!(*Arc::try_unwrap(x).unwrap_err(), 4); + /// ``` + #[inline] + #[stable(feature = "arc_unique", since = "1.4.0")] + pub fn try_unwrap(this: Self) -> Result<T, Self> { + if this.inner().strong.compare_exchange(1, 0, Relaxed, Relaxed).is_err() { + return Err(this); + } + + acquire!(this.inner().strong); + + unsafe { + let elem = ptr::read(&this.ptr.as_ref().data); + + // Make a weak pointer to clean up the implicit strong-weak reference + let _weak = Weak { ptr: this.ptr }; + mem::forget(this); + + Ok(elem) + } + } +} + +impl<T> Arc<[T]> { + /// Constructs a new atomically reference-counted slice with uninitialized contents. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut values = Arc::<[u32]>::new_uninit_slice(3); + /// + /// // Deferred initialization: + /// let data = Arc::get_mut(&mut values).unwrap(); + /// data[0].write(1); + /// data[1].write(2); + /// data[2].write(3); + /// + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_uninit_slice(len: usize) -> Arc<[mem::MaybeUninit<T>]> { + unsafe { Arc::from_ptr(Arc::allocate_for_slice(len)) } + } + + /// Constructs a new atomically reference-counted slice with uninitialized contents, with the memory being + /// filled with `0` bytes. + /// + /// See [`MaybeUninit::zeroed`][zeroed] for examples of correct and + /// incorrect usage of this method. + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// + /// use std::sync::Arc; + /// + /// let values = Arc::<[u32]>::new_zeroed_slice(3); + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [0, 0, 0]) + /// ``` + /// + /// [zeroed]: mem::MaybeUninit::zeroed + #[cfg(not(no_global_oom_handling))] + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use] + pub fn new_zeroed_slice(len: usize) -> Arc<[mem::MaybeUninit<T>]> { + unsafe { + Arc::from_ptr(Arc::allocate_for_layout( + Layout::array::<T>(len).unwrap(), + |layout| Global.allocate_zeroed(layout), + |mem| { + ptr::slice_from_raw_parts_mut(mem as *mut T, len) + as *mut ArcInner<[mem::MaybeUninit<T>]> + }, + )) + } + } +} + +impl<T> Arc<mem::MaybeUninit<T>> { + /// Converts to `Arc<T>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the inner value + /// really is in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut five = Arc::<u32>::new_uninit(); + /// + /// // Deferred initialization: + /// Arc::get_mut(&mut five).unwrap().write(5); + /// + /// let five = unsafe { five.assume_init() }; + /// + /// assert_eq!(*five, 5) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use = "`self` will be dropped if the result is not used"] + #[inline] + pub unsafe fn assume_init(self) -> Arc<T> { + unsafe { Arc::from_inner(mem::ManuallyDrop::new(self).ptr.cast()) } + } +} + +impl<T> Arc<[mem::MaybeUninit<T>]> { + /// Converts to `Arc<[T]>`. + /// + /// # Safety + /// + /// As with [`MaybeUninit::assume_init`], + /// it is up to the caller to guarantee that the inner value + /// really is in an initialized state. + /// Calling this when the content is not yet fully initialized + /// causes immediate undefined behavior. + /// + /// [`MaybeUninit::assume_init`]: mem::MaybeUninit::assume_init + /// + /// # Examples + /// + /// ``` + /// #![feature(new_uninit)] + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut values = Arc::<[u32]>::new_uninit_slice(3); + /// + /// // Deferred initialization: + /// let data = Arc::get_mut(&mut values).unwrap(); + /// data[0].write(1); + /// data[1].write(2); + /// data[2].write(3); + /// + /// let values = unsafe { values.assume_init() }; + /// + /// assert_eq!(*values, [1, 2, 3]) + /// ``` + #[unstable(feature = "new_uninit", issue = "63291")] + #[must_use = "`self` will be dropped if the result is not used"] + #[inline] + pub unsafe fn assume_init(self) -> Arc<[T]> { + unsafe { Arc::from_ptr(mem::ManuallyDrop::new(self).ptr.as_ptr() as _) } + } +} + +impl<T: ?Sized> Arc<T> { + /// Consumes the `Arc`, returning the wrapped pointer. + /// + /// To avoid a memory leak the pointer must be converted back to an `Arc` using + /// [`Arc::from_raw`]. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let x = Arc::new("hello".to_owned()); + /// let x_ptr = Arc::into_raw(x); + /// assert_eq!(unsafe { &*x_ptr }, "hello"); + /// ``` + #[must_use = "losing the pointer will leak memory"] + #[stable(feature = "rc_raw", since = "1.17.0")] + pub fn into_raw(this: Self) -> *const T { + let ptr = Self::as_ptr(&this); + mem::forget(this); + ptr + } + + /// Provides a raw pointer to the data. + /// + /// The counts are not affected in any way and the `Arc` is not consumed. The pointer is valid for + /// as long as there are strong counts in the `Arc`. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let x = Arc::new("hello".to_owned()); + /// let y = Arc::clone(&x); + /// let x_ptr = Arc::as_ptr(&x); + /// assert_eq!(x_ptr, Arc::as_ptr(&y)); + /// assert_eq!(unsafe { &*x_ptr }, "hello"); + /// ``` + #[must_use] + #[stable(feature = "rc_as_ptr", since = "1.45.0")] + pub fn as_ptr(this: &Self) -> *const T { + let ptr: *mut ArcInner<T> = NonNull::as_ptr(this.ptr); + + // SAFETY: This cannot go through Deref::deref or RcBoxPtr::inner because + // this is required to retain raw/mut provenance such that e.g. `get_mut` can + // write through the pointer after the Rc is recovered through `from_raw`. + unsafe { ptr::addr_of_mut!((*ptr).data) } + } + + /// Constructs an `Arc<T>` from a raw pointer. + /// + /// The raw pointer must have been previously returned by a call to + /// [`Arc<U>::into_raw`][into_raw] where `U` must have the same size and + /// alignment as `T`. This is trivially true if `U` is `T`. + /// Note that if `U` is not `T` but has the same size and alignment, this is + /// basically like transmuting references of different types. See + /// [`mem::transmute`][transmute] for more information on what + /// restrictions apply in this case. + /// + /// The user of `from_raw` has to make sure a specific value of `T` is only + /// dropped once. + /// + /// This function is unsafe because improper use may lead to memory unsafety, + /// even if the returned `Arc<T>` is never accessed. + /// + /// [into_raw]: Arc::into_raw + /// [transmute]: core::mem::transmute + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let x = Arc::new("hello".to_owned()); + /// let x_ptr = Arc::into_raw(x); + /// + /// unsafe { + /// // Convert back to an `Arc` to prevent leak. + /// let x = Arc::from_raw(x_ptr); + /// assert_eq!(&*x, "hello"); + /// + /// // Further calls to `Arc::from_raw(x_ptr)` would be memory-unsafe. + /// } + /// + /// // The memory was freed when `x` went out of scope above, so `x_ptr` is now dangling! + /// ``` + #[stable(feature = "rc_raw", since = "1.17.0")] + pub unsafe fn from_raw(ptr: *const T) -> Self { + unsafe { + let offset = data_offset(ptr); + + // Reverse the offset to find the original ArcInner. + let arc_ptr = ptr.byte_sub(offset) as *mut ArcInner<T>; + + Self::from_ptr(arc_ptr) + } + } + + /// Creates a new [`Weak`] pointer to this allocation. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// let weak_five = Arc::downgrade(&five); + /// ``` + #[must_use = "this returns a new `Weak` pointer, \ + without modifying the original `Arc`"] + #[stable(feature = "arc_weak", since = "1.4.0")] + pub fn downgrade(this: &Self) -> Weak<T> { + // This Relaxed is OK because we're checking the value in the CAS + // below. + let mut cur = this.inner().weak.load(Relaxed); + + loop { + // check if the weak counter is currently "locked"; if so, spin. + if cur == usize::MAX { + hint::spin_loop(); + cur = this.inner().weak.load(Relaxed); + continue; + } + + // NOTE: this code currently ignores the possibility of overflow + // into usize::MAX; in general both Rc and Arc need to be adjusted + // to deal with overflow. + + // Unlike with Clone(), we need this to be an Acquire read to + // synchronize with the write coming from `is_unique`, so that the + // events prior to that write happen before this read. + match this.inner().weak.compare_exchange_weak(cur, cur + 1, Acquire, Relaxed) { + Ok(_) => { + // Make sure we do not create a dangling Weak + debug_assert!(!is_dangling(this.ptr.as_ptr())); + return Weak { ptr: this.ptr }; + } + Err(old) => cur = old, + } + } + } + + /// Gets the number of [`Weak`] pointers to this allocation. + /// + /// # Safety + /// + /// This method by itself is safe, but using it correctly requires extra care. + /// Another thread can change the weak count at any time, + /// including potentially between calling this method and acting on the result. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// let _weak_five = Arc::downgrade(&five); + /// + /// // This assertion is deterministic because we haven't shared + /// // the `Arc` or `Weak` between threads. + /// assert_eq!(1, Arc::weak_count(&five)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "arc_counts", since = "1.15.0")] + pub fn weak_count(this: &Self) -> usize { + let cnt = this.inner().weak.load(Acquire); + // If the weak count is currently locked, the value of the + // count was 0 just before taking the lock. + if cnt == usize::MAX { 0 } else { cnt - 1 } + } + + /// Gets the number of strong (`Arc`) pointers to this allocation. + /// + /// # Safety + /// + /// This method by itself is safe, but using it correctly requires extra care. + /// Another thread can change the strong count at any time, + /// including potentially between calling this method and acting on the result. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// let _also_five = Arc::clone(&five); + /// + /// // This assertion is deterministic because we haven't shared + /// // the `Arc` between threads. + /// assert_eq!(2, Arc::strong_count(&five)); + /// ``` + #[inline] + #[must_use] + #[stable(feature = "arc_counts", since = "1.15.0")] + pub fn strong_count(this: &Self) -> usize { + this.inner().strong.load(Acquire) + } + + /// Increments the strong reference count on the `Arc<T>` associated with the + /// provided pointer by one. + /// + /// # Safety + /// + /// The pointer must have been obtained through `Arc::into_raw`, and the + /// associated `Arc` instance must be valid (i.e. the strong count must be at + /// least 1) for the duration of this method. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// unsafe { + /// let ptr = Arc::into_raw(five); + /// Arc::increment_strong_count(ptr); + /// + /// // This assertion is deterministic because we haven't shared + /// // the `Arc` between threads. + /// let five = Arc::from_raw(ptr); + /// assert_eq!(2, Arc::strong_count(&five)); + /// } + /// ``` + #[inline] + #[stable(feature = "arc_mutate_strong_count", since = "1.51.0")] + pub unsafe fn increment_strong_count(ptr: *const T) { + // Retain Arc, but don't touch refcount by wrapping in ManuallyDrop + let arc = unsafe { mem::ManuallyDrop::new(Arc::<T>::from_raw(ptr)) }; + // Now increase refcount, but don't drop new refcount either + let _arc_clone: mem::ManuallyDrop<_> = arc.clone(); + } + + /// Decrements the strong reference count on the `Arc<T>` associated with the + /// provided pointer by one. + /// + /// # Safety + /// + /// The pointer must have been obtained through `Arc::into_raw`, and the + /// associated `Arc` instance must be valid (i.e. the strong count must be at + /// least 1) when invoking this method. This method can be used to release the final + /// `Arc` and backing storage, but **should not** be called after the final `Arc` has been + /// released. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// unsafe { + /// let ptr = Arc::into_raw(five); + /// Arc::increment_strong_count(ptr); + /// + /// // Those assertions are deterministic because we haven't shared + /// // the `Arc` between threads. + /// let five = Arc::from_raw(ptr); + /// assert_eq!(2, Arc::strong_count(&five)); + /// Arc::decrement_strong_count(ptr); + /// assert_eq!(1, Arc::strong_count(&five)); + /// } + /// ``` + #[inline] + #[stable(feature = "arc_mutate_strong_count", since = "1.51.0")] + pub unsafe fn decrement_strong_count(ptr: *const T) { + unsafe { mem::drop(Arc::from_raw(ptr)) }; + } + + #[inline] + fn inner(&self) -> &ArcInner<T> { + // This unsafety is ok because while this arc is alive we're guaranteed + // that the inner pointer is valid. Furthermore, we know that the + // `ArcInner` structure itself is `Sync` because the inner data is + // `Sync` as well, so we're ok loaning out an immutable pointer to these + // contents. + unsafe { self.ptr.as_ref() } + } + + // Non-inlined part of `drop`. + #[inline(never)] + unsafe fn drop_slow(&mut self) { + // Destroy the data at this time, even though we must not free the box + // allocation itself (there might still be weak pointers lying around). + unsafe { ptr::drop_in_place(Self::get_mut_unchecked(self)) }; + + // Drop the weak ref collectively held by all strong references + drop(Weak { ptr: self.ptr }); + } + + /// Returns `true` if the two `Arc`s point to the same allocation + /// (in a vein similar to [`ptr::eq`]). + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// let same_five = Arc::clone(&five); + /// let other_five = Arc::new(5); + /// + /// assert!(Arc::ptr_eq(&five, &same_five)); + /// assert!(!Arc::ptr_eq(&five, &other_five)); + /// ``` + /// + /// [`ptr::eq`]: core::ptr::eq "ptr::eq" + #[inline] + #[must_use] + #[stable(feature = "ptr_eq", since = "1.17.0")] + pub fn ptr_eq(this: &Self, other: &Self) -> bool { + this.ptr.as_ptr() == other.ptr.as_ptr() + } +} + +impl<T: ?Sized> Arc<T> { + /// Allocates an `ArcInner<T>` with sufficient space for + /// a possibly-unsized inner value where the value has the layout provided. + /// + /// The function `mem_to_arcinner` is called with the data pointer + /// and must return back a (potentially fat)-pointer for the `ArcInner<T>`. + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_layout( + value_layout: Layout, + allocate: impl FnOnce(Layout) -> Result<NonNull<[u8]>, AllocError>, + mem_to_arcinner: impl FnOnce(*mut u8) -> *mut ArcInner<T>, + ) -> *mut ArcInner<T> { + // Calculate layout using the given value layout. + // Previously, layout was calculated on the expression + // `&*(ptr as *const ArcInner<T>)`, but this created a misaligned + // reference (see #54908). + let layout = Layout::new::<ArcInner<()>>().extend(value_layout).unwrap().0.pad_to_align(); + unsafe { + Arc::try_allocate_for_layout(value_layout, allocate, mem_to_arcinner) + .unwrap_or_else(|_| handle_alloc_error(layout)) + } + } + + /// Allocates an `ArcInner<T>` with sufficient space for + /// a possibly-unsized inner value where the value has the layout provided, + /// returning an error if allocation fails. + /// + /// The function `mem_to_arcinner` is called with the data pointer + /// and must return back a (potentially fat)-pointer for the `ArcInner<T>`. + unsafe fn try_allocate_for_layout( + value_layout: Layout, + allocate: impl FnOnce(Layout) -> Result<NonNull<[u8]>, AllocError>, + mem_to_arcinner: impl FnOnce(*mut u8) -> *mut ArcInner<T>, + ) -> Result<*mut ArcInner<T>, AllocError> { + // Calculate layout using the given value layout. + // Previously, layout was calculated on the expression + // `&*(ptr as *const ArcInner<T>)`, but this created a misaligned + // reference (see #54908). + let layout = Layout::new::<ArcInner<()>>().extend(value_layout).unwrap().0.pad_to_align(); + + let ptr = allocate(layout)?; + + // Initialize the ArcInner + let inner = mem_to_arcinner(ptr.as_non_null_ptr().as_ptr()); + debug_assert_eq!(unsafe { Layout::for_value(&*inner) }, layout); + + unsafe { + ptr::write(&mut (*inner).strong, atomic::AtomicUsize::new(1)); + ptr::write(&mut (*inner).weak, atomic::AtomicUsize::new(1)); + } + + Ok(inner) + } + + /// Allocates an `ArcInner<T>` with sufficient space for an unsized inner value. + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_ptr(ptr: *const T) -> *mut ArcInner<T> { + // Allocate for the `ArcInner<T>` using the given value. + unsafe { + Self::allocate_for_layout( + Layout::for_value(&*ptr), + |layout| Global.allocate(layout), + |mem| mem.with_metadata_of(ptr as *mut ArcInner<T>), + ) + } + } + + #[cfg(not(no_global_oom_handling))] + fn from_box(v: Box<T>) -> Arc<T> { + unsafe { + let (box_unique, alloc) = Box::into_unique(v); + let bptr = box_unique.as_ptr(); + + let value_size = size_of_val(&*bptr); + let ptr = Self::allocate_for_ptr(bptr); + + // Copy value as bytes + ptr::copy_nonoverlapping( + bptr as *const T as *const u8, + &mut (*ptr).data as *mut _ as *mut u8, + value_size, + ); + + // Free the allocation without dropping its contents + box_free(box_unique, alloc); + + Self::from_ptr(ptr) + } + } +} + +impl<T> Arc<[T]> { + /// Allocates an `ArcInner<[T]>` with the given length. + #[cfg(not(no_global_oom_handling))] + unsafe fn allocate_for_slice(len: usize) -> *mut ArcInner<[T]> { + unsafe { + Self::allocate_for_layout( + Layout::array::<T>(len).unwrap(), + |layout| Global.allocate(layout), + |mem| ptr::slice_from_raw_parts_mut(mem as *mut T, len) as *mut ArcInner<[T]>, + ) + } + } + + /// Copy elements from slice into newly allocated Arc<\[T\]> + /// + /// Unsafe because the caller must either take ownership or bind `T: Copy`. + #[cfg(not(no_global_oom_handling))] + unsafe fn copy_from_slice(v: &[T]) -> Arc<[T]> { + unsafe { + let ptr = Self::allocate_for_slice(v.len()); + + ptr::copy_nonoverlapping(v.as_ptr(), &mut (*ptr).data as *mut [T] as *mut T, v.len()); + + Self::from_ptr(ptr) + } + } + + /// Constructs an `Arc<[T]>` from an iterator known to be of a certain size. + /// + /// Behavior is undefined should the size be wrong. + #[cfg(not(no_global_oom_handling))] + unsafe fn from_iter_exact(iter: impl iter::Iterator<Item = T>, len: usize) -> Arc<[T]> { + // Panic guard while cloning T elements. + // In the event of a panic, elements that have been written + // into the new ArcInner will be dropped, then the memory freed. + struct Guard<T> { + mem: NonNull<u8>, + elems: *mut T, + layout: Layout, + n_elems: usize, + } + + impl<T> Drop for Guard<T> { + fn drop(&mut self) { + unsafe { + let slice = from_raw_parts_mut(self.elems, self.n_elems); + ptr::drop_in_place(slice); + + Global.deallocate(self.mem, self.layout); + } + } + } + + unsafe { + let ptr = Self::allocate_for_slice(len); + + let mem = ptr as *mut _ as *mut u8; + let layout = Layout::for_value(&*ptr); + + // Pointer to first element + let elems = &mut (*ptr).data as *mut [T] as *mut T; + + let mut guard = Guard { mem: NonNull::new_unchecked(mem), elems, layout, n_elems: 0 }; + + for (i, item) in iter.enumerate() { + ptr::write(elems.add(i), item); + guard.n_elems += 1; + } + + // All clear. Forget the guard so it doesn't free the new ArcInner. + mem::forget(guard); + + Self::from_ptr(ptr) + } + } +} + +/// Specialization trait used for `From<&[T]>`. +#[cfg(not(no_global_oom_handling))] +trait ArcFromSlice<T> { + fn from_slice(slice: &[T]) -> Self; +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Clone> ArcFromSlice<T> for Arc<[T]> { + #[inline] + default fn from_slice(v: &[T]) -> Self { + unsafe { Self::from_iter_exact(v.iter().cloned(), v.len()) } + } +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Copy> ArcFromSlice<T> for Arc<[T]> { + #[inline] + fn from_slice(v: &[T]) -> Self { + unsafe { Arc::copy_from_slice(v) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> Clone for Arc<T> { + /// Makes a clone of the `Arc` pointer. + /// + /// This creates another pointer to the same allocation, increasing the + /// strong reference count. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// let _ = Arc::clone(&five); + /// ``` + #[inline] + fn clone(&self) -> Arc<T> { + // Using a relaxed ordering is alright here, as knowledge of the + // original reference prevents other threads from erroneously deleting + // the object. + // + // As explained in the [Boost documentation][1], Increasing the + // reference counter can always be done with memory_order_relaxed: New + // references to an object can only be formed from an existing + // reference, and passing an existing reference from one thread to + // another must already provide any required synchronization. + // + // [1]: (www.boost.org/doc/libs/1_55_0/doc/html/atomic/usage_examples.html) + let old_size = self.inner().strong.fetch_add(1, Relaxed); + + // However we need to guard against massive refcounts in case someone is `mem::forget`ing + // Arcs. If we don't do this the count can overflow and users will use-after free. This + // branch will never be taken in any realistic program. We abort because such a program is + // incredibly degenerate, and we don't care to support it. + // + // This check is not 100% water-proof: we error when the refcount grows beyond `isize::MAX`. + // But we do that check *after* having done the increment, so there is a chance here that + // the worst already happened and we actually do overflow the `usize` counter. However, that + // requires the counter to grow from `isize::MAX` to `usize::MAX` between the increment + // above and the `abort` below, which seems exceedingly unlikely. + if old_size > MAX_REFCOUNT { + abort(); + } + + unsafe { Self::from_inner(self.ptr) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> Deref for Arc<T> { + type Target = T; + + #[inline] + fn deref(&self) -> &T { + &self.inner().data + } +} + +#[unstable(feature = "receiver_trait", issue = "none")] +impl<T: ?Sized> Receiver for Arc<T> {} + +impl<T: Clone> Arc<T> { + /// Makes a mutable reference into the given `Arc`. + /// + /// If there are other `Arc` pointers to the same allocation, then `make_mut` will + /// [`clone`] the inner value to a new allocation to ensure unique ownership. This is also + /// referred to as clone-on-write. + /// + /// However, if there are no other `Arc` pointers to this allocation, but some [`Weak`] + /// pointers, then the [`Weak`] pointers will be dissociated and the inner value will not + /// be cloned. + /// + /// See also [`get_mut`], which will fail rather than cloning the inner value + /// or dissociating [`Weak`] pointers. + /// + /// [`clone`]: Clone::clone + /// [`get_mut`]: Arc::get_mut + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let mut data = Arc::new(5); + /// + /// *Arc::make_mut(&mut data) += 1; // Won't clone anything + /// let mut other_data = Arc::clone(&data); // Won't clone inner data + /// *Arc::make_mut(&mut data) += 1; // Clones inner data + /// *Arc::make_mut(&mut data) += 1; // Won't clone anything + /// *Arc::make_mut(&mut other_data) *= 2; // Won't clone anything + /// + /// // Now `data` and `other_data` point to different allocations. + /// assert_eq!(*data, 8); + /// assert_eq!(*other_data, 12); + /// ``` + /// + /// [`Weak`] pointers will be dissociated: + /// + /// ``` + /// use std::sync::Arc; + /// + /// let mut data = Arc::new(75); + /// let weak = Arc::downgrade(&data); + /// + /// assert!(75 == *data); + /// assert!(75 == *weak.upgrade().unwrap()); + /// + /// *Arc::make_mut(&mut data) += 1; + /// + /// assert!(76 == *data); + /// assert!(weak.upgrade().is_none()); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "arc_unique", since = "1.4.0")] + pub fn make_mut(this: &mut Self) -> &mut T { + // Note that we hold both a strong reference and a weak reference. + // Thus, releasing our strong reference only will not, by itself, cause + // the memory to be deallocated. + // + // Use Acquire to ensure that we see any writes to `weak` that happen + // before release writes (i.e., decrements) to `strong`. Since we hold a + // weak count, there's no chance the ArcInner itself could be + // deallocated. + if this.inner().strong.compare_exchange(1, 0, Acquire, Relaxed).is_err() { + // Another strong pointer exists, so we must clone. + // Pre-allocate memory to allow writing the cloned value directly. + let mut arc = Self::new_uninit(); + unsafe { + let data = Arc::get_mut_unchecked(&mut arc); + (**this).write_clone_into_raw(data.as_mut_ptr()); + *this = arc.assume_init(); + } + } else if this.inner().weak.load(Relaxed) != 1 { + // Relaxed suffices in the above because this is fundamentally an + // optimization: we are always racing with weak pointers being + // dropped. Worst case, we end up allocated a new Arc unnecessarily. + + // We removed the last strong ref, but there are additional weak + // refs remaining. We'll move the contents to a new Arc, and + // invalidate the other weak refs. + + // Note that it is not possible for the read of `weak` to yield + // usize::MAX (i.e., locked), since the weak count can only be + // locked by a thread with a strong reference. + + // Materialize our own implicit weak pointer, so that it can clean + // up the ArcInner as needed. + let _weak = Weak { ptr: this.ptr }; + + // Can just steal the data, all that's left is Weaks + let mut arc = Self::new_uninit(); + unsafe { + let data = Arc::get_mut_unchecked(&mut arc); + data.as_mut_ptr().copy_from_nonoverlapping(&**this, 1); + ptr::write(this, arc.assume_init()); + } + } else { + // We were the sole reference of either kind; bump back up the + // strong ref count. + this.inner().strong.store(1, Release); + } + + // As with `get_mut()`, the unsafety is ok because our reference was + // either unique to begin with, or became one upon cloning the contents. + unsafe { Self::get_mut_unchecked(this) } + } + + /// If we have the only reference to `T` then unwrap it. Otherwise, clone `T` and return the + /// clone. + /// + /// Assuming `arc_t` is of type `Arc<T>`, this function is functionally equivalent to + /// `(*arc_t).clone()`, but will avoid cloning the inner value where possible. + /// + /// # Examples + /// + /// ``` + /// #![feature(arc_unwrap_or_clone)] + /// # use std::{ptr, sync::Arc}; + /// let inner = String::from("test"); + /// let ptr = inner.as_ptr(); + /// + /// let arc = Arc::new(inner); + /// let inner = Arc::unwrap_or_clone(arc); + /// // The inner value was not cloned + /// assert!(ptr::eq(ptr, inner.as_ptr())); + /// + /// let arc = Arc::new(inner); + /// let arc2 = arc.clone(); + /// let inner = Arc::unwrap_or_clone(arc); + /// // Because there were 2 references, we had to clone the inner value. + /// assert!(!ptr::eq(ptr, inner.as_ptr())); + /// // `arc2` is the last reference, so when we unwrap it we get back + /// // the original `String`. + /// let inner = Arc::unwrap_or_clone(arc2); + /// assert!(ptr::eq(ptr, inner.as_ptr())); + /// ``` + #[inline] + #[unstable(feature = "arc_unwrap_or_clone", issue = "93610")] + pub fn unwrap_or_clone(this: Self) -> T { + Arc::try_unwrap(this).unwrap_or_else(|arc| (*arc).clone()) + } +} + +impl<T: ?Sized> Arc<T> { + /// Returns a mutable reference into the given `Arc`, if there are + /// no other `Arc` or [`Weak`] pointers to the same allocation. + /// + /// Returns [`None`] otherwise, because it is not safe to + /// mutate a shared value. + /// + /// See also [`make_mut`][make_mut], which will [`clone`][clone] + /// the inner value when there are other `Arc` pointers. + /// + /// [make_mut]: Arc::make_mut + /// [clone]: Clone::clone + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let mut x = Arc::new(3); + /// *Arc::get_mut(&mut x).unwrap() = 4; + /// assert_eq!(*x, 4); + /// + /// let _y = Arc::clone(&x); + /// assert!(Arc::get_mut(&mut x).is_none()); + /// ``` + #[inline] + #[stable(feature = "arc_unique", since = "1.4.0")] + pub fn get_mut(this: &mut Self) -> Option<&mut T> { + if this.is_unique() { + // This unsafety is ok because we're guaranteed that the pointer + // returned is the *only* pointer that will ever be returned to T. Our + // reference count is guaranteed to be 1 at this point, and we required + // the Arc itself to be `mut`, so we're returning the only possible + // reference to the inner data. + unsafe { Some(Arc::get_mut_unchecked(this)) } + } else { + None + } + } + + /// Returns a mutable reference into the given `Arc`, + /// without any check. + /// + /// See also [`get_mut`], which is safe and does appropriate checks. + /// + /// [`get_mut`]: Arc::get_mut + /// + /// # Safety + /// + /// Any other `Arc` or [`Weak`] pointers to the same allocation must not be dereferenced + /// for the duration of the returned borrow. + /// This is trivially the case if no such pointers exist, + /// for example immediately after `Arc::new`. + /// + /// # Examples + /// + /// ``` + /// #![feature(get_mut_unchecked)] + /// + /// use std::sync::Arc; + /// + /// let mut x = Arc::new(String::new()); + /// unsafe { + /// Arc::get_mut_unchecked(&mut x).push_str("foo") + /// } + /// assert_eq!(*x, "foo"); + /// ``` + #[inline] + #[unstable(feature = "get_mut_unchecked", issue = "63292")] + pub unsafe fn get_mut_unchecked(this: &mut Self) -> &mut T { + // We are careful to *not* create a reference covering the "count" fields, as + // this would alias with concurrent access to the reference counts (e.g. by `Weak`). + unsafe { &mut (*this.ptr.as_ptr()).data } + } + + /// Determine whether this is the unique reference (including weak refs) to + /// the underlying data. + /// + /// Note that this requires locking the weak ref count. + fn is_unique(&mut self) -> bool { + // lock the weak pointer count if we appear to be the sole weak pointer + // holder. + // + // The acquire label here ensures a happens-before relationship with any + // writes to `strong` (in particular in `Weak::upgrade`) prior to decrements + // of the `weak` count (via `Weak::drop`, which uses release). If the upgraded + // weak ref was never dropped, the CAS here will fail so we do not care to synchronize. + if self.inner().weak.compare_exchange(1, usize::MAX, Acquire, Relaxed).is_ok() { + // This needs to be an `Acquire` to synchronize with the decrement of the `strong` + // counter in `drop` -- the only access that happens when any but the last reference + // is being dropped. + let unique = self.inner().strong.load(Acquire) == 1; + + // The release write here synchronizes with a read in `downgrade`, + // effectively preventing the above read of `strong` from happening + // after the write. + self.inner().weak.store(1, Release); // release the lock + unique + } else { + false + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T: ?Sized> Drop for Arc<T> { + /// Drops the `Arc`. + /// + /// This will decrement the strong reference count. If the strong reference + /// count reaches zero then the only other references (if any) are + /// [`Weak`], so we `drop` the inner value. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// struct Foo; + /// + /// impl Drop for Foo { + /// fn drop(&mut self) { + /// println!("dropped!"); + /// } + /// } + /// + /// let foo = Arc::new(Foo); + /// let foo2 = Arc::clone(&foo); + /// + /// drop(foo); // Doesn't print anything + /// drop(foo2); // Prints "dropped!" + /// ``` + #[inline] + fn drop(&mut self) { + // Because `fetch_sub` is already atomic, we do not need to synchronize + // with other threads unless we are going to delete the object. This + // same logic applies to the below `fetch_sub` to the `weak` count. + if self.inner().strong.fetch_sub(1, Release) != 1 { + return; + } + + // This fence is needed to prevent reordering of use of the data and + // deletion of the data. Because it is marked `Release`, the decreasing + // of the reference count synchronizes with this `Acquire` fence. This + // means that use of the data happens before decreasing the reference + // count, which happens before this fence, which happens before the + // deletion of the data. + // + // As explained in the [Boost documentation][1], + // + // > It is important to enforce any possible access to the object in one + // > thread (through an existing reference) to *happen before* deleting + // > the object in a different thread. This is achieved by a "release" + // > operation after dropping a reference (any access to the object + // > through this reference must obviously happened before), and an + // > "acquire" operation before deleting the object. + // + // In particular, while the contents of an Arc are usually immutable, it's + // possible to have interior writes to something like a Mutex<T>. Since a + // Mutex is not acquired when it is deleted, we can't rely on its + // synchronization logic to make writes in thread A visible to a destructor + // running in thread B. + // + // Also note that the Acquire fence here could probably be replaced with an + // Acquire load, which could improve performance in highly-contended + // situations. See [2]. + // + // [1]: (www.boost.org/doc/libs/1_55_0/doc/html/atomic/usage_examples.html) + // [2]: (https://github.com/rust-lang/rust/pull/41714) + acquire!(self.inner().strong); + + unsafe { + self.drop_slow(); + } + } +} + +impl Arc<dyn Any + Send + Sync> { + /// Attempt to downcast the `Arc<dyn Any + Send + Sync>` to a concrete type. + /// + /// # Examples + /// + /// ``` + /// use std::any::Any; + /// use std::sync::Arc; + /// + /// fn print_if_string(value: Arc<dyn Any + Send + Sync>) { + /// if let Ok(string) = value.downcast::<String>() { + /// println!("String ({}): {}", string.len(), string); + /// } + /// } + /// + /// let my_string = "Hello World".to_string(); + /// print_if_string(Arc::new(my_string)); + /// print_if_string(Arc::new(0i8)); + /// ``` + #[inline] + #[stable(feature = "rc_downcast", since = "1.29.0")] + pub fn downcast<T>(self) -> Result<Arc<T>, Self> + where + T: Any + Send + Sync, + { + if (*self).is::<T>() { + unsafe { + let ptr = self.ptr.cast::<ArcInner<T>>(); + mem::forget(self); + Ok(Arc::from_inner(ptr)) + } + } else { + Err(self) + } + } + + /// Downcasts the `Arc<dyn Any + Send + Sync>` to a concrete type. + /// + /// For a safe alternative see [`downcast`]. + /// + /// # Examples + /// + /// ``` + /// #![feature(downcast_unchecked)] + /// + /// use std::any::Any; + /// use std::sync::Arc; + /// + /// let x: Arc<dyn Any + Send + Sync> = Arc::new(1_usize); + /// + /// unsafe { + /// assert_eq!(*x.downcast_unchecked::<usize>(), 1); + /// } + /// ``` + /// + /// # Safety + /// + /// The contained value must be of type `T`. Calling this method + /// with the incorrect type is *undefined behavior*. + /// + /// + /// [`downcast`]: Self::downcast + #[inline] + #[unstable(feature = "downcast_unchecked", issue = "90850")] + pub unsafe fn downcast_unchecked<T>(self) -> Arc<T> + where + T: Any + Send + Sync, + { + unsafe { + let ptr = self.ptr.cast::<ArcInner<T>>(); + mem::forget(self); + Arc::from_inner(ptr) + } + } +} + +impl<T> Weak<T> { + /// Constructs a new `Weak<T>`, without allocating any memory. + /// Calling [`upgrade`] on the return value always gives [`None`]. + /// + /// [`upgrade`]: Weak::upgrade + /// + /// # Examples + /// + /// ``` + /// use std::sync::Weak; + /// + /// let empty: Weak<i64> = Weak::new(); + /// assert!(empty.upgrade().is_none()); + /// ``` + #[stable(feature = "downgraded_weak", since = "1.10.0")] + #[rustc_const_unstable(feature = "const_weak_new", issue = "95091", reason = "recently added")] + #[must_use] + pub const fn new() -> Weak<T> { + Weak { ptr: unsafe { NonNull::new_unchecked(ptr::invalid_mut::<ArcInner<T>>(usize::MAX)) } } + } +} + +/// Helper type to allow accessing the reference counts without +/// making any assertions about the data field. +struct WeakInner<'a> { + weak: &'a atomic::AtomicUsize, + strong: &'a atomic::AtomicUsize, +} + +impl<T: ?Sized> Weak<T> { + /// Returns a raw pointer to the object `T` pointed to by this `Weak<T>`. + /// + /// The pointer is valid only if there are some strong references. The pointer may be dangling, + /// unaligned or even [`null`] otherwise. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// use std::ptr; + /// + /// let strong = Arc::new("hello".to_owned()); + /// let weak = Arc::downgrade(&strong); + /// // Both point to the same object + /// assert!(ptr::eq(&*strong, weak.as_ptr())); + /// // The strong here keeps it alive, so we can still access the object. + /// assert_eq!("hello", unsafe { &*weak.as_ptr() }); + /// + /// drop(strong); + /// // But not any more. We can do weak.as_ptr(), but accessing the pointer would lead to + /// // undefined behaviour. + /// // assert_eq!("hello", unsafe { &*weak.as_ptr() }); + /// ``` + /// + /// [`null`]: core::ptr::null "ptr::null" + #[must_use] + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub fn as_ptr(&self) -> *const T { + let ptr: *mut ArcInner<T> = NonNull::as_ptr(self.ptr); + + if is_dangling(ptr) { + // If the pointer is dangling, we return the sentinel directly. This cannot be + // a valid payload address, as the payload is at least as aligned as ArcInner (usize). + ptr as *const T + } else { + // SAFETY: if is_dangling returns false, then the pointer is dereferenceable. + // The payload may be dropped at this point, and we have to maintain provenance, + // so use raw pointer manipulation. + unsafe { ptr::addr_of_mut!((*ptr).data) } + } + } + + /// Consumes the `Weak<T>` and turns it into a raw pointer. + /// + /// This converts the weak pointer into a raw pointer, while still preserving the ownership of + /// one weak reference (the weak count is not modified by this operation). It can be turned + /// back into the `Weak<T>` with [`from_raw`]. + /// + /// The same restrictions of accessing the target of the pointer as with + /// [`as_ptr`] apply. + /// + /// # Examples + /// + /// ``` + /// use std::sync::{Arc, Weak}; + /// + /// let strong = Arc::new("hello".to_owned()); + /// let weak = Arc::downgrade(&strong); + /// let raw = weak.into_raw(); + /// + /// assert_eq!(1, Arc::weak_count(&strong)); + /// assert_eq!("hello", unsafe { &*raw }); + /// + /// drop(unsafe { Weak::from_raw(raw) }); + /// assert_eq!(0, Arc::weak_count(&strong)); + /// ``` + /// + /// [`from_raw`]: Weak::from_raw + /// [`as_ptr`]: Weak::as_ptr + #[must_use = "`self` will be dropped if the result is not used"] + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub fn into_raw(self) -> *const T { + let result = self.as_ptr(); + mem::forget(self); + result + } + + /// Converts a raw pointer previously created by [`into_raw`] back into `Weak<T>`. + /// + /// This can be used to safely get a strong reference (by calling [`upgrade`] + /// later) or to deallocate the weak count by dropping the `Weak<T>`. + /// + /// It takes ownership of one weak reference (with the exception of pointers created by [`new`], + /// as these don't own anything; the method still works on them). + /// + /// # Safety + /// + /// The pointer must have originated from the [`into_raw`] and must still own its potential + /// weak reference. + /// + /// It is allowed for the strong count to be 0 at the time of calling this. Nevertheless, this + /// takes ownership of one weak reference currently represented as a raw pointer (the weak + /// count is not modified by this operation) and therefore it must be paired with a previous + /// call to [`into_raw`]. + /// # Examples + /// + /// ``` + /// use std::sync::{Arc, Weak}; + /// + /// let strong = Arc::new("hello".to_owned()); + /// + /// let raw_1 = Arc::downgrade(&strong).into_raw(); + /// let raw_2 = Arc::downgrade(&strong).into_raw(); + /// + /// assert_eq!(2, Arc::weak_count(&strong)); + /// + /// assert_eq!("hello", &*unsafe { Weak::from_raw(raw_1) }.upgrade().unwrap()); + /// assert_eq!(1, Arc::weak_count(&strong)); + /// + /// drop(strong); + /// + /// // Decrement the last weak count. + /// assert!(unsafe { Weak::from_raw(raw_2) }.upgrade().is_none()); + /// ``` + /// + /// [`new`]: Weak::new + /// [`into_raw`]: Weak::into_raw + /// [`upgrade`]: Weak::upgrade + #[stable(feature = "weak_into_raw", since = "1.45.0")] + pub unsafe fn from_raw(ptr: *const T) -> Self { + // See Weak::as_ptr for context on how the input pointer is derived. + + let ptr = if is_dangling(ptr as *mut T) { + // This is a dangling Weak. + ptr as *mut ArcInner<T> + } else { + // Otherwise, we're guaranteed the pointer came from a nondangling Weak. + // SAFETY: data_offset is safe to call, as ptr references a real (potentially dropped) T. + let offset = unsafe { data_offset(ptr) }; + // Thus, we reverse the offset to get the whole RcBox. + // SAFETY: the pointer originated from a Weak, so this offset is safe. + unsafe { ptr.byte_sub(offset) as *mut ArcInner<T> } + }; + + // SAFETY: we now have recovered the original Weak pointer, so can create the Weak. + Weak { ptr: unsafe { NonNull::new_unchecked(ptr) } } + } +} + +impl<T: ?Sized> Weak<T> { + /// Attempts to upgrade the `Weak` pointer to an [`Arc`], delaying + /// dropping of the inner value if successful. + /// + /// Returns [`None`] if the inner value has since been dropped. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// let weak_five = Arc::downgrade(&five); + /// + /// let strong_five: Option<Arc<_>> = weak_five.upgrade(); + /// assert!(strong_five.is_some()); + /// + /// // Destroy all strong pointers. + /// drop(strong_five); + /// drop(five); + /// + /// assert!(weak_five.upgrade().is_none()); + /// ``` + #[must_use = "this returns a new `Arc`, \ + without modifying the original weak pointer"] + #[stable(feature = "arc_weak", since = "1.4.0")] + pub fn upgrade(&self) -> Option<Arc<T>> { + // We use a CAS loop to increment the strong count instead of a + // fetch_add as this function should never take the reference count + // from zero to one. + let inner = self.inner()?; + + // Relaxed load because any write of 0 that we can observe + // leaves the field in a permanently zero state (so a + // "stale" read of 0 is fine), and any other value is + // confirmed via the CAS below. + let mut n = inner.strong.load(Relaxed); + + loop { + if n == 0 { + return None; + } + + // See comments in `Arc::clone` for why we do this (for `mem::forget`). + if n > MAX_REFCOUNT { + abort(); + } + + // Relaxed is fine for the failure case because we don't have any expectations about the new state. + // Acquire is necessary for the success case to synchronise with `Arc::new_cyclic`, when the inner + // value can be initialized after `Weak` references have already been created. In that case, we + // expect to observe the fully initialized value. + match inner.strong.compare_exchange_weak(n, n + 1, Acquire, Relaxed) { + Ok(_) => return Some(unsafe { Arc::from_inner(self.ptr) }), // null checked above + Err(old) => n = old, + } + } + } + + /// Gets the number of strong (`Arc`) pointers pointing to this allocation. + /// + /// If `self` was created using [`Weak::new`], this will return 0. + #[must_use] + #[stable(feature = "weak_counts", since = "1.41.0")] + pub fn strong_count(&self) -> usize { + if let Some(inner) = self.inner() { inner.strong.load(Acquire) } else { 0 } + } + + /// Gets an approximation of the number of `Weak` pointers pointing to this + /// allocation. + /// + /// If `self` was created using [`Weak::new`], or if there are no remaining + /// strong pointers, this will return 0. + /// + /// # Accuracy + /// + /// Due to implementation details, the returned value can be off by 1 in + /// either direction when other threads are manipulating any `Arc`s or + /// `Weak`s pointing to the same allocation. + #[must_use] + #[stable(feature = "weak_counts", since = "1.41.0")] + pub fn weak_count(&self) -> usize { + self.inner() + .map(|inner| { + let weak = inner.weak.load(Acquire); + let strong = inner.strong.load(Acquire); + if strong == 0 { + 0 + } else { + // Since we observed that there was at least one strong pointer + // after reading the weak count, we know that the implicit weak + // reference (present whenever any strong references are alive) + // was still around when we observed the weak count, and can + // therefore safely subtract it. + weak - 1 + } + }) + .unwrap_or(0) + } + + /// Returns `None` when the pointer is dangling and there is no allocated `ArcInner`, + /// (i.e., when this `Weak` was created by `Weak::new`). + #[inline] + fn inner(&self) -> Option<WeakInner<'_>> { + if is_dangling(self.ptr.as_ptr()) { + None + } else { + // We are careful to *not* create a reference covering the "data" field, as + // the field may be mutated concurrently (for example, if the last `Arc` + // is dropped, the data field will be dropped in-place). + Some(unsafe { + let ptr = self.ptr.as_ptr(); + WeakInner { strong: &(*ptr).strong, weak: &(*ptr).weak } + }) + } + } + + /// Returns `true` if the two `Weak`s point to the same allocation (similar to + /// [`ptr::eq`]), or if both don't point to any allocation + /// (because they were created with `Weak::new()`). + /// + /// # Notes + /// + /// Since this compares pointers it means that `Weak::new()` will equal each + /// other, even though they don't point to any allocation. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let first_rc = Arc::new(5); + /// let first = Arc::downgrade(&first_rc); + /// let second = Arc::downgrade(&first_rc); + /// + /// assert!(first.ptr_eq(&second)); + /// + /// let third_rc = Arc::new(5); + /// let third = Arc::downgrade(&third_rc); + /// + /// assert!(!first.ptr_eq(&third)); + /// ``` + /// + /// Comparing `Weak::new`. + /// + /// ``` + /// use std::sync::{Arc, Weak}; + /// + /// let first = Weak::new(); + /// let second = Weak::new(); + /// assert!(first.ptr_eq(&second)); + /// + /// let third_rc = Arc::new(()); + /// let third = Arc::downgrade(&third_rc); + /// assert!(!first.ptr_eq(&third)); + /// ``` + /// + /// [`ptr::eq`]: core::ptr::eq "ptr::eq" + #[inline] + #[must_use] + #[stable(feature = "weak_ptr_eq", since = "1.39.0")] + pub fn ptr_eq(&self, other: &Self) -> bool { + self.ptr.as_ptr() == other.ptr.as_ptr() + } +} + +#[stable(feature = "arc_weak", since = "1.4.0")] +impl<T: ?Sized> Clone for Weak<T> { + /// Makes a clone of the `Weak` pointer that points to the same allocation. + /// + /// # Examples + /// + /// ``` + /// use std::sync::{Arc, Weak}; + /// + /// let weak_five = Arc::downgrade(&Arc::new(5)); + /// + /// let _ = Weak::clone(&weak_five); + /// ``` + #[inline] + fn clone(&self) -> Weak<T> { + let inner = if let Some(inner) = self.inner() { + inner + } else { + return Weak { ptr: self.ptr }; + }; + // See comments in Arc::clone() for why this is relaxed. This can use a + // fetch_add (ignoring the lock) because the weak count is only locked + // where are *no other* weak pointers in existence. (So we can't be + // running this code in that case). + let old_size = inner.weak.fetch_add(1, Relaxed); + + // See comments in Arc::clone() for why we do this (for mem::forget). + if old_size > MAX_REFCOUNT { + abort(); + } + + Weak { ptr: self.ptr } + } +} + +#[stable(feature = "downgraded_weak", since = "1.10.0")] +impl<T> Default for Weak<T> { + /// Constructs a new `Weak<T>`, without allocating memory. + /// Calling [`upgrade`] on the return value always + /// gives [`None`]. + /// + /// [`upgrade`]: Weak::upgrade + /// + /// # Examples + /// + /// ``` + /// use std::sync::Weak; + /// + /// let empty: Weak<i64> = Default::default(); + /// assert!(empty.upgrade().is_none()); + /// ``` + fn default() -> Weak<T> { + Weak::new() + } +} + +#[stable(feature = "arc_weak", since = "1.4.0")] +unsafe impl<#[may_dangle] T: ?Sized> Drop for Weak<T> { + /// Drops the `Weak` pointer. + /// + /// # Examples + /// + /// ``` + /// use std::sync::{Arc, Weak}; + /// + /// struct Foo; + /// + /// impl Drop for Foo { + /// fn drop(&mut self) { + /// println!("dropped!"); + /// } + /// } + /// + /// let foo = Arc::new(Foo); + /// let weak_foo = Arc::downgrade(&foo); + /// let other_weak_foo = Weak::clone(&weak_foo); + /// + /// drop(weak_foo); // Doesn't print anything + /// drop(foo); // Prints "dropped!" + /// + /// assert!(other_weak_foo.upgrade().is_none()); + /// ``` + fn drop(&mut self) { + // If we find out that we were the last weak pointer, then its time to + // deallocate the data entirely. See the discussion in Arc::drop() about + // the memory orderings + // + // It's not necessary to check for the locked state here, because the + // weak count can only be locked if there was precisely one weak ref, + // meaning that drop could only subsequently run ON that remaining weak + // ref, which can only happen after the lock is released. + let inner = if let Some(inner) = self.inner() { inner } else { return }; + + if inner.weak.fetch_sub(1, Release) == 1 { + acquire!(inner.weak); + unsafe { Global.deallocate(self.ptr.cast(), Layout::for_value_raw(self.ptr.as_ptr())) } + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +trait ArcEqIdent<T: ?Sized + PartialEq> { + fn eq(&self, other: &Arc<T>) -> bool; + fn ne(&self, other: &Arc<T>) -> bool; +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialEq> ArcEqIdent<T> for Arc<T> { + #[inline] + default fn eq(&self, other: &Arc<T>) -> bool { + **self == **other + } + #[inline] + default fn ne(&self, other: &Arc<T>) -> bool { + **self != **other + } +} + +/// We're doing this specialization here, and not as a more general optimization on `&T`, because it +/// would otherwise add a cost to all equality checks on refs. We assume that `Arc`s are used to +/// store large values, that are slow to clone, but also heavy to check for equality, causing this +/// cost to pay off more easily. It's also more likely to have two `Arc` clones, that point to +/// the same value, than two `&T`s. +/// +/// We can only do this when `T: Eq` as a `PartialEq` might be deliberately irreflexive. +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + crate::rc::MarkerEq> ArcEqIdent<T> for Arc<T> { + #[inline] + fn eq(&self, other: &Arc<T>) -> bool { + Arc::ptr_eq(self, other) || **self == **other + } + + #[inline] + fn ne(&self, other: &Arc<T>) -> bool { + !Arc::ptr_eq(self, other) && **self != **other + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialEq> PartialEq for Arc<T> { + /// Equality for two `Arc`s. + /// + /// Two `Arc`s are equal if their inner values are equal, even if they are + /// stored in different allocation. + /// + /// If `T` also implements `Eq` (implying reflexivity of equality), + /// two `Arc`s that point to the same allocation are always equal. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five == Arc::new(5)); + /// ``` + #[inline] + fn eq(&self, other: &Arc<T>) -> bool { + ArcEqIdent::eq(self, other) + } + + /// Inequality for two `Arc`s. + /// + /// Two `Arc`s are unequal if their inner values are unequal. + /// + /// If `T` also implements `Eq` (implying reflexivity of equality), + /// two `Arc`s that point to the same value are never unequal. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five != Arc::new(6)); + /// ``` + #[inline] + fn ne(&self, other: &Arc<T>) -> bool { + ArcEqIdent::ne(self, other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + PartialOrd> PartialOrd for Arc<T> { + /// Partial comparison for two `Arc`s. + /// + /// The two are compared by calling `partial_cmp()` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// use std::cmp::Ordering; + /// + /// let five = Arc::new(5); + /// + /// assert_eq!(Some(Ordering::Less), five.partial_cmp(&Arc::new(6))); + /// ``` + fn partial_cmp(&self, other: &Arc<T>) -> Option<Ordering> { + (**self).partial_cmp(&**other) + } + + /// Less-than comparison for two `Arc`s. + /// + /// The two are compared by calling `<` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five < Arc::new(6)); + /// ``` + fn lt(&self, other: &Arc<T>) -> bool { + *(*self) < *(*other) + } + + /// 'Less than or equal to' comparison for two `Arc`s. + /// + /// The two are compared by calling `<=` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five <= Arc::new(5)); + /// ``` + fn le(&self, other: &Arc<T>) -> bool { + *(*self) <= *(*other) + } + + /// Greater-than comparison for two `Arc`s. + /// + /// The two are compared by calling `>` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five > Arc::new(4)); + /// ``` + fn gt(&self, other: &Arc<T>) -> bool { + *(*self) > *(*other) + } + + /// 'Greater than or equal to' comparison for two `Arc`s. + /// + /// The two are compared by calling `>=` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let five = Arc::new(5); + /// + /// assert!(five >= Arc::new(5)); + /// ``` + fn ge(&self, other: &Arc<T>) -> bool { + *(*self) >= *(*other) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Ord> Ord for Arc<T> { + /// Comparison for two `Arc`s. + /// + /// The two are compared by calling `cmp()` on their inner values. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// use std::cmp::Ordering; + /// + /// let five = Arc::new(5); + /// + /// assert_eq!(Ordering::Less, five.cmp(&Arc::new(6))); + /// ``` + fn cmp(&self, other: &Arc<T>) -> Ordering { + (**self).cmp(&**other) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Eq> Eq for Arc<T> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + fmt::Display> fmt::Display for Arc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Display::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + fmt::Debug> fmt::Debug for Arc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> fmt::Pointer for Arc<T> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Pointer::fmt(&(&**self as *const T), f) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Default> Default for Arc<T> { + /// Creates a new `Arc<T>`, with the `Default` value for `T`. + /// + /// # Examples + /// + /// ``` + /// use std::sync::Arc; + /// + /// let x: Arc<i32> = Default::default(); + /// assert_eq!(*x, 0); + /// ``` + fn default() -> Arc<T> { + Arc::new(Default::default()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized + Hash> Hash for Arc<T> { + fn hash<H: Hasher>(&self, state: &mut H) { + (**self).hash(state) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "from_for_ptrs", since = "1.6.0")] +impl<T> From<T> for Arc<T> { + /// Converts a `T` into an `Arc<T>` + /// + /// The conversion moves the value into a + /// newly allocated `Arc`. It is equivalent to + /// calling `Arc::new(t)`. + /// + /// # Example + /// ```rust + /// # use std::sync::Arc; + /// let x = 5; + /// let arc = Arc::new(5); + /// + /// assert_eq!(Arc::from(x), arc); + /// ``` + fn from(t: T) -> Self { + Arc::new(t) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T: Clone> From<&[T]> for Arc<[T]> { + /// Allocate a reference-counted slice and fill it by cloning `v`'s items. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let original: &[i32] = &[1, 2, 3]; + /// let shared: Arc<[i32]> = Arc::from(original); + /// assert_eq!(&[1, 2, 3], &shared[..]); + /// ``` + #[inline] + fn from(v: &[T]) -> Arc<[T]> { + <Self as ArcFromSlice<T>>::from_slice(v) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl From<&str> for Arc<str> { + /// Allocate a reference-counted `str` and copy `v` into it. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let shared: Arc<str> = Arc::from("eggplant"); + /// assert_eq!("eggplant", &shared[..]); + /// ``` + #[inline] + fn from(v: &str) -> Arc<str> { + let arc = Arc::<[u8]>::from(v.as_bytes()); + unsafe { Arc::from_raw(Arc::into_raw(arc) as *const str) } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl From<String> for Arc<str> { + /// Allocate a reference-counted `str` and copy `v` into it. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let unique: String = "eggplant".to_owned(); + /// let shared: Arc<str> = Arc::from(unique); + /// assert_eq!("eggplant", &shared[..]); + /// ``` + #[inline] + fn from(v: String) -> Arc<str> { + Arc::from(&v[..]) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T: ?Sized> From<Box<T>> for Arc<T> { + /// Move a boxed object to a new, reference-counted allocation. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let unique: Box<str> = Box::from("eggplant"); + /// let shared: Arc<str> = Arc::from(unique); + /// assert_eq!("eggplant", &shared[..]); + /// ``` + #[inline] + fn from(v: Box<T>) -> Arc<T> { + Arc::from_box(v) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_slice", since = "1.21.0")] +impl<T> From<Vec<T>> for Arc<[T]> { + /// Allocate a reference-counted slice and move `v`'s items into it. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let unique: Vec<i32> = vec![1, 2, 3]; + /// let shared: Arc<[i32]> = Arc::from(unique); + /// assert_eq!(&[1, 2, 3], &shared[..]); + /// ``` + #[inline] + fn from(mut v: Vec<T>) -> Arc<[T]> { + unsafe { + let arc = Arc::copy_from_slice(&v); + + // Allow the Vec to free its memory, but not destroy its contents + v.set_len(0); + + arc + } + } +} + +#[stable(feature = "shared_from_cow", since = "1.45.0")] +impl<'a, B> From<Cow<'a, B>> for Arc<B> +where + B: ToOwned + ?Sized, + Arc<B>: From<&'a B> + From<B::Owned>, +{ + /// Create an atomically reference-counted pointer from + /// a clone-on-write pointer by copying its content. + /// + /// # Example + /// + /// ```rust + /// # use std::sync::Arc; + /// # use std::borrow::Cow; + /// let cow: Cow<str> = Cow::Borrowed("eggplant"); + /// let shared: Arc<str> = Arc::from(cow); + /// assert_eq!("eggplant", &shared[..]); + /// ``` + #[inline] + fn from(cow: Cow<'a, B>) -> Arc<B> { + match cow { + Cow::Borrowed(s) => Arc::from(s), + Cow::Owned(s) => Arc::from(s), + } + } +} + +#[stable(feature = "shared_from_str", since = "1.62.0")] +impl From<Arc<str>> for Arc<[u8]> { + /// Converts an atomically reference-counted string slice into a byte slice. + /// + /// # Example + /// + /// ``` + /// # use std::sync::Arc; + /// let string: Arc<str> = Arc::from("eggplant"); + /// let bytes: Arc<[u8]> = Arc::from(string); + /// assert_eq!("eggplant".as_bytes(), bytes.as_ref()); + /// ``` + #[inline] + fn from(rc: Arc<str>) -> Self { + // SAFETY: `str` has the same layout as `[u8]`. + unsafe { Arc::from_raw(Arc::into_raw(rc) as *const [u8]) } + } +} + +#[stable(feature = "boxed_slice_try_from", since = "1.43.0")] +impl<T, const N: usize> TryFrom<Arc<[T]>> for Arc<[T; N]> { + type Error = Arc<[T]>; + + fn try_from(boxed_slice: Arc<[T]>) -> Result<Self, Self::Error> { + if boxed_slice.len() == N { + Ok(unsafe { Arc::from_raw(Arc::into_raw(boxed_slice) as *mut [T; N]) }) + } else { + Err(boxed_slice) + } + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "shared_from_iter", since = "1.37.0")] +impl<T> iter::FromIterator<T> for Arc<[T]> { + /// Takes each element in the `Iterator` and collects it into an `Arc<[T]>`. + /// + /// # Performance characteristics + /// + /// ## The general case + /// + /// In the general case, collecting into `Arc<[T]>` is done by first + /// collecting into a `Vec<T>`. That is, when writing the following: + /// + /// ```rust + /// # use std::sync::Arc; + /// let evens: Arc<[u8]> = (0..10).filter(|&x| x % 2 == 0).collect(); + /// # assert_eq!(&*evens, &[0, 2, 4, 6, 8]); + /// ``` + /// + /// this behaves as if we wrote: + /// + /// ```rust + /// # use std::sync::Arc; + /// let evens: Arc<[u8]> = (0..10).filter(|&x| x % 2 == 0) + /// .collect::<Vec<_>>() // The first set of allocations happens here. + /// .into(); // A second allocation for `Arc<[T]>` happens here. + /// # assert_eq!(&*evens, &[0, 2, 4, 6, 8]); + /// ``` + /// + /// This will allocate as many times as needed for constructing the `Vec<T>` + /// and then it will allocate once for turning the `Vec<T>` into the `Arc<[T]>`. + /// + /// ## Iterators of known length + /// + /// When your `Iterator` implements `TrustedLen` and is of an exact size, + /// a single allocation will be made for the `Arc<[T]>`. For example: + /// + /// ```rust + /// # use std::sync::Arc; + /// let evens: Arc<[u8]> = (0..10).collect(); // Just a single allocation happens here. + /// # assert_eq!(&*evens, &*(0..10).collect::<Vec<_>>()); + /// ``` + fn from_iter<I: iter::IntoIterator<Item = T>>(iter: I) -> Self { + ToArcSlice::to_arc_slice(iter.into_iter()) + } +} + +/// Specialization trait used for collecting into `Arc<[T]>`. +trait ToArcSlice<T>: Iterator<Item = T> + Sized { + fn to_arc_slice(self) -> Arc<[T]>; +} + +#[cfg(not(no_global_oom_handling))] +impl<T, I: Iterator<Item = T>> ToArcSlice<T> for I { + default fn to_arc_slice(self) -> Arc<[T]> { + self.collect::<Vec<T>>().into() + } +} + +#[cfg(not(no_global_oom_handling))] +impl<T, I: iter::TrustedLen<Item = T>> ToArcSlice<T> for I { + fn to_arc_slice(self) -> Arc<[T]> { + // This is the case for a `TrustedLen` iterator. + let (low, high) = self.size_hint(); + if let Some(high) = high { + debug_assert_eq!( + low, + high, + "TrustedLen iterator's size hint is not exact: {:?}", + (low, high) + ); + + unsafe { + // SAFETY: We need to ensure that the iterator has an exact length and we have. + Arc::from_iter_exact(self, low) + } + } else { + // TrustedLen contract guarantees that `upper_bound == `None` implies an iterator + // length exceeding `usize::MAX`. + // The default implementation would collect into a vec which would panic. + // Thus we panic here immediately without invoking `Vec` code. + panic!("capacity overflow"); + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: ?Sized> borrow::Borrow<T> for Arc<T> { + fn borrow(&self) -> &T { + &**self + } +} + +#[stable(since = "1.5.0", feature = "smart_ptr_as_ref")] +impl<T: ?Sized> AsRef<T> for Arc<T> { + fn as_ref(&self) -> &T { + &**self + } +} + +#[stable(feature = "pin", since = "1.33.0")] +impl<T: ?Sized> Unpin for Arc<T> {} + +/// Get the offset within an `ArcInner` for the payload behind a pointer. +/// +/// # Safety +/// +/// The pointer must point to (and have valid metadata for) a previously +/// valid instance of T, but the T is allowed to be dropped. +unsafe fn data_offset<T: ?Sized>(ptr: *const T) -> usize { + // Align the unsized value to the end of the ArcInner. + // Because RcBox is repr(C), it will always be the last field in memory. + // SAFETY: since the only unsized types possible are slices, trait objects, + // and extern types, the input safety requirement is currently enough to + // satisfy the requirements of align_of_val_raw; this is an implementation + // detail of the language that must not be relied upon outside of std. + unsafe { data_offset_align(align_of_val_raw(ptr)) } +} + +#[inline] +fn data_offset_align(align: usize) -> usize { + let layout = Layout::new::<ArcInner<()>>(); + layout.size() + layout.padding_needed_for(align) +} diff --git a/library/alloc/src/sync/tests.rs b/library/alloc/src/sync/tests.rs new file mode 100644 index 000000000..202d0e7f0 --- /dev/null +++ b/library/alloc/src/sync/tests.rs @@ -0,0 +1,620 @@ +use super::*; + +use std::boxed::Box; +use std::clone::Clone; +use std::convert::{From, TryInto}; +use std::mem::drop; +use std::ops::Drop; +use std::option::Option::{self, None, Some}; +use std::sync::atomic::{ + self, + Ordering::{Acquire, SeqCst}, +}; +use std::sync::mpsc::channel; +use std::sync::Mutex; +use std::thread; + +use crate::vec::Vec; + +struct Canary(*mut atomic::AtomicUsize); + +impl Drop for Canary { + fn drop(&mut self) { + unsafe { + match *self { + Canary(c) => { + (*c).fetch_add(1, SeqCst); + } + } + } + } +} + +#[test] +#[cfg_attr(target_os = "emscripten", ignore)] +fn manually_share_arc() { + let v = vec![1, 2, 3, 4, 5, 6, 7, 8, 9, 10]; + let arc_v = Arc::new(v); + + let (tx, rx) = channel(); + + let _t = thread::spawn(move || { + let arc_v: Arc<Vec<i32>> = rx.recv().unwrap(); + assert_eq!((*arc_v)[3], 4); + }); + + tx.send(arc_v.clone()).unwrap(); + + assert_eq!((*arc_v)[2], 3); + assert_eq!((*arc_v)[4], 5); +} + +#[test] +fn test_arc_get_mut() { + let mut x = Arc::new(3); + *Arc::get_mut(&mut x).unwrap() = 4; + assert_eq!(*x, 4); + let y = x.clone(); + assert!(Arc::get_mut(&mut x).is_none()); + drop(y); + assert!(Arc::get_mut(&mut x).is_some()); + let _w = Arc::downgrade(&x); + assert!(Arc::get_mut(&mut x).is_none()); +} + +#[test] +fn weak_counts() { + assert_eq!(Weak::weak_count(&Weak::<u64>::new()), 0); + assert_eq!(Weak::strong_count(&Weak::<u64>::new()), 0); + + let a = Arc::new(0); + let w = Arc::downgrade(&a); + assert_eq!(Weak::strong_count(&w), 1); + assert_eq!(Weak::weak_count(&w), 1); + let w2 = w.clone(); + assert_eq!(Weak::strong_count(&w), 1); + assert_eq!(Weak::weak_count(&w), 2); + assert_eq!(Weak::strong_count(&w2), 1); + assert_eq!(Weak::weak_count(&w2), 2); + drop(w); + assert_eq!(Weak::strong_count(&w2), 1); + assert_eq!(Weak::weak_count(&w2), 1); + let a2 = a.clone(); + assert_eq!(Weak::strong_count(&w2), 2); + assert_eq!(Weak::weak_count(&w2), 1); + drop(a2); + drop(a); + assert_eq!(Weak::strong_count(&w2), 0); + assert_eq!(Weak::weak_count(&w2), 0); + drop(w2); +} + +#[test] +fn try_unwrap() { + let x = Arc::new(3); + assert_eq!(Arc::try_unwrap(x), Ok(3)); + let x = Arc::new(4); + let _y = x.clone(); + assert_eq!(Arc::try_unwrap(x), Err(Arc::new(4))); + let x = Arc::new(5); + let _w = Arc::downgrade(&x); + assert_eq!(Arc::try_unwrap(x), Ok(5)); +} + +#[test] +fn into_from_raw() { + let x = Arc::new(Box::new("hello")); + let y = x.clone(); + + let x_ptr = Arc::into_raw(x); + drop(y); + unsafe { + assert_eq!(**x_ptr, "hello"); + + let x = Arc::from_raw(x_ptr); + assert_eq!(**x, "hello"); + + assert_eq!(Arc::try_unwrap(x).map(|x| *x), Ok("hello")); + } +} + +#[test] +fn test_into_from_raw_unsized() { + use std::fmt::Display; + use std::string::ToString; + + let arc: Arc<str> = Arc::from("foo"); + + let ptr = Arc::into_raw(arc.clone()); + let arc2 = unsafe { Arc::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }, "foo"); + assert_eq!(arc, arc2); + + let arc: Arc<dyn Display> = Arc::new(123); + + let ptr = Arc::into_raw(arc.clone()); + let arc2 = unsafe { Arc::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }.to_string(), "123"); + assert_eq!(arc2.to_string(), "123"); +} + +#[test] +fn into_from_weak_raw() { + let x = Arc::new(Box::new("hello")); + let y = Arc::downgrade(&x); + + let y_ptr = Weak::into_raw(y); + unsafe { + assert_eq!(**y_ptr, "hello"); + + let y = Weak::from_raw(y_ptr); + let y_up = Weak::upgrade(&y).unwrap(); + assert_eq!(**y_up, "hello"); + drop(y_up); + + assert_eq!(Arc::try_unwrap(x).map(|x| *x), Ok("hello")); + } +} + +#[test] +fn test_into_from_weak_raw_unsized() { + use std::fmt::Display; + use std::string::ToString; + + let arc: Arc<str> = Arc::from("foo"); + let weak: Weak<str> = Arc::downgrade(&arc); + + let ptr = Weak::into_raw(weak.clone()); + let weak2 = unsafe { Weak::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }, "foo"); + assert!(weak.ptr_eq(&weak2)); + + let arc: Arc<dyn Display> = Arc::new(123); + let weak: Weak<dyn Display> = Arc::downgrade(&arc); + + let ptr = Weak::into_raw(weak.clone()); + let weak2 = unsafe { Weak::from_raw(ptr) }; + + assert_eq!(unsafe { &*ptr }.to_string(), "123"); + assert!(weak.ptr_eq(&weak2)); +} + +#[test] +fn test_cowarc_clone_make_mut() { + let mut cow0 = Arc::new(75); + let mut cow1 = cow0.clone(); + let mut cow2 = cow1.clone(); + + assert!(75 == *Arc::make_mut(&mut cow0)); + assert!(75 == *Arc::make_mut(&mut cow1)); + assert!(75 == *Arc::make_mut(&mut cow2)); + + *Arc::make_mut(&mut cow0) += 1; + *Arc::make_mut(&mut cow1) += 2; + *Arc::make_mut(&mut cow2) += 3; + + assert!(76 == *cow0); + assert!(77 == *cow1); + assert!(78 == *cow2); + + // none should point to the same backing memory + assert!(*cow0 != *cow1); + assert!(*cow0 != *cow2); + assert!(*cow1 != *cow2); +} + +#[test] +fn test_cowarc_clone_unique2() { + let mut cow0 = Arc::new(75); + let cow1 = cow0.clone(); + let cow2 = cow1.clone(); + + assert!(75 == *cow0); + assert!(75 == *cow1); + assert!(75 == *cow2); + + *Arc::make_mut(&mut cow0) += 1; + assert!(76 == *cow0); + assert!(75 == *cow1); + assert!(75 == *cow2); + + // cow1 and cow2 should share the same contents + // cow0 should have a unique reference + assert!(*cow0 != *cow1); + assert!(*cow0 != *cow2); + assert!(*cow1 == *cow2); +} + +#[test] +fn test_cowarc_clone_weak() { + let mut cow0 = Arc::new(75); + let cow1_weak = Arc::downgrade(&cow0); + + assert!(75 == *cow0); + assert!(75 == *cow1_weak.upgrade().unwrap()); + + *Arc::make_mut(&mut cow0) += 1; + + assert!(76 == *cow0); + assert!(cow1_weak.upgrade().is_none()); +} + +#[test] +fn test_live() { + let x = Arc::new(5); + let y = Arc::downgrade(&x); + assert!(y.upgrade().is_some()); +} + +#[test] +fn test_dead() { + let x = Arc::new(5); + let y = Arc::downgrade(&x); + drop(x); + assert!(y.upgrade().is_none()); +} + +#[test] +fn weak_self_cyclic() { + struct Cycle { + x: Mutex<Option<Weak<Cycle>>>, + } + + let a = Arc::new(Cycle { x: Mutex::new(None) }); + let b = Arc::downgrade(&a.clone()); + *a.x.lock().unwrap() = Some(b); + + // hopefully we don't double-free (or leak)... +} + +#[test] +fn drop_arc() { + let mut canary = atomic::AtomicUsize::new(0); + let x = Arc::new(Canary(&mut canary as *mut atomic::AtomicUsize)); + drop(x); + assert!(canary.load(Acquire) == 1); +} + +#[test] +fn drop_arc_weak() { + let mut canary = atomic::AtomicUsize::new(0); + let arc = Arc::new(Canary(&mut canary as *mut atomic::AtomicUsize)); + let arc_weak = Arc::downgrade(&arc); + assert!(canary.load(Acquire) == 0); + drop(arc); + assert!(canary.load(Acquire) == 1); + drop(arc_weak); +} + +#[test] +fn test_strong_count() { + let a = Arc::new(0); + assert!(Arc::strong_count(&a) == 1); + let w = Arc::downgrade(&a); + assert!(Arc::strong_count(&a) == 1); + let b = w.upgrade().expect(""); + assert!(Arc::strong_count(&b) == 2); + assert!(Arc::strong_count(&a) == 2); + drop(w); + drop(a); + assert!(Arc::strong_count(&b) == 1); + let c = b.clone(); + assert!(Arc::strong_count(&b) == 2); + assert!(Arc::strong_count(&c) == 2); +} + +#[test] +fn test_weak_count() { + let a = Arc::new(0); + assert!(Arc::strong_count(&a) == 1); + assert!(Arc::weak_count(&a) == 0); + let w = Arc::downgrade(&a); + assert!(Arc::strong_count(&a) == 1); + assert!(Arc::weak_count(&a) == 1); + let x = w.clone(); + assert!(Arc::weak_count(&a) == 2); + drop(w); + drop(x); + assert!(Arc::strong_count(&a) == 1); + assert!(Arc::weak_count(&a) == 0); + let c = a.clone(); + assert!(Arc::strong_count(&a) == 2); + assert!(Arc::weak_count(&a) == 0); + let d = Arc::downgrade(&c); + assert!(Arc::weak_count(&c) == 1); + assert!(Arc::strong_count(&c) == 2); + + drop(a); + drop(c); + drop(d); +} + +#[test] +fn show_arc() { + let a = Arc::new(5); + assert_eq!(format!("{a:?}"), "5"); +} + +// Make sure deriving works with Arc<T> +#[derive(Eq, Ord, PartialEq, PartialOrd, Clone, Debug, Default)] +struct Foo { + inner: Arc<i32>, +} + +#[test] +fn test_unsized() { + let x: Arc<[i32]> = Arc::new([1, 2, 3]); + assert_eq!(format!("{x:?}"), "[1, 2, 3]"); + let y = Arc::downgrade(&x.clone()); + drop(x); + assert!(y.upgrade().is_none()); +} + +#[test] +fn test_maybe_thin_unsized() { + // If/when custom thin DSTs exist, this test should be updated to use one + use std::ffi::{CStr, CString}; + + let x: Arc<CStr> = Arc::from(CString::new("swordfish").unwrap().into_boxed_c_str()); + assert_eq!(format!("{x:?}"), "\"swordfish\""); + let y: Weak<CStr> = Arc::downgrade(&x); + drop(x); + + // At this point, the weak points to a dropped DST + assert!(y.upgrade().is_none()); + // But we still need to be able to get the alloc layout to drop. + // CStr has no drop glue, but custom DSTs might, and need to work. + drop(y); +} + +#[test] +fn test_from_owned() { + let foo = 123; + let foo_arc = Arc::from(foo); + assert!(123 == *foo_arc); +} + +#[test] +fn test_new_weak() { + let foo: Weak<usize> = Weak::new(); + assert!(foo.upgrade().is_none()); +} + +#[test] +fn test_ptr_eq() { + let five = Arc::new(5); + let same_five = five.clone(); + let other_five = Arc::new(5); + + assert!(Arc::ptr_eq(&five, &same_five)); + assert!(!Arc::ptr_eq(&five, &other_five)); +} + +#[test] +#[cfg_attr(target_os = "emscripten", ignore)] +fn test_weak_count_locked() { + let mut a = Arc::new(atomic::AtomicBool::new(false)); + let a2 = a.clone(); + let t = thread::spawn(move || { + // Miri is too slow + let count = if cfg!(miri) { 1000 } else { 1000000 }; + for _i in 0..count { + Arc::get_mut(&mut a); + } + a.store(true, SeqCst); + }); + + while !a2.load(SeqCst) { + let n = Arc::weak_count(&a2); + assert!(n < 2, "bad weak count: {}", n); + #[cfg(miri)] // Miri's scheduler does not guarantee liveness, and thus needs this hint. + std::hint::spin_loop(); + } + t.join().unwrap(); +} + +#[test] +fn test_from_str() { + let r: Arc<str> = Arc::from("foo"); + + assert_eq!(&r[..], "foo"); +} + +#[test] +fn test_copy_from_slice() { + let s: &[u32] = &[1, 2, 3]; + let r: Arc<[u32]> = Arc::from(s); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_clone_from_slice() { + #[derive(Clone, Debug, Eq, PartialEq)] + struct X(u32); + + let s: &[X] = &[X(1), X(2), X(3)]; + let r: Arc<[X]> = Arc::from(s); + + assert_eq!(&r[..], s); +} + +#[test] +#[should_panic] +fn test_clone_from_slice_panic() { + use std::string::{String, ToString}; + + struct Fail(u32, String); + + impl Clone for Fail { + fn clone(&self) -> Fail { + if self.0 == 2 { + panic!(); + } + Fail(self.0, self.1.clone()) + } + } + + let s: &[Fail] = + &[Fail(0, "foo".to_string()), Fail(1, "bar".to_string()), Fail(2, "baz".to_string())]; + + // Should panic, but not cause memory corruption + let _r: Arc<[Fail]> = Arc::from(s); +} + +#[test] +fn test_from_box() { + let b: Box<u32> = Box::new(123); + let r: Arc<u32> = Arc::from(b); + + assert_eq!(*r, 123); +} + +#[test] +fn test_from_box_str() { + use std::string::String; + + let s = String::from("foo").into_boxed_str(); + let r: Arc<str> = Arc::from(s); + + assert_eq!(&r[..], "foo"); +} + +#[test] +fn test_from_box_slice() { + let s = vec![1, 2, 3].into_boxed_slice(); + let r: Arc<[u32]> = Arc::from(s); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_from_box_trait() { + use std::fmt::Display; + use std::string::ToString; + + let b: Box<dyn Display> = Box::new(123); + let r: Arc<dyn Display> = Arc::from(b); + + assert_eq!(r.to_string(), "123"); +} + +#[test] +fn test_from_box_trait_zero_sized() { + use std::fmt::Debug; + + let b: Box<dyn Debug> = Box::new(()); + let r: Arc<dyn Debug> = Arc::from(b); + + assert_eq!(format!("{r:?}"), "()"); +} + +#[test] +fn test_from_vec() { + let v = vec![1, 2, 3]; + let r: Arc<[u32]> = Arc::from(v); + + assert_eq!(&r[..], [1, 2, 3]); +} + +#[test] +fn test_downcast() { + use std::any::Any; + + let r1: Arc<dyn Any + Send + Sync> = Arc::new(i32::MAX); + let r2: Arc<dyn Any + Send + Sync> = Arc::new("abc"); + + assert!(r1.clone().downcast::<u32>().is_err()); + + let r1i32 = r1.downcast::<i32>(); + assert!(r1i32.is_ok()); + assert_eq!(r1i32.unwrap(), Arc::new(i32::MAX)); + + assert!(r2.clone().downcast::<i32>().is_err()); + + let r2str = r2.downcast::<&'static str>(); + assert!(r2str.is_ok()); + assert_eq!(r2str.unwrap(), Arc::new("abc")); +} + +#[test] +fn test_array_from_slice() { + let v = vec![1, 2, 3]; + let r: Arc<[u32]> = Arc::from(v); + + let a: Result<Arc<[u32; 3]>, _> = r.clone().try_into(); + assert!(a.is_ok()); + + let a: Result<Arc<[u32; 2]>, _> = r.clone().try_into(); + assert!(a.is_err()); +} + +#[test] +fn test_arc_cyclic_with_zero_refs() { + struct ZeroRefs { + inner: Weak<ZeroRefs>, + } + let zero_refs = Arc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + ZeroRefs { inner: Weak::new() } + }); + + assert_eq!(Arc::strong_count(&zero_refs), 1); + assert_eq!(Arc::weak_count(&zero_refs), 0); + assert_eq!(zero_refs.inner.strong_count(), 0); + assert_eq!(zero_refs.inner.weak_count(), 0); +} + +#[test] +fn test_arc_new_cyclic_one_ref() { + struct OneRef { + inner: Weak<OneRef>, + } + let one_ref = Arc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + OneRef { inner: inner.clone() } + }); + + assert_eq!(Arc::strong_count(&one_ref), 1); + assert_eq!(Arc::weak_count(&one_ref), 1); + + let one_ref2 = Weak::upgrade(&one_ref.inner).unwrap(); + assert!(Arc::ptr_eq(&one_ref, &one_ref2)); + + assert_eq!(Arc::strong_count(&one_ref), 2); + assert_eq!(Arc::weak_count(&one_ref), 1); +} + +#[test] +fn test_arc_cyclic_two_refs() { + struct TwoRefs { + inner1: Weak<TwoRefs>, + inner2: Weak<TwoRefs>, + } + let two_refs = Arc::new_cyclic(|inner| { + assert_eq!(inner.strong_count(), 0); + assert!(inner.upgrade().is_none()); + + let inner1 = inner.clone(); + let inner2 = inner1.clone(); + + TwoRefs { inner1, inner2 } + }); + + assert_eq!(Arc::strong_count(&two_refs), 1); + assert_eq!(Arc::weak_count(&two_refs), 2); + + let two_refs1 = Weak::upgrade(&two_refs.inner1).unwrap(); + assert!(Arc::ptr_eq(&two_refs, &two_refs1)); + + let two_refs2 = Weak::upgrade(&two_refs.inner2).unwrap(); + assert!(Arc::ptr_eq(&two_refs, &two_refs2)); + + assert_eq!(Arc::strong_count(&two_refs), 3); + assert_eq!(Arc::weak_count(&two_refs), 2); +} diff --git a/library/alloc/src/task.rs b/library/alloc/src/task.rs new file mode 100644 index 000000000..528ee4ff1 --- /dev/null +++ b/library/alloc/src/task.rs @@ -0,0 +1,147 @@ +#![stable(feature = "wake_trait", since = "1.51.0")] +//! Types and Traits for working with asynchronous tasks. +use core::mem::ManuallyDrop; +use core::task::{RawWaker, RawWakerVTable, Waker}; + +use crate::sync::Arc; + +/// The implementation of waking a task on an executor. +/// +/// This trait can be used to create a [`Waker`]. An executor can define an +/// implementation of this trait, and use that to construct a Waker to pass +/// to the tasks that are executed on that executor. +/// +/// This trait is a memory-safe and ergonomic alternative to constructing a +/// [`RawWaker`]. It supports the common executor design in which the data used +/// to wake up a task is stored in an [`Arc`]. Some executors (especially +/// those for embedded systems) cannot use this API, which is why [`RawWaker`] +/// exists as an alternative for those systems. +/// +/// [arc]: ../../std/sync/struct.Arc.html +/// +/// # Examples +/// +/// A basic `block_on` function that takes a future and runs it to completion on +/// the current thread. +/// +/// **Note:** This example trades correctness for simplicity. In order to prevent +/// deadlocks, production-grade implementations will also need to handle +/// intermediate calls to `thread::unpark` as well as nested invocations. +/// +/// ```rust +/// use std::future::Future; +/// use std::sync::Arc; +/// use std::task::{Context, Poll, Wake}; +/// use std::thread::{self, Thread}; +/// +/// /// A waker that wakes up the current thread when called. +/// struct ThreadWaker(Thread); +/// +/// impl Wake for ThreadWaker { +/// fn wake(self: Arc<Self>) { +/// self.0.unpark(); +/// } +/// } +/// +/// /// Run a future to completion on the current thread. +/// fn block_on<T>(fut: impl Future<Output = T>) -> T { +/// // Pin the future so it can be polled. +/// let mut fut = Box::pin(fut); +/// +/// // Create a new context to be passed to the future. +/// let t = thread::current(); +/// let waker = Arc::new(ThreadWaker(t)).into(); +/// let mut cx = Context::from_waker(&waker); +/// +/// // Run the future to completion. +/// loop { +/// match fut.as_mut().poll(&mut cx) { +/// Poll::Ready(res) => return res, +/// Poll::Pending => thread::park(), +/// } +/// } +/// } +/// +/// block_on(async { +/// println!("Hi from inside a future!"); +/// }); +/// ``` +#[stable(feature = "wake_trait", since = "1.51.0")] +pub trait Wake { + /// Wake this task. + #[stable(feature = "wake_trait", since = "1.51.0")] + fn wake(self: Arc<Self>); + + /// Wake this task without consuming the waker. + /// + /// If an executor supports a cheaper way to wake without consuming the + /// waker, it should override this method. By default, it clones the + /// [`Arc`] and calls [`wake`] on the clone. + /// + /// [`wake`]: Wake::wake + #[stable(feature = "wake_trait", since = "1.51.0")] + fn wake_by_ref(self: &Arc<Self>) { + self.clone().wake(); + } +} + +#[stable(feature = "wake_trait", since = "1.51.0")] +impl<W: Wake + Send + Sync + 'static> From<Arc<W>> for Waker { + /// Use a `Wake`-able type as a `Waker`. + /// + /// No heap allocations or atomic operations are used for this conversion. + fn from(waker: Arc<W>) -> Waker { + // SAFETY: This is safe because raw_waker safely constructs + // a RawWaker from Arc<W>. + unsafe { Waker::from_raw(raw_waker(waker)) } + } +} + +#[stable(feature = "wake_trait", since = "1.51.0")] +impl<W: Wake + Send + Sync + 'static> From<Arc<W>> for RawWaker { + /// Use a `Wake`-able type as a `RawWaker`. + /// + /// No heap allocations or atomic operations are used for this conversion. + fn from(waker: Arc<W>) -> RawWaker { + raw_waker(waker) + } +} + +// NB: This private function for constructing a RawWaker is used, rather than +// inlining this into the `From<Arc<W>> for RawWaker` impl, to ensure that +// the safety of `From<Arc<W>> for Waker` does not depend on the correct +// trait dispatch - instead both impls call this function directly and +// explicitly. +#[inline(always)] +fn raw_waker<W: Wake + Send + Sync + 'static>(waker: Arc<W>) -> RawWaker { + // Increment the reference count of the arc to clone it. + unsafe fn clone_waker<W: Wake + Send + Sync + 'static>(waker: *const ()) -> RawWaker { + unsafe { Arc::increment_strong_count(waker as *const W) }; + RawWaker::new( + waker as *const (), + &RawWakerVTable::new(clone_waker::<W>, wake::<W>, wake_by_ref::<W>, drop_waker::<W>), + ) + } + + // Wake by value, moving the Arc into the Wake::wake function + unsafe fn wake<W: Wake + Send + Sync + 'static>(waker: *const ()) { + let waker = unsafe { Arc::from_raw(waker as *const W) }; + <W as Wake>::wake(waker); + } + + // Wake by reference, wrap the waker in ManuallyDrop to avoid dropping it + unsafe fn wake_by_ref<W: Wake + Send + Sync + 'static>(waker: *const ()) { + let waker = unsafe { ManuallyDrop::new(Arc::from_raw(waker as *const W)) }; + <W as Wake>::wake_by_ref(&waker); + } + + // Decrement the reference count of the Arc on drop + unsafe fn drop_waker<W: Wake + Send + Sync + 'static>(waker: *const ()) { + unsafe { Arc::decrement_strong_count(waker as *const W) }; + } + + RawWaker::new( + Arc::into_raw(waker) as *const (), + &RawWakerVTable::new(clone_waker::<W>, wake::<W>, wake_by_ref::<W>, drop_waker::<W>), + ) +} diff --git a/library/alloc/src/tests.rs b/library/alloc/src/tests.rs new file mode 100644 index 000000000..299ed156a --- /dev/null +++ b/library/alloc/src/tests.rs @@ -0,0 +1,151 @@ +//! Test for `boxed` mod. + +use core::any::Any; +use core::clone::Clone; +use core::convert::TryInto; +use core::ops::Deref; +use core::result::Result::{Err, Ok}; + +use std::boxed::Box; + +#[test] +fn test_owned_clone() { + let a = Box::new(5); + let b: Box<i32> = a.clone(); + assert!(a == b); +} + +#[derive(PartialEq, Eq)] +struct Test; + +#[test] +fn any_move() { + let a = Box::new(8) as Box<dyn Any>; + let b = Box::new(Test) as Box<dyn Any>; + + match a.downcast::<i32>() { + Ok(a) => { + assert!(a == Box::new(8)); + } + Err(..) => panic!(), + } + match b.downcast::<Test>() { + Ok(a) => { + assert!(a == Box::new(Test)); + } + Err(..) => panic!(), + } + + let a = Box::new(8) as Box<dyn Any>; + let b = Box::new(Test) as Box<dyn Any>; + + assert!(a.downcast::<Box<Test>>().is_err()); + assert!(b.downcast::<Box<i32>>().is_err()); +} + +#[test] +fn test_show() { + let a = Box::new(8) as Box<dyn Any>; + let b = Box::new(Test) as Box<dyn Any>; + let a_str = format!("{a:?}"); + let b_str = format!("{b:?}"); + assert_eq!(a_str, "Any { .. }"); + assert_eq!(b_str, "Any { .. }"); + + static EIGHT: usize = 8; + static TEST: Test = Test; + let a = &EIGHT as &dyn Any; + let b = &TEST as &dyn Any; + let s = format!("{a:?}"); + assert_eq!(s, "Any { .. }"); + let s = format!("{b:?}"); + assert_eq!(s, "Any { .. }"); +} + +#[test] +fn deref() { + fn homura<T: Deref<Target = i32>>(_: T) {} + homura(Box::new(765)); +} + +#[test] +fn raw_sized() { + let x = Box::new(17); + let p = Box::into_raw(x); + unsafe { + assert_eq!(17, *p); + *p = 19; + let y = Box::from_raw(p); + assert_eq!(19, *y); + } +} + +#[test] +fn raw_trait() { + trait Foo { + fn get(&self) -> u32; + fn set(&mut self, value: u32); + } + + struct Bar(u32); + + impl Foo for Bar { + fn get(&self) -> u32 { + self.0 + } + + fn set(&mut self, value: u32) { + self.0 = value; + } + } + + let x: Box<dyn Foo> = Box::new(Bar(17)); + let p = Box::into_raw(x); + unsafe { + assert_eq!(17, (*p).get()); + (*p).set(19); + let y: Box<dyn Foo> = Box::from_raw(p); + assert_eq!(19, y.get()); + } +} + +#[test] +fn f64_slice() { + let slice: &[f64] = &[-1.0, 0.0, 1.0, f64::INFINITY]; + let boxed: Box<[f64]> = Box::from(slice); + assert_eq!(&*boxed, slice) +} + +#[test] +fn i64_slice() { + let slice: &[i64] = &[i64::MIN, -2, -1, 0, 1, 2, i64::MAX]; + let boxed: Box<[i64]> = Box::from(slice); + assert_eq!(&*boxed, slice) +} + +#[test] +fn str_slice() { + let s = "Hello, world!"; + let boxed: Box<str> = Box::from(s); + assert_eq!(&*boxed, s) +} + +#[test] +fn boxed_slice_from_iter() { + let iter = 0..100; + let boxed: Box<[u32]> = iter.collect(); + assert_eq!(boxed.len(), 100); + assert_eq!(boxed[7], 7); +} + +#[test] +fn test_array_from_slice() { + let v = vec![1, 2, 3]; + let r: Box<[u32]> = v.into_boxed_slice(); + + let a: Result<Box<[u32; 3]>, _> = r.clone().try_into(); + assert!(a.is_ok()); + + let a: Result<Box<[u32; 2]>, _> = r.clone().try_into(); + assert!(a.is_err()); +} diff --git a/library/alloc/src/vec/cow.rs b/library/alloc/src/vec/cow.rs new file mode 100644 index 000000000..64943a273 --- /dev/null +++ b/library/alloc/src/vec/cow.rs @@ -0,0 +1,53 @@ +use crate::borrow::Cow; +use core::iter::FromIterator; + +use super::Vec; + +#[stable(feature = "cow_from_vec", since = "1.8.0")] +impl<'a, T: Clone> From<&'a [T]> for Cow<'a, [T]> { + /// Creates a [`Borrowed`] variant of [`Cow`] + /// from a slice. + /// + /// This conversion does not allocate or clone the data. + /// + /// [`Borrowed`]: crate::borrow::Cow::Borrowed + fn from(s: &'a [T]) -> Cow<'a, [T]> { + Cow::Borrowed(s) + } +} + +#[stable(feature = "cow_from_vec", since = "1.8.0")] +impl<'a, T: Clone> From<Vec<T>> for Cow<'a, [T]> { + /// Creates an [`Owned`] variant of [`Cow`] + /// from an owned instance of [`Vec`]. + /// + /// This conversion does not allocate or clone the data. + /// + /// [`Owned`]: crate::borrow::Cow::Owned + fn from(v: Vec<T>) -> Cow<'a, [T]> { + Cow::Owned(v) + } +} + +#[stable(feature = "cow_from_vec_ref", since = "1.28.0")] +impl<'a, T: Clone> From<&'a Vec<T>> for Cow<'a, [T]> { + /// Creates a [`Borrowed`] variant of [`Cow`] + /// from a reference to [`Vec`]. + /// + /// This conversion does not allocate or clone the data. + /// + /// [`Borrowed`]: crate::borrow::Cow::Borrowed + fn from(v: &'a Vec<T>) -> Cow<'a, [T]> { + Cow::Borrowed(v.as_slice()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T> FromIterator<T> for Cow<'a, [T]> +where + T: Clone, +{ + fn from_iter<I: IntoIterator<Item = T>>(it: I) -> Cow<'a, [T]> { + Cow::Owned(FromIterator::from_iter(it)) + } +} diff --git a/library/alloc/src/vec/drain.rs b/library/alloc/src/vec/drain.rs new file mode 100644 index 000000000..5cdee0bd4 --- /dev/null +++ b/library/alloc/src/vec/drain.rs @@ -0,0 +1,184 @@ +use crate::alloc::{Allocator, Global}; +use core::fmt; +use core::iter::{FusedIterator, TrustedLen}; +use core::mem; +use core::ptr::{self, NonNull}; +use core::slice::{self}; + +use super::Vec; + +/// A draining iterator for `Vec<T>`. +/// +/// This `struct` is created by [`Vec::drain`]. +/// See its documentation for more. +/// +/// # Example +/// +/// ``` +/// let mut v = vec![0, 1, 2]; +/// let iter: std::vec::Drain<_> = v.drain(..); +/// ``` +#[stable(feature = "drain", since = "1.6.0")] +pub struct Drain< + 'a, + T: 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + 'a = Global, +> { + /// Index of tail to preserve + pub(super) tail_start: usize, + /// Length of tail + pub(super) tail_len: usize, + /// Current remaining range to remove + pub(super) iter: slice::Iter<'a, T>, + pub(super) vec: NonNull<Vec<T, A>>, +} + +#[stable(feature = "collection_debug", since = "1.17.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for Drain<'_, T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Drain").field(&self.iter.as_slice()).finish() + } +} + +impl<'a, T, A: Allocator> Drain<'a, T, A> { + /// Returns the remaining items of this iterator as a slice. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec!['a', 'b', 'c']; + /// let mut drain = vec.drain(..); + /// assert_eq!(drain.as_slice(), &['a', 'b', 'c']); + /// let _ = drain.next().unwrap(); + /// assert_eq!(drain.as_slice(), &['b', 'c']); + /// ``` + #[must_use] + #[stable(feature = "vec_drain_as_slice", since = "1.46.0")] + pub fn as_slice(&self) -> &[T] { + self.iter.as_slice() + } + + /// Returns a reference to the underlying allocator. + #[unstable(feature = "allocator_api", issue = "32838")] + #[must_use] + #[inline] + pub fn allocator(&self) -> &A { + unsafe { self.vec.as_ref().allocator() } + } +} + +#[stable(feature = "vec_drain_as_slice", since = "1.46.0")] +impl<'a, T, A: Allocator> AsRef<[T]> for Drain<'a, T, A> { + fn as_ref(&self) -> &[T] { + self.as_slice() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl<T: Sync, A: Sync + Allocator> Sync for Drain<'_, T, A> {} +#[stable(feature = "drain", since = "1.6.0")] +unsafe impl<T: Send, A: Send + Allocator> Send for Drain<'_, T, A> {} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> Iterator for Drain<'_, T, A> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + self.iter.next().map(|elt| unsafe { ptr::read(elt as *const _) }) + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.iter.size_hint() + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> DoubleEndedIterator for Drain<'_, T, A> { + #[inline] + fn next_back(&mut self) -> Option<T> { + self.iter.next_back().map(|elt| unsafe { ptr::read(elt as *const _) }) + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> Drop for Drain<'_, T, A> { + fn drop(&mut self) { + /// Moves back the un-`Drain`ed elements to restore the original `Vec`. + struct DropGuard<'r, 'a, T, A: Allocator>(&'r mut Drain<'a, T, A>); + + impl<'r, 'a, T, A: Allocator> Drop for DropGuard<'r, 'a, T, A> { + fn drop(&mut self) { + if self.0.tail_len > 0 { + unsafe { + let source_vec = self.0.vec.as_mut(); + // memmove back untouched tail, update to new length + let start = source_vec.len(); + let tail = self.0.tail_start; + if tail != start { + let src = source_vec.as_ptr().add(tail); + let dst = source_vec.as_mut_ptr().add(start); + ptr::copy(src, dst, self.0.tail_len); + } + source_vec.set_len(start + self.0.tail_len); + } + } + } + } + + let iter = mem::replace(&mut self.iter, (&mut []).iter()); + let drop_len = iter.len(); + + let mut vec = self.vec; + + if mem::size_of::<T>() == 0 { + // ZSTs have no identity, so we don't need to move them around, we only need to drop the correct amount. + // this can be achieved by manipulating the Vec length instead of moving values out from `iter`. + unsafe { + let vec = vec.as_mut(); + let old_len = vec.len(); + vec.set_len(old_len + drop_len + self.tail_len); + vec.truncate(old_len + self.tail_len); + } + + return; + } + + // ensure elements are moved back into their appropriate places, even when drop_in_place panics + let _guard = DropGuard(self); + + if drop_len == 0 { + return; + } + + // as_slice() must only be called when iter.len() is > 0 because + // vec::Splice modifies vec::Drain fields and may grow the vec which would invalidate + // the iterator's internal pointers. Creating a reference to deallocated memory + // is invalid even when it is zero-length + let drop_ptr = iter.as_slice().as_ptr(); + + unsafe { + // drop_ptr comes from a slice::Iter which only gives us a &[T] but for drop_in_place + // a pointer with mutable provenance is necessary. Therefore we must reconstruct + // it from the original vec but also avoid creating a &mut to the front since that could + // invalidate raw pointers to it which some unsafe code might rely on. + let vec_ptr = vec.as_mut().as_mut_ptr(); + let drop_offset = drop_ptr.sub_ptr(vec_ptr); + let to_drop = ptr::slice_from_raw_parts_mut(vec_ptr.add(drop_offset), drop_len); + ptr::drop_in_place(to_drop); + } + } +} + +#[stable(feature = "drain", since = "1.6.0")] +impl<T, A: Allocator> ExactSizeIterator for Drain<'_, T, A> { + fn is_empty(&self) -> bool { + self.iter.is_empty() + } +} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T, A: Allocator> TrustedLen for Drain<'_, T, A> {} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T, A: Allocator> FusedIterator for Drain<'_, T, A> {} diff --git a/library/alloc/src/vec/drain_filter.rs b/library/alloc/src/vec/drain_filter.rs new file mode 100644 index 000000000..3c37c92ae --- /dev/null +++ b/library/alloc/src/vec/drain_filter.rs @@ -0,0 +1,143 @@ +use crate::alloc::{Allocator, Global}; +use core::ptr::{self}; +use core::slice::{self}; + +use super::Vec; + +/// An iterator which uses a closure to determine if an element should be removed. +/// +/// This struct is created by [`Vec::drain_filter`]. +/// See its documentation for more. +/// +/// # Example +/// +/// ``` +/// #![feature(drain_filter)] +/// +/// let mut v = vec![0, 1, 2]; +/// let iter: std::vec::DrainFilter<_, _> = v.drain_filter(|x| *x % 2 == 0); +/// ``` +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +#[derive(Debug)] +pub struct DrainFilter< + 'a, + T, + F, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +> where + F: FnMut(&mut T) -> bool, +{ + pub(super) vec: &'a mut Vec<T, A>, + /// The index of the item that will be inspected by the next call to `next`. + pub(super) idx: usize, + /// The number of items that have been drained (removed) thus far. + pub(super) del: usize, + /// The original length of `vec` prior to draining. + pub(super) old_len: usize, + /// The filter test predicate. + pub(super) pred: F, + /// A flag that indicates a panic has occurred in the filter test predicate. + /// This is used as a hint in the drop implementation to prevent consumption + /// of the remainder of the `DrainFilter`. Any unprocessed items will be + /// backshifted in the `vec`, but no further items will be dropped or + /// tested by the filter predicate. + pub(super) panic_flag: bool, +} + +impl<T, F, A: Allocator> DrainFilter<'_, T, F, A> +where + F: FnMut(&mut T) -> bool, +{ + /// Returns a reference to the underlying allocator. + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn allocator(&self) -> &A { + self.vec.allocator() + } +} + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +impl<T, F, A: Allocator> Iterator for DrainFilter<'_, T, F, A> +where + F: FnMut(&mut T) -> bool, +{ + type Item = T; + + fn next(&mut self) -> Option<T> { + unsafe { + while self.idx < self.old_len { + let i = self.idx; + let v = slice::from_raw_parts_mut(self.vec.as_mut_ptr(), self.old_len); + self.panic_flag = true; + let drained = (self.pred)(&mut v[i]); + self.panic_flag = false; + // Update the index *after* the predicate is called. If the index + // is updated prior and the predicate panics, the element at this + // index would be leaked. + self.idx += 1; + if drained { + self.del += 1; + return Some(ptr::read(&v[i])); + } else if self.del > 0 { + let del = self.del; + let src: *const T = &v[i]; + let dst: *mut T = &mut v[i - del]; + ptr::copy_nonoverlapping(src, dst, 1); + } + } + None + } + } + + fn size_hint(&self) -> (usize, Option<usize>) { + (0, Some(self.old_len - self.idx)) + } +} + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +impl<T, F, A: Allocator> Drop for DrainFilter<'_, T, F, A> +where + F: FnMut(&mut T) -> bool, +{ + fn drop(&mut self) { + struct BackshiftOnDrop<'a, 'b, T, F, A: Allocator> + where + F: FnMut(&mut T) -> bool, + { + drain: &'b mut DrainFilter<'a, T, F, A>, + } + + impl<'a, 'b, T, F, A: Allocator> Drop for BackshiftOnDrop<'a, 'b, T, F, A> + where + F: FnMut(&mut T) -> bool, + { + fn drop(&mut self) { + unsafe { + if self.drain.idx < self.drain.old_len && self.drain.del > 0 { + // This is a pretty messed up state, and there isn't really an + // obviously right thing to do. We don't want to keep trying + // to execute `pred`, so we just backshift all the unprocessed + // elements and tell the vec that they still exist. The backshift + // is required to prevent a double-drop of the last successfully + // drained item prior to a panic in the predicate. + let ptr = self.drain.vec.as_mut_ptr(); + let src = ptr.add(self.drain.idx); + let dst = src.sub(self.drain.del); + let tail_len = self.drain.old_len - self.drain.idx; + src.copy_to(dst, tail_len); + } + self.drain.vec.set_len(self.drain.old_len - self.drain.del); + } + } + } + + let backshift = BackshiftOnDrop { drain: self }; + + // Attempt to consume any remaining elements if the filter predicate + // has not yet panicked. We'll backshift any remaining elements + // whether we've already panicked or if the consumption here panics. + if !backshift.drain.panic_flag { + backshift.drain.for_each(drop); + } + } +} diff --git a/library/alloc/src/vec/in_place_collect.rs b/library/alloc/src/vec/in_place_collect.rs new file mode 100644 index 000000000..55dcb84ad --- /dev/null +++ b/library/alloc/src/vec/in_place_collect.rs @@ -0,0 +1,302 @@ +//! Inplace iterate-and-collect specialization for `Vec` +//! +//! Note: This documents Vec internals, some of the following sections explain implementation +//! details and are best read together with the source of this module. +//! +//! The specialization in this module applies to iterators in the shape of +//! `source.adapter().adapter().adapter().collect::<Vec<U>>()` +//! where `source` is an owning iterator obtained from [`Vec<T>`], [`Box<[T]>`][box] (by conversion to `Vec`) +//! or [`BinaryHeap<T>`], the adapters each consume one or more items per step +//! (represented by [`InPlaceIterable`]), provide transitive access to `source` (via [`SourceIter`]) +//! and thus the underlying allocation. And finally the layouts of `T` and `U` must +//! have the same size and alignment, this is currently ensured via const eval instead of trait bounds +//! in the specialized [`SpecFromIter`] implementation. +//! +//! [`BinaryHeap<T>`]: crate::collections::BinaryHeap +//! [box]: crate::boxed::Box +//! +//! By extension some other collections which use `collect::<Vec<_>>()` internally in their +//! `FromIterator` implementation benefit from this too. +//! +//! Access to the underlying source goes through a further layer of indirection via the private +//! trait [`AsVecIntoIter`] to hide the implementation detail that other collections may use +//! `vec::IntoIter` internally. +//! +//! In-place iteration depends on the interaction of several unsafe traits, implementation +//! details of multiple parts in the iterator pipeline and often requires holistic reasoning +//! across multiple structs since iterators are executed cooperatively rather than having +//! a central evaluator/visitor struct executing all iterator components. +//! +//! # Reading from and writing to the same allocation +//! +//! By its nature collecting in place means that the reader and writer side of the iterator +//! use the same allocation. Since `try_fold()` (used in [`SpecInPlaceCollect`]) takes a +//! reference to the iterator for the duration of the iteration that means we can't interleave +//! the step of reading a value and getting a reference to write to. Instead raw pointers must be +//! used on the reader and writer side. +//! +//! That writes never clobber a yet-to-be-read item is ensured by the [`InPlaceIterable`] requirements. +//! +//! # Layout constraints +//! +//! [`Allocator`] requires that `allocate()` and `deallocate()` have matching alignment and size. +//! Additionally this specialization doesn't make sense for ZSTs as there is no reallocation to +//! avoid and it would make pointer arithmetic more difficult. +//! +//! [`Allocator`]: core::alloc::Allocator +//! +//! # Drop- and panic-safety +//! +//! Iteration can panic, requiring dropping the already written parts but also the remainder of +//! the source. Iteration can also leave some source items unconsumed which must be dropped. +//! All those drops in turn can panic which then must either leak the allocation or abort to avoid +//! double-drops. +//! +//! This is handled by the [`InPlaceDrop`] guard for sink items (`U`) and by +//! [`vec::IntoIter::forget_allocation_drop_remaining()`] for remaining source items (`T`). +//! +//! [`vec::IntoIter::forget_allocation_drop_remaining()`]: super::IntoIter::forget_allocation_drop_remaining() +//! +//! # O(1) collect +//! +//! The main iteration itself is further specialized when the iterator implements +//! [`TrustedRandomAccessNoCoerce`] to let the optimizer see that it is a counted loop with a single +//! [induction variable]. This can turn some iterators into a noop, i.e. it reduces them from O(n) to +//! O(1). This particular optimization is quite fickle and doesn't always work, see [#79308] +//! +//! [#79308]: https://github.com/rust-lang/rust/issues/79308 +//! [induction variable]: https://en.wikipedia.org/wiki/Induction_variable +//! +//! Since unchecked accesses through that trait do not advance the read pointer of `IntoIter` +//! this would interact unsoundly with the requirements about dropping the tail described above. +//! But since the normal `Drop` implementation of `IntoIter` would suffer from the same problem it +//! is only correct for `TrustedRandomAccessNoCoerce` to be implemented when the items don't +//! have a destructor. Thus that implicit requirement also makes the specialization safe to use for +//! in-place collection. +//! Note that this safety concern is about the correctness of `impl Drop for IntoIter`, +//! not the guarantees of `InPlaceIterable`. +//! +//! # Adapter implementations +//! +//! The invariants for adapters are documented in [`SourceIter`] and [`InPlaceIterable`], but +//! getting them right can be rather subtle for multiple, sometimes non-local reasons. +//! For example `InPlaceIterable` would be valid to implement for [`Peekable`], except +//! that it is stateful, cloneable and `IntoIter`'s clone implementation shortens the underlying +//! allocation which means if the iterator has been peeked and then gets cloned there no longer is +//! enough room, thus breaking an invariant ([#85322]). +//! +//! [#85322]: https://github.com/rust-lang/rust/issues/85322 +//! [`Peekable`]: core::iter::Peekable +//! +//! +//! # Examples +//! +//! Some cases that are optimized by this specialization, more can be found in the `Vec` +//! benchmarks: +//! +//! ```rust +//! # #[allow(dead_code)] +//! /// Converts a usize vec into an isize one. +//! pub fn cast(vec: Vec<usize>) -> Vec<isize> { +//! // Does not allocate, free or panic. On optlevel>=2 it does not loop. +//! // Of course this particular case could and should be written with `into_raw_parts` and +//! // `from_raw_parts` instead. +//! vec.into_iter().map(|u| u as isize).collect() +//! } +//! ``` +//! +//! ```rust +//! # #[allow(dead_code)] +//! /// Drops remaining items in `src` and if the layouts of `T` and `U` match it +//! /// returns an empty Vec backed by the original allocation. Otherwise it returns a new +//! /// empty vec. +//! pub fn recycle_allocation<T, U>(src: Vec<T>) -> Vec<U> { +//! src.into_iter().filter_map(|_| None).collect() +//! } +//! ``` +//! +//! ```rust +//! let vec = vec![13usize; 1024]; +//! let _ = vec.into_iter() +//! .enumerate() +//! .filter_map(|(idx, val)| if idx % 2 == 0 { Some(val+idx) } else {None}) +//! .collect::<Vec<_>>(); +//! +//! // is equivalent to the following, but doesn't require bounds checks +//! +//! let mut vec = vec![13usize; 1024]; +//! let mut write_idx = 0; +//! for idx in 0..vec.len() { +//! if idx % 2 == 0 { +//! vec[write_idx] = vec[idx] + idx; +//! write_idx += 1; +//! } +//! } +//! vec.truncate(write_idx); +//! ``` +use core::iter::{InPlaceIterable, SourceIter, TrustedRandomAccessNoCoerce}; +use core::mem::{self, ManuallyDrop}; +use core::ptr::{self}; + +use super::{InPlaceDrop, SpecFromIter, SpecFromIterNested, Vec}; + +/// Specialization marker for collecting an iterator pipeline into a Vec while reusing the +/// source allocation, i.e. executing the pipeline in place. +#[rustc_unsafe_specialization_marker] +pub(super) trait InPlaceIterableMarker {} + +impl<T> InPlaceIterableMarker for T where T: InPlaceIterable {} + +impl<T, I> SpecFromIter<T, I> for Vec<T> +where + I: Iterator<Item = T> + SourceIter<Source: AsVecIntoIter> + InPlaceIterableMarker, +{ + default fn from_iter(mut iterator: I) -> Self { + // See "Layout constraints" section in the module documentation. We rely on const + // optimization here since these conditions currently cannot be expressed as trait bounds + if mem::size_of::<T>() == 0 + || mem::size_of::<T>() + != mem::size_of::<<<I as SourceIter>::Source as AsVecIntoIter>::Item>() + || mem::align_of::<T>() + != mem::align_of::<<<I as SourceIter>::Source as AsVecIntoIter>::Item>() + { + // fallback to more generic implementations + return SpecFromIterNested::from_iter(iterator); + } + + let (src_buf, src_ptr, dst_buf, dst_end, cap) = unsafe { + let inner = iterator.as_inner().as_into_iter(); + ( + inner.buf.as_ptr(), + inner.ptr, + inner.buf.as_ptr() as *mut T, + inner.end as *const T, + inner.cap, + ) + }; + + let len = SpecInPlaceCollect::collect_in_place(&mut iterator, dst_buf, dst_end); + + let src = unsafe { iterator.as_inner().as_into_iter() }; + // check if SourceIter contract was upheld + // caveat: if they weren't we might not even make it to this point + debug_assert_eq!(src_buf, src.buf.as_ptr()); + // check InPlaceIterable contract. This is only possible if the iterator advanced the + // source pointer at all. If it uses unchecked access via TrustedRandomAccess + // then the source pointer will stay in its initial position and we can't use it as reference + if src.ptr != src_ptr { + debug_assert!( + unsafe { dst_buf.add(len) as *const _ } <= src.ptr, + "InPlaceIterable contract violation, write pointer advanced beyond read pointer" + ); + } + + // Drop any remaining values at the tail of the source but prevent drop of the allocation + // itself once IntoIter goes out of scope. + // If the drop panics then we also leak any elements collected into dst_buf. + // + // Note: This access to the source wouldn't be allowed by the TrustedRandomIteratorNoCoerce + // contract (used by SpecInPlaceCollect below). But see the "O(1) collect" section in the + // module documenttation why this is ok anyway. + src.forget_allocation_drop_remaining(); + + let vec = unsafe { Vec::from_raw_parts(dst_buf, len, cap) }; + + vec + } +} + +fn write_in_place_with_drop<T>( + src_end: *const T, +) -> impl FnMut(InPlaceDrop<T>, T) -> Result<InPlaceDrop<T>, !> { + move |mut sink, item| { + unsafe { + // the InPlaceIterable contract cannot be verified precisely here since + // try_fold has an exclusive reference to the source pointer + // all we can do is check if it's still in range + debug_assert!(sink.dst as *const _ <= src_end, "InPlaceIterable contract violation"); + ptr::write(sink.dst, item); + // Since this executes user code which can panic we have to bump the pointer + // after each step. + sink.dst = sink.dst.add(1); + } + Ok(sink) + } +} + +/// Helper trait to hold specialized implementations of the in-place iterate-collect loop +trait SpecInPlaceCollect<T, I>: Iterator<Item = T> { + /// Collects an iterator (`self`) into the destination buffer (`dst`) and returns the number of items + /// collected. `end` is the last writable element of the allocation and used for bounds checks. + /// + /// This method is specialized and one of its implementations makes use of + /// `Iterator::__iterator_get_unchecked` calls with a `TrustedRandomAccessNoCoerce` bound + /// on `I` which means the caller of this method must take the safety conditions + /// of that trait into consideration. + fn collect_in_place(&mut self, dst: *mut T, end: *const T) -> usize; +} + +impl<T, I> SpecInPlaceCollect<T, I> for I +where + I: Iterator<Item = T>, +{ + #[inline] + default fn collect_in_place(&mut self, dst_buf: *mut T, end: *const T) -> usize { + // use try-fold since + // - it vectorizes better for some iterator adapters + // - unlike most internal iteration methods, it only takes a &mut self + // - it lets us thread the write pointer through its innards and get it back in the end + let sink = InPlaceDrop { inner: dst_buf, dst: dst_buf }; + let sink = + self.try_fold::<_, _, Result<_, !>>(sink, write_in_place_with_drop(end)).unwrap(); + // iteration succeeded, don't drop head + unsafe { ManuallyDrop::new(sink).dst.sub_ptr(dst_buf) } + } +} + +impl<T, I> SpecInPlaceCollect<T, I> for I +where + I: Iterator<Item = T> + TrustedRandomAccessNoCoerce, +{ + #[inline] + fn collect_in_place(&mut self, dst_buf: *mut T, end: *const T) -> usize { + let len = self.size(); + let mut drop_guard = InPlaceDrop { inner: dst_buf, dst: dst_buf }; + for i in 0..len { + // Safety: InplaceIterable contract guarantees that for every element we read + // one slot in the underlying storage will have been freed up and we can immediately + // write back the result. + unsafe { + let dst = dst_buf.offset(i as isize); + debug_assert!(dst as *const _ <= end, "InPlaceIterable contract violation"); + ptr::write(dst, self.__iterator_get_unchecked(i)); + // Since this executes user code which can panic we have to bump the pointer + // after each step. + drop_guard.dst = dst.add(1); + } + } + mem::forget(drop_guard); + len + } +} + +/// Internal helper trait for in-place iteration specialization. +/// +/// Currently this is only implemented by [`vec::IntoIter`] - returning a reference to itself - and +/// [`binary_heap::IntoIter`] which returns a reference to its inner representation. +/// +/// Since this is an internal trait it hides the implementation detail `binary_heap::IntoIter` +/// uses `vec::IntoIter` internally. +/// +/// [`vec::IntoIter`]: super::IntoIter +/// [`binary_heap::IntoIter`]: crate::collections::binary_heap::IntoIter +/// +/// # Safety +/// +/// In-place iteration relies on implementation details of `vec::IntoIter`, most importantly that +/// it does not create references to the whole allocation during iteration, only raw pointers +#[rustc_specialization_trait] +pub(crate) unsafe trait AsVecIntoIter { + type Item; + fn as_into_iter(&mut self) -> &mut super::IntoIter<Self::Item>; +} diff --git a/library/alloc/src/vec/in_place_drop.rs b/library/alloc/src/vec/in_place_drop.rs new file mode 100644 index 000000000..1b1ef9130 --- /dev/null +++ b/library/alloc/src/vec/in_place_drop.rs @@ -0,0 +1,24 @@ +use core::ptr::{self}; +use core::slice::{self}; + +// A helper struct for in-place iteration that drops the destination slice of iteration, +// i.e. the head. The source slice (the tail) is dropped by IntoIter. +pub(super) struct InPlaceDrop<T> { + pub(super) inner: *mut T, + pub(super) dst: *mut T, +} + +impl<T> InPlaceDrop<T> { + fn len(&self) -> usize { + unsafe { self.dst.sub_ptr(self.inner) } + } +} + +impl<T> Drop for InPlaceDrop<T> { + #[inline] + fn drop(&mut self) { + unsafe { + ptr::drop_in_place(slice::from_raw_parts_mut(self.inner, self.len())); + } + } +} diff --git a/library/alloc/src/vec/into_iter.rs b/library/alloc/src/vec/into_iter.rs new file mode 100644 index 000000000..1b483e3fc --- /dev/null +++ b/library/alloc/src/vec/into_iter.rs @@ -0,0 +1,401 @@ +#[cfg(not(no_global_oom_handling))] +use super::AsVecIntoIter; +use crate::alloc::{Allocator, Global}; +use crate::raw_vec::RawVec; +use core::array; +use core::fmt; +use core::intrinsics::arith_offset; +use core::iter::{ + FusedIterator, InPlaceIterable, SourceIter, TrustedLen, TrustedRandomAccessNoCoerce, +}; +use core::marker::PhantomData; +use core::mem::{self, ManuallyDrop, MaybeUninit}; +#[cfg(not(no_global_oom_handling))] +use core::ops::Deref; +use core::ptr::{self, NonNull}; +use core::slice::{self}; + +/// An iterator that moves out of a vector. +/// +/// This `struct` is created by the `into_iter` method on [`Vec`](super::Vec) +/// (provided by the [`IntoIterator`] trait). +/// +/// # Example +/// +/// ``` +/// let v = vec![0, 1, 2]; +/// let iter: std::vec::IntoIter<_> = v.into_iter(); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_insignificant_dtor] +pub struct IntoIter< + T, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global, +> { + pub(super) buf: NonNull<T>, + pub(super) phantom: PhantomData<T>, + pub(super) cap: usize, + // the drop impl reconstructs a RawVec from buf, cap and alloc + // to avoid dropping the allocator twice we need to wrap it into ManuallyDrop + pub(super) alloc: ManuallyDrop<A>, + pub(super) ptr: *const T, + pub(super) end: *const T, +} + +#[stable(feature = "vec_intoiter_debug", since = "1.13.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for IntoIter<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("IntoIter").field(&self.as_slice()).finish() + } +} + +impl<T, A: Allocator> IntoIter<T, A> { + /// Returns the remaining items of this iterator as a slice. + /// + /// # Examples + /// + /// ``` + /// let vec = vec!['a', 'b', 'c']; + /// let mut into_iter = vec.into_iter(); + /// assert_eq!(into_iter.as_slice(), &['a', 'b', 'c']); + /// let _ = into_iter.next().unwrap(); + /// assert_eq!(into_iter.as_slice(), &['b', 'c']); + /// ``` + #[stable(feature = "vec_into_iter_as_slice", since = "1.15.0")] + pub fn as_slice(&self) -> &[T] { + unsafe { slice::from_raw_parts(self.ptr, self.len()) } + } + + /// Returns the remaining items of this iterator as a mutable slice. + /// + /// # Examples + /// + /// ``` + /// let vec = vec!['a', 'b', 'c']; + /// let mut into_iter = vec.into_iter(); + /// assert_eq!(into_iter.as_slice(), &['a', 'b', 'c']); + /// into_iter.as_mut_slice()[2] = 'z'; + /// assert_eq!(into_iter.next().unwrap(), 'a'); + /// assert_eq!(into_iter.next().unwrap(), 'b'); + /// assert_eq!(into_iter.next().unwrap(), 'z'); + /// ``` + #[stable(feature = "vec_into_iter_as_slice", since = "1.15.0")] + pub fn as_mut_slice(&mut self) -> &mut [T] { + unsafe { &mut *self.as_raw_mut_slice() } + } + + /// Returns a reference to the underlying allocator. + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn allocator(&self) -> &A { + &self.alloc + } + + fn as_raw_mut_slice(&mut self) -> *mut [T] { + ptr::slice_from_raw_parts_mut(self.ptr as *mut T, self.len()) + } + + /// Drops remaining elements and relinquishes the backing allocation. + /// + /// This is roughly equivalent to the following, but more efficient + /// + /// ``` + /// # let mut into_iter = Vec::<u8>::with_capacity(10).into_iter(); + /// (&mut into_iter).for_each(core::mem::drop); + /// unsafe { core::ptr::write(&mut into_iter, Vec::new().into_iter()); } + /// ``` + /// + /// This method is used by in-place iteration, refer to the vec::in_place_collect + /// documentation for an overview. + #[cfg(not(no_global_oom_handling))] + pub(super) fn forget_allocation_drop_remaining(&mut self) { + let remaining = self.as_raw_mut_slice(); + + // overwrite the individual fields instead of creating a new + // struct and then overwriting &mut self. + // this creates less assembly + self.cap = 0; + self.buf = unsafe { NonNull::new_unchecked(RawVec::NEW.ptr()) }; + self.ptr = self.buf.as_ptr(); + self.end = self.buf.as_ptr(); + + unsafe { + ptr::drop_in_place(remaining); + } + } + + /// Forgets to Drop the remaining elements while still allowing the backing allocation to be freed. + pub(crate) fn forget_remaining_elements(&mut self) { + self.ptr = self.end; + } +} + +#[stable(feature = "vec_intoiter_as_ref", since = "1.46.0")] +impl<T, A: Allocator> AsRef<[T]> for IntoIter<T, A> { + fn as_ref(&self) -> &[T] { + self.as_slice() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Send, A: Allocator + Send> Send for IntoIter<T, A> {} +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<T: Sync, A: Allocator + Sync> Sync for IntoIter<T, A> {} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Iterator for IntoIter<T, A> { + type Item = T; + + #[inline] + fn next(&mut self) -> Option<T> { + if self.ptr as *const _ == self.end { + None + } else if mem::size_of::<T>() == 0 { + // purposefully don't use 'ptr.offset' because for + // vectors with 0-size elements this would return the + // same pointer. + self.ptr = unsafe { arith_offset(self.ptr as *const i8, 1) as *mut T }; + + // Make up a value of this ZST. + Some(unsafe { mem::zeroed() }) + } else { + let old = self.ptr; + self.ptr = unsafe { self.ptr.offset(1) }; + + Some(unsafe { ptr::read(old) }) + } + } + + #[inline] + fn size_hint(&self) -> (usize, Option<usize>) { + let exact = if mem::size_of::<T>() == 0 { + self.end.addr().wrapping_sub(self.ptr.addr()) + } else { + unsafe { self.end.sub_ptr(self.ptr) } + }; + (exact, Some(exact)) + } + + #[inline] + fn advance_by(&mut self, n: usize) -> Result<(), usize> { + let step_size = self.len().min(n); + let to_drop = ptr::slice_from_raw_parts_mut(self.ptr as *mut T, step_size); + if mem::size_of::<T>() == 0 { + // SAFETY: due to unchecked casts of unsigned amounts to signed offsets the wraparound + // effectively results in unsigned pointers representing positions 0..usize::MAX, + // which is valid for ZSTs. + self.ptr = unsafe { arith_offset(self.ptr as *const i8, step_size as isize) as *mut T } + } else { + // SAFETY: the min() above ensures that step_size is in bounds + self.ptr = unsafe { self.ptr.add(step_size) }; + } + // SAFETY: the min() above ensures that step_size is in bounds + unsafe { + ptr::drop_in_place(to_drop); + } + if step_size < n { + return Err(step_size); + } + Ok(()) + } + + #[inline] + fn count(self) -> usize { + self.len() + } + + #[inline] + fn next_chunk<const N: usize>(&mut self) -> Result<[T; N], core::array::IntoIter<T, N>> { + let mut raw_ary = MaybeUninit::uninit_array(); + + let len = self.len(); + + if mem::size_of::<T>() == 0 { + if len < N { + self.forget_remaining_elements(); + // Safety: ZSTs can be conjured ex nihilo, only the amount has to be correct + return Err(unsafe { array::IntoIter::new_unchecked(raw_ary, 0..len) }); + } + + self.ptr = unsafe { arith_offset(self.ptr as *const i8, N as isize) as *mut T }; + // Safety: ditto + return Ok(unsafe { MaybeUninit::array_assume_init(raw_ary) }); + } + + if len < N { + // Safety: `len` indicates that this many elements are available and we just checked that + // it fits into the array. + unsafe { + ptr::copy_nonoverlapping(self.ptr, raw_ary.as_mut_ptr() as *mut T, len); + self.forget_remaining_elements(); + return Err(array::IntoIter::new_unchecked(raw_ary, 0..len)); + } + } + + // Safety: `len` is larger than the array size. Copy a fixed amount here to fully initialize + // the array. + return unsafe { + ptr::copy_nonoverlapping(self.ptr, raw_ary.as_mut_ptr() as *mut T, N); + self.ptr = self.ptr.add(N); + Ok(MaybeUninit::array_assume_init(raw_ary)) + }; + } + + unsafe fn __iterator_get_unchecked(&mut self, i: usize) -> Self::Item + where + Self: TrustedRandomAccessNoCoerce, + { + // SAFETY: the caller must guarantee that `i` is in bounds of the + // `Vec<T>`, so `i` cannot overflow an `isize`, and the `self.ptr.add(i)` + // is guaranteed to pointer to an element of the `Vec<T>` and + // thus guaranteed to be valid to dereference. + // + // Also note the implementation of `Self: TrustedRandomAccess` requires + // that `T: Copy` so reading elements from the buffer doesn't invalidate + // them for `Drop`. + unsafe { + if mem::size_of::<T>() == 0 { mem::zeroed() } else { ptr::read(self.ptr.add(i)) } + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> DoubleEndedIterator for IntoIter<T, A> { + #[inline] + fn next_back(&mut self) -> Option<T> { + if self.end == self.ptr { + None + } else if mem::size_of::<T>() == 0 { + // See above for why 'ptr.offset' isn't used + self.end = unsafe { arith_offset(self.end as *const i8, -1) as *mut T }; + + // Make up a value of this ZST. + Some(unsafe { mem::zeroed() }) + } else { + self.end = unsafe { self.end.offset(-1) }; + + Some(unsafe { ptr::read(self.end) }) + } + } + + #[inline] + fn advance_back_by(&mut self, n: usize) -> Result<(), usize> { + let step_size = self.len().min(n); + if mem::size_of::<T>() == 0 { + // SAFETY: same as for advance_by() + self.end = unsafe { + arith_offset(self.end as *const i8, step_size.wrapping_neg() as isize) as *mut T + } + } else { + // SAFETY: same as for advance_by() + self.end = unsafe { self.end.offset(step_size.wrapping_neg() as isize) }; + } + let to_drop = ptr::slice_from_raw_parts_mut(self.end as *mut T, step_size); + // SAFETY: same as for advance_by() + unsafe { + ptr::drop_in_place(to_drop); + } + if step_size < n { + return Err(step_size); + } + Ok(()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> ExactSizeIterator for IntoIter<T, A> { + fn is_empty(&self) -> bool { + self.ptr == self.end + } +} + +#[stable(feature = "fused", since = "1.26.0")] +impl<T, A: Allocator> FusedIterator for IntoIter<T, A> {} + +#[unstable(feature = "trusted_len", issue = "37572")] +unsafe impl<T, A: Allocator> TrustedLen for IntoIter<T, A> {} + +#[doc(hidden)] +#[unstable(issue = "none", feature = "std_internals")] +#[rustc_unsafe_specialization_marker] +pub trait NonDrop {} + +// T: Copy as approximation for !Drop since get_unchecked does not advance self.ptr +// and thus we can't implement drop-handling +#[unstable(issue = "none", feature = "std_internals")] +impl<T: Copy> NonDrop for T {} + +#[doc(hidden)] +#[unstable(issue = "none", feature = "std_internals")] +// TrustedRandomAccess (without NoCoerce) must not be implemented because +// subtypes/supertypes of `T` might not be `NonDrop` +unsafe impl<T, A: Allocator> TrustedRandomAccessNoCoerce for IntoIter<T, A> +where + T: NonDrop, +{ + const MAY_HAVE_SIDE_EFFECT: bool = false; +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "vec_into_iter_clone", since = "1.8.0")] +impl<T: Clone, A: Allocator + Clone> Clone for IntoIter<T, A> { + #[cfg(not(test))] + fn clone(&self) -> Self { + self.as_slice().to_vec_in(self.alloc.deref().clone()).into_iter() + } + #[cfg(test)] + fn clone(&self) -> Self { + crate::slice::to_vec(self.as_slice(), self.alloc.deref().clone()).into_iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T, A: Allocator> Drop for IntoIter<T, A> { + fn drop(&mut self) { + struct DropGuard<'a, T, A: Allocator>(&'a mut IntoIter<T, A>); + + impl<T, A: Allocator> Drop for DropGuard<'_, T, A> { + fn drop(&mut self) { + unsafe { + // `IntoIter::alloc` is not used anymore after this and will be dropped by RawVec + let alloc = ManuallyDrop::take(&mut self.0.alloc); + // RawVec handles deallocation + let _ = RawVec::from_raw_parts_in(self.0.buf.as_ptr(), self.0.cap, alloc); + } + } + } + + let guard = DropGuard(self); + // destroy the remaining elements + unsafe { + ptr::drop_in_place(guard.0.as_raw_mut_slice()); + } + // now `guard` will be dropped and do the rest + } +} + +// In addition to the SAFETY invariants of the following three unsafe traits +// also refer to the vec::in_place_collect module documentation to get an overview +#[unstable(issue = "none", feature = "inplace_iteration")] +#[doc(hidden)] +unsafe impl<T, A: Allocator> InPlaceIterable for IntoIter<T, A> {} + +#[unstable(issue = "none", feature = "inplace_iteration")] +#[doc(hidden)] +unsafe impl<T, A: Allocator> SourceIter for IntoIter<T, A> { + type Source = Self; + + #[inline] + unsafe fn as_inner(&mut self) -> &mut Self::Source { + self + } +} + +#[cfg(not(no_global_oom_handling))] +unsafe impl<T> AsVecIntoIter for IntoIter<T> { + type Item = T; + + fn as_into_iter(&mut self) -> &mut IntoIter<Self::Item> { + self + } +} diff --git a/library/alloc/src/vec/is_zero.rs b/library/alloc/src/vec/is_zero.rs new file mode 100644 index 000000000..92a32779b --- /dev/null +++ b/library/alloc/src/vec/is_zero.rs @@ -0,0 +1,146 @@ +use crate::boxed::Box; + +#[rustc_specialization_trait] +pub(super) unsafe trait IsZero { + /// Whether this value's representation is all zeros + fn is_zero(&self) -> bool; +} + +macro_rules! impl_is_zero { + ($t:ty, $is_zero:expr) => { + unsafe impl IsZero for $t { + #[inline] + fn is_zero(&self) -> bool { + $is_zero(*self) + } + } + }; +} + +impl_is_zero!(i8, |x| x == 0); // It is needed to impl for arrays and tuples of i8. +impl_is_zero!(i16, |x| x == 0); +impl_is_zero!(i32, |x| x == 0); +impl_is_zero!(i64, |x| x == 0); +impl_is_zero!(i128, |x| x == 0); +impl_is_zero!(isize, |x| x == 0); + +impl_is_zero!(u8, |x| x == 0); // It is needed to impl for arrays and tuples of u8. +impl_is_zero!(u16, |x| x == 0); +impl_is_zero!(u32, |x| x == 0); +impl_is_zero!(u64, |x| x == 0); +impl_is_zero!(u128, |x| x == 0); +impl_is_zero!(usize, |x| x == 0); + +impl_is_zero!(bool, |x| x == false); +impl_is_zero!(char, |x| x == '\0'); + +impl_is_zero!(f32, |x: f32| x.to_bits() == 0); +impl_is_zero!(f64, |x: f64| x.to_bits() == 0); + +unsafe impl<T> IsZero for *const T { + #[inline] + fn is_zero(&self) -> bool { + (*self).is_null() + } +} + +unsafe impl<T> IsZero for *mut T { + #[inline] + fn is_zero(&self) -> bool { + (*self).is_null() + } +} + +unsafe impl<T: IsZero, const N: usize> IsZero for [T; N] { + #[inline] + fn is_zero(&self) -> bool { + // Because this is generated as a runtime check, it's not obvious that + // it's worth doing if the array is really long. The threshold here + // is largely arbitrary, but was picked because as of 2022-07-01 LLVM + // fails to const-fold the check in `vec![[1; 32]; n]` + // See https://github.com/rust-lang/rust/pull/97581#issuecomment-1166628022 + // Feel free to tweak if you have better evidence. + + N <= 16 && self.iter().all(IsZero::is_zero) + } +} + +// This is recursive macro. +macro_rules! impl_for_tuples { + // Stopper + () => { + // No use for implementing for empty tuple because it is ZST. + }; + ($first_arg:ident $(,$rest:ident)*) => { + unsafe impl <$first_arg: IsZero, $($rest: IsZero,)*> IsZero for ($first_arg, $($rest,)*){ + #[inline] + fn is_zero(&self) -> bool{ + // Destructure tuple to N references + // Rust allows to hide generic params by local variable names. + #[allow(non_snake_case)] + let ($first_arg, $($rest,)*) = self; + + $first_arg.is_zero() + $( && $rest.is_zero() )* + } + } + + impl_for_tuples!($($rest),*); + } +} + +impl_for_tuples!(A, B, C, D, E, F, G, H); + +// `Option<&T>` and `Option<Box<T>>` are guaranteed to represent `None` as null. +// For fat pointers, the bytes that would be the pointer metadata in the `Some` +// variant are padding in the `None` variant, so ignoring them and +// zero-initializing instead is ok. +// `Option<&mut T>` never implements `Clone`, so there's no need for an impl of +// `SpecFromElem`. + +unsafe impl<T: ?Sized> IsZero for Option<&T> { + #[inline] + fn is_zero(&self) -> bool { + self.is_none() + } +} + +unsafe impl<T: ?Sized> IsZero for Option<Box<T>> { + #[inline] + fn is_zero(&self) -> bool { + self.is_none() + } +} + +// `Option<num::NonZeroU32>` and similar have a representation guarantee that +// they're the same size as the corresponding `u32` type, as well as a guarantee +// that transmuting between `NonZeroU32` and `Option<num::NonZeroU32>` works. +// While the documentation officially makes it UB to transmute from `None`, +// we're the standard library so we can make extra inferences, and we know that +// the only niche available to represent `None` is the one that's all zeros. + +macro_rules! impl_is_zero_option_of_nonzero { + ($($t:ident,)+) => {$( + unsafe impl IsZero for Option<core::num::$t> { + #[inline] + fn is_zero(&self) -> bool { + self.is_none() + } + } + )+}; +} + +impl_is_zero_option_of_nonzero!( + NonZeroU8, + NonZeroU16, + NonZeroU32, + NonZeroU64, + NonZeroU128, + NonZeroI8, + NonZeroI16, + NonZeroI32, + NonZeroI64, + NonZeroI128, + NonZeroUsize, + NonZeroIsize, +); diff --git a/library/alloc/src/vec/mod.rs b/library/alloc/src/vec/mod.rs new file mode 100644 index 000000000..fa9f2131c --- /dev/null +++ b/library/alloc/src/vec/mod.rs @@ -0,0 +1,3157 @@ +//! A contiguous growable array type with heap-allocated contents, written +//! `Vec<T>`. +//! +//! Vectors have *O*(1) indexing, amortized *O*(1) push (to the end) and +//! *O*(1) pop (from the end). +//! +//! Vectors ensure they never allocate more than `isize::MAX` bytes. +//! +//! # Examples +//! +//! You can explicitly create a [`Vec`] with [`Vec::new`]: +//! +//! ``` +//! let v: Vec<i32> = Vec::new(); +//! ``` +//! +//! ...or by using the [`vec!`] macro: +//! +//! ``` +//! let v: Vec<i32> = vec![]; +//! +//! let v = vec![1, 2, 3, 4, 5]; +//! +//! let v = vec![0; 10]; // ten zeroes +//! ``` +//! +//! You can [`push`] values onto the end of a vector (which will grow the vector +//! as needed): +//! +//! ``` +//! let mut v = vec![1, 2]; +//! +//! v.push(3); +//! ``` +//! +//! Popping values works in much the same way: +//! +//! ``` +//! let mut v = vec![1, 2]; +//! +//! let two = v.pop(); +//! ``` +//! +//! Vectors also support indexing (through the [`Index`] and [`IndexMut`] traits): +//! +//! ``` +//! let mut v = vec![1, 2, 3]; +//! let three = v[2]; +//! v[1] = v[1] + 5; +//! ``` +//! +//! [`push`]: Vec::push + +#![stable(feature = "rust1", since = "1.0.0")] + +#[cfg(not(no_global_oom_handling))] +use core::cmp; +use core::cmp::Ordering; +use core::convert::TryFrom; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::intrinsics::{arith_offset, assume}; +use core::iter; +#[cfg(not(no_global_oom_handling))] +use core::iter::FromIterator; +use core::marker::PhantomData; +use core::mem::{self, ManuallyDrop, MaybeUninit}; +use core::ops::{self, Index, IndexMut, Range, RangeBounds}; +use core::ptr::{self, NonNull}; +use core::slice::{self, SliceIndex}; + +use crate::alloc::{Allocator, Global}; +use crate::borrow::{Cow, ToOwned}; +use crate::boxed::Box; +use crate::collections::TryReserveError; +use crate::raw_vec::RawVec; + +#[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] +pub use self::drain_filter::DrainFilter; + +mod drain_filter; + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "vec_splice", since = "1.21.0")] +pub use self::splice::Splice; + +#[cfg(not(no_global_oom_handling))] +mod splice; + +#[stable(feature = "drain", since = "1.6.0")] +pub use self::drain::Drain; + +mod drain; + +#[cfg(not(no_global_oom_handling))] +mod cow; + +#[cfg(not(no_global_oom_handling))] +pub(crate) use self::in_place_collect::AsVecIntoIter; +#[stable(feature = "rust1", since = "1.0.0")] +pub use self::into_iter::IntoIter; + +mod into_iter; + +#[cfg(not(no_global_oom_handling))] +use self::is_zero::IsZero; + +mod is_zero; + +#[cfg(not(no_global_oom_handling))] +mod in_place_collect; + +mod partial_eq; + +#[cfg(not(no_global_oom_handling))] +use self::spec_from_elem::SpecFromElem; + +#[cfg(not(no_global_oom_handling))] +mod spec_from_elem; + +#[cfg(not(no_global_oom_handling))] +use self::set_len_on_drop::SetLenOnDrop; + +#[cfg(not(no_global_oom_handling))] +mod set_len_on_drop; + +#[cfg(not(no_global_oom_handling))] +use self::in_place_drop::InPlaceDrop; + +#[cfg(not(no_global_oom_handling))] +mod in_place_drop; + +#[cfg(not(no_global_oom_handling))] +use self::spec_from_iter_nested::SpecFromIterNested; + +#[cfg(not(no_global_oom_handling))] +mod spec_from_iter_nested; + +#[cfg(not(no_global_oom_handling))] +use self::spec_from_iter::SpecFromIter; + +#[cfg(not(no_global_oom_handling))] +mod spec_from_iter; + +#[cfg(not(no_global_oom_handling))] +use self::spec_extend::SpecExtend; + +#[cfg(not(no_global_oom_handling))] +mod spec_extend; + +/// A contiguous growable array type, written as `Vec<T>`, short for 'vector'. +/// +/// # Examples +/// +/// ``` +/// let mut vec = Vec::new(); +/// vec.push(1); +/// vec.push(2); +/// +/// assert_eq!(vec.len(), 2); +/// assert_eq!(vec[0], 1); +/// +/// assert_eq!(vec.pop(), Some(2)); +/// assert_eq!(vec.len(), 1); +/// +/// vec[0] = 7; +/// assert_eq!(vec[0], 7); +/// +/// vec.extend([1, 2, 3].iter().copied()); +/// +/// for x in &vec { +/// println!("{x}"); +/// } +/// assert_eq!(vec, [7, 1, 2, 3]); +/// ``` +/// +/// The [`vec!`] macro is provided for convenient initialization: +/// +/// ``` +/// let mut vec1 = vec![1, 2, 3]; +/// vec1.push(4); +/// let vec2 = Vec::from([1, 2, 3, 4]); +/// assert_eq!(vec1, vec2); +/// ``` +/// +/// It can also initialize each element of a `Vec<T>` with a given value. +/// This may be more efficient than performing allocation and initialization +/// in separate steps, especially when initializing a vector of zeros: +/// +/// ``` +/// let vec = vec![0; 5]; +/// assert_eq!(vec, [0, 0, 0, 0, 0]); +/// +/// // The following is equivalent, but potentially slower: +/// let mut vec = Vec::with_capacity(5); +/// vec.resize(5, 0); +/// assert_eq!(vec, [0, 0, 0, 0, 0]); +/// ``` +/// +/// For more information, see +/// [Capacity and Reallocation](#capacity-and-reallocation). +/// +/// Use a `Vec<T>` as an efficient stack: +/// +/// ``` +/// let mut stack = Vec::new(); +/// +/// stack.push(1); +/// stack.push(2); +/// stack.push(3); +/// +/// while let Some(top) = stack.pop() { +/// // Prints 3, 2, 1 +/// println!("{top}"); +/// } +/// ``` +/// +/// # Indexing +/// +/// The `Vec` type allows to access values by index, because it implements the +/// [`Index`] trait. An example will be more explicit: +/// +/// ``` +/// let v = vec![0, 2, 4, 6]; +/// println!("{}", v[1]); // it will display '2' +/// ``` +/// +/// However be careful: if you try to access an index which isn't in the `Vec`, +/// your software will panic! You cannot do this: +/// +/// ```should_panic +/// let v = vec![0, 2, 4, 6]; +/// println!("{}", v[6]); // it will panic! +/// ``` +/// +/// Use [`get`] and [`get_mut`] if you want to check whether the index is in +/// the `Vec`. +/// +/// # Slicing +/// +/// A `Vec` can be mutable. On the other hand, slices are read-only objects. +/// To get a [slice][prim@slice], use [`&`]. Example: +/// +/// ``` +/// fn read_slice(slice: &[usize]) { +/// // ... +/// } +/// +/// let v = vec![0, 1]; +/// read_slice(&v); +/// +/// // ... and that's all! +/// // you can also do it like this: +/// let u: &[usize] = &v; +/// // or like this: +/// let u: &[_] = &v; +/// ``` +/// +/// In Rust, it's more common to pass slices as arguments rather than vectors +/// when you just want to provide read access. The same goes for [`String`] and +/// [`&str`]. +/// +/// # Capacity and reallocation +/// +/// The capacity of a vector is the amount of space allocated for any future +/// elements that will be added onto the vector. This is not to be confused with +/// the *length* of a vector, which specifies the number of actual elements +/// within the vector. If a vector's length exceeds its capacity, its capacity +/// will automatically be increased, but its elements will have to be +/// reallocated. +/// +/// For example, a vector with capacity 10 and length 0 would be an empty vector +/// with space for 10 more elements. Pushing 10 or fewer elements onto the +/// vector will not change its capacity or cause reallocation to occur. However, +/// if the vector's length is increased to 11, it will have to reallocate, which +/// can be slow. For this reason, it is recommended to use [`Vec::with_capacity`] +/// whenever possible to specify how big the vector is expected to get. +/// +/// # Guarantees +/// +/// Due to its incredibly fundamental nature, `Vec` makes a lot of guarantees +/// about its design. This ensures that it's as low-overhead as possible in +/// the general case, and can be correctly manipulated in primitive ways +/// by unsafe code. Note that these guarantees refer to an unqualified `Vec<T>`. +/// If additional type parameters are added (e.g., to support custom allocators), +/// overriding their defaults may change the behavior. +/// +/// Most fundamentally, `Vec` is and always will be a (pointer, capacity, length) +/// triplet. No more, no less. The order of these fields is completely +/// unspecified, and you should use the appropriate methods to modify these. +/// The pointer will never be null, so this type is null-pointer-optimized. +/// +/// However, the pointer might not actually point to allocated memory. In particular, +/// if you construct a `Vec` with capacity 0 via [`Vec::new`], [`vec![]`][`vec!`], +/// [`Vec::with_capacity(0)`][`Vec::with_capacity`], or by calling [`shrink_to_fit`] +/// on an empty Vec, it will not allocate memory. Similarly, if you store zero-sized +/// types inside a `Vec`, it will not allocate space for them. *Note that in this case +/// the `Vec` might not report a [`capacity`] of 0*. `Vec` will allocate if and only +/// if <code>[mem::size_of::\<T>]\() * [capacity]\() > 0</code>. In general, `Vec`'s allocation +/// details are very subtle --- if you intend to allocate memory using a `Vec` +/// and use it for something else (either to pass to unsafe code, or to build your +/// own memory-backed collection), be sure to deallocate this memory by using +/// `from_raw_parts` to recover the `Vec` and then dropping it. +/// +/// If a `Vec` *has* allocated memory, then the memory it points to is on the heap +/// (as defined by the allocator Rust is configured to use by default), and its +/// pointer points to [`len`] initialized, contiguous elements in order (what +/// you would see if you coerced it to a slice), followed by <code>[capacity] - [len]</code> +/// logically uninitialized, contiguous elements. +/// +/// A vector containing the elements `'a'` and `'b'` with capacity 4 can be +/// visualized as below. The top part is the `Vec` struct, it contains a +/// pointer to the head of the allocation in the heap, length and capacity. +/// The bottom part is the allocation on the heap, a contiguous memory block. +/// +/// ```text +/// ptr len capacity +/// +--------+--------+--------+ +/// | 0x0123 | 2 | 4 | +/// +--------+--------+--------+ +/// | +/// v +/// Heap +--------+--------+--------+--------+ +/// | 'a' | 'b' | uninit | uninit | +/// +--------+--------+--------+--------+ +/// ``` +/// +/// - **uninit** represents memory that is not initialized, see [`MaybeUninit`]. +/// - Note: the ABI is not stable and `Vec` makes no guarantees about its memory +/// layout (including the order of fields). +/// +/// `Vec` will never perform a "small optimization" where elements are actually +/// stored on the stack for two reasons: +/// +/// * It would make it more difficult for unsafe code to correctly manipulate +/// a `Vec`. The contents of a `Vec` wouldn't have a stable address if it were +/// only moved, and it would be more difficult to determine if a `Vec` had +/// actually allocated memory. +/// +/// * It would penalize the general case, incurring an additional branch +/// on every access. +/// +/// `Vec` will never automatically shrink itself, even if completely empty. This +/// ensures no unnecessary allocations or deallocations occur. Emptying a `Vec` +/// and then filling it back up to the same [`len`] should incur no calls to +/// the allocator. If you wish to free up unused memory, use +/// [`shrink_to_fit`] or [`shrink_to`]. +/// +/// [`push`] and [`insert`] will never (re)allocate if the reported capacity is +/// sufficient. [`push`] and [`insert`] *will* (re)allocate if +/// <code>[len] == [capacity]</code>. That is, the reported capacity is completely +/// accurate, and can be relied on. It can even be used to manually free the memory +/// allocated by a `Vec` if desired. Bulk insertion methods *may* reallocate, even +/// when not necessary. +/// +/// `Vec` does not guarantee any particular growth strategy when reallocating +/// when full, nor when [`reserve`] is called. The current strategy is basic +/// and it may prove desirable to use a non-constant growth factor. Whatever +/// strategy is used will of course guarantee *O*(1) amortized [`push`]. +/// +/// `vec![x; n]`, `vec![a, b, c, d]`, and +/// [`Vec::with_capacity(n)`][`Vec::with_capacity`], will all produce a `Vec` +/// with exactly the requested capacity. If <code>[len] == [capacity]</code>, +/// (as is the case for the [`vec!`] macro), then a `Vec<T>` can be converted to +/// and from a [`Box<[T]>`][owned slice] without reallocating or moving the elements. +/// +/// `Vec` will not specifically overwrite any data that is removed from it, +/// but also won't specifically preserve it. Its uninitialized memory is +/// scratch space that it may use however it wants. It will generally just do +/// whatever is most efficient or otherwise easy to implement. Do not rely on +/// removed data to be erased for security purposes. Even if you drop a `Vec`, its +/// buffer may simply be reused by another allocation. Even if you zero a `Vec`'s memory +/// first, that might not actually happen because the optimizer does not consider +/// this a side-effect that must be preserved. There is one case which we will +/// not break, however: using `unsafe` code to write to the excess capacity, +/// and then increasing the length to match, is always valid. +/// +/// Currently, `Vec` does not guarantee the order in which elements are dropped. +/// The order has changed in the past and may change again. +/// +/// [`get`]: ../../std/vec/struct.Vec.html#method.get +/// [`get_mut`]: ../../std/vec/struct.Vec.html#method.get_mut +/// [`String`]: crate::string::String +/// [`&str`]: type@str +/// [`shrink_to_fit`]: Vec::shrink_to_fit +/// [`shrink_to`]: Vec::shrink_to +/// [capacity]: Vec::capacity +/// [`capacity`]: Vec::capacity +/// [mem::size_of::\<T>]: core::mem::size_of +/// [len]: Vec::len +/// [`len`]: Vec::len +/// [`push`]: Vec::push +/// [`insert`]: Vec::insert +/// [`reserve`]: Vec::reserve +/// [`MaybeUninit`]: core::mem::MaybeUninit +/// [owned slice]: Box +#[stable(feature = "rust1", since = "1.0.0")] +#[cfg_attr(not(test), rustc_diagnostic_item = "Vec")] +#[rustc_insignificant_dtor] +pub struct Vec<T, #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global> { + buf: RawVec<T, A>, + len: usize, +} + +//////////////////////////////////////////////////////////////////////////////// +// Inherent methods +//////////////////////////////////////////////////////////////////////////////// + +impl<T> Vec<T> { + /// Constructs a new, empty `Vec<T>`. + /// + /// The vector will not allocate until elements are pushed onto it. + /// + /// # Examples + /// + /// ``` + /// # #![allow(unused_mut)] + /// let mut vec: Vec<i32> = Vec::new(); + /// ``` + #[inline] + #[rustc_const_stable(feature = "const_vec_new", since = "1.39.0")] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub const fn new() -> Self { + Vec { buf: RawVec::NEW, len: 0 } + } + + /// Constructs a new, empty `Vec<T>` with at least the specified capacity. + /// + /// The vector will be able to hold at least `capacity` elements without + /// reallocating. This method is allowed to allocate for more elements than + /// `capacity`. If `capacity` is 0, the vector will not allocate. + /// + /// It is important to note that although the returned vector has the + /// minimum *capacity* specified, the vector will have a zero *length*. For + /// an explanation of the difference between length and capacity, see + /// *[Capacity and reallocation]*. + /// + /// If it is imporant to know the exact allocated capacity of a `Vec`, + /// always use the [`capacity`] method after construction. + /// + /// For `Vec<T>` where `T` is a zero-sized type, there will be no allocation + /// and the capacity will always be `usize::MAX`. + /// + /// [Capacity and reallocation]: #capacity-and-reallocation + /// [`capacity`]: Vec::capacity + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// let mut vec = Vec::with_capacity(10); + /// + /// // The vector contains no items, even though it has capacity for more + /// assert_eq!(vec.len(), 0); + /// assert!(vec.capacity() >= 10); + /// + /// // These are all done without reallocating... + /// for i in 0..10 { + /// vec.push(i); + /// } + /// assert_eq!(vec.len(), 10); + /// assert!(vec.capacity() >= 10); + /// + /// // ...but this may make the vector reallocate + /// vec.push(11); + /// assert_eq!(vec.len(), 11); + /// assert!(vec.capacity() >= 11); + /// + /// // A vector of a zero-sized type will always over-allocate, since no + /// // allocation is necessary + /// let vec_units = Vec::<()>::with_capacity(10); + /// assert_eq!(vec_units.capacity(), usize::MAX); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + #[must_use] + pub fn with_capacity(capacity: usize) -> Self { + Self::with_capacity_in(capacity, Global) + } + + /// Creates a `Vec<T>` directly from the raw components of another vector. + /// + /// # Safety + /// + /// This is highly unsafe, due to the number of invariants that aren't + /// checked: + /// + /// * `ptr` needs to have been previously allocated via [`String`]/`Vec<T>` + /// (at least, it's highly likely to be incorrect if it wasn't). + /// * `T` needs to have the same alignment as what `ptr` was allocated with. + /// (`T` having a less strict alignment is not sufficient, the alignment really + /// needs to be equal to satisfy the [`dealloc`] requirement that memory must be + /// allocated and deallocated with the same layout.) + /// * The size of `T` times the `capacity` (ie. the allocated size in bytes) needs + /// to be the same size as the pointer was allocated with. (Because similar to + /// alignment, [`dealloc`] must be called with the same layout `size`.) + /// * `length` needs to be less than or equal to `capacity`. + /// + /// Violating these may cause problems like corrupting the allocator's + /// internal data structures. For example it is normally **not** safe + /// to build a `Vec<u8>` from a pointer to a C `char` array with length + /// `size_t`, doing so is only safe if the array was initially allocated by + /// a `Vec` or `String`. + /// It's also not safe to build one from a `Vec<u16>` and its length, because + /// the allocator cares about the alignment, and these two types have different + /// alignments. The buffer was allocated with alignment 2 (for `u16`), but after + /// turning it into a `Vec<u8>` it'll be deallocated with alignment 1. To avoid + /// these issues, it is often preferable to do casting/transmuting using + /// [`slice::from_raw_parts`] instead. + /// + /// The ownership of `ptr` is effectively transferred to the + /// `Vec<T>` which may then deallocate, reallocate or change the + /// contents of memory pointed to by the pointer at will. Ensure + /// that nothing else uses the pointer after calling this + /// function. + /// + /// [`String`]: crate::string::String + /// [`dealloc`]: crate::alloc::GlobalAlloc::dealloc + /// + /// # Examples + /// + /// ``` + /// use std::ptr; + /// use std::mem; + /// + /// let v = vec![1, 2, 3]; + /// + // FIXME Update this when vec_into_raw_parts is stabilized + /// // Prevent running `v`'s destructor so we are in complete control + /// // of the allocation. + /// let mut v = mem::ManuallyDrop::new(v); + /// + /// // Pull out the various important pieces of information about `v` + /// let p = v.as_mut_ptr(); + /// let len = v.len(); + /// let cap = v.capacity(); + /// + /// unsafe { + /// // Overwrite memory with 4, 5, 6 + /// for i in 0..len as isize { + /// ptr::write(p.offset(i), 4 + i); + /// } + /// + /// // Put everything back together into a Vec + /// let rebuilt = Vec::from_raw_parts(p, len, cap); + /// assert_eq!(rebuilt, [4, 5, 6]); + /// } + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn from_raw_parts(ptr: *mut T, length: usize, capacity: usize) -> Self { + unsafe { Self::from_raw_parts_in(ptr, length, capacity, Global) } + } +} + +impl<T, A: Allocator> Vec<T, A> { + /// Constructs a new, empty `Vec<T, A>`. + /// + /// The vector will not allocate until elements are pushed onto it. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// # #[allow(unused_mut)] + /// let mut vec: Vec<i32, _> = Vec::new_in(System); + /// ``` + #[inline] + #[unstable(feature = "allocator_api", issue = "32838")] + pub const fn new_in(alloc: A) -> Self { + Vec { buf: RawVec::new_in(alloc), len: 0 } + } + + /// Constructs a new, empty `Vec<T, A>` with at least the specified capacity + /// with the provided allocator. + /// + /// The vector will be able to hold at least `capacity` elements without + /// reallocating. This method is allowed to allocate for more elements than + /// `capacity`. If `capacity` is 0, the vector will not allocate. + /// + /// It is important to note that although the returned vector has the + /// minimum *capacity* specified, the vector will have a zero *length*. For + /// an explanation of the difference between length and capacity, see + /// *[Capacity and reallocation]*. + /// + /// If it is imporant to know the exact allocated capacity of a `Vec`, + /// always use the [`capacity`] method after construction. + /// + /// For `Vec<T, A>` where `T` is a zero-sized type, there will be no allocation + /// and the capacity will always be `usize::MAX`. + /// + /// [Capacity and reallocation]: #capacity-and-reallocation + /// [`capacity`]: Vec::capacity + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// let mut vec = Vec::with_capacity_in(10, System); + /// + /// // The vector contains no items, even though it has capacity for more + /// assert_eq!(vec.len(), 0); + /// assert_eq!(vec.capacity(), 10); + /// + /// // These are all done without reallocating... + /// for i in 0..10 { + /// vec.push(i); + /// } + /// assert_eq!(vec.len(), 10); + /// assert_eq!(vec.capacity(), 10); + /// + /// // ...but this may make the vector reallocate + /// vec.push(11); + /// assert_eq!(vec.len(), 11); + /// assert!(vec.capacity() >= 11); + /// + /// // A vector of a zero-sized type will always over-allocate, since no + /// // allocation is necessary + /// let vec_units = Vec::<(), System>::with_capacity_in(10, System); + /// assert_eq!(vec_units.capacity(), usize::MAX); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[unstable(feature = "allocator_api", issue = "32838")] + pub fn with_capacity_in(capacity: usize, alloc: A) -> Self { + Vec { buf: RawVec::with_capacity_in(capacity, alloc), len: 0 } + } + + /// Creates a `Vec<T, A>` directly from the raw components of another vector. + /// + /// # Safety + /// + /// This is highly unsafe, due to the number of invariants that aren't + /// checked: + /// + /// * `ptr` needs to have been previously allocated via [`String`]/`Vec<T>` + /// (at least, it's highly likely to be incorrect if it wasn't). + /// * `T` needs to have the same size and alignment as what `ptr` was allocated with. + /// (`T` having a less strict alignment is not sufficient, the alignment really + /// needs to be equal to satisfy the [`dealloc`] requirement that memory must be + /// allocated and deallocated with the same layout.) + /// * `length` needs to be less than or equal to `capacity`. + /// * `capacity` needs to be the capacity that the pointer was allocated with. + /// + /// Violating these may cause problems like corrupting the allocator's + /// internal data structures. For example it is **not** safe + /// to build a `Vec<u8>` from a pointer to a C `char` array with length `size_t`. + /// It's also not safe to build one from a `Vec<u16>` and its length, because + /// the allocator cares about the alignment, and these two types have different + /// alignments. The buffer was allocated with alignment 2 (for `u16`), but after + /// turning it into a `Vec<u8>` it'll be deallocated with alignment 1. + /// + /// The ownership of `ptr` is effectively transferred to the + /// `Vec<T>` which may then deallocate, reallocate or change the + /// contents of memory pointed to by the pointer at will. Ensure + /// that nothing else uses the pointer after calling this + /// function. + /// + /// [`String`]: crate::string::String + /// [`dealloc`]: crate::alloc::GlobalAlloc::dealloc + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api)] + /// + /// use std::alloc::System; + /// + /// use std::ptr; + /// use std::mem; + /// + /// let mut v = Vec::with_capacity_in(3, System); + /// v.push(1); + /// v.push(2); + /// v.push(3); + /// + // FIXME Update this when vec_into_raw_parts is stabilized + /// // Prevent running `v`'s destructor so we are in complete control + /// // of the allocation. + /// let mut v = mem::ManuallyDrop::new(v); + /// + /// // Pull out the various important pieces of information about `v` + /// let p = v.as_mut_ptr(); + /// let len = v.len(); + /// let cap = v.capacity(); + /// let alloc = v.allocator(); + /// + /// unsafe { + /// // Overwrite memory with 4, 5, 6 + /// for i in 0..len as isize { + /// ptr::write(p.offset(i), 4 + i); + /// } + /// + /// // Put everything back together into a Vec + /// let rebuilt = Vec::from_raw_parts_in(p, len, cap, alloc.clone()); + /// assert_eq!(rebuilt, [4, 5, 6]); + /// } + /// ``` + #[inline] + #[unstable(feature = "allocator_api", issue = "32838")] + pub unsafe fn from_raw_parts_in(ptr: *mut T, length: usize, capacity: usize, alloc: A) -> Self { + unsafe { Vec { buf: RawVec::from_raw_parts_in(ptr, capacity, alloc), len: length } } + } + + /// Decomposes a `Vec<T>` into its raw components. + /// + /// Returns the raw pointer to the underlying data, the length of + /// the vector (in elements), and the allocated capacity of the + /// data (in elements). These are the same arguments in the same + /// order as the arguments to [`from_raw_parts`]. + /// + /// After calling this function, the caller is responsible for the + /// memory previously managed by the `Vec`. The only way to do + /// this is to convert the raw pointer, length, and capacity back + /// into a `Vec` with the [`from_raw_parts`] function, allowing + /// the destructor to perform the cleanup. + /// + /// [`from_raw_parts`]: Vec::from_raw_parts + /// + /// # Examples + /// + /// ``` + /// #![feature(vec_into_raw_parts)] + /// let v: Vec<i32> = vec![-1, 0, 1]; + /// + /// let (ptr, len, cap) = v.into_raw_parts(); + /// + /// let rebuilt = unsafe { + /// // We can now make changes to the components, such as + /// // transmuting the raw pointer to a compatible type. + /// let ptr = ptr as *mut u32; + /// + /// Vec::from_raw_parts(ptr, len, cap) + /// }; + /// assert_eq!(rebuilt, [4294967295, 0, 1]); + /// ``` + #[unstable(feature = "vec_into_raw_parts", reason = "new API", issue = "65816")] + pub fn into_raw_parts(self) -> (*mut T, usize, usize) { + let mut me = ManuallyDrop::new(self); + (me.as_mut_ptr(), me.len(), me.capacity()) + } + + /// Decomposes a `Vec<T>` into its raw components. + /// + /// Returns the raw pointer to the underlying data, the length of the vector (in elements), + /// the allocated capacity of the data (in elements), and the allocator. These are the same + /// arguments in the same order as the arguments to [`from_raw_parts_in`]. + /// + /// After calling this function, the caller is responsible for the + /// memory previously managed by the `Vec`. The only way to do + /// this is to convert the raw pointer, length, and capacity back + /// into a `Vec` with the [`from_raw_parts_in`] function, allowing + /// the destructor to perform the cleanup. + /// + /// [`from_raw_parts_in`]: Vec::from_raw_parts_in + /// + /// # Examples + /// + /// ``` + /// #![feature(allocator_api, vec_into_raw_parts)] + /// + /// use std::alloc::System; + /// + /// let mut v: Vec<i32, System> = Vec::new_in(System); + /// v.push(-1); + /// v.push(0); + /// v.push(1); + /// + /// let (ptr, len, cap, alloc) = v.into_raw_parts_with_alloc(); + /// + /// let rebuilt = unsafe { + /// // We can now make changes to the components, such as + /// // transmuting the raw pointer to a compatible type. + /// let ptr = ptr as *mut u32; + /// + /// Vec::from_raw_parts_in(ptr, len, cap, alloc) + /// }; + /// assert_eq!(rebuilt, [4294967295, 0, 1]); + /// ``` + #[unstable(feature = "allocator_api", issue = "32838")] + // #[unstable(feature = "vec_into_raw_parts", reason = "new API", issue = "65816")] + pub fn into_raw_parts_with_alloc(self) -> (*mut T, usize, usize, A) { + let mut me = ManuallyDrop::new(self); + let len = me.len(); + let capacity = me.capacity(); + let ptr = me.as_mut_ptr(); + let alloc = unsafe { ptr::read(me.allocator()) }; + (ptr, len, capacity, alloc) + } + + /// Returns the number of elements the vector can hold without + /// reallocating. + /// + /// # Examples + /// + /// ``` + /// let vec: Vec<i32> = Vec::with_capacity(10); + /// assert_eq!(vec.capacity(), 10); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn capacity(&self) -> usize { + self.buf.capacity() + } + + /// Reserves capacity for at least `additional` more elements to be inserted + /// in the given `Vec<T>`. The collection may reserve more space to + /// speculatively avoid frequent reallocations. After calling `reserve`, + /// capacity will be greater than or equal to `self.len() + additional`. + /// Does nothing if capacity is already sufficient. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1]; + /// vec.reserve(10); + /// assert!(vec.capacity() >= 11); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve(&mut self, additional: usize) { + self.buf.reserve(self.len, additional); + } + + /// Reserves the minimum capacity for at least `additional` more elements to + /// be inserted in the given `Vec<T>`. Unlike [`reserve`], this will not + /// deliberately over-allocate to speculatively avoid frequent allocations. + /// After calling `reserve_exact`, capacity will be greater than or equal to + /// `self.len() + additional`. Does nothing if the capacity is already + /// sufficient. + /// + /// Note that the allocator may give the collection more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`reserve`] if future insertions are expected. + /// + /// [`reserve`]: Vec::reserve + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1]; + /// vec.reserve_exact(10); + /// assert!(vec.capacity() >= 11); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn reserve_exact(&mut self, additional: usize) { + self.buf.reserve_exact(self.len, additional); + } + + /// Tries to reserve capacity for at least `additional` more elements to be inserted + /// in the given `Vec<T>`. The collection may reserve more space to speculatively avoid + /// frequent reallocations. After calling `try_reserve`, capacity will be + /// greater than or equal to `self.len() + additional` if it returns + /// `Ok(())`. Does nothing if capacity is already sufficient. + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// + /// fn process_data(data: &[u32]) -> Result<Vec<u32>, TryReserveError> { + /// let mut output = Vec::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// output.extend(data.iter().map(|&val| { + /// val * 2 + 5 // very complicated + /// })); + /// + /// Ok(output) + /// } + /// # process_data(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.buf.try_reserve(self.len, additional) + } + + /// Tries to reserve the minimum capacity for at least `additional` + /// elements to be inserted in the given `Vec<T>`. Unlike [`try_reserve`], + /// this will not deliberately over-allocate to speculatively avoid frequent + /// allocations. After calling `try_reserve_exact`, capacity will be greater + /// than or equal to `self.len() + additional` if it returns `Ok(())`. + /// Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the collection more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: Vec::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + /// + /// # Examples + /// + /// ``` + /// use std::collections::TryReserveError; + /// + /// fn process_data(data: &[u32]) -> Result<Vec<u32>, TryReserveError> { + /// let mut output = Vec::new(); + /// + /// // Pre-reserve the memory, exiting if we can't + /// output.try_reserve_exact(data.len())?; + /// + /// // Now we know this can't OOM in the middle of our complex work + /// output.extend(data.iter().map(|&val| { + /// val * 2 + 5 // very complicated + /// })); + /// + /// Ok(output) + /// } + /// # process_data(&[1, 2, 3]).expect("why is the test harness OOMing on 12 bytes?"); + /// ``` + #[stable(feature = "try_reserve", since = "1.57.0")] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.buf.try_reserve_exact(self.len, additional) + } + + /// Shrinks the capacity of the vector as much as possible. + /// + /// It will drop down as close as possible to the length but the allocator + /// may still inform the vector that there is space for a few more elements. + /// + /// # Examples + /// + /// ``` + /// let mut vec = Vec::with_capacity(10); + /// vec.extend([1, 2, 3]); + /// assert_eq!(vec.capacity(), 10); + /// vec.shrink_to_fit(); + /// assert!(vec.capacity() >= 3); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn shrink_to_fit(&mut self) { + // The capacity is never less than the length, and there's nothing to do when + // they are equal, so we can avoid the panic case in `RawVec::shrink_to_fit` + // by only calling it with a greater capacity. + if self.capacity() > self.len { + self.buf.shrink_to_fit(self.len); + } + } + + /// Shrinks the capacity of the vector with a lower bound. + /// + /// The capacity will remain at least as large as both the length + /// and the supplied value. + /// + /// If the current capacity is less than the lower limit, this is a no-op. + /// + /// # Examples + /// + /// ``` + /// let mut vec = Vec::with_capacity(10); + /// vec.extend([1, 2, 3]); + /// assert_eq!(vec.capacity(), 10); + /// vec.shrink_to(4); + /// assert!(vec.capacity() >= 4); + /// vec.shrink_to(0); + /// assert!(vec.capacity() >= 3); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "shrink_to", since = "1.56.0")] + pub fn shrink_to(&mut self, min_capacity: usize) { + if self.capacity() > min_capacity { + self.buf.shrink_to_fit(cmp::max(self.len, min_capacity)); + } + } + + /// Converts the vector into [`Box<[T]>`][owned slice]. + /// + /// Note that this will drop any excess capacity. + /// + /// [owned slice]: Box + /// + /// # Examples + /// + /// ``` + /// let v = vec![1, 2, 3]; + /// + /// let slice = v.into_boxed_slice(); + /// ``` + /// + /// Any excess capacity is removed: + /// + /// ``` + /// let mut vec = Vec::with_capacity(10); + /// vec.extend([1, 2, 3]); + /// + /// assert_eq!(vec.capacity(), 10); + /// let slice = vec.into_boxed_slice(); + /// assert_eq!(slice.into_vec().capacity(), 3); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn into_boxed_slice(mut self) -> Box<[T], A> { + unsafe { + self.shrink_to_fit(); + let me = ManuallyDrop::new(self); + let buf = ptr::read(&me.buf); + let len = me.len(); + buf.into_box(len).assume_init() + } + } + + /// Shortens the vector, keeping the first `len` elements and dropping + /// the rest. + /// + /// If `len` is greater than the vector's current length, this has no + /// effect. + /// + /// The [`drain`] method can emulate `truncate`, but causes the excess + /// elements to be returned instead of dropped. + /// + /// Note that this method has no effect on the allocated capacity + /// of the vector. + /// + /// # Examples + /// + /// Truncating a five element vector to two elements: + /// + /// ``` + /// let mut vec = vec![1, 2, 3, 4, 5]; + /// vec.truncate(2); + /// assert_eq!(vec, [1, 2]); + /// ``` + /// + /// No truncation occurs when `len` is greater than the vector's current + /// length: + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// vec.truncate(8); + /// assert_eq!(vec, [1, 2, 3]); + /// ``` + /// + /// Truncating when `len == 0` is equivalent to calling the [`clear`] + /// method. + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// vec.truncate(0); + /// assert_eq!(vec, []); + /// ``` + /// + /// [`clear`]: Vec::clear + /// [`drain`]: Vec::drain + #[stable(feature = "rust1", since = "1.0.0")] + pub fn truncate(&mut self, len: usize) { + // This is safe because: + // + // * the slice passed to `drop_in_place` is valid; the `len > self.len` + // case avoids creating an invalid slice, and + // * the `len` of the vector is shrunk before calling `drop_in_place`, + // such that no value will be dropped twice in case `drop_in_place` + // were to panic once (if it panics twice, the program aborts). + unsafe { + // Note: It's intentional that this is `>` and not `>=`. + // Changing it to `>=` has negative performance + // implications in some cases. See #78884 for more. + if len > self.len { + return; + } + let remaining_len = self.len - len; + let s = ptr::slice_from_raw_parts_mut(self.as_mut_ptr().add(len), remaining_len); + self.len = len; + ptr::drop_in_place(s); + } + } + + /// Extracts a slice containing the entire vector. + /// + /// Equivalent to `&s[..]`. + /// + /// # Examples + /// + /// ``` + /// use std::io::{self, Write}; + /// let buffer = vec![1, 2, 3, 5, 8]; + /// io::sink().write(buffer.as_slice()).unwrap(); + /// ``` + #[inline] + #[stable(feature = "vec_as_slice", since = "1.7.0")] + pub fn as_slice(&self) -> &[T] { + self + } + + /// Extracts a mutable slice of the entire vector. + /// + /// Equivalent to `&mut s[..]`. + /// + /// # Examples + /// + /// ``` + /// use std::io::{self, Read}; + /// let mut buffer = vec![0; 3]; + /// io::repeat(0b101).read_exact(buffer.as_mut_slice()).unwrap(); + /// ``` + #[inline] + #[stable(feature = "vec_as_slice", since = "1.7.0")] + pub fn as_mut_slice(&mut self) -> &mut [T] { + self + } + + /// Returns a raw pointer to the vector's buffer, or a dangling raw pointer + /// valid for zero sized reads if the vector didn't allocate. + /// + /// The caller must ensure that the vector outlives the pointer this + /// function returns, or else it will end up pointing to garbage. + /// Modifying the vector may cause its buffer to be reallocated, + /// which would also make any pointers to it invalid. + /// + /// The caller must also ensure that the memory the pointer (non-transitively) points to + /// is never written to (except inside an `UnsafeCell`) using this pointer or any pointer + /// derived from it. If you need to mutate the contents of the slice, use [`as_mut_ptr`]. + /// + /// # Examples + /// + /// ``` + /// let x = vec![1, 2, 4]; + /// let x_ptr = x.as_ptr(); + /// + /// unsafe { + /// for i in 0..x.len() { + /// assert_eq!(*x_ptr.add(i), 1 << i); + /// } + /// } + /// ``` + /// + /// [`as_mut_ptr`]: Vec::as_mut_ptr + #[stable(feature = "vec_as_ptr", since = "1.37.0")] + #[inline] + pub fn as_ptr(&self) -> *const T { + // We shadow the slice method of the same name to avoid going through + // `deref`, which creates an intermediate reference. + let ptr = self.buf.ptr(); + unsafe { + assume(!ptr.is_null()); + } + ptr + } + + /// Returns an unsafe mutable pointer to the vector's buffer, or a dangling + /// raw pointer valid for zero sized reads if the vector didn't allocate. + /// + /// The caller must ensure that the vector outlives the pointer this + /// function returns, or else it will end up pointing to garbage. + /// Modifying the vector may cause its buffer to be reallocated, + /// which would also make any pointers to it invalid. + /// + /// # Examples + /// + /// ``` + /// // Allocate vector big enough for 4 elements. + /// let size = 4; + /// let mut x: Vec<i32> = Vec::with_capacity(size); + /// let x_ptr = x.as_mut_ptr(); + /// + /// // Initialize elements via raw pointer writes, then set length. + /// unsafe { + /// for i in 0..size { + /// *x_ptr.add(i) = i as i32; + /// } + /// x.set_len(size); + /// } + /// assert_eq!(&*x, &[0, 1, 2, 3]); + /// ``` + #[stable(feature = "vec_as_ptr", since = "1.37.0")] + #[inline] + pub fn as_mut_ptr(&mut self) -> *mut T { + // We shadow the slice method of the same name to avoid going through + // `deref_mut`, which creates an intermediate reference. + let ptr = self.buf.ptr(); + unsafe { + assume(!ptr.is_null()); + } + ptr + } + + /// Returns a reference to the underlying allocator. + #[unstable(feature = "allocator_api", issue = "32838")] + #[inline] + pub fn allocator(&self) -> &A { + self.buf.allocator() + } + + /// Forces the length of the vector to `new_len`. + /// + /// This is a low-level operation that maintains none of the normal + /// invariants of the type. Normally changing the length of a vector + /// is done using one of the safe operations instead, such as + /// [`truncate`], [`resize`], [`extend`], or [`clear`]. + /// + /// [`truncate`]: Vec::truncate + /// [`resize`]: Vec::resize + /// [`extend`]: Extend::extend + /// [`clear`]: Vec::clear + /// + /// # Safety + /// + /// - `new_len` must be less than or equal to [`capacity()`]. + /// - The elements at `old_len..new_len` must be initialized. + /// + /// [`capacity()`]: Vec::capacity + /// + /// # Examples + /// + /// This method can be useful for situations in which the vector + /// is serving as a buffer for other code, particularly over FFI: + /// + /// ```no_run + /// # #![allow(dead_code)] + /// # // This is just a minimal skeleton for the doc example; + /// # // don't use this as a starting point for a real library. + /// # pub struct StreamWrapper { strm: *mut std::ffi::c_void } + /// # const Z_OK: i32 = 0; + /// # extern "C" { + /// # fn deflateGetDictionary( + /// # strm: *mut std::ffi::c_void, + /// # dictionary: *mut u8, + /// # dictLength: *mut usize, + /// # ) -> i32; + /// # } + /// # impl StreamWrapper { + /// pub fn get_dictionary(&self) -> Option<Vec<u8>> { + /// // Per the FFI method's docs, "32768 bytes is always enough". + /// let mut dict = Vec::with_capacity(32_768); + /// let mut dict_length = 0; + /// // SAFETY: When `deflateGetDictionary` returns `Z_OK`, it holds that: + /// // 1. `dict_length` elements were initialized. + /// // 2. `dict_length` <= the capacity (32_768) + /// // which makes `set_len` safe to call. + /// unsafe { + /// // Make the FFI call... + /// let r = deflateGetDictionary(self.strm, dict.as_mut_ptr(), &mut dict_length); + /// if r == Z_OK { + /// // ...and update the length to what was initialized. + /// dict.set_len(dict_length); + /// Some(dict) + /// } else { + /// None + /// } + /// } + /// } + /// # } + /// ``` + /// + /// While the following example is sound, there is a memory leak since + /// the inner vectors were not freed prior to the `set_len` call: + /// + /// ``` + /// let mut vec = vec![vec![1, 0, 0], + /// vec![0, 1, 0], + /// vec![0, 0, 1]]; + /// // SAFETY: + /// // 1. `old_len..0` is empty so no elements need to be initialized. + /// // 2. `0 <= capacity` always holds whatever `capacity` is. + /// unsafe { + /// vec.set_len(0); + /// } + /// ``` + /// + /// Normally, here, one would use [`clear`] instead to correctly drop + /// the contents and thus not leak memory. + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub unsafe fn set_len(&mut self, new_len: usize) { + debug_assert!(new_len <= self.capacity()); + + self.len = new_len; + } + + /// Removes an element from the vector and returns it. + /// + /// The removed element is replaced by the last element of the vector. + /// + /// This does not preserve ordering, but is *O*(1). + /// If you need to preserve the element order, use [`remove`] instead. + /// + /// [`remove`]: Vec::remove + /// + /// # Panics + /// + /// Panics if `index` is out of bounds. + /// + /// # Examples + /// + /// ``` + /// let mut v = vec!["foo", "bar", "baz", "qux"]; + /// + /// assert_eq!(v.swap_remove(1), "bar"); + /// assert_eq!(v, ["foo", "qux", "baz"]); + /// + /// assert_eq!(v.swap_remove(0), "foo"); + /// assert_eq!(v, ["baz", "qux"]); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn swap_remove(&mut self, index: usize) -> T { + #[cold] + #[inline(never)] + fn assert_failed(index: usize, len: usize) -> ! { + panic!("swap_remove index (is {index}) should be < len (is {len})"); + } + + let len = self.len(); + if index >= len { + assert_failed(index, len); + } + unsafe { + // We replace self[index] with the last element. Note that if the + // bounds check above succeeds there must be a last element (which + // can be self[index] itself). + let value = ptr::read(self.as_ptr().add(index)); + let base_ptr = self.as_mut_ptr(); + ptr::copy(base_ptr.add(len - 1), base_ptr.add(index), 1); + self.set_len(len - 1); + value + } + } + + /// Inserts an element at position `index` within the vector, shifting all + /// elements after it to the right. + /// + /// # Panics + /// + /// Panics if `index > len`. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// vec.insert(1, 4); + /// assert_eq!(vec, [1, 4, 2, 3]); + /// vec.insert(4, 5); + /// assert_eq!(vec, [1, 4, 2, 3, 5]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn insert(&mut self, index: usize, element: T) { + #[cold] + #[inline(never)] + fn assert_failed(index: usize, len: usize) -> ! { + panic!("insertion index (is {index}) should be <= len (is {len})"); + } + + let len = self.len(); + + // space for the new element + if len == self.buf.capacity() { + self.reserve(1); + } + + unsafe { + // infallible + // The spot to put the new value + { + let p = self.as_mut_ptr().add(index); + if index < len { + // Shift everything over to make space. (Duplicating the + // `index`th element into two consecutive places.) + ptr::copy(p, p.offset(1), len - index); + } else if index == len { + // No elements need shifting. + } else { + assert_failed(index, len); + } + // Write it in, overwriting the first copy of the `index`th + // element. + ptr::write(p, element); + } + self.set_len(len + 1); + } + } + + /// Removes and returns the element at position `index` within the vector, + /// shifting all elements after it to the left. + /// + /// Note: Because this shifts over the remaining elements, it has a + /// worst-case performance of *O*(*n*). If you don't need the order of elements + /// to be preserved, use [`swap_remove`] instead. If you'd like to remove + /// elements from the beginning of the `Vec`, consider using + /// [`VecDeque::pop_front`] instead. + /// + /// [`swap_remove`]: Vec::swap_remove + /// [`VecDeque::pop_front`]: crate::collections::VecDeque::pop_front + /// + /// # Panics + /// + /// Panics if `index` is out of bounds. + /// + /// # Examples + /// + /// ``` + /// let mut v = vec![1, 2, 3]; + /// assert_eq!(v.remove(1), 2); + /// assert_eq!(v, [1, 3]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[track_caller] + pub fn remove(&mut self, index: usize) -> T { + #[cold] + #[inline(never)] + #[track_caller] + fn assert_failed(index: usize, len: usize) -> ! { + panic!("removal index (is {index}) should be < len (is {len})"); + } + + let len = self.len(); + if index >= len { + assert_failed(index, len); + } + unsafe { + // infallible + let ret; + { + // the place we are taking from. + let ptr = self.as_mut_ptr().add(index); + // copy it out, unsafely having a copy of the value on + // the stack and in the vector at the same time. + ret = ptr::read(ptr); + + // Shift everything down to fill in that spot. + ptr::copy(ptr.offset(1), ptr, len - index - 1); + } + self.set_len(len - 1); + ret + } + } + + /// Retains only the elements specified by the predicate. + /// + /// In other words, remove all elements `e` for which `f(&e)` returns `false`. + /// This method operates in place, visiting each element exactly once in the + /// original order, and preserves the order of the retained elements. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3, 4]; + /// vec.retain(|&x| x % 2 == 0); + /// assert_eq!(vec, [2, 4]); + /// ``` + /// + /// Because the elements are visited exactly once in the original order, + /// external state may be used to decide which elements to keep. + /// + /// ``` + /// let mut vec = vec![1, 2, 3, 4, 5]; + /// let keep = [false, true, true, false, true]; + /// let mut iter = keep.iter(); + /// vec.retain(|_| *iter.next().unwrap()); + /// assert_eq!(vec, [2, 3, 5]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn retain<F>(&mut self, mut f: F) + where + F: FnMut(&T) -> bool, + { + self.retain_mut(|elem| f(elem)); + } + + /// Retains only the elements specified by the predicate, passing a mutable reference to it. + /// + /// In other words, remove all elements `e` such that `f(&mut e)` returns `false`. + /// This method operates in place, visiting each element exactly once in the + /// original order, and preserves the order of the retained elements. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3, 4]; + /// vec.retain_mut(|x| if *x <= 3 { + /// *x += 1; + /// true + /// } else { + /// false + /// }); + /// assert_eq!(vec, [2, 3, 4]); + /// ``` + #[stable(feature = "vec_retain_mut", since = "1.61.0")] + pub fn retain_mut<F>(&mut self, mut f: F) + where + F: FnMut(&mut T) -> bool, + { + let original_len = self.len(); + // Avoid double drop if the drop guard is not executed, + // since we may make some holes during the process. + unsafe { self.set_len(0) }; + + // Vec: [Kept, Kept, Hole, Hole, Hole, Hole, Unchecked, Unchecked] + // |<- processed len ->| ^- next to check + // |<- deleted cnt ->| + // |<- original_len ->| + // Kept: Elements which predicate returns true on. + // Hole: Moved or dropped element slot. + // Unchecked: Unchecked valid elements. + // + // This drop guard will be invoked when predicate or `drop` of element panicked. + // It shifts unchecked elements to cover holes and `set_len` to the correct length. + // In cases when predicate and `drop` never panick, it will be optimized out. + struct BackshiftOnDrop<'a, T, A: Allocator> { + v: &'a mut Vec<T, A>, + processed_len: usize, + deleted_cnt: usize, + original_len: usize, + } + + impl<T, A: Allocator> Drop for BackshiftOnDrop<'_, T, A> { + fn drop(&mut self) { + if self.deleted_cnt > 0 { + // SAFETY: Trailing unchecked items must be valid since we never touch them. + unsafe { + ptr::copy( + self.v.as_ptr().add(self.processed_len), + self.v.as_mut_ptr().add(self.processed_len - self.deleted_cnt), + self.original_len - self.processed_len, + ); + } + } + // SAFETY: After filling holes, all items are in contiguous memory. + unsafe { + self.v.set_len(self.original_len - self.deleted_cnt); + } + } + } + + let mut g = BackshiftOnDrop { v: self, processed_len: 0, deleted_cnt: 0, original_len }; + + fn process_loop<F, T, A: Allocator, const DELETED: bool>( + original_len: usize, + f: &mut F, + g: &mut BackshiftOnDrop<'_, T, A>, + ) where + F: FnMut(&mut T) -> bool, + { + while g.processed_len != original_len { + // SAFETY: Unchecked element must be valid. + let cur = unsafe { &mut *g.v.as_mut_ptr().add(g.processed_len) }; + if !f(cur) { + // Advance early to avoid double drop if `drop_in_place` panicked. + g.processed_len += 1; + g.deleted_cnt += 1; + // SAFETY: We never touch this element again after dropped. + unsafe { ptr::drop_in_place(cur) }; + // We already advanced the counter. + if DELETED { + continue; + } else { + break; + } + } + if DELETED { + // SAFETY: `deleted_cnt` > 0, so the hole slot must not overlap with current element. + // We use copy for move, and never touch this element again. + unsafe { + let hole_slot = g.v.as_mut_ptr().add(g.processed_len - g.deleted_cnt); + ptr::copy_nonoverlapping(cur, hole_slot, 1); + } + } + g.processed_len += 1; + } + } + + // Stage 1: Nothing was deleted. + process_loop::<F, T, A, false>(original_len, &mut f, &mut g); + + // Stage 2: Some elements were deleted. + process_loop::<F, T, A, true>(original_len, &mut f, &mut g); + + // All item are processed. This can be optimized to `set_len` by LLVM. + drop(g); + } + + /// Removes all but the first of consecutive elements in the vector that resolve to the same + /// key. + /// + /// If the vector is sorted, this removes all duplicates. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![10, 20, 21, 30, 20]; + /// + /// vec.dedup_by_key(|i| *i / 10); + /// + /// assert_eq!(vec, [10, 20, 30, 20]); + /// ``` + #[stable(feature = "dedup_by", since = "1.16.0")] + #[inline] + pub fn dedup_by_key<F, K>(&mut self, mut key: F) + where + F: FnMut(&mut T) -> K, + K: PartialEq, + { + self.dedup_by(|a, b| key(a) == key(b)) + } + + /// Removes all but the first of consecutive elements in the vector satisfying a given equality + /// relation. + /// + /// The `same_bucket` function is passed references to two elements from the vector and + /// must determine if the elements compare equal. The elements are passed in opposite order + /// from their order in the slice, so if `same_bucket(a, b)` returns `true`, `a` is removed. + /// + /// If the vector is sorted, this removes all duplicates. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec!["foo", "bar", "Bar", "baz", "bar"]; + /// + /// vec.dedup_by(|a, b| a.eq_ignore_ascii_case(b)); + /// + /// assert_eq!(vec, ["foo", "bar", "baz", "bar"]); + /// ``` + #[stable(feature = "dedup_by", since = "1.16.0")] + pub fn dedup_by<F>(&mut self, mut same_bucket: F) + where + F: FnMut(&mut T, &mut T) -> bool, + { + let len = self.len(); + if len <= 1 { + return; + } + + /* INVARIANT: vec.len() > read >= write > write-1 >= 0 */ + struct FillGapOnDrop<'a, T, A: core::alloc::Allocator> { + /* Offset of the element we want to check if it is duplicate */ + read: usize, + + /* Offset of the place where we want to place the non-duplicate + * when we find it. */ + write: usize, + + /* The Vec that would need correction if `same_bucket` panicked */ + vec: &'a mut Vec<T, A>, + } + + impl<'a, T, A: core::alloc::Allocator> Drop for FillGapOnDrop<'a, T, A> { + fn drop(&mut self) { + /* This code gets executed when `same_bucket` panics */ + + /* SAFETY: invariant guarantees that `read - write` + * and `len - read` never overflow and that the copy is always + * in-bounds. */ + unsafe { + let ptr = self.vec.as_mut_ptr(); + let len = self.vec.len(); + + /* How many items were left when `same_bucket` panicked. + * Basically vec[read..].len() */ + let items_left = len.wrapping_sub(self.read); + + /* Pointer to first item in vec[write..write+items_left] slice */ + let dropped_ptr = ptr.add(self.write); + /* Pointer to first item in vec[read..] slice */ + let valid_ptr = ptr.add(self.read); + + /* Copy `vec[read..]` to `vec[write..write+items_left]`. + * The slices can overlap, so `copy_nonoverlapping` cannot be used */ + ptr::copy(valid_ptr, dropped_ptr, items_left); + + /* How many items have been already dropped + * Basically vec[read..write].len() */ + let dropped = self.read.wrapping_sub(self.write); + + self.vec.set_len(len - dropped); + } + } + } + + let mut gap = FillGapOnDrop { read: 1, write: 1, vec: self }; + let ptr = gap.vec.as_mut_ptr(); + + /* Drop items while going through Vec, it should be more efficient than + * doing slice partition_dedup + truncate */ + + /* SAFETY: Because of the invariant, read_ptr, prev_ptr and write_ptr + * are always in-bounds and read_ptr never aliases prev_ptr */ + unsafe { + while gap.read < len { + let read_ptr = ptr.add(gap.read); + let prev_ptr = ptr.add(gap.write.wrapping_sub(1)); + + if same_bucket(&mut *read_ptr, &mut *prev_ptr) { + // Increase `gap.read` now since the drop may panic. + gap.read += 1; + /* We have found duplicate, drop it in-place */ + ptr::drop_in_place(read_ptr); + } else { + let write_ptr = ptr.add(gap.write); + + /* Because `read_ptr` can be equal to `write_ptr`, we either + * have to use `copy` or conditional `copy_nonoverlapping`. + * Looks like the first option is faster. */ + ptr::copy(read_ptr, write_ptr, 1); + + /* We have filled that place, so go further */ + gap.write += 1; + gap.read += 1; + } + } + + /* Technically we could let `gap` clean up with its Drop, but + * when `same_bucket` is guaranteed to not panic, this bloats a little + * the codegen, so we just do it manually */ + gap.vec.set_len(gap.write); + mem::forget(gap); + } + } + + /// Appends an element to the back of a collection. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2]; + /// vec.push(3); + /// assert_eq!(vec, [1, 2, 3]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn push(&mut self, value: T) { + // This will panic or abort if we would allocate > isize::MAX bytes + // or if the length increment would overflow for zero-sized types. + if self.len == self.buf.capacity() { + self.buf.reserve_for_push(self.len); + } + unsafe { + let end = self.as_mut_ptr().add(self.len); + ptr::write(end, value); + self.len += 1; + } + } + + /// Removes the last element from a vector and returns it, or [`None`] if it + /// is empty. + /// + /// If you'd like to pop the first element, consider using + /// [`VecDeque::pop_front`] instead. + /// + /// [`VecDeque::pop_front`]: crate::collections::VecDeque::pop_front + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// assert_eq!(vec.pop(), Some(3)); + /// assert_eq!(vec, [1, 2]); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn pop(&mut self) -> Option<T> { + if self.len == 0 { + None + } else { + unsafe { + self.len -= 1; + Some(ptr::read(self.as_ptr().add(self.len()))) + } + } + } + + /// Moves all the elements of `other` into `self`, leaving `other` empty. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// let mut vec2 = vec![4, 5, 6]; + /// vec.append(&mut vec2); + /// assert_eq!(vec, [1, 2, 3, 4, 5, 6]); + /// assert_eq!(vec2, []); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "append", since = "1.4.0")] + pub fn append(&mut self, other: &mut Self) { + unsafe { + self.append_elements(other.as_slice() as _); + other.set_len(0); + } + } + + /// Appends elements to `self` from other buffer. + #[cfg(not(no_global_oom_handling))] + #[inline] + unsafe fn append_elements(&mut self, other: *const [T]) { + let count = unsafe { (*other).len() }; + self.reserve(count); + let len = self.len(); + unsafe { ptr::copy_nonoverlapping(other as *const T, self.as_mut_ptr().add(len), count) }; + self.len += count; + } + + /// Removes the specified range from the vector in bulk, returning all + /// removed elements as an iterator. If the iterator is dropped before + /// being fully consumed, it drops the remaining removed elements. + /// + /// The returned iterator keeps a mutable borrow on the vector to optimize + /// its implementation. + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the vector. + /// + /// # Leaking + /// + /// If the returned iterator goes out of scope without being dropped (due to + /// [`mem::forget`], for example), the vector may have lost and leaked + /// elements arbitrarily, including elements outside the range. + /// + /// # Examples + /// + /// ``` + /// let mut v = vec![1, 2, 3]; + /// let u: Vec<_> = v.drain(1..).collect(); + /// assert_eq!(v, &[1]); + /// assert_eq!(u, &[2, 3]); + /// + /// // A full range clears the vector, like `clear()` does + /// v.drain(..); + /// assert_eq!(v, &[]); + /// ``` + #[stable(feature = "drain", since = "1.6.0")] + pub fn drain<R>(&mut self, range: R) -> Drain<'_, T, A> + where + R: RangeBounds<usize>, + { + // Memory safety + // + // When the Drain is first created, it shortens the length of + // the source vector to make sure no uninitialized or moved-from elements + // are accessible at all if the Drain's destructor never gets to run. + // + // Drain will ptr::read out the values to remove. + // When finished, remaining tail of the vec is copied back to cover + // the hole, and the vector length is restored to the new length. + // + let len = self.len(); + let Range { start, end } = slice::range(range, ..len); + + unsafe { + // set self.vec length's to start, to be safe in case Drain is leaked + self.set_len(start); + // Use the borrow in the IterMut to indicate borrowing behavior of the + // whole Drain iterator (like &mut T). + let range_slice = slice::from_raw_parts_mut(self.as_mut_ptr().add(start), end - start); + Drain { + tail_start: end, + tail_len: len - end, + iter: range_slice.iter(), + vec: NonNull::from(self), + } + } + } + + /// Clears the vector, removing all values. + /// + /// Note that this method has no effect on the allocated capacity + /// of the vector. + /// + /// # Examples + /// + /// ``` + /// let mut v = vec![1, 2, 3]; + /// + /// v.clear(); + /// + /// assert!(v.is_empty()); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn clear(&mut self) { + let elems: *mut [T] = self.as_mut_slice(); + + // SAFETY: + // - `elems` comes directly from `as_mut_slice` and is therefore valid. + // - Setting `self.len` before calling `drop_in_place` means that, + // if an element's `Drop` impl panics, the vector's `Drop` impl will + // do nothing (leaking the rest of the elements) instead of dropping + // some twice. + unsafe { + self.len = 0; + ptr::drop_in_place(elems); + } + } + + /// Returns the number of elements in the vector, also referred to + /// as its 'length'. + /// + /// # Examples + /// + /// ``` + /// let a = vec![1, 2, 3]; + /// assert_eq!(a.len(), 3); + /// ``` + #[inline] + #[stable(feature = "rust1", since = "1.0.0")] + pub fn len(&self) -> usize { + self.len + } + + /// Returns `true` if the vector contains no elements. + /// + /// # Examples + /// + /// ``` + /// let mut v = Vec::new(); + /// assert!(v.is_empty()); + /// + /// v.push(1); + /// assert!(!v.is_empty()); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + pub fn is_empty(&self) -> bool { + self.len() == 0 + } + + /// Splits the collection into two at the given index. + /// + /// Returns a newly allocated vector containing the elements in the range + /// `[at, len)`. After the call, the original vector will be left containing + /// the elements `[0, at)` with its previous capacity unchanged. + /// + /// # Panics + /// + /// Panics if `at > len`. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// let vec2 = vec.split_off(1); + /// assert_eq!(vec, [1]); + /// assert_eq!(vec2, [2, 3]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[must_use = "use `.truncate()` if you don't need the other half"] + #[stable(feature = "split_off", since = "1.4.0")] + pub fn split_off(&mut self, at: usize) -> Self + where + A: Clone, + { + #[cold] + #[inline(never)] + fn assert_failed(at: usize, len: usize) -> ! { + panic!("`at` split index (is {at}) should be <= len (is {len})"); + } + + if at > self.len() { + assert_failed(at, self.len()); + } + + if at == 0 { + // the new vector can take over the original buffer and avoid the copy + return mem::replace( + self, + Vec::with_capacity_in(self.capacity(), self.allocator().clone()), + ); + } + + let other_len = self.len - at; + let mut other = Vec::with_capacity_in(other_len, self.allocator().clone()); + + // Unsafely `set_len` and copy items to `other`. + unsafe { + self.set_len(at); + other.set_len(other_len); + + ptr::copy_nonoverlapping(self.as_ptr().add(at), other.as_mut_ptr(), other.len()); + } + other + } + + /// Resizes the `Vec` in-place so that `len` is equal to `new_len`. + /// + /// If `new_len` is greater than `len`, the `Vec` is extended by the + /// difference, with each additional slot filled with the result of + /// calling the closure `f`. The return values from `f` will end up + /// in the `Vec` in the order they have been generated. + /// + /// If `new_len` is less than `len`, the `Vec` is simply truncated. + /// + /// This method uses a closure to create new values on every push. If + /// you'd rather [`Clone`] a given value, use [`Vec::resize`]. If you + /// want to use the [`Default`] trait to generate values, you can + /// pass [`Default::default`] as the second argument. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 3]; + /// vec.resize_with(5, Default::default); + /// assert_eq!(vec, [1, 2, 3, 0, 0]); + /// + /// let mut vec = vec![]; + /// let mut p = 1; + /// vec.resize_with(4, || { p *= 2; p }); + /// assert_eq!(vec, [2, 4, 8, 16]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "vec_resize_with", since = "1.33.0")] + pub fn resize_with<F>(&mut self, new_len: usize, f: F) + where + F: FnMut() -> T, + { + let len = self.len(); + if new_len > len { + self.extend_with(new_len - len, ExtendFunc(f)); + } else { + self.truncate(new_len); + } + } + + /// Consumes and leaks the `Vec`, returning a mutable reference to the contents, + /// `&'a mut [T]`. Note that the type `T` must outlive the chosen lifetime + /// `'a`. If the type has only static references, or none at all, then this + /// may be chosen to be `'static`. + /// + /// As of Rust 1.57, this method does not reallocate or shrink the `Vec`, + /// so the leaked allocation may include unused capacity that is not part + /// of the returned slice. + /// + /// This function is mainly useful for data that lives for the remainder of + /// the program's life. Dropping the returned reference will cause a memory + /// leak. + /// + /// # Examples + /// + /// Simple usage: + /// + /// ``` + /// let x = vec![1, 2, 3]; + /// let static_ref: &'static mut [usize] = x.leak(); + /// static_ref[0] += 1; + /// assert_eq!(static_ref, &[2, 2, 3]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "vec_leak", since = "1.47.0")] + #[inline] + pub fn leak<'a>(self) -> &'a mut [T] + where + A: 'a, + { + let mut me = ManuallyDrop::new(self); + unsafe { slice::from_raw_parts_mut(me.as_mut_ptr(), me.len) } + } + + /// Returns the remaining spare capacity of the vector as a slice of + /// `MaybeUninit<T>`. + /// + /// The returned slice can be used to fill the vector with data (e.g. by + /// reading from a file) before marking the data as initialized using the + /// [`set_len`] method. + /// + /// [`set_len`]: Vec::set_len + /// + /// # Examples + /// + /// ``` + /// // Allocate vector big enough for 10 elements. + /// let mut v = Vec::with_capacity(10); + /// + /// // Fill in the first 3 elements. + /// let uninit = v.spare_capacity_mut(); + /// uninit[0].write(0); + /// uninit[1].write(1); + /// uninit[2].write(2); + /// + /// // Mark the first 3 elements of the vector as being initialized. + /// unsafe { + /// v.set_len(3); + /// } + /// + /// assert_eq!(&v, &[0, 1, 2]); + /// ``` + #[stable(feature = "vec_spare_capacity", since = "1.60.0")] + #[inline] + pub fn spare_capacity_mut(&mut self) -> &mut [MaybeUninit<T>] { + // Note: + // This method is not implemented in terms of `split_at_spare_mut`, + // to prevent invalidation of pointers to the buffer. + unsafe { + slice::from_raw_parts_mut( + self.as_mut_ptr().add(self.len) as *mut MaybeUninit<T>, + self.buf.capacity() - self.len, + ) + } + } + + /// Returns vector content as a slice of `T`, along with the remaining spare + /// capacity of the vector as a slice of `MaybeUninit<T>`. + /// + /// The returned spare capacity slice can be used to fill the vector with data + /// (e.g. by reading from a file) before marking the data as initialized using + /// the [`set_len`] method. + /// + /// [`set_len`]: Vec::set_len + /// + /// Note that this is a low-level API, which should be used with care for + /// optimization purposes. If you need to append data to a `Vec` + /// you can use [`push`], [`extend`], [`extend_from_slice`], + /// [`extend_from_within`], [`insert`], [`append`], [`resize`] or + /// [`resize_with`], depending on your exact needs. + /// + /// [`push`]: Vec::push + /// [`extend`]: Vec::extend + /// [`extend_from_slice`]: Vec::extend_from_slice + /// [`extend_from_within`]: Vec::extend_from_within + /// [`insert`]: Vec::insert + /// [`append`]: Vec::append + /// [`resize`]: Vec::resize + /// [`resize_with`]: Vec::resize_with + /// + /// # Examples + /// + /// ``` + /// #![feature(vec_split_at_spare)] + /// + /// let mut v = vec![1, 1, 2]; + /// + /// // Reserve additional space big enough for 10 elements. + /// v.reserve(10); + /// + /// let (init, uninit) = v.split_at_spare_mut(); + /// let sum = init.iter().copied().sum::<u32>(); + /// + /// // Fill in the next 4 elements. + /// uninit[0].write(sum); + /// uninit[1].write(sum * 2); + /// uninit[2].write(sum * 3); + /// uninit[3].write(sum * 4); + /// + /// // Mark the 4 elements of the vector as being initialized. + /// unsafe { + /// let len = v.len(); + /// v.set_len(len + 4); + /// } + /// + /// assert_eq!(&v, &[1, 1, 2, 4, 8, 12, 16]); + /// ``` + #[unstable(feature = "vec_split_at_spare", issue = "81944")] + #[inline] + pub fn split_at_spare_mut(&mut self) -> (&mut [T], &mut [MaybeUninit<T>]) { + // SAFETY: + // - len is ignored and so never changed + let (init, spare, _) = unsafe { self.split_at_spare_mut_with_len() }; + (init, spare) + } + + /// Safety: changing returned .2 (&mut usize) is considered the same as calling `.set_len(_)`. + /// + /// This method provides unique access to all vec parts at once in `extend_from_within`. + unsafe fn split_at_spare_mut_with_len( + &mut self, + ) -> (&mut [T], &mut [MaybeUninit<T>], &mut usize) { + let ptr = self.as_mut_ptr(); + // SAFETY: + // - `ptr` is guaranteed to be valid for `self.len` elements + // - but the allocation extends out to `self.buf.capacity()` elements, possibly + // uninitialized + let spare_ptr = unsafe { ptr.add(self.len) }; + let spare_ptr = spare_ptr.cast::<MaybeUninit<T>>(); + let spare_len = self.buf.capacity() - self.len; + + // SAFETY: + // - `ptr` is guaranteed to be valid for `self.len` elements + // - `spare_ptr` is pointing one element past the buffer, so it doesn't overlap with `initialized` + unsafe { + let initialized = slice::from_raw_parts_mut(ptr, self.len); + let spare = slice::from_raw_parts_mut(spare_ptr, spare_len); + + (initialized, spare, &mut self.len) + } + } +} + +impl<T: Clone, A: Allocator> Vec<T, A> { + /// Resizes the `Vec` in-place so that `len` is equal to `new_len`. + /// + /// If `new_len` is greater than `len`, the `Vec` is extended by the + /// difference, with each additional slot filled with `value`. + /// If `new_len` is less than `len`, the `Vec` is simply truncated. + /// + /// This method requires `T` to implement [`Clone`], + /// in order to be able to clone the passed value. + /// If you need more flexibility (or want to rely on [`Default`] instead of + /// [`Clone`]), use [`Vec::resize_with`]. + /// If you only need to resize to a smaller size, use [`Vec::truncate`]. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec!["hello"]; + /// vec.resize(3, "world"); + /// assert_eq!(vec, ["hello", "world", "world"]); + /// + /// let mut vec = vec![1, 2, 3, 4]; + /// vec.resize(2, 0); + /// assert_eq!(vec, [1, 2]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "vec_resize", since = "1.5.0")] + pub fn resize(&mut self, new_len: usize, value: T) { + let len = self.len(); + + if new_len > len { + self.extend_with(new_len - len, ExtendElement(value)) + } else { + self.truncate(new_len); + } + } + + /// Clones and appends all elements in a slice to the `Vec`. + /// + /// Iterates over the slice `other`, clones each element, and then appends + /// it to this `Vec`. The `other` slice is traversed in-order. + /// + /// Note that this function is same as [`extend`] except that it is + /// specialized to work with slices instead. If and when Rust gets + /// specialization this function will likely be deprecated (but still + /// available). + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1]; + /// vec.extend_from_slice(&[2, 3, 4]); + /// assert_eq!(vec, [1, 2, 3, 4]); + /// ``` + /// + /// [`extend`]: Vec::extend + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "vec_extend_from_slice", since = "1.6.0")] + pub fn extend_from_slice(&mut self, other: &[T]) { + self.spec_extend(other.iter()) + } + + /// Copies elements from `src` range to the end of the vector. + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the vector. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![0, 1, 2, 3, 4]; + /// + /// vec.extend_from_within(2..); + /// assert_eq!(vec, [0, 1, 2, 3, 4, 2, 3, 4]); + /// + /// vec.extend_from_within(..2); + /// assert_eq!(vec, [0, 1, 2, 3, 4, 2, 3, 4, 0, 1]); + /// + /// vec.extend_from_within(4..8); + /// assert_eq!(vec, [0, 1, 2, 3, 4, 2, 3, 4, 0, 1, 4, 2, 3, 4]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[stable(feature = "vec_extend_from_within", since = "1.53.0")] + pub fn extend_from_within<R>(&mut self, src: R) + where + R: RangeBounds<usize>, + { + let range = slice::range(src, ..self.len()); + self.reserve(range.len()); + + // SAFETY: + // - `slice::range` guarantees that the given range is valid for indexing self + unsafe { + self.spec_extend_from_within(range); + } + } +} + +impl<T, A: Allocator, const N: usize> Vec<[T; N], A> { + /// Takes a `Vec<[T; N]>` and flattens it into a `Vec<T>`. + /// + /// # Panics + /// + /// Panics if the length of the resulting vector would overflow a `usize`. + /// + /// This is only possible when flattening a vector of arrays of zero-sized + /// types, and thus tends to be irrelevant in practice. If + /// `size_of::<T>() > 0`, this will never panic. + /// + /// # Examples + /// + /// ``` + /// #![feature(slice_flatten)] + /// + /// let mut vec = vec![[1, 2, 3], [4, 5, 6], [7, 8, 9]]; + /// assert_eq!(vec.pop(), Some([7, 8, 9])); + /// + /// let mut flattened = vec.into_flattened(); + /// assert_eq!(flattened.pop(), Some(6)); + /// ``` + #[unstable(feature = "slice_flatten", issue = "95629")] + pub fn into_flattened(self) -> Vec<T, A> { + let (ptr, len, cap, alloc) = self.into_raw_parts_with_alloc(); + let (new_len, new_cap) = if mem::size_of::<T>() == 0 { + (len.checked_mul(N).expect("vec len overflow"), usize::MAX) + } else { + // SAFETY: + // - `cap * N` cannot overflow because the allocation is already in + // the address space. + // - Each `[T; N]` has `N` valid elements, so there are `len * N` + // valid elements in the allocation. + unsafe { (len.unchecked_mul(N), cap.unchecked_mul(N)) } + }; + // SAFETY: + // - `ptr` was allocated by `self` + // - `ptr` is well-aligned because `[T; N]` has the same alignment as `T`. + // - `new_cap` refers to the same sized allocation as `cap` because + // `new_cap * size_of::<T>()` == `cap * size_of::<[T; N]>()` + // - `len` <= `cap`, so `len * N` <= `cap * N`. + unsafe { Vec::<T, A>::from_raw_parts_in(ptr.cast(), new_len, new_cap, alloc) } + } +} + +// This code generalizes `extend_with_{element,default}`. +trait ExtendWith<T> { + fn next(&mut self) -> T; + fn last(self) -> T; +} + +struct ExtendElement<T>(T); +impl<T: Clone> ExtendWith<T> for ExtendElement<T> { + fn next(&mut self) -> T { + self.0.clone() + } + fn last(self) -> T { + self.0 + } +} + +struct ExtendFunc<F>(F); +impl<T, F: FnMut() -> T> ExtendWith<T> for ExtendFunc<F> { + fn next(&mut self) -> T { + (self.0)() + } + fn last(mut self) -> T { + (self.0)() + } +} + +impl<T, A: Allocator> Vec<T, A> { + #[cfg(not(no_global_oom_handling))] + /// Extend the vector by `n` values, using the given generator. + fn extend_with<E: ExtendWith<T>>(&mut self, n: usize, mut value: E) { + self.reserve(n); + + unsafe { + let mut ptr = self.as_mut_ptr().add(self.len()); + // Use SetLenOnDrop to work around bug where compiler + // might not realize the store through `ptr` through self.set_len() + // don't alias. + let mut local_len = SetLenOnDrop::new(&mut self.len); + + // Write all elements except the last one + for _ in 1..n { + ptr::write(ptr, value.next()); + ptr = ptr.offset(1); + // Increment the length in every step in case next() panics + local_len.increment_len(1); + } + + if n > 0 { + // We can write the last element directly without cloning needlessly + ptr::write(ptr, value.last()); + local_len.increment_len(1); + } + + // len set by scope guard + } + } +} + +impl<T: PartialEq, A: Allocator> Vec<T, A> { + /// Removes consecutive repeated elements in the vector according to the + /// [`PartialEq`] trait implementation. + /// + /// If the vector is sorted, this removes all duplicates. + /// + /// # Examples + /// + /// ``` + /// let mut vec = vec![1, 2, 2, 3, 2]; + /// + /// vec.dedup(); + /// + /// assert_eq!(vec, [1, 2, 3, 2]); + /// ``` + #[stable(feature = "rust1", since = "1.0.0")] + #[inline] + pub fn dedup(&mut self) { + self.dedup_by(|a, b| a == b) + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Internal methods and functions +//////////////////////////////////////////////////////////////////////////////// + +#[doc(hidden)] +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +pub fn from_elem<T: Clone>(elem: T, n: usize) -> Vec<T> { + <T as SpecFromElem>::from_elem(elem, n, Global) +} + +#[doc(hidden)] +#[cfg(not(no_global_oom_handling))] +#[unstable(feature = "allocator_api", issue = "32838")] +pub fn from_elem_in<T: Clone, A: Allocator>(elem: T, n: usize, alloc: A) -> Vec<T, A> { + <T as SpecFromElem>::from_elem(elem, n, alloc) +} + +trait ExtendFromWithinSpec { + /// # Safety + /// + /// - `src` needs to be valid index + /// - `self.capacity() - self.len()` must be `>= src.len()` + unsafe fn spec_extend_from_within(&mut self, src: Range<usize>); +} + +impl<T: Clone, A: Allocator> ExtendFromWithinSpec for Vec<T, A> { + default unsafe fn spec_extend_from_within(&mut self, src: Range<usize>) { + // SAFETY: + // - len is increased only after initializing elements + let (this, spare, len) = unsafe { self.split_at_spare_mut_with_len() }; + + // SAFETY: + // - caller guaratees that src is a valid index + let to_clone = unsafe { this.get_unchecked(src) }; + + iter::zip(to_clone, spare) + .map(|(src, dst)| dst.write(src.clone())) + // Note: + // - Element was just initialized with `MaybeUninit::write`, so it's ok to increase len + // - len is increased after each element to prevent leaks (see issue #82533) + .for_each(|_| *len += 1); + } +} + +impl<T: Copy, A: Allocator> ExtendFromWithinSpec for Vec<T, A> { + unsafe fn spec_extend_from_within(&mut self, src: Range<usize>) { + let count = src.len(); + { + let (init, spare) = self.split_at_spare_mut(); + + // SAFETY: + // - caller guaratees that `src` is a valid index + let source = unsafe { init.get_unchecked(src) }; + + // SAFETY: + // - Both pointers are created from unique slice references (`&mut [_]`) + // so they are valid and do not overlap. + // - Elements are :Copy so it's OK to copy them, without doing + // anything with the original values + // - `count` is equal to the len of `source`, so source is valid for + // `count` reads + // - `.reserve(count)` guarantees that `spare.len() >= count` so spare + // is valid for `count` writes + unsafe { ptr::copy_nonoverlapping(source.as_ptr(), spare.as_mut_ptr() as _, count) }; + } + + // SAFETY: + // - The elements were just initialized by `copy_nonoverlapping` + self.len += count; + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Common trait implementations for Vec +//////////////////////////////////////////////////////////////////////////////// + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> ops::Deref for Vec<T, A> { + type Target = [T]; + + #[inline] + fn deref(&self) -> &[T] { + unsafe { slice::from_raw_parts(self.as_ptr(), self.len) } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> ops::DerefMut for Vec<T, A> { + #[inline] + fn deref_mut(&mut self) -> &mut [T] { + unsafe { slice::from_raw_parts_mut(self.as_mut_ptr(), self.len) } + } +} + +#[cfg(not(no_global_oom_handling))] +trait SpecCloneFrom { + fn clone_from(this: &mut Self, other: &Self); +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Clone, A: Allocator> SpecCloneFrom for Vec<T, A> { + default fn clone_from(this: &mut Self, other: &Self) { + // drop anything that will not be overwritten + this.truncate(other.len()); + + // self.len <= other.len due to the truncate above, so the + // slices here are always in-bounds. + let (init, tail) = other.split_at(this.len()); + + // reuse the contained values' allocations/resources. + this.clone_from_slice(init); + this.extend_from_slice(tail); + } +} + +#[cfg(not(no_global_oom_handling))] +impl<T: Copy, A: Allocator> SpecCloneFrom for Vec<T, A> { + fn clone_from(this: &mut Self, other: &Self) { + this.clear(); + this.extend_from_slice(other); + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone, A: Allocator + Clone> Clone for Vec<T, A> { + #[cfg(not(test))] + fn clone(&self) -> Self { + let alloc = self.allocator().clone(); + <[T]>::to_vec_in(&**self, alloc) + } + + // HACK(japaric): with cfg(test) the inherent `[T]::to_vec` method, which is + // required for this method definition, is not available. Instead use the + // `slice::to_vec` function which is only available with cfg(test) + // NB see the slice::hack module in slice.rs for more information + #[cfg(test)] + fn clone(&self) -> Self { + let alloc = self.allocator().clone(); + crate::slice::to_vec(&**self, alloc) + } + + fn clone_from(&mut self, other: &Self) { + SpecCloneFrom::clone_from(self, other) + } +} + +/// The hash of a vector is the same as that of the corresponding slice, +/// as required by the `core::borrow::Borrow` implementation. +/// +/// ``` +/// #![feature(build_hasher_simple_hash_one)] +/// use std::hash::BuildHasher; +/// +/// let b = std::collections::hash_map::RandomState::new(); +/// let v: Vec<u8> = vec![0xa8, 0x3c, 0x09]; +/// let s: &[u8] = &[0xa8, 0x3c, 0x09]; +/// assert_eq!(b.hash_one(v), b.hash_one(s)); +/// ``` +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Hash, A: Allocator> Hash for Vec<T, A> { + #[inline] + fn hash<H: Hasher>(&self, state: &mut H) { + Hash::hash(&**self, state) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_on_unimplemented( + message = "vector indices are of type `usize` or ranges of `usize`", + label = "vector indices are of type `usize` or ranges of `usize`" +)] +impl<T, I: SliceIndex<[T]>, A: Allocator> Index<I> for Vec<T, A> { + type Output = I::Output; + + #[inline] + fn index(&self, index: I) -> &Self::Output { + Index::index(&**self, index) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_on_unimplemented( + message = "vector indices are of type `usize` or ranges of `usize`", + label = "vector indices are of type `usize` or ranges of `usize`" +)] +impl<T, I: SliceIndex<[T]>, A: Allocator> IndexMut<I> for Vec<T, A> { + #[inline] + fn index_mut(&mut self, index: I) -> &mut Self::Output { + IndexMut::index_mut(&mut **self, index) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T> FromIterator<T> for Vec<T> { + #[inline] + fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> Vec<T> { + <Self as SpecFromIter<T, I::IntoIter>>::from_iter(iter.into_iter()) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> IntoIterator for Vec<T, A> { + type Item = T; + type IntoIter = IntoIter<T, A>; + + /// Creates a consuming iterator, that is, one that moves each value out of + /// the vector (from start to end). The vector cannot be used after calling + /// this. + /// + /// # Examples + /// + /// ``` + /// let v = vec!["a".to_string(), "b".to_string()]; + /// let mut v_iter = v.into_iter(); + /// + /// let first_element: Option<String> = v_iter.next(); + /// + /// assert_eq!(first_element, Some("a".to_string())); + /// assert_eq!(v_iter.next(), Some("b".to_string())); + /// assert_eq!(v_iter.next(), None); + /// ``` + #[inline] + fn into_iter(self) -> IntoIter<T, A> { + unsafe { + let mut me = ManuallyDrop::new(self); + let alloc = ManuallyDrop::new(ptr::read(me.allocator())); + let begin = me.as_mut_ptr(); + let end = if mem::size_of::<T>() == 0 { + arith_offset(begin as *const i8, me.len() as isize) as *const T + } else { + begin.add(me.len()) as *const T + }; + let cap = me.buf.capacity(); + IntoIter { + buf: NonNull::new_unchecked(begin), + phantom: PhantomData, + cap, + alloc, + ptr: begin, + end, + } + } + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T, A: Allocator> IntoIterator for &'a Vec<T, A> { + type Item = &'a T; + type IntoIter = slice::Iter<'a, T>; + + fn into_iter(self) -> slice::Iter<'a, T> { + self.iter() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<'a, T, A: Allocator> IntoIterator for &'a mut Vec<T, A> { + type Item = &'a mut T; + type IntoIter = slice::IterMut<'a, T>; + + fn into_iter(self) -> slice::IterMut<'a, T> { + self.iter_mut() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> Extend<T> for Vec<T, A> { + #[inline] + fn extend<I: IntoIterator<Item = T>>(&mut self, iter: I) { + <Self as SpecExtend<T, I::IntoIter>>::spec_extend(self, iter.into_iter()) + } + + #[inline] + fn extend_one(&mut self, item: T) { + self.push(item); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +impl<T, A: Allocator> Vec<T, A> { + // leaf method to which various SpecFrom/SpecExtend implementations delegate when + // they have no further optimizations to apply + #[cfg(not(no_global_oom_handling))] + fn extend_desugared<I: Iterator<Item = T>>(&mut self, mut iterator: I) { + // This is the case for a general iterator. + // + // This function should be the moral equivalent of: + // + // for item in iterator { + // self.push(item); + // } + while let Some(element) = iterator.next() { + let len = self.len(); + if len == self.capacity() { + let (lower, _) = iterator.size_hint(); + self.reserve(lower.saturating_add(1)); + } + unsafe { + ptr::write(self.as_mut_ptr().add(len), element); + // Since next() executes user code which can panic we have to bump the length + // after each step. + // NB can't overflow since we would have had to alloc the address space + self.set_len(len + 1); + } + } + } + + /// Creates a splicing iterator that replaces the specified range in the vector + /// with the given `replace_with` iterator and yields the removed items. + /// `replace_with` does not need to be the same length as `range`. + /// + /// `range` is removed even if the iterator is not consumed until the end. + /// + /// It is unspecified how many elements are removed from the vector + /// if the `Splice` value is leaked. + /// + /// The input iterator `replace_with` is only consumed when the `Splice` value is dropped. + /// + /// This is optimal if: + /// + /// * The tail (elements in the vector after `range`) is empty, + /// * or `replace_with` yields fewer or equal elements than `range`’s length + /// * or the lower bound of its `size_hint()` is exact. + /// + /// Otherwise, a temporary vector is allocated and the tail is moved twice. + /// + /// # Panics + /// + /// Panics if the starting point is greater than the end point or if + /// the end point is greater than the length of the vector. + /// + /// # Examples + /// + /// ``` + /// let mut v = vec![1, 2, 3, 4]; + /// let new = [7, 8, 9]; + /// let u: Vec<_> = v.splice(1..3, new).collect(); + /// assert_eq!(v, &[1, 7, 8, 9, 4]); + /// assert_eq!(u, &[2, 3]); + /// ``` + #[cfg(not(no_global_oom_handling))] + #[inline] + #[stable(feature = "vec_splice", since = "1.21.0")] + pub fn splice<R, I>(&mut self, range: R, replace_with: I) -> Splice<'_, I::IntoIter, A> + where + R: RangeBounds<usize>, + I: IntoIterator<Item = T>, + { + Splice { drain: self.drain(range), replace_with: replace_with.into_iter() } + } + + /// Creates an iterator which uses a closure to determine if an element should be removed. + /// + /// If the closure returns true, then the element is removed and yielded. + /// If the closure returns false, the element will remain in the vector and will not be yielded + /// by the iterator. + /// + /// Using this method is equivalent to the following code: + /// + /// ``` + /// # let some_predicate = |x: &mut i32| { *x == 2 || *x == 3 || *x == 6 }; + /// # let mut vec = vec![1, 2, 3, 4, 5, 6]; + /// let mut i = 0; + /// while i < vec.len() { + /// if some_predicate(&mut vec[i]) { + /// let val = vec.remove(i); + /// // your code here + /// } else { + /// i += 1; + /// } + /// } + /// + /// # assert_eq!(vec, vec![1, 4, 5]); + /// ``` + /// + /// But `drain_filter` is easier to use. `drain_filter` is also more efficient, + /// because it can backshift the elements of the array in bulk. + /// + /// Note that `drain_filter` also lets you mutate every element in the filter closure, + /// regardless of whether you choose to keep or remove it. + /// + /// # Examples + /// + /// Splitting an array into evens and odds, reusing the original allocation: + /// + /// ``` + /// #![feature(drain_filter)] + /// let mut numbers = vec![1, 2, 3, 4, 5, 6, 8, 9, 11, 13, 14, 15]; + /// + /// let evens = numbers.drain_filter(|x| *x % 2 == 0).collect::<Vec<_>>(); + /// let odds = numbers; + /// + /// assert_eq!(evens, vec![2, 4, 6, 8, 14]); + /// assert_eq!(odds, vec![1, 3, 5, 9, 11, 13, 15]); + /// ``` + #[unstable(feature = "drain_filter", reason = "recently added", issue = "43244")] + pub fn drain_filter<F>(&mut self, filter: F) -> DrainFilter<'_, T, F, A> + where + F: FnMut(&mut T) -> bool, + { + let old_len = self.len(); + + // Guard against us getting leaked (leak amplification) + unsafe { + self.set_len(0); + } + + DrainFilter { vec: self, idx: 0, del: 0, old_len, pred: filter, panic_flag: false } + } +} + +/// Extend implementation that copies elements out of references before pushing them onto the Vec. +/// +/// This implementation is specialized for slice iterators, where it uses [`copy_from_slice`] to +/// append the entire slice at once. +/// +/// [`copy_from_slice`]: slice::copy_from_slice +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "extend_ref", since = "1.2.0")] +impl<'a, T: Copy + 'a, A: Allocator + 'a> Extend<&'a T> for Vec<T, A> { + fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iter: I) { + self.spec_extend(iter.into_iter()) + } + + #[inline] + fn extend_one(&mut self, &item: &'a T) { + self.push(item); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + self.reserve(additional); + } +} + +/// Implements comparison of vectors, [lexicographically](core::cmp::Ord#lexicographical-comparison). +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: PartialOrd, A: Allocator> PartialOrd for Vec<T, A> { + #[inline] + fn partial_cmp(&self, other: &Self) -> Option<Ordering> { + PartialOrd::partial_cmp(&**self, &**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Eq, A: Allocator> Eq for Vec<T, A> {} + +/// Implements ordering of vectors, [lexicographically](core::cmp::Ord#lexicographical-comparison). +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Ord, A: Allocator> Ord for Vec<T, A> { + #[inline] + fn cmp(&self, other: &Self) -> Ordering { + Ord::cmp(&**self, &**other) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +unsafe impl<#[may_dangle] T, A: Allocator> Drop for Vec<T, A> { + fn drop(&mut self) { + unsafe { + // use drop for [T] + // use a raw slice to refer to the elements of the vector as weakest necessary type; + // could avoid questions of validity in certain cases + ptr::drop_in_place(ptr::slice_from_raw_parts_mut(self.as_mut_ptr(), self.len)) + } + // RawVec handles deallocation + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +#[rustc_const_unstable(feature = "const_default_impls", issue = "87864")] +impl<T> const Default for Vec<T> { + /// Creates an empty `Vec<T>`. + fn default() -> Vec<T> { + Vec::new() + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: fmt::Debug, A: Allocator> fmt::Debug for Vec<T, A> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, f) + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> AsRef<Vec<T, A>> for Vec<T, A> { + fn as_ref(&self) -> &Vec<T, A> { + self + } +} + +#[stable(feature = "vec_as_mut", since = "1.5.0")] +impl<T, A: Allocator> AsMut<Vec<T, A>> for Vec<T, A> { + fn as_mut(&mut self) -> &mut Vec<T, A> { + self + } +} + +#[stable(feature = "rust1", since = "1.0.0")] +impl<T, A: Allocator> AsRef<[T]> for Vec<T, A> { + fn as_ref(&self) -> &[T] { + self + } +} + +#[stable(feature = "vec_as_mut", since = "1.5.0")] +impl<T, A: Allocator> AsMut<[T]> for Vec<T, A> { + fn as_mut(&mut self) -> &mut [T] { + self + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl<T: Clone> From<&[T]> for Vec<T> { + /// Allocate a `Vec<T>` and fill it by cloning `s`'s items. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(Vec::from(&[1, 2, 3][..]), vec![1, 2, 3]); + /// ``` + #[cfg(not(test))] + fn from(s: &[T]) -> Vec<T> { + s.to_vec() + } + #[cfg(test)] + fn from(s: &[T]) -> Vec<T> { + crate::slice::to_vec(s, Global) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "vec_from_mut", since = "1.19.0")] +impl<T: Clone> From<&mut [T]> for Vec<T> { + /// Allocate a `Vec<T>` and fill it by cloning `s`'s items. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(Vec::from(&mut [1, 2, 3][..]), vec![1, 2, 3]); + /// ``` + #[cfg(not(test))] + fn from(s: &mut [T]) -> Vec<T> { + s.to_vec() + } + #[cfg(test)] + fn from(s: &mut [T]) -> Vec<T> { + crate::slice::to_vec(s, Global) + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "vec_from_array", since = "1.44.0")] +impl<T, const N: usize> From<[T; N]> for Vec<T> { + /// Allocate a `Vec<T>` and move `s`'s items into it. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(Vec::from([1, 2, 3]), vec![1, 2, 3]); + /// ``` + #[cfg(not(test))] + fn from(s: [T; N]) -> Vec<T> { + <[T]>::into_vec( + #[rustc_box] + Box::new(s), + ) + } + + #[cfg(test)] + fn from(s: [T; N]) -> Vec<T> { + crate::slice::into_vec(Box::new(s)) + } +} + +#[stable(feature = "vec_from_cow_slice", since = "1.14.0")] +impl<'a, T> From<Cow<'a, [T]>> for Vec<T> +where + [T]: ToOwned<Owned = Vec<T>>, +{ + /// Convert a clone-on-write slice into a vector. + /// + /// If `s` already owns a `Vec<T>`, it will be returned directly. + /// If `s` is borrowing a slice, a new `Vec<T>` will be allocated and + /// filled by cloning `s`'s items into it. + /// + /// # Examples + /// + /// ``` + /// # use std::borrow::Cow; + /// let o: Cow<[i32]> = Cow::Owned(vec![1, 2, 3]); + /// let b: Cow<[i32]> = Cow::Borrowed(&[1, 2, 3]); + /// assert_eq!(Vec::from(o), Vec::from(b)); + /// ``` + fn from(s: Cow<'a, [T]>) -> Vec<T> { + s.into_owned() + } +} + +// note: test pulls in libstd, which causes errors here +#[cfg(not(test))] +#[stable(feature = "vec_from_box", since = "1.18.0")] +impl<T, A: Allocator> From<Box<[T], A>> for Vec<T, A> { + /// Convert a boxed slice into a vector by transferring ownership of + /// the existing heap allocation. + /// + /// # Examples + /// + /// ``` + /// let b: Box<[i32]> = vec![1, 2, 3].into_boxed_slice(); + /// assert_eq!(Vec::from(b), vec![1, 2, 3]); + /// ``` + fn from(s: Box<[T], A>) -> Self { + s.into_vec() + } +} + +// note: test pulls in libstd, which causes errors here +#[cfg(not(no_global_oom_handling))] +#[cfg(not(test))] +#[stable(feature = "box_from_vec", since = "1.20.0")] +impl<T, A: Allocator> From<Vec<T, A>> for Box<[T], A> { + /// Convert a vector into a boxed slice. + /// + /// If `v` has excess capacity, its items will be moved into a + /// newly-allocated buffer with exactly the right capacity. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(Box::from(vec![1, 2, 3]), vec![1, 2, 3].into_boxed_slice()); + /// ``` + fn from(v: Vec<T, A>) -> Self { + v.into_boxed_slice() + } +} + +#[cfg(not(no_global_oom_handling))] +#[stable(feature = "rust1", since = "1.0.0")] +impl From<&str> for Vec<u8> { + /// Allocate a `Vec<u8>` and fill it with a UTF-8 string. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(Vec::from("123"), vec![b'1', b'2', b'3']); + /// ``` + fn from(s: &str) -> Vec<u8> { + From::from(s.as_bytes()) + } +} + +#[stable(feature = "array_try_from_vec", since = "1.48.0")] +impl<T, A: Allocator, const N: usize> TryFrom<Vec<T, A>> for [T; N] { + type Error = Vec<T, A>; + + /// Gets the entire contents of the `Vec<T>` as an array, + /// if its size exactly matches that of the requested array. + /// + /// # Examples + /// + /// ``` + /// assert_eq!(vec![1, 2, 3].try_into(), Ok([1, 2, 3])); + /// assert_eq!(<Vec<i32>>::new().try_into(), Ok([])); + /// ``` + /// + /// If the length doesn't match, the input comes back in `Err`: + /// ``` + /// let r: Result<[i32; 4], _> = (0..10).collect::<Vec<_>>().try_into(); + /// assert_eq!(r, Err(vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9])); + /// ``` + /// + /// If you're fine with just getting a prefix of the `Vec<T>`, + /// you can call [`.truncate(N)`](Vec::truncate) first. + /// ``` + /// let mut v = String::from("hello world").into_bytes(); + /// v.sort(); + /// v.truncate(2); + /// let [a, b]: [_; 2] = v.try_into().unwrap(); + /// assert_eq!(a, b' '); + /// assert_eq!(b, b'd'); + /// ``` + fn try_from(mut vec: Vec<T, A>) -> Result<[T; N], Vec<T, A>> { + if vec.len() != N { + return Err(vec); + } + + // SAFETY: `.set_len(0)` is always sound. + unsafe { vec.set_len(0) }; + + // SAFETY: A `Vec`'s pointer is always aligned properly, and + // the alignment the array needs is the same as the items. + // We checked earlier that we have sufficient items. + // The items will not double-drop as the `set_len` + // tells the `Vec` not to also drop them. + let array = unsafe { ptr::read(vec.as_ptr() as *const [T; N]) }; + Ok(array) + } +} diff --git a/library/alloc/src/vec/partial_eq.rs b/library/alloc/src/vec/partial_eq.rs new file mode 100644 index 000000000..b0cf72577 --- /dev/null +++ b/library/alloc/src/vec/partial_eq.rs @@ -0,0 +1,47 @@ +use crate::alloc::Allocator; +#[cfg(not(no_global_oom_handling))] +use crate::borrow::Cow; + +use super::Vec; + +macro_rules! __impl_slice_eq1 { + ([$($vars:tt)*] $lhs:ty, $rhs:ty $(where $ty:ty: $bound:ident)?, #[$stability:meta]) => { + #[$stability] + impl<T, U, $($vars)*> PartialEq<$rhs> for $lhs + where + T: PartialEq<U>, + $($ty: $bound)? + { + #[inline] + fn eq(&self, other: &$rhs) -> bool { self[..] == other[..] } + #[inline] + fn ne(&self, other: &$rhs) -> bool { self[..] != other[..] } + } + } +} + +__impl_slice_eq1! { [A1: Allocator, A2: Allocator] Vec<T, A1>, Vec<U, A2>, #[stable(feature = "rust1", since = "1.0.0")] } +__impl_slice_eq1! { [A: Allocator] Vec<T, A>, &[U], #[stable(feature = "rust1", since = "1.0.0")] } +__impl_slice_eq1! { [A: Allocator] Vec<T, A>, &mut [U], #[stable(feature = "rust1", since = "1.0.0")] } +__impl_slice_eq1! { [A: Allocator] &[T], Vec<U, A>, #[stable(feature = "partialeq_vec_for_ref_slice", since = "1.46.0")] } +__impl_slice_eq1! { [A: Allocator] &mut [T], Vec<U, A>, #[stable(feature = "partialeq_vec_for_ref_slice", since = "1.46.0")] } +__impl_slice_eq1! { [A: Allocator] Vec<T, A>, [U], #[stable(feature = "partialeq_vec_for_slice", since = "1.48.0")] } +__impl_slice_eq1! { [A: Allocator] [T], Vec<U, A>, #[stable(feature = "partialeq_vec_for_slice", since = "1.48.0")] } +#[cfg(not(no_global_oom_handling))] +__impl_slice_eq1! { [A: Allocator] Cow<'_, [T]>, Vec<U, A> where T: Clone, #[stable(feature = "rust1", since = "1.0.0")] } +#[cfg(not(no_global_oom_handling))] +__impl_slice_eq1! { [] Cow<'_, [T]>, &[U] where T: Clone, #[stable(feature = "rust1", since = "1.0.0")] } +#[cfg(not(no_global_oom_handling))] +__impl_slice_eq1! { [] Cow<'_, [T]>, &mut [U] where T: Clone, #[stable(feature = "rust1", since = "1.0.0")] } +__impl_slice_eq1! { [A: Allocator, const N: usize] Vec<T, A>, [U; N], #[stable(feature = "rust1", since = "1.0.0")] } +__impl_slice_eq1! { [A: Allocator, const N: usize] Vec<T, A>, &[U; N], #[stable(feature = "rust1", since = "1.0.0")] } + +// NOTE: some less important impls are omitted to reduce code bloat +// FIXME(Centril): Reconsider this? +//__impl_slice_eq1! { [const N: usize] Vec<A>, &mut [B; N], } +//__impl_slice_eq1! { [const N: usize] [A; N], Vec<B>, } +//__impl_slice_eq1! { [const N: usize] &[A; N], Vec<B>, } +//__impl_slice_eq1! { [const N: usize] &mut [A; N], Vec<B>, } +//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, [B; N], } +//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, &[B; N], } +//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, &mut [B; N], } diff --git a/library/alloc/src/vec/set_len_on_drop.rs b/library/alloc/src/vec/set_len_on_drop.rs new file mode 100644 index 000000000..8b66bc812 --- /dev/null +++ b/library/alloc/src/vec/set_len_on_drop.rs @@ -0,0 +1,28 @@ +// Set the length of the vec when the `SetLenOnDrop` value goes out of scope. +// +// The idea is: The length field in SetLenOnDrop is a local variable +// that the optimizer will see does not alias with any stores through the Vec's data +// pointer. This is a workaround for alias analysis issue #32155 +pub(super) struct SetLenOnDrop<'a> { + len: &'a mut usize, + local_len: usize, +} + +impl<'a> SetLenOnDrop<'a> { + #[inline] + pub(super) fn new(len: &'a mut usize) -> Self { + SetLenOnDrop { local_len: *len, len } + } + + #[inline] + pub(super) fn increment_len(&mut self, increment: usize) { + self.local_len += increment; + } +} + +impl Drop for SetLenOnDrop<'_> { + #[inline] + fn drop(&mut self) { + *self.len = self.local_len; + } +} diff --git a/library/alloc/src/vec/spec_extend.rs b/library/alloc/src/vec/spec_extend.rs new file mode 100644 index 000000000..506ee0ecf --- /dev/null +++ b/library/alloc/src/vec/spec_extend.rs @@ -0,0 +1,87 @@ +use crate::alloc::Allocator; +use core::iter::TrustedLen; +use core::ptr::{self}; +use core::slice::{self}; + +use super::{IntoIter, SetLenOnDrop, Vec}; + +// Specialization trait used for Vec::extend +pub(super) trait SpecExtend<T, I> { + fn spec_extend(&mut self, iter: I); +} + +impl<T, I, A: Allocator> SpecExtend<T, I> for Vec<T, A> +where + I: Iterator<Item = T>, +{ + default fn spec_extend(&mut self, iter: I) { + self.extend_desugared(iter) + } +} + +impl<T, I, A: Allocator> SpecExtend<T, I> for Vec<T, A> +where + I: TrustedLen<Item = T>, +{ + default fn spec_extend(&mut self, iterator: I) { + // This is the case for a TrustedLen iterator. + let (low, high) = iterator.size_hint(); + if let Some(additional) = high { + debug_assert_eq!( + low, + additional, + "TrustedLen iterator's size hint is not exact: {:?}", + (low, high) + ); + self.reserve(additional); + unsafe { + let mut ptr = self.as_mut_ptr().add(self.len()); + let mut local_len = SetLenOnDrop::new(&mut self.len); + iterator.for_each(move |element| { + ptr::write(ptr, element); + ptr = ptr.offset(1); + // Since the loop executes user code which can panic we have to bump the pointer + // after each step. + // NB can't overflow since we would have had to alloc the address space + local_len.increment_len(1); + }); + } + } else { + // Per TrustedLen contract a `None` upper bound means that the iterator length + // truly exceeds usize::MAX, which would eventually lead to a capacity overflow anyway. + // Since the other branch already panics eagerly (via `reserve()`) we do the same here. + // This avoids additional codegen for a fallback code path which would eventually + // panic anyway. + panic!("capacity overflow"); + } + } +} + +impl<T, A: Allocator> SpecExtend<T, IntoIter<T>> for Vec<T, A> { + fn spec_extend(&mut self, mut iterator: IntoIter<T>) { + unsafe { + self.append_elements(iterator.as_slice() as _); + } + iterator.forget_remaining_elements(); + } +} + +impl<'a, T: 'a, I, A: Allocator + 'a> SpecExtend<&'a T, I> for Vec<T, A> +where + I: Iterator<Item = &'a T>, + T: Clone, +{ + default fn spec_extend(&mut self, iterator: I) { + self.spec_extend(iterator.cloned()) + } +} + +impl<'a, T: 'a, A: Allocator + 'a> SpecExtend<&'a T, slice::Iter<'a, T>> for Vec<T, A> +where + T: Copy, +{ + fn spec_extend(&mut self, iterator: slice::Iter<'a, T>) { + let slice = iterator.as_slice(); + unsafe { self.append_elements(slice) }; + } +} diff --git a/library/alloc/src/vec/spec_from_elem.rs b/library/alloc/src/vec/spec_from_elem.rs new file mode 100644 index 000000000..ff364c033 --- /dev/null +++ b/library/alloc/src/vec/spec_from_elem.rs @@ -0,0 +1,61 @@ +use core::ptr; + +use crate::alloc::Allocator; +use crate::raw_vec::RawVec; + +use super::{ExtendElement, IsZero, Vec}; + +// Specialization trait used for Vec::from_elem +pub(super) trait SpecFromElem: Sized { + fn from_elem<A: Allocator>(elem: Self, n: usize, alloc: A) -> Vec<Self, A>; +} + +impl<T: Clone> SpecFromElem for T { + default fn from_elem<A: Allocator>(elem: Self, n: usize, alloc: A) -> Vec<Self, A> { + let mut v = Vec::with_capacity_in(n, alloc); + v.extend_with(n, ExtendElement(elem)); + v + } +} + +impl<T: Clone + IsZero> SpecFromElem for T { + #[inline] + default fn from_elem<A: Allocator>(elem: T, n: usize, alloc: A) -> Vec<T, A> { + if elem.is_zero() { + return Vec { buf: RawVec::with_capacity_zeroed_in(n, alloc), len: n }; + } + let mut v = Vec::with_capacity_in(n, alloc); + v.extend_with(n, ExtendElement(elem)); + v + } +} + +impl SpecFromElem for i8 { + #[inline] + fn from_elem<A: Allocator>(elem: i8, n: usize, alloc: A) -> Vec<i8, A> { + if elem == 0 { + return Vec { buf: RawVec::with_capacity_zeroed_in(n, alloc), len: n }; + } + unsafe { + let mut v = Vec::with_capacity_in(n, alloc); + ptr::write_bytes(v.as_mut_ptr(), elem as u8, n); + v.set_len(n); + v + } + } +} + +impl SpecFromElem for u8 { + #[inline] + fn from_elem<A: Allocator>(elem: u8, n: usize, alloc: A) -> Vec<u8, A> { + if elem == 0 { + return Vec { buf: RawVec::with_capacity_zeroed_in(n, alloc), len: n }; + } + unsafe { + let mut v = Vec::with_capacity_in(n, alloc); + ptr::write_bytes(v.as_mut_ptr(), elem, n); + v.set_len(n); + v + } + } +} diff --git a/library/alloc/src/vec/spec_from_iter.rs b/library/alloc/src/vec/spec_from_iter.rs new file mode 100644 index 000000000..efa686847 --- /dev/null +++ b/library/alloc/src/vec/spec_from_iter.rs @@ -0,0 +1,64 @@ +use core::mem::ManuallyDrop; +use core::ptr::{self}; + +use super::{IntoIter, SpecExtend, SpecFromIterNested, Vec}; + +/// Specialization trait used for Vec::from_iter +/// +/// ## The delegation graph: +/// +/// ```text +/// +-------------+ +/// |FromIterator | +/// +-+-----------+ +/// | +/// v +/// +-+-------------------------------+ +---------------------+ +/// |SpecFromIter +---->+SpecFromIterNested | +/// |where I: | | |where I: | +/// | Iterator (default)----------+ | | Iterator (default) | +/// | vec::IntoIter | | | TrustedLen | +/// | SourceIterMarker---fallback-+ | +---------------------+ +/// +---------------------------------+ +/// ``` +pub(super) trait SpecFromIter<T, I> { + fn from_iter(iter: I) -> Self; +} + +impl<T, I> SpecFromIter<T, I> for Vec<T> +where + I: Iterator<Item = T>, +{ + default fn from_iter(iterator: I) -> Self { + SpecFromIterNested::from_iter(iterator) + } +} + +impl<T> SpecFromIter<T, IntoIter<T>> for Vec<T> { + fn from_iter(iterator: IntoIter<T>) -> Self { + // A common case is passing a vector into a function which immediately + // re-collects into a vector. We can short circuit this if the IntoIter + // has not been advanced at all. + // When it has been advanced We can also reuse the memory and move the data to the front. + // But we only do so when the resulting Vec wouldn't have more unused capacity + // than creating it through the generic FromIterator implementation would. That limitation + // is not strictly necessary as Vec's allocation behavior is intentionally unspecified. + // But it is a conservative choice. + let has_advanced = iterator.buf.as_ptr() as *const _ != iterator.ptr; + if !has_advanced || iterator.len() >= iterator.cap / 2 { + unsafe { + let it = ManuallyDrop::new(iterator); + if has_advanced { + ptr::copy(it.ptr, it.buf.as_ptr(), it.len()); + } + return Vec::from_raw_parts(it.buf.as_ptr(), it.len(), it.cap); + } + } + + let mut vec = Vec::new(); + // must delegate to spec_extend() since extend() itself delegates + // to spec_from for empty Vecs + vec.spec_extend(iterator); + vec + } +} diff --git a/library/alloc/src/vec/spec_from_iter_nested.rs b/library/alloc/src/vec/spec_from_iter_nested.rs new file mode 100644 index 000000000..f915ebb86 --- /dev/null +++ b/library/alloc/src/vec/spec_from_iter_nested.rs @@ -0,0 +1,65 @@ +use core::cmp; +use core::iter::TrustedLen; +use core::ptr; + +use crate::raw_vec::RawVec; + +use super::{SpecExtend, Vec}; + +/// Another specialization trait for Vec::from_iter +/// necessary to manually prioritize overlapping specializations +/// see [`SpecFromIter`](super::SpecFromIter) for details. +pub(super) trait SpecFromIterNested<T, I> { + fn from_iter(iter: I) -> Self; +} + +impl<T, I> SpecFromIterNested<T, I> for Vec<T> +where + I: Iterator<Item = T>, +{ + default fn from_iter(mut iterator: I) -> Self { + // Unroll the first iteration, as the vector is going to be + // expanded on this iteration in every case when the iterable is not + // empty, but the loop in extend_desugared() is not going to see the + // vector being full in the few subsequent loop iterations. + // So we get better branch prediction. + let mut vector = match iterator.next() { + None => return Vec::new(), + Some(element) => { + let (lower, _) = iterator.size_hint(); + let initial_capacity = + cmp::max(RawVec::<T>::MIN_NON_ZERO_CAP, lower.saturating_add(1)); + let mut vector = Vec::with_capacity(initial_capacity); + unsafe { + // SAFETY: We requested capacity at least 1 + ptr::write(vector.as_mut_ptr(), element); + vector.set_len(1); + } + vector + } + }; + // must delegate to spec_extend() since extend() itself delegates + // to spec_from for empty Vecs + <Vec<T> as SpecExtend<T, I>>::spec_extend(&mut vector, iterator); + vector + } +} + +impl<T, I> SpecFromIterNested<T, I> for Vec<T> +where + I: TrustedLen<Item = T>, +{ + fn from_iter(iterator: I) -> Self { + let mut vector = match iterator.size_hint() { + (_, Some(upper)) => Vec::with_capacity(upper), + // TrustedLen contract guarantees that `size_hint() == (_, None)` means that there + // are more than `usize::MAX` elements. + // Since the previous branch would eagerly panic if the capacity is too large + // (via `with_capacity`) we do the same here. + _ => panic!("capacity overflow"), + }; + // reuse extend specialization for TrustedLen + vector.spec_extend(iterator); + vector + } +} diff --git a/library/alloc/src/vec/splice.rs b/library/alloc/src/vec/splice.rs new file mode 100644 index 000000000..bad765c7f --- /dev/null +++ b/library/alloc/src/vec/splice.rs @@ -0,0 +1,133 @@ +use crate::alloc::{Allocator, Global}; +use core::ptr::{self}; +use core::slice::{self}; + +use super::{Drain, Vec}; + +/// A splicing iterator for `Vec`. +/// +/// This struct is created by [`Vec::splice()`]. +/// See its documentation for more. +/// +/// # Example +/// +/// ``` +/// let mut v = vec![0, 1, 2]; +/// let new = [7, 8]; +/// let iter: std::vec::Splice<_> = v.splice(1.., new); +/// ``` +#[derive(Debug)] +#[stable(feature = "vec_splice", since = "1.21.0")] +pub struct Splice< + 'a, + I: Iterator + 'a, + #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator + 'a = Global, +> { + pub(super) drain: Drain<'a, I::Item, A>, + pub(super) replace_with: I, +} + +#[stable(feature = "vec_splice", since = "1.21.0")] +impl<I: Iterator, A: Allocator> Iterator for Splice<'_, I, A> { + type Item = I::Item; + + fn next(&mut self) -> Option<Self::Item> { + self.drain.next() + } + + fn size_hint(&self) -> (usize, Option<usize>) { + self.drain.size_hint() + } +} + +#[stable(feature = "vec_splice", since = "1.21.0")] +impl<I: Iterator, A: Allocator> DoubleEndedIterator for Splice<'_, I, A> { + fn next_back(&mut self) -> Option<Self::Item> { + self.drain.next_back() + } +} + +#[stable(feature = "vec_splice", since = "1.21.0")] +impl<I: Iterator, A: Allocator> ExactSizeIterator for Splice<'_, I, A> {} + +#[stable(feature = "vec_splice", since = "1.21.0")] +impl<I: Iterator, A: Allocator> Drop for Splice<'_, I, A> { + fn drop(&mut self) { + self.drain.by_ref().for_each(drop); + + unsafe { + if self.drain.tail_len == 0 { + self.drain.vec.as_mut().extend(self.replace_with.by_ref()); + return; + } + + // First fill the range left by drain(). + if !self.drain.fill(&mut self.replace_with) { + return; + } + + // There may be more elements. Use the lower bound as an estimate. + // FIXME: Is the upper bound a better guess? Or something else? + let (lower_bound, _upper_bound) = self.replace_with.size_hint(); + if lower_bound > 0 { + self.drain.move_tail(lower_bound); + if !self.drain.fill(&mut self.replace_with) { + return; + } + } + + // Collect any remaining elements. + // This is a zero-length vector which does not allocate if `lower_bound` was exact. + let mut collected = self.replace_with.by_ref().collect::<Vec<I::Item>>().into_iter(); + // Now we have an exact count. + if collected.len() > 0 { + self.drain.move_tail(collected.len()); + let filled = self.drain.fill(&mut collected); + debug_assert!(filled); + debug_assert_eq!(collected.len(), 0); + } + } + // Let `Drain::drop` move the tail back if necessary and restore `vec.len`. + } +} + +/// Private helper methods for `Splice::drop` +impl<T, A: Allocator> Drain<'_, T, A> { + /// The range from `self.vec.len` to `self.tail_start` contains elements + /// that have been moved out. + /// Fill that range as much as possible with new elements from the `replace_with` iterator. + /// Returns `true` if we filled the entire range. (`replace_with.next()` didn’t return `None`.) + unsafe fn fill<I: Iterator<Item = T>>(&mut self, replace_with: &mut I) -> bool { + let vec = unsafe { self.vec.as_mut() }; + let range_start = vec.len; + let range_end = self.tail_start; + let range_slice = unsafe { + slice::from_raw_parts_mut(vec.as_mut_ptr().add(range_start), range_end - range_start) + }; + + for place in range_slice { + if let Some(new_item) = replace_with.next() { + unsafe { ptr::write(place, new_item) }; + vec.len += 1; + } else { + return false; + } + } + true + } + + /// Makes room for inserting more elements before the tail. + unsafe fn move_tail(&mut self, additional: usize) { + let vec = unsafe { self.vec.as_mut() }; + let len = self.tail_start + self.tail_len; + vec.buf.reserve(len, additional); + + let new_tail_start = self.tail_start + additional; + unsafe { + let src = vec.as_ptr().add(self.tail_start); + let dst = vec.as_mut_ptr().add(new_tail_start); + ptr::copy(src, dst, self.tail_len); + } + self.tail_start = new_tail_start; + } +} |