diff options
Diffstat (limited to 'vendor/text-size/src')
-rw-r--r-- | vendor/text-size/src/lib.rs | 32 | ||||
-rw-r--r-- | vendor/text-size/src/range.rs | 456 | ||||
-rw-r--r-- | vendor/text-size/src/serde_impls.rs | 48 | ||||
-rw-r--r-- | vendor/text-size/src/size.rs | 161 | ||||
-rw-r--r-- | vendor/text-size/src/traits.rs | 36 |
5 files changed, 733 insertions, 0 deletions
diff --git a/vendor/text-size/src/lib.rs b/vendor/text-size/src/lib.rs new file mode 100644 index 000000000..92bd36b19 --- /dev/null +++ b/vendor/text-size/src/lib.rs @@ -0,0 +1,32 @@ +//! Newtypes for working with text sizes/ranges in a more type-safe manner. +//! +//! This library can help with two things: +//! * Reducing storage requirements for offsets and ranges, under the +//! assumption that 32 bits is enough. +//! * Providing standard vocabulary types for applications where text ranges +//! are pervasive. +//! +//! However, you should not use this library simply because you work with +//! strings. In the overwhelming majority of cases, using `usize` and +//! `std::ops::Range<usize>` is better. In particular, if you are publishing a +//! library, using only std types in the interface would make it more +//! interoperable. Similarly, if you are writing something like a lexer, which +//! produces, but does not *store* text ranges, then sticking to `usize` would +//! be better. +//! +//! Minimal Supported Rust Version: latest stable. + +#![forbid(unsafe_code)] +#![warn(missing_debug_implementations, missing_docs)] + +mod range; +mod size; +mod traits; + +#[cfg(feature = "serde")] +mod serde_impls; + +pub use crate::{range::TextRange, size::TextSize, traits::TextLen}; + +#[cfg(target_pointer_width = "16")] +compile_error!("text-size assumes usize >= u32 and does not work on 16-bit targets"); diff --git a/vendor/text-size/src/range.rs b/vendor/text-size/src/range.rs new file mode 100644 index 000000000..4a98deec5 --- /dev/null +++ b/vendor/text-size/src/range.rs @@ -0,0 +1,456 @@ +use cmp::Ordering; + +use { + crate::TextSize, + std::{ + cmp, fmt, + ops::{Add, AddAssign, Bound, Index, IndexMut, Range, RangeBounds, Sub, SubAssign}, + }, +}; + +/// A range in text, represented as a pair of [`TextSize`][struct@TextSize]. +/// +/// It is a logic error for `start` to be greater than `end`. +#[derive(Default, Copy, Clone, Eq, PartialEq, Hash)] +pub struct TextRange { + // Invariant: start <= end + start: TextSize, + end: TextSize, +} + +impl fmt::Debug for TextRange { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(f, "{}..{}", self.start().raw, self.end().raw) + } +} + +impl TextRange { + /// Creates a new `TextRange` with the given `start` and `end` (`start..end`). + /// + /// # Panics + /// + /// Panics if `end < start`. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let start = TextSize::from(5); + /// let end = TextSize::from(10); + /// let range = TextRange::new(start, end); + /// + /// assert_eq!(range.start(), start); + /// assert_eq!(range.end(), end); + /// assert_eq!(range.len(), end - start); + /// ``` + #[inline] + pub fn new(start: TextSize, end: TextSize) -> TextRange { + assert!(start <= end); + TextRange { start, end } + } + + /// Create a new `TextRange` with the given `offset` and `len` (`offset..offset + len`). + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let text = "0123456789"; + /// + /// let offset = TextSize::from(2); + /// let length = TextSize::from(5); + /// let range = TextRange::at(offset, length); + /// + /// assert_eq!(range, TextRange::new(offset, offset + length)); + /// assert_eq!(&text[range], "23456") + /// ``` + #[inline] + pub fn at(offset: TextSize, len: TextSize) -> TextRange { + TextRange::new(offset, offset + len) + } + + /// Create a zero-length range at the specified offset (`offset..offset`). + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let point: TextSize; + /// # point = TextSize::from(3); + /// let range = TextRange::empty(point); + /// assert!(range.is_empty()); + /// assert_eq!(range, TextRange::new(point, point)); + /// ``` + #[inline] + pub fn empty(offset: TextSize) -> TextRange { + TextRange { + start: offset, + end: offset, + } + } + + /// Create a range up to the given end (`..end`). + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let point: TextSize; + /// # point = TextSize::from(12); + /// let range = TextRange::up_to(point); + /// + /// assert_eq!(range.len(), point); + /// assert_eq!(range, TextRange::new(0.into(), point)); + /// assert_eq!(range, TextRange::at(0.into(), point)); + /// ``` + #[inline] + pub fn up_to(end: TextSize) -> TextRange { + TextRange { + start: 0.into(), + end, + } + } +} + +/// Identity methods. +impl TextRange { + /// The start point of this range. + #[inline] + pub const fn start(self) -> TextSize { + self.start + } + + /// The end point of this range. + #[inline] + pub const fn end(self) -> TextSize { + self.end + } + + /// The size of this range. + #[inline] + pub const fn len(self) -> TextSize { + // HACK for const fn: math on primitives only + TextSize { + raw: self.end().raw - self.start().raw, + } + } + + /// Check if this range is empty. + #[inline] + pub const fn is_empty(self) -> bool { + // HACK for const fn: math on primitives only + self.start().raw == self.end().raw + } +} + +/// Manipulation methods. +impl TextRange { + /// Check if this range contains an offset. + /// + /// The end index is considered excluded. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let (start, end): (TextSize, TextSize); + /// # start = 10.into(); end = 20.into(); + /// let range = TextRange::new(start, end); + /// assert!(range.contains(start)); + /// assert!(!range.contains(end)); + /// ``` + #[inline] + pub fn contains(self, offset: TextSize) -> bool { + self.start() <= offset && offset < self.end() + } + + /// Check if this range contains an offset. + /// + /// The end index is considered included. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let (start, end): (TextSize, TextSize); + /// # start = 10.into(); end = 20.into(); + /// let range = TextRange::new(start, end); + /// assert!(range.contains_inclusive(start)); + /// assert!(range.contains_inclusive(end)); + /// ``` + #[inline] + pub fn contains_inclusive(self, offset: TextSize) -> bool { + self.start() <= offset && offset <= self.end() + } + + /// Check if this range completely contains another range. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let larger = TextRange::new(0.into(), 20.into()); + /// let smaller = TextRange::new(5.into(), 15.into()); + /// assert!(larger.contains_range(smaller)); + /// assert!(!smaller.contains_range(larger)); + /// + /// // a range always contains itself + /// assert!(larger.contains_range(larger)); + /// assert!(smaller.contains_range(smaller)); + /// ``` + #[inline] + pub fn contains_range(self, other: TextRange) -> bool { + self.start() <= other.start() && other.end() <= self.end() + } + + /// The range covered by both ranges, if it exists. + /// If the ranges touch but do not overlap, the output range is empty. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// assert_eq!( + /// TextRange::intersect( + /// TextRange::new(0.into(), 10.into()), + /// TextRange::new(5.into(), 15.into()), + /// ), + /// Some(TextRange::new(5.into(), 10.into())), + /// ); + /// ``` + #[inline] + pub fn intersect(self, other: TextRange) -> Option<TextRange> { + let start = cmp::max(self.start(), other.start()); + let end = cmp::min(self.end(), other.end()); + if end < start { + return None; + } + Some(TextRange::new(start, end)) + } + + /// Extends the range to cover `other` as well. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// assert_eq!( + /// TextRange::cover( + /// TextRange::new(0.into(), 5.into()), + /// TextRange::new(15.into(), 20.into()), + /// ), + /// TextRange::new(0.into(), 20.into()), + /// ); + /// ``` + #[inline] + pub fn cover(self, other: TextRange) -> TextRange { + let start = cmp::min(self.start(), other.start()); + let end = cmp::max(self.end(), other.end()); + TextRange::new(start, end) + } + + /// Extends the range to cover `other` offsets as well. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// assert_eq!( + /// TextRange::empty(0.into()).cover_offset(20.into()), + /// TextRange::new(0.into(), 20.into()), + /// ) + /// ``` + #[inline] + pub fn cover_offset(self, offset: TextSize) -> TextRange { + self.cover(TextRange::empty(offset)) + } + + /// Add an offset to this range. + /// + /// Note that this is not appropriate for changing where a `TextRange` is + /// within some string; rather, it is for changing the reference anchor + /// that the `TextRange` is measured against. + /// + /// The unchecked version (`Add::add`) will _always_ panic on overflow, + /// in contrast to primitive integers, which check in debug mode only. + #[inline] + pub fn checked_add(self, offset: TextSize) -> Option<TextRange> { + Some(TextRange { + start: self.start.checked_add(offset)?, + end: self.end.checked_add(offset)?, + }) + } + + /// Subtract an offset from this range. + /// + /// Note that this is not appropriate for changing where a `TextRange` is + /// within some string; rather, it is for changing the reference anchor + /// that the `TextRange` is measured against. + /// + /// The unchecked version (`Sub::sub`) will _always_ panic on overflow, + /// in contrast to primitive integers, which check in debug mode only. + #[inline] + pub fn checked_sub(self, offset: TextSize) -> Option<TextRange> { + Some(TextRange { + start: self.start.checked_sub(offset)?, + end: self.end.checked_sub(offset)?, + }) + } + + /// Relative order of the two ranges (overlapping ranges are considered + /// equal). + /// + /// + /// This is useful when, for example, binary searching an array of disjoint + /// ranges. + /// + /// # Examples + /// + /// ``` + /// # use text_size::*; + /// # use std::cmp::Ordering; + /// + /// let a = TextRange::new(0.into(), 3.into()); + /// let b = TextRange::new(4.into(), 5.into()); + /// assert_eq!(a.ordering(b), Ordering::Less); + /// + /// let a = TextRange::new(0.into(), 3.into()); + /// let b = TextRange::new(3.into(), 5.into()); + /// assert_eq!(a.ordering(b), Ordering::Less); + /// + /// let a = TextRange::new(0.into(), 3.into()); + /// let b = TextRange::new(2.into(), 5.into()); + /// assert_eq!(a.ordering(b), Ordering::Equal); + /// + /// let a = TextRange::new(0.into(), 3.into()); + /// let b = TextRange::new(2.into(), 2.into()); + /// assert_eq!(a.ordering(b), Ordering::Equal); + /// + /// let a = TextRange::new(2.into(), 3.into()); + /// let b = TextRange::new(2.into(), 2.into()); + /// assert_eq!(a.ordering(b), Ordering::Greater); + /// ``` + #[inline] + pub fn ordering(self, other: TextRange) -> Ordering { + if self.end() <= other.start() { + Ordering::Less + } else if other.end() <= self.start() { + Ordering::Greater + } else { + Ordering::Equal + } + } +} + +impl Index<TextRange> for str { + type Output = str; + #[inline] + fn index(&self, index: TextRange) -> &str { + &self[Range::<usize>::from(index)] + } +} + +impl Index<TextRange> for String { + type Output = str; + #[inline] + fn index(&self, index: TextRange) -> &str { + &self[Range::<usize>::from(index)] + } +} + +impl IndexMut<TextRange> for str { + #[inline] + fn index_mut(&mut self, index: TextRange) -> &mut str { + &mut self[Range::<usize>::from(index)] + } +} + +impl IndexMut<TextRange> for String { + #[inline] + fn index_mut(&mut self, index: TextRange) -> &mut str { + &mut self[Range::<usize>::from(index)] + } +} + +impl RangeBounds<TextSize> for TextRange { + fn start_bound(&self) -> Bound<&TextSize> { + Bound::Included(&self.start) + } + + fn end_bound(&self) -> Bound<&TextSize> { + Bound::Excluded(&self.end) + } +} + +impl<T> From<TextRange> for Range<T> +where + T: From<TextSize>, +{ + #[inline] + fn from(r: TextRange) -> Self { + r.start().into()..r.end().into() + } +} + +macro_rules! ops { + (impl $Op:ident for TextRange by fn $f:ident = $op:tt) => { + impl $Op<&TextSize> for TextRange { + type Output = TextRange; + #[inline] + fn $f(self, other: &TextSize) -> TextRange { + self $op *other + } + } + impl<T> $Op<T> for &TextRange + where + TextRange: $Op<T, Output=TextRange>, + { + type Output = TextRange; + #[inline] + fn $f(self, other: T) -> TextRange { + *self $op other + } + } + }; +} + +impl Add<TextSize> for TextRange { + type Output = TextRange; + #[inline] + fn add(self, offset: TextSize) -> TextRange { + self.checked_add(offset) + .expect("TextRange +offset overflowed") + } +} + +impl Sub<TextSize> for TextRange { + type Output = TextRange; + #[inline] + fn sub(self, offset: TextSize) -> TextRange { + self.checked_sub(offset) + .expect("TextRange -offset overflowed") + } +} + +ops!(impl Add for TextRange by fn add = +); +ops!(impl Sub for TextRange by fn sub = -); + +impl<A> AddAssign<A> for TextRange +where + TextRange: Add<A, Output = TextRange>, +{ + #[inline] + fn add_assign(&mut self, rhs: A) { + *self = *self + rhs + } +} + +impl<S> SubAssign<S> for TextRange +where + TextRange: Sub<S, Output = TextRange>, +{ + #[inline] + fn sub_assign(&mut self, rhs: S) { + *self = *self - rhs + } +} diff --git a/vendor/text-size/src/serde_impls.rs b/vendor/text-size/src/serde_impls.rs new file mode 100644 index 000000000..a94bee956 --- /dev/null +++ b/vendor/text-size/src/serde_impls.rs @@ -0,0 +1,48 @@ +use { + crate::{TextRange, TextSize}, + serde::{de, Deserialize, Deserializer, Serialize, Serializer}, +}; + +impl Serialize for TextSize { + fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error> + where + S: Serializer, + { + self.raw.serialize(serializer) + } +} + +impl<'de> Deserialize<'de> for TextSize { + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: Deserializer<'de>, + { + u32::deserialize(deserializer).map(TextSize::from) + } +} + +impl Serialize for TextRange { + fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error> + where + S: Serializer, + { + (self.start(), self.end()).serialize(serializer) + } +} + +impl<'de> Deserialize<'de> for TextRange { + #[allow(clippy::nonminimal_bool)] + fn deserialize<D>(deserializer: D) -> Result<Self, D::Error> + where + D: Deserializer<'de>, + { + let (start, end) = Deserialize::deserialize(deserializer)?; + if !(start <= end) { + return Err(de::Error::custom(format!( + "invalid range: {:?}..{:?}", + start, end + ))); + } + Ok(TextRange::new(start, end)) + } +} diff --git a/vendor/text-size/src/size.rs b/vendor/text-size/src/size.rs new file mode 100644 index 000000000..ab2ec9a73 --- /dev/null +++ b/vendor/text-size/src/size.rs @@ -0,0 +1,161 @@ +use { + crate::TextLen, + std::{ + convert::TryFrom, + fmt, iter, + num::TryFromIntError, + ops::{Add, AddAssign, Sub, SubAssign}, + u32, + }, +}; + +/// A measure of text length. Also, equivalently, an index into text. +/// +/// This is a UTF-8 bytes offset stored as `u32`, but +/// most clients should treat it as an opaque measure. +/// +/// For cases that need to escape `TextSize` and return to working directly +/// with primitive integers, `TextSize` can be converted losslessly to/from +/// `u32` via [`From`] conversions as well as losslessly be converted [`Into`] +/// `usize`. The `usize -> TextSize` direction can be done via [`TryFrom`]. +/// +/// These escape hatches are primarily required for unit testing and when +/// converting from UTF-8 size to another coordinate space, such as UTF-16. +#[derive(Clone, Copy, Default, PartialEq, Eq, PartialOrd, Ord, Hash)] +pub struct TextSize { + pub(crate) raw: u32, +} + +impl fmt::Debug for TextSize { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(f, "{}", self.raw) + } +} + +impl TextSize { + /// The text size of some primitive text-like object. + /// + /// Accepts `char`, `&str`, and `&String`. + /// + /// # Examples + /// + /// ```rust + /// # use text_size::*; + /// let char_size = TextSize::of('🦀'); + /// assert_eq!(char_size, TextSize::from(4)); + /// + /// let str_size = TextSize::of("rust-analyzer"); + /// assert_eq!(str_size, TextSize::from(13)); + /// ``` + #[inline] + pub fn of<T: TextLen>(text: T) -> TextSize { + text.text_len() + } +} + +/// Methods to act like a primitive integer type, where reasonably applicable. +// Last updated for parity with Rust 1.42.0. +impl TextSize { + /// Checked addition. Returns `None` if overflow occurred. + #[inline] + pub fn checked_add(self, rhs: TextSize) -> Option<TextSize> { + self.raw.checked_add(rhs.raw).map(|raw| TextSize { raw }) + } + + /// Checked subtraction. Returns `None` if overflow occurred. + #[inline] + pub fn checked_sub(self, rhs: TextSize) -> Option<TextSize> { + self.raw.checked_sub(rhs.raw).map(|raw| TextSize { raw }) + } +} + +impl From<u32> for TextSize { + #[inline] + fn from(raw: u32) -> Self { + TextSize { raw } + } +} + +impl From<TextSize> for u32 { + #[inline] + fn from(value: TextSize) -> Self { + value.raw + } +} + +impl TryFrom<usize> for TextSize { + type Error = TryFromIntError; + #[inline] + fn try_from(value: usize) -> Result<Self, TryFromIntError> { + Ok(u32::try_from(value)?.into()) + } +} + +impl From<TextSize> for usize { + #[inline] + fn from(value: TextSize) -> Self { + value.raw as usize + } +} + +macro_rules! ops { + (impl $Op:ident for TextSize by fn $f:ident = $op:tt) => { + impl $Op<TextSize> for TextSize { + type Output = TextSize; + #[inline] + fn $f(self, other: TextSize) -> TextSize { + TextSize { raw: self.raw $op other.raw } + } + } + impl $Op<&TextSize> for TextSize { + type Output = TextSize; + #[inline] + fn $f(self, other: &TextSize) -> TextSize { + self $op *other + } + } + impl<T> $Op<T> for &TextSize + where + TextSize: $Op<T, Output=TextSize>, + { + type Output = TextSize; + #[inline] + fn $f(self, other: T) -> TextSize { + *self $op other + } + } + }; +} + +ops!(impl Add for TextSize by fn add = +); +ops!(impl Sub for TextSize by fn sub = -); + +impl<A> AddAssign<A> for TextSize +where + TextSize: Add<A, Output = TextSize>, +{ + #[inline] + fn add_assign(&mut self, rhs: A) { + *self = *self + rhs + } +} + +impl<S> SubAssign<S> for TextSize +where + TextSize: Sub<S, Output = TextSize>, +{ + #[inline] + fn sub_assign(&mut self, rhs: S) { + *self = *self - rhs + } +} + +impl<A> iter::Sum<A> for TextSize +where + TextSize: Add<A, Output = TextSize>, +{ + #[inline] + fn sum<I: Iterator<Item = A>>(iter: I) -> TextSize { + iter.fold(0.into(), Add::add) + } +} diff --git a/vendor/text-size/src/traits.rs b/vendor/text-size/src/traits.rs new file mode 100644 index 000000000..d0bb6c1f6 --- /dev/null +++ b/vendor/text-size/src/traits.rs @@ -0,0 +1,36 @@ +use {crate::TextSize, std::convert::TryInto}; + +use priv_in_pub::Sealed; +mod priv_in_pub { + pub trait Sealed {} +} + +/// Primitives with a textual length that can be passed to [`TextSize::of`]. +pub trait TextLen: Copy + Sealed { + /// The textual length of this primitive. + fn text_len(self) -> TextSize; +} + +impl Sealed for &'_ str {} +impl TextLen for &'_ str { + #[inline] + fn text_len(self) -> TextSize { + self.len().try_into().unwrap() + } +} + +impl Sealed for &'_ String {} +impl TextLen for &'_ String { + #[inline] + fn text_len(self) -> TextSize { + self.as_str().text_len() + } +} + +impl Sealed for char {} +impl TextLen for char { + #[inline] + fn text_len(self) -> TextSize { + (self.len_utf8() as u32).into() + } +} |