diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:20:29 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:20:29 +0000 |
commit | 631cd5845e8de329d0e227aaa707d7ea228b8f8f (patch) | |
tree | a1b87c8f8cad01cf18f7c5f57a08f102771ed303 /vendor/clap_lex | |
parent | Adding debian version 1.69.0+dfsg1-1. (diff) | |
download | rustc-631cd5845e8de329d0e227aaa707d7ea228b8f8f.tar.xz rustc-631cd5845e8de329d0e227aaa707d7ea228b8f8f.zip |
Merging upstream version 1.70.0+dfsg1.
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'vendor/clap_lex')
-rw-r--r-- | vendor/clap_lex/.cargo-checksum.json | 2 | ||||
-rw-r--r-- | vendor/clap_lex/Cargo.toml | 10 | ||||
-rw-r--r-- | vendor/clap_lex/README.md | 8 | ||||
-rw-r--r-- | vendor/clap_lex/src/ext.rs | 354 | ||||
-rw-r--r-- | vendor/clap_lex/src/lib.rs | 128 |
5 files changed, 415 insertions, 87 deletions
diff --git a/vendor/clap_lex/.cargo-checksum.json b/vendor/clap_lex/.cargo-checksum.json index a493abc0f..096f62c28 100644 --- a/vendor/clap_lex/.cargo-checksum.json +++ b/vendor/clap_lex/.cargo-checksum.json @@ -1 +1 @@ -{"files":{"Cargo.toml":"f2eca81e3cfadfd80956be964ce0d06c7cb1ad735679c38ed86836cd2780e26f","LICENSE-APACHE":"c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4","LICENSE-MIT":"0d687e1f07b58fe68bda74668ff6326125e5e5efa184cce755cd84ac535b7058","README.md":"d8a7c26d4ab47ac8ca387aada755d4dc375e1e85ed52b15575bb67ec118e7cb4","src/lib.rs":"2604cd6a948e0c56da4e8da3f28932476e3243e8c8a9a357dcccf6dddd55ce9c"},"package":"0d4198f73e42b4936b35b5bb248d81d2b595ecb170da0bac7655c54eedfa8da8"}
\ No newline at end of file +{"files":{"Cargo.toml":"c9eb0f99077d05483dfe7737502d90522d1d3271c98d1ac53f6e5ae7a5d4f582","LICENSE-APACHE":"c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4","LICENSE-MIT":"0d687e1f07b58fe68bda74668ff6326125e5e5efa184cce755cd84ac535b7058","README.md":"d1f4162770bfbd32c84dc9d914dceeb6e5050ab8be5037e7eb4539b8bbbb8e73","src/ext.rs":"7eae476b1f6919949e2c72331117a34e11eb3c21eebc050410870e7bbd06092c","src/lib.rs":"233a2661c7cfb61d95848e0fc2d37477912c29474a8db0c858ac71bc6af86d57"},"package":"8a2dd5a6fe8c6e3502f568a6353e5273bbb15193ad9a89e457b9970798efbea1"}
\ No newline at end of file diff --git a/vendor/clap_lex/Cargo.toml b/vendor/clap_lex/Cargo.toml index d6859256b..5fa2b9292 100644 --- a/vendor/clap_lex/Cargo.toml +++ b/vendor/clap_lex/Cargo.toml @@ -11,9 +11,9 @@ [package] edition = "2021" -rust-version = "1.60.0" +rust-version = "1.64.0" name = "clap_lex" -version = "0.3.0" +version = "0.4.1" include = [ "build.rs", "src/**/*", @@ -35,7 +35,6 @@ keywords = [ categories = ["command-line-interface"] license = "MIT OR Apache-2.0" repository = "https://github.com/clap-rs/clap/tree/master/clap_lex" -resolver = "2" [[package.metadata.release.pre-release-replacements]] file = "CHANGELOG.md" @@ -81,8 +80,3 @@ prerelease = true [lib] bench = false - -[dependencies.os_str_bytes] -version = "6.0" -features = ["raw_os_str"] -default-features = false diff --git a/vendor/clap_lex/README.md b/vendor/clap_lex/README.md index eebd4c571..18d856205 100644 --- a/vendor/clap_lex/README.md +++ b/vendor/clap_lex/README.md @@ -5,15 +5,15 @@ [![Crates.io](https://img.shields.io/crates/v/clap_lex?style=flat-square)](https://crates.io/crates/clap_lex) [![Crates.io](https://img.shields.io/crates/d/clap_lex?style=flat-square)](https://crates.io/crates/clap_lex) -[![License](https://img.shields.io/badge/license-Apache%202.0-blue?style=flat-square)](https://github.com/clap-rs/clap/blob/clap_lex-v0.3.0/LICENSE-APACHE) -[![License](https://img.shields.io/badge/license-MIT-blue?style=flat-square)](https://github.com/clap-rs/clap/blob/clap_lex-v0.3.0/LICENSE-MIT) +[![License](https://img.shields.io/badge/license-Apache%202.0-blue?style=flat-square)](https://github.com/clap-rs/clap/blob/clap_lex-v0.4.1/LICENSE-APACHE) +[![License](https://img.shields.io/badge/license-MIT-blue?style=flat-square)](https://github.com/clap-rs/clap/blob/clap_lex-v0.4.1/LICENSE-MIT) Dual-licensed under [Apache 2.0](LICENSE-APACHE) or [MIT](LICENSE-MIT). 1. [About](#about) 2. [API Reference](https://docs.rs/clap_lex) 3. [Questions & Discussions](https://github.com/clap-rs/clap/discussions) -4. [CONTRIBUTING](https://github.com/clap-rs/clap/blob/clap_lex-v0.3.0/clap_lex/CONTRIBUTING.md) -5. [Sponsors](https://github.com/clap-rs/clap/blob/clap_lex-v0.3.0/README.md#sponsors) +4. [CONTRIBUTING](https://github.com/clap-rs/clap/blob/clap_lex-v0.4.1/clap_lex/CONTRIBUTING.md) +5. [Sponsors](https://github.com/clap-rs/clap/blob/clap_lex-v0.4.1/README.md#sponsors) ## About diff --git a/vendor/clap_lex/src/ext.rs b/vendor/clap_lex/src/ext.rs new file mode 100644 index 000000000..0e6b33ebb --- /dev/null +++ b/vendor/clap_lex/src/ext.rs @@ -0,0 +1,354 @@ +use std::ffi::OsStr; + +pub trait OsStrExt: private::Sealed { + /// Converts to a string slice. + fn try_str(&self) -> Result<&str, std::str::Utf8Error>; + /// Returns `true` if the given pattern matches a sub-slice of + /// this string slice. + /// + /// Returns `false` if it does not. + /// + /// # Examples + /// + /// ```rust + /// use clap_lex::OsStrExt as _; + /// let bananas = std::ffi::OsStr::new("bananas"); + /// + /// assert!(bananas.contains("nana")); + /// assert!(!bananas.contains("apples")); + /// ``` + fn contains(&self, needle: &str) -> bool; + /// Returns the byte index of the first character of this string slice that + /// matches the pattern. + /// + /// Returns [`None`] if the pattern doesn't match. + /// + /// # Examples + /// + /// ```rust + /// use clap_lex::OsStrExt as _; + /// let s = std::ffi::OsStr::new("Löwe 老虎 Léopard Gepardi"); + /// + /// assert_eq!(s.find("L"), Some(0)); + /// assert_eq!(s.find("é"), Some(14)); + /// assert_eq!(s.find("par"), Some(17)); + /// ``` + /// + /// Not finding the pattern: + /// + /// ```rust + /// use clap_lex::OsStrExt as _; + /// let s = std::ffi::OsStr::new("Löwe 老虎 Léopard"); + /// + /// assert_eq!(s.find("1"), None); + /// ``` + fn find(&self, needle: &str) -> Option<usize>; + /// Returns a string slice with the prefix removed. + /// + /// If the string starts with the pattern `prefix`, returns substring after the prefix, wrapped + /// in `Some`. + /// + /// If the string does not start with `prefix`, returns `None`. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// assert_eq!(OsStr::new("foo:bar").strip_prefix("foo:"), Some(OsStr::new("bar"))); + /// assert_eq!(OsStr::new("foo:bar").strip_prefix("bar"), None); + /// assert_eq!(OsStr::new("foofoo").strip_prefix("foo"), Some(OsStr::new("foo"))); + /// ``` + fn strip_prefix(&self, prefix: &str) -> Option<&OsStr>; + /// Returns `true` if the given pattern matches a prefix of this + /// string slice. + /// + /// Returns `false` if it does not. + /// + /// # Examples + /// + /// ``` + /// use clap_lex::OsStrExt as _; + /// let bananas = std::ffi::OsStr::new("bananas"); + /// + /// assert!(bananas.starts_with("bana")); + /// assert!(!bananas.starts_with("nana")); + /// ``` + fn starts_with(&self, prefix: &str) -> bool; + /// An iterator over substrings of this string slice, separated by + /// characters matched by a pattern. + /// + /// # Examples + /// + /// Simple patterns: + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let v: Vec<_> = OsStr::new("Mary had a little lamb").split(" ").collect(); + /// assert_eq!(v, [OsStr::new("Mary"), OsStr::new("had"), OsStr::new("a"), OsStr::new("little"), OsStr::new("lamb")]); + /// + /// let v: Vec<_> = OsStr::new("").split("X").collect(); + /// assert_eq!(v, [OsStr::new("")]); + /// + /// let v: Vec<_> = OsStr::new("lionXXtigerXleopard").split("X").collect(); + /// assert_eq!(v, [OsStr::new("lion"), OsStr::new(""), OsStr::new("tiger"), OsStr::new("leopard")]); + /// + /// let v: Vec<_> = OsStr::new("lion::tiger::leopard").split("::").collect(); + /// assert_eq!(v, [OsStr::new("lion"), OsStr::new("tiger"), OsStr::new("leopard")]); + /// ``` + /// + /// If a string contains multiple contiguous separators, you will end up + /// with empty strings in the output: + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let x = OsStr::new("||||a||b|c"); + /// let d: Vec<_> = x.split("|").collect(); + /// + /// assert_eq!(d, &[OsStr::new(""), OsStr::new(""), OsStr::new(""), OsStr::new(""), OsStr::new("a"), OsStr::new(""), OsStr::new("b"), OsStr::new("c")]); + /// ``` + /// + /// Contiguous separators are separated by the empty string. + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let x = OsStr::new("(///)"); + /// let d: Vec<_> = x.split("/").collect(); + /// + /// assert_eq!(d, &[OsStr::new("("), OsStr::new(""), OsStr::new(""), OsStr::new(")")]); + /// ``` + /// + /// Separators at the start or end of a string are neighbored + /// by empty strings. + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let d: Vec<_> = OsStr::new("010").split("0").collect(); + /// assert_eq!(d, &[OsStr::new(""), OsStr::new("1"), OsStr::new("")]); + /// ``` + /// + /// When the empty string is used as a separator, it panics + /// + /// ```should_panic + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let f: Vec<_> = OsStr::new("rust").split("").collect(); + /// assert_eq!(f, &[OsStr::new(""), OsStr::new("r"), OsStr::new("u"), OsStr::new("s"), OsStr::new("t"), OsStr::new("")]); + /// ``` + /// + /// Contiguous separators can lead to possibly surprising behavior + /// when whitespace is used as the separator. This code is correct: + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// let x = OsStr::new(" a b c"); + /// let d: Vec<_> = x.split(" ").collect(); + /// + /// assert_eq!(d, &[OsStr::new(""), OsStr::new(""), OsStr::new(""), OsStr::new(""), OsStr::new("a"), OsStr::new(""), OsStr::new("b"), OsStr::new("c")]); + /// ``` + /// + /// It does _not_ give you: + /// + /// ```,ignore + /// assert_eq!(d, &[OsStr::new("a"), OsStr::new("b"), OsStr::new("c")]); + /// ``` + /// + /// Use [`split_whitespace`] for this behavior. + /// + /// [`split_whitespace`]: str::split_whitespace + fn split<'s, 'n>(&'s self, needle: &'n str) -> Split<'s, 'n>; + /// Divide one string slice into two at an index. + /// + /// The argument, `mid`, should be a byte offset from the start of the + /// string. It must also be on the boundary of a UTF-8 code point. + /// + /// The two slices returned go from the start of the string slice to `mid`, + /// and from `mid` to the end of the string slice. + /// + /// To get mutable string slices instead, see the [`split_at_mut`] + /// method. + /// + /// [`split_at_mut`]: str::split_at_mut + /// + /// # Panics + /// + /// Panics if `mid` is not on a UTF-8 code point boundary, or if it is + /// past the end of the last code point of the string slice. + /// + /// # Examples + /// + /// Basic usage: + /// + /// ``` + /// use clap_lex::OsStrExt as _; + /// let s = std::ffi::OsStr::new("Per Martin-Löf"); + /// + /// let (first, last) = s.split_at(3); + /// + /// assert_eq!("Per", first); + /// assert_eq!(" Martin-Löf", last); + /// ``` + #[deprecated(since = "4.1.0", note = "This is not sound for all `index`")] + fn split_at(&self, index: usize) -> (&OsStr, &OsStr); + /// Splits the string on the first occurrence of the specified delimiter and + /// returns prefix before delimiter and suffix after delimiter. + /// + /// # Examples + /// + /// ``` + /// use std::ffi::OsStr; + /// use clap_lex::OsStrExt as _; + /// assert_eq!(OsStr::new("cfg").split_once("="), None); + /// assert_eq!(OsStr::new("cfg=").split_once("="), Some((OsStr::new("cfg"), OsStr::new("")))); + /// assert_eq!(OsStr::new("cfg=foo").split_once("="), Some((OsStr::new("cfg"), OsStr::new("foo")))); + /// assert_eq!(OsStr::new("cfg=foo=bar").split_once("="), Some((OsStr::new("cfg"), OsStr::new("foo=bar")))); + /// ``` + fn split_once(&self, needle: &'_ str) -> Option<(&OsStr, &OsStr)>; +} + +impl OsStrExt for OsStr { + fn try_str(&self) -> Result<&str, std::str::Utf8Error> { + let bytes = to_bytes(self); + std::str::from_utf8(bytes) + } + + fn contains(&self, needle: &str) -> bool { + self.find(needle).is_some() + } + + fn find(&self, needle: &str) -> Option<usize> { + let bytes = to_bytes(self); + (0..=self.len().checked_sub(needle.len())?) + .find(|&x| bytes[x..].starts_with(needle.as_bytes())) + } + + fn strip_prefix(&self, prefix: &str) -> Option<&OsStr> { + let bytes = to_bytes(self); + bytes.strip_prefix(prefix.as_bytes()).map(|s| { + // SAFETY: + // - This came from `to_bytes` + // - Since `prefix` is `&str`, any split will be along UTF-8 boundarie + unsafe { to_os_str_unchecked(s) } + }) + } + fn starts_with(&self, prefix: &str) -> bool { + let bytes = to_bytes(self); + bytes.starts_with(prefix.as_bytes()) + } + + fn split<'s, 'n>(&'s self, needle: &'n str) -> Split<'s, 'n> { + assert_ne!(needle, ""); + Split { + haystack: Some(self), + needle, + } + } + + fn split_at(&self, index: usize) -> (&OsStr, &OsStr) { + let bytes = to_bytes(self); + unsafe { + // BUG: This is unsafe and has been deprecated + let (first, second) = bytes.split_at(index); + (to_os_str_unchecked(first), to_os_str_unchecked(second)) + } + } + + fn split_once(&self, needle: &'_ str) -> Option<(&OsStr, &OsStr)> { + let start = self.find(needle)?; + let end = start + needle.len(); + let haystack = to_bytes(self); + let first = &haystack[0..start]; + let second = &haystack[end..]; + // SAFETY: + // - This came from `to_bytes` + // - Since `needle` is `&str`, any split will be along UTF-8 boundarie + unsafe { Some((to_os_str_unchecked(first), to_os_str_unchecked(second))) } + } +} + +mod private { + pub trait Sealed {} + + impl Sealed for std::ffi::OsStr {} +} + +/// Allow access to raw bytes +/// +/// As the non-UTF8 encoding is not defined, the bytes only make sense when compared with +/// 7-bit ASCII or `&str` +/// +/// # Compatibility +/// +/// There is no guarantee how non-UTF8 bytes will be encoded, even within versions of this crate +/// (since its dependent on rustc) +fn to_bytes(s: &OsStr) -> &[u8] { + // SAFETY: + // - Lifetimes are the same + // - Types are compatible (`OsStr` is effectively a transparent wrapper for `[u8]`) + // - The primary contract is that the encoding for invalid surrogate code points is not + // guaranteed which isn't a problem here + // + // There is a proposal to support this natively (https://github.com/rust-lang/rust/pull/95290) + // but its in limbo + unsafe { std::mem::transmute(s) } +} + +/// Restore raw bytes as `OsStr` +/// +/// # Safety +/// +/// - `&[u8]` must either by a `&str` or originated with `to_bytes` within the same binary +/// - Any splits of the original `&[u8]` must be done along UTF-8 boundaries +unsafe fn to_os_str_unchecked(s: &[u8]) -> &OsStr { + // SAFETY: + // - Lifetimes are the same + // - Types are compatible (`OsStr` is effectively a transparent wrapper for `[u8]`) + // - The primary contract is that the encoding for invalid surrogate code points is not + // guaranteed which isn't a problem here + // + // There is a proposal to support this natively (https://github.com/rust-lang/rust/pull/95290) + // but its in limbo + std::mem::transmute(s) +} + +pub struct Split<'s, 'n> { + haystack: Option<&'s OsStr>, + needle: &'n str, +} + +impl<'s, 'n> Iterator for Split<'s, 'n> { + type Item = &'s OsStr; + + fn next(&mut self) -> Option<Self::Item> { + let haystack = self.haystack?; + match haystack.split_once(self.needle) { + Some((first, second)) => { + if !haystack.is_empty() { + debug_assert_ne!(haystack, second); + } + self.haystack = Some(second); + Some(first) + } + None => { + self.haystack = None; + Some(haystack) + } + } + } +} + +/// Split an `OsStr` +/// +/// # Safety +/// +/// `index` must be at a valid UTF-8 boundary +pub(crate) unsafe fn split_at(os: &OsStr, index: usize) -> (&OsStr, &OsStr) { + let bytes = to_bytes(os); + let (first, second) = bytes.split_at(index); + (to_os_str_unchecked(first), to_os_str_unchecked(second)) +} diff --git a/vendor/clap_lex/src/lib.rs b/vendor/clap_lex/src/lib.rs index bc4d4ba65..b349fba51 100644 --- a/vendor/clap_lex/src/lib.rs +++ b/vendor/clap_lex/src/lib.rs @@ -7,8 +7,11 @@ //! # Examples //! //! ```rust -//! # use std::path::PathBuf; -//! # type BoxedError = Box<dyn std::error::Error + Send + Sync>; +//! use std::path::PathBuf; +//! use std::ffi::OsStr; +//! +//! type BoxedError = Box<dyn std::error::Error + Send + Sync>; +//! //! #[derive(Debug)] //! struct Args { //! paths: Vec<PathBuf>, @@ -24,7 +27,7 @@ //! } //! //! impl Color { -//! fn parse(s: Option<&clap_lex::RawOsStr>) -> Result<Self, BoxedError> { +//! fn parse(s: Option<&OsStr>) -> Result<Self, BoxedError> { //! let s = s.map(|s| s.to_str().ok_or(s)); //! match s { //! Some(Ok("always")) | Some(Ok("")) | None => { @@ -91,12 +94,12 @@ //! return Err(format!("Unexpected flag: -{}", c).into()); //! } //! Err(e) => { -//! return Err(format!("Unexpected flag: -{}", e.to_str_lossy()).into()); +//! return Err(format!("Unexpected flag: -{}", e.to_string_lossy()).into()); //! } //! } //! } //! } else { -//! args.paths.push(PathBuf::from(arg.to_value_os().to_os_str().into_owned())); +//! args.paths.push(PathBuf::from(arg.to_value_os().to_owned())); //! } //! } //! @@ -107,13 +110,14 @@ //! println!("{:?}", args); //! ``` +mod ext; + use std::ffi::OsStr; use std::ffi::OsString; pub use std::io::SeekFrom; -pub use os_str_bytes::RawOsStr; -pub use os_str_bytes::RawOsString; +pub use ext::OsStrExt; /// Command-line arguments #[derive(Default, Clone, Debug, PartialEq, Eq)] @@ -273,30 +277,27 @@ impl ArgCursor { /// Command-line Argument #[derive(Clone, Debug, PartialEq, Eq, PartialOrd, Ord, Hash)] pub struct ParsedArg<'s> { - inner: std::borrow::Cow<'s, RawOsStr>, - utf8: Option<&'s str>, + inner: &'s OsStr, } impl<'s> ParsedArg<'s> { fn new(inner: &'s OsStr) -> Self { - let utf8 = inner.to_str(); - let inner = RawOsStr::new(inner); - Self { inner, utf8 } + Self { inner } } /// Argument is length of 0 pub fn is_empty(&self) -> bool { - self.inner.as_ref().is_empty() + self.inner.is_empty() } /// Does the argument look like a stdio argument (`-`) pub fn is_stdio(&self) -> bool { - self.inner.as_ref() == "-" + self.inner == "-" } /// Does the argument look like an argument escape (`--`) pub fn is_escape(&self) -> bool { - self.inner.as_ref() == "--" + self.inner == "--" } /// Does the argument look like a number @@ -307,56 +308,38 @@ impl<'s> ParsedArg<'s> { } /// Treat as a long-flag - pub fn to_long(&self) -> Option<(Result<&str, &RawOsStr>, Option<&RawOsStr>)> { - if let Some(raw) = self.utf8 { - let remainder = raw.strip_prefix("--")?; - if remainder.is_empty() { - debug_assert!(self.is_escape()); - return None; - } + pub fn to_long(&self) -> Option<(Result<&str, &OsStr>, Option<&OsStr>)> { + let raw = self.inner; + let remainder = raw.strip_prefix("--")?; + if remainder.is_empty() { + debug_assert!(self.is_escape()); + return None; + } - let (flag, value) = if let Some((p0, p1)) = remainder.split_once('=') { - (p0, Some(p1)) - } else { - (remainder, None) - }; - let flag = Ok(flag); - let value = value.map(RawOsStr::from_str); - Some((flag, value)) + let (flag, value) = if let Some((p0, p1)) = remainder.split_once("=") { + (p0, Some(p1)) } else { - let raw = self.inner.as_ref(); - let remainder = raw.strip_prefix("--")?; - if remainder.is_empty() { - debug_assert!(self.is_escape()); - return None; - } - - let (flag, value) = if let Some((p0, p1)) = remainder.split_once('=') { - (p0, Some(p1)) - } else { - (remainder, None) - }; - let flag = flag.to_str().ok_or(flag); - Some((flag, value)) - } + (remainder, None) + }; + let flag = flag.to_str().ok_or(flag); + Some((flag, value)) } /// Can treat as a long-flag pub fn is_long(&self) -> bool { - self.inner.as_ref().starts_with("--") && !self.is_escape() + self.inner.starts_with("--") && !self.is_escape() } /// Treat as a short-flag pub fn to_short(&self) -> Option<ShortFlags<'_>> { - if let Some(remainder_os) = self.inner.as_ref().strip_prefix('-') { - if remainder_os.starts_with('-') { + if let Some(remainder_os) = self.inner.strip_prefix("-") { + if remainder_os.starts_with("-") { None } else if remainder_os.is_empty() { debug_assert!(self.is_stdio()); None } else { - let remainder = self.utf8.map(|s| &s[1..]); - Some(ShortFlags::new(remainder_os, remainder)) + Some(ShortFlags::new(remainder_os)) } } else { None @@ -365,48 +348,42 @@ impl<'s> ParsedArg<'s> { /// Can treat as a short-flag pub fn is_short(&self) -> bool { - self.inner.as_ref().starts_with('-') - && !self.is_stdio() - && !self.inner.as_ref().starts_with("--") + self.inner.starts_with("-") && !self.is_stdio() && !self.inner.starts_with("--") } /// Treat as a value /// /// **NOTE:** May return a flag or an escape. - pub fn to_value_os(&self) -> &RawOsStr { - self.inner.as_ref() + pub fn to_value_os(&self) -> &OsStr { + self.inner } /// Treat as a value /// /// **NOTE:** May return a flag or an escape. - pub fn to_value(&self) -> Result<&str, &RawOsStr> { - self.utf8.ok_or_else(|| self.inner.as_ref()) + pub fn to_value(&self) -> Result<&str, &OsStr> { + self.inner.to_str().ok_or(self.inner) } /// Safely print an argument that may contain non-UTF8 content /// /// This may perform lossy conversion, depending on the platform. If you would like an implementation which escapes the path please use Debug instead. pub fn display(&self) -> impl std::fmt::Display + '_ { - self.inner.to_str_lossy() + self.inner.to_string_lossy() } } /// Walk through short flags within a [`ParsedArg`] #[derive(Clone, Debug)] pub struct ShortFlags<'s> { - inner: &'s RawOsStr, + inner: &'s OsStr, utf8_prefix: std::str::CharIndices<'s>, - invalid_suffix: Option<&'s RawOsStr>, + invalid_suffix: Option<&'s OsStr>, } impl<'s> ShortFlags<'s> { - fn new(inner: &'s RawOsStr, utf8: Option<&'s str>) -> Self { - let (utf8_prefix, invalid_suffix) = if let Some(utf8) = utf8 { - (utf8, None) - } else { - split_nonutf8_once(inner) - }; + fn new(inner: &'s OsStr) -> Self { + let (utf8_prefix, invalid_suffix) = split_nonutf8_once(inner); let utf8_prefix = utf8_prefix.char_indices(); Self { inner, @@ -438,7 +415,7 @@ impl<'s> ShortFlags<'s> { /// Advance the iterator, returning the next short flag on success /// /// On error, returns the invalid-UTF8 value - pub fn next_flag(&mut self) -> Option<Result<char, &'s RawOsStr>> { + pub fn next_flag(&mut self) -> Option<Result<char, &'s OsStr>> { if let Some((_, flag)) = self.utf8_prefix.next() { return Some(Ok(flag)); } @@ -452,11 +429,13 @@ impl<'s> ShortFlags<'s> { } /// Advance the iterator, returning everything left as a value - pub fn next_value_os(&mut self) -> Option<&'s RawOsStr> { + pub fn next_value_os(&mut self) -> Option<&'s OsStr> { if let Some((index, _)) = self.utf8_prefix.next() { self.utf8_prefix = "".char_indices(); self.invalid_suffix = None; - return Some(&self.inner[index..]); + // SAFETY: `char_indices` ensures `index` is at a valid UTF-8 boundary + let remainder = unsafe { ext::split_at(self.inner, index).1 }; + return Some(remainder); } if let Some(suffix) = self.invalid_suffix { @@ -469,19 +448,20 @@ impl<'s> ShortFlags<'s> { } impl<'s> Iterator for ShortFlags<'s> { - type Item = Result<char, &'s RawOsStr>; + type Item = Result<char, &'s OsStr>; fn next(&mut self) -> Option<Self::Item> { self.next_flag() } } -fn split_nonutf8_once(b: &RawOsStr) -> (&str, Option<&RawOsStr>) { - match std::str::from_utf8(b.as_raw_bytes()) { +fn split_nonutf8_once(b: &OsStr) -> (&str, Option<&OsStr>) { + match b.try_str() { Ok(s) => (s, None), Err(err) => { - let (valid, after_valid) = b.split_at(err.valid_up_to()); - let valid = std::str::from_utf8(valid.as_raw_bytes()).unwrap(); + // SAFETY: `char_indices` ensures `index` is at a valid UTF-8 boundary + let (valid, after_valid) = unsafe { ext::split_at(b, err.valid_up_to()) }; + let valid = valid.try_str().unwrap(); (valid, Some(after_valid)) } } |