diff options
author | Valentin Popov <valentin@popov.link> | 2024-01-08 00:21:28 +0300 |
---|---|---|
committer | Valentin Popov <valentin@popov.link> | 2024-01-08 00:21:28 +0300 |
commit | 1b6a04ca5504955c571d1c97504fb45ea0befee4 (patch) | |
tree | 7579f518b23313e8a9748a88ab6173d5e030b227 /vendor/clap_lex/src | |
parent | 5ecd8cf2cba827454317368b68571df0d13d7842 (diff) | |
download | fparkan-1b6a04ca5504955c571d1c97504fb45ea0befee4.tar.xz fparkan-1b6a04ca5504955c571d1c97504fb45ea0befee4.zip |
Initial vendor packages
Signed-off-by: Valentin Popov <valentin@popov.link>
Diffstat (limited to 'vendor/clap_lex/src')
-rw-r--r-- | vendor/clap_lex/src/ext.rs | 312 | ||||
-rw-r--r-- | vendor/clap_lex/src/lib.rs | 504 |
2 files changed, 816 insertions, 0 deletions
diff --git a/vendor/clap_lex/src/ext.rs b/vendor/clap_lex/src/ext.rs new file mode 100644 index 0000000..a2de707 --- /dev/null +++ b/vendor/clap_lex/src/ext.rs @@ -0,0 +1,312 @@ +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>; + /// 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_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 new file mode 100644 index 0000000..581595a --- /dev/null +++ b/vendor/clap_lex/src/lib.rs @@ -0,0 +1,504 @@ +//! Minimal, flexible command-line parser +//! +//! As opposed to a declarative parser, this processes arguments as a stream of tokens. As lexing +//! a command-line is not context-free, we rely on the caller to decide how to interpret the +//! arguments. +//! +//! # Examples +//! +//! ```rust +//! use std::path::PathBuf; +//! use std::ffi::OsStr; +//! +//! type BoxedError = Box<dyn std::error::Error + Send + Sync>; +//! +//! #[derive(Debug)] +//! struct Args { +//! paths: Vec<PathBuf>, +//! color: Color, +//! verbosity: usize, +//! } +//! +//! #[derive(Debug)] +//! enum Color { +//! Always, +//! Auto, +//! Never, +//! } +//! +//! impl Color { +//! 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 => { +//! Ok(Color::Always) +//! } +//! Some(Ok("auto")) => { +//! Ok(Color::Auto) +//! } +//! Some(Ok("never")) => { +//! Ok(Color::Never) +//! } +//! Some(invalid) => { +//! Err(format!("Invalid value for `--color`, {invalid:?}").into()) +//! } +//! } +//! } +//! } +//! +//! fn parse_args( +//! raw: impl IntoIterator<Item=impl Into<std::ffi::OsString>> +//! ) -> Result<Args, BoxedError> { +//! let mut args = Args { +//! paths: Vec::new(), +//! color: Color::Auto, +//! verbosity: 0, +//! }; +//! +//! let raw = clap_lex::RawArgs::new(raw); +//! let mut cursor = raw.cursor(); +//! raw.next(&mut cursor); // Skip the bin +//! while let Some(arg) = raw.next(&mut cursor) { +//! if arg.is_escape() { +//! args.paths.extend(raw.remaining(&mut cursor).map(PathBuf::from)); +//! } else if arg.is_stdio() { +//! args.paths.push(PathBuf::from("-")); +//! } else if let Some((long, value)) = arg.to_long() { +//! match long { +//! Ok("verbose") => { +//! if let Some(value) = value { +//! return Err(format!("`--verbose` does not take a value, got `{value:?}`").into()); +//! } +//! args.verbosity += 1; +//! } +//! Ok("color") => { +//! args.color = Color::parse(value)?; +//! } +//! _ => { +//! return Err( +//! format!("Unexpected flag: --{}", arg.display()).into() +//! ); +//! } +//! } +//! } else if let Some(mut shorts) = arg.to_short() { +//! while let Some(short) = shorts.next_flag() { +//! match short { +//! Ok('v') => { +//! args.verbosity += 1; +//! } +//! Ok('c') => { +//! let value = shorts.next_value_os(); +//! args.color = Color::parse(value)?; +//! } +//! Ok(c) => { +//! return Err(format!("Unexpected flag: -{c}").into()); +//! } +//! Err(e) => { +//! return Err(format!("Unexpected flag: -{}", e.to_string_lossy()).into()); +//! } +//! } +//! } +//! } else { +//! args.paths.push(PathBuf::from(arg.to_value_os().to_owned())); +//! } +//! } +//! +//! Ok(args) +//! } +//! +//! let args = parse_args(["bin", "--hello", "world"]); +//! println!("{args:?}"); +//! ``` + +mod ext; + +use std::ffi::OsStr; +use std::ffi::OsString; + +pub use std::io::SeekFrom; + +pub use ext::OsStrExt; + +/// Command-line arguments +#[derive(Default, Clone, Debug, PartialEq, Eq)] +pub struct RawArgs { + items: Vec<OsString>, +} + +impl RawArgs { + //// Create an argument list to parse + /// + /// **NOTE:** The argument returned will be the current binary. + /// + /// # Example + /// + /// ```rust,no_run + /// # use std::path::PathBuf; + /// let raw = clap_lex::RawArgs::from_args(); + /// let mut cursor = raw.cursor(); + /// let _bin = raw.next_os(&mut cursor); + /// + /// let mut paths = raw.remaining(&mut cursor).map(PathBuf::from).collect::<Vec<_>>(); + /// println!("{paths:?}"); + /// ``` + pub fn from_args() -> Self { + Self::new(std::env::args_os()) + } + + //// Create an argument list to parse + /// + /// # Example + /// + /// ```rust,no_run + /// # use std::path::PathBuf; + /// let raw = clap_lex::RawArgs::new(["bin", "foo.txt"]); + /// let mut cursor = raw.cursor(); + /// let _bin = raw.next_os(&mut cursor); + /// + /// let mut paths = raw.remaining(&mut cursor).map(PathBuf::from).collect::<Vec<_>>(); + /// println!("{paths:?}"); + /// ``` + pub fn new(iter: impl IntoIterator<Item = impl Into<std::ffi::OsString>>) -> Self { + let iter = iter.into_iter(); + Self::from(iter) + } + + /// Create a cursor for walking the arguments + /// + /// # Example + /// + /// ```rust,no_run + /// # use std::path::PathBuf; + /// let raw = clap_lex::RawArgs::new(["bin", "foo.txt"]); + /// let mut cursor = raw.cursor(); + /// let _bin = raw.next_os(&mut cursor); + /// + /// let mut paths = raw.remaining(&mut cursor).map(PathBuf::from).collect::<Vec<_>>(); + /// println!("{paths:?}"); + /// ``` + pub fn cursor(&self) -> ArgCursor { + ArgCursor::new() + } + + /// Advance the cursor, returning the next [`ParsedArg`] + pub fn next(&self, cursor: &mut ArgCursor) -> Option<ParsedArg<'_>> { + self.next_os(cursor).map(ParsedArg::new) + } + + /// Advance the cursor, returning a raw argument value. + pub fn next_os(&self, cursor: &mut ArgCursor) -> Option<&OsStr> { + let next = self.items.get(cursor.cursor).map(|s| s.as_os_str()); + cursor.cursor = cursor.cursor.saturating_add(1); + next + } + + /// Return the next [`ParsedArg`] + pub fn peek(&self, cursor: &ArgCursor) -> Option<ParsedArg<'_>> { + self.peek_os(cursor).map(ParsedArg::new) + } + + /// Return a raw argument value. + pub fn peek_os(&self, cursor: &ArgCursor) -> Option<&OsStr> { + self.items.get(cursor.cursor).map(|s| s.as_os_str()) + } + + /// Return all remaining raw arguments, advancing the cursor to the end + /// + /// # Example + /// + /// ```rust,no_run + /// # use std::path::PathBuf; + /// let raw = clap_lex::RawArgs::new(["bin", "foo.txt"]); + /// let mut cursor = raw.cursor(); + /// let _bin = raw.next_os(&mut cursor); + /// + /// let mut paths = raw.remaining(&mut cursor).map(PathBuf::from).collect::<Vec<_>>(); + /// println!("{paths:?}"); + /// ``` + pub fn remaining(&self, cursor: &mut ArgCursor) -> impl Iterator<Item = &OsStr> { + let remaining = self.items[cursor.cursor..].iter().map(|s| s.as_os_str()); + cursor.cursor = self.items.len(); + remaining + } + + /// Adjust the cursor's position + pub fn seek(&self, cursor: &mut ArgCursor, pos: SeekFrom) { + let pos = match pos { + SeekFrom::Start(pos) => pos, + SeekFrom::End(pos) => (self.items.len() as i64).saturating_add(pos).max(0) as u64, + SeekFrom::Current(pos) => (cursor.cursor as i64).saturating_add(pos).max(0) as u64, + }; + let pos = (pos as usize).min(self.items.len()); + cursor.cursor = pos; + } + + /// Inject arguments before the [`RawArgs::next`] + pub fn insert( + &mut self, + cursor: &ArgCursor, + insert_items: impl IntoIterator<Item = impl Into<OsString>>, + ) { + self.items.splice( + cursor.cursor..cursor.cursor, + insert_items.into_iter().map(Into::into), + ); + } + + /// Any remaining args? + pub fn is_end(&self, cursor: &ArgCursor) -> bool { + self.peek_os(cursor).is_none() + } +} + +impl<I, T> From<I> for RawArgs +where + I: Iterator<Item = T>, + T: Into<OsString>, +{ + fn from(val: I) -> Self { + Self { + items: val.map(|x| x.into()).collect(), + } + } +} + +/// Position within [`RawArgs`] +#[derive(Clone, Debug, PartialEq, Eq, PartialOrd, Ord)] +pub struct ArgCursor { + cursor: usize, +} + +impl ArgCursor { + fn new() -> Self { + Self { cursor: 0 } + } +} + +/// Command-line Argument +#[derive(Clone, Debug, PartialEq, Eq, PartialOrd, Ord, Hash)] +pub struct ParsedArg<'s> { + inner: &'s OsStr, +} + +impl<'s> ParsedArg<'s> { + fn new(inner: &'s OsStr) -> Self { + Self { inner } + } + + /// Argument is length of 0 + pub fn is_empty(&self) -> bool { + self.inner.is_empty() + } + + /// Does the argument look like a stdio argument (`-`) + pub fn is_stdio(&self) -> bool { + self.inner == "-" + } + + /// Does the argument look like an argument escape (`--`) + pub fn is_escape(&self) -> bool { + self.inner == "--" + } + + /// Does the argument look like a negative number? + /// + /// This won't parse the number in full but attempts to see if this looks + /// like something along the lines of `-3`, `-0.3`, or `-33.03` + pub fn is_negative_number(&self) -> bool { + self.to_value() + .ok() + .and_then(|s| Some(is_number(s.strip_prefix('-')?))) + .unwrap_or_default() + } + + /// Treat as a long-flag + 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 = flag.to_str().ok_or(flag); + Some((flag, value)) + } + + /// Can treat as a long-flag + pub fn is_long(&self) -> bool { + 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.strip_prefix("-") { + if remainder_os.starts_with("-") { + None + } else if remainder_os.is_empty() { + debug_assert!(self.is_stdio()); + None + } else { + Some(ShortFlags::new(remainder_os)) + } + } else { + None + } + } + + /// Can treat as a short-flag + pub fn is_short(&self) -> bool { + 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) -> &OsStr { + self.inner + } + + /// Treat as a value + /// + /// **NOTE:** May return a flag or an escape. + 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_string_lossy() + } +} + +/// Walk through short flags within a [`ParsedArg`] +#[derive(Clone, Debug)] +pub struct ShortFlags<'s> { + inner: &'s OsStr, + utf8_prefix: std::str::CharIndices<'s>, + invalid_suffix: Option<&'s OsStr>, +} + +impl<'s> ShortFlags<'s> { + fn new(inner: &'s OsStr) -> Self { + let (utf8_prefix, invalid_suffix) = split_nonutf8_once(inner); + let utf8_prefix = utf8_prefix.char_indices(); + Self { + inner, + utf8_prefix, + invalid_suffix, + } + } + + /// Move the iterator forward by `n` short flags + pub fn advance_by(&mut self, n: usize) -> Result<(), usize> { + for i in 0..n { + self.next().ok_or(i)?.map_err(|_| i)?; + } + Ok(()) + } + + /// No short flags left + pub fn is_empty(&self) -> bool { + self.invalid_suffix.is_none() && self.utf8_prefix.as_str().is_empty() + } + + /// Does the short flag look like a number + /// + /// Ideally call this before doing any iterator + pub fn is_negative_number(&self) -> bool { + self.invalid_suffix.is_none() && is_number(self.utf8_prefix.as_str()) + } + + /// 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 OsStr>> { + if let Some((_, flag)) = self.utf8_prefix.next() { + return Some(Ok(flag)); + } + + if let Some(suffix) = self.invalid_suffix { + self.invalid_suffix = None; + return Some(Err(suffix)); + } + + None + } + + /// Advance the iterator, returning everything left as a value + 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; + // 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 { + self.invalid_suffix = None; + return Some(suffix); + } + + None + } +} + +impl<'s> Iterator for ShortFlags<'s> { + type Item = Result<char, &'s OsStr>; + + fn next(&mut self) -> Option<Self::Item> { + self.next_flag() + } +} + +fn split_nonutf8_once(b: &OsStr) -> (&str, Option<&OsStr>) { + match b.try_str() { + Ok(s) => (s, None), + Err(err) => { + // 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)) + } + } +} + +fn is_number(arg: &str) -> bool { + // Return true if this looks like an integer or a float where it's all + // digits plus an optional single dot after some digits. + // + // For floats allow forms such as `1.`, `1.2`, `1.2e10`, etc. + let mut seen_dot = false; + let mut position_of_e = None; + for (i, c) in arg.as_bytes().iter().enumerate() { + match c { + // Digits are always valid + b'0'..=b'9' => {} + + // Allow a `.`, but only one, only if it comes before an + // optional exponent, and only if it's not the first character. + b'.' if !seen_dot && position_of_e.is_none() && i > 0 => seen_dot = true, + + // Allow an exponent `e` but only at most one after the first + // character. + b'e' if position_of_e.is_none() && i > 0 => position_of_e = Some(i), + + _ => return false, + } + } + + // Disallow `-1e` which isn't a valid float since it doesn't actually have + // an exponent. + match position_of_e { + Some(i) => i != arg.len() - 1, + None => true, + } +} |