From 0c8a465e930f3aabfda39540e9eae22425c1cfe2 Mon Sep 17 00:00:00 2001 From: Fenrir Date: Fri, 12 Aug 2016 15:59:22 -0700 Subject: [PATCH] Add Path and OsString APIs --- src/ascii.rs | 554 ++++++++ src/ffi/mod.rs | 3 + src/ffi/os_str.rs | 576 ++++++++ src/lib.rs | 23 +- src/path.rs | 3378 +++++++++++++++++++++++++++++++++++++++++++++ src/sys/mod.rs | 25 + src/sys/wtf8.rs | 1199 ++++++++++++++++ 7 files changed, 5754 insertions(+), 4 deletions(-) create mode 100644 src/ascii.rs create mode 100644 src/ffi/mod.rs create mode 100644 src/ffi/os_str.rs create mode 100644 src/path.rs create mode 100644 src/sys/mod.rs create mode 100644 src/sys/wtf8.rs diff --git a/src/ascii.rs b/src/ascii.rs new file mode 100644 index 0000000..93f447b --- /dev/null +++ b/src/ascii.rs @@ -0,0 +1,554 @@ +// Copyright 2013-2014 The Rust Project Developers. See the COPYRIGHT +// file at the top-level directory of this distribution and at +// http://rust-lang.org/COPYRIGHT. +// +// Licensed under the Apache License, Version 2.0 or the MIT license +// , at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Operations on ASCII strings and characters. + +use core::mem; +use core::ops::Range; +use collections::{String, Vec}; + +/// Extension methods for ASCII-subset only operations on string slices. +/// +/// Be aware that operations on seemingly non-ASCII characters can sometimes +/// have unexpected results. Consider this example: +/// +/// ``` +/// use std::ascii::AsciiExt; +/// +/// assert_eq!("café".to_ascii_uppercase(), "CAFÉ"); +/// assert_eq!("café".to_ascii_uppercase(), "CAFé"); +/// ``` +/// +/// In the first example, the lowercased string is represented `"cafe\u{301}"` +/// (the last character is an acute accent [combining character]). Unlike the +/// other characters in the string, the combining character will not get mapped +/// to an uppercase variant, resulting in `"CAFE\u{301}"`. In the second +/// example, the lowercased string is represented `"caf\u{e9}"` (the last +/// character is a single Unicode character representing an 'e' with an acute +/// accent). Since the last character is defined outside the scope of ASCII, +/// it will not get mapped to an uppercase variant, resulting in `"CAF\u{e9}"`. +/// +/// [combining character]: https://en.wikipedia.org/wiki/Combining_character +pub trait AsciiExt { + /// Container type for copied ASCII characters. + type Owned; + + /// Checks if the value is within the ASCII range. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let ascii = 'a'; + /// let utf8 = '❤'; + /// + /// assert!(ascii.is_ascii()); + /// assert!(!utf8.is_ascii()); + /// ``` + fn is_ascii(&self) -> bool; + + /// Makes a copy of the string in ASCII upper case. + /// + /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z', + /// but non-ASCII letters are unchanged. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let ascii = 'a'; + /// let utf8 = '❤'; + /// + /// assert_eq!('A', ascii.to_ascii_uppercase()); + /// assert_eq!('❤', utf8.to_ascii_uppercase()); + /// ``` + fn to_ascii_uppercase(&self) -> Self::Owned; + + /// Makes a copy of the string in ASCII lower case. + /// + /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z', + /// but non-ASCII letters are unchanged. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let ascii = 'A'; + /// let utf8 = '❤'; + /// + /// assert_eq!('a', ascii.to_ascii_lowercase()); + /// assert_eq!('❤', utf8.to_ascii_lowercase()); + /// ``` + fn to_ascii_lowercase(&self) -> Self::Owned; + + /// Checks that two strings are an ASCII case-insensitive match. + /// + /// Same as `to_ascii_lowercase(a) == to_ascii_lowercase(b)`, + /// but without allocating and copying temporary strings. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let ascii1 = 'A'; + /// let ascii2 = 'a'; + /// let ascii3 = 'A'; + /// let ascii4 = 'z'; + /// + /// assert!(ascii1.eq_ignore_ascii_case(&ascii2)); + /// assert!(ascii1.eq_ignore_ascii_case(&ascii3)); + /// assert!(!ascii1.eq_ignore_ascii_case(&ascii4)); + /// ``` + fn eq_ignore_ascii_case(&self, other: &Self) -> bool; + + /// Converts this type to its ASCII upper case equivalent in-place. + /// + /// See `to_ascii_uppercase` for more information. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let mut ascii = 'a'; + /// + /// ascii.make_ascii_uppercase(); + /// + /// assert_eq!('A', ascii); + /// ``` + fn make_ascii_uppercase(&mut self); + + /// Converts this type to its ASCII lower case equivalent in-place. + /// + /// See `to_ascii_lowercase` for more information. + /// + /// # Examples + /// + /// ``` + /// use std::ascii::AsciiExt; + /// + /// let mut ascii = 'A'; + /// + /// ascii.make_ascii_lowercase(); + /// + /// assert_eq!('a', ascii); + /// ``` + fn make_ascii_lowercase(&mut self); +} + +impl AsciiExt for str { + type Owned = String; + + #[inline] + fn is_ascii(&self) -> bool { + self.bytes().all(|b| b.is_ascii()) + } + + #[inline] + 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) } + } + + #[inline] + fn to_ascii_lowercase(&self) -> String { + let mut bytes = self.as_bytes().to_vec(); + bytes.make_ascii_lowercase(); + // make_ascii_uppercase() preserves the UTF-8 invariant. + unsafe { String::from_utf8_unchecked(bytes) } + } + + #[inline] + fn eq_ignore_ascii_case(&self, other: &str) -> bool { + self.as_bytes().eq_ignore_ascii_case(other.as_bytes()) + } + + fn make_ascii_uppercase(&mut self) { + let me: &mut [u8] = unsafe { mem::transmute(self) }; + me.make_ascii_uppercase() + } + + fn make_ascii_lowercase(&mut self) { + let me: &mut [u8] = unsafe { mem::transmute(self) }; + me.make_ascii_lowercase() + } +} + +impl AsciiExt for [u8] { + type Owned = Vec; + #[inline] + fn is_ascii(&self) -> bool { + self.iter().all(|b| b.is_ascii()) + } + + #[inline] + fn to_ascii_uppercase(&self) -> Vec { + let mut me = self.to_vec(); + me.make_ascii_uppercase(); + return me + } + + #[inline] + fn to_ascii_lowercase(&self) -> Vec { + let mut me = self.to_vec(); + me.make_ascii_lowercase(); + return me + } + + #[inline] + fn eq_ignore_ascii_case(&self, other: &[u8]) -> bool { + self.len() == other.len() && + self.iter().zip(other).all(|(a, b)| { + a.eq_ignore_ascii_case(b) + }) + } + + fn make_ascii_uppercase(&mut self) { + for byte in self { + byte.make_ascii_uppercase(); + } + } + + fn make_ascii_lowercase(&mut self) { + for byte in self { + byte.make_ascii_lowercase(); + } + } +} + +impl AsciiExt for u8 { + type Owned = u8; + #[inline] + fn is_ascii(&self) -> bool { *self & 128 == 0 } + #[inline] + fn to_ascii_uppercase(&self) -> u8 { ASCII_UPPERCASE_MAP[*self as usize] } + #[inline] + fn to_ascii_lowercase(&self) -> u8 { ASCII_LOWERCASE_MAP[*self as usize] } + #[inline] + fn eq_ignore_ascii_case(&self, other: &u8) -> bool { + self.to_ascii_lowercase() == other.to_ascii_lowercase() + } + #[inline] + fn make_ascii_uppercase(&mut self) { *self = self.to_ascii_uppercase(); } + #[inline] + fn make_ascii_lowercase(&mut self) { *self = self.to_ascii_lowercase(); } +} + +impl AsciiExt for char { + type Owned = char; + #[inline] + fn is_ascii(&self) -> bool { + *self as u32 <= 0x7F + } + + #[inline] + fn to_ascii_uppercase(&self) -> char { + if self.is_ascii() { + (*self as u8).to_ascii_uppercase() as char + } else { + *self + } + } + + #[inline] + fn to_ascii_lowercase(&self) -> char { + if self.is_ascii() { + (*self as u8).to_ascii_lowercase() as char + } else { + *self + } + } + + #[inline] + fn eq_ignore_ascii_case(&self, other: &char) -> bool { + self.to_ascii_lowercase() == other.to_ascii_lowercase() + } + + #[inline] + fn make_ascii_uppercase(&mut self) { *self = self.to_ascii_uppercase(); } + #[inline] + fn make_ascii_lowercase(&mut self) { *self = self.to_ascii_lowercase(); } +} + +/// An iterator over the escaped version of a byte, constructed via +/// `std::ascii::escape_default`. +pub struct EscapeDefault { + range: Range, + data: [u8; 4], +} + +/// Returns an iterator that produces an escaped version of a `u8`. +/// +/// The default is chosen with a bias toward producing literals that are +/// legal in a variety of languages, including C++11 and similar C-family +/// languages. The exact rules are: +/// +/// - Tab, CR and LF are escaped as '\t', '\r' and '\n' respectively. +/// - Single-quote, double-quote and backslash chars are backslash-escaped. +/// - Any other chars in the range [0x20,0x7e] are not escaped. +/// - Any other chars are given hex escapes of the form '\xNN'. +/// - Unicode escapes are never generated by this function. +/// +/// # Examples +/// +/// ``` +/// use std::ascii; +/// +/// let escaped = ascii::escape_default(b'0').next().unwrap(); +/// assert_eq!(b'0', escaped); +/// +/// let mut escaped = ascii::escape_default(b'\t'); +/// +/// assert_eq!(b'\\', escaped.next().unwrap()); +/// assert_eq!(b't', escaped.next().unwrap()); +/// ``` +pub fn escape_default(c: u8) -> EscapeDefault { + let (data, len) = match c { + b'\t' => ([b'\\', b't', 0, 0], 2), + b'\r' => ([b'\\', b'r', 0, 0], 2), + b'\n' => ([b'\\', b'n', 0, 0], 2), + b'\\' => ([b'\\', b'\\', 0, 0], 2), + b'\'' => ([b'\\', b'\'', 0, 0], 2), + b'"' => ([b'\\', b'"', 0, 0], 2), + b'\x20' ... b'\x7e' => ([c, 0, 0, 0], 1), + _ => ([b'\\', b'x', hexify(c >> 4), hexify(c & 0xf)], 4), + }; + + return EscapeDefault { range: (0.. len), data: data }; + + fn hexify(b: u8) -> u8 { + match b { + 0 ... 9 => b'0' + b, + _ => b'a' + b - 10, + } + } +} + +impl Iterator for EscapeDefault { + type Item = u8; + fn next(&mut self) -> Option { self.range.next().map(|i| self.data[i]) } + fn size_hint(&self) -> (usize, Option) { self.range.size_hint() } +} +impl DoubleEndedIterator for EscapeDefault { + fn next_back(&mut self) -> Option { + self.range.next_back().map(|i| self.data[i]) + } +} +impl ExactSizeIterator for EscapeDefault {} + +static ASCII_LOWERCASE_MAP: [u8; 256] = [ + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, + 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, + 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, + 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, + b' ', b'!', b'"', b'#', b'$', b'%', b'&', b'\'', + b'(', b')', b'*', b'+', b',', b'-', b'.', b'/', + b'0', b'1', b'2', b'3', b'4', b'5', b'6', b'7', + b'8', b'9', b':', b';', b'<', b'=', b'>', b'?', + b'@', + + b'a', b'b', b'c', b'd', b'e', b'f', b'g', + b'h', b'i', b'j', b'k', b'l', b'm', b'n', b'o', + b'p', b'q', b'r', b's', b't', b'u', b'v', b'w', + b'x', b'y', b'z', + + b'[', b'\\', b']', b'^', b'_', + b'`', b'a', b'b', b'c', b'd', b'e', b'f', b'g', + b'h', b'i', b'j', b'k', b'l', b'm', b'n', b'o', + b'p', b'q', b'r', b's', b't', b'u', b'v', b'w', + b'x', b'y', b'z', b'{', b'|', b'}', b'~', 0x7f, + 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, + 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, + 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, + 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, + 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, + 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, + 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, + 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, + 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, + 0xd0, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, + 0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0xdd, 0xde, 0xdf, + 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, + 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, + 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, + 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, +]; + +static ASCII_UPPERCASE_MAP: [u8; 256] = [ + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, + 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, + 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, + 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, + b' ', b'!', b'"', b'#', b'$', b'%', b'&', b'\'', + b'(', b')', b'*', b'+', b',', b'-', b'.', b'/', + b'0', b'1', b'2', b'3', b'4', b'5', b'6', b'7', + b'8', b'9', b':', b';', b'<', b'=', b'>', b'?', + b'@', b'A', b'B', b'C', b'D', b'E', b'F', b'G', + b'H', b'I', b'J', b'K', b'L', b'M', b'N', b'O', + b'P', b'Q', b'R', b'S', b'T', b'U', b'V', b'W', + b'X', b'Y', b'Z', b'[', b'\\', b']', b'^', b'_', + b'`', + + b'A', b'B', b'C', b'D', b'E', b'F', b'G', + b'H', b'I', b'J', b'K', b'L', b'M', b'N', b'O', + b'P', b'Q', b'R', b'S', b'T', b'U', b'V', b'W', + b'X', b'Y', b'Z', + + b'{', b'|', b'}', b'~', 0x7f, + 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, + 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, + 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, + 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, + 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, + 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, + 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, + 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, + 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, + 0xd0, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, + 0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0xdd, 0xde, 0xdf, + 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, + 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, + 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, + 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, +]; + + +#[cfg(test)] +mod tests { + use super::*; + use rustc_unicode::char::from_u32; + use collections::string::ToString; + + #[test] + fn test_is_ascii() { + assert!(b"".is_ascii()); + assert!(b"banana\0\x7F".is_ascii()); + assert!(b"banana\0\x7F".iter().all(|b| b.is_ascii())); + assert!(!b"Vi\xe1\xbb\x87t Nam".is_ascii()); + assert!(!b"Vi\xe1\xbb\x87t Nam".iter().all(|b| b.is_ascii())); + assert!(!b"\xe1\xbb\x87".iter().any(|b| b.is_ascii())); + + assert!("".is_ascii()); + assert!("banana\0\u{7F}".is_ascii()); + assert!("banana\0\u{7F}".chars().all(|c| c.is_ascii())); + assert!(!"ประเทศไทย中华Việt Nam".chars().all(|c| c.is_ascii())); + + // NOTE: This test fails for some reason. + assert!(!"ประเทศไทย中华ệ ".chars().any(|c| c.is_ascii())); + } + + #[test] + fn test_to_ascii_uppercase() { + assert_eq!("url()URL()uRl()ürl".to_ascii_uppercase(), "URL()URL()URL()üRL"); + assert_eq!("hıKß".to_ascii_uppercase(), "HıKß"); + + for i in 0..501 { + let upper = if 'a' as u32 <= i && i <= 'z' as u32 { i + 'A' as u32 - 'a' as u32 } + else { i }; + assert_eq!((from_u32(i).unwrap()).to_string().to_ascii_uppercase(), + (from_u32(upper).unwrap()).to_string()); + } + } + + #[test] + fn test_to_ascii_lowercase() { + assert_eq!("url()URL()uRl()Ürl".to_ascii_lowercase(), "url()url()url()Ürl"); + // Dotted capital I, Kelvin sign, Sharp S. + assert_eq!("HİKß".to_ascii_lowercase(), "hİKß"); + + for i in 0..501 { + let lower = if 'A' as u32 <= i && i <= 'Z' as u32 { i + 'a' as u32 - 'A' as u32 } + else { i }; + assert_eq!((from_u32(i).unwrap()).to_string().to_ascii_lowercase(), + (from_u32(lower).unwrap()).to_string()); + } + } + + #[test] + fn test_make_ascii_lower_case() { + macro_rules! test { + ($from: expr, $to: expr) => { + { + let mut x = $from; + x.make_ascii_lowercase(); + assert_eq!(x, $to); + } + } + } + test!(b'A', b'a'); + test!(b'a', b'a'); + test!(b'!', b'!'); + test!('A', 'a'); + test!('À', 'À'); + test!('a', 'a'); + test!('!', '!'); + test!(b"H\xc3\x89".to_vec(), b"h\xc3\x89"); + test!("HİKß".to_string(), "hİKß"); + } + + + #[test] + fn test_make_ascii_upper_case() { + macro_rules! test { + ($from: expr, $to: expr) => { + { + let mut x = $from; + x.make_ascii_uppercase(); + assert_eq!(x, $to); + } + } + } + test!(b'a', b'A'); + test!(b'A', b'A'); + test!(b'!', b'!'); + test!('a', 'A'); + test!('à', 'à'); + test!('A', 'A'); + test!('!', '!'); + test!(b"h\xc3\xa9".to_vec(), b"H\xc3\xa9"); + test!("hıKß".to_string(), "HıKß"); + + let mut x = "Hello".to_string(); + x[..3].make_ascii_uppercase(); // Test IndexMut on String. + assert_eq!(x, "HELlo") + } + + #[test] + fn test_eq_ignore_ascii_case() { + assert!("url()URL()uRl()Ürl".eq_ignore_ascii_case("url()url()url()Ürl")); + assert!(!"Ürl".eq_ignore_ascii_case("ürl")); + // Dotted capital I, Kelvin sign, Sharp S. + assert!("HİKß".eq_ignore_ascii_case("hİKß")); + assert!(!"İ".eq_ignore_ascii_case("i")); + assert!(!"K".eq_ignore_ascii_case("k")); + assert!(!"ß".eq_ignore_ascii_case("s")); + + for i in 0..501 { + let lower = if 'A' as u32 <= i && i <= 'Z' as u32 { i + 'a' as u32 - 'A' as u32 } + else { i }; + assert!((from_u32(i).unwrap()).to_string().eq_ignore_ascii_case( + &from_u32(lower).unwrap().to_string())); + } + } + + #[test] + fn inference_works() { + let x = "a".to_string(); + x.eq_ignore_ascii_case("A"); + } +} diff --git a/src/ffi/mod.rs b/src/ffi/mod.rs new file mode 100644 index 0000000..3a2d5f1 --- /dev/null +++ b/src/ffi/mod.rs @@ -0,0 +1,3 @@ +pub use self::os_str::{OsString, OsStr}; + +mod os_str; diff --git a/src/ffi/os_str.rs b/src/ffi/os_str.rs new file mode 100644 index 0000000..d062d55 --- /dev/null +++ b/src/ffi/os_str.rs @@ -0,0 +1,576 @@ +// Copyright 2015 The Rust Project Developers. See the COPYRIGHT +// file at the top-level directory of this distribution and at +// http://rust-lang.org/COPYRIGHT. +// +// Licensed under the Apache License, Version 2.0 or the MIT license +// , at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +use collections::borrow::{Borrow, Cow, ToOwned}; +use core::fmt::{self, Debug}; +use core::mem; +use collections::String; +use core::ops; +use core::cmp; +use core::hash::{Hash, Hasher}; +use collections::Vec; + +use sys::wtf8::{Wtf8, Wtf8Buf}; +use sys::{AsInner, IntoInner, FromInner}; + +pub use sys::wtf8::EncodeWide; + +/// A type that can represent owned, mutable platform-native strings, but is +/// cheaply inter-convertible with Rust strings. +/// +/// The need for this type arises from the fact that: +/// +/// * On Unix systems, strings are often arbitrary sequences of non-zero +/// bytes, in many cases interpreted as UTF-8. +/// +/// * On Windows, strings are often arbitrary sequences of non-zero 16-bit +/// values, interpreted as UTF-16 when it is valid to do so. +/// +/// * In Rust, strings are always valid UTF-8, but may contain zeros. +/// +/// `OsString` and `OsStr` bridge this gap by simultaneously representing Rust +/// and platform-native string values, and in particular allowing a Rust string +/// to be converted into an "OS" string with no cost. +#[derive(Clone)] +pub struct OsString { + inner: Wtf8Buf +} + +/// Slices into OS strings (see `OsString`). +pub struct OsStr { + inner: Wtf8 +} + +impl OsString { + /// Constructs a new empty `OsString`. + pub fn new() -> OsString { + OsString { inner: Wtf8Buf::from_string(String::new()) } + } + + fn _from_bytes(vec: Vec) -> Option { + String::from_utf8(vec).ok().map(OsString::from) + } + + /// Converts to an `OsStr` slice. + pub fn as_os_str(&self) -> &OsStr { + self + } + + /// Converts the `OsString` into a `String` if it contains valid Unicode data. + /// + /// On failure, ownership of the original `OsString` is returned. + pub fn into_string(self) -> Result { + self.inner.into_string().map_err(|buf| OsString { inner: buf} ) + } + + /// Extends the string with the given `&OsStr` slice. + pub fn push>(&mut self, s: T) { + self.inner.push_wtf8(&s.as_ref().inner) + } + + /// Creates a new `OsString` with the given capacity. + /// + /// The string will be able to hold exactly `capacity` lenth units of other + /// OS strings without reallocating. If `capacity` is 0, the string will not + /// allocate. + /// + /// See main `OsString` documentation information about encoding. + pub fn with_capacity(capacity: usize) -> OsString { + OsString { + inner: Wtf8Buf::with_capacity(capacity) + } + } + + /// Truncates the `OsString` to zero length. + pub fn clear(&mut self) { + self.inner.clear() + } + + /// Returns the capacity this `OsString` can hold without reallocating. + /// + /// See `OsString` introduction for information about encoding. + pub fn capacity(&self) -> usize { + self.inner.capacity() + } + + /// Reserves capacity for at least `additional` more capacity to be inserted + /// in the given `OsString`. + /// + /// The collection may reserve more space to avoid frequent reallocations. + pub fn reserve(&mut self, additional: usize) { + self.inner.reserve(additional) + } + + /// Reserves the minimum capacity for exactly `additional` more capacity to + /// be inserted in the given `OsString`. 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. + pub fn reserve_exact(&mut self, additional: usize) { + self.inner.reserve_exact(additional) + } + + /// Creates an `OsString` from a potentially ill-formed UTF-16 slice of + /// 16-bit code units. + /// + /// This is lossless: calling `.encode_wide()` on the resulting string + /// will always return the original code units. + /// + /// NOTE: This function was copied from the windows implementation of OsStringExt + pub fn from_wide(wide: &[u16]) -> OsString { + OsString { inner: Wtf8Buf::from_wide(wide) } + } +} + +impl From for OsString { + fn from(s: String) -> OsString { + OsString { inner: Wtf8Buf::from_string(s) } + } +} + +impl<'a, T: ?Sized + AsRef> From<&'a T> for OsString { + fn from(s: &'a T) -> OsString { + s.as_ref().to_os_string() + } +} + +impl ops::Index for OsString { + type Output = OsStr; + + #[inline] + fn index(&self, _index: ops::RangeFull) -> &OsStr { + OsStr::from_inner(self.inner.as_slice()) + } +} + +impl ops::Deref for OsString { + type Target = OsStr; + + #[inline] + fn deref(&self) -> &OsStr { + &self[..] + } +} + +impl Default for OsString { + #[inline] + fn default() -> OsString { + OsString::new() + } +} + +impl Debug for OsString { + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + fmt::Debug::fmt(&**self, formatter) + } +} + +impl PartialEq for OsString { + fn eq(&self, other: &OsString) -> bool { + &**self == &**other + } +} + +impl PartialEq for OsString { + fn eq(&self, other: &str) -> bool { + &**self == other + } +} + +impl PartialEq for str { + fn eq(&self, other: &OsString) -> bool { + &**other == self + } +} + +impl Eq for OsString {} + +impl PartialOrd for OsString { + #[inline] + fn partial_cmp(&self, other: &OsString) -> Option { + (&**self).partial_cmp(&**other) + } + #[inline] + fn lt(&self, other: &OsString) -> bool { &**self < &**other } + #[inline] + fn le(&self, other: &OsString) -> bool { &**self <= &**other } + #[inline] + fn gt(&self, other: &OsString) -> bool { &**self > &**other } + #[inline] + fn ge(&self, other: &OsString) -> bool { &**self >= &**other } +} + +impl PartialOrd for OsString { + #[inline] + fn partial_cmp(&self, other: &str) -> Option { + (&**self).partial_cmp(other) + } +} + +impl Ord for OsString { + #[inline] + fn cmp(&self, other: &OsString) -> cmp::Ordering { + (&**self).cmp(&**other) + } +} + +impl Hash for OsString { + #[inline] + fn hash(&self, state: &mut H) { + (&**self).hash(state) + } +} + +impl OsStr { + /// Coerces into an `OsStr` slice. + pub fn new + ?Sized>(s: &S) -> &OsStr { + s.as_ref() + } + + fn from_inner(inner: &Wtf8) -> &OsStr { + unsafe { mem::transmute(inner) } + } + + /// Yields a `&str` slice if the `OsStr` is valid Unicode. + /// + /// This conversion may entail doing a check for UTF-8 validity. + pub fn to_str(&self) -> Option<&str> { + self.inner.as_str() + } + + /// Converts an `OsStr` to a `Cow`. + /// + /// Any non-Unicode sequences are replaced with U+FFFD REPLACEMENT CHARACTER. + pub fn to_string_lossy(&self) -> Cow { + self.inner.to_string_lossy() + } + + /// Copies the slice into an owned `OsString`. + pub fn to_os_string(&self) -> OsString { + let mut buf = Wtf8Buf::with_capacity(self.inner.len()); + buf.push_wtf8(&self.inner); + OsString { inner: buf } + } + + /// Checks whether the `OsStr` is empty. + pub fn is_empty(&self) -> bool { + self.inner.is_empty() + } + + /// Returns the length of this `OsStr`. + /// + /// Note that this does **not** return the number of bytes in this string + /// as, for example, OS strings on Windows are encoded as a list of `u16` + /// rather than a list of bytes. This number is simply useful for passing to + /// other methods like `OsString::with_capacity` to avoid reallocations. + /// + /// See `OsStr` introduction for more information about encoding. + pub fn len(&self) -> usize { + self.inner.len() + } + + /// Gets the underlying byte representation. + /// + /// Note: it is *crucial* that this API is private, to avoid + /// revealing the internal, platform-specific encodings. + fn bytes(&self) -> &[u8] { + unsafe { mem::transmute(&self.inner) } + } + + /// Re-encodes an `OsStr` as a wide character sequence, + /// i.e. potentially ill-formed UTF-16. + /// This is lossless. Note that the encoding does not include a final + /// null. + /// + /// NOTE: This function was copied from the windows implementation of OsStrExt + pub fn encode_wide(&self) -> EncodeWide { + self.inner.encode_wide() + } + +} + +impl<'a> Default for &'a OsStr { + #[inline] + fn default() -> &'a OsStr { + OsStr::new("") + } +} + +impl PartialEq for OsStr { + fn eq(&self, other: &OsStr) -> bool { + self.bytes().eq(other.bytes()) + } +} + +impl PartialEq for OsStr { + fn eq(&self, other: &str) -> bool { + *self == *OsStr::new(other) + } +} + +impl PartialEq for str { + fn eq(&self, other: &OsStr) -> bool { + *other == *OsStr::new(self) + } +} + +impl Eq for OsStr {} + +impl PartialOrd for OsStr { + #[inline] + fn partial_cmp(&self, other: &OsStr) -> Option { + self.bytes().partial_cmp(other.bytes()) + } + #[inline] + fn lt(&self, other: &OsStr) -> bool { self.bytes().lt(other.bytes()) } + #[inline] + fn le(&self, other: &OsStr) -> bool { self.bytes().le(other.bytes()) } + #[inline] + fn gt(&self, other: &OsStr) -> bool { self.bytes().gt(other.bytes()) } + #[inline] + fn ge(&self, other: &OsStr) -> bool { self.bytes().ge(other.bytes()) } +} + +impl PartialOrd for OsStr { + #[inline] + fn partial_cmp(&self, other: &str) -> Option { + self.partial_cmp(OsStr::new(other)) + } +} + +// FIXME (#19470): cannot provide PartialOrd for str until we +// have more flexible coherence rules. + +impl Ord for OsStr { + #[inline] + fn cmp(&self, other: &OsStr) -> cmp::Ordering { self.bytes().cmp(other.bytes()) } +} + +macro_rules! impl_cmp { + ($lhs:ty, $rhs: ty) => { + impl<'a, 'b> PartialEq<$rhs> for $lhs { + #[inline] + fn eq(&self, other: &$rhs) -> bool { ::eq(self, other) } + } + + impl<'a, 'b> PartialEq<$lhs> for $rhs { + #[inline] + fn eq(&self, other: &$lhs) -> bool { ::eq(self, other) } + } + + impl<'a, 'b> PartialOrd<$rhs> for $lhs { + #[inline] + fn partial_cmp(&self, other: &$rhs) -> Option { + ::partial_cmp(self, other) + } + } + + impl<'a, 'b> PartialOrd<$lhs> for $rhs { + #[inline] + fn partial_cmp(&self, other: &$lhs) -> Option { + ::partial_cmp(self, other) + } + } + } +} + +impl_cmp!(OsString, OsStr); +impl_cmp!(OsString, &'a OsStr); +impl_cmp!(Cow<'a, OsStr>, OsStr); +impl_cmp!(Cow<'a, OsStr>, &'b OsStr); +impl_cmp!(Cow<'a, OsStr>, OsString); + +impl Hash for OsStr { + #[inline] + fn hash(&self, state: &mut H) { + self.bytes().hash(state) + } +} + +impl Debug for OsStr { + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + self.inner.fmt(formatter) + } +} + +impl Borrow for OsString { + fn borrow(&self) -> &OsStr { &self[..] } +} + +impl ToOwned for OsStr { + type Owned = OsString; + fn to_owned(&self) -> OsString { self.to_os_string() } +} + +impl AsRef for OsStr { + fn as_ref(&self) -> &OsStr { + self + } +} + +impl AsRef for OsString { + fn as_ref(&self) -> &OsStr { + self + } +} + +impl AsRef for str { + fn as_ref(&self) -> &OsStr { + OsStr::from_inner(Wtf8::from_str(self)) + } +} + +impl AsRef for String { + fn as_ref(&self) -> &OsStr { + (&**self).as_ref() + } +} + +impl FromInner for OsString { + fn from_inner(buf: Wtf8Buf) -> OsString { + OsString { inner: buf } + } +} + +impl IntoInner for OsString { + fn into_inner(self) -> Wtf8Buf { + self.inner + } +} + +impl AsInner for OsStr { + fn as_inner(&self) -> &Wtf8 { + &self.inner + } +} + +#[cfg(test)] +mod tests { + use super::*; + use sys::{AsInner, IntoInner}; + + #[test] + fn test_os_string_with_capacity() { + let os_string = OsString::with_capacity(0); + assert_eq!(0, os_string.inner.capacity()); + + let os_string = OsString::with_capacity(10); + assert_eq!(10, os_string.inner.capacity()); + + let mut os_string = OsString::with_capacity(0); + os_string.push("abc"); + assert!(os_string.inner.capacity() >= 3); + } + + #[test] + fn test_os_string_clear() { + let mut os_string = OsString::from("abc"); + assert_eq!(3, os_string.inner.len()); + + os_string.clear(); + assert_eq!(&os_string, ""); + assert_eq!(0, os_string.inner.len()); + } + + #[test] + fn test_os_string_capacity() { + let os_string = OsString::with_capacity(0); + assert_eq!(0, os_string.capacity()); + + let os_string = OsString::with_capacity(10); + assert_eq!(10, os_string.capacity()); + + let mut os_string = OsString::with_capacity(0); + os_string.push("abc"); + assert!(os_string.capacity() >= 3); + } + + #[test] + fn test_os_string_reserve() { + let mut os_string = OsString::new(); + assert_eq!(os_string.capacity(), 0); + + os_string.reserve(2); + assert!(os_string.capacity() >= 2); + + for _ in 0..16 { + os_string.push("a"); + } + + assert!(os_string.capacity() >= 16); + os_string.reserve(16); + assert!(os_string.capacity() >= 32); + + os_string.push("a"); + + os_string.reserve(16); + assert!(os_string.capacity() >= 33) + } + + #[test] + fn test_os_string_reserve_exact() { + let mut os_string = OsString::new(); + assert_eq!(os_string.capacity(), 0); + + os_string.reserve_exact(2); + assert!(os_string.capacity() >= 2); + + for _ in 0..16 { + os_string.push("a"); + } + + assert!(os_string.capacity() >= 16); + os_string.reserve_exact(16); + assert!(os_string.capacity() >= 32); + + os_string.push("a"); + + os_string.reserve_exact(16); + assert!(os_string.capacity() >= 33) + } + + #[test] + fn test_os_string_default() { + let os_string: OsString = Default::default(); + assert_eq!("", &os_string); + } + + #[test] + fn test_os_str_is_empty() { + let mut os_string = OsString::new(); + assert!(os_string.is_empty()); + + os_string.push("abc"); + assert!(!os_string.is_empty()); + + os_string.clear(); + assert!(os_string.is_empty()); + } + + #[test] + fn test_os_str_len() { + let mut os_string = OsString::new(); + assert_eq!(0, os_string.len()); + + os_string.push("abc"); + assert_eq!(3, os_string.len()); + + os_string.clear(); + assert_eq!(0, os_string.len()); + } + + #[test] + fn test_os_str_default() { + let os_str: &OsStr = Default::default(); + assert_eq!("", os_str); + } +} diff --git a/src/lib.rs b/src/lib.rs index dc225c6..7fa8597 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -1,20 +1,35 @@ -#![feature(alloc, collections, lang_items)] +#![feature(alloc)] +#![feature(collections)] +#![feature(char_escape_debug)] +#![feature(lang_items)] +#![feature(question_mark)] +#![feature(slice_patterns)] +#![feature(str_internals)] +#![feature(unicode)] + #![no_std] + #![crate_type = "rlib"] #![crate_name = "ctru"] -extern crate ctru_sys as libctru; - extern crate alloc; extern crate collections; +extern crate rustc_unicode; + +extern crate ctru_sys as libctru; pub mod console; pub mod srv; pub mod gfx; +pub mod services; pub mod sdmc; -pub mod services; +pub mod ascii; +pub mod ffi; pub mod panic; +pub mod path; + +mod sys; pub use srv::Srv; pub use gfx::Gfx; diff --git a/src/path.rs b/src/path.rs new file mode 100644 index 0000000..3fb6357 --- /dev/null +++ b/src/path.rs @@ -0,0 +1,3378 @@ +// Copyright 2015 The Rust Project Developers. See the COPYRIGHT +// file at the top-level directory of this distribution and at +// http://rust-lang.org/COPYRIGHT. +// +// Licensed under the Apache License, Version 2.0 or the MIT license +// , at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Cross-platform path manipulation. +//! +//! This module provides two types, `PathBuf` and `Path` (akin to `String` and +//! `str`), for working with paths abstractly. These types are thin wrappers +//! around `OsString` and `OsStr` respectively, meaning that they work directly +//! on strings according to the local platform's path syntax. +//! +//! ## Simple usage +//! +//! Path manipulation includes both parsing components from slices and building +//! new owned paths. +//! +//! To parse a path, you can create a `Path` slice from a `str` +//! slice and start asking questions: +//! +//! ```rust +//! use std::path::Path; +//! +//! let path = Path::new("/tmp/foo/bar.txt"); +//! let file = path.file_name(); +//! let extension = path.extension(); +//! let parent_dir = path.parent(); +//! ``` +//! +//! To build or modify paths, use `PathBuf`: +//! +//! ```rust +//! use std::path::PathBuf; +//! +//! let mut path = PathBuf::from("c:\\"); +//! path.push("windows"); +//! path.push("system32"); +//! path.set_extension("dll"); +//! ``` +//! +//! ## Path components and normalization +//! +//! The path APIs are built around the notion of "components", which roughly +//! correspond to the substrings between path separators (`/` and, on Windows, +//! `\`). The APIs for path parsing are largely specified in terms of the path's +//! components, so it's important to clearly understand how those are +//! determined. +//! +//! A path can always be reconstructed into an *equivalent* path by +//! putting together its components via `push`. Syntactically, the +//! paths may differ by the normalization described below. +//! +//! ### Component types +//! +//! Components come in several types: +//! +//! * Normal components are the default: standard references to files or +//! directories. The path `a/b` has two normal components, `a` and `b`. +//! +//! * Current directory components represent the `.` character. For example, +//! `./a` has a current directory component and a normal component `a`. +//! +//! * The root directory component represents a separator that designates +//! starting from root. For example, `/a/b` has a root directory component +//! followed by normal components `a` and `b`. +//! +//! On Windows, an additional component type comes into play: +//! +//! * Prefix components, of which there is a large variety. For example, `C:` +//! and `\\server\share` are prefixes. The path `C:windows` has a prefix +//! component `C:` and a normal component `windows`; the path `C:\windows` has a +//! prefix component `C:`, a root directory component, and a normal component +//! `windows`. +//! +//! ### Normalization +//! +//! Aside from splitting on the separator(s), there is a small amount of +//! "normalization": +//! +//! * Repeated separators are ignored: `a/b` and `a//b` both have components `a` +//! and `b`. +//! +//! * Occurrences of `.` are normalized away, *except* if they are at +//! the beginning of the path (in which case they are often meaningful +//! in terms of path searching). So, for example, `a/./b`, `a/b/`, +//! `/a/b/.` and `a/b` all have components `a` and `b`, but `./a/b` +//! has a leading current directory component. +//! +//! No other normalization takes place by default. In particular, +//! `a/c` and `a/b/../c` are distinct, to account for the possibility +//! that `b` is a symbolic link (so its parent isn't `a`). Further +//! normalization is possible to build on top of the components APIs, +//! and will be included in this library in the near future. + +use ascii::*; +use collections::borrow::{Borrow, ToOwned, Cow}; +use core::cmp; +//use error::Error; +use core::fmt; +//use fs; +use core::hash::{Hash, Hasher}; +//use io; +use core::iter; +use core::mem; +use core::ops::{self, Deref}; +use collections::String; +use collections::Vec; + +use ffi::{OsStr, OsString}; + +use self::platform::{is_sep_byte, is_verbatim_sep, MAIN_SEP_STR, parse_prefix}; + +//////////////////////////////////////////////////////////////////////////////// +// GENERAL NOTES +//////////////////////////////////////////////////////////////////////////////// +// +// Parsing in this module is done by directly transmuting OsStr to [u8] slices, +// taking advantage of the fact that OsStr always encodes ASCII characters +// as-is. Eventually, this transmutation should be replaced by direct uses of +// OsStr APIs for parsing, but it will take a while for those to become +// available. + +//////////////////////////////////////////////////////////////////////////////// +// Platform-specific definitions +//////////////////////////////////////////////////////////////////////////////// + +// The following modules give the most basic tools for parsing paths on various +// platforms. The bulk of the code is devoted to parsing prefixes on Windows. + +#[cfg(unix)] +mod platform { + use super::Prefix; + use ffi::OsStr; + + #[inline] + pub fn is_sep_byte(b: u8) -> bool { + b == b'/' + } + + #[inline] + pub fn is_verbatim_sep(b: u8) -> bool { + b == b'/' + } + + pub fn parse_prefix(_: &OsStr) -> Option { + None + } + + pub const MAIN_SEP_STR: &'static str = "/"; + pub const MAIN_SEP: char = '/'; +} + +#[cfg(windows)] +mod platform { + use ascii::*; + + use super::{os_str_as_u8_slice, u8_slice_as_os_str, Prefix}; + use ffi::OsStr; + + #[inline] + pub fn is_sep_byte(b: u8) -> bool { + b == b'/' || b == b'\\' + } + + #[inline] + pub fn is_verbatim_sep(b: u8) -> bool { + b == b'\\' + } + + pub fn parse_prefix<'a>(path: &'a OsStr) -> Option { + use super::Prefix::*; + unsafe { + // The unsafety here stems from converting between &OsStr and &[u8] + // and back. This is safe to do because (1) we only look at ASCII + // contents of the encoding and (2) new &OsStr values are produced + // only from ASCII-bounded slices of existing &OsStr values. + let mut path = os_str_as_u8_slice(path); + + if path.starts_with(br"\\") { + // \\ + path = &path[2..]; + if path.starts_with(br"?\") { + // \\?\ + path = &path[2..]; + if path.starts_with(br"UNC\") { + // \\?\UNC\server\share + path = &path[4..]; + let (server, share) = match parse_two_comps(path, is_verbatim_sep) { + Some((server, share)) => + (u8_slice_as_os_str(server), u8_slice_as_os_str(share)), + None => (u8_slice_as_os_str(path), u8_slice_as_os_str(&[])), + }; + return Some(VerbatimUNC(server, share)); + } else { + // \\?\path + let idx = path.iter().position(|&b| b == b'\\'); + if idx == Some(2) && path[1] == b':' { + let c = path[0]; + if c.is_ascii() && (c as char).is_alphabetic() { + // \\?\C:\ path + return Some(VerbatimDisk(c.to_ascii_uppercase())); + } + } + let slice = &path[..idx.unwrap_or(path.len())]; + return Some(Verbatim(u8_slice_as_os_str(slice))); + } + } else if path.starts_with(b".\\") { + // \\.\path + path = &path[2..]; + let pos = path.iter().position(|&b| b == b'\\'); + let slice = &path[..pos.unwrap_or(path.len())]; + return Some(DeviceNS(u8_slice_as_os_str(slice))); + } + match parse_two_comps(path, is_sep_byte) { + Some((server, share)) if !server.is_empty() && !share.is_empty() => { + // \\server\share + return Some(UNC(u8_slice_as_os_str(server), u8_slice_as_os_str(share))); + } + _ => (), + } + } else if path.get(1) == Some(& b':') { + // C: + let c = path[0]; + if c.is_ascii() && (c as char).is_alphabetic() { + return Some(Disk(c.to_ascii_uppercase())); + } + } + return None; + } + + fn parse_two_comps(mut path: &[u8], f: fn(u8) -> bool) -> Option<(&[u8], &[u8])> { + let first = match path.iter().position(|x| f(*x)) { + None => return None, + Some(x) => &path[..x], + }; + path = &path[(first.len() + 1)..]; + let idx = path.iter().position(|x| f(*x)); + let second = &path[..idx.unwrap_or(path.len())]; + Some((first, second)) + } + } + + pub const MAIN_SEP_STR: &'static str = "\\"; + pub const MAIN_SEP: char = '\\'; +} + +//////////////////////////////////////////////////////////////////////////////// +// Windows Prefixes +//////////////////////////////////////////////////////////////////////////////// + +/// Path prefixes (Windows only). +/// +/// Windows uses a variety of path styles, including references to drive +/// volumes (like `C:`), network shared folders (like `\\server\share`) and +/// others. In addition, some path prefixes are "verbatim", in which case +/// `/` is *not* treated as a separator and essentially no normalization is +/// performed. +#[derive(Copy, Clone, Debug, Hash, PartialOrd, Ord, PartialEq, Eq)] +pub enum Prefix<'a> { + /// Prefix `\\?\`, together with the given component immediately following it. + Verbatim(&'a OsStr), + /// Prefix `\\?\UNC\`, with the "server" and "share" components following it. + VerbatimUNC( + &'a OsStr, + &'a OsStr, + ), + + /// Prefix like `\\?\C:\`, for the given drive letter + VerbatimDisk(u8), + + /// Prefix `\\.\`, together with the given component immediately following it. + DeviceNS(&'a OsStr), + + /// Prefix `\\server\share`, with the given "server" and "share" components. + UNC( + &'a OsStr, + &'a OsStr, + ), + + /// Prefix `C:` for the given disk drive. + Disk(u8), +} + +impl<'a> Prefix<'a> { + #[inline] + fn len(&self) -> usize { + use self::Prefix::*; + fn os_str_len(s: &OsStr) -> usize { + os_str_as_u8_slice(s).len() + } + match *self { + Verbatim(x) => 4 + os_str_len(x), + VerbatimUNC(x, y) => { + 8 + os_str_len(x) + + if os_str_len(y) > 0 { + 1 + os_str_len(y) + } else { + 0 + } + }, + VerbatimDisk(_) => 6, + UNC(x, y) => { + 2 + os_str_len(x) + + if os_str_len(y) > 0 { + 1 + os_str_len(y) + } else { + 0 + } + }, + DeviceNS(x) => 4 + os_str_len(x), + Disk(_) => 2, + } + + } + + /// Determines if the prefix is verbatim, i.e. begins with `\\?\`. + #[inline] + pub fn is_verbatim(&self) -> bool { + use self::Prefix::*; + match *self { + Verbatim(_) | VerbatimDisk(_) | VerbatimUNC(_, _) => true, + _ => false, + } + } + + #[inline] + fn is_drive(&self) -> bool { + match *self { + Prefix::Disk(_) => true, + _ => false, + } + } + + #[inline] + fn has_implicit_root(&self) -> bool { + !self.is_drive() + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Exposed parsing helpers +//////////////////////////////////////////////////////////////////////////////// + +/// Determines whether the character is one of the permitted path +/// separators for the current platform. +/// +/// # Examples +/// +/// ``` +/// use std::path; +/// +/// assert!(path::is_separator('/')); +/// assert!(!path::is_separator('❤')); +/// ``` +pub fn is_separator(c: char) -> bool { + c.is_ascii() && is_sep_byte(c as u8) +} + +/// The primary separator for the current platform +pub const MAIN_SEPARATOR: char = platform::MAIN_SEP; + +//////////////////////////////////////////////////////////////////////////////// +// Misc helpers +//////////////////////////////////////////////////////////////////////////////// + +// Iterate through `iter` while it matches `prefix`; return `None` if `prefix` +// is not a prefix of `iter`, otherwise return `Some(iter_after_prefix)` giving +// `iter` after having exhausted `prefix`. +fn iter_after(mut iter: I, mut prefix: J) -> Option + where I: Iterator + Clone, + J: Iterator, + A: PartialEq +{ + loop { + let mut iter_next = iter.clone(); + match (iter_next.next(), prefix.next()) { + (Some(ref x), Some(ref y)) if x == y => (), + (Some(_), Some(_)) => return None, + (Some(_), None) => return Some(iter), + (None, None) => return Some(iter), + (None, Some(_)) => return None, + } + iter = iter_next; + } +} + +// See note at the top of this module to understand why these are used: +fn os_str_as_u8_slice(s: &OsStr) -> &[u8] { + unsafe { mem::transmute(s) } +} +unsafe fn u8_slice_as_os_str(s: &[u8]) -> &OsStr { + mem::transmute(s) +} + +//////////////////////////////////////////////////////////////////////////////// +// Cross-platform, iterator-independent parsing +//////////////////////////////////////////////////////////////////////////////// + +/// Says whether the first byte after the prefix is a separator. +fn has_physical_root(s: &[u8], prefix: Option) -> bool { + let path = if let Some(p) = prefix { + &s[p.len()..] + } else { + s + }; + !path.is_empty() && is_sep_byte(path[0]) +} + +// basic workhorse for splitting stem and extension +fn split_file_at_dot(file: &OsStr) -> (Option<&OsStr>, Option<&OsStr>) { + unsafe { + if os_str_as_u8_slice(file) == b".." { + return (Some(file), None); + } + + // The unsafety here stems from converting between &OsStr and &[u8] + // and back. This is safe to do because (1) we only look at ASCII + // contents of the encoding and (2) new &OsStr values are produced + // only from ASCII-bounded slices of existing &OsStr values. + + let mut iter = os_str_as_u8_slice(file).rsplitn(2, |b| *b == b'.'); + let after = iter.next(); + let before = iter.next(); + if before == Some(b"") { + (Some(file), None) + } else { + (before.map(|s| u8_slice_as_os_str(s)), + after.map(|s| u8_slice_as_os_str(s))) + } + } +} + +//////////////////////////////////////////////////////////////////////////////// +// The core iterators +//////////////////////////////////////////////////////////////////////////////// + +/// Component parsing works by a double-ended state machine; the cursors at the +/// front and back of the path each keep track of what parts of the path have +/// been consumed so far. +/// +/// Going front to back, a path is made up of a prefix, a starting +/// directory component, and a body (of normal components) +#[derive(Copy, Clone, PartialEq, PartialOrd, Debug)] +enum State { + Prefix = 0, // c: + StartDir = 1, // / or . or nothing + Body = 2, // foo/bar/baz + Done = 3, +} + +/// A Windows path prefix, e.g. `C:` or `\\server\share`. +/// +/// Does not occur on Unix. +#[derive(Copy, Clone, Eq, Debug)] +pub struct PrefixComponent<'a> { + /// The prefix as an unparsed `OsStr` slice. + raw: &'a OsStr, + + /// The parsed prefix data. + parsed: Prefix<'a>, +} + +impl<'a> PrefixComponent<'a> { + /// The parsed prefix data. + pub fn kind(&self) -> Prefix<'a> { + self.parsed + } + + /// The raw `OsStr` slice for this prefix. + pub fn as_os_str(&self) -> &'a OsStr { + self.raw + } +} + +impl<'a> cmp::PartialEq for PrefixComponent<'a> { + fn eq(&self, other: &PrefixComponent<'a>) -> bool { + cmp::PartialEq::eq(&self.parsed, &other.parsed) + } +} + +impl<'a> cmp::PartialOrd for PrefixComponent<'a> { + fn partial_cmp(&self, other: &PrefixComponent<'a>) -> Option { + cmp::PartialOrd::partial_cmp(&self.parsed, &other.parsed) + } +} + +impl<'a> cmp::Ord for PrefixComponent<'a> { + fn cmp(&self, other: &PrefixComponent<'a>) -> cmp::Ordering { + cmp::Ord::cmp(&self.parsed, &other.parsed) + } +} + +impl<'a> Hash for PrefixComponent<'a> { + fn hash(&self, h: &mut H) { + self.parsed.hash(h); + } +} + +/// A single component of a path. +/// +/// See the module documentation for an in-depth explanation of components and +/// their role in the API. +/// +/// This `enum` is created from iterating over the [`path::Components`] +/// `struct`. +/// +/// # Examples +/// +/// ```rust +/// use std::path::{Component, Path}; +/// +/// let path = Path::new("/tmp/foo/bar.txt"); +/// let components = path.components().collect::>(); +/// assert_eq!(&components, &[ +/// Component::RootDir, +/// Component::Normal("tmp".as_ref()), +/// Component::Normal("foo".as_ref()), +/// Component::Normal("bar.txt".as_ref()), +/// ]); +/// ``` +/// +/// [`path::Components`]: struct.Components.html +#[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Hash, Debug)] +pub enum Component<'a> { + /// A Windows path prefix, e.g. `C:` or `\\server\share`. + /// + /// Does not occur on Unix. + Prefix( + PrefixComponent<'a> + ), + + /// The root directory component, appears after any prefix and before anything else + RootDir, + + /// A reference to the current directory, i.e. `.` + CurDir, + + /// A reference to the parent directory, i.e. `..` + ParentDir, + + /// A normal component, i.e. `a` and `b` in `a/b` + Normal(&'a OsStr), +} + +impl<'a> Component<'a> { + /// Extracts the underlying `OsStr` slice + pub fn as_os_str(self) -> &'a OsStr { + match self { + Component::Prefix(p) => p.as_os_str(), + Component::RootDir => OsStr::new(MAIN_SEP_STR), + Component::CurDir => OsStr::new("."), + Component::ParentDir => OsStr::new(".."), + Component::Normal(path) => path, + } + } +} + +impl<'a> AsRef for Component<'a> { + fn as_ref(&self) -> &OsStr { + self.as_os_str() + } +} + +/// The core iterator giving the components of a path. +/// +/// See the module documentation for an in-depth explanation of components and +/// their role in the API. +/// +/// This `struct` is created by the [`path::Path::components`] method. +/// +/// # Examples +/// +/// ``` +/// use std::path::Path; +/// +/// let path = Path::new("/tmp/foo/bar.txt"); +/// +/// for component in path.components() { +/// println!("{:?}", component); +/// } +/// ``` +/// +/// [`path::Path::components`]: struct.Path.html#method.components +#[derive(Clone)] +pub struct Components<'a> { + // The path left to parse components from + path: &'a [u8], + + // The prefix as it was originally parsed, if any + prefix: Option>, + + // true if path *physically* has a root separator; for most Windows + // prefixes, it may have a "logical" rootseparator for the purposes of + // normalization, e.g. \\server\share == \\server\share\. + has_physical_root: bool, + + // The iterator is double-ended, and these two states keep track of what has + // been produced from either end + front: State, + back: State, +} + +/// An iterator over the components of a path, as `OsStr` slices. +#[derive(Clone)] +pub struct Iter<'a> { + inner: Components<'a>, +} + +impl<'a> Components<'a> { + // how long is the prefix, if any? + #[inline] + fn prefix_len(&self) -> usize { + self.prefix.as_ref().map(Prefix::len).unwrap_or(0) + } + + #[inline] + fn prefix_verbatim(&self) -> bool { + self.prefix.as_ref().map(Prefix::is_verbatim).unwrap_or(false) + } + + /// how much of the prefix is left from the point of view of iteration? + #[inline] + fn prefix_remaining(&self) -> usize { + if self.front == State::Prefix { + self.prefix_len() + } else { + 0 + } + } + + // Given the iteration so far, how much of the pre-State::Body path is left? + #[inline] + fn len_before_body(&self) -> usize { + let root = if self.front <= State::StartDir && self.has_physical_root { + 1 + } else { + 0 + }; + let cur_dir = if self.front <= State::StartDir && self.include_cur_dir() { + 1 + } else { + 0 + }; + self.prefix_remaining() + root + cur_dir + } + + // is the iteration complete? + #[inline] + fn finished(&self) -> bool { + self.front == State::Done || self.back == State::Done || self.front > self.back + } + + #[inline] + fn is_sep_byte(&self, b: u8) -> bool { + if self.prefix_verbatim() { + is_verbatim_sep(b) + } else { + is_sep_byte(b) + } + } + + /// Extracts a slice corresponding to the portion of the path remaining for iteration. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let mut components = Path::new("/tmp/foo/bar.txt").components(); + /// components.next(); + /// components.next(); + /// + /// assert_eq!(Path::new("foo/bar.txt"), components.as_path()); + /// ``` + pub fn as_path(&self) -> &'a Path { + let mut comps = self.clone(); + if comps.front == State::Body { + comps.trim_left(); + } + if comps.back == State::Body { + comps.trim_right(); + } + unsafe { Path::from_u8_slice(comps.path) } + } + + /// Is the *original* path rooted? + fn has_root(&self) -> bool { + if self.has_physical_root { + return true; + } + if let Some(p) = self.prefix { + if p.has_implicit_root() { + return true; + } + } + false + } + + /// Should the normalized path include a leading . ? + fn include_cur_dir(&self) -> bool { + if self.has_root() { + return false; + } + let mut iter = self.path[self.prefix_len()..].iter(); + match (iter.next(), iter.next()) { + (Some(&b'.'), None) => true, + (Some(&b'.'), Some(&b)) => self.is_sep_byte(b), + _ => false, + } + } + + // parse a given byte sequence into the corresponding path component + fn parse_single_component<'b>(&self, comp: &'b [u8]) -> Option> { + match comp { + b"." if self.prefix_verbatim() => Some(Component::CurDir), + b"." => None, // . components are normalized away, except at + // the beginning of a path, which is treated + // separately via `include_cur_dir` + b".." => Some(Component::ParentDir), + b"" => None, + _ => Some(Component::Normal(unsafe { u8_slice_as_os_str(comp) })), + } + } + + // parse a component from the left, saying how many bytes to consume to + // remove the component + fn parse_next_component(&self) -> (usize, Option>) { + debug_assert!(self.front == State::Body); + let (extra, comp) = match self.path.iter().position(|b| self.is_sep_byte(*b)) { + None => (0, self.path), + Some(i) => (1, &self.path[..i]), + }; + (comp.len() + extra, self.parse_single_component(comp)) + } + + // parse a component from the right, saying how many bytes to consume to + // remove the component + fn parse_next_component_back(&self) -> (usize, Option>) { + debug_assert!(self.back == State::Body); + let start = self.len_before_body(); + let (extra, comp) = match self.path[start..].iter().rposition(|b| self.is_sep_byte(*b)) { + None => (0, &self.path[start..]), + Some(i) => (1, &self.path[start + i + 1..]), + }; + (comp.len() + extra, self.parse_single_component(comp)) + } + + // trim away repeated separators (i.e. empty components) on the left + fn trim_left(&mut self) { + while !self.path.is_empty() { + let (size, comp) = self.parse_next_component(); + if comp.is_some() { + return; + } else { + self.path = &self.path[size..]; + } + } + } + + // trim away repeated separators (i.e. empty components) on the right + fn trim_right(&mut self) { + while self.path.len() > self.len_before_body() { + let (size, comp) = self.parse_next_component_back(); + if comp.is_some() { + return; + } else { + self.path = &self.path[..self.path.len() - size]; + } + } + } +} + +impl<'a> AsRef for Components<'a> { + fn as_ref(&self) -> &Path { + self.as_path() + } +} + +impl<'a> AsRef for Components<'a> { + fn as_ref(&self) -> &OsStr { + self.as_path().as_os_str() + } +} + +impl<'a> Iter<'a> { + /// Extracts a slice corresponding to the portion of the path remaining for iteration. + pub fn as_path(&self) -> &'a Path { + self.inner.as_path() + } +} + +impl<'a> AsRef for Iter<'a> { + fn as_ref(&self) -> &Path { + self.as_path() + } +} + +impl<'a> AsRef for Iter<'a> { + fn as_ref(&self) -> &OsStr { + self.as_path().as_os_str() + } +} + +impl<'a> Iterator for Iter<'a> { + type Item = &'a OsStr; + + fn next(&mut self) -> Option<&'a OsStr> { + self.inner.next().map(Component::as_os_str) + } +} + +impl<'a> DoubleEndedIterator for Iter<'a> { + fn next_back(&mut self) -> Option<&'a OsStr> { + self.inner.next_back().map(Component::as_os_str) + } +} + +impl<'a> Iterator for Components<'a> { + type Item = Component<'a>; + + fn next(&mut self) -> Option> { + while !self.finished() { + match self.front { + State::Prefix if self.prefix_len() > 0 => { + self.front = State::StartDir; + debug_assert!(self.prefix_len() <= self.path.len()); + let raw = &self.path[..self.prefix_len()]; + self.path = &self.path[self.prefix_len()..]; + return Some(Component::Prefix(PrefixComponent { + raw: unsafe { u8_slice_as_os_str(raw) }, + parsed: self.prefix.unwrap(), + })); + } + State::Prefix => { + self.front = State::StartDir; + } + State::StartDir => { + self.front = State::Body; + if self.has_physical_root { + debug_assert!(!self.path.is_empty()); + self.path = &self.path[1..]; + return Some(Component::RootDir); + } else if let Some(p) = self.prefix { + if p.has_implicit_root() && !p.is_verbatim() { + return Some(Component::RootDir); + } + } else if self.include_cur_dir() { + debug_assert!(!self.path.is_empty()); + self.path = &self.path[1..]; + return Some(Component::CurDir); + } + } + State::Body if !self.path.is_empty() => { + let (size, comp) = self.parse_next_component(); + self.path = &self.path[size..]; + if comp.is_some() { + return comp; + } + } + State::Body => { + self.front = State::Done; + } + State::Done => unreachable!(), + } + } + None + } +} + +impl<'a> DoubleEndedIterator for Components<'a> { + fn next_back(&mut self) -> Option> { + while !self.finished() { + match self.back { + State::Body if self.path.len() > self.len_before_body() => { + let (size, comp) = self.parse_next_component_back(); + self.path = &self.path[..self.path.len() - size]; + if comp.is_some() { + return comp; + } + } + State::Body => { + self.back = State::StartDir; + } + State::StartDir => { + self.back = State::Prefix; + if self.has_physical_root { + self.path = &self.path[..self.path.len() - 1]; + return Some(Component::RootDir); + } else if let Some(p) = self.prefix { + if p.has_implicit_root() && !p.is_verbatim() { + return Some(Component::RootDir); + } + } else if self.include_cur_dir() { + self.path = &self.path[..self.path.len() - 1]; + return Some(Component::CurDir); + } + } + State::Prefix if self.prefix_len() > 0 => { + self.back = State::Done; + return Some(Component::Prefix(PrefixComponent { + raw: unsafe { u8_slice_as_os_str(self.path) }, + parsed: self.prefix.unwrap(), + })); + } + State::Prefix => { + self.back = State::Done; + return None; + } + State::Done => unreachable!(), + } + } + None + } +} + +impl<'a> cmp::PartialEq for Components<'a> { + fn eq(&self, other: &Components<'a>) -> bool { + Iterator::eq(self.clone(), other.clone()) + } +} + +impl<'a> cmp::Eq for Components<'a> {} + +impl<'a> cmp::PartialOrd for Components<'a> { + fn partial_cmp(&self, other: &Components<'a>) -> Option { + Iterator::partial_cmp(self.clone(), other.clone()) + } +} + +impl<'a> cmp::Ord for Components<'a> { + fn cmp(&self, other: &Components<'a>) -> cmp::Ordering { + Iterator::cmp(self.clone(), other.clone()) + } +} + +//////////////////////////////////////////////////////////////////////////////// +// Basic types and traits +//////////////////////////////////////////////////////////////////////////////// + +/// An owned, mutable path (akin to `String`). +/// +/// This type provides methods like `push` and `set_extension` that mutate the +/// path in place. It also implements `Deref` to `Path`, meaning that all +/// methods on `Path` slices are available on `PathBuf` values as well. +/// +/// More details about the overall approach can be found in +/// the module documentation. +/// +/// # Examples +/// +/// ``` +/// use std::path::PathBuf; +/// +/// let mut path = PathBuf::from("c:\\"); +/// path.push("windows"); +/// path.push("system32"); +/// path.set_extension("dll"); +/// ``` +#[derive(Clone)] +pub struct PathBuf { + inner: OsString, +} + +impl PathBuf { + fn as_mut_vec(&mut self) -> &mut Vec { + unsafe { &mut *(self as *mut PathBuf as *mut Vec) } + } + + /// Allocates an empty `PathBuf`. + pub fn new() -> PathBuf { + PathBuf { inner: OsString::new() } + } + + /// Coerces to a `Path` slice. + pub fn as_path(&self) -> &Path { + self + } + + /// Extends `self` with `path`. + /// + /// If `path` is absolute, it replaces the current path. + /// + /// On Windows: + /// + /// * if `path` has a root but no prefix (e.g. `\windows`), it + /// replaces everything except for the prefix (if any) of `self`. + /// * if `path` has a prefix but no root, it replaces `self`. + /// + /// # Examples + /// + /// ``` + /// use std::path::PathBuf; + /// + /// let mut path = PathBuf::new(); + /// path.push("/tmp"); + /// path.push("file.bk"); + /// assert_eq!(path, PathBuf::from("/tmp/file.bk")); + /// + /// // Pushing an absolute path replaces the current path + /// path.push("/etc/passwd"); + /// assert_eq!(path, PathBuf::from("/etc/passwd")); + /// ``` + pub fn push>(&mut self, path: P) { + self._push(path.as_ref()) + } + + fn _push(&mut self, path: &Path) { + // in general, a separator is needed if the rightmost byte is not a separator + let mut need_sep = self.as_mut_vec().last().map(|c| !is_sep_byte(*c)).unwrap_or(false); + + // in the special case of `C:` on Windows, do *not* add a separator + { + let comps = self.components(); + if comps.prefix_len() > 0 && comps.prefix_len() == comps.path.len() && + comps.prefix.unwrap().is_drive() { + need_sep = false + } + } + + // absolute `path` replaces `self` + if path.is_absolute() || path.prefix().is_some() { + self.as_mut_vec().truncate(0); + + // `path` has a root but no prefix, e.g. `\windows` (Windows only) + } else if path.has_root() { + let prefix_len = self.components().prefix_remaining(); + self.as_mut_vec().truncate(prefix_len); + + // `path` is a pure relative path + } else if need_sep { + self.inner.push(MAIN_SEP_STR); + } + + self.inner.push(path); + } + + /// Truncate `self` to `self.parent()`. + /// + /// Returns false and does nothing if `self.file_name()` is `None`. + /// Otherwise, returns `true`. + pub fn pop(&mut self) -> bool { + match self.parent().map(|p| p.as_u8_slice().len()) { + Some(len) => { + self.as_mut_vec().truncate(len); + true + } + None => false, + } + } + + /// Updates `self.file_name()` to `file_name`. + /// + /// If `self.file_name()` was `None`, this is equivalent to pushing + /// `file_name`. + /// + /// # Examples + /// + /// ``` + /// use std::path::PathBuf; + /// + /// let mut buf = PathBuf::from("/"); + /// assert!(buf.file_name() == None); + /// buf.set_file_name("bar"); + /// assert!(buf == PathBuf::from("/bar")); + /// assert!(buf.file_name().is_some()); + /// buf.set_file_name("baz.txt"); + /// assert!(buf == PathBuf::from("/baz.txt")); + /// ``` + pub fn set_file_name>(&mut self, file_name: S) { + self._set_file_name(file_name.as_ref()) + } + + fn _set_file_name(&mut self, file_name: &OsStr) { + if self.file_name().is_some() { + let popped = self.pop(); + debug_assert!(popped); + } + self.push(file_name); + } + + /// Updates `self.extension()` to `extension`. + /// + /// If `self.file_name()` is `None`, does nothing and returns `false`. + /// + /// Otherwise, returns `true`; if `self.extension()` is `None`, the extension + /// is added; otherwise it is replaced. + pub fn set_extension>(&mut self, extension: S) -> bool { + self._set_extension(extension.as_ref()) + } + + fn _set_extension(&mut self, extension: &OsStr) -> bool { + if self.file_name().is_none() { + return false; + } + + let mut stem = match self.file_stem() { + Some(stem) => stem.to_os_string(), + None => OsString::new(), + }; + + if !os_str_as_u8_slice(extension).is_empty() { + stem.push("."); + stem.push(extension); + } + self.set_file_name(&stem); + + true + } + + /// Consumes the `PathBuf`, yielding its internal `OsString` storage. + pub fn into_os_string(self) -> OsString { + self.inner + } +} + +impl<'a, T: ?Sized + AsRef> From<&'a T> for PathBuf { + fn from(s: &'a T) -> PathBuf { + PathBuf::from(s.as_ref().to_os_string()) + } +} + +impl From for PathBuf { + fn from(s: OsString) -> PathBuf { + PathBuf { inner: s } + } +} + +impl From for PathBuf { + fn from(s: String) -> PathBuf { + PathBuf::from(OsString::from(s)) + } +} + +impl> iter::FromIterator

for PathBuf { + fn from_iter>(iter: I) -> PathBuf { + let mut buf = PathBuf::new(); + buf.extend(iter); + buf + } +} + +impl> iter::Extend

for PathBuf { + fn extend>(&mut self, iter: I) { + for p in iter { + self.push(p.as_ref()) + } + } +} + +impl fmt::Debug for PathBuf { + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + fmt::Debug::fmt(&**self, formatter) + } +} + +impl ops::Deref for PathBuf { + type Target = Path; + + fn deref(&self) -> &Path { + Path::new(&self.inner) + } +} + +impl Borrow for PathBuf { + fn borrow(&self) -> &Path { + self.deref() + } +} + +impl<'a> From<&'a Path> for Cow<'a, Path> { + #[inline] + fn from(s: &'a Path) -> Cow<'a, Path> { + Cow::Borrowed(s) + } +} + +impl<'a> From for Cow<'a, Path> { + #[inline] + fn from(s: PathBuf) -> Cow<'a, Path> { + Cow::Owned(s) + } +} + +impl ToOwned for Path { + type Owned = PathBuf; + fn to_owned(&self) -> PathBuf { + self.to_path_buf() + } +} + +impl cmp::PartialEq for PathBuf { + fn eq(&self, other: &PathBuf) -> bool { + self.components() == other.components() + } +} + +impl Hash for PathBuf { + fn hash(&self, h: &mut H) { + self.as_path().hash(h) + } +} + +impl cmp::Eq for PathBuf {} + +impl cmp::PartialOrd for PathBuf { + fn partial_cmp(&self, other: &PathBuf) -> Option { + self.components().partial_cmp(other.components()) + } +} + +impl cmp::Ord for PathBuf { + fn cmp(&self, other: &PathBuf) -> cmp::Ordering { + self.components().cmp(other.components()) + } +} + +impl AsRef for PathBuf { + fn as_ref(&self) -> &OsStr { + &self.inner[..] + } +} + +impl Into for PathBuf { + fn into(self) -> OsString { + self.inner + } +} + +/// A slice of a path (akin to `str`). +/// +/// This type supports a number of operations for inspecting a path, including +/// breaking the path into its components (separated by `/` or `\`, depending on +/// the platform), extracting the file name, determining whether the path is +/// absolute, and so on. More details about the overall approach can be found in +/// the module documentation. +/// +/// This is an *unsized* type, meaning that it must always be used behind a +/// pointer like `&` or `Box`. +/// +/// # Examples +/// +/// ``` +/// use std::path::Path; +/// +/// let path = Path::new("/tmp/foo/bar.txt"); +/// let file = path.file_name(); +/// let extension = path.extension(); +/// let parent_dir = path.parent(); +/// ``` +/// +pub struct Path { + inner: OsStr, +} + +/// An error returned from the `Path::strip_prefix` method indicating that the +/// prefix was not found in `self`. +#[derive(Debug, Clone, PartialEq, Eq)] +pub struct StripPrefixError(()); + +impl Path { + // The following (private!) function allows construction of a path from a u8 + // slice, which is only safe when it is known to follow the OsStr encoding. + unsafe fn from_u8_slice(s: &[u8]) -> &Path { + Path::new(u8_slice_as_os_str(s)) + } + // The following (private!) function reveals the byte encoding used for OsStr. + fn as_u8_slice(&self) -> &[u8] { + os_str_as_u8_slice(&self.inner) + } + + /// Directly wrap a string slice as a `Path` slice. + /// + /// This is a cost-free conversion. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// Path::new("foo.txt"); + /// ``` + /// + /// You can create `Path`s from `String`s, or even other `Path`s: + /// + /// ``` + /// use std::path::Path; + /// + /// let string = String::from("foo.txt"); + /// let from_string = Path::new(&string); + /// let from_path = Path::new(&from_string); + /// assert_eq!(from_string, from_path); + /// ``` + pub fn new + ?Sized>(s: &S) -> &Path { + unsafe { mem::transmute(s.as_ref()) } + } + + /// Yields the underlying `OsStr` slice. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let os_str = Path::new("foo.txt").as_os_str(); + /// assert_eq!(os_str, std::ffi::OsStr::new("foo.txt")); + /// ``` + pub fn as_os_str(&self) -> &OsStr { + &self.inner + } + + /// Yields a `&str` slice if the `Path` is valid unicode. + /// + /// This conversion may entail doing a check for UTF-8 validity. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path_str = Path::new("foo.txt").to_str(); + /// assert_eq!(path_str, Some("foo.txt")); + /// ``` + pub fn to_str(&self) -> Option<&str> { + self.inner.to_str() + } + + /// Converts a `Path` to a `Cow`. + /// + /// Any non-Unicode sequences are replaced with U+FFFD REPLACEMENT CHARACTER. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path_str = Path::new("foo.txt").to_string_lossy(); + /// assert_eq!(path_str, "foo.txt"); + /// ``` + pub fn to_string_lossy(&self) -> Cow { + self.inner.to_string_lossy() + } + + /// Converts a `Path` to an owned `PathBuf`. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path_buf = Path::new("foo.txt").to_path_buf(); + /// assert_eq!(path_buf, std::path::PathBuf::from("foo.txt")); + /// ``` + pub fn to_path_buf(&self) -> PathBuf { + PathBuf::from(self.inner.to_os_string()) + } + + /// A path is *absolute* if it is independent of the current directory. + /// + /// * On Unix, a path is absolute if it starts with the root, so + /// `is_absolute` and `has_root` are equivalent. + /// + /// * On Windows, a path is absolute if it has a prefix and starts with the + /// root: `c:\windows` is absolute, while `c:temp` and `\temp` are not. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// assert!(!Path::new("foo.txt").is_absolute()); + /// ``` + #[allow(deprecated)] + pub fn is_absolute(&self) -> bool { + self.has_root() && (cfg!(unix) || self.prefix().is_some()) + } + + /// A path is *relative* if it is not absolute. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// assert!(Path::new("foo.txt").is_relative()); + /// ``` + pub fn is_relative(&self) -> bool { + !self.is_absolute() + } + + fn prefix(&self) -> Option { + self.components().prefix + } + + /// A path has a root if the body of the path begins with the directory separator. + /// + /// * On Unix, a path has a root if it begins with `/`. + /// + /// * On Windows, a path has a root if it: + /// * has no prefix and begins with a separator, e.g. `\\windows` + /// * has a prefix followed by a separator, e.g. `c:\windows` but not `c:windows` + /// * has any non-disk prefix, e.g. `\\server\share` + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// assert!(Path::new("/etc/passwd").has_root()); + /// ``` + pub fn has_root(&self) -> bool { + self.components().has_root() + } + + /// The path without its final component, if any. + /// + /// Returns `None` if the path terminates in a root or prefix. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("/foo/bar"); + /// let parent = path.parent().unwrap(); + /// assert_eq!(parent, Path::new("/foo")); + /// + /// let grand_parent = parent.parent().unwrap(); + /// assert_eq!(grand_parent, Path::new("/")); + /// assert_eq!(grand_parent.parent(), None); + /// ``` + pub fn parent(&self) -> Option<&Path> { + let mut comps = self.components(); + let comp = comps.next_back(); + comp.and_then(|p| { + match p { + Component::Normal(_) | + Component::CurDir | + Component::ParentDir => Some(comps.as_path()), + _ => None, + } + }) + } + + /// The final component of the path, if it is a normal file. + /// + /// If the path terminates in `..`, `file_name` will return `None`. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// use std::ffi::OsStr; + /// + /// let path = Path::new("foo.txt"); + /// let os_str = OsStr::new("foo.txt"); + /// + /// assert_eq!(Some(os_str), path.file_name()); + /// ``` + /// + /// # Other examples + /// + /// ``` + /// use std::path::Path; + /// use std::ffi::OsStr; + /// + /// assert_eq!(Some(OsStr::new("foo.txt")), Path::new("foo.txt/.").file_name()); + /// assert_eq!(Some(OsStr::new("foo.txt")), Path::new("foo.txt/.//").file_name()); + /// assert_eq!(None, Path::new("foo.txt/..").file_name()); + /// ``` + pub fn file_name(&self) -> Option<&OsStr> { + self.components().next_back().and_then(|p| { + match p { + Component::Normal(p) => Some(p.as_ref()), + _ => None, + } + }) + } + + /// Returns a path that, when joined onto `base`, yields `self`. + /// + /// # Errors + /// + /// If `base` is not a prefix of `self` (i.e. `starts_with` + /// returns `false`), returns `Err`. + pub fn strip_prefix<'a, P: ?Sized>(&'a self, base: &'a P) + -> Result<&'a Path, StripPrefixError> + where P: AsRef + { + self._strip_prefix(base.as_ref()) + } + + fn _strip_prefix<'a>(&'a self, base: &'a Path) + -> Result<&'a Path, StripPrefixError> { + iter_after(self.components(), base.components()) + .map(|c| c.as_path()) + .ok_or(StripPrefixError(())) + } + + /// Determines whether `base` is a prefix of `self`. + /// + /// Only considers whole path components to match. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("/etc/passwd"); + /// + /// assert!(path.starts_with("/etc")); + /// + /// assert!(!path.starts_with("/e")); + /// ``` + pub fn starts_with>(&self, base: P) -> bool { + self._starts_with(base.as_ref()) + } + + fn _starts_with(&self, base: &Path) -> bool { + iter_after(self.components(), base.components()).is_some() + } + + /// Determines whether `child` is a suffix of `self`. + /// + /// Only considers whole path components to match. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("/etc/passwd"); + /// + /// assert!(path.ends_with("passwd")); + /// ``` + pub fn ends_with>(&self, child: P) -> bool { + self._ends_with(child.as_ref()) + } + + fn _ends_with(&self, child: &Path) -> bool { + iter_after(self.components().rev(), child.components().rev()).is_some() + } + + /// Extracts the stem (non-extension) portion of `self.file_name()`. + /// + /// The stem is: + /// + /// * None, if there is no file name; + /// * The entire file name if there is no embedded `.`; + /// * The entire file name if the file name begins with `.` and has no other `.`s within; + /// * Otherwise, the portion of the file name before the final `.` + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("foo.rs"); + /// + /// assert_eq!("foo", path.file_stem().unwrap()); + /// ``` + pub fn file_stem(&self) -> Option<&OsStr> { + self.file_name().map(split_file_at_dot).and_then(|(before, after)| before.or(after)) + } + + /// Extracts the extension of `self.file_name()`, if possible. + /// + /// The extension is: + /// + /// * None, if there is no file name; + /// * None, if there is no embedded `.`; + /// * None, if the file name begins with `.` and has no other `.`s within; + /// * Otherwise, the portion of the file name after the final `.` + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("foo.rs"); + /// + /// assert_eq!("rs", path.extension().unwrap()); + /// ``` + pub fn extension(&self) -> Option<&OsStr> { + self.file_name().map(split_file_at_dot).and_then(|(before, after)| before.and(after)) + } + + /// Creates an owned `PathBuf` with `path` adjoined to `self`. + /// + /// See `PathBuf::push` for more details on what it means to adjoin a path. + /// + /// # Examples + /// + /// ``` + /// use std::path::{Path, PathBuf}; + /// + /// assert_eq!(Path::new("/etc").join("passwd"), PathBuf::from("/etc/passwd")); + /// ``` + pub fn join>(&self, path: P) -> PathBuf { + self._join(path.as_ref()) + } + + fn _join(&self, path: &Path) -> PathBuf { + let mut buf = self.to_path_buf(); + buf.push(path); + buf + } + + /// Creates an owned `PathBuf` like `self` but with the given file name. + /// + /// See `PathBuf::set_file_name` for more details. + /// + /// # Examples + /// + /// ``` + /// use std::path::{Path, PathBuf}; + /// + /// let path = Path::new("/tmp/foo.txt"); + /// assert_eq!(path.with_file_name("bar.txt"), PathBuf::from("/tmp/bar.txt")); + /// ``` + pub fn with_file_name>(&self, file_name: S) -> PathBuf { + self._with_file_name(file_name.as_ref()) + } + + fn _with_file_name(&self, file_name: &OsStr) -> PathBuf { + let mut buf = self.to_path_buf(); + buf.set_file_name(file_name); + buf + } + + /// Creates an owned `PathBuf` like `self` but with the given extension. + /// + /// See `PathBuf::set_extension` for more details. + /// + /// # Examples + /// + /// ``` + /// use std::path::{Path, PathBuf}; + /// + /// let path = Path::new("foo.rs"); + /// assert_eq!(path.with_extension("txt"), PathBuf::from("foo.txt")); + /// ``` + pub fn with_extension>(&self, extension: S) -> PathBuf { + self._with_extension(extension.as_ref()) + } + + fn _with_extension(&self, extension: &OsStr) -> PathBuf { + let mut buf = self.to_path_buf(); + buf.set_extension(extension); + buf + } + + /// Produce an iterator over the components of the path. + /// + /// # Examples + /// + /// ``` + /// use std::path::{Path, Component}; + /// use std::ffi::OsStr; + /// + /// let mut components = Path::new("/tmp/foo.txt").components(); + /// + /// assert_eq!(components.next(), Some(Component::RootDir)); + /// assert_eq!(components.next(), Some(Component::Normal(OsStr::new("tmp")))); + /// assert_eq!(components.next(), Some(Component::Normal(OsStr::new("foo.txt")))); + /// assert_eq!(components.next(), None) + /// ``` + pub fn components(&self) -> Components { + let prefix = parse_prefix(self.as_os_str()); + Components { + path: self.as_u8_slice(), + prefix: prefix, + has_physical_root: has_physical_root(self.as_u8_slice(), prefix), + front: State::Prefix, + back: State::Body, + } + } + + /// Produce an iterator over the path's components viewed as `OsStr` slices. + /// + /// # Examples + /// + /// ``` + /// use std::path::{self, Path}; + /// use std::ffi::OsStr; + /// + /// let mut it = Path::new("/tmp/foo.txt").iter(); + /// assert_eq!(it.next(), Some(OsStr::new(&path::MAIN_SEPARATOR.to_string()))); + /// assert_eq!(it.next(), Some(OsStr::new("tmp"))); + /// assert_eq!(it.next(), Some(OsStr::new("foo.txt"))); + /// assert_eq!(it.next(), None) + /// ``` + pub fn iter(&self) -> Iter { + Iter { inner: self.components() } + } + + /// Returns an object that implements `Display` for safely printing paths + /// that may contain non-Unicode data. + /// + /// # Examples + /// + /// ``` + /// use std::path::Path; + /// + /// let path = Path::new("/tmp/foo.rs"); + /// + /// println!("{}", path.display()); + /// ``` + pub fn display(&self) -> Display { + Display { path: self } + } + + + //NOTE: The following functions rely on filesystem functionality that + //probably have to be implemented in ctru-rs instead of this library, + //and thus are commented out + + /* + /// Query the file system to get information about a file, directory, etc. + /// + /// This function will traverse symbolic links to query information about the + /// destination file. + /// + /// This is an alias to [`fs::metadata`]. + /// + /// [`fs::metadata`]: ../fs/fn.metadata.html + pub fn metadata(&self) -> io::Result { + fs::metadata(self) + } + + /// Query the metadata about a file without following symlinks. + /// + /// This is an alias to [`fs::symlink_metadata`]. + /// + /// [`fs::symlink_metadata`]: ../fs/fn.symlink_metadata.html + + pub fn symlink_metadata(&self) -> io::Result { + fs::symlink_metadata(self) + } + + + /// Returns the canonical form of the path with all intermediate components + /// normalized and symbolic links resolved. + /// + /// This is an alias to [`fs::canonicalize`]. + /// + /// [`fs::canonicalize`]: ../fs/fn.canonicalize.html + + pub fn canonicalize(&self) -> io::Result { + fs::canonicalize(self) + } + + + /// Reads a symbolic link, returning the file that the link points to. + /// + /// This is an alias to [`fs::read_link`]. + /// + /// [`fs::read_link`]: ../fs/fn.read_link.html + + pub fn read_link(&self) -> io::Result { + fs::read_link(self) + } + + + /// Returns an iterator over the entries within a directory. + /// + /// The iterator will yield instances of `io::Result`. New errors may + /// be encountered after an iterator is initially constructed. + /// + /// This is an alias to [`fs::read_dir`]. + /// + /// [`fs::read_dir`]: ../fs/fn.read_dir.html + + pub fn read_dir(&self) -> io::Result { + fs::read_dir(self) + } + + + /// Returns whether the path points at an existing entity. + /// + /// This function will traverse symbolic links to query information about the + /// destination file. In case of broken symbolic links this will return `false`. + /// + /// # Examples + /// + /// ```no_run + /// use std::path::Path; + /// assert_eq!(Path::new("does_not_exist.txt").exists(), false); + /// ``` + + pub fn exists(&self) -> bool { + fs::metadata(self).is_ok() + } + + + /// Returns whether the path is pointing at a regular file. + /// + /// This function will traverse symbolic links to query information about the + /// destination file. In case of broken symbolic links this will return `false`. + /// + /// # Examples + /// + /// ```no_run + /// use std::path::Path; + /// assert_eq!(Path::new("./is_a_directory/").is_file(), false); + /// assert_eq!(Path::new("a_file.txt").is_file(), true); + /// ``` + + pub fn is_file(&self) -> bool { + fs::metadata(self).map(|m| m.is_file()).unwrap_or(false) + } + + + /// Returns whether the path is pointing at a directory. + /// + /// This function will traverse symbolic links to query information about the + /// destination file. In case of broken symbolic links this will return `false`. + /// + /// # Examples + /// + /// ```no_run + /// use std::path::Path; + /// assert_eq!(Path::new("./is_a_directory/").is_dir(), true); + /// assert_eq!(Path::new("a_file.txt").is_dir(), false); + /// ``` + + pub fn is_dir(&self) -> bool { + fs::metadata(self).map(|m| m.is_dir()).unwrap_or(false) + } + */ +} + +impl AsRef for Path { + fn as_ref(&self) -> &OsStr { + &self.inner + } +} + +impl fmt::Debug for Path { + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + self.inner.fmt(formatter) + } +} + +/// Helper struct for safely printing paths with `format!()` and `{}` +pub struct Display<'a> { + path: &'a Path, +} + +impl<'a> fmt::Debug for Display<'a> { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + fmt::Debug::fmt(&self.path.to_string_lossy(), f) + } +} + +impl<'a> fmt::Display for Display<'a> { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + fmt::Display::fmt(&self.path.to_string_lossy(), f) + } +} + +impl cmp::PartialEq for Path { + fn eq(&self, other: &Path) -> bool { + self.components().eq(other.components()) + } +} + +impl Hash for Path { + fn hash(&self, h: &mut H) { + for component in self.components() { + component.hash(h); + } + } +} + +impl cmp::Eq for Path {} + +impl cmp::PartialOrd for Path { + fn partial_cmp(&self, other: &Path) -> Option { + self.components().partial_cmp(other.components()) + } +} + +impl cmp::Ord for Path { + fn cmp(&self, other: &Path) -> cmp::Ordering { + self.components().cmp(other.components()) + } +} + +impl AsRef for Path { + fn as_ref(&self) -> &Path { + self + } +} + +impl AsRef for OsStr { + fn as_ref(&self) -> &Path { + Path::new(self) + } +} + +impl<'a> AsRef for Cow<'a, OsStr> { + fn as_ref(&self) -> &Path { + Path::new(self) + } +} + +impl AsRef for OsString { + fn as_ref(&self) -> &Path { + Path::new(self) + } +} + +impl AsRef for str { + fn as_ref(&self) -> &Path { + Path::new(self) + } +} + +impl AsRef for String { + fn as_ref(&self) -> &Path { + Path::new(self) + } +} + +impl AsRef for PathBuf { + fn as_ref(&self) -> &Path { + self + } +} + +impl<'a> IntoIterator for &'a PathBuf { + type Item = &'a OsStr; + type IntoIter = Iter<'a>; + fn into_iter(self) -> Iter<'a> { self.iter() } +} + +impl<'a> IntoIterator for &'a Path { + type Item = &'a OsStr; + type IntoIter = Iter<'a>; + fn into_iter(self) -> Iter<'a> { self.iter() } +} + +macro_rules! impl_cmp { + ($lhs:ty, $rhs: ty) => { + impl<'a, 'b> PartialEq<$rhs> for $lhs { + #[inline] + fn eq(&self, other: &$rhs) -> bool { ::eq(self, other) } + } + + impl<'a, 'b> PartialEq<$lhs> for $rhs { + #[inline] + fn eq(&self, other: &$lhs) -> bool { ::eq(self, other) } + } + + impl<'a, 'b> PartialOrd<$rhs> for $lhs { + #[inline] + fn partial_cmp(&self, other: &$rhs) -> Option { + ::partial_cmp(self, other) + } + } + + impl<'a, 'b> PartialOrd<$lhs> for $rhs { + #[inline] + fn partial_cmp(&self, other: &$lhs) -> Option { + ::partial_cmp(self, other) + } + } + } +} + +impl_cmp!(PathBuf, Path); +impl_cmp!(PathBuf, &'a Path); +impl_cmp!(Cow<'a, Path>, Path); +impl_cmp!(Cow<'a, Path>, &'b Path); +impl_cmp!(Cow<'a, Path>, PathBuf); + +macro_rules! impl_cmp_os_str { + ($lhs:ty, $rhs: ty) => { + impl<'a, 'b> PartialEq<$rhs> for $lhs { + #[inline] + fn eq(&self, other: &$rhs) -> bool { ::eq(self, other.as_ref()) } + } + + impl<'a, 'b> PartialEq<$lhs> for $rhs { + #[inline] + fn eq(&self, other: &$lhs) -> bool { ::eq(self.as_ref(), other) } + } + + impl<'a, 'b> PartialOrd<$rhs> for $lhs { + #[inline] + fn partial_cmp(&self, other: &$rhs) -> Option { + ::partial_cmp(self, other.as_ref()) + } + } + + impl<'a, 'b> PartialOrd<$lhs> for $rhs { + #[inline] + fn partial_cmp(&self, other: &$lhs) -> Option { + ::partial_cmp(self.as_ref(), other) + } + } + } +} + +impl_cmp_os_str!(PathBuf, OsStr); +impl_cmp_os_str!(PathBuf, &'a OsStr); +impl_cmp_os_str!(PathBuf, Cow<'a, OsStr>); +impl_cmp_os_str!(PathBuf, OsString); +impl_cmp_os_str!(Path, OsStr); +impl_cmp_os_str!(Path, &'a OsStr); +impl_cmp_os_str!(Path, Cow<'a, OsStr>); +impl_cmp_os_str!(Path, OsString); +impl_cmp_os_str!(&'a Path, OsStr); +impl_cmp_os_str!(&'a Path, Cow<'b, OsStr>); +impl_cmp_os_str!(&'a Path, OsString); +impl_cmp_os_str!(Cow<'a, Path>, OsStr); +impl_cmp_os_str!(Cow<'a, Path>, &'b OsStr); +impl_cmp_os_str!(Cow<'a, Path>, OsString); + +/* +impl fmt::Display for StripPrefixError { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + self.description().fmt(f) + } +} +*/ + +/* +impl Error for StripPrefixError { + fn description(&self) -> &str { "prefix not found" } +} +*/ + +#[cfg(test)] +mod tests { + use super::*; + use collections::string::{ToString, String}; + use collections::borrow; + use collections::Vec; + + macro_rules! t( + ($path:expr, iter: $iter:expr) => ( + { + let path = Path::new($path); + + // Forward iteration + let comps = path.iter() + .map(|p| p.to_string_lossy().into_owned()) + .collect::>(); + let exp: &[&str] = &$iter; + let exps = exp.iter().map(|s| s.to_string()).collect::>(); + assert!(comps == exps, "iter: Expected {:?}, found {:?}", + exps, comps); + + // Reverse iteration + let comps = Path::new($path).iter().rev() + .map(|p| p.to_string_lossy().into_owned()) + .collect::>(); + let exps = exps.into_iter().rev().collect::>(); + assert!(comps == exps, "iter().rev(): Expected {:?}, found {:?}", + exps, comps); + } + ); + + ($path:expr, has_root: $has_root:expr, is_absolute: $is_absolute:expr) => ( + { + let path = Path::new($path); + + let act_root = path.has_root(); + assert!(act_root == $has_root, "has_root: Expected {:?}, found {:?}", + $has_root, act_root); + + let act_abs = path.is_absolute(); + assert!(act_abs == $is_absolute, "is_absolute: Expected {:?}, found {:?}", + $is_absolute, act_abs); + } + ); + + ($path:expr, parent: $parent:expr, file_name: $file:expr) => ( + { + let path = Path::new($path); + + let parent = path.parent().map(|p| p.to_str().unwrap()); + let exp_parent: Option<&str> = $parent; + assert!(parent == exp_parent, "parent: Expected {:?}, found {:?}", + exp_parent, parent); + + let file = path.file_name().map(|p| p.to_str().unwrap()); + let exp_file: Option<&str> = $file; + assert!(file == exp_file, "file_name: Expected {:?}, found {:?}", + exp_file, file); + } + ); + + ($path:expr, file_stem: $file_stem:expr, extension: $extension:expr) => ( + { + let path = Path::new($path); + + let stem = path.file_stem().map(|p| p.to_str().unwrap()); + let exp_stem: Option<&str> = $file_stem; + assert!(stem == exp_stem, "file_stem: Expected {:?}, found {:?}", + exp_stem, stem); + + let ext = path.extension().map(|p| p.to_str().unwrap()); + let exp_ext: Option<&str> = $extension; + assert!(ext == exp_ext, "extension: Expected {:?}, found {:?}", + exp_ext, ext); + } + ); + + ($path:expr, iter: $iter:expr, + has_root: $has_root:expr, is_absolute: $is_absolute:expr, + parent: $parent:expr, file_name: $file:expr, + file_stem: $file_stem:expr, extension: $extension:expr) => ( + { + t!($path, iter: $iter); + t!($path, has_root: $has_root, is_absolute: $is_absolute); + t!($path, parent: $parent, file_name: $file); + t!($path, file_stem: $file_stem, extension: $extension); + } + ); + ); + + #[test] + fn into() { + use collections::borrow::Cow; + + let static_path = Path::new("/home/foo"); + let static_cow_path: Cow<'static, Path> = static_path.into(); + let pathbuf = PathBuf::from("/home/foo"); + + { + let path: &Path = &pathbuf; + let borrowed_cow_path: Cow = path.into(); + + assert_eq!(static_cow_path, borrowed_cow_path); + } + + let owned_cow_path: Cow<'static, Path> = pathbuf.into(); + + assert_eq!(static_cow_path, owned_cow_path); + } + + #[test] + #[cfg(unix)] + pub fn test_decompositions_unix() { + t!("", + iter: [], + has_root: false, + is_absolute: false, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("/", + iter: ["/"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("/foo", + iter: ["/", "foo"], + has_root: true, + is_absolute: true, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("/foo/", + iter: ["/", "foo"], + has_root: true, + is_absolute: true, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/bar", + iter: ["foo", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("/foo/bar", + iter: ["/", "foo", "bar"], + has_root: true, + is_absolute: true, + parent: Some("/foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("///foo///", + iter: ["/", "foo"], + has_root: true, + is_absolute: true, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("///foo///bar", + iter: ["/", "foo", "bar"], + has_root: true, + is_absolute: true, + parent: Some("///foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("./.", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("/..", + iter: ["/", ".."], + has_root: true, + is_absolute: true, + parent: Some("/"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("../", + iter: [".."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/.", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/..", + iter: ["foo", ".."], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/./", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/./bar", + iter: ["foo", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("foo/../", + iter: ["foo", ".."], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/../bar", + iter: ["foo", "..", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo/.."), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("./a", + iter: [".", "a"], + has_root: false, + is_absolute: false, + parent: Some("."), + file_name: Some("a"), + file_stem: Some("a"), + extension: None + ); + + t!(".", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("./", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("a/b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a//b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a/./b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a/b/c", + iter: ["a", "b", "c"], + has_root: false, + is_absolute: false, + parent: Some("a/b"), + file_name: Some("c"), + file_stem: Some("c"), + extension: None + ); + + t!(".foo", + iter: [".foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some(".foo"), + file_stem: Some(".foo"), + extension: None + ); + } + + #[test] + #[cfg(windows)] + pub fn test_decompositions_windows() { + t!("", + iter: [], + has_root: false, + is_absolute: false, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("/", + iter: ["\\"], + has_root: true, + is_absolute: false, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\", + iter: ["\\"], + has_root: true, + is_absolute: false, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("c:", + iter: ["c:"], + has_root: false, + is_absolute: false, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("c:\\", + iter: ["c:", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("c:/", + iter: ["c:", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("/foo", + iter: ["\\", "foo"], + has_root: true, + is_absolute: false, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("/foo/", + iter: ["\\", "foo"], + has_root: true, + is_absolute: false, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/bar", + iter: ["foo", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("/foo/bar", + iter: ["\\", "foo", "bar"], + has_root: true, + is_absolute: false, + parent: Some("/foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("///foo///", + iter: ["\\", "foo"], + has_root: true, + is_absolute: false, + parent: Some("/"), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("///foo///bar", + iter: ["\\", "foo", "bar"], + has_root: true, + is_absolute: false, + parent: Some("///foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("./.", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("/..", + iter: ["\\", ".."], + has_root: true, + is_absolute: false, + parent: Some("/"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("../", + iter: [".."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/.", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/..", + iter: ["foo", ".."], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/./", + iter: ["foo"], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: Some("foo"), + file_stem: Some("foo"), + extension: None + ); + + t!("foo/./bar", + iter: ["foo", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("foo/../", + iter: ["foo", ".."], + has_root: false, + is_absolute: false, + parent: Some("foo"), + file_name: None, + file_stem: None, + extension: None + ); + + t!("foo/../bar", + iter: ["foo", "..", "bar"], + has_root: false, + is_absolute: false, + parent: Some("foo/.."), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + t!("./a", + iter: [".", "a"], + has_root: false, + is_absolute: false, + parent: Some("."), + file_name: Some("a"), + file_stem: Some("a"), + extension: None + ); + + t!(".", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("./", + iter: ["."], + has_root: false, + is_absolute: false, + parent: Some(""), + file_name: None, + file_stem: None, + extension: None + ); + + t!("a/b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a//b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a/./b", + iter: ["a", "b"], + has_root: false, + is_absolute: false, + parent: Some("a"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + + t!("a/b/c", + iter: ["a", "b", "c"], + has_root: false, + is_absolute: false, + parent: Some("a/b"), + file_name: Some("c"), + file_stem: Some("c"), + extension: None); + + t!("a\\b\\c", + iter: ["a", "b", "c"], + has_root: false, + is_absolute: false, + parent: Some("a\\b"), + file_name: Some("c"), + file_stem: Some("c"), + extension: None + ); + + t!("\\a", + iter: ["\\", "a"], + has_root: true, + is_absolute: false, + parent: Some("\\"), + file_name: Some("a"), + file_stem: Some("a"), + extension: None + ); + + t!("c:\\foo.txt", + iter: ["c:", "\\", "foo.txt"], + has_root: true, + is_absolute: true, + parent: Some("c:\\"), + file_name: Some("foo.txt"), + file_stem: Some("foo"), + extension: Some("txt") + ); + + t!("\\\\server\\share\\foo.txt", + iter: ["\\\\server\\share", "\\", "foo.txt"], + has_root: true, + is_absolute: true, + parent: Some("\\\\server\\share\\"), + file_name: Some("foo.txt"), + file_stem: Some("foo"), + extension: Some("txt") + ); + + t!("\\\\server\\share", + iter: ["\\\\server\\share", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\server", + iter: ["\\", "server"], + has_root: true, + is_absolute: false, + parent: Some("\\"), + file_name: Some("server"), + file_stem: Some("server"), + extension: None + ); + + t!("\\\\?\\bar\\foo.txt", + iter: ["\\\\?\\bar", "\\", "foo.txt"], + has_root: true, + is_absolute: true, + parent: Some("\\\\?\\bar\\"), + file_name: Some("foo.txt"), + file_stem: Some("foo"), + extension: Some("txt") + ); + + t!("\\\\?\\bar", + iter: ["\\\\?\\bar"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\?\\", + iter: ["\\\\?\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\?\\UNC\\server\\share\\foo.txt", + iter: ["\\\\?\\UNC\\server\\share", "\\", "foo.txt"], + has_root: true, + is_absolute: true, + parent: Some("\\\\?\\UNC\\server\\share\\"), + file_name: Some("foo.txt"), + file_stem: Some("foo"), + extension: Some("txt") + ); + + t!("\\\\?\\UNC\\server", + iter: ["\\\\?\\UNC\\server"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\?\\UNC\\", + iter: ["\\\\?\\UNC\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\?\\C:\\foo.txt", + iter: ["\\\\?\\C:", "\\", "foo.txt"], + has_root: true, + is_absolute: true, + parent: Some("\\\\?\\C:\\"), + file_name: Some("foo.txt"), + file_stem: Some("foo"), + extension: Some("txt") + ); + + + t!("\\\\?\\C:\\", + iter: ["\\\\?\\C:", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\?\\C:", + iter: ["\\\\?\\C:"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\?\\foo/bar", + iter: ["\\\\?\\foo/bar"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\?\\C:/foo", + iter: ["\\\\?\\C:/foo"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\.\\foo\\bar", + iter: ["\\\\.\\foo", "\\", "bar"], + has_root: true, + is_absolute: true, + parent: Some("\\\\.\\foo\\"), + file_name: Some("bar"), + file_stem: Some("bar"), + extension: None + ); + + + t!("\\\\.\\foo", + iter: ["\\\\.\\foo", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\.\\foo/bar", + iter: ["\\\\.\\foo/bar", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + + t!("\\\\.\\foo\\bar/baz", + iter: ["\\\\.\\foo", "\\", "bar", "baz"], + has_root: true, + is_absolute: true, + parent: Some("\\\\.\\foo\\bar"), + file_name: Some("baz"), + file_stem: Some("baz"), + extension: None + ); + + + t!("\\\\.\\", + iter: ["\\\\.\\", "\\"], + has_root: true, + is_absolute: true, + parent: None, + file_name: None, + file_stem: None, + extension: None + ); + + t!("\\\\?\\a\\b\\", + iter: ["\\\\?\\a", "\\", "b"], + has_root: true, + is_absolute: true, + parent: Some("\\\\?\\a\\"), + file_name: Some("b"), + file_stem: Some("b"), + extension: None + ); + } + + #[test] + pub fn test_stem_ext() { + t!("foo", + file_stem: Some("foo"), + extension: None + ); + + t!("foo.", + file_stem: Some("foo"), + extension: Some("") + ); + + t!(".foo", + file_stem: Some(".foo"), + extension: None + ); + + t!("foo.txt", + file_stem: Some("foo"), + extension: Some("txt") + ); + + t!("foo.bar.txt", + file_stem: Some("foo.bar"), + extension: Some("txt") + ); + + t!("foo.bar.", + file_stem: Some("foo.bar"), + extension: Some("") + ); + + t!(".", + file_stem: None, + extension: None + ); + + t!("..", + file_stem: None, + extension: None + ); + + t!("", + file_stem: None, + extension: None + ); + } + + #[test] + pub fn test_push() { + macro_rules! tp( + ($path:expr, $push:expr, $expected:expr) => ( { + let mut actual = PathBuf::from($path); + actual.push($push); + assert!(actual.to_str() == Some($expected), + "pushing {:?} onto {:?}: Expected {:?}, got {:?}", + $push, $path, $expected, actual.to_str().unwrap()); + }); + ); + + if cfg!(unix) { + tp!("", "foo", "foo"); + tp!("foo", "bar", "foo/bar"); + tp!("foo/", "bar", "foo/bar"); + tp!("foo//", "bar", "foo//bar"); + tp!("foo/.", "bar", "foo/./bar"); + tp!("foo./.", "bar", "foo././bar"); + tp!("foo", "", "foo/"); + tp!("foo", ".", "foo/."); + tp!("foo", "..", "foo/.."); + tp!("foo", "/", "/"); + tp!("/foo/bar", "/", "/"); + tp!("/foo/bar", "/baz", "/baz"); + tp!("/foo/bar", "./baz", "/foo/bar/./baz"); + } else { + tp!("", "foo", "foo"); + tp!("foo", "bar", r"foo\bar"); + tp!("foo/", "bar", r"foo/bar"); + tp!(r"foo\", "bar", r"foo\bar"); + tp!("foo//", "bar", r"foo//bar"); + tp!(r"foo\\", "bar", r"foo\\bar"); + tp!("foo/.", "bar", r"foo/.\bar"); + tp!("foo./.", "bar", r"foo./.\bar"); + tp!(r"foo\.", "bar", r"foo\.\bar"); + tp!(r"foo.\.", "bar", r"foo.\.\bar"); + tp!("foo", "", "foo\\"); + tp!("foo", ".", r"foo\."); + tp!("foo", "..", r"foo\.."); + tp!("foo", "/", "/"); + tp!("foo", r"\", r"\"); + tp!("/foo/bar", "/", "/"); + tp!(r"\foo\bar", r"\", r"\"); + tp!("/foo/bar", "/baz", "/baz"); + tp!("/foo/bar", r"\baz", r"\baz"); + tp!("/foo/bar", "./baz", r"/foo/bar\./baz"); + tp!("/foo/bar", r".\baz", r"/foo/bar\.\baz"); + + tp!("c:\\", "windows", "c:\\windows"); + tp!("c:", "windows", "c:windows"); + + tp!("a\\b\\c", "d", "a\\b\\c\\d"); + tp!("\\a\\b\\c", "d", "\\a\\b\\c\\d"); + tp!("a\\b", "c\\d", "a\\b\\c\\d"); + tp!("a\\b", "\\c\\d", "\\c\\d"); + tp!("a\\b", ".", "a\\b\\."); + tp!("a\\b", "..\\c", "a\\b\\..\\c"); + tp!("a\\b", "C:a.txt", "C:a.txt"); + tp!("a\\b", "C:\\a.txt", "C:\\a.txt"); + tp!("C:\\a", "C:\\b.txt", "C:\\b.txt"); + tp!("C:\\a\\b\\c", "C:d", "C:d"); + tp!("C:a\\b\\c", "C:d", "C:d"); + tp!("C:", r"a\b\c", r"C:a\b\c"); + tp!("C:", r"..\a", r"C:..\a"); + tp!("\\\\server\\share\\foo", + "bar", + "\\\\server\\share\\foo\\bar"); + tp!("\\\\server\\share\\foo", "C:baz", "C:baz"); + tp!("\\\\?\\C:\\a\\b", "C:c\\d", "C:c\\d"); + tp!("\\\\?\\C:a\\b", "C:c\\d", "C:c\\d"); + tp!("\\\\?\\C:\\a\\b", "C:\\c\\d", "C:\\c\\d"); + tp!("\\\\?\\foo\\bar", "baz", "\\\\?\\foo\\bar\\baz"); + tp!("\\\\?\\UNC\\server\\share\\foo", + "bar", + "\\\\?\\UNC\\server\\share\\foo\\bar"); + tp!("\\\\?\\UNC\\server\\share", "C:\\a", "C:\\a"); + tp!("\\\\?\\UNC\\server\\share", "C:a", "C:a"); + + // Note: modified from old path API + tp!("\\\\?\\UNC\\server", "foo", "\\\\?\\UNC\\server\\foo"); + + tp!("C:\\a", + "\\\\?\\UNC\\server\\share", + "\\\\?\\UNC\\server\\share"); + tp!("\\\\.\\foo\\bar", "baz", "\\\\.\\foo\\bar\\baz"); + tp!("\\\\.\\foo\\bar", "C:a", "C:a"); + // again, not sure about the following, but I'm assuming \\.\ should be verbatim + tp!("\\\\.\\foo", "..\\bar", "\\\\.\\foo\\..\\bar"); + + tp!("\\\\?\\C:", "foo", "\\\\?\\C:\\foo"); // this is a weird one + } + } + + #[test] + pub fn test_pop() { + macro_rules! tp( + ($path:expr, $expected:expr, $output:expr) => ( { + let mut actual = PathBuf::from($path); + let output = actual.pop(); + assert!(actual.to_str() == Some($expected) && output == $output, + "popping from {:?}: Expected {:?}/{:?}, got {:?}/{:?}", + $path, $expected, $output, + actual.to_str().unwrap(), output); + }); + ); + + tp!("", "", false); + tp!("/", "/", false); + tp!("foo", "", true); + tp!(".", "", true); + tp!("/foo", "/", true); + tp!("/foo/bar", "/foo", true); + tp!("foo/bar", "foo", true); + tp!("foo/.", "", true); + tp!("foo//bar", "foo", true); + + if cfg!(windows) { + tp!("a\\b\\c", "a\\b", true); + tp!("\\a", "\\", true); + tp!("\\", "\\", false); + + tp!("C:\\a\\b", "C:\\a", true); + tp!("C:\\a", "C:\\", true); + tp!("C:\\", "C:\\", false); + tp!("C:a\\b", "C:a", true); + tp!("C:a", "C:", true); + tp!("C:", "C:", false); + tp!("\\\\server\\share\\a\\b", "\\\\server\\share\\a", true); + tp!("\\\\server\\share\\a", "\\\\server\\share\\", true); + tp!("\\\\server\\share", "\\\\server\\share", false); + tp!("\\\\?\\a\\b\\c", "\\\\?\\a\\b", true); + tp!("\\\\?\\a\\b", "\\\\?\\a\\", true); + tp!("\\\\?\\a", "\\\\?\\a", false); + tp!("\\\\?\\C:\\a\\b", "\\\\?\\C:\\a", true); + tp!("\\\\?\\C:\\a", "\\\\?\\C:\\", true); + tp!("\\\\?\\C:\\", "\\\\?\\C:\\", false); + tp!("\\\\?\\UNC\\server\\share\\a\\b", + "\\\\?\\UNC\\server\\share\\a", + true); + tp!("\\\\?\\UNC\\server\\share\\a", + "\\\\?\\UNC\\server\\share\\", + true); + tp!("\\\\?\\UNC\\server\\share", + "\\\\?\\UNC\\server\\share", + false); + tp!("\\\\.\\a\\b\\c", "\\\\.\\a\\b", true); + tp!("\\\\.\\a\\b", "\\\\.\\a\\", true); + tp!("\\\\.\\a", "\\\\.\\a", false); + + tp!("\\\\?\\a\\b\\", "\\\\?\\a\\", true); + } + } + + #[test] + pub fn test_set_file_name() { + macro_rules! tfn( + ($path:expr, $file:expr, $expected:expr) => ( { + let mut p = PathBuf::from($path); + p.set_file_name($file); + assert!(p.to_str() == Some($expected), + "setting file name of {:?} to {:?}: Expected {:?}, got {:?}", + $path, $file, $expected, + p.to_str().unwrap()); + }); + ); + + tfn!("foo", "foo", "foo"); + tfn!("foo", "bar", "bar"); + tfn!("foo", "", ""); + tfn!("", "foo", "foo"); + if cfg!(unix) { + tfn!(".", "foo", "./foo"); + tfn!("foo/", "bar", "bar"); + tfn!("foo/.", "bar", "bar"); + tfn!("..", "foo", "../foo"); + tfn!("foo/..", "bar", "foo/../bar"); + tfn!("/", "foo", "/foo"); + } else { + tfn!(".", "foo", r".\foo"); + tfn!(r"foo\", "bar", r"bar"); + tfn!(r"foo\.", "bar", r"bar"); + tfn!("..", "foo", r"..\foo"); + tfn!(r"foo\..", "bar", r"foo\..\bar"); + tfn!(r"\", "foo", r"\foo"); + } + } + + #[test] + pub fn test_set_extension() { + macro_rules! tfe( + ($path:expr, $ext:expr, $expected:expr, $output:expr) => ( { + let mut p = PathBuf::from($path); + let output = p.set_extension($ext); + assert!(p.to_str() == Some($expected) && output == $output, + "setting extension of {:?} to {:?}: Expected {:?}/{:?}, got {:?}/{:?}", + $path, $ext, $expected, $output, + p.to_str().unwrap(), output); + }); + ); + + tfe!("foo", "txt", "foo.txt", true); + tfe!("foo.bar", "txt", "foo.txt", true); + tfe!("foo.bar.baz", "txt", "foo.bar.txt", true); + tfe!(".test", "txt", ".test.txt", true); + tfe!("foo.txt", "", "foo", true); + tfe!("foo", "", "foo", true); + tfe!("", "foo", "", false); + tfe!(".", "foo", ".", false); + tfe!("foo/", "bar", "foo.bar", true); + tfe!("foo/.", "bar", "foo.bar", true); + tfe!("..", "foo", "..", false); + tfe!("foo/..", "bar", "foo/..", false); + tfe!("/", "foo", "/", false); + } + + #[test] + fn test_eq_recievers() { + use collections::borrow::Cow; + + let borrowed: &Path = Path::new("foo/bar"); + let mut owned: PathBuf = PathBuf::new(); + owned.push("foo"); + owned.push("bar"); + let borrowed_cow: Cow = borrowed.into(); + let owned_cow: Cow = owned.clone().into(); + + macro_rules! t { + ($($current:expr),+) => { + $( + assert_eq!($current, borrowed); + assert_eq!($current, owned); + assert_eq!($current, borrowed_cow); + assert_eq!($current, owned_cow); + )+ + } + } + + t!(borrowed, owned, borrowed_cow, owned_cow); + } + + #[test] + pub fn test_compare() { + use core::hash::{Hash, Hasher, SipHasher}; + + fn hash(t: T) -> u64 { + let mut s = SipHasher::new_with_keys(0, 0); + t.hash(&mut s); + s.finish() + } + + macro_rules! tc( + ($path1:expr, $path2:expr, eq: $eq:expr, + starts_with: $starts_with:expr, ends_with: $ends_with:expr, + relative_from: $relative_from:expr) => ({ + let path1 = Path::new($path1); + let path2 = Path::new($path2); + + let eq = path1 == path2; + assert!(eq == $eq, "{:?} == {:?}, expected {:?}, got {:?}", + $path1, $path2, $eq, eq); + assert!($eq == (hash(path1) == hash(path2)), + "{:?} == {:?}, expected {:?}, got {} and {}", + $path1, $path2, $eq, hash(path1), hash(path2)); + + let starts_with = path1.starts_with(path2); + assert!(starts_with == $starts_with, + "{:?}.starts_with({:?}), expected {:?}, got {:?}", $path1, $path2, + $starts_with, starts_with); + + let ends_with = path1.ends_with(path2); + assert!(ends_with == $ends_with, + "{:?}.ends_with({:?}), expected {:?}, got {:?}", $path1, $path2, + $ends_with, ends_with); + + let relative_from = path1.strip_prefix(path2) + .map(|p| p.to_str().unwrap()) + .ok(); + let exp: Option<&str> = $relative_from; + assert!(relative_from == exp, + "{:?}.strip_prefix({:?}), expected {:?}, got {:?}", + $path1, $path2, exp, relative_from); + }); + ); + + tc!("", "", + eq: true, + starts_with: true, + ends_with: true, + relative_from: Some("") + ); + + tc!("foo", "", + eq: false, + starts_with: true, + ends_with: true, + relative_from: Some("foo") + ); + + tc!("", "foo", + eq: false, + starts_with: false, + ends_with: false, + relative_from: None + ); + + tc!("foo", "foo", + eq: true, + starts_with: true, + ends_with: true, + relative_from: Some("") + ); + + tc!("foo/", "foo", + eq: true, + starts_with: true, + ends_with: true, + relative_from: Some("") + ); + + tc!("foo/bar", "foo", + eq: false, + starts_with: true, + ends_with: false, + relative_from: Some("bar") + ); + + tc!("foo/bar/baz", "foo/bar", + eq: false, + starts_with: true, + ends_with: false, + relative_from: Some("baz") + ); + + tc!("foo/bar", "foo/bar/baz", + eq: false, + starts_with: false, + ends_with: false, + relative_from: None + ); + + tc!("./foo/bar/", ".", + eq: false, + starts_with: true, + ends_with: false, + relative_from: Some("foo/bar") + ); + + if cfg!(windows) { + tc!(r"C:\src\rust\cargo-test\test\Cargo.toml", + r"c:\src\rust\cargo-test\test", + eq: false, + starts_with: true, + ends_with: false, + relative_from: Some("Cargo.toml") + ); + + tc!(r"c:\foo", r"C:\foo", + eq: true, + starts_with: true, + ends_with: true, + relative_from: Some("") + ); + } + } +} diff --git a/src/sys/mod.rs b/src/sys/mod.rs new file mode 100644 index 0000000..86f49e5 --- /dev/null +++ b/src/sys/mod.rs @@ -0,0 +1,25 @@ +/// A trait for viewing representations from std types +#[doc(hidden)] +pub trait AsInner { + fn as_inner(&self) -> &Inner; +} + +/// A trait for viewing representations from std types +#[doc(hidden)] +pub trait AsInnerMut { + fn as_inner_mut(&mut self) -> &mut Inner; +} + +/// A trait for extracting representations from std types +#[doc(hidden)] +pub trait IntoInner { + fn into_inner(self) -> Inner; +} + +/// A trait for creating std types from internal representations +#[doc(hidden)] +pub trait FromInner { + fn from_inner(inner: Inner) -> Self; +} + +pub mod wtf8; diff --git a/src/sys/wtf8.rs b/src/sys/wtf8.rs new file mode 100644 index 0000000..25c21c7 --- /dev/null +++ b/src/sys/wtf8.rs @@ -0,0 +1,1199 @@ +// Copyright 2015 The Rust Project Developers. See the COPYRIGHT +// file at the top-level directory of this distribution and at +// http://rust-lang.org/COPYRIGHT. +// +// Licensed under the Apache License, Version 2.0 or the MIT license +// , at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/). +//! +//! This library uses Rust’s type system to maintain +//! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed), +//! like the `String` and `&str` types do for UTF-8. +//! +//! Since [WTF-8 must not be used +//! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience), +//! this library deliberately does not provide access to the underlying bytes +//! of WTF-8 strings, +//! nor can it decode WTF-8 from arbitrary bytes. +//! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. + +// this module is imported from @SimonSapin's repo and has tons of dead code on +// unix (it's mostly used on windows), so don't worry about dead code here. +#![allow(dead_code)] + +use core::str::next_code_point; + +use ascii::*; +use collections::borrow::Cow; +use rustc_unicode::char; +use core::fmt; +use core::hash::{Hash, Hasher}; +use core::iter::FromIterator; +use core::mem; +use core::ops; +use collections::slice; +use core::str; +use collections::String; +use sys::AsInner; +use collections::Vec; + +const UTF8_REPLACEMENT_CHARACTER: &'static [u8] = b"\xEF\xBF\xBD"; + +/// A Unicode code point: from U+0000 to U+10FFFF. +/// +/// Compare with the `char` type, +/// which represents a Unicode scalar value: +/// a code point that is not a surrogate (U+D800 to U+DFFF). +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] +pub struct CodePoint { + value: u32, +} + +/// Format the code point as `U+` followed by four to six hexadecimal digits. +/// Example: `U+1F4A9` +impl fmt::Debug for CodePoint { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + write!(formatter, "U+{:04X}", self.value) + } +} + +impl CodePoint { + /// Unsafely creates a new `CodePoint` without checking the value. + /// + /// Only use when `value` is known to be less than or equal to 0x10FFFF. + #[inline] + pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { + CodePoint { value: value } + } + + /// Creates a new `CodePoint` if the value is a valid code point. + /// + /// Returns `None` if `value` is above 0x10FFFF. + #[inline] + pub fn from_u32(value: u32) -> Option { + match value { + 0...0x10FFFF => Some(CodePoint { value: value }), + _ => None, + } + } + + /// Creates a new `CodePoint` from a `char`. + /// + /// Since all Unicode scalar values are code points, this always succeeds. + #[inline] + pub fn from_char(value: char) -> CodePoint { + CodePoint { value: value as u32 } + } + + /// Returns the numeric value of the code point. + #[inline] + pub fn to_u32(&self) -> u32 { + self.value + } + + /// Optionally returns a Unicode scalar value for the code point. + /// + /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char(&self) -> Option { + match self.value { + 0xD800...0xDFFF => None, + _ => Some(unsafe { char::from_u32_unchecked(self.value) }), + } + } + + /// Returns a Unicode scalar value for the code point. + /// + /// Returns `'\u{FFFD}'` (the replacement character “�”) + /// if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char_lossy(&self) -> char { + self.to_char().unwrap_or('\u{FFFD}') + } +} + +/// An owned, growable string of well-formed WTF-8 data. +/// +/// Similar to `String`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] +pub struct Wtf8Buf { + bytes: Vec, +} + +impl ops::Deref for Wtf8Buf { + type Target = Wtf8; + + fn deref(&self) -> &Wtf8 { + self.as_slice() + } +} + +/// Format the string with double quotes, +/// and surrogates as `\u` followed by four hexadecimal digits. +/// Example: `"a\u{D800}"` for a string with code points [U+0061, U+D800] +impl fmt::Debug for Wtf8Buf { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter) -> Result<(), fmt::Error> { + fmt::Debug::fmt(&**self, formatter) + } +} + +impl Wtf8Buf { + /// Creates a new, empty WTF-8 string. + #[inline] + pub fn new() -> Wtf8Buf { + Wtf8Buf { bytes: Vec::new() } + } + + /// Creates a new, empty WTF-8 string with pre-allocated capacity for `n` bytes. + #[inline] + pub fn with_capacity(n: usize) -> Wtf8Buf { + Wtf8Buf { bytes: Vec::with_capacity(n) } + } + + /// Creates a WTF-8 string from a UTF-8 `String`. + /// + /// This takes ownership of the `String` and does not copy. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_string(string: String) -> Wtf8Buf { + Wtf8Buf { bytes: string.into_bytes() } + } + + /// Creates a WTF-8 string from a UTF-8 `&str` slice. + /// + /// This copies the content of the slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(str: &str) -> Wtf8Buf { + Wtf8Buf { bytes: <[_]>::to_vec(str.as_bytes()) } + } + + pub fn clear(&mut self) { + self.bytes.clear() + } + + /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units. + /// + /// This is lossless: calling `.encode_wide()` on the resulting string + /// will always return the original code units. + pub fn from_wide(v: &[u16]) -> Wtf8Buf { + let mut string = Wtf8Buf::with_capacity(v.len()); + for item in char::decode_utf16(v.iter().cloned()) { + match item { + Ok(ch) => string.push_char(ch), + Err(surrogate) => { + let surrogate = surrogate.unpaired_surrogate(); + // Surrogates are known to be in the code point range. + let code_point = unsafe { CodePoint::from_u32_unchecked(surrogate as u32) }; + // Skip the WTF-8 concatenation check, + // surrogate pairs are already decoded by decode_utf16 + string.push_code_point_unchecked(code_point) + } + } + } + string + } + + /// Copied from String::push + /// This does **not** include the WTF-8 concatenation check. + fn push_code_point_unchecked(&mut self, code_point: CodePoint) { + let bytes = unsafe { char::from_u32_unchecked(code_point.value).encode_utf8() }; + self.bytes.extend_from_slice(bytes.as_slice()); + } + + #[inline] + pub fn as_slice(&self) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(&self.bytes) } + } + + /// Reserves capacity for at least `additional` more bytes to be inserted + /// in the given `Wtf8Buf`. + /// The collection may reserve more space to avoid frequent reallocations. + /// + /// # Panics + /// + /// Panics if the new capacity overflows `usize`. + #[inline] + pub fn reserve(&mut self, additional: usize) { + self.bytes.reserve(additional) + } + + #[inline] + pub fn reserve_exact(&mut self, additional: usize) { + self.bytes.reserve_exact(additional) + } + + /// Returns the number of bytes that this string buffer can hold without reallocating. + #[inline] + pub fn capacity(&self) -> usize { + self.bytes.capacity() + } + + /// Append a UTF-8 slice at the end of the string. + #[inline] + pub fn push_str(&mut self, other: &str) { + self.bytes.extend_from_slice(other.as_bytes()) + } + + /// Append a WTF-8 slice at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push_wtf8(&mut self, other: &Wtf8) { + match ((&*self).final_lead_surrogate(), other.initial_trail_surrogate()) { + // Replace newly paired surrogates by a supplementary code point. + (Some(lead), Some(trail)) => { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + let other_without_trail_surrogate = &other.bytes[3..]; + // 4 bytes for the supplementary code point + self.bytes.reserve(4 + other_without_trail_surrogate.len()); + self.push_char(decode_surrogate_pair(lead, trail)); + self.bytes.extend_from_slice(other_without_trail_surrogate); + } + _ => self.bytes.extend_from_slice(&other.bytes), + } + } + + /// Append a Unicode scalar value at the end of the string. + #[inline] + pub fn push_char(&mut self, c: char) { + self.push_code_point_unchecked(CodePoint::from_char(c)) + } + + /// Append a code point at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push(&mut self, code_point: CodePoint) { + if let trail @ 0xDC00...0xDFFF = code_point.to_u32() { + if let Some(lead) = (&*self).final_lead_surrogate() { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + self.push_char(decode_surrogate_pair(lead, trail as u16)); + return; + } + } + + // No newly paired surrogates at the boundary. + self.push_code_point_unchecked(code_point) + } + + /// Shortens a string to the specified length. + /// + /// # Panics + /// + /// Panics if `new_len` > current length, + /// or if `new_len` is not a code point boundary. + #[inline] + pub fn truncate(&mut self, new_len: usize) { + assert!(is_code_point_boundary(self, new_len)); + self.bytes.truncate(new_len) + } + + /// Consumes the WTF-8 string and tries to convert it to UTF-8. + /// + /// This does not copy the data. + /// + /// If the contents are not well-formed UTF-8 + /// (that is, if the string contains surrogates), + /// the original WTF-8 string is returned instead. + pub fn into_string(self) -> Result { + match self.next_surrogate(0) { + None => Ok(unsafe { String::from_utf8_unchecked(self.bytes) }), + Some(_) => Err(self), + } + } + + /// Consumes the WTF-8 string and converts it lossily to UTF-8. + /// + /// This does not copy the data (but may overwrite parts of it in place). + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”) + pub fn into_string_lossy(mut self) -> String { + let mut pos = 0; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + pos = surrogate_pos + 3; + self.bytes[surrogate_pos..pos].copy_from_slice(UTF8_REPLACEMENT_CHARACTER); + } + None => return unsafe { String::from_utf8_unchecked(self.bytes) }, + } + } + } +} + +/// Create a new WTF-8 string from an iterator of code points. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl FromIterator for Wtf8Buf { + fn from_iter>(iter: T) -> Wtf8Buf { + let mut string = Wtf8Buf::new(); + string.extend(iter); + string + } +} + +/// Append code points from an iterator to the string. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl Extend for Wtf8Buf { + fn extend>(&mut self, iter: T) { + let iterator = iter.into_iter(); + let (low, _high) = iterator.size_hint(); + // Lower bound of one byte per code point (ASCII only) + self.bytes.reserve(low); + for code_point in iterator { + self.push(code_point); + } + } +} + +/// A borrowed slice of well-formed WTF-8 data. +/// +/// Similar to `&str`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, Ord, PartialEq, PartialOrd)] +pub struct Wtf8 { + bytes: [u8], +} + +impl AsInner<[u8]> for Wtf8 { + fn as_inner(&self) -> &[u8] { + &self.bytes + } +} + +/// Format the slice with double quotes, +/// and surrogates as `\u` followed by four hexadecimal digits. +/// Example: `"a\u{D800}"` for a slice with code points [U+0061, U+D800] +impl fmt::Debug for Wtf8 { + fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { + fn write_str_escaped(f: &mut fmt::Formatter, s: &str) -> fmt::Result { + use core::fmt::Write; + for c in s.chars().flat_map(|c| c.escape_debug()) { + f.write_char(c)? + } + Ok(()) + } + + formatter.write_str("\"")?; + let mut pos = 0; + loop { + match self.next_surrogate(pos) { + None => break, + Some((surrogate_pos, surrogate)) => { + write_str_escaped(formatter, unsafe { + str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) + }) + ?; + write!(formatter, "\\u{{{:x}}}", surrogate)?; + pos = surrogate_pos + 3; + } + } + } + write_str_escaped(formatter, + unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) }) + ?; + formatter.write_str("\"") + } +} + +impl Wtf8 { + /// Creates a WTF-8 slice from a UTF-8 `&str` slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(value: &str) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } + } + + /// Creates a WTF-8 slice from a WTF-8 byte slice. + /// + /// Since the byte slice is not checked for valid WTF-8, this functions is + /// marked unsafe. + #[inline] + unsafe fn from_bytes_unchecked(value: &[u8]) -> &Wtf8 { + mem::transmute(value) + } + + /// Returns the length, in WTF-8 bytes. + #[inline] + pub fn len(&self) -> usize { + self.bytes.len() + } + + #[inline] + pub fn is_empty(&self) -> bool { + self.bytes.is_empty() + } + + /// Returns the code point at `position` if it is in the ASCII range, + /// or `b'\xFF' otherwise. + /// + /// # Panics + /// + /// Panics if `position` is beyond the end of the string. + #[inline] + pub fn ascii_byte_at(&self, position: usize) -> u8 { + match self.bytes[position] { + ascii_byte @ 0x00...0x7F => ascii_byte, + _ => 0xFF, + } + } + + /// Returns an iterator for the string’s code points. + #[inline] + pub fn code_points(&self) -> Wtf8CodePoints { + Wtf8CodePoints { bytes: self.bytes.iter() } + } + + /// Tries to convert the string to UTF-8 and return a `&str` slice. + /// + /// Returns `None` if the string contains surrogates. + /// + /// This does not copy the data. + #[inline] + pub fn as_str(&self) -> Option<&str> { + // Well-formed WTF-8 is also well-formed UTF-8 + // if and only if it contains no surrogate. + match self.next_surrogate(0) { + None => Some(unsafe { str::from_utf8_unchecked(&self.bytes) }), + Some(_) => None, + } + } + + /// Lossily converts the string to UTF-8. + /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). + /// + /// This only copies the data if necessary (if it contains any surrogate). + pub fn to_string_lossy(&self) -> Cow { + let surrogate_pos = match self.next_surrogate(0) { + None => return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }), + Some((pos, _)) => pos, + }; + let wtf8_bytes = &self.bytes; + let mut utf8_bytes = Vec::with_capacity(self.len()); + utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER); + let mut pos = surrogate_pos + 3; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER); + pos = surrogate_pos + 3; + } + None => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); + return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); + } + } + } + } + + /// Converts the WTF-8 string to potentially ill-formed UTF-16 + /// and return an iterator of 16-bit code units. + /// + /// This is lossless: + /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units + /// would always return the original WTF-8 string. + #[inline] + pub fn encode_wide(&self) -> EncodeWide { + EncodeWide { + code_points: self.code_points(), + extra: 0, + } + } + + #[inline] + fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { + let mut iter = self.bytes[pos..].iter(); + loop { + let b = match iter.next() { + None => return None, + Some(&b) => b, + }; + if b < 0x80 { + pos += 1; + } else if b < 0xE0 { + iter.next(); + pos += 2; + } else if b == 0xED { + match (iter.next(), iter.next()) { + (Some(&b2), Some(&b3)) if b2 >= 0xA0 => { + return Some((pos, decode_surrogate(b2, b3))) + } + _ => pos += 3, + } + } else if b < 0xF0 { + iter.next(); + iter.next(); + pos += 3; + } else { + iter.next(); + iter.next(); + iter.next(); + pos += 4; + } + } + } + + #[inline] + fn final_lead_surrogate(&self) -> Option { + let len = self.len(); + if len < 3 { + return None; + } + match &self.bytes[(len - 3)..] { + &[0xED, b2 @ 0xA0...0xAF, b3] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } + + #[inline] + fn initial_trail_surrogate(&self) -> Option { + let len = self.len(); + if len < 3 { + return None; + } + match &self.bytes[..3] { + &[0xED, b2 @ 0xB0...0xBF, b3] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } +} + + +/// Return a slice of the given string for the byte range [`begin`..`end`). +/// +/// # Panics +/// +/// Panics when `begin` and `end` do not point to code point boundaries, +/// or point beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::Range) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if range.start <= range.end && is_code_point_boundary(self, range.start) && + is_code_point_boundary(self, range.end) { + unsafe { slice_unchecked(self, range.start, range.end) } + } else { + slice_error_fail(self, range.start, range.end) + } + } +} + +/// Return a slice of the given string from byte `begin` to its end. +/// +/// # Panics +/// +/// Panics when `begin` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeFrom) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.start) { + unsafe { slice_unchecked(self, range.start, self.len()) } + } else { + slice_error_fail(self, range.start, self.len()) + } + } +} + +/// Return a slice of the given string from its beginning to byte `end`. +/// +/// # Panics +/// +/// Panics when `end` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeTo) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.end) { + unsafe { slice_unchecked(self, 0, range.end) } + } else { + slice_error_fail(self, 0, range.end) + } + } +} + +impl ops::Index for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, _range: ops::RangeFull) -> &Wtf8 { + self + } +} + +#[inline] +fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { + // The first byte is assumed to be 0xED + 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F +} + +#[inline] +fn decode_surrogate_pair(lead: u16, trail: u16) -> char { + let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); + unsafe { char::from_u32_unchecked(code_point) } +} + +/// Copied from core::str::StrPrelude::is_char_boundary +#[inline] +pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { + if index == slice.len() { + return true; + } + match slice.bytes.get(index) { + None => false, + Some(&b) => b < 128 || b >= 192, + } +} + +/// Copied from core::str::raw::slice_unchecked +#[inline] +pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { + // memory layout of an &[u8] and &Wtf8 are the same + Wtf8::from_bytes_unchecked(slice::from_raw_parts(s.bytes.as_ptr().offset(begin as isize), + end - begin)) +} + +/// Copied from core::str::raw::slice_error_fail +#[inline(never)] +pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { + assert!(begin <= end); + panic!("index {} and/or {} in `{:?}` do not lie on character boundary", + begin, + end, + s); +} + +/// Iterator for the code points of a WTF-8 string. +/// +/// Created with the method `.code_points()`. +#[derive(Clone)] +pub struct Wtf8CodePoints<'a> { + bytes: slice::Iter<'a, u8>, +} + +impl<'a> Iterator for Wtf8CodePoints<'a> { + type Item = CodePoint; + + #[inline] + fn next(&mut self) -> Option { + next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let len = self.bytes.len(); + (len.saturating_add(3) / 4, Some(len)) + } +} + +#[derive(Clone)] +pub struct EncodeWide<'a> { + code_points: Wtf8CodePoints<'a>, + extra: u16, +} + +// Copied from libunicode/u_str.rs +impl<'a> Iterator for EncodeWide<'a> { + type Item = u16; + + #[inline] + fn next(&mut self) -> Option { + if self.extra != 0 { + let tmp = self.extra; + self.extra = 0; + return Some(tmp); + } + + self.code_points.next().map(|code_point| { + let n = unsafe { char::from_u32_unchecked(code_point.value).encode_utf16() }; + let n = n.as_slice(); + if n.len() == 2 { + self.extra = n[1]; + } + n[0] + }) + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let (low, high) = self.code_points.size_hint(); + // every code point gets either one u16 or two u16, + // so this iterator is between 1 or 2 times as + // long as the underlying iterator. + (low, high.and_then(|n| n.checked_mul(2))) + } +} + +impl Hash for CodePoint { + #[inline] + fn hash(&self, state: &mut H) { + self.value.hash(state) + } +} + +impl Hash for Wtf8Buf { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +impl Hash for Wtf8 { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +impl AsciiExt for Wtf8 { + type Owned = Wtf8Buf; + + fn is_ascii(&self) -> bool { + self.bytes.is_ascii() + } + fn to_ascii_uppercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_uppercase() } + } + fn to_ascii_lowercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_lowercase() } + } + fn eq_ignore_ascii_case(&self, other: &Wtf8) -> bool { + self.bytes.eq_ignore_ascii_case(&other.bytes) + } + + fn make_ascii_uppercase(&mut self) { + self.bytes.make_ascii_uppercase() + } + fn make_ascii_lowercase(&mut self) { + self.bytes.make_ascii_lowercase() + } +} + +#[cfg(test)] +mod tests { + use collections::borrow::Cow; + use collections::{String, Vec}; + use super::*; + + #[test] + fn code_point_from_u32() { + assert!(CodePoint::from_u32(0).is_some()); + assert!(CodePoint::from_u32(0xD800).is_some()); + assert!(CodePoint::from_u32(0x10FFFF).is_some()); + assert!(CodePoint::from_u32(0x110000).is_none()); + } + + #[test] + fn code_point_to_u32() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0).to_u32(), 0); + assert_eq!(c(0xD800).to_u32(), 0xD800); + assert_eq!(c(0x10FFFF).to_u32(), 0x10FFFF); + } + + #[test] + fn code_point_from_char() { + assert_eq!(CodePoint::from_char('a').to_u32(), 0x61); + assert_eq!(CodePoint::from_char('💩').to_u32(), 0x1F4A9); + } + + #[test] + fn code_point_to_string() { + assert_eq!(format!("{:?}", CodePoint::from_char('a')), "U+0061"); + assert_eq!(format!("{:?}", CodePoint::from_char('💩')), "U+1F4A9"); + } + + #[test] + fn code_point_to_char() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0x61).to_char(), Some('a')); + assert_eq!(c(0x1F4A9).to_char(), Some('💩')); + assert_eq!(c(0xD800).to_char(), None); + } + + #[test] + fn code_point_to_char_lossy() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0x61).to_char_lossy(), 'a'); + assert_eq!(c(0x1F4A9).to_char_lossy(), '💩'); + assert_eq!(c(0xD800).to_char_lossy(), '\u{FFFD}'); + } + + #[test] + fn wtf8buf_new() { + assert_eq!(Wtf8Buf::new().bytes, b""); + } + + #[test] + fn wtf8buf_from_str() { + assert_eq!(Wtf8Buf::from_str("").bytes, b""); + assert_eq!(Wtf8Buf::from_str("aé 💩").bytes, + b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8buf_from_string() { + assert_eq!(Wtf8Buf::from_string(String::from("")).bytes, b""); + assert_eq!(Wtf8Buf::from_string(String::from("aé 💩")).bytes, + b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8buf_from_wide() { + assert_eq!(Wtf8Buf::from_wide(&[]).bytes, b""); + assert_eq!(Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]).bytes, + b"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8buf_push_str() { + let mut string = Wtf8Buf::new(); + assert_eq!(string.bytes, b""); + string.push_str("aé 💩"); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8buf_push_char() { + let mut string = Wtf8Buf::from_str("aé "); + assert_eq!(string.bytes, b"a\xC3\xA9 "); + string.push_char('💩'); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8buf_push() { + let mut string = Wtf8Buf::from_str("aé "); + assert_eq!(string.bytes, b"a\xC3\xA9 "); + string.push(CodePoint::from_char('💩')); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + + let mut string = Wtf8Buf::new(); + string.push(c(0xD83D)); // lead + string.push(c(0xDCA9)); // trail + assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + + let mut string = Wtf8Buf::new(); + string.push(c(0xD83D)); // lead + string.push(c(0x20)); // not surrogate + string.push(c(0xDCA9)); // trail + assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD800)); // lead + string.push(c(0xDBFF)); // lead + assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD800)); // lead + string.push(c(0xE000)); // not surrogate + assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD7FF)); // not surrogate + string.push(c(0xDC00)); // trail + assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0x61)); // not surrogate, < 3 bytes + string.push(c(0xDC00)); // trail + assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xDC00)); // trail + assert_eq!(string.bytes, b"\xED\xB0\x80"); + } + + #[test] + fn wtf8buf_push_wtf8() { + let mut string = Wtf8Buf::from_str("aé"); + assert_eq!(string.bytes, b"a\xC3\xA9"); + string.push_wtf8(Wtf8::from_str(" 💩")); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + + fn w(v: &[u8]) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(v) } + } + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\xBD")); // lead + string.push_wtf8(w(b"\xED\xB2\xA9")); // trail + assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\xBD")); // lead + string.push_wtf8(w(b" ")); // not surrogate + string.push_wtf8(w(b"\xED\xB2\xA9")); // trail + assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\x80")); // lead + string.push_wtf8(w(b"\xED\xAF\xBF")); // lead + assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\x80")); // lead + string.push_wtf8(w(b"\xEE\x80\x80")); // not surrogate + assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\x9F\xBF")); // not surrogate + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"a")); // not surrogate, < 3 bytes + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\xED\xB0\x80"); + } + + #[test] + fn wtf8buf_truncate() { + let mut string = Wtf8Buf::from_str("aé"); + string.truncate(1); + assert_eq!(string.bytes, b"a"); + } + + #[test] + #[should_panic] + fn wtf8buf_truncate_fail_code_point_boundary() { + let mut string = Wtf8Buf::from_str("aé"); + string.truncate(2); + } + + #[test] + #[should_panic] + fn wtf8buf_truncate_fail_longer() { + let mut string = Wtf8Buf::from_str("aé"); + string.truncate(4); + } + + #[test] + fn wtf8buf_into_string() { + let mut string = Wtf8Buf::from_str("aé 💩"); + assert_eq!(string.clone().into_string(), Ok(String::from("aé 💩"))); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(string.clone().into_string(), Err(string)); + } + + #[test] + fn wtf8buf_into_string_lossy() { + let mut string = Wtf8Buf::from_str("aé 💩"); + assert_eq!(string.clone().into_string_lossy(), String::from("aé 💩")); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(string.clone().into_string_lossy(), + String::from("aé 💩�")); + } + + #[test] + fn wtf8buf_from_iterator() { + fn f(values: &[u32]) -> Wtf8Buf { + values.iter().map(|&c| CodePoint::from_u32(c).unwrap()).collect::() + } + assert_eq!(f(&[0x61, 0xE9, 0x20, 0x1F4A9]).bytes, + b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + + assert_eq!(f(&[0xD83D, 0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(f(&[0xD83D, 0x20, 0xDCA9]).bytes, + b"\xED\xA0\xBD \xED\xB2\xA9"); + assert_eq!(f(&[0xD800, 0xDBFF]).bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + assert_eq!(f(&[0xD800, 0xE000]).bytes, b"\xED\xA0\x80\xEE\x80\x80"); + assert_eq!(f(&[0xD7FF, 0xDC00]).bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + assert_eq!(f(&[0x61, 0xDC00]).bytes, b"\x61\xED\xB0\x80"); + assert_eq!(f(&[0xDC00]).bytes, b"\xED\xB0\x80"); + } + + #[test] + fn wtf8buf_extend() { + fn e(initial: &[u32], extended: &[u32]) -> Wtf8Buf { + fn c(value: &u32) -> CodePoint { + CodePoint::from_u32(*value).unwrap() + } + let mut string = initial.iter().map(c).collect::(); + string.extend(extended.iter().map(c)); + string + } + + assert_eq!(e(&[0x61, 0xE9], &[0x20, 0x1F4A9]).bytes, + b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + + assert_eq!(e(&[0xD83D], &[0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(e(&[0xD83D, 0x20], &[0xDCA9]).bytes, + b"\xED\xA0\xBD \xED\xB2\xA9"); + assert_eq!(e(&[0xD800], &[0xDBFF]).bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + assert_eq!(e(&[0xD800], &[0xE000]).bytes, b"\xED\xA0\x80\xEE\x80\x80"); + assert_eq!(e(&[0xD7FF], &[0xDC00]).bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + assert_eq!(e(&[0x61], &[0xDC00]).bytes, b"\x61\xED\xB0\x80"); + assert_eq!(e(&[], &[0xDC00]).bytes, b"\xED\xB0\x80"); + } + + #[test] + fn wtf8buf_show() { + let mut string = Wtf8Buf::from_str("a\té \u{7f}💩\r"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(format!("{:?}", string), + "\"a\\té \\u{7f}\u{1f4a9}\\r\\u{d800}\""); + } + + #[test] + fn wtf8buf_as_slice() { + assert_eq!(Wtf8Buf::from_str("aé").as_slice(), Wtf8::from_str("aé")); + } + + #[test] + fn wtf8buf_show_str() { + let text = "a\té 💩\r"; + let string = Wtf8Buf::from_str(text); + assert_eq!(format!("{:?}", text), format!("{:?}", string)); + } + + #[test] + fn wtf8_from_str() { + assert_eq!(&Wtf8::from_str("").bytes, b""); + assert_eq!(&Wtf8::from_str("aé 💩").bytes, + b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + fn wtf8_len() { + assert_eq!(Wtf8::from_str("").len(), 0); + assert_eq!(Wtf8::from_str("aé 💩").len(), 8); + } + + #[test] + fn wtf8_slice() { + assert_eq!(&Wtf8::from_str("aé 💩")[1..4].bytes, b"\xC3\xA9 "); + } + + #[test] + #[should_panic] + fn wtf8_slice_not_code_point_boundary() { + &Wtf8::from_str("aé 💩")[2..4]; + } + + #[test] + fn wtf8_slice_from() { + assert_eq!(&Wtf8::from_str("aé 💩")[1..].bytes, + b"\xC3\xA9 \xF0\x9F\x92\xA9"); + } + + #[test] + #[should_panic] + fn wtf8_slice_from_not_code_point_boundary() { + &Wtf8::from_str("aé 💩")[2..]; + } + + #[test] + fn wtf8_slice_to() { + assert_eq!(&Wtf8::from_str("aé 💩")[..4].bytes, b"a\xC3\xA9 "); + } + + #[test] + #[should_panic] + fn wtf8_slice_to_not_code_point_boundary() { + &Wtf8::from_str("aé 💩")[5..]; + } + + #[test] + fn wtf8_ascii_byte_at() { + let slice = Wtf8::from_str("aé 💩"); + assert_eq!(slice.ascii_byte_at(0), b'a'); + assert_eq!(slice.ascii_byte_at(1), b'\xFF'); + assert_eq!(slice.ascii_byte_at(2), b'\xFF'); + assert_eq!(slice.ascii_byte_at(3), b' '); + assert_eq!(slice.ascii_byte_at(4), b'\xFF'); + } + + #[test] + fn wtf8_code_points() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + fn cp(string: &Wtf8Buf) -> Vec> { + string.code_points().map(|c| c.to_char()).collect::>() + } + let mut string = Wtf8Buf::from_str("é "); + assert_eq!(cp(&string), [Some('é'), Some(' ')]); + string.push(c(0xD83D)); + assert_eq!(cp(&string), [Some('é'), Some(' '), None]); + string.push(c(0xDCA9)); + assert_eq!(cp(&string), [Some('é'), Some(' '), Some('💩')]); + } + + #[test] + fn wtf8_as_str() { + assert_eq!(Wtf8::from_str("").as_str(), Some("")); + assert_eq!(Wtf8::from_str("aé 💩").as_str(), Some("aé 💩")); + let mut string = Wtf8Buf::new(); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(string.as_str(), None); + } + + #[test] + fn wtf8_to_string_lossy() { + assert_eq!(Wtf8::from_str("").to_string_lossy(), Cow::Borrowed("")); + assert_eq!(Wtf8::from_str("aé 💩").to_string_lossy(), + Cow::Borrowed("aé 💩")); + let mut string = Wtf8Buf::from_str("aé 💩"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + let expected: Cow = Cow::Owned(String::from("aé 💩�")); + assert_eq!(string.to_string_lossy(), expected); + } + + #[test] + fn wtf8_encode_wide() { + let mut string = Wtf8Buf::from_str("aé "); + string.push(CodePoint::from_u32(0xD83D).unwrap()); + string.push_char('💩'); + assert_eq!(string.encode_wide().collect::>(), + vec![0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]); + } +}