std/os/windows/
ffi.rs

1//! Windows-specific extensions to primitives in the [`std::ffi`] module.
2//!
3//! # Overview
4//!
5//! For historical reasons, the Windows API uses a form of potentially
6//! ill-formed UTF-16 encoding for strings. Specifically, the 16-bit
7//! code units in Windows strings may contain [isolated surrogate code
8//! points which are not paired together][ill-formed-utf-16]. The
9//! Unicode standard requires that surrogate code points (those in the
10//! range U+D800 to U+DFFF) always be *paired*, because in the UTF-16
11//! encoding a *surrogate code unit pair* is used to encode a single
12//! character. For compatibility with code that does not enforce
13//! these pairings, Windows does not enforce them, either.
14//!
15//! While it is not always possible to convert such a string losslessly into
16//! a valid UTF-16 string (or even UTF-8), it is often desirable to be
17//! able to round-trip such a string from and to Windows APIs
18//! losslessly. For example, some Rust code may be "bridging" some
19//! Windows APIs together, just passing `WCHAR` strings among those
20//! APIs without ever really looking into the strings.
21//!
22//! If Rust code *does* need to look into those strings, it can
23//! convert them to valid UTF-8, possibly lossily, by substituting
24//! invalid sequences with [`U+FFFD REPLACEMENT CHARACTER`][U+FFFD], as is
25//! conventionally done in other Rust APIs that deal with string
26//! encodings.
27//!
28//! # `OsStringExt` and `OsStrExt`
29//!
30//! [`OsString`] is the Rust wrapper for owned strings in the
31//! preferred representation of the operating system. On Windows,
32//! this struct gets augmented with an implementation of the
33//! [`OsStringExt`] trait, which has an [`OsStringExt::from_wide`] method. This
34//! lets you create an [`OsString`] from a `&[u16]` slice; presumably
35//! you get such a slice out of a `WCHAR` Windows API.
36//!
37//! Similarly, [`OsStr`] is the Rust wrapper for borrowed strings from
38//! preferred representation of the operating system. On Windows, the
39//! [`OsStrExt`] trait provides the [`OsStrExt::encode_wide`] method, which
40//! outputs an [`EncodeWide`] iterator. You can [`collect`] this
41//! iterator, for example, to obtain a `Vec<u16>`; you can later get a
42//! pointer to this vector's contents and feed it to Windows APIs.
43//!
44//! These traits, along with [`OsString`] and [`OsStr`], work in
45//! conjunction so that it is possible to **round-trip** strings from
46//! Windows and back, with no loss of data, even if the strings are
47//! ill-formed UTF-16.
48//!
49//! [ill-formed-utf-16]: https://simonsapin.github.io/wtf-8/#ill-formed-utf-16
50//! [`collect`]: crate::iter::Iterator::collect
51//! [U+FFFD]: crate::char::REPLACEMENT_CHARACTER
52//! [`std::ffi`]: crate::ffi
53
54#![stable(feature = "rust1", since = "1.0.0")]
55
56use alloc::wtf8::Wtf8Buf;
57
58use crate::ffi::{OsStr, OsString};
59use crate::fmt;
60use crate::iter::FusedIterator;
61use crate::sys::os_str::Buf;
62use crate::sys::{AsInner, FromInner};
63
64/// Windows-specific extensions to [`OsString`].
65#[stable(feature = "rust1", since = "1.0.0")]
66pub impl(self) trait OsStringExt {
67    /// Creates an `OsString` from a potentially ill-formed UTF-16 slice of
68    /// 16-bit code units.
69    ///
70    /// This is lossless: calling [`OsStrExt::encode_wide`] on the resulting string
71    /// will always return the original code units.
72    ///
73    /// # Examples
74    ///
75    #[cfg_attr(windows, doc = "```no_run")]
76    #[cfg_attr(not(windows), doc = "```ignore (needs windows)")]
77    /// use std::ffi::OsString;
78    /// use std::os::windows::prelude::*;
79    ///
80    /// // UTF-16 encoding for "Unicode".
81    /// let source = [0x0055, 0x006E, 0x0069, 0x0063, 0x006F, 0x0064, 0x0065];
82    ///
83    /// let string = OsString::from_wide(&source[..]);
84    /// ```
85    #[stable(feature = "rust1", since = "1.0.0")]
86    fn from_wide(wide: &[u16]) -> Self;
87}
88
89#[stable(feature = "rust1", since = "1.0.0")]
90impl OsStringExt for OsString {
91    fn from_wide(wide: &[u16]) -> OsString {
92        FromInner::from_inner(Buf { inner: Wtf8Buf::from_wide(wide) })
93    }
94}
95
96/// Windows-specific extensions to [`OsStr`].
97#[stable(feature = "rust1", since = "1.0.0")]
98pub impl(self) trait OsStrExt {
99    /// Re-encodes an `OsStr` as a wide character sequence, i.e., potentially
100    /// ill-formed UTF-16.
101    ///
102    /// This is lossless: calling [`OsStringExt::from_wide`] and then
103    /// `encode_wide` on the result will yield the original code units.
104    /// Note that the encoding does not add a final null terminator.
105    ///
106    /// # Examples
107    ///
108    #[cfg_attr(windows, doc = "```no_run")]
109    #[cfg_attr(not(windows), doc = "```ignore (needs windows)")]
110    /// use std::ffi::OsString;
111    /// use std::os::windows::prelude::*;
112    ///
113    /// // UTF-16 encoding for "Unicode".
114    /// let source = [0x0055, 0x006E, 0x0069, 0x0063, 0x006F, 0x0064, 0x0065];
115    ///
116    /// let string = OsString::from_wide(&source[..]);
117    ///
118    /// let result: Vec<u16> = string.encode_wide().collect();
119    /// assert_eq!(&source[..], &result[..]);
120    /// ```
121    #[stable(feature = "rust1", since = "1.0.0")]
122    fn encode_wide(&self) -> EncodeWide<'_>;
123}
124
125#[stable(feature = "rust1", since = "1.0.0")]
126impl OsStrExt for OsStr {
127    #[inline]
128    fn encode_wide(&self) -> EncodeWide<'_> {
129        EncodeWide { inner: self.as_inner().inner.encode_wide() }
130    }
131}
132
133/// Iterator returned by [`OsStrExt::encode_wide`].
134#[stable(feature = "rust1", since = "1.0.0")]
135#[derive(Clone)]
136pub struct EncodeWide<'a> {
137    inner: alloc::wtf8::EncodeWide<'a>,
138}
139#[stable(feature = "encode_wide_debug", since = "1.91.0")]
140impl fmt::Debug for EncodeWide<'_> {
141    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
142        fmt::Debug::fmt(&self.inner, f)
143    }
144}
145#[stable(feature = "rust1", since = "1.0.0")]
146impl Iterator for EncodeWide<'_> {
147    type Item = u16;
148
149    #[inline]
150    fn next(&mut self) -> Option<u16> {
151        self.inner.next()
152    }
153
154    #[inline]
155    fn size_hint(&self) -> (usize, Option<usize>) {
156        self.inner.size_hint()
157    }
158}
159#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")]
160impl FusedIterator for EncodeWide<'_> {}
std/os/windows/ffi.rs

std/os/windows/
ffi.rs
