library/std/src/sys/windows/ext/ffi.rs

   1 //! Windows-specific extensions to the 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 a [`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
  53 #![stable(feature = "rust1", since = "1.0.0")]
  54
  55 use crate::ffi::{OsStr, OsString};
  56 use crate::sys::os_str::Buf;
  57 use crate::sys_common::wtf8::Wtf8Buf;
  58 use crate::sys_common::{AsInner, FromInner};
  59
  60 #[stable(feature = "rust1", since = "1.0.0")]
  61 pub use crate::sys_common::wtf8::EncodeWide;
  62
  63 /// Windows-specific extensions to [`OsString`].
  64 #[stable(feature = "rust1", since = "1.0.0")]
  65 pub trait OsStringExt {
  66     /// Creates an `OsString` from a potentially ill-formed UTF-16 slice of
  67     /// 16-bit code units.
  68     ///
  69     /// This is lossless: calling [`OsStrExt::encode_wide`] on the resulting string
  70     /// will always return the original code units.
  71     ///
  72     /// # Examples
  73     ///
  74     /// ```
  75     /// use std::ffi::OsString;
  76     /// use std::os::windows::prelude::*;
  77     ///
  78     /// // UTF-16 encoding for "Unicode".
  79     /// let source = [0x0055, 0x006E, 0x0069, 0x0063, 0x006F, 0x0064, 0x0065];
  80     ///
  81     /// let string = OsString::from_wide(&source[..]);
  82     /// ```
  83     #[stable(feature = "rust1", since = "1.0.0")]
  84     fn from_wide(wide: &[u16]) -> Self;
  85 }
  86
  87 #[stable(feature = "rust1", since = "1.0.0")]
  88 impl OsStringExt for OsString {
  89     fn from_wide(wide: &[u16]) -> OsString {
  90         FromInner::from_inner(Buf { inner: Wtf8Buf::from_wide(wide) })
  91     }
  92 }
  93
  94 /// Windows-specific extensions to [`OsStr`].
  95 #[stable(feature = "rust1", since = "1.0.0")]
  96 pub trait OsStrExt {
  97     /// Re-encodes an `OsStr` as a wide character sequence, i.e., potentially
  98     /// ill-formed UTF-16.
  99     ///
 100     /// This is lossless: calling [`OsStringExt::from_wide`] and then
 101     /// `encode_wide` on the result will yield the original code units.
 102     /// Note that the encoding does not add a final null terminator.
 103     ///
 104     /// # Examples
 105     ///
 106     /// ```
 107     /// use std::ffi::OsString;
 108     /// use std::os::windows::prelude::*;
 109     ///
 110     /// // UTF-16 encoding for "Unicode".
 111     /// let source = [0x0055, 0x006E, 0x0069, 0x0063, 0x006F, 0x0064, 0x0065];
 112     ///
 113     /// let string = OsString::from_wide(&source[..]);
 114     ///
 115     /// let result: Vec<u16> = string.encode_wide().collect();
 116     /// assert_eq!(&source[..], &result[..]);
 117     /// ```
 118     #[stable(feature = "rust1", since = "1.0.0")]
 119     fn encode_wide(&self) -> EncodeWide<'_>;
 120 }
 121
 122 #[stable(feature = "rust1", since = "1.0.0")]
 123 impl OsStrExt for OsStr {
 124     fn encode_wide(&self) -> EncodeWide<'_> {
 125         self.as_inner().inner.encode_wide()
 126     }
 127 }