Rollup merge of rust-lang#58136 - abonander:doc-win-stdio-unicode, r=…

…dtolnay

Improve error message and docs for non-UTF-8 bytes in stdio on Windows

This should make debugging problems like abonander/multipart#106 significantly more straightforward in the future.

cc rust-lang#23344, @retep998 @alexcrichton

Not sure who do r? so I'll let rust-highfive pick one.

src/libstd/io/stdio.rs

@@ -131,6 +131,11 @@ fn handle_ebadf<T>(r: io::Result<T>, default: T) -> io::Result<T> {
 ///
 /// [`io::stdin`]: fn.stdin.html
 /// [`BufRead`]: trait.BufRead.html
+///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to read bytes that are not valid UTF-8 will return
+/// an error.
 #[stable(feature = "rust1", since = "1.0.0")]
 pub struct Stdin {
     inner: Arc<Mutex<BufReader<Maybe<StdinRaw>>>>,
@@ -144,6 +149,11 @@ pub struct Stdin {
 /// [`Read`]: trait.Read.html
 /// [`BufRead`]: trait.BufRead.html
 /// [`Stdin::lock`]: struct.Stdin.html#method.lock
+///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to read bytes that are not valid UTF-8 will return
+/// an error.
 #[stable(feature = "rust1", since = "1.0.0")]
 pub struct StdinLock<'a> {
     inner: MutexGuard<'a, BufReader<Maybe<StdinRaw>>>,
@@ -157,6 +167,11 @@ pub struct StdinLock<'a> {
 ///
 /// [lock]: struct.Stdin.html#method.lock
 ///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to read bytes that are not valid UTF-8 will return
+/// an error.
+///
 /// # Examples
 ///
 /// Using implicit synchronization:
@@ -328,6 +343,11 @@ impl<'a> fmt::Debug for StdinLock<'a> {
 ///
 /// Created by the [`io::stdout`] method.
 ///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
+///
 /// [`lock`]: #method.lock
 /// [`io::stdout`]: fn.stdout.html
 #[stable(feature = "rust1", since = "1.0.0")]
@@ -343,6 +363,11 @@ pub struct Stdout {
 /// This handle implements the [`Write`] trait, and is constructed via
 /// the [`Stdout::lock`] method.
 ///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
+///
 /// [`Write`]: trait.Write.html
 /// [`Stdout::lock`]: struct.Stdout.html#method.lock
 #[stable(feature = "rust1", since = "1.0.0")]
@@ -358,6 +383,11 @@ pub struct StdoutLock<'a> {
 ///
 /// [Stdout::lock]: struct.Stdout.html#method.lock
 ///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
+///
 /// # Examples
 ///
 /// Using implicit synchronization:
@@ -476,6 +506,11 @@ impl<'a> fmt::Debug for StdoutLock<'a> {
 /// For more information, see the [`io::stderr`] method.
 ///
 /// [`io::stderr`]: fn.stderr.html
+///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
 #[stable(feature = "rust1", since = "1.0.0")]
 pub struct Stderr {
     inner: Arc<ReentrantMutex<RefCell<Maybe<StderrRaw>>>>,
@@ -487,6 +522,11 @@ pub struct Stderr {
 /// the [`Stderr::lock`] method.
 ///
 /// [`Stderr::lock`]: struct.Stderr.html#method.lock
+///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
 #[stable(feature = "rust1", since = "1.0.0")]
 pub struct StderrLock<'a> {
     inner: ReentrantMutexGuard<'a, RefCell<Maybe<StderrRaw>>>,
@@ -496,6 +536,11 @@ pub struct StderrLock<'a> {
 ///
 /// This handle is not buffered.
 ///
+/// ### Note: Windows Portability Consideration
+/// When operating in a console, the Windows implementation of this stream does not support
+/// non-UTF-8 byte sequences. Attempting to write bytes that are not valid UTF-8 will return
+/// an error.
+///
 /// # Examples
 ///
 /// Using implicit synchronization:

src/libstd/sys/windows/stdio.rs

@@ -188,7 +188,9 @@ impl Output {
 }
 
 fn invalid_encoding() -> io::Error {
-    io::Error::new(io::ErrorKind::InvalidData, "text was not valid unicode")
+    io::Error::new(io::ErrorKind::InvalidData,
+                   "Windows stdio in console mode does not support non-UTF-8 byte sequences; \
+                    see https://github.com/rust-lang/rust/issues/23344")
 }
 
 fn readconsole_input_control(wakeup_mask: c::ULONG) -> c::CONSOLE_READCONSOLE_CONTROL {