Rollup merge of #134622 - ChrisDenton:write-file-utf8, r=Mark-Simulacrum

Windows: Use WriteFile to write to a UTF-8 console

If the console code page is UTF-8 then we can simply write to it without needing to convert to UTF-16 and calling `WriteConsole`.

3 files changed, 26 insertions, 1 deletions

diff --git a/library/std/src/sys/pal/windows/c/bindings.txt b/library/std/src/sys/pal/windows/c/bindings.txt
index d98e90eedfe..c06f274685c 100644
--- a/library/std/src/sys/pal/windows/c/bindings.txt
+++ b/library/std/src/sys/pal/windows/c/bindings.txt
@@ -2426,6 +2426,7 @@ Windows.Win32.System.Console.ENABLE_VIRTUAL_TERMINAL_PROCESSING
 Windows.Win32.System.Console.ENABLE_WINDOW_INPUT
 Windows.Win32.System.Console.ENABLE_WRAP_AT_EOL_OUTPUT
 Windows.Win32.System.Console.GetConsoleMode
+Windows.Win32.System.Console.GetConsoleOutputCP
 Windows.Win32.System.Console.GetStdHandle
 Windows.Win32.System.Console.ReadConsoleW
 Windows.Win32.System.Console.STD_ERROR_HANDLE
diff --git a/library/std/src/sys/pal/windows/c/windows_sys.rs b/library/std/src/sys/pal/windows/c/windows_sys.rs
index ed29f3d264c..79513d33a1a 100644
--- a/library/std/src/sys/pal/windows/c/windows_sys.rs
+++ b/library/std/src/sys/pal/windows/c/windows_sys.rs
@@ -34,6 +34,7 @@ windows_targets::link!("kernel32.dll" "system" fn FreeEnvironmentStringsW(penv :
 windows_targets::link!("kernel32.dll" "system" fn GetActiveProcessorCount(groupnumber : u16) -> u32);
 windows_targets::link!("kernel32.dll" "system" fn GetCommandLineW() -> PCWSTR);
 windows_targets::link!("kernel32.dll" "system" fn GetConsoleMode(hconsolehandle : HANDLE, lpmode : *mut CONSOLE_MODE) -> BOOL);
+windows_targets::link!("kernel32.dll" "system" fn GetConsoleOutputCP() -> u32);
 windows_targets::link!("kernel32.dll" "system" fn GetCurrentDirectoryW(nbufferlength : u32, lpbuffer : PWSTR) -> u32);
 windows_targets::link!("kernel32.dll" "system" fn GetCurrentProcess() -> HANDLE);
 windows_targets::link!("kernel32.dll" "system" fn GetCurrentProcessId() -> u32);
@@ -3333,6 +3334,7 @@ pub struct XSAVE_FORMAT {
     pub XmmRegisters: [M128A; 8],
     pub Reserved4: [u8; 224],
 }
+
 #[cfg(target_arch = "arm")]
 #[repr(C)]
 pub struct WSADATA {
diff --git a/library/std/src/sys/pal/windows/stdio.rs b/library/std/src/sys/pal/windows/stdio.rs
index 642c8bc4df7..1b735e7f0cb 100644
--- a/library/std/src/sys/pal/windows/stdio.rs
+++ b/library/std/src/sys/pal/windows/stdio.rs
@@ -84,21 +84,43 @@ fn is_console(handle: c::HANDLE) -> bool {
     unsafe { c::GetConsoleMode(handle, &mut mode) != 0 }
 }
 
+/// Returns true if the attached console's code page is currently UTF-8.
+#[cfg(not(target_vendor = "win7"))]
+fn is_utf8_console() -> bool {
+    unsafe { c::GetConsoleOutputCP() == c::CP_UTF8 }
+}
+
+#[cfg(target_vendor = "win7")]
+fn is_utf8_console() -> bool {
+    // Windows 7 has a fun "feature" where WriteFile on a console handle will return
+    // the number of UTF-16 code units written and not the number of bytes from the input string.
+    // So we always claim the console isn't UTF-8 to trigger the WriteConsole fallback code.
+    false
+}
+
 fn write(handle_id: u32, data: &[u8], incomplete_utf8: &mut IncompleteUtf8) -> io::Result<usize> {
     if data.is_empty() {
         return Ok(0);
     }
 
     let handle = get_handle(handle_id)?;
-    if !is_console(handle) {
+    if !is_console(handle) || is_utf8_console() {
         unsafe {
             let handle = Handle::from_raw_handle(handle);
             let ret = handle.write(data);
             let _ = handle.into_raw_handle(); // Don't close the handle
             return ret;
         }
+    } else {
+        write_console_utf16(data, incomplete_utf8, handle)
     }
+}
 
+fn write_console_utf16(
+    data: &[u8],
+    incomplete_utf8: &mut IncompleteUtf8,
+    handle: c::HANDLE,
+) -> io::Result<usize> {
     if incomplete_utf8.len > 0 {
         assert!(
             incomplete_utf8.len < 4,