Fix uutils UTF-8 locale handling on Redox

This commit is contained in:
Ziwen Li(Frank) 2026-05-20 09:45:15 +12:00
parent 8a28ce8da2
commit 30fd08a340

View File

@ -48,6 +48,66 @@ index fd1f30303..c508f6b9b 100644
#[cfg(test)] #[cfg(test)]
mod tests { mod tests {
// Note this useful idiom: importing names from outer (for mod tests) scope. // Note this useful idiom: importing names from outer (for mod tests) scope.
diff --git a/src/uucore/src/lib/features/i18n/mod.rs b/src/uucore/src/lib/features/i18n/mod.rs
index 282baf2e7..a38ccb30e 100644
--- a/src/uucore/src/lib/features/i18n/mod.rs
+++ b/src/uucore/src/lib/features/i18n/mod.rs
@@ -45,8 +45,21 @@ pub fn get_locale_from_env(locale_name: &str) -> (Locale, UEncoding) {
if let Some(simple) = split.next() {
// Handle explicit C and POSIX locales - these should always use byte comparison
+ let encoding = split.next();
+ let is_utf8 = encoding.is_some_and(|enc| {
+ let lower = enc.to_lowercase();
+ lower == "utf-8" || lower == "utf8"
+ });
+
if simple == "C" || simple == "POSIX" {
- return (DEFAULT_LOCALE, UEncoding::Ascii);
+ return (
+ DEFAULT_LOCALE,
+ if is_utf8 {
+ UEncoding::Utf8
+ } else {
+ UEncoding::Ascii
+ },
+ );
}
// Naively convert the locale name to BCP47 tag format.
@@ -55,14 +68,9 @@ pub fn get_locale_from_env(locale_name: &str) -> (Locale, UEncoding) {
let bcp47 = simple.replace('_', "-");
let locale = Locale::try_from_str(&bcp47).unwrap_or(DEFAULT_LOCALE);
- // If locale parsing failed, parse the encoding part of the
- // locale. Treat the special case of the given locale being "C"
- // which becomes the default locale.
- let encoding = if (locale != DEFAULT_LOCALE || bcp47 == "C")
- && split.next().is_some_and(|enc| {
- let lower = enc.to_lowercase();
- lower == "utf-8" || lower == "utf8"
- }) {
+ // If the locale name was parsed successfully, use the encoding suffix
+ // to decide whether filenames should be treated as UTF-8.
+ let encoding = if (locale != DEFAULT_LOCALE || bcp47 == "C") && is_utf8 {
UEncoding::Utf8
} else {
UEncoding::Ascii
@@ -70,8 +78,12 @@ pub fn get_locale_from_env(locale_name: &str) -> (Locale, UEncoding) {
return (locale, encoding);
}
}
- // Default POSIX locale representing LC_ALL=C
- (DEFAULT_LOCALE, UEncoding::Ascii)
+
+ #[cfg(target_os = "redox")]
+ return (DEFAULT_LOCALE, UEncoding::Utf8);
+
+ #[cfg(not(target_os = "redox"))]
+ return (DEFAULT_LOCALE, UEncoding::Ascii);
}
/// Get the collating locale from the environment
diff --git a/src/uucore/src/lib/mods/locale.rs b/src/uucore/src/lib/mods/locale.rs diff --git a/src/uucore/src/lib/mods/locale.rs b/src/uucore/src/lib/mods/locale.rs
index b670f8976..a4ff9f983 100644 index b670f8976..a4ff9f983 100644
--- a/src/uucore/src/lib/mods/locale.rs --- a/src/uucore/src/lib/mods/locale.rs