uucore: retrieve the Locale encoding from env variables

Cargo.toml

@@ -1,7 +1,7 @@
 # coreutils (uutils)
 # * see the repository LICENSE, README, and CONTRIBUTING files for more information
 
-# spell-checker:ignore (libs) bigdecimal datetime serde bincode gethostid kqueue libselinux mangen memmap uuhelp startswith constness expl
+# spell-checker:ignore (libs) bigdecimal datetime serde bincode gethostid kqueue libselinux mangen memmap uuhelp startswith constness expl unnested
 
 [package]
 name = "coreutils"
@@ -313,6 +313,8 @@ gcd = "2.3"
 glob = "0.3.1"
 half = "2.4.1"
 hostname = "0.4"
+icu_collator = "2.0.0"
+icu_locale = "2.0.0"
 indicatif = "0.17.8"
 itertools = "0.14.0"
 jiff = { version = "0.2.10", default-features = false, features = [

src/uucore/Cargo.toml

@@ -27,6 +27,10 @@ dns-lookup = { workspace = true, optional = true }
 dunce = { version = "1.0.4", optional = true }
 wild = "2.2.1"
 glob = { workspace = true, optional = true }
+icu_collator = { workspace = true, optional = true, features = [
+  "compiled_data",
+] }
+icu_locale = { workspace = true, optional = true, features = ["compiled_data"] }
 itertools = { workspace = true, optional = true }
 time = { workspace = true, optional = true, features = [
   "formatting",
@@ -106,6 +110,7 @@ format = [
   "num-traits",
   "quoting-style",
 ]
+i18n = ["icu_collator", "icu_locale"]
 mode = ["libc"]
 perms = ["entries", "libc", "walkdir"]
 buf-copy = []
@@ -113,7 +118,7 @@ parser = ["extendedbigdecimal", "glob", "num-traits"]
 pipes = []
 process = ["libc"]
 proc-info = ["tty", "walkdir"]
-quoting-style = []
+quoting-style = ["i18n"]
 ranges = []
 ringbuffer = []
 selinux = ["dep:selinux"]

src/uucore/src/lib/features.rs

@@ -26,6 +26,8 @@ pub mod format;
 pub mod fs;
 #[cfg(feature = "fsext")]
 pub mod fsext;
+#[cfg(feature = "i18n")]
+pub mod i18n;
 #[cfg(feature = "lines")]
 pub mod lines;
 #[cfg(feature = "parser")]

src/uucore/src/lib/features/i18n/mod.rs

@@ -0,0 +1,62 @@
+use std::sync::OnceLock;
+
+use icu_locale::{Locale, locale};
+
+/// The encoding specified by the locale, if specified
+/// Currently only supports ASCII and UTF-8 for the sake of simplicity.
+#[derive(Debug, PartialEq, Eq, Clone, Copy)]
+pub enum UEncoding {
+    Ascii,
+    Utf8,
+}
+
+const DEFAULT_LOCALE: Locale = locale!("en-US-posix");
+
+/// Deduce the locale from the current environment
+fn get_collating_locale() -> &'static (Locale, UEncoding) {
+    static COLLATING_LOCALE: OnceLock<(Locale, UEncoding)> = OnceLock::new();
+
+    COLLATING_LOCALE.get_or_init(|| {
+        // Look at 3 environment variables in the following order
+        //
+        // 1. LC_ALL
+        // 2. LC_COLLATE
+        // 3. LANG
+        //
+        // Or fallback on Posix locale, with ASCII encoding.
+
+        let locale_var = std::env::var("LC_ALL")
+            .or_else(|_| std::env::var("LC_COLLATE"))
+            .or_else(|_| std::env::var("LANG"));
+
+        if let Ok(locale_var_str) = locale_var {
+            let mut split = locale_var_str.split(&['.', '@']);
+
+            if let Some(simple) = split.next() {
+                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() == Some("UTF-8")
+                {
+                    UEncoding::Utf8
+                } else {
+                    UEncoding::Ascii
+                };
+                return (locale, encoding);
+            } else {
+                return (DEFAULT_LOCALE, UEncoding::Ascii);
+            };
+        }
+        // Default POSIX locale representing LC_ALL=C
+        (DEFAULT_LOCALE, UEncoding::Ascii)
+    })
+}
+
+/// Return the encoding deduced from the locale environment variable.
+pub fn get_locale_encoding() -> UEncoding {
+    get_collating_locale().1
+}

src/uucore/src/lib/lib.rs

@@ -51,6 +51,8 @@ pub use crate::features::fast_inc;
 pub use crate::features::format;
 #[cfg(feature = "fs")]
 pub use crate::features::fs;
+#[cfg(feature = "i18n")]
+pub use crate::features::i18n;
 #[cfg(feature = "lines")]
 pub use crate::features::lines;
 #[cfg(feature = "parser")]