i18n: small refactor, add decimal stuff

Cargo.lock

@@ -990,6 +990,17 @@ dependencies = [
  "windows-sys 0.59.0",
 ]
 
+[[package]]
+name = "fixed_decimal"
+version = "0.7.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "35943d22b2f19c0cb198ecf915910a8158e94541c89dcc63300d7799d46c2c5e"
+dependencies = [
+ "displaydoc",
+ "smallvec",
+ "writeable",
+]
+
 [[package]]
 name = "flate2"
 version = "1.1.2"
@@ -1290,6 +1301,29 @@ dependencies = [
  "zerovec",
 ]
 
+[[package]]
+name = "icu_decimal"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fec61c43fdc4e368a9f450272833123a8ef0d7083a44597660ce94d791b8a2e2"
+dependencies = [
+ "displaydoc",
+ "fixed_decimal",
+ "icu_decimal_data",
+ "icu_locale",
+ "icu_locale_core",
+ "icu_provider",
+ "tinystr",
+ "writeable",
+ "zerovec",
+]
+
+[[package]]
+name = "icu_decimal_data"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b70963bc35f9bdf1bc66a5c1f458f4991c1dc71760e00fa06016b2c76b2738d5"
+
 [[package]]
 name = "icu_locale"
 version = "2.0.0"
@@ -3946,7 +3980,9 @@ dependencies = [
  "glob",
  "hex",
  "icu_collator",
+ "icu_decimal",
  "icu_locale",
+ "icu_provider",
  "itertools 0.14.0",
  "libc",
  "md-5",

Cargo.toml

@@ -314,7 +314,9 @@ glob = "0.3.1"
 half = "2.4.1"
 hostname = "0.4"
 icu_collator = "2.0.0"
+icu_decimal = "2.0.0"
 icu_locale = "2.0.0"
+icu_provider = "2.0.0"
 indicatif = "0.18.0"
 itertools = "0.14.0"
 jiff = { version = "0.2.10", default-features = false, features = [

src/uucore/Cargo.toml

@@ -27,10 +27,6 @@ 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",
@@ -59,6 +55,16 @@ bigdecimal = { workspace = true, optional = true }
 num-traits = { workspace = true, optional = true }
 selinux = { workspace = true, optional = true }
 
+# icu stuff
+icu_collator = { workspace = true, optional = true, features = [
+  "compiled_data",
+] }
+icu_decimal = { workspace = true, optional = true, features = [
+  "compiled_data",
+] }
+icu_locale = { workspace = true, optional = true, features = ["compiled_data"] }
+icu_provider = { workspace = true, optional = true }
+
 # Fluent dependencies
 fluent = { workspace = true }
 fluent-syntax = { workspace = true }
@@ -108,7 +114,9 @@ format = [
   "num-traits",
   "quoting-style",
 ]
-i18n = ["icu_collator", "icu_locale"]
+i18n-all = ["i18n-decimal"]
+i18n-common = ["icu_locale", "icu_provider"]
+i18n-decimal = ["i18n-common", "icu_decimal", "icu_locale", "icu_provider"]
 mode = ["libc"]
 perms = ["entries", "libc", "walkdir"]
 buf-copy = []
@@ -116,7 +124,7 @@ parser = ["extendedbigdecimal", "glob", "num-traits"]
 pipes = []
 process = ["libc"]
 proc-info = ["tty", "walkdir"]
-quoting-style = ["i18n"]
+quoting-style = ["i18n-common"]
 ranges = []
 ringbuffer = []
 selinux = ["dep:selinux"]

src/uucore/src/lib/features.rs

@@ -26,7 +26,7 @@ pub mod format;
 pub mod fs;
 #[cfg(feature = "fsext")]
 pub mod fsext;
-#[cfg(feature = "i18n")]
+#[cfg(feature = "i18n-common")]
 pub mod i18n;
 #[cfg(feature = "lines")]
 pub mod lines;

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

@@ -0,0 +1,51 @@
+// This file is part of the uutils coreutils package.
+//
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use std::sync::OnceLock;
+
+use icu_decimal::provider::DecimalSymbolsV1;
+use icu_locale::Locale;
+use icu_provider::prelude::*;
+
+use crate::i18n::get_numeric_locale;
+
+/// Return the decimal separator for the given locale
+fn get_decimal_separator(loc: Locale) -> String {
+    let data_locale = DataLocale::from(loc);
+
+    let request = DataRequest {
+        id: DataIdentifierBorrowed::for_locale(&data_locale),
+        metadata: DataRequestMetadata::default(),
+    };
+
+    let response: DataResponse<DecimalSymbolsV1> =
+        icu_decimal::provider::Baked.load(request).unwrap();
+
+    response.payload.get().decimal_separator().to_string()
+}
+
+/// Return the decimal separator from the language we're working with.
+/// Example:
+///  Say we need to format 1000.5
+///     en_US: 1,000.5 -> decimal separator is '.'
+///     fr_FR: 1 000,5 -> decimal separator is ','
+pub fn locale_decimal_separator() -> &'static str {
+    static DECIMAL_SEP: OnceLock<String> = OnceLock::new();
+
+    DECIMAL_SEP.get_or_init(|| get_decimal_separator(get_numeric_locale().0.clone()))
+}
+
+#[cfg(test)]
+mod tests {
+    use icu_locale::locale;
+
+    use super::get_decimal_separator;
+
+    #[test]
+    fn test_simple_separator() {
+        assert_eq!(get_decimal_separator(locale!("en")), ".");
+        assert_eq!(get_decimal_separator(locale!("fr")), ",");
+    }
+}

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

@@ -1,7 +1,15 @@
+// This file is part of the uutils coreutils package.
+//
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
 use std::sync::OnceLock;
 
 use icu_locale::{Locale, locale};
 
+#[cfg(feature = "i18n-decimal")]
+pub mod decimal;
+
 /// 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)]
@@ -12,48 +20,59 @@ pub enum UEncoding {
 
 const DEFAULT_LOCALE: Locale = locale!("en-US-posix");
 
-/// Deduce the locale from the current environment
+/// Look at 3 environment variables in the following order
+///
+/// 1. LC_ALL
+/// 2. `locale_name`
+/// 3. LANG
+///
+/// Or fallback on Posix locale, with ASCII encoding.
+fn get_locale_from_env(locale_name: &str) -> (Locale, UEncoding) {
+    let locale_var = ["LC_ALL", locale_name, "LANG"]
+        .iter()
+        .find_map(|&key| std::env::var(key).ok());
+
+    if let Some(locale_var_str) = locale_var {
+        let mut split = locale_var_str.split(&['.', '@']);
+
+        if let Some(simple) = split.next() {
+            // Naively convert the locale name to BCP47 tag format.
+            //
+            // See https://en.wikipedia.org/wiki/IETF_language_tag
+            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| enc.to_lowercase() == "utf-8")
+            {
+                UEncoding::Utf8
+            } else {
+                UEncoding::Ascii
+            };
+            return (locale, encoding);
+        }
+    }
+    // Default POSIX locale representing LC_ALL=C
+    (DEFAULT_LOCALE, UEncoding::Ascii)
+}
+
+/// Get the collating locale from the 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.
+    COLLATING_LOCALE.get_or_init(|| {get_locale_from_env("LC_COLLATE")})
+}
 
-        let locale_var = std::env::var("LC_ALL")
-            .or_else(|_| std::env::var("LC_COLLATE"))
-            .or_else(|_| std::env::var("LANG"));
+/// Get the numeric locale from the environment
+pub fn get_numeric_locale() -> &'static (Locale, UEncoding) {
+    static NUMERIC_LOCALE: OnceLock<(Locale, UEncoding)> = OnceLock::new();
 
-        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)
-    })
+    NUMERIC_LOCALE.get_or_init(|| {get_locale_from_env("LC_NUMERIC")})
 }
 
 /// Return the encoding deduced from the locale environment variable.

src/uucore/src/lib/lib.rs

@@ -51,7 +51,7 @@ pub use crate::features::fast_inc;
 pub use crate::features::format;
 #[cfg(feature = "fs")]
 pub use crate::features::fs;
-#[cfg(feature = "i18n")]
+#[cfg(feature = "i18n-common")]
 pub use crate::features::i18n;
 #[cfg(feature = "lines")]
 pub use crate::features::lines;