i18n: Add charset related settings (#408307)

Author: doronbehar

doc/release-notes/rl-2505.section.md

@@ -374,6 +374,7 @@
   - `i18n.extraLocales` should now be the preferred way to install additional locales.
   - `i18n.supportedLocales` is now considered an implementation detail and will be hidden from the documentation. But the option will still continue to work.
   - `i18n.supportedLocales` will now trigger a warning when it omits any locale set in `i18n.defaultLocale`, `i18n.extraLocales` or `i18n.extraLocaleSettings`.
+  - The options `i18n.defaultCharset` & `i18n.localeCharsets` were added, and they complement `i18n.defaultLocale` & `i18n.extraLocaleSettings` respectively - allowing to control the character set used per locale setting.
 
 - `titaniumenv`, `titanium`, and `titanium-alloy` have been removed due to lack of maintenance in Nixpkgs []{#sec-nixpkgs-release-25.05-incompatibilities-titanium-removed}.
 

nixos/modules/config/i18n.nix

@@ -5,15 +5,19 @@
   ...
 }:
 let
+  sanitizeUTF8Capitalization =
+    lang: (lib.replaceStrings [ "utf8" "utf-8" "UTF8" ] [ "UTF-8" "UTF-8" "UTF-8" ] lang);
   aggregatedLocales =
-    (builtins.map
-      (l: (lib.replaceStrings [ "utf8" "utf-8" "UTF8" ] [ "UTF-8" "UTF-8" "UTF-8" ] l) + "/UTF-8")
-      (
-        [ config.i18n.defaultLocale ]
-        ++ (lib.optionals (builtins.isList config.i18n.extraLocales) config.i18n.extraLocales)
-        ++ (lib.attrValues (lib.filterAttrs (n: v: n != "LANGUAGE") config.i18n.extraLocaleSettings))
-      )
-    )
+    [
+      "${config.i18n.defaultLocale}/${config.i18n.defaultCharset}"
+    ]
+    ++ lib.pipe config.i18n.extraLocaleSettings [
+      (lib.mapAttrs (n: v: (sanitizeUTF8Capitalization v)))
+      (lib.mapAttrsToList (LCRole: lang: lang + "/" + (config.i18n.localeCharsets.${LCRole} or "UTF-8")))
+    ]
+    ++ (builtins.map sanitizeUTF8Capitalization (
+      lib.optionals (builtins.isList config.i18n.extraLocales) config.i18n.extraLocales
+    ))
     ++ (lib.optional (builtins.isString config.i18n.extraLocales) config.i18n.extraLocales);
 in
 {
@@ -48,16 +52,24 @@ in
         default = "en_US.UTF-8";
         example = "nl_NL.UTF-8";
         description = ''
-          The default locale.  It determines the language for program
-          messages, the format for dates and times, sort order, and so on.
-          It also determines the character set, such as UTF-8.
+          The default locale. It determines the language for program messages,
+          the format for dates and times, sort order, and so on. Setting the
+          default character set is done via {option}`i18n.defaultCharset`.
+        '';
+      };
+      defaultCharset = lib.mkOption {
+        type = lib.types.str;
+        default = "UTF-8";
+        example = "ISO-8859-8";
+        description = ''
+          The default locale character set.
         '';
       };
 
       extraLocales = lib.mkOption {
         type = lib.types.either (lib.types.listOf lib.types.str) (lib.types.enum [ "all" ]);
         default = [ ];
-        example = [ "nl_NL.UTF-8" ];
+        example = [ "nl_NL.UTF-8/UTF-8" ];
         description = ''
           Additional locales that the system should support, besides the ones
           configured with {option}`i18n.defaultLocale` and
@@ -74,9 +86,24 @@ in
           LC_TIME = "de_DE.UTF-8";
         };
         description = ''
-          A set of additional system-wide locale settings other than
-          `LANG` which can be configured with
-          {option}`i18n.defaultLocale`.
+          A set of additional system-wide locale settings other than `LANG`
+          which can be configured with {option}`i18n.defaultLocale`. Note that
+          the `/UTF-8` suffix used in {option}`i18n.extraLocales` indicates a
+          character set, and it must not be added manually here. To use a
+          non-`UTF-8` character set such as ISO-XXXX-8, the
+          {option}`i18n.localeCharsets` can be used.
+        '';
+      };
+      localeCharsets = lib.mkOption {
+        type = lib.types.attrsOf lib.types.str;
+        default = { };
+        example = {
+          LC_MESSAGES = "ISO-8859-15";
+          LC_TIME = "ISO-8859-1";
+        };
+        description = ''
+          Per each {option}`i18n.extraLocaleSettings`, choose the character set
+          to use for it. Essentially defaults to UTF-8 for all of them.
         '';
       };
 

nixos/tests/all-tests.nix

@@ -994,6 +994,7 @@ in
   orthanc = runTest ./orthanc.nix;
   owncast = handleTest ./owncast.nix { };
   outline = handleTest ./outline.nix { };
+  i18n = runTest ./i18n.nix;
   image-contents = handleTest ./image-contents.nix { };
   openvscode-server = handleTest ./openvscode-server.nix { };
   open-webui = runTest ./open-webui.nix;

nixos/tests/i18n.nix

@@ -0,0 +1,43 @@
+{ lib, ... }:
+{
+  name = "glibLocales-custom-builds";
+  meta.maintainers = with lib.maintainers; [ doronbehar ];
+
+  nodes = {
+    nonUTF8Charset = {
+      i18n = {
+        defaultLocale = "en_US";
+        defaultCharset = "ISO-8859-1";
+      };
+    };
+    extraLocales1 = {
+      i18n = {
+        defaultLocale = "en_US.UTF-8";
+        extraLocales = [
+          "nl_NL.UTF-8/UTF-8"
+        ];
+      };
+    };
+    extraLocaleSettings = {
+      i18n = {
+        defaultLocale = "en_US.UTF-8";
+        extraLocaleSettings = {
+          LC_MESSAGES = "en_US.UTF-8";
+          LC_TIME = "de_DE.UTF-8";
+        };
+      };
+    };
+    localeCharsets = {
+      i18n = {
+        defaultLocale = "en_US.UTF-8";
+        extraLocaleSettings = {
+          LC_TIME = "de_DE";
+        };
+        localeCharsets = {
+          LC_TIME = "ISO-8859-1";
+        };
+      };
+    };
+  };
+  testScript = { nodes, ... }: "";
+}