fix(i18n): prevent Traditional Chinese locales from falling back to S…

.changeset/fix-chinese-locales-fallback.md

@@ -0,0 +1,7 @@
+---
+'@rocket.chat/meteor': patch
+'@rocket.chat/i18n': patch
+'@rocket.chat/tools': patch
+---
+
+Fix regression where Traditional Chinese (zh-TW, zh-HK) locales resolve to Simplified Chinese (zh) resources and browser language detection collapses Traditional Chinese variants into Simplified Chinese.

apps/meteor/client/providers/TranslationProvider.tsx

@@ -6,6 +6,7 @@ import {
 	availableTranslationNamespaces,
 	defaultTranslationNamespace,
 	extractTranslationNamespaces,
+	defaultFallbackLng,
 } from '@rocket.chat/i18n';
 import languages from '@rocket.chat/i18n/dist/languages';
 import en from '@rocket.chat/i18n/dist/resources/en.i18n.json';
@@ -72,7 +73,8 @@ const useI18next = (lng: string): typeof i18next => {
 		isI18nInitialized = true;
 		i18n.init({
 			lng,
-			fallbackLng: 'en',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
 			ns: availableTranslationNamespaces,
 			defaultNS: defaultTranslationNamespace,
 			nsSeparator: '.',

apps/meteor/server/lib/i18n.ts

@@ -1,4 +1,9 @@
-import { availableTranslationNamespaces, defaultTranslationNamespace, extractTranslationNamespaces } from '@rocket.chat/i18n';
+import {
+	availableTranslationNamespaces,
+	defaultFallbackLng,
+	defaultTranslationNamespace,
+	extractTranslationNamespaces,
+} from '@rocket.chat/i18n';
 import languages from '@rocket.chat/i18n/dist/languages';
 
 import { i18n } from '../../app/utils/lib/i18n';
@@ -79,7 +84,8 @@ if (false) {
 
 void i18n.init({
 	lng: 'en',
-	fallbackLng: 'en',
+	fallbackLng: defaultFallbackLng,
+	load: 'currentOnly',
 	defaultNS: defaultTranslationNamespace,
 	ns: availableTranslationNamespaces,
 	nsSeparator: '.',

packages/i18n/src/fallbackLng.spec.ts

@@ -0,0 +1,111 @@
+import i18next from 'i18next';
+
+import { defaultFallbackLng } from './index';
+
+describe('defaultFallbackLng', () => {
+	it('should resolve zh-TW to [zh-TW, zh-HK, en] without zh / Simplified Chinese', () => {
+		const i18n = i18next.createInstance();
+		i18n.init({
+			lng: 'zh-TW',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+
+		expect(i18n.languages).toEqual(['zh-TW', 'zh-HK', 'en']);
+		expect(i18n.languages).not.toContain('zh');
+		expect(i18n.languages).not.toContain('zh-CN');
+	});
+
+	it('should resolve zh-HK to [zh-HK, zh-TW, en] without zh / Simplified Chinese', () => {
+		const i18n = i18next.createInstance();
+		i18n.init({
+			lng: 'zh-HK',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+
+		expect(i18n.languages).toEqual(['zh-HK', 'zh-TW', 'en']);
+		expect(i18n.languages).not.toContain('zh');
+		expect(i18n.languages).not.toContain('zh-CN');
+	});
+
+	it('should resolve zh (Simplified Chinese) to [zh, en]', () => {
+		const i18n = i18next.createInstance();
+		i18n.init({
+			lng: 'zh',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+
+		expect(i18n.languages).toEqual(['zh', 'en']);
+	});
+
+	it('should resolve zh-CN to [zh-CN, zh, en]', () => {
+		const i18n = i18next.createInstance();
+		i18n.init({
+			lng: 'zh-CN',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+
+		expect(i18n.languages).toEqual(['zh-CN', 'zh', 'en']);
+	});
+
+	it('should resolve regional variants like pt-BR and de-AT appropriately', () => {
+		const i18nPt = i18next.createInstance();
+		i18nPt.init({
+			lng: 'pt-BR',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+		expect(i18nPt.languages).toEqual(['pt-BR', 'pt', 'en']);
+
+		const i18nDe = i18next.createInstance();
+		i18nDe.init({
+			lng: 'de-AT',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+		});
+		expect(i18nDe.languages).toEqual(['de-AT', 'de', 'en']);
+	});
+
+	it('should resolve translation keys from zh-TW and fall back to en when missing in zh-TW', () => {
+		const i18n = i18next.createInstance();
+		i18n.init({
+			lng: 'zh-TW',
+			fallbackLng: defaultFallbackLng,
+			load: 'currentOnly',
+			initImmediate: false,
+			resources: {
+				'zh-TW': {
+					translation: {
+						Users: '使用者',
+					},
+				},
+				'zh': {
+					translation: {
+						Users: '用户',
+						private: '私有',
+					},
+				},
+				'en': {
+					translation: {
+						Users: 'Users',
+						private: 'private',
+					},
+				},
+			},
+		});
+
+		// Present in zh-TW
+		expect(i18n.t('Users')).toBe('使用者');
+		// Missing in zh-TW -> should fall back to en, NOT zh (Simplified Chinese)
+		expect(i18n.t('private')).toBe('private');
+	});
+});

packages/i18n/src/index.ts

@@ -149,3 +149,20 @@ export const applyCustomTranslations = (
 		}
 	}
 };
+
+export const defaultFallbackLng = {
+	'zh-TW': ['zh-HK', 'en'],
+	'zh-HK': ['zh-TW', 'en'],
+	'zh-Hant': ['zh-TW', 'zh-HK', 'en'],
+	'zh-Hans': ['zh', 'en'],
+	'zh-CN': ['zh', 'en'],
+	'zh-SG': ['zh', 'en'],
+	'zh-MO': ['zh-HK', 'zh-TW', 'en'],
+	'pt-BR': ['pt', 'en'],
+	'de-AT': ['de', 'en'],
+	'de-IN': ['de', 'en'],
+	'bn-BD': ['bn-IN', 'en'],
+	'bn-IN': ['bn-BD', 'en'],
+	'hi-IN': ['hi', 'en'],
+	'default': ['en'],
+};

packages/tools/src/normalizeLanguage.spec.ts

@@ -0,0 +1,74 @@
+import { normalizeLanguage } from './normalizeLanguage';
+
+describe('normalizeLanguage', () => {
+	it('should return empty/falsy language as is', () => {
+		expect(normalizeLanguage('')).toBe('');
+	});
+
+	it('should normalize lowercase region codes to uppercase', () => {
+		expect(normalizeLanguage('pt-br')).toBe('pt-BR');
+		expect(normalizeLanguage('en-us')).toBe('en-US');
+		expect(normalizeLanguage('de-at')).toBe('de-AT');
+	});
+
+	it('should support 3-digit UN M49 region codes', () => {
+		expect(normalizeLanguage('es-419')).toBe('es-419');
+	});
+
+	it('should normalize script subtags with title casing rather than treating them as region codes', () => {
+		expect(normalizeLanguage('sr-latn')).toBe('sr-Latn');
+		expect(normalizeLanguage('sr-Latn')).toBe('sr-Latn');
+		expect(normalizeLanguage('sr-LATN')).toBe('sr-Latn');
+		expect(normalizeLanguage('az-latn')).toBe('az-Latn');
+		expect(normalizeLanguage('uz-cyrl')).toBe('uz-Cyrl');
+	});
+
+	it('should handle underscore delimiter', () => {
+		expect(normalizeLanguage('pt_br')).toBe('pt-BR');
+		expect(normalizeLanguage('zh_tw')).toBe('zh-TW');
+		expect(normalizeLanguage('zh_hk')).toBe('zh-HK');
+		expect(normalizeLanguage('zh_cn')).toBe('zh');
+		expect(normalizeLanguage('sr_latn')).toBe('sr-Latn');
+	});
+
+	it('should resolve Traditional Chinese variants to zh-TW', () => {
+		expect(normalizeLanguage('zh-TW')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-tw')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-Hant')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-hant')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-Hant-TW')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-hant-tw')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-CHT')).toBe('zh-TW');
+		expect(normalizeLanguage('zh-cht')).toBe('zh-TW');
+	});
+
+	it('should resolve Hong Kong / Macau Traditional Chinese variants to zh-HK', () => {
+		expect(normalizeLanguage('zh-HK')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-hk')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-Hant-HK')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-hant-hk')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-MO')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-mo')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-Hant-MO')).toBe('zh-HK');
+		expect(normalizeLanguage('zh-hant-mo')).toBe('zh-HK');
+	});
+
+	it('should resolve Simplified Chinese variants to zh', () => {
+		expect(normalizeLanguage('zh')).toBe('zh');
+		expect(normalizeLanguage('zh-CN')).toBe('zh');
+		expect(normalizeLanguage('zh-cn')).toBe('zh');
+		expect(normalizeLanguage('zh-Hans')).toBe('zh');
+		expect(normalizeLanguage('zh-hans')).toBe('zh');
+		expect(normalizeLanguage('zh-Hans-CN')).toBe('zh');
+		expect(normalizeLanguage('zh-hans-cn')).toBe('zh');
+		expect(normalizeLanguage('zh-SG')).toBe('zh');
+		expect(normalizeLanguage('zh-CHS')).toBe('zh');
+	});
+
+	it('should keep other languages unchanged', () => {
+		expect(normalizeLanguage('en')).toBe('en');
+		expect(normalizeLanguage('fr')).toBe('fr');
+		expect(normalizeLanguage('de')).toBe('de');
+		expect(normalizeLanguage('ja')).toBe('ja');
+	});
+});

packages/tools/src/normalizeLanguage.ts

@@ -1,9 +1,31 @@
 export const normalizeLanguage = (language: string): string => {
+	if (!language) {
+		return language;
+	}
+
+	const lang = language.replace(/_/g, '-');
+
+	if (/^zh-(hant-hk|hk|hant-mo|mo)/i.test(lang)) {
+		return 'zh-HK';
+	}
+	if (/^zh-(hant(-tw)?|cht|tw)/i.test(lang)) {
+		return 'zh-TW';
+	}
+	if (/^zh-(hans|cn|sg|chs)/i.test(lang) || lang.toLowerCase() === 'zh') {
+		return 'zh';
+	}
+
 	// Fix browsers having all-lowercase language settings eg. pt-br, en-us
-	const regex = /([a-z]{2,3})-([a-z]{2,4})/;
-	const matches = regex.exec(language);
+	const regex = /^([a-z]{2,3})-([a-z]{2}|[0-9]{3})$/i;
+	const matches = regex.exec(lang);
 	if (matches) {
-		return `${matches[1]}-${matches[2].toUpperCase()}`;
+		return `${matches[1].toLowerCase()}-${matches[2].toUpperCase()}`;
+	}
+
+	const scriptRegex = /^([a-z]{2,3})-([a-z]{4})$/i;
+	const scriptMatches = scriptRegex.exec(lang);
+	if (scriptMatches) {
+		return `${scriptMatches[1].toLowerCase()}-${scriptMatches[2][0].toUpperCase()}${scriptMatches[2].slice(1).toLowerCase()}`;
 	}
 
 	return language;