feat(i18n): add Russian (ru) locale to Hermes Desktop

[saimon4eg]

Russian localization for Hermes Desktop

Uses the defineLocale() partial-locale pattern (same as ja.ts) — any missing keys fall back to English, so the locale stays healthy as upstream adds new keys.

What is included

• Desktop locale (apps/desktop/src/i18n/ru.ts) — 2654 lines, all sections translated
• Settings fields — fieldLabels and fieldDescriptions populated via defineFieldCopy() (all config keys)
• Web locale (web/src/i18n/ru.ts) — updated Russian translations
• Registration — ru in Locale type, catalog.ts, languages.ts, and locale aliases (ru, ru-ru, ru_ru)

What this fixes from the previous PR

• ✅ fieldLabels and fieldDescriptions now populated (were empty {})
• ✅ Uses defineLocale() partial pattern (was rigid Translations type — breaks on new upstream keys)
• ✅ No install.sh changes — upstream URLs preserved
• ✅ No ACP/IDE changes mixed in
• ✅ No .idea/ files

Structure

apps/desktop/src/i18n/
├── ru.ts          ← new (defineLocale + defineFieldCopy)
├── types.ts       ← ru in Locale
├── catalog.ts     ← ru import + TRANSLATIONS entry
├── languages.ts   ← ru locale option + aliases

web/src/i18n/
└── ru.ts          ← updated

[alt-glitch]

This was generated by AI during triage.

Correction: this PR is not a duplicate. Its broad Desktop Russian locale and separate web translations have concrete scope/completeness differences from the ongoing #41677, #40701, and #50083 cluster. The maintainer sweep marked it keep_open; it remains related and needs a maintainer decision.

[saimon4eg]

Why this PR should be the canonical ru-locale PR

Compared to #40701 (the earliest ru-PR marked as canonical by the triage bot), this PR is objectively more complete and closer to merge-ready:

	#40701	#64929 (this)
Pattern	defineLocale() ✅	defineLocale() ✅
fieldLabels	❌ empty	✅ fully populated
fieldDescriptions	❌ empty	✅ fully populated
Translation keys	963	3,216 (3.3× more)
Web locale	❌	✅
settings.gateway (Cloud mode, 35+ keys)	❌	✅
settings.mcp (catalog, 22 keys)	❌	✅
skills.hub (browsing, 50+ keys)	❌	✅
commandCenter.maintenance (40 keys)	❌	✅
onboarding.fireworks	❌	✅
Bad aliases (be/uk/kk→ru)	❌	✅ only ru/ru-ru/ru_ru

This PR is a superset: every key from #40701 is also present here, plus 2,200+ additional translations covering all upstream sections added since #40701 was opened.

Recent update

Just pushed a comprehensive update adding all 249 missing upstream keys (commandCenter, gateway Cloud, MCP, skills hub, updates, etc.). The locale now covers every section of the Hermes Desktop UI.

Merci!

[teknium1]

Thanks for implementing the Desktop locale with the existing partial-locale pattern. Current origin/main still lacks ru in the Desktop catalog (apps/desktop/src/i18n/catalog.ts:7-12), locale union (apps/desktop/src/i18n/types.ts:8), and alias map (apps/desktop/src/i18n/languages.ts:42-68). The PR wires all of those paths and defines the locale with defineLocale() plus defineFieldCopy() (apps/desktop/src/i18n/ru.ts:1-4, :464-616).

Problems

• No test file is changed. The existing resolver contract covers aliases, exact locale IDs, and config persistence in apps/desktop/src/i18n/languages.test.ts:6-42; add equivalent ru coverage so the new registration cannot silently regress.

Suggested changes

• Test ru, ru-RU, and ru_ru normalization; test exact/support behavior and localeConfigValue('ru').
• Add one runtime catalog assertion for a Russian translated string, following apps/desktop/src/i18n/runtime.test.ts:18-54.

Automated hermes-sweeper review.

[antonbru]

Suggested tests for ru locale registration

Per @teknium1's review, here are the missing tests for apps/desktop/src/i18n/languages.test.ts.

Changes to the existing test suite

Additions to apps/desktop/src/i18n/languages.test.ts:

1. normalizeLocale — Russian aliases (in the existing "normalizes supported locale aliases" test):

expect(normalizeLocale('ru')).toBe('ru')
expect(normalizeLocale('ru-RU')).toBe('ru')
expect(normalizeLocale('ru_ru')).toBe('ru')

2. isSupportedLocaleValue / isLocale — Russian (in the existing "distinguishes exact locale ids" test):

expect(isSupportedLocaleValue('ru-RU')).toBe(true)
expect(isLocale('ru-RU')).toBe(false)
expect(isLocale('ru')).toBe(true)

3. localeConfigValue — Russian (in the existing "returns the persisted config value" test):

expect(localeConfigValue('ru')).toBe('ru')

4. New test: runtime catalog assertion — verifies the TRANSLATIONS['ru'] entry is wired through catalog.ts and produces actual translated strings:

import { TRANSLATIONS } from './catalog'

it('provides Russian translations for common keys', () => {
  const ru = TRANSLATIONS['ru']
  expect(ru.common.apply).toBe('Применить')
  expect(ru.common.cancel).toBe('Отмена')
  expect(ru.common.save).toBe('Сохранить')
  expect(ru.common.delete).toBe('Удалить')
  expect(ru.common.send).toBe('Отправить')
  expect(ru.common.close).toBe('Закрыть')
  expect(ru.common.loading).toBe('Загрузка...')
  expect(ru.common.error).toBe('Ошибка')
})

it('provides Russian translations for desktop-specific keys', () => {
  const ru = TRANSLATIONS['ru']
  expect(ru.boot.ready).toBe('Hermes Desktop готов')
  expect(ru.desktop.createSessionFailed).toBe('Не удалось создать сессию')
  expect(ru.desktop.promptFailed).toBe('Не удалось отправить запрос')
})

Happy to open a PR against this branch if that makes it easier to land the review feedback.

[saimon4eg]

Superseded by #70673 — clean rebase on fresh upstream main.