Commit Graph

5 Commits

Author SHA1 Message Date
gsxdsm
75213bbb08 fix(i18n): address PR review feedback
- config: explicit Hans script wins over HK/MO region (zh-Hans-HK -> zh-CN),
  with regression tests for script-vs-region precedence
- dashboard i18n: detection.caches [] so the detector's init-time auto-persist
  can't masquerade as a user choice and suppress server-settings hydration
- LanguageSelector: role=group (radiogroup conflicted with aria-pressed)
- bin: validate --lang against SUPPORTED_LOCALES (fail loudly, not silent
  fallback); help text clarifies the flag is terminal-UI-only
- cli i18n test: assert a real fr catalog lookup (defaultValue could mask a
  catalog that never loaded); comment the async changeLanguage re-init seam
- assert-locale-chunks: only dirs containing common.json count as locales
- plan doc: merge duplicate Ink 6.8->7.0 risk bullets
- AGENTS.md/solution doc: frontmatter field list + normalizer excerpt synced

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 12:29:02 -07:00
gsxdsm
4f6013ff3a test+refactor(i18n): close code-review residuals
- Extract a shared normalizeToSupportedLocale helper in @fusion/i18n used by
  both the CLI env detection and the dashboard navigator detection
  (convertDetectedLanguage), fixing multi-subtag navigator tags (zh-Hans-CN)
  and unifying Chinese script resolution.
- Add a cross-tab storage listener to useLanguage so a language change in one
  tab propagates to others.
- Dedup the namespace lists into packages/i18n/namespaces.json, consumed by
  config.ts and all three build scripts (no more 3-way drift risk).
- Add tests: dashboard i18n/index.ts (document.lang mirror, init non-blocking),
  useLanguage concurrent-hydration race + cross-tab adoption,
  normalizeToSupportedLocale, namespace-source consistency.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 09:03:16 -07:00
gsxdsm
f707b62155 fix(i18n): address code-review findings
- Validate GlobalSettings.language at the write boundary (store.ts) via
  validateLocale, so invalid locales are dropped not persisted (api-contract P1).
- Fix detectEnvLocale: Traditional-script env tags (zh_Hant/zh_Hant_TW/zh_HK/
  zh_MO) now resolve to zh-TW instead of Simplified; use replaceAll for
  multi-underscore POSIX tags (adversarial P2). Add coverage.
- Agent-native parity: add 'language' to the CLI settings allowlist
  (VALID_SETTINGS + GLOBAL_ONLY + enum) so 'fn settings set language' works
  like the dashboard switcher.
- Document --lang in the bin.ts help table (api-contract P3).
- Derive the expected locale-chunk count from the locales dir instead of a
  hardcoded 5 (maintainability), and add prebuild:client so a raw vite build
  on a fresh clone still syncs catalogs (adversarial P2).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 08:51:41 -07:00
gsxdsm
959557ade8 chore(i18n): satisfy lint on node scripts; mark plan completed
Add /* global */ directives to the i18n node scripts (matching the existing
prepare-publish-manifest.mjs convention) and flip the plan status to completed.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 08:35:49 -07:00
gsxdsm
cd71b92930 feat(dashboard): add i18n runtime with per-locale code-splitting (U3)
Initialize the browser i18next instance with lazy per-locale catalog loading,
localStorage->navigator->htmlTag detection, script-aware zh fallback, and the
shared @fusion/i18n config. Catalogs are synced from @fusion/i18n into a
gitignored app/locales/ (predev/prebuild) and imported app-relative so Vite
emits one chunk per locale/namespace (verified: 15 chunks, none inlined into
the main bundle). First paint is gated on i18nReady to avoid raw-key flashes;
<I18nextProvider> wraps the App provider stack; vendor-i18n manualChunk added.

A build-assertion script (verify:locale-chunks) guards the KTD3a splitting
invariant. Fallback/namespace behavior is covered by @fusion/i18n config tests;
the live instance is verified via the build assertion rather than a unit test
(the dynamic catalog backend is impractical to exercise in vitest).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 08:16:08 -07:00