feat(i18n): port check-keys.py + add phase 6 unit tests - #327
Merged
Conversation
Cross-repo backfill from niac-go#723 (check-keys.py) and niac-go#724
(Phase 6 test coverage). Brings stem to parity with niac.
What lands:
- scripts/i18n/check-keys.py — read-only static analysis that
cross-references t() calls in ui/src against committed EN locale
JSON. Catches typo'd key names and dead JSON keys. Identical to
the niac script; added --ratchet flag so newly-added repos can
absorb it without an immediate cleanup burden (stem currently has
48 t() calls referencing non-existent keys — pre-existing bugs
masked by i18next's fallback semantics, queued for cleanup).
- scripts/i18n/validate.sh — pulled in the niac improvements:
(a) check_key_usage step delegating to check-keys.py with ratchet
propagation
(b) word-boundary \\bt\\( regex fix for the fallback-pattern check
so identifiers ending in `t` (headers.set('Accept',
'application/json')) don't false-match
- ui/vitest.config.ts — adds @locales path alias so i18n/index.ts
can be imported under test, plus dedupe entries for i18next/
react-i18next/react to encourage module deduplication.
- ui/src/hooks/useLocale.test.ts — 3 unit tests for the BCP-47
normalization (en->en-US, es->es-ES, unknown->raw passthrough).
Uses vi.mock('react-i18next') because vitest's jsdom + Vite's
module resolution otherwise produces separate i18next instances
for direct imports vs react-i18next's chain, breaking React
Context subscription assertions.
- ui/src/i18n/i18n.test.ts — i18next configuration smoke + key
resolution + <html lang> sync (WCAG 3.1.1/3.1.2). 11 assertions
covering declared langs (en/es), declared namespaces (9), default
ns = common, every namespace has EN+ES bundles loaded, top-level/
pluralized key resolution, plus document.documentElement.lang
update on language change.
Validation: vitest run 136 passed (124 prior + 12 new); validator
--ratchet OK with 4 warnings (3 unchanged + 48 new t() call errors
demoted to warning by check-keys.py --ratchet — queued for cleanup).
E2E spec for language switching deferred to a separate PR — stem's
UI shell pattern differs enough from niac's that the spec needs
fresh markers, and this PR is already at a reasonable size.
Cleanup follow-up: 48 t() calls in stem reference keys that don't
exist in the EN locale (most are wrong-namespace prefixes like
t('common:settings.interface.title') when the key actually lives in
settings.json). i18next's fallback shows the key as English text,
so the bugs are invisible to users — but ES translations don't
resolve. Will land in a follow-up batch.
krisarmstrong
enabled auto-merge (squash)
May 27, 2026 14:07
Qodo reviews are paused for this user.Troubleshooting steps vary by plan Learn more → On a Teams plan? Using GitHub Enterprise Server, GitLab Self-Managed, or Bitbucket Data Center? |
Contributor
License Compliance ReportAll dependencies pass license compliance checksGo Dependencies
npm DependenciesSee full report in workflow artifacts Allowed Licenses: MIT, Apache-2.0, BSD-*, ISC, CC0-1.0, MPL-2.0 |
`local args=()` + `${args[@]}` errors on macOS bash 3.2 under
`set -u`. Use `local extra=""` + `$extra` instead — works on both
bash 3.2 and 5.x. Same patch already in seed PR #1203.
Detected when running ./scripts/i18n/validate.sh --ratchet locally
on macOS; the validator FAILED instead of demoting the
check_key_usage errors as intended by --ratchet.
No behavior change in CI (Linux runners use bash 5.x where
${empty[@]} expands fine); fixes local-dev workflow.
Contributor
License Compliance ReportAll dependencies pass license compliance checksGo Dependencies
npm DependenciesSee full report in workflow artifacts Allowed Licenses: MIT, Apache-2.0, BSD-*, ISC, CC0-1.0, MPL-2.0 |
5 tasks
krisarmstrong
added a commit
that referenced
this pull request
May 27, 2026
) Stem had 48 t() call sites referencing keys that didn't exist in the EN locale JSON. i18next's fallback semantics (second arg to t()) showed the key as English text so users didn't see broken UI — but ES translations never resolved, leaving the Spanish version permanently showing English for these labels. Surfaced by check-keys.py landing via PR #327; cleanup of the backlog so the --ratchet flag can eventually be dropped from CI. Two-pronged fix: (1) Namespace prefix corrections (24 call sites) — keys actually exist in settings.json but t() calls referenced them via the default common namespace. Fixed by changing literal call sites to use the settings: cross-namespace prefix: t('settings.interface.title', 'Interface') → t('settings:interface.title', 'Interface') Affected: SettingsDrawer, HeaderInterfaceSelector, InterfaceSection, ModeSection, ReflectorSection, RFC2544Section, RFC2889Section, RFC6349Section, Y1564Section, Y1731Section, MEFSection, TrafficGenSection, TSNSection. (2) Missing keys added to locale JSON (24 keys + ES translations): common.json: - accessibility.{closeSettings, openHistory, refreshInterfaces, selectProfile} - interface.{networkInterfaces, noInterfaces, select} - profile.{current, manage, noProfiles, select} - tooltips.header.{refresh, history, help, settings, logout} - status.{clickToReconnect, tapToReconnect, tapToReconnectHint} settings.json: - viewModuleHint, viewStandardHint - tests.{rfc2544, rfc2889, rfc6349, y1564, y1731, mef, trafficgen, tsn}.title (8 test category headings) setup.json: - buttons.copyTooltip ES translations follow I18N_TRANSLATION_MEMORY.md (formal usted, sentence case, glossary terms preserved verbatim — Reflector, MEF, RFC numbers, Y.1564 etc. stay English). Validation: tsc clean; biome clean; vitest 136 passed; check-keys strict mode now passes (was 48 errors, now 0); validator strict still fails on the pre-existing 355 fallback-pattern check (separate cleanup) and i18next 26.2 → 26.3 version bump (PR #317). Drop the --ratchet flag from .github/workflows/ci.yml's check_key_usage invocation in a separate PR once the related fallback-pattern cleanup also lands. Co-authored-by: Kris Armstrong <kris.armstrong@icloud.com>
This was referenced Jul 8, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Cross-repo backfill from niac-go#723 (check-keys.py) and niac-go#724
(Phase 6 test coverage). Brings stem to parity with niac.
What lands
`scripts/i18n/check-keys.py` — read-only static analysis that
cross-references `t()` calls in `ui/src` against committed EN
locale JSON. Catches typo'd key names and dead JSON keys.
Identical to the niac script; added `--ratchet` flag so newly-
added repos can absorb it without an immediate cleanup burden
(stem currently has 48 t() calls referencing non-existent keys —
pre-existing bugs masked by i18next's fallback semantics, queued
for cleanup).
`scripts/i18n/validate.sh` — pulled in the niac improvements:
propagation
so identifiers ending in `t` (`headers.set('Accept',
'application/json')`) don't false-match.
`ui/vitest.config.ts` — adds `@locales` path alias so
`i18n/index.ts` can be imported under test, plus dedupe entries
for i18next/react-i18next/react to encourage module deduplication.
`ui/src/hooks/useLocale.test.ts` — 3 unit tests for the BCP-47
normalization (`en` → `en-US`, `es` → `es-ES`, unknown → raw
passthrough). Uses `vi.mock('react-i18next')` because vitest's
jsdom + Vite's module resolution otherwise produces separate
i18next instances for direct imports vs react-i18next's chain,
breaking React Context subscription assertions.
`ui/src/i18n/i18n.test.ts` — i18next configuration smoke + key
resolution + `` sync (WCAG 3.1.1/3.1.2). 11 assertions
covering declared langs (en/es), declared namespaces (9), default
ns = common, every namespace has EN+ES bundles loaded, top-level
`document.documentElement.lang` update on language change.
Test plan
(3 unchanged + 48 new t() call errors demoted to warning by
check-keys.py `--ratchet`)
Out of scope
UI shell pattern differs enough from niac's that the spec needs
fresh markers, and this PR is already at a reasonable size.
bugs, e.g. `t('common:settings.interface.title')` when the key
lives in `settings.json`). i18next's fallback shows the key as
English text so the bugs are invisible to users — but ES
translations don't resolve. Will land in a follow-up batch.