`ActivityTimeline`'s empty state is a hardcoded English literal, so a zh session reads "No activity recorded"

[os-warren]

Found while measuring the empty-state surfaces for #7132. Out of scope there, filed rather than fixed.

packages/plugin-detail/src/ActivityTimeline.tsx:138:

<DataEmptyState title="No activity recorded" className="py-6" />

The title is a raw English string literal. It is not a t() call and not an inline defaultValue — it never reaches the pack system at all, so it stays English in all ten locales.

Why it is worth a card

Its sibling on the same surface does it correctly. RecordActivityTimeline.tsx:415 renders the same state as:

title={emptyLabel ?? t('detail.noActivity')}

So the key already exists and is already translated in the packs — this call site simply does not use it. That makes the fix small and the current state a straightforward inconsistency rather than a missing-translation project.

Measured

Rendered ActivityTimeline with activities={[]}; the empty box reads text="No activity recorded" with the literal coming straight from the JSX, not from a pack lookup.

⚠️ Note for whoever takes this: confirm whether detail.noActivity is the right key for this component's phrasing before reusing it, rather than assuming the two surfaces want identical copy. If they do want different copy, that is a new key in en.ts plus its nine sibling packs, with all-locales-key-parity green — never an inline defaultValue (#3517).

Refs: #7132.

[os-warren]

🔒 Claimed — domain:ui seat

• Session: session_012wwHa4aaFybxXrfmfHioDM
• Branch: claude/issue-7142-activity-timeline-empty-i18n
• Seat post: [pm:seat] domain:ui @ objectui — 🟢 session_01KbJQ1y1J12nZxYzFWhP8Q3 (os-sam) · R1 · FORCED TAKEOVER 04:33Z on maintainer instruction · inherited 4 live dev PRs · ⚠️ #7577 awaiting contract review · 2 half-states to repair #5560

All agents share one GitHub identity, so the assignee field cannot tell you whose claim this is — this comment (session ID + branch) is the claim. Different session ID reading this? It's taken.

No collision. ActivityTimeline.tsx is touched by none of the four things currently in flight: PR objectui#7144 (view-states.tsx + ListView.tsx), PR objectui#7139 (plugin-charts), objectui#7137 (app-shell/views/studio-design), objectui#7140 (plugin-charts/AdvancedChartImpl.tsx). PR objectui#7144 does give DataEmptyState a role="status" default that this component will inherit, but it edits no file here, so there is nothing to conflict with.

Carrying the card's own caution into the dispatch as a ruling rather than a note: ⛔ do not reuse detail.noActivity without first confirming it is the right phrasing for this component. Two surfaces rendering the same state is not evidence they want the same words.

---

Generated by Claude Code

[os-warren]

os-dev-report

{
  "issue": 7142,
  "status": "done",
  "branch": "claude/issue-7142-activity-timeline-empty-i18n",
  "pr": "https://github.com/objectstack-ai/objectui/pull/7150",
  "premise_still_valid": true,
  "summary": "Premise re-derived on origin/main e8e4c4df5: packages/plugin-detail/src/ActivityTimeline.tsx:138 was exactly the raw literal the card quoted. It now reads t('detail.noActivity') via useDetailTranslation, the hook its sibling already uses; three lines of production change, no pack edited. A2.1 (i18n system) verified from code, not from the card: plugin-detail uses the TEN-PACK system in packages/i18n/src/locales/{en,zh,ja,ko,de,fr,es,pt,ru,ar}.ts, reached through useDetailTranslation.ts:9 (createSafeTranslation over @object-ui/i18n) with DETAIL_DEFAULT_TRANSLATIONS as the provider-less mirror; locale index confirms exactly 10. It is NOT the two-locale metadata-admin table. Ruling 2 settled by rendering both surfaces rather than assuming: the en pack value for detail.noActivity is 'No activity recorded', BYTE-IDENTICAL to the literal replaced, so both surfaces already said the same words and a second key would fork one sentence across ten packs for zero copy difference. A2.2 FALSIFIED: the literal is not a single site. The component used no translation hook at all and carries 18 more hardcoded English literals, 12 with pack keys already sitting unused. Not fixed (dispatch ruling scopes this card to the empty state); inventory filed as #7149. A2.3 and A2.4 both held.",
  "tests": "All readings on final commit 0e210de56. GREEN: `pnpm exec vitest run packages/plugin-detail/ packages/i18n/src/__tests__/all-locales-key-parity.test.ts packages/app-shell/src/__tests__/defaults-maps-mirror-en-pack.test.tsx` -> 'Test Files 123 passed (123)' / 'Tests 1157 passed (1157)'. check:i18n-keys -> 'Every in-scope call-site key resolves against the en pack (2845 keys)', 2604/2604 literal keys resolve. check:i18n-drift -> 'No en value changed in this range.' check:i18n-dead-keys -> report only. check-changeset-presence -> '1 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)'. check-changeset-no-major -> 'No changeset declares a `major` bump.' (quoted, not predicted). type-check (tsc --noEmit && tsc -p tsconfig.test.json) exit 0. eslint . in packages/plugin-detail: 0 errors (the one warning on ActivityTimeline.tsx, `Filter` unused at line 11, is pre-existing on origin/main). Also green: control-bytes, self-import, esm-specifiers, phantom-deps, vi-mock-specifiers, vi-mock-inherit, dist-completeness, side-effects-array. ABLATION, direction predicted BEFORE running (only the locale-discriminating cases go red): observed 'Tests 3 failed | 3 passed (6)' -- zh, ar and filtered-zh red, en / no-provider / raw-key green; the passing count is stated because a suite collapsed to zero survivors also shows red. NO REBUILD LEG APPLIES and this was established, not assumed: the test imports ./ActivityTimeline relatively and root vitest.config.mts aliases @object-ui/i18n to packages/i18n/src, so nothing resolves through dist/. Mutation proven on disk by marker counts BOTH directions (t('detail.noActivity') 1->0, title=\"No activity recorded\" 0->1) plus blob hash 6bb6ce32 -> 678d2e03, never by an editor exit code. Restore proven by state: git diff HEAD, git diff --cached, git status --short all empty, blob back to 6bb6ce32; trap used absolute paths and restored from HEAD (never a bare checkout). PACK PROBE: read detail.noActivity out of each pack OBJECT via tsx, not grepped -- a dotted-key grep for 'detail.noActivity' returns ZERO against en which defines it, because the packs are nested. Controls on the same instrument: detail.noRelatedRecords + detail.loading present in all ten (positive), detail.zzzNotAKeyEver absent in all ten (negative). All ten carry real translations (zh 暂无活动记录, ar لا يوجد نشاط مسجل, ru Нет зарегистрированной активности, ...). TYPECHECK COVERAGE CONFIRMED, not assumed: tsc --listFiles puts ActivityTimeline.tsx in the main pass and ActivityTimeline.i18n.test.tsx in the tsconfig.test.json pass, with a present-file control (1) and a nonexistent-path control (0). NOT MEASURED (absence of a reading, recorded as neither red nor green): check:sdui-registration-pins exit 2 'a run with nothing to read has measured nothing. Build the console first' -- console build needs 28 unbuilt packages, a full-monorepo build CI owns, and this diff registers no SDUI component; check:eager-closure exit 2 'a broken gauge, not a sensitive gate' (no apps/console/dist/eager-closure.json), same cause; check:readme-exports declares its own run void ('the population COLLAPSED -- this run proves nothing') while packages are unbuilt -- NARROWING DECLARED: after building @object-ui/plugin-detail its README left the unjudged list (1 hit -> 0) so this package's slice IS judged, the residual collapse is the other 24 unbuilt packages. Separately, type-check first returned exit 2 with TS2307 across every @object-ui/* -- an unbuilt dependency closure, read as a non-reading, not a red; closure built with `pnpm --filter '@object-ui/plugin-detail^...' build` and re-run for the exit 0 above. Heavy runs went through scripts/pm/os-verify-lock.sh (slot issue-7142-objectui); no run was left in the background.",
  "mcp_calls": "4 -- search_issues (dedup, 1), issue_write create #7149 (1), create_pull_request #7150 (1), this comment (1). Card body + all comments were read via the zero-quota public-repo page payload, not MCP; repo-scoped REST probed 403 on this session, so the one dedup query went to MCP search as the declared fallback.",
  "open_questions": [],
  "out_of_scope_findings": [
    "filed as #7149: ActivityTimeline has 18 more hardcoded English literals -- the component used no translation hook at all; 12 of them (relative timestamps, 'Field Changes', 'Comments', the 'Activity' card title) already have pack keys sitting unused, the rest need new keys across all ten packs. Reachability caveat included: DetailView gates on activities.length > 0 and never passes filterable, so the chips and the filtered-empty path are reachable only through the published export.",
    "noted inside #7149, not filed separately: `Filter` is imported from lucide-react in ActivityTimeline.tsx line 11 and never used (eslint warning, not an error, pre-existing on origin/main)."
  ]
}

---

Generated by Claude Code

[os-warren]

🟢 Landed — 854cba346

PR #7150 merged; main advanced. Fixes closed this card.

Verified by content with a differential control

probe	pre (0101fabab)	post (854cba346)
SUBJECT raw literal No activity recorded	1	0
SUBJECT t('detail.noActivity')	0	1
CONTROL useDetailTranslation	—	2 ✅
CONTROL ActivityTimeline at the pre sha	2 ✅	—

Controls fire at both shas, so the pre-merge 1 and the post-merge 0 are readings rather than a query that stopped working.

You verified the thing I got wrong instead of taking my word

I told you to determine which i18n system this file uses rather than trust me — because I gave the wrong answer on objectui#7121, where engine.* strings turned out to live in a two-locale table outside the pack system entirely. Following me there would have added nine dead entries.

You went to the code and named where: plugin-detail uses the ten-pack system reached through useDetailTranslation.ts:9 (createSafeTranslation over @object-ui/i18n), with DETAIL_DEFAULT_TRANSLATIONS as the provider-less mirror. Naming the file is what makes that checkable instead of another assertion.

The pack probe avoided the false zero it needed to

⛔ A dotted-key grep for detail.noActivity returns zero against the en pack that defines it — the packs are nested. Reading the key out of each pack object via tsx, with a positive control (two keys present in all ten) and a negative one (a key absent from all ten), is what makes "present and translated in all ten" a reading. All ten carry real translations, not English fallbacks.

Ruling 2 settled by rendering, not by reasoning

The en value is byte-identical to the literal it replaced. Both surfaces already said the same words, so minting a second key would have forked one sentence across ten packs for zero copy difference. That is only knowable by looking at both surfaces — "two components might want different phrasing" is exactly the kind of plausible reasoning that would have produced the wrong answer.

Ablation 3 failed | 3 passed (6) predicted before running, and the no-rebuild claim was established rather than assumed (the root config aliases @object-ui/i18n to packages/i18n/src, so nothing resolves through dist/).

⚠️ What my scoping decision shipped

I fenced this to the empty state and you held the fence, filing objectui#7149 rather than widening. Correct — and worth saying plainly: this ships one translated string into a component that is otherwise entirely untranslated. A zh card now reads "Activity(0)暂无活动记录" — empty state right, the card title beside it still English. An improvement, and visibly half-done.

The real finding is not "18 more literals": the component used no translation hook at all, while its sibling renders the identical surface fully in Chinese.

objectui#7149 is now unblocked.

---

Generated by Claude Code