chore(plugin-timeline): retire the five dead timeline.relative.* default rows (objectui#7874) - #7887

Fixes #7874

The PM ruling is delete, as residue: the five timeline.relative.* rows in TIMELINE_DEFAULT_TRANSLATIONS have no reachable path, and no product behaviour attaches to code nothing can reach. Wiring relative dates through t() would be a new feature (copy, plural rules, ten-pack coverage), not the other half of this question.

One row deleted from packages/plugin-timeline/src/useTimelineTranslation.ts, a retirement pin added beside it, and an empty-frontmatter changeset declaring that nothing releases. Measured on origin/main at 83fe6e741; branch head 3ac057309.

1. The five rows, transcribed verbatim

The deletion is safe because it is reversible, and it is reversible because the content is written down here (and in the pin test, and in the changeset) - not because "it is in git history somewhere":

'timeline.relative.today': 'Today',
'timeline.relative.tomorrow': 'Tomorrow',
'timeline.relative.yesterday': 'Yesterday',
'timeline.relative.inDays': 'In {{n}} days',
'timeline.relative.daysAgo': '{{n}} days ago',

2. Escape hatch: "zero callers" re-verified on current main, by PREFIX, with a positive control

Re-measured on this branch's base rather than inherited from the card.

By prefix, not by the five full key names, because a dynamically assembled key (t(`timeline.relative.${x}`)) is invisible to a literal search for the complete names. Two passes, both returning exactly the five declaration lines and nothing else:

Positive control - the same command shape, on keys that DO have callers. A search that has not been shown to fire is not a measurement:

Dynamic assembly, checked as its own question, not as a by-product of the prefix search:

• No template-literal translator call exists anywhere in the tree: git grep -nE '\bt\(' -- 'packages//src' 'apps//src'` returns nothing (exit 1).
• No concatenated or template-built timeline.* key: `timeline., 'timeline.' +, "timeline." +, timeline.${ - every hit is prose in a comment or a spec key like timeline.startDateField, none of them a key handed to t().
• The suffix side too, in case a prefix lived in a variable: the only tree-wide hit for a bare 'daysAgo' is apps/console/src/pages/system/ApprovalsInboxPage.tsx:653, whose local tr resolves t(`approvalsInbox.${key}`) - a different namespace, and its own module-local defaults.
• Nobody reads the table wholesale either: every importer of TIMELINE_DEFAULT_TRANSLATIONS indexes a named key (timeline.scale.*, timeline.gantt.*); none iterates it or pins its size.

Why they are residue, positively rather than by absence. Day-granularity relative phrases are produced by formatRelativeDate / formatRelativeDays in @object-ui/core (utils/date-display.ts) through Intl.RelativeTimeFormat, whose output for the en locale is literally "Today" / "Tomorrow" / "Yesterday" / "in 3 days" / "3 days ago". A formatter that derives the words from the locale needs no copy row in any pack - that is what left these five behind.

Also worth recording: TIMELINE_DEFAULT_TRANSLATIONS and translateTimelineDefault are not re-exported from the package entry (src/index.tsx re-exports ./renderer, which imports them without re-exporting). No consumer of the published tarball can name them, so the deletion cannot move a published API surface either.

3. The three instruments: none of them moved

Expected none, measured none. Before and after are the same commands on the same tree, run either side of the edit:

That is exactly the card's reading, and it is why these rows survived: class 1 judges call sites (they had none), parity compares packs to each other (ten packs identically lacking a key is full parity), and the reverse sweep judges pack keys (these were in no pack). They were caught by objectui#7567's census printing its abstention count - 5 of 846 factory-table rows - not by any rule judging them.

Two neighbours measured for the same reason, also unmoved: check:i18n-drift reports 0 en value(s) changed (0 key(s) added, 0 removed), and fallback-placeholder-spelling-3512.test.ts - which walks this exact table - stays green.

One reading the card did not have, and it goes the other way. There IS a rule in this repo that rejects a defaults row whose key no pack defines: defaults-maps-mirror-en-pack.test.tsx (objectui#4401), "every row names a key the en pack actually defines". It did not fire because its MAPS list is three hand-written entries - detail, list, designer - and TIMELINE_DEFAULT_TRANSLATIONS is not among them. So the blind spot is a fourth gate's coverage, not a fourth gate's absence. Filed separately as #7884; out of scope here, and it changes nothing about this deletion.

4. Changeset: the checker's verdict, quoted

Not guessed. node scripts/check-changeset-presence.mjs first refused the change with no changeset:

❌  1 source file(s) of 1 released package(s) changed, and this change adds no changeset:

      @object-ui/plugin-timeline
          packages/plugin-timeline/src/useTimelineTranslation.ts
...
    If this change really should release nothing, say so — that is a pass, not a
    workaround. [quote trimmed here: the next line names a changeset file with an
    EMPTY frontmatter, spelled with angle brackets that GitHub's body sanitizer eats]

.changeset/7874-timeline-relative-defaults-retired.md was added with an empty frontmatter, because nothing reachable changes and nothing published moves (see the export note in section 2). The checker then answered, verbatim:

Compared the working tree with 83fe6e741 (merge-base with origin/main): 1 file(s) changed, 1 of them published source of a package the release covers, 0 of them a manifest whose published contract moved, 0 under a package changesets ignores, 1 changeset(s) added.
✅  1 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s): .changeset/7874-timeline-relative-defaults-retired.md.
    Every one of them has an EMPTY frontmatter — declared as releasing nothing, which
    is the explicit exemption and a complete answer to this gate.

check-changeset-no-major.mjs: ✅ No changeset declares a major bump.

5. Clause-② stays no

No key added, no published export moved, no gate predicate or scan population touched. The governed-surface check agrees, on this commit's three paths:

✅ NOT GOVERNED — 3 path(s) checked against 5 governed surface(s); none matched.
   An ordinary pull request: the normal review and merge-queue route applies.

The pin, and the ablation that shows it fires

packages/plugin-timeline/src/__tests__/timeline-relative-defaults-retired-7874.test.ts follows the retirement-pin series (objectui#4392 / #4730 / #5504 / #6310 / #7125): the retired rows asserted absent by key and by prefix, the exported translateTimelineDefault asserted to answer a retired key with the raw key rather than its old string, and surviving-sibling assertions so a green cannot be bought by deleting the neighbourhood. timeline.bucket.today and timeline.bucket.tomorrow carry the same two strings the retired rows did, so a sweep by VALUE takes them out as collateral - they are pinned for exactly that reason.

The pin exists because no gate can see a dead row come back: each of them runs call site to key, pack to pack, or pack to call site.

Ablation - the rows re-inserted, then restored, with both legs proven on disk rather than by an exit code. No rebuild leg is needed here and none was run: the test imports ../useTimelineTranslation by relative path inside its own package, so vitest compiles the source, with no exports/dist indirection to go stale.

HEAD blob for packages/plugin-timeline/src/useTimelineTranslation.ts = c02661e28d40040d0168a2c7650bd00c114beb66
mutation written
injected rows on disk = 5 ; blob now 67ff97460aa6b3a621d8a4e506a57bea37a4fe71 (HEAD c02661e28d40040d0168a2c7650bd00c114beb66)
ABLATION_TEST_EXIT=1 (expect non-zero = the pin fires)
     × carries no `timeline.relative.*` row, by key and by prefix
     × answers a retired key with the key itself, not with its old string
 Test Files  1 failed (1)
      Tests  2 failed | 2 passed (4)
restored blob = c02661e28d40040d0168a2c7650bd00c114beb66 ; rows on disk = 0 ; git diff HEAD = ''
RESTORE VERIFIED: blob matches HEAD, git diff HEAD empty, 0 retired rows on disk

The mutating script carried trap restore EXIT INT TERM with an absolute path resolved from git rev-parse --show-toplevel, and both legs were verified by state (blob hash, row count, empty git diff HEAD), never by the editor's exit code.

What ran, on the pushed SHA 3ac057309 with git diff HEAD empty

On the lint narrowing (two files rather than the repo sweep, which CI runs anyway): the population came from eslint's own resolution of the two paths, the file count is read off --format json, and this repo's flat config is not type-aware - tseslint.configs.recommended, no parserOptions.project and no projectService - so this diff cannot move the verdict on any file it does not touch. That makes the narrowing a measurement rather than a skip.

🤖 Generated with Claude Code

https://claude.ai/code/session_01KbJQ1y1J12nZxYzFWhP8Q3

Generated by Claude Code