Repository navigation
docs(types): cite the 'agenda' retirement's landing commits, not the dead issue numbers (objectui#10701) - #10707
Conversation
…dead issue numbers The 'agenda' retirement text cited two issue numbers and, in the landing commits' titles, two PR numbers that all answer 404 in objectui. The same numbers resolve in objectstack, but to unrelated subjects, so a repo-qualified citation would mislead. The durable record is the two landing commits, b55a346 (CalendarViewMode) and ed8df3e (defaultView). - The three zod describe strings drop the pointer: "('agenda' was retired)" stands alone, and a runtime reader of the zod description has no repository to resolve a bare sha against. - The docblocks (types, zod, plugin-calendar), the zod line comment, the schema-reference CalendarViewMode row and the i18n batch changeset cite the landing commit. - The two retirement changesets drop their self-pointer: changelog-git stamps each entry with the commit that added the file, which is the landing commit. Frontmatter is byte-identical. Wording only: no accepted value, type or behaviour changes. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BA3nKVUwKQJf8DBxrSVtNC
|
changeset-claim-re-read
|
…ad-key batch objectui#5783 answers 404, like the retirement citations repaired in the previous commit, and the i18n dead-key batch changeset publishes it verbatim into the CHANGELOG at the next release. The sentence stands without it: `calendar.agenda` is one of the batch's retired keys, so the parenthesis now ends "`calendar.agenda` included". Only that one line changes; the frontmatter is byte-identical. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BA3nKVUwKQJf8DBxrSVtNC
✅ Console Performance Budget
The eager closure is every chunk the entry reaches through static imports — what the browser fetches and parses before the app renders. The entry chunk on its own is a small fraction of it. 📦 Bundle Size Report
Size Limits
|
1 similar comment
✅ Console Performance Budget
The eager closure is every chunk the entry reaches through static imports — what the browser fetches and parses before the app renders. The entry chunk on its own is a small fraction of it. 📦 Bundle Size Report
Size Limits
|
Fixes #10701
Clause-②: no (wording only; no accept set, type or behaviour moves)
Dispatched implementation of the
domain:uiseat 1 claim (comment5838463073) on objectui#10701, sessionhttps://claude.ai/code/session_01BA3nKVUwKQJf8DBxrSVtNC. Wording only: no accepted value, type or runtime behaviour changes.Premise, re-measured first
REST
GET /repos/objectstack-ai/objectui/issues/{n}, 2026-09-25:None of the four answers 200 in objectui, so the premise holds. The sister-repo numbers resolve, but to unrelated subjects, so a repo-qualified citation would send a reader to the wrong record. This PR cites the landing commits instead. Both are ancestors of
main(git merge-base --is-ancestor, exit 0 for each):b55a34647feat(types): retire 'agenda' from CalendarViewMode and the zod enumed8df3e50feat(types): retire 'agenda' from defaultView on all three declaration facesCitation form, site by site
.describe()strings (CalendarViewSchema.view, the list-viewcalendarconfig'sdefaultView,ObjectCalendarSchema.defaultView): the pointer is dropped, and they now end('agenda' was retired). That sentence stands alone. These strings ship indistand reach whatever reads the zoddescriptionat runtime, a reader who has no repository to resolve a bare sha against. No.describe()string underpackages/types/src/zod/carries a commit citation today, so there was no runtime-string precedent to follow.89d7b35a7" inpackages/data-objectstack/src/metadata-client.tsand "deleted from the repository in8c5d20455" in.changeset/6342-dead-migration-guide-pointers.md.schema-reference.md, theCalendarViewModerow: "retired inb55a34647and now fails validation".@changesets/cli/changelogis@changesets/changelog-git, which prefixes each CHANGELOG entry with the short hash of the commit that ADDED the file (git log --diff-filter=A). That is the landing commit, so the released entry already carriesb55a346/ed8df3e. Where thedefaultViewchangeset points at theCalendarViewModeretirement, it now citesb55a34647.4730-retire-dead-locale-key-batch.md): "labelled a view modeb55a34647retired fromCalendarViewMode".check-changeset-overwriteprintsdeclared at baseequal todeclares nowfor all three. The filenames keep their5740/5784suffixes: a filename is not a citation, and a rename would read as deleting one declaration and adding another.Which sites actually publish: the card's two classes, measured against
distI built
@object-ui/typesand@object-ui/plugin-calendarand greppeddist:complex.zod.tsdocblocks onCalendarViewModeSchemaandCalendarViewSchemaship indist/zod/complex.zod.d.tsand.js. Theobjectql.zod.tsline comment above the calendar config'sdefaultViewships indist/zod/objectql.zod.js, because tsc emits withremoveComments: false. This is published text, so this PR repairs it along with the rest. All three are the optional adjacent comments the claim's file surface already lists.plugin-calendar/src/index.tsxappears inpackages/plugin-calendar/dist. The same docblock'sobjectui#8201is missing too, while a code string from the same file (thedefaultViewinput description) is present as the control. The site is repaired anyway because the triage names it, and it is declared comment-only.@object-ui/typesdistcontains noobjectui#5740orobjectui#5784.Census
git grep -nE 'objectui#5740|objectui#5784|#5740|#5784|#5782|#5806'overpackages/,content/and.changeset/at the base finds 26 lines, all in files the card names. Outside those three roots it finds 0 (control:#7853hits 5 files). After this PR there are 0 non-test hits. The 10 lines that remain are docblocks and comments in four test files. Those do not ship and are optional per the triage (see Acceptance notes).Changeset
.changeset/10701-agenda-retirement-citations.mddeclares@object-ui/types: patchfor the.describe()wording (precedent:10084-grid-ceiling-describe.md). Theplugin-calendaredit is comment-only, as measured above. This repo's precedent for comment-onlysrc/edits is an empty declaration (for example10159-action-preview-reference-citation.md). Instead of adding a second file, the new changeset's body saysplugin-calendardeclares no release.check-changeset-presencerequires one added declaration, and this one satisfies it.Pins
grep -rln "agenda' was retired" packages/returns only the two zod source files. No test snapshots or reads the olddescribetext, so no pin changed. No ablation was run, because nothing executable changes.Verification, all on
045a05c26(the patch round's re-runs atf730dc423are under Acceptance notes)pnpm exec vitest run packages/types/src/__tests__/:Test Files 240 passed (240),Tests 5276 passed (5276).pnpm exec vitest run packages/plugin-calendar/:Test Files 46 passed (46),Tests 328 passed (328).node scripts/markdown-test-inputs.mjs --list(classes.changeset/**,content/docs/**andcontent/docs/api/schema-reference.md): the 15 readers underscripts/__tests__/giveTest Files 15 passed (15),Tests 682 passed (682). The readers underpackages/typesran in the 240 above.type-check:@object-ui/types(tsc on the main, examples and test configs) and@object-ui/plugin-calendar(after building its dependency closure). Both exit 0.pnpm docs:check-linksprints "Links are valid across 17 scan roots.";pnpm check:doc-typesprints "Every documented component type is registered.";pnpm check:doc-fences,check-doc-example-ids,check-doc-expression-carriageandpnpm check:doc-example-readersexit 0.pnpm check:control-bytesis OK.pnpm check:new-line-citationsprintsVERDICT new-cross-file-line-citations: 0 new citation(s).node scripts/check-changeset-presence.mjsprints "5 source file(s) of 2 released package(s) changed, and this change declares 1 changeset(s)".pnpm changeset:checkexits 0.pnpm check:pending-changeset-literalsprints "No test source names a pending changeset."pnpm check:changeset-claims(report-only): 50 pending changesets name a file this PR touches (the sharedcomplex.ts/objectql.ts/ zod files andschema-reference.md). I read them against the diff. No pending changeset quotes the changed text, and the two retirement changesets' closing sentences about thedescribestrings stay true.check-changeset-overwrite(report-only): 3 files modified, each its case 2 (a deliberate prose correction), with declarations unchanged.--no-inline-config --format json) over the five touched source files: 5 files, 0 errors, 49 pre-existing warnings, none on an edited line.eslint.config.jsenables no type-aware linting, so a comment or string diff cannot change another file's result.check-doc-snippet-typesandcheck-doc-example-types. Reason: both exit 2 with PREREQUISITE NOT MET (they need a 34-package scoped build). They type-check code fences, and the only doc edit here is a table cell.Acceptance notes
objectui#5783, folded in the PM patch round (headf730dc423). It also answers 404 (control:#5781answers 200), and it was in a changeset this PR already edits, so the seat extended the claim's surface to that one pointer..changeset/4730-retire-dead-locale-key-batch.mdline 7 now reads`calendar.agenda` included).with no replacement citation, as a single hunk@@ -7 +7 @@. The frontmatter sha256 is unchanged (57e32eec…). The patch round re-ran control bytes, changeset presence /changeset:check/ claims / overwrite / pending literals, and the twoscripts/__tests__suites that read.changeset/**(2 files / 44 tests): all exit 0. The one residual is the header docblock ofdead-key-batch-retired-4730.test.ts, which does not ship and sits with the other comment-only test sites below.calendar-view-mode-agenda-retired.test.ts,default-view-agenda-retired.test.ts,dead-key-batch-retired-4730.test.tsandscalarKeysAreDeclaredAndHonoured-8201.test.tsstill cite the dead numbers. They do not ship and are optional per the triage.distgrep settles that classification for future cards of this class.Generated by Claude Code