Skip to content

feat(spec): retire the nine dead contributes members and kinds[].globs (family: #10724 + #11169) - #12001

Merged
os-support-ai merged 6 commits into
mainfrom
claude/issue-10724-contributes-retirements
Aug 25, 2026
Merged

feat(spec): retire the nine dead contributes members and kinds[].globs (family: #10724 + #11169)#12001
os-support-ai merged 6 commits into
mainfrom
claude/issue-10724-contributes-retirements

Conversation

@claude

@claudeclaudeBot commented Aug 25, 2026

Copy link
Copy Markdown
Contributor

Fixes#10724
Fixes#11169

Family dispatch (chain head #10724): retires the nine dead members of the plugin-manifest contributes block, and the inert contributes.kinds[].globs sub-field, under ADR-0049 enforce-or-remove. One commit per card — each card's work is independently identifiable.

Census re-verification (PM mechanism assumption — re-measured at claim time)

Measured on current merged main (ac59744), monorepo-wide, non-test, with the card's own controls:

  • objectstack: exactly ONE read of manifest.contributespackages/objectql/src/engine.ts:4604-4606, member kinds (line drift from the card's :4499 re-derived). Control: the same regex finds those three lines.
  • objectui (origin/main @ 65d3e76): 0 property reads (\.contributes\b|contributes\?\.|contributes\[ over *.ts,*.tsx); control: the bare word contributes appears in 63 files, so the zero is about the path, not the corpus.
  • cloud: delivered board reading used as an existing fact (2026-08-24, comment 5396673984): origin/main @ 5b5925a, zero manifest.contributes reads, controls held. Not re-run per dispatch.
  • contributes.kinds.globs is declared-but-unenforced: nothing reads it, and the file-type discovery it advertises globs filePatterns off a registry contributes.kinds does not extend #11169's one-line executable criterion, re-run with its positive control: git grep -n "\.globs" -- packages | grep -v -e '\.test\.' -e 'zod.ts' -e 'registry.ts' → empty; unfiltered, the query returns the schema declaration (manifest.zod.ts:329) and the two type positions (registry.ts:3748registerKind param, :3752getAllKinds return — anchors moved from the card's engine.ts citations). Zero is a reading, not a broken query.

premise_still_valid: true for both cards, with one named sub-item exception (below).

Card 1 — #10724: nine retiredKey() tombstones (commit 1)

  • events / menus / themes / translations / actions / drivers / fieldTypes / functions / commands are retiredKey() tombstones in manifest.zod.ts, mirroring loading — the block is not .strict(), so a plain deletion would silently strip.
  • kinds (live reader) and routes (open fork contributes.routes has zero readers, yet a published skill, an ADR row and the dispatcher's own doc all present it as the way to serve a code-handler endpoint — enforce-or-remove fork #10726) untouched.
  • Nine retired-key entries (kernel/Manifest:contributes.<member>, major 18) + ONE ADR-0087 D3 SemanticMigrationplugin-manifest-contributes-dead-members-retired (NOT a D2 conversion: a manifest is not a stack collection member — PLURAL_TO_SINGULAR has no packages/plugins entry — so a conversion would be a transform with no seam that ever runs; recorded verbatim in the entries).
  • Generated upgrade-guide row + regenerated artifacts; liveness ledger rows updated (tombstone rows STAY, per the loading/orphans asymmetry); minor changeset with FROM → TO per member and <!-- adr-0087: registered … --> marker.
  • In-repo authors corrected: plugin-hono-server (contributes.events — already subscribes imperatively) and driver-memory (contributes.drivers — wired via the driver.* service), each with an in-place comment.
  • Author-facing corrections: content/docs/protocol/kernel/i18n-standard.mdx Plugin Integration section rewritten to the real mechanism (defineTranslationBundle + stack translations); the manifest.zod.ts Commander.js comment replaced by the commands tombstone (now consistent with cli-extension.zod.ts); the same false crediting in metadata-plugin.zod.ts's ADR-0088 comment corrected (bounded in-place fix, same defect class).
  • Premise exception (one named correction): content/docs/plugins/development.mdx:334/:388 — on current main those contributes blocks belong to defineStudioPlugin (the live Studio-plugin contribution surface, studio/plugin.zod.ts), not ManifestSchema; the file's only ManifestSchema example (line ~65) carries no contributes block. Nothing to correct there; evidence in the report.
  • ADR-0088:41 row correction is the separate sidecar PR (governed surface): docs(adr): correct ADR-0088's function delivered-form row — contributes.functions was never a delivered form #12000

Card 2 — #11169: contributes.kinds[].globs retired (commit 2)

Maintainer-ruled 2026-08-24 (「接受你的建议。」): remove via the full ADR-0049 ceremony.

  • globs is a retiredKey() tombstone inside the kinds item; kinds bucket and id untouched; the kinds JSDoc no longer promises file-type discovery (metadata-plugin.zod.ts:56-61 records that discovery globs the metadata type registry's filePatterns, which contributes.kinds does not extend).
  • Consumer types updated: registry.registerKind / getAllKinds drop globs from their declared shapes (nothing read the value; the parameter widens).
  • Retired-key entry kernel/Manifest:contributes.kinds.globs + D3 semantic entry plugin-manifest-kind-globs-retired; breaking changeset naming the removed key; ledger kinds row note updated.
  • Rejection pin asserts the specific zod issue (path + prescription), not just success: false; positive pin: kinds: [{ id, description }] still parses and registers.

Runtime behaviour unchanged — evidence

Nothing read the nine members or globs (census above), so removing them removes no behaviour. Verified: @objectstack/objectql full suite green (engine manifest reads / kinds registration included), @objectstack/spec full suite green; the one live contributes read (kinds) still registers and its logging pin (engine-kind-registration-log.test.ts) still passes with globs-free fixtures. A stored manifest carrying a retired key degrades to one [metadata_spec_invalid] log line at registration (registry validate() is a diagnostic, not a gate), not a boot failure.

Gates

Initial readings at 38c37ade; after the pre-PR merge of origin/main@497ded78 (the card's binding registry constraint) the union re-ran at the final head c415a378 — every exit captured before any pipe; verdict lines quoted from the tools themselves.

Union at final head c415a378 (post-merge, post-regeneration): @objectstack/spec FULL suite + typecheck (locked) — Test Files 424 passed (424) / Tests 11292 passed (11292), os-verify-lock: VERDICT command-exit 0; the 12-gate generated-artifact battery (incl. strictness-ledger, variant-docs) ALL PASS on the identical tree; family entries and tombstones verified present post-merge (git grep counts: 18 exact-key mentions, 10 entry files, 12 retiredKey sites). The merge brought one spec change (#11963, ui/object-master-detail — disjoint files); objectql was untouched by the incoming commits, so its 232/232 · 4113/4113 reading at 38c37ade stands (declared).

  • Build: spec dependency closure + spec rebuilt after every source change — os-verify-lock: VERDICT command-exit 0 (build2, held 124s).
  • @objectstack/spec FULL suite + typecheck (locked): Test Files 424 passed (424) / Tests 11285 passed (11285) + tsc --noEmitos-verify-lock: VERDICT command-exit 0.
  • @objectstack/objectql suite + typecheck (locked): Test Files 232 passed (232) / Tests 4113 passed (4113)os-verify-lock: VERDICT command-exit 0.
  • @objectstack/plugin-hono-server / @objectstack/driver-memorytypecheck (script name verified, run echoed): both exit 0.
  • Generated-artifact battery (spec): check:migration-registry · check:liveness · check:empty-state · check:authorable-surface · check:docs · check:api-surface · check:spec-changes · check:upgrade-guide · check:skill-refs · check:skill-docs — ALL PASS. check:skill-examples: ✅ 256 prose examples type-check across 3 surface(s) (the new i18n-standard.mdx block included).
  • Derived-family subset (from node scripts/pm/dispatch-gates.mjs, no paths — 28-path change set derived at merge-base ac59744, provenance line: "gate list derived from the tree of 'objectstack-ai/objectstack' at commit 38c37ad"): check:driver-memory-census · check:spec-parsed-alias · check:test-source-alias · check:cross-package-test-inputs · check:adr-0087-registration · check:adr-anchors · check:doc-anchors · check:role-word · check:strictness-ledger · check:variant-docs — ALL PASS. check:nul-bytes: OK (scanned 6661 text file(s) … no raw ASCII control bytes).

Declared narrowings (each a measurement, not a skip; CI runs the full farm regardless):

  1. Under sustained verify-lock contention (two 9-minute queue-timeouts; three sibling agents cycling multi-minute spec suites), the single-threaded gen:*/check:* tsx scripts (2–11s each) ran directly instead of inside the lock; every BUILD and TEST run stayed under the lock (verdict lines above).
  2. check:i18n refused its prerequisite (workspace CLI not built in this fresh worktree): "PREREQUISITE NOT MET — the workspace CLI is not built … Nothing was checked". Not run locally; my diff touches no i18n-extraction input (no *.form.ts, no bundles, no platform-objects metadata) — CI owns the fresh-build run.
  3. Repo-wide sweep families from the derived list with no input in this diff (published-files, merge-driver, type-check-coverage/debt, engine-double-contract, etc.) ride CI.

Expected-zero readings, stated so they are not misread: authorable-surface/ and json-schema.manifest/ are byte-identical — the retired keys live BELOW the one-level surface walk (kernel/Manifest:contributes is the walked key), so the four ratchets are structurally blind to them; the ledger drill-down rows and RETIRED_KEYS_BY_MAJOR entries carry the record instead, and gate (b2) accepts entries for below-the-walk keys. spec-changes.json / docs/protocol-upgrade-guide.md are byte-identical because both project majors ≤ 17 today — no shipped major-18 retirement appears in either yet; the rows materialize at the protocol-18 bump from the entries this PR registers. api-surface/ unchanged: no export added or removed (the registerKind narrowing is a type-only edit inside objectql, not a spec export).


Generated by Claude Code


Generated by Claude Code

…ibutes block (#10724)
ADR-0049 enforce-or-remove, triage graded 2026-08-21, cloud census leg
discharged clean 2026-08-24. #10627 measured exactly ONE non-test read of
manifest.contributes monorepo-wide (engine.ts, member kinds), controlled and
re-verified on current main at claim time (objectstack main@ac59744:
engine.ts:4604-4606 only; objectui origin/main@65d3e76: 0 property reads,
control 63 files carry the bare word; cloud: delivered clean reading
2026-08-24 @5b5925a).
- events / menus / themes / translations / actions / drivers / fieldTypes /
functions / commands are retiredKey() tombstones mirroring loading (the
block is not .strict() — a plain deletion would silently strip)
- kinds (live reader) and routes (open fork #10726) untouched
- nine retired-key entries kernel/Manifest:contributes.<member> at major 18 +
ONE D3 SemanticMigration plugin-manifest-contributes-dead-members-retired
(no D2 conversion: a manifest is not a stack collection member — no seam
would ever run it)
- liveness ledger rows updated (tombstone rows STAY, the loading/orphans
asymmetry) + README row; cloud-census provenance refreshed
- in-repo authors corrected: plugin-hono-server (events — subscribes
imperatively already), driver-memory (drivers — wired via driver.* service)
- author-facing corrections: i18n-standard.mdx Plugin Integration rewritten
to defineTranslationBundle + stack translations; the Commander.js JSDoc
replaced by the commands tombstone; metadata-plugin.zod.ts ADR-0088 comment
stops crediting contributes.functions
- rejection pins assert the specific zod issue (path + prescription) per
member; survivors pinned parsing
- minor changeset with FROM → TO per member + adr-0087 registered marker
Upgrade-guide/spec-changes projections cover majors <=17 today; the major-18
row materializes at the protocol bump, as with every shipped major-18
retirement (entries are registered now).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K93P8PbH7kVjsAXkqzH1zY
…ch patterns nothing ever read (#11169)
ADR-0049 enforce-or-remove; maintainer ruling 2026-08-24 (verbatim provenance
on the card: 「接受你的建议。」). The schema promised that declaring globs
enables file-type discovery; real discovery globs filePatterns off the
metadata type registry — which contributes.kinds does not extend, as
metadata-plugin.zod.ts records outright — so an authored globs was accepted,
stored, served back through GET /metadata/kind, and never consulted.
Liveness dead confirmation re-run at claim with the card's positive control:
filtered criterion empty; unfiltered returns the schema declaration
(manifest.zod.ts:329) and the two type positions (registry.ts:3748/:3752 —
anchors moved from the card's engine.ts citations).
- globs is a retiredKey() tombstone inside the kinds item (non-strict shape);
the kind bucket and id are NOT touched; kinds JSDoc stops promising
discovery
- registerKind/getAllKinds drop globs from their declared shapes (type-only;
the parameter widens; nothing read the value)
- retired-key entry kernel/Manifest:contributes.kinds.globs at major 18 + D3
SemanticMigration plugin-manifest-kind-globs-retired (no D2 conversion — no
seam would ever run it)
- rejection pin asserts the specific zod issue (path
contributes.kinds[0].globs + prescription); positive pin: { id,
description } still parses and keeps registering
- breaking changeset naming the removed key + adr-0087 registered marker;
ledger kinds row updated
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K93P8PbH7kVjsAXkqzH1zY
…contributes retirements (#10724, #11169)
gen:strictness-ledger — kernel/ non-strict accepted-key count 296 -> 288: the
nine tombstoned contributes members and kinds[].globs leave the accept set.
Caught by scripts/strictness-ledger-doc.test.ts on the full spec suite run.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K93P8PbH7kVjsAXkqzH1zY
@github-actions

github-actionsBot commented Aug 25, 2026

Copy link
Copy Markdown
Contributor

📓 Docs Drift Check

This PR changes 4 package(s): @objectstack/driver-memory, @objectstack/objectql, @objectstack/plugin-hono-server, @objectstack/spec, touching 11 documentable anchor(s). ⚠️11 changed file(s) yielded no anchor (packages/spec/liveness/README.md, packages/spec/liveness/manifest.json, packages/spec/src/migrations/entries/retired-keys/18.kernel__Manifest__contributes.actions.ts, …), so the pages documenting them are NOT COVERED by this run — this is not a clean bill of health for those files.

11 hand-written doc(s) NAME something this change touched and may need an implementation-accuracy re-verification:

  • content/docs/concepts/metadata-lifecycle.mdx(via SchemaRegistry (symbol))
  • content/docs/getting-started/quick-reference.mdx(via ManifestSchema (symbol))
  • content/docs/kernel/cluster.mdx(via /metadata/:type (route))
  • content/docs/kernel/services-checklist.mdx(via SchemaRegistry (symbol))
  • content/docs/permissions/authentication.mdx(via HonoServerPlugin (symbol))
  • content/docs/plugins/adding-a-metadata-type.mdx(via SchemaRegistry (symbol))
  • content/docs/plugins/development.mdx(via ManifestSchema (symbol))
  • content/docs/plugins/packages.mdx(via /metadata/:type (route))
  • content/docs/protocol/kernel/index.mdx(via HonoServerPlugin (symbol))
  • content/docs/protocol/kernel/metadata-service.mdx(via /metadata/:type (route))
  • content/docs/protocol/kernel/plugin-spec.mdx(via ManifestSchema (symbol))

3 release-owned page(s) also name something this change touched. These are read-only:

  • content/docs/releases/v12.mdx(via MetadataTypeSchema (symbol))
  • content/docs/releases/v15.mdx(via ManifestSchema (symbol))
  • content/docs/releases/v17.mdx(via HonoServerPlugin (symbol), SchemaRegistry (symbol))

content/docs/releases/ is RELEASE-OWNED (AGENTS.md "Documentation Guardrails"): release
notes are written centrally at release time, and a code PR that edits them is the exact PR
that guardrail exists to stop. They are still audited — read-only. If one of them is actually
wrong, file an issue or open a dedicated docs-only PR; do not edit it here.

What this run could not see
  • 11 changed file(s) yielded no anchor (packages/spec/liveness/README.md, packages/spec/liveness/manifest.json, packages/spec/src/migrations/entries/retired-keys/18.kernel__Manifest__contributes.actions.ts, …) — pages documenting those are invisible to this run
  • 6 name(s) were too generic to anchor anything (single lowercase words)
  • the SDK route bridge reached 45 of 222 client-bound route-ledger rows — the other 177 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
  • a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.

Coarse fallback — 128 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 227c1788d55da8f94127137bcf75ba2bf08b15d8packageMentionDocs.

Which tree this was computed on

This run read content/docs from ef28e212bc4016260ceb02fe55dc11d4f8a62f64 — the merge of head c415a378508e7372d362d0e38a6227d50037a9b4 into base 227c1788d55da8f94127137bcf75ba2bf08b15d8, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin ef28e212bc4016260ceb02fe55dc11d4f8a62f64 && git checkout ef28e212bc4016260ceb02fe55dc11d4f8a62f64
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 227c1788d55da8f94127137bcf75ba2bf08b15d8 c415a378508e7372d362d0e38a6227d50037a9b4 && git checkout -B drift-repro 227c1788d55da8f94127137bcf75ba2bf08b15d8 && git merge --no-ff c415a378508e7372d362d0e38a6227d50037a9b4
node scripts/docs-audit/affected-docs.mjs --json 227c1788d55da8f94127137bcf75ba2bf08b15d8

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

Advisory only, and a precision-first one (#9192): a page is listed because it names a
symbol, wire route or SDK method this diff touched — not because it mentions a changed
package. Each row says which anchor put it there, so a wrong row is reportable rather than
merely annoying. To re-verify, run the docs-accuracy-audit workflow scoped to these files:
node scripts/docs-audit/affected-docs.mjs 227c1788d55da8f94127137bcf75ba2bf08b15d8 → pass the list as
args.docs, on the commit named under Which tree this was computed on.

@github-actionsgithub-actionsBot added documentation Improvements or additions to documentation tests tooling labels Aug 25, 2026
…neration commit)
os-regen-merge step 4: spec rebuilt on the merge of origin/main@497ded78,
whole gen chain re-run; the two artifacts the driver took from main's side
(references/kernel/manifest.mdx, strictness-ledger counts) now describe the
merged source. Battery green: migration-registry / liveness / empty-state /
authorable-surface / docs / api-surface / spec-changes / upgrade-guide /
skill-refs / skill-docs / strictness-ledger / variant-docs all PASS. Family
entries and tombstones verified present post-merge.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K93P8PbH7kVjsAXkqzH1zY
@os-support-ai
os-support-ai marked this pull request as ready for review August 25, 2026 06:35
@os-support-ai
os-support-ai added this pull request to the merge queueAug 25, 2026
Merged via the queue into main with commit be21955Aug 25, 2026
36 checks passed
@os-support-ai
os-support-ai deleted the claude/issue-10724-contributes-retirements branch August 25, 2026 06:50
os-support-ai pushed a commit that referenced this pull request Aug 25, 2026
…os-regen deferral)
os-regen-merge.sh sequence: merge committed first, then the full gen chain on
the merged source. check:generated --fix regenerated the 4 provably stale
artifacts; the audit-door entries (this PR) and the #12001/#12006 sibling
entries coexist in the regenerated set.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K93P8PbH7kVjsAXkqzH1zY
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment