Repository navigation
deps(pin): move .objectstack-sha from the next.3 release commit to an objectstack main commit that carries 586705d85d, 1e001c7df and e15897c0f1, and follow the contract changes since next.3 (the first pin bump under ruling B on objectui#12955) #12971
Description
Unblocked: #12957 closed completed when PR #12963 merged as 749ed760; re-derived with no new blocker (unlock comment below).
Filing gate ④, a coordination node: one change unblocks several held cards. It was split out of objectui#12957 by the seat's item 6 amendment (6250606447). Filed and dispatched by domain:ui seat 1 (session_0188TsgPV8kQXfuRuKmkj2ks) on the maintainer's order 「你负责派发处理相关的issue，包括 objectstack 和 objectui 并测试确认。」. ⛔ Not a claim.
Why
objectui#12957 lands the pin mechanism with the first pin at 2c9bf9430a, the @objectstack/spec@18.0.0-next.3 release commit. That pin deliberately reproduces npm next.3. This card makes the first real bump under ruling B. It moves the pin to an objectstack main commit that contains the three back-end halves whose front-end halves wait today:
- objectui#12932 (
586705d85d,emptyStateon bound record blocks); - objectui#12929 (
1e001c7df,publishPackageDraftsbackfill); - objectui#12484 (
e15897c0f1, the approval request page's highlights).
All three are Blocked-by: objectstack-ai/objectstack#24168 today, the next.4 Version Packages PR. This card lets them go ahead without waiting for that release.
Done when
scripts/bump-objectstack.shmoves.objectstack-shato an objectstackmaincommit.git merge-base --is-ancestorproves that commit contains all three shas.pnpm-lock.yamlis regenerated by pnpm.- objectui follows every contract change between
2c9bf943and the new pin, with no tolerant fallbacks. objectui#12957's dev measured the drift at2ee5861: 15 failing cases in 5 files. Re-measure them at the new pin:registry-inputs-spec-parity: a newlist-viewrow, counts 60→61 and 40→41,marketplace:installed-listretired,emptyState;block-config.test.ts:record:jsonis no longer a specPageComponentType, so the stale exclusion goes;record-highlights-layout-9187:emptyState;spec-object-refinements-7715:checkPageRecordBindingPlacement;ObjectTree.schemaTyped-8655.
- Each followed change is named in one changeset, with
Clause-②read per change. A narrowed accept set readsClause-②: no (narrowing)and earns its contract review. - On landing, the seat re-points objectui#12932, Studio 变更 drawer: on publish, offer 补默认值 per new field that carries a default, through the protocol's opt-in backfill (ruling A on cloud#2924) #12929 and approvals B1 closing card (ruling C on objectui#12438): the inbox opens a request on the served sys_approval_request page where the side sheet is, deletes the sheet, and lists each retired pin against its user report #12484 from chore: version packages (next) objectstack#24168 to this card's merge, and releases them.
- The two other consumers of objectui's pin keep their own order:
- tooling(build-console): when the console build installs objectui at
.objectui-sha, objectui's@objectstack/*resolve to this framework tree, not objectui's own.objectstack-shabuild (ruling B on objectstack-ai/objectui#12955, step 3) objectstack#24306 must land before objectstack's.objectui-shamoves past objectui#12957's merge; - objectstack-ai/cloud#3023 must land before cloud's
.objectui-shadoes. - Neither ordering is this card's to change.
- tooling(build-console): when the console build installs objectui at
Generated by Claude Code
Activity
objectstack-fleet commentedon Oct 18, 2026
Cross-repo note from objectstack
domain:specseat 2 (objectstack-ai/objectstack#18549) ·zhuangjianguo· sessionsession_019oFY6Hi35J7atvgVFnmoaK· 2026-10-18T15:59Z · ⛔ Not a claim. Information for this pin bump; no reply is owed.objectstack-ai/objectstack PR #24362 (objectstack-ai/objectstack#24326, in contract review, not yet merged) narrows
PageSchema: a region namedmaintakes onlywidth: 'full', and'small','medium'and'large'there are refused atregions.N.width. No layout at the console pin reads those three values onmain.If this bump's new
.objectstack-shacontains that PR's merge, two objectui fixtures that pin amainregion withwidth: 'large'as accepted turn red. Read at objectui0707b4ae37and atorigin/main:packages/types/src/__tests__/passing-keys-arms-11440.test.ts(the page fixture about:83;SpecPageSchema.safeParse(page).successis expectedtrueat about:104);packages/cli/src/__tests__/validate-passing-keys-11440.test.ts(the same document, about:98).The fix is a one-key delete of
width: 'large'from themainregion in each fixture. It renders the same, and'full'stays accepted. The #24326 dev probed the first fixture against the PR's builtPageSchemaand saw it refused atregions.1.width. The second one was not measured. If the new pin predates the merge, nothing changes for this card.Generated by Claude Code
objectstack-fleet commentedon Oct 18, 2026
Addendum to
6252002879· objectstackdomain:specseat 2 · sessionsession_019oFY6Hi35J7atvgVFnmoaK· 2026-10-18T16:10Z · ⛔ Not a claim; no reply is owed.From the contract review of objectstack-ai/objectstack PR #24362 (
6252093173, PASS), re-read by the seat at objectuiorigin/main. objectui's authoring face rebuilds the page schema fromSpecPageSchema.shape(packages/types/src/zod/layout.zod.ts, about:1069), which drops the spec's refinements. It re-attaches them by identifier:checkPageSourceCompleteness,checkPageRequiresKind,checkPageSlotPairandcheckPageRegionPlacement, at about:1390–:1393and:1455–:1456.PR #24362 adds and exports one more page refinement for exactly this kind of mirror:
checkPageMainRegionWidthfrom@objectstack/spec/ui. If this bump's new pin contains that PR's merge, attach it at both sites besidecheckPageRegionPlacement. Until then,objectui validatekeeps acceptingwidth: 'large'onmainwhile the spec refuses it. This is in addition to the two fixture deletes the first note names.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Unlock:
pm:blocked→pm:queue·domain:uiseat 1 ·session_0188TsgPV8kQXfuRuKmkj2ks· 2026-10-19T00:35Z⛔ Not a claim.
@objectstack/*by commit —.objectstack-sha, a bump script, and dev and CI built from that commit's source, cached by the pin (ruling B on objectui#12955, step 1) #12957 closedcompletedwhen objectstack-ai/objectui PR feat(tooling): objectui consumes @objectstack/* by commit — .objectstack-sha, bump-objectstack.sh, dev and CI built from the pin (objectui#12957) #12963 merged as749ed760(landing note on that card). objectuimainnow carries.objectstack-sha=2c9bf9430a,scripts/objectstack-pin.mjs(with theOBJECTSTACK_ROOTinput) andpnpm run setup.scripts/bump-objectstack.shis onmain. The three shas this card names are on objectstackmain. The pinned checkout is hermetic, so the spec build no longer judges the pin against a newerorigin/main(spec build: build-schemas.ts judges a commit's published schemas againstorigin/main, so rebuilding a released commit (e.g. the 18.0.0-next.3 one) from a clone with a newerorigin/mainfails — not reproducible objectstack#24359 is not a precondition here).Blocked-by:line in the body is replaced with anUnblocked:line in the same act.@objectstack/*by commit —.objectstack-sha, a bump script, and dev and CI built from that commit's source, cached by the pin (ruling B on objectui#12955, step 1) #12957 unblocked one at a time (the maintainer's 「后续改为串行派发」), in the order its seat post [PM seat] domain:ui @ objectui — 🟢 huangyiirene · session_014MXhk1yjWh37d634hLy3p7 · serial (maintainer) #5560 records.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Triage: priority set, p1 ·
target:v18stays ·domain:ui·pm:queue· triage seat ·session_01AavokzJ5DndAwitDXvKy4U· 2026-10-19T00:55ZTriage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015). ⛔ Not a claim, ⛔ not a dispatch.
priority:*was set when this card was filed straight intopm:blocked. Triage grades it now.origin/main749ed76076carries.objectstack-sha=2c9bf9430a. All three named shas (586705d85d,1e001c7df,e15897c0f1) are on objectstackorigin/main(git merge-base --is-ancestor, exit 0 each).Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
This card is now the one gate for five held objectui cards · re-pointed from the npm release · triage seat ·
session_01AavokzJ5DndAwitDXvKy4U· 2026-10-19T05:08ZTriage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015). ⛔ Not a claim, ⛔ not a dispatch.
Why: objectui consumes
@objectstack/*by commit since tooling: objectui consumes@objectstack/*by commit —.objectstack-sha, a bump script, and dev and CI built from that commit's source, cached by the pin (ruling B on objectui#12955, step 1) #12957. A held card therefore waits on.objectstack-sha, not on an npm release. chore: version packages (next) objectstack#24168 merged as725a5b1311, but the release gate refused it (Spec Shape Gate, ci(spec-shape-gate): red since objectui#12957 — the gate runspnpm install --frozen-lockfilein objectui, whose @objectstack/* now resolve from a pin build onlypnpm run setupcreates; the release gate reads this run, so next.4 is held objectstack#24424), so next.4 is not on npm. Its closed state would have unlocked these cards with nothing in the pin.Re-pointed to this card in the same act:
586705d85dvisibleWhencan follow tier and count (新建环境 on paid tiers, the 🔒 line on Free, the 生产 tag with a second environment) (ruling B on cloud#2774) #12931ef65646843, which is not in this card's list of three1e001c7dfe15897c0f1success: falsewithfailures, after #24380) shows no message, and the job is still listed as undoable #13015adb7910f81(PR fix(rest)!: an import undo runs under the object's declared controls for every caller but the system principal objectstack#24431,UndoImportJobResponseSchema.failures)Cascading behind them: approvals B1 (child of objectui#2763): the
sys_approval_requestdetail page on the standard record renderer, composed from A1's data source, A2's record card, A3's progress widget, the action timeline and DeclaredActionsBar #12045 (on approvals B1 closing card (ruling C on objectui#12438): the inbox opens a request on the served sys_approval_request page where the side sheet is, deletes the sheet, and lists each retired pin against its user report #12484), and cloud's objectstack-ai/cloud#2924 (on Studio 变更 drawer: on publish, offer 补默认值 per new field that carries a default, through the protocol's opt-in backfill (ruling A on cloud#2924) #12929).Read now: the pin is
2c9bf9430a(next.3), and none of the five shas is in it. All five are on objectstackmain, so a bump to the newest green objectstackmaincarries every one.Done-when addition: item 1's
merge-base --is-ancestorproof covers all five shas, not three.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Claim: PM loop round 9 · on the maintainer's order, verbatim 「13022 12971 两个p1 插队，并发临时调到3，这两个处理完再下班」 (
priority:p1; triage6257373035and6260132160) · ruling B (6248859734) on objectui#12955, the first real pin bumpSession:
session_0188TsgPV8kQXfuRuKmkj2ksAccount:
sunhaolin(the seat's linked user asGET /useranswers it; the card's assignee)Branch:
claude/issue-12971-pin-bump-mainWorktree:
objectui-issue-12971Domain:
domain:uiSeat:
domain:ui#1File surface:
.objectstack-sha, andpnpm-lock.yamlonly if pnpm regenerates it;2c9bf9430aand the new pin needs. That means the tests, fixtures, mirrors and registryinputsthe drift reds, among them:apps/console/src/__tests__/registry-inputs-spec-parity.test.ts,block-config.test.ts,record-highlights-layout-9187,spec-object-refinements-7715andObjectTree.schemaTyped-8655;packages/types/src/zod/layout.zod.ts(attachcheckPageMainRegionWidthbesidecheckPageRegionPlacement,6252114202);passing-keys-11440fixtures (6252002879);inputsblocks ofpackages/components/src/renderers/action/action-button.tsxandaction-icon.tsx(stop publishing the rowpatchinput that objectstack PR #24442 retired), and the inputs-list pins that read them;packages/core/src/utils/predicate-fields.ts,ActionRunner.ts) and the four renderers'patchforwarding. Those are objectui#13022's, dispatched in parallel onclaude/issue-13022-action-row-patch.visibleWhencan follow tier and count (新建环境 on paid tiers, the 🔒 line on Free, the 生产 tag with a second environment) (ruling B on cloud#2774) #12931, Studio 变更 drawer: on publish, offer 补默认值 per new field that carries a default, through the protocol's opt-in backfill (ruling A on cloud#2924) #12929, approvals B1 closing card (ruling C on objectui#12438): the inbox opens a request on the served sys_approval_request page where the side sheet is, deletes the sheet, and lists each retired pin against its user report #12484, objectui Import Wizard: an import undo that leaves rows unreverted (success: falsewithfailures, after #24380) shows no message, and the job is still listed as undoable #13015). This card makes their commits available; it does not adopt them.Stop on breach; explain in the report.
Container & model:
M,mode:subagent,model: opus(default tier, with aCONTRACT_REVIEW_TIERreview before ACCEPT on the narrowed faces)Clause-②: no
Clause-② reason: a narrowing only. The registry stops publishing the retired row
patchinput onaction:buttonandaction:icon, and@object-ui/typesattachescheckPageMainRegionWidth, soobjectui validaterefuses amainregion that is not full width. Both follow a spec narrowing that already landed, and nothing widens. The changeset carriesClause-②: no (narrowing).Responsibility: n/a — not a defect card
Thread-read: 6260132160
Serial constraints cleared: read at objectui
main47508f3b.domain:devxseat 1, in flight) touchesscripts/objectstack-pin.mjsand Dependabot configuration, not this surface.patch— the Undo snapshot must read the named action'spatch, and the renderers and registry stop forwarding and publishing the row key #13022 mergesmainafter this one, then goes ready.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
os-dev-report
{
"issue": 12971,
"status": "needs_decision",
"branch": "claude/issue-12971-pin-bump-main",
"pr": "#13043",
"session": "session_0188TsgPV8kQXfuRuKmkj2ks",
"premise_still_valid": true,
"summary": "The pin is now objectstack main f4755a68a797, the newest commit that was green when the pin moved (150 check-runs completed, none failed). It contains all five held-card commits plus 69f342b924 and 3a734e0888. CI at head 8f9f299 is green except three reds that each need a ruling: shard 2/16 (the registry-inputs parity test: the list-view row, and emptyState on three record blocks) and Type Check (one test owned by objectui#13022).\n\nPin history (PREMISE CORRECTION): the first move went to 2ec6827962. Its packages read 18.0.0-next.4, which objectstack never published or tagged. That left the live-e2e backend pin (backend.env, held equal to the spec version by ci-cd-pipeline-doc.test.ts) with no published version to pair with. Since then objectstack released 18.0.0-next.5 (820fd5aef0), so I re-ran bump-objectstack.sh to the newest green main commit, f4755a68a7, which reads next.5. backend.env is now next.5, and CI's Live E2E passes with it. pnpm-lock.yaml is unchanged at both pins, because the seven packages declare the same dependency names. Main moved under the branch and added objectui#13008's committed .objectstack-pin manifests, so I merged origin/main and regenerated the manifests with --write-manifests.\n\nChanges followed (narrowings, with the spec's own message):\n- checkPageMainRegionWidth is attached on both page nodes.\n- checkPageRecordBindingPlacement is attached on the page-kind node only, and recorded as not attachable on the page node.\n- action:button and action:icon stop publishing the retired row patch input. This touches the inputs blocks only; forwarding and Undo are left to objectui#13022.\n\nSpec-fact pins re-pointed: list-view joins the judged parity set (60->61 / 40->41); record:json and marketplace:installed-list read RETIRED UPSTREAM; block-config drops record:json; highlights key list gains emptyState; ObjectTree navigation/tree rows gain list-view; ObjectSchema key counts 45->46 (defaultRowScope); the two new checks are registered as refinement exceptions in imported-defaults-8317; the passing-keys-11440 fixtures (2) drop main width large.\n\nNav and claims: a navigation entry's dataSource is mirrored by reference (NavigationItemSchema had been stripping it). The installed-pin claims at 34 sites were re-measured with a node probe and restamped to 18.0.0-next.5, and 2 self-test fixture lines are ledgered as historical. The changeset and the layout guide are written.",
"tests": "Pre-fix measurement at the bump-only commit a3bc0f6: the known-drift files plus packages/components/src/renderers/action/tests gave 983 tests, 20 failed in 6 files (parity 11, ObjectTree 2, passing-keys 3, highlights 1, refinements 2, block-config 1). action-button-undoable-11168 failed only at type-check.\n\nWhole-workspace type-check with --continue at bbf5d4b (old pin 2ec): Tasks 80/81, the single failure being components/action-button-undoable-11168.test.tsx(64,14) TS2353. Re-run at head 8f9f299 (new pin): Tasks 80/81, the same single error, TC_EXIT=2.\n\nLocal vitest shards 1-7 of 16 at bbf5d4b: 2428 files, 31091 tests. Reds were only the parity file and MetadataService.specKeyObjectPayload / retiredObjectEnabled (both since fixed).\n\nLOCAL RUNS REPLACED BY CI (coordinator's order): shards 8-16 were stopped mid-shard-8 (my own PIDs, lock released). A first single-process full run had also been OOM-killed at 11.5 GB RSS after 2h38m and produced nothing. CI ran the full 16-shard battery on every push (the pin file triggers a full run).\n\nTargeted runs at the head, after each fix:\n- nav set (22 files, types and metadata-admin inspectors): 684/684 pass.\n- spec-object-refinements-7715, imported-defaults-8317, passing-keys-arms-11440, validate-passing-keys-11440, ci-cd-pipeline-doc, check-installed-spec-pin-claims, objectstack-pin, bump-objectstack, both MetadataService pins, parity: 825 tests, 6 failed, all of them the parity decision items.\n- pnpm --filter @object-ui/types run type-check: exit 0.\n- type-check:scripts: TCS_EXIT=0.\n- type-check:vitest-config: TCV_EXIT=0.\n\nAblation (committed first, at 37906b1): scripts/ablation-replace.mjs --delete --expect 2 on '.superRefine(checkPageMainRegionWidth)' in layout.zod.ts. Anchor went x2->x0, blob 1e46e7fa45cd->ed50b46f958d. Run of spec-object-refinements-7715, passing-keys-arms-11440 and validate-passing-keys-11440: Tests 3 failed | 97 passed. The three failures are exactly the main-width refusal rows. Expected direction red, observed red. Restore: blob back to 1e46e7fa45cd, equal to HEAD, and git diff HEAD empty.",
"gates": "CI check-runs at head 8f9f299. Success: Build & E2E, Build Docs, Bundle Analysis, Changeset Bump Policy, Changeset Claim Re-read, Changeset Declaration, Changeset Fixed Group Check, Changeset Overwrite Report, Governed Surface Queue Guard, Line Citation Gate, Lint, Live E2E (informational, now on backend 18.0.0-next.5), Repo Gates, Test (dist pins), Test shards 1 and 3-16, label. Skipped: coverage x2, dependabot. FAILURE: Test (shard 2/16), only registry-inputs-spec-parity open items 1+2. FAILURE: Type Check, only open item 3, and CI stops at the first failing package, so the local continue-mode run above is the reading for the other 80 packages. FAILURE: Test (aggregate).\n\nLocal at the head, exit code captured before any pipe:\n- node scripts/check-governed-queue-guard.mjs --test over the PR's paths: NOT GOVERNED, exit 0.\n- pnpm changeset:check: 0.\n- pnpm check:changeset-claims: 0.\n- node scripts/check-changeset-presence.mjs: 0 (36 source files of 13 released packages, 1 changeset).\n- pnpm check:installed-pin-claims: OK, 0.\n- pnpm check:lockfile-integrity: VERDICT clean, 0.\n- pnpm check:lockfile-dedupe: VERDICT deduped, 0. One earlier live-registry reading was red on the unchanged lockfile; the re-run was deduped.\n- check:new-line-citations: 0 new.\n- check:control-bytes: OK.\n- check:action-forward-parity, check:handler-key-reads, check:test-path-roots, check:doc-fences, check:element-data-source-declaration, check:registry-bare-names, check:entry-guard: all 0.\n- check:readme-exports in the built measure tree: 600 real, 0 wrong. 6 plugin-ai blocks NOT MEASURED (that package was not built).\n- NOT MEASURED locally (need built dist and the console build; CI's Repo Gates and Lint passed): check:sdui-registration-pins, check:doc-json, check:doc-examples, check:doc-snippets, check:spec-floors.\n\nobjectui has no dispatch-gates.mjs, so the families were derived by hand and there is no --ran reconciliation.",
"mcp_calls": "0 — no MCP GitHub tool called",
"api_writes": "2 — both through the fleet-write relay (POST /repos/objectstack-ai/objectstack/dispatches, executed as objectstack-fleet[bot]): (1) pr_create = POST /repos/objectstack-ai/objectui/pulls (draft) + POST /repos//issues/13043/assignees (sunhaolin); read-back 12474 bytes sent and stored identical; (2) this os-dev-report comment = POST /repos//issues/12971/comments. git push not counted. No label, assignee or PATCH writes.",
"open_questions": [
{
"question": "D1. The spec's new list-view ComponentPropsMap row (objectstack#22444, 194a42d1f7) makes the registry's list-view judged. It publishes 6 inputs (data, type, columns, filter, sort, options). 38 live spec keys are unpublished (name, label, searchableFields, pagination, kanban, calendar, rowActions, emptyState …), and 6 member pins are owed (list-view.data, columns, filter, sort, options, dataSource). How is this resolved? (Four-axis analysis.) Business need, measured: the inputs feed objectstack's sdui manifest JSX gate and Studio's palette, and a key that is not published stays refused at the JSX gate. Long term: declare only what ListViewBlock is measured to honour. AI errors: refusing an unmeasured key beats advertising one. Startup focus: this card unblocks five held cards.",
"options": [
"A — book every difference OWED to a new owner card with an expiry, under objectui#11111 decision 3 = B (applied to 17.6.0 by objectui#11438 ruling A′), raising the caps from 0; the bump declares nothing. Cost: one ruling, one card, about 44 ledger entries here",
"B — declare in this PR: measure each of the 38 keys against ListViewBlock/ListView, publish the honoured ones, and write the 6 member pins. Cost: card-sized (objectui#11536 took a card for 10 keys); delays the five held cards",
"C — publish all 38 unmeasured. Rejected: it advertises inputs nobody verified the renderer honours, the failure decision 3 = B exists to stop"
],
"recommendation": "A. It is the existing doctrine for a bump surfacing a newly judged block. It keeps the unmeasured keys refused at the JSX gate, which is safer for AI-authored metadata. It lets this p1 coordination card land now. The owner card does B's measurement."
},
{
"question": "D2. emptyState on record:details, record:highlights and record:quick_actions (586705d85d) is unpublished in the registry. The renderer half is held card objectui#12932, which waits on this pin. (Axes.) Business need: real, it is the next p1 card. Long term: the renderer and the input land together. AI errors: publishing now would declare an undrawn key. Startup focus: do not fold feature work into the pin bump.",
"options": [
"A — book three UNPUBLISHED_EXEMPTIONS entries naming objectui#12932 as owner, the precedent objectui#12370 set for record:related_list.parentDataSource (owner objectui#12294); #12932 strikes them when it publishes the input with its renderer",
"B — fold objectui#12932's renderer work into this PR",
"C — publish the input without a renderer (rejected: a declared capability the runtime does not honour)"
],
"recommendation": "A, because it is the accepted next.1 precedent, it keeps the bump mechanical, and #12932 is first in line after this lands. It is not done here only because the dispatch forbade adding an exemption."
},
{
"question": "D3. packages/components/src/renderers/action/tests/action-button-undoable-11168.test.tsx is objectui#13022's Undo pin. Its fixture authors a row patch, which at the new pin is a retiredKey type, so it fails with TS2353 (Type Check red; it still passes at runtime). The claim orders this PR to land first, so it cannot go green alone. (Axes.) Business need: Undo on update rows is real and is #13022's subject. Long term: one owner rewrites the pin to the named action's patch. AI errors: no workaround cast. Startup focus: the cheapest sequencing.",
"options": [
"A — flip the landing order: objectui#13022 lands first at the old pin (its rewrite authors no row patch, which compiles at both pins), then this PR merges main and goes green",
"B — re-partition: hand that test file to this PR, which rewrites the fixture without a row patch. This collides with #13022's in-flight edit of the same file and needs its Undo change to keep the assertions",
"C — keep the order and land #13022's test edit in this PR as a minimal compile fix (cast or retype). Rejected: a workaround on a retired key"
],
"recommendation": "A, because #13022's change compiles on both pins, and it removes the only Type Check red without either branch editing the other's file."
},
{
"question": "D4. Clause-② reading. The claim says 'no — nothing widens', and the changeset carries 'Clause-②: no (narrowing)' as instructed. But at the new pin the strict authoring face also accepts three spec-owned keys by reference that it refused as undeclared before, and each has an unbuilt or planned renderer half: emptyState on the three record blocks (#12932); record on a page node for any pageType (#12931; the spec refuses non-app, and the check cannot read pageType); dataSource on navigation entries (spec 'planned'). Should the contract review read this PR as Clause-② yes (widening by reference) as well?",
"options": [
"A — keep 'no (narrowing)': by-reference spec keys are the spec's contract, and objectui authored no widening",
"B — read 'yes', with both arms recorded in the changeset"
],
"recommendation": "A, with the three listed in the changeset's 'Keys that arrive with the spec, by reference' paragraph (already written). That is the seat's contract-review call."
}
],
"out_of_scope_findings": [
"class: b · reach: public door, measured at head 8f9f299. The spec's NavigationItemSchema refuses { id: 'nav_envs', type: 'object', objectName: 'sys_environment', visible: 'data.total == 1' } (no dataSource) with custom@visible. objectui's NavigationItemSchema and StrictAnyComponentSchema (an app document carrying that entry, the face objectui validate uses) ACCEPT it. Lit control: the same entry with a dataSource is accepted by all three. · evidence: objectstack 8016a9d9dd attaches navItemVisibleDataNeedsDataSource with .superRefine on the spec's nav-item union but does not export it, so objectui's own flat nav object cannot attach it by identifier (the objectui#7715 rule forbids restating it). · Seam: spec:navItemVisibleDataNeedsDataSource (unexported) → renderer: @object-ui/types NavigationItemSchema · fix lands in objectstack (export it), then objectui attaches it · dedupe words: navItemVisibleDataNeedsDataSource export, nav visible data dataSource refused, NavigationItemSchema object-level check",
"carrier: domain:ui seat 1 · noted, not filed: the PR body was written once at head 9310c0e and is stale. Corrections for the seat: (1) the pin is f4755a68a797d5a000cb8bd1c942cb2a9862108e, not 2ec6827962; give the reason (next.4 unpublished, next.5 released at 820fd5aef0) and that it was green with 150 check-runs; the ancestry proof holds at f4755 (all seven shas exit 0; negative control 5b14b7f3ad exit 1; 201 commits from 2c9bf9430a). (2) Add to the red table: ci-cd-pipeline-doc (backend.env OBJECTSTACK_VERSION 18.0.0-next.3 -> 18.0.0-next.5); navigation-spec-parity plus nav-target.navEntryTypes-11790 (objectstack#24332, 26501ae24f: nav-item base dataSource, now mirrored by reference in NavigationItemSchema, the NavigationItem type and NAV_ENTRY_BASE_KEYS); the .objectstack-pin manifests regenerated after merging main (objectui#13008). (3) Installed-pin claims stamp is 18.0.0-next.5, not next.4. (4) Widenings list gains nav dataSource. (5) Gates section: CI is now green except shard 2 and Type Check.",
"承接者:无 · Acceptance notes only: RecordDetailView.recordOpenWaves-13010.test.tsx (3 fixtures) and page-single-h1.test.tsx still author width 'large' on main; nothing validates them, so nothing reds. check:installed-pin-claims and ci-cd-pipeline-doc compare version strings, not commits, so a pin that moves within one version string prompts no re-measurement, and the live-e2e lane pairs a published backend with a console built from a later commit. bump-objectstack.sh publishes each build into the per-machine store /root/.cache/objectstack-pin (objectui#12972 design, about 200 MB per pin)."
]
}
Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Seat decisions on the dev's four questions (report
6267554433) ·domain:uiseat 1 ·session_0188TsgPV8kQXfuRuKmkj2ks· 2026-10-19T12:51Z⛔ Not a maintainer ruling. D1 and D2 apply an existing one. D3 and D4 are the seat's own calls, on sequencing and on its contract-review reading.
list-viewrow): A, under objectui#11111 decision 3 = B (ruling5902351047). Book everylist-viewdifference (the 38 unpublished spec keys and the 6 member pins) as OWED, with an expiry, to the new owner card objectui#13047 (Blocked-by: #12971). The bump declares nothing. This is the doctrine objectui#11438 ruling A′ (5965062249) also ran under. ⛔ C (publishing unmeasured keys) is rejected.emptyStateonrecord:details,record:highlights,record:quick_actions): A. ThreeUNPUBLISHED_EXEMPTIONSentries name objectui#12932 as owner (the objectui#12370 precedent forrecord:related_list.parentDataSource). objectui#12932 strikes them when it publishes the input with its renderer. This is booking under the ruled doctrine, not tolerance, so it does not breach the brief's "no exemption to make a red go away".action-button-undoable-11168TS2353): A, the landing order flips. objectui#13022's PR fix(core,components): an update's Undo covers its named declaration'spatch, and no action renderer forwards a rowpatch(objectui#13022) #13033 (accepted, green ata5cd587dc1) lands first. PR chore(deps): build @objectstack/* from objectstack main 2ec6827962 and follow its contract (objectui#12971) #13043 then mergesmainand goes green, and neither branch edits the other's file.Clause-②: yes. At the new pin the strict authoring face accepts three spec-owned keys it refused before:emptyStateon the three record blocks,recordon any page node, anddataSourceon navigation entries. An author sees that as a widening. The narrowings stand beside it:checkPageMainRegionWidth,checkPageRecordBindingPlacement, and the retired rowpatchinput.Clause-②: yes, with both directions listed (the widening keys by reference, the narrowings with their spec messages).minor, unchanged.CONTRACT_REVIEW_TIERreview judges the declaration.navItemVisibleDataNeedsDataSourceso that objectui can attach it).Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Landed ·
domain:uiseat 1 ·session_0188TsgPV8kQXfuRuKmkj2ks· 2026-10-19T14:26ZPR #13043 merged through objectui's merge queue as
8aa8e3e1e5. Its one parent is988aa9fa15, and the merge is an ancestor oforigin/main.git patch-id --stable)4a292e16(843a09a0; 52 files, +448 / −131).objectstack-shaonmainf4755a68a797d5a000cb8bd1c942cb2a9862108e(objectstackmain,18.0.0-next.5)State change:
Fixes #12971closed this cardcompleted, and this act removespm:dispatched.mainf4755a68a797. That commit carries586705d85d,ef65646843,1e001c7df,e15897c0f1andadb7910f81with no npm release.visibleWhencan follow tier and count (新建环境 on paid tiers, the 🔒 line on Free, the 生产 tag with a second environment) (ruling B on cloud#2774) #12931, Studio 变更 drawer: on publish, offer 补默认值 per new field that carries a default, through the protocol's opt-in backfill (ruling A on cloud#2924) #12929, approvals B1 closing card (ruling C on objectui#12438): the inbox opens a request on the served sys_approval_request page where the side sheet is, deletes the sheet, and lists each retired pin against its user report #12484 and objectui Import Wizard: an import undo that leaves rows unreverted (success: falsewithfailures, after #24380) shows no message, and the job is still listed as undoable #13015 (allBlocked-by: #12971, as the triage re-pointing6260132160set them) are re-derived and released in their own acts, together with the owner card list-view: the spec's new ComponentPropsMap row judges list-view — measure its 38 unpublished spec keys against ListViewBlock, publish the ones it honours, and strike the OWED rows objectui#12971 books #13047.patchinput) is now onmain, so that card's acceptance is complete.Records:
6261969552;6267645900;6267554433;6268625896(theClause-②arm, the seat's own error), then PASS6268975517;6268998510.Generated by Claude Code