registry: protocol placeholders overwrite the view:calendar and view:timeline bare-name fallbacks that pending lazy stubs already claim, so resolution depends on chunk load order

[objectstack-fleet]

Path: ② the capabilities an end user meets in the app — view rendering | 缺项 | P2

Split by the triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) · session_01AavokzJ5DndAwitDXvKy4U from objectui#11672 (the 2026-10-06 console dogfood sweep, showcase, framework ce577ec4), item 8. Re-read on objectui main (848ba0e12f) before filing: still live, with no open duplicate. ⛔ Not a claim, ⛔ not a dispatch.

Triage: lands in packages/components/src/renderers/placeholders.tsx (about :137–:141) and packages/components/src/registry/Registry.ts (about :576–:588, and lazyStubFullType) ⇒ domain:ui; rationale: the placeholder's guard asks get(), which ignores pending lazy stubs.

What happens

Every full load logs the registry's own race detector: Component "view:calendar" bare-name fallback is being overwritten by "protocol-placeholder:view:calendar", which a pending lazy stub already claims for "view:view:calendar" (and the same for view:timeline). The message also doubles the namespace (view:view:…).

Done when

• A placeholder never claims a slot that a registered lazy stub owns. The guard checks pending stubs as well as loaded components.
• lazyStubFullType stops adding the namespace a second time.
• Pin: a full load logs no race warning, and an authored view:calendar resolves to the same component whether or not its chunk has loaded.

---

Generated by Claude Code

[objectstack-fleet]

Claim: PM loop round 1
Session: session_01FngvPpdrnhHMdHHq6vwwju
Account: os-justin
Branch: claude/issue-11680-placeholder-lazy-stub-guard
Worktree: objectui-issue-11680
Domain: domain:ui
Seat: domain:ui#2
File surface: the placeholder guard packages/components/src/renderers/placeholders.tsx (registerPlaceholder, about :137–:141); the registry packages/core/src/registry/Registry.ts (lazyStubFullType about :391, and the bare-name fallback contest about :555–:590, only as much as a stub-aware ownership query needs; the card's packages/components/src/registry/Registry.ts path does not exist, and the file lives in @object-ui/core); the tests beside them; and .changeset/11680-*.md (stop on breach; explain in the report)
Container & model: M, mode:subagent, model: opus (dispatch-gates --tier --repo objectstack-ai/objectui: no path-derived mandate; default tier)
Clause-②: no
Thread-read: none
Serial constraints cleared: none blocking. No open PR touches placeholders.tsx or packages/core/src/registry/**. Open PRs objectui#11713 (app-shell), objectui#11715 (@object-ui/types), objectui#11069 and the release PR objectui#11600 were read. This seat's in-flight objectui#11676 remainder is on plugin-timeline's defaults row and the locale packs. Read 2026-10-06T12:39Z.

Why Clause-②: no: the fix changes which registration a bare view:calendar / view:timeline key resolves to while a lazy stub is pending (the registered renderer, never the placeholder), and the text of one warning. No export is added or removed, and no accepted input is widened or narrowed. If the dev adds a public query (for example has() counting pending stubs) to ComponentRegistry's published surface, it reports the built-declaration reading, and the seat amends this line to yes before any review.

---

Generated by Claude Code

[objectstack-fleet]

os-dev-report
{
"issue": 11680,
"status": "done",
"branch": "claude/issue-11680-placeholder-lazy-stub-guard",
"pr": "#11718",
"session": "session_01FngvPpdrnhHMdHHq6vwwju — subagent run, so this is the parent's (domain:ui seat 2 PM) session id",
"premise_still_valid": true,
"summary": "The card's mechanism held on main 9ca3cac. The registry lives in @object-ui/core, not components. registerPlaceholder guarded with get(), which ignores pending stubs, so after the console's registerLazy('calendar' / 'timeline', namespace 'view') the placeholder took the bare key view:calendar, which is the stub's own FULL key, and the eager door then cleared that stub. Fix: registerPlaceholder returns early on the existing public ComponentRegistry.hasLazy(type), with the get() line kept verbatim because check:registry-bare-names re-reads it. LazyEntry now records fullType where registerLazy keys it, and lazyStubFullType is removed. All three readers (the eager door's warning, the lazy door's warning, unregister's ownership test) read entry.fullType, so no reader doubles the namespace. Measured once on the real console registrations: 2 race warnings, with view:calendar resolving to the placeholder, before the fix; 0 warnings and both keys staying with their stubs after it, then resolving to the real Calendar View and Timeline View once the chunks load. Published .d.ts is byte-identical, so Clause-②: no holds.",
"tests": "All at head 409496f unless noted. vitest run packages/core/ exit 0 (197 files, 3915 passed, 27 skipped). vitest run packages/components/ exit 0 (361 files passed, 1 skipped; 3679 tests passed, 24 skipped). 22 cross-package registration and script pins exit 0 (548 passed). New pins: components placeholder-lazy-stub-ownership-11680.test.tsx, 5 tests; Registry.test.ts block 'a stub found under its FULL key names that full type once (objectui#11680)', 4 tests. ABLATION, guard reverted via objectstack scripts/ablation-replace.mjs (anchor 'if (ComponentRegistry.hasLazy(type)) return;' 1 to 0, blob 202f3062ae59 to e8a66e1276cb, anchor count 0 read inside the wrapped run). Predicted: both console-boot rows red, the 3 controls green. Observed at 409496f: 'Tests 2 failed | 3 passed (5)', each failing on 'registering the placeholders raced the stub: expected [ Array(1) ] to deeply equal []'. Restored with blob == HEAD 202f3062ae59 and git diff HEAD empty. The same direction showed at 1ec4369 and 18b8a39. REVERSE VERIFICATION of the doubling fix: Registry.ts replaced by its 9ca3cac blob 69457e1f5bb2 under an EXIT/INT/TERM trap on an absolute path, at 18b8a39. Predicted and observed: 4 new pins red ('expected view:view:calendar to be view:calendar' twice, the silent guard 'expected [] to have a length of 1', and unregister deleting the stub), 'Tests 4 failed | 55 passed (59)'. Restored with blob == HEAD c7d525e29bc9 and git diff HEAD empty. vitest aliases @object-ui/core and @object-ui/react to src, so neither leg needed a dist rebuild. Clause-② evidence: the core Registry.d.ts emitted at base and at head is cmp-equal (exit 0, sha256 prefix 893dd7070a53757b both).",
"gates": [
{
"cmd": "pnpm --workspace-concurrency=2 --filter '@object-ui/components...' build",
"exit": 0,
"head": "18b8a39",
"note": "later commits touch only a test file"
},
{
"cmd": "pnpm --workspace-concurrency=2 --filter @object-ui/core --filter @object-ui/components type-check",
"exit": 0,
"head": "409496f",
"note": "tsc --noEmit && tsc -p tsconfig.test.json; --listFilesOnly shows both touched test files in the test projects"
},
{
"cmd": "pnpm --workspace-concurrency=2 --filter @object-ui/core --filter @object-ui/components lint",
"exit": 0,
"head": "409496f",
"note": "0 errors; no warning on a line this diff added"
},
{
"cmd": "pnpm exec vitest run packages/core/ --maxWorkers=2",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm exec vitest run packages/components/ --maxWorkers=2",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm exec vitest run (22 registration and script pin files, listed in the PR body)",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:registry-bare-names",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:doc-types",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:prompt-keys",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:new-line-citations",
"exit": 0,
"head": "409496f",
"note": "0 new citations"
},
{
"cmd": "pnpm check:control-bytes",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:test-path-roots",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:changeset-claims",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm check:pending-changeset-literals",
"exit": 0,
"head": "409496f"
},
{
"cmd": "pnpm changeset:check (check-changeset-fixed + check-changeset-no-major)",
"exit": 0,
"head": "409496f"
},
{
"cmd": "node scripts/check-changeset-presence.mjs",
"exit": 0,
"head": "409496f"
},
{
"cmd": "node scripts/check-governed-queue-guard.mjs --test (the 5 changed paths)",
"exit": 0,
"head": "409496f",
"note": "NOT GOVERNED"
},
{
"cmd": "CI on PR head",
"exit": null,
"head": "409496f",
"note": "in_progress at report time: 42 check runs, 20 success, 3 skipped, 19 in_progress, 0 failing. Convergence is the PM's"
}
],
"line_budget": "n/a — no skills/** file and no governed line-ratchet ledger is touched",
"files_changed": [
".changeset/11680-placeholder-lazy-stub-guard.md (+12, patch for @object-ui/components and @object-ui/core)",
"packages/components/src/renderers/placeholders.tsx (+21 -1)",
"packages/core/src/registry/Registry.ts (+22 -22)",
"packages/core/src/registry/tests/Registry.test.ts (+76)",
"packages/components/src/tests/placeholder-lazy-stub-ownership-11680.test.tsx (+164, new)"
],
"deviations": [
"Done-when 'a full load logs no race warning' is pinned as a replay, beside placeholders.tsx: the console's two stub declarations, with their real arguments, run through the real registerPlaceholders() and the real SchemaRenderer. It is not a literal console full-load pin. That pin would sit in apps/console/src/tests, outside the claim's file surface, which says to stop on breach. The real full load was measured once instead, with a scratch probe deleted before commit (before: 2 warnings, after: 0). See open_questions.",
"Order Zone 2 #4 falsified: the eager PALETTE_PLACEHOLDER_BLOCKS side effect is not the producer for view:calendar / view:timeline (it holds nav:menu, nav:breadcrumb, global:search, ai:suggestion). The producer is registerPlaceholders(), so the pin replays that. 'Both load orders' is read as stub-pending-at-placeholder-time vs chunk-loaded-first. A placeholder-before-stub order has no producer on this tree (Acceptance note 1).",
"Order Zone 2 #1 confirmed: the card's packages/components/src/registry/Registry.ts does not exist; the file is packages/core/src/registry/Registry.ts.",
"Commit trailers use objectui AGENTS.md's model-free pair (Claude-Session plus 'Co-authored-by: Claude'), not the harness reminder's model-named Co-Authored-By. The PR body ends with os-dev.md's session-URL footer, not the harness reminder's form. Both by stated precedence; nothing was rewritten.",
"No labels written: the order names none, and objectui's path labels belong to labeler.yml, which applied package: core, package: components and tests itself."
],
"mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "3 — each one repository_dispatch to objectstack-ai/objectstack, executed by the fleet-write relay as objectstack-fleet[bot]: (1) pr_create, POST /repos/objectstack-ai/objectui/pulls (draft forced), giving #11718, read back with 12226 bytes sent and stored identical; (2) label-write --assign os-justin, POST /repos//issues/11718/assignees, read back as matching; (3) this os-dev-report comment, POST /repos//issues/11680/comments via post-stamped. Plus 4 git pushes (not REST): empty-branch probe, 1ec4369, 18b8a39, 409496f.",
"open_questions": [
{
"question": "Should a live console full-load pin (import apps/console/src/register-plugins.ts, call registerPlaceholders(), assert zero bare-name race warnings) be added as a file-surface extension?",
"options": [
"A — add it in a patch round on this claim, with the surface amended to include apps/console/src/tests/",
"B — leave the replay pin plus the one-time measurement as delivered"
],
"recommendation": "A, because the replay pin only sees the two keys it names. A future console stub on another protocol key would race the placeholder registrar again, and only a pin over the real registration list would see it. The cost is one console-project test file of about 30 lines, about 18 s under the lock."
}
],
"out_of_scope_findings": [
"carrier: 承接者:无 · noted in PR Acceptance notes, not filed — the placeholder guard is checked at registration time, so a stub declared AFTER a placeholder still loses the key until its chunk loads. No producer: the palette side effect's 4 keys are stubbed nowhere, and registerPlaceholders()'s only caller runs after the console's stubs. Dedupe words: placeholder, registerLazy, order, hasLazy",
"carrier: 承接者:无 · noted, not filed — Registry.register's fallback branch deletes the stub stored under a FULL key when a registration's bare key equals another stub's full type (how the placeholder erased view:calendar's stub). Dormant once the placeholder is fixed: no other colon-shaped bare key on this tree equals a stub's full type. Dedupe words: bare-name fallback, lazyEntries.delete, full key, stub",
"carrier: 承接者:无 · noted, not filed — Registry.getContractConfig re-derives a stub's canonical type with its own startsWith test instead of reading entry.fullType. Same class, but it agrees with fullType for every stub declared here and diverges only for registerLazy('ns:x', …, { namespace: 'ns' }), which nothing declares. Dedupe words: getContractConfig, canonical, lazy stub, namespace",
"carrier: 承接者:无 · observation — view:calendar / view:timeline are not DeclaredNode types in @object-ui/types, so the new pin declares them through CustomNodeRegistry as an application would. Dedupe words: DeclaredNode, view:calendar, CustomNodeRegistry"
]
}

---

Generated by Claude Code

[objectstack-fleet]

os-dev-report
{
"issue": 11680,
"round": "patch round 1",
"status": "done",
"branch": "claude/issue-11680-placeholder-lazy-stub-guard",
"pr": "#11718",
"session": "session_01FngvPpdrnhHMdHHq6vwwju — subagent run, so this is the parent's (domain:ui seat 2 PM) session id",
"premise_still_valid": true,
"summary": "Round 1, two parts. (1) The live console full-load pin was added: apps/console/src/tests/full-load-placeholder-race-11680.test.ts replays main.tsx's boot over the real stub list (../register-plugins), then the real registerPlaceholders(), and asserts no bare-name race warning across the whole load. The capture is installed in vi.hoisted, ahead of the imports. Positive control, chosen in-file: every key a stub held before the placeholders is still held after them (hasLazy), and that set contains view:calendar and view:timeline. The ablation leg was also run. (2) CI red at 409496f (Test (shard 6/8)) was fixed in the stub, not in production: ReportRenderer.test.tsx's hand-built ComponentRegistry stub gains hasLazy, a vi.fn answering false. The palette placeholder side effect calls it at module load, so the suite died at import with 'ComponentRegistry.hasLazy is not a function'. Before that, origin/main 89cc738 was merged in (merge bd6196d, no conflicts). No console source and no production file changed this round, and the changeset stays as is, since none of its sentences went false.",
"tests": "ReportRenderer.test.tsx red reproduced locally before the stub fix: 'TypeError: ComponentRegistry.hasLazy is not a function' at registerPlaceholder ← the palette side effect, 'Test Files 1 failed (1)'. The placeholders.tsx blob and the suite file are byte-equal at bdcd1a9 and 409496f. Green after, at 14498a5: 'Tests 3 passed (3)'. plugin-report package at 14498a5: type-check 0, lint 0 (no new warning), vitest run packages/plugin-report/ 0 (26 files, 239 passed, 7 skipped). New console pin green at bdcd1a9 and at 14498a5. ABLATION of the guard against the console pin, at bdcd1a9, via objectstack scripts/ablation-replace.mjs: anchor 'if (ComponentRegistry.hasLazy(type)) return;' 1 to 0, blob 202f3062ae59 to e8a66e1276cb, anchor count 0 read inside the wrapped run. Predicted: red on the race assertion naming both keys. Observed: 'Tests 1 failed (1)', 'a registration took a bare key another declaration holds: expected [ …(2) ] to deeply equal []', the two captured warnings being the view:calendar and view:timeline races, each now naming the stub's claim once ('claims for "view:calendar"'). Restore proven: blob == HEAD 202f3062ae59 and git diff HEAD empty. Neighbours at bdcd1a9: the 20 console suites importing register-plugins (the new pin among them), plus the components pin, Registry.test.ts and palette-placeholder-blocks: 23 files, 601 passed. At 14498a5: the console pin, ReportRenderer, the components pin and Registry.test.ts: 4 files, 68 passed.",
"gates": [
{
"cmd": "pnpm --workspace-concurrency=2 --filter '@object-ui/console^...' build",
"exit": 0,
"head": "bdcd1a9",
"note": "console dependency closure, 35 of 47 workspace projects; built because the console tsconfig resolves @object-ui/* through dist"
},
{
"cmd": "pnpm --filter @object-ui/console type-check",
"exit": 0,
"head": "bdcd1a9",
"note": "tsc --noEmit && tsc -b tsconfig.node.json --force; --listFilesOnly shows the new pin in the program. 14498a5 touches only a plugin-report test file, outside this program"
},
{
"cmd": "pnpm --filter @object-ui/console lint",
"exit": 0,
"head": "bdcd1a9",
"note": "0 errors, no warning in the new pin"
},
{
"cmd": "pnpm --filter @object-ui/plugin-report type-check",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm --filter @object-ui/plugin-report lint",
"exit": 0,
"head": "14498a5",
"note": "0 errors; warnings only on pre-existing lines"
},
{
"cmd": "pnpm exec vitest run packages/plugin-report/ --maxWorkers=2",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm exec vitest run (20 console register-plugins importers, the components pin, Registry.test.ts, palette-placeholder-blocks)",
"exit": 0,
"head": "bdcd1a9"
},
{
"cmd": "pnpm exec vitest run (console full-load pin, ReportRenderer.test.tsx, components pin, Registry.test.ts)",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:vi-mock-specifiers",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:vi-mock-inherit",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:vi-mock-override-shape",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:new-line-citations",
"exit": 0,
"head": "14498a5",
"note": "0 new citations"
},
{
"cmd": "pnpm check:control-bytes",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:test-path-roots",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:registry-bare-names",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm check:changeset-claims",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "pnpm changeset:check",
"exit": 0,
"head": "14498a5"
},
{
"cmd": "node scripts/check-changeset-presence.mjs",
"exit": 0,
"head": "14498a5",
"note": "6 source files of 4 released packages, 1 changeset declared"
},
{
"cmd": "node scripts/check-governed-queue-guard.mjs --test (the 2 round-1 paths)",
"exit": 0,
"head": "14498a5",
"note": "NOT GOVERNED"
},
{
"cmd": "CI on PR head",
"exit": null,
"head": "14498a5",
"note": "in_progress at report time: 42 check runs, 20 success, 3 skipped, 19 in_progress, 0 failing. At 409496f it was red on exactly Test (shard 6/8) and the aggregate Test. Convergence is the PM's"
}
],
"line_budget": "n/a — no skills/** file and no governed line-ratchet ledger is touched",
"files_changed": [
"apps/console/src/tests/full-load-placeholder-race-11680.test.ts (+51, new; surface widened by the seat)",
"packages/plugin-report/src/tests/ReportRenderer.test.tsx (+4; surface widened by the seat)",
"merge bd6196d of origin/main 89cc738 (no conflicts, no hand edits)",
"unchanged from round 0: .changeset/11680-placeholder-lazy-stub-guard.md, placeholders.tsx, Registry.ts, Registry.test.ts, placeholder-lazy-stub-ownership-11680.test.tsx"
],
"deviations": [
"The console pin is 51 lines against the order's 'about 30': the license header and docblock take 21 lines, and the code is 30.",
"Round-0 miss, named because it is the class of the CI red: round 0's targeted set ran plugin-report's report-bare-key-ownership but not ReportRenderer.test.tsx, the one suite that hand-builds ComponentRegistry inside vi.mock('@object-ui/core') and imports @object-ui/components. A producer that starts calling a public method breaks every hand-built mock of that object, and the targeted set was derived from who reads the touched files, not from who mocks the module they call. The seat's census was re-measured: 3 files vi.mock '@object-ui/core', the other 2 spread the real module, and ReportRenderer.test.tsx is the only one with a hand-built ComponentRegistry. 0 doMock.",
"Console build, type-check, lint and the 23-file neighbour run were measured at bdcd1a9, not 14498a5. 14498a5 adds only the 4-line plugin-report test-stub edit, which is not in the console tsconfig program and not imported by those suites. The 4-file pin set and every check:* gate were re-run at 14498a5.",
"The core and components package suites, type-check and lint were not re-run after the main merge (which brings @object-ui/types and @object-ui/app-shell changes this diff does not touch). Their round-0 results stand at 409496f; CI re-runs them on 14498a5.",
"The merge commit was amended before its first push to carry the model-free trailer pair, so no pushed history was rewritten."
],
"pr_body_patch": {
"instructions": "Two replacements, applied by the seat (the dev did not PATCH the PR). (1) Replace from the line '## Pins' through the line that starts 'Declared to CI, not run locally:' (inclusive). That span covers the old Pins, 'Ablation and reverse verification' and 'Gates' sections. Replace it with replace_1. (2) Replace Acceptance note 4, the paragraph that starts '4. A live full-load pin', with replace_2. Everything else stays byte-for-byte. Neither text contains an angle bracket or a closing keyword.",
"replace_1": "## Pins\n\n- apps/console/src/__tests__/full-load-placeholder-race-11680.test.ts (console project; patch round 1): the live full-load pin. It replays the boot main.tsx performs over the console's REAL stub list, ../register-plugins, and then calls the real registerPlaceholders(). It asserts that the whole load emits no bare-name race warning. The capture is installed in vi.hoisted, ahead of the imports, because the stubs register while register-plugins is imported. Positive control, in the same file: every key a stub held before the placeholders registered is still held after them (hasLazy), and that set includes view:calendar and view:timeline, so the silence covers the two keys the card measured. Because it reads the real list, the next console stub that a placeholder races turns this pin red, with no new card. No console source is edited.\n- packages/components/src/__tests__/placeholder-lazy-stub-ownership-11680.test.tsx (dom project, isolated): stubs declared with the console's arguments, each chunk gated so it lands on command and registers the way the plugin does. Uses the real registerPlaceholders() and the real SchemaRenderer. Console boot order, for view:calendar and view:timeline: no race warning when the placeholders register or when the chunk lands. Before the load, get() is not PlaceholderRenderer and hasLazy() is true, and an authored node never shows "Component Placeholder" and renders the plugin once the chunk lands. After the load, get() is the plugin's renderer. Controls: chunk-loaded-first order (unchanged behaviour), and a protocol key nothing owns (view:kanban) still gets the placeholder. The two node types are declared through CustomNodeRegistry, the way an application declares a type it registers, rather than cast.\n- packages/core/src/registry/__tests__/Registry.test.ts, block a stub found under its FULL key names that full type once (objectui#11680): the eager door names the stub view:calendar, with no view:view:. The eager door still reports a registration whose own full type is the doubled spelling: before, the doubled claim equalled it, so the guard said nothing. The lazy door names the prior stub view:calendar. unregister('view:calendar', 'view') leaves the console's stub: before, it deleted it.\n- packages/plugin-report/src/__tests__/ReportRenderer.test.tsx (patch round 1, the CI red at 409496f): the suite replaces ComponentRegistry with a hand-built partial stub, { get, register }. The palette placeholder side effect, which @object-ui/components runs at module load, now calls hasLazy() first, so the suite failed at import with TypeError: ComponentRegistry.hasLazy is not a function (Test (shard 6/8)). The stub gains hasLazy, a vi.fn that answers false, because no stub is pending in that suite. Its assertions are unchanged, and the registrar does not tolerate a missing method. It is the only vi.mock('@object-ui/core') in the tree that hand-builds ComponentRegistry: the other two spread the real module. Red reproduced locally (the placeholder and suite bytes at bdcd1a9 equal those at 409496f): TypeError: ComponentRegistry.hasLazy is not a function, Test Files 1 failed (1). Green after: Tests 3 passed (3).\n\n## Ablation and reverse verification (one-shot, all from committed state)\n\n- Guard reverted, against the components pin (anchor if (ComponentRegistry.hasLazy(type)) return; deleted through ablation-replace.mjs: anchor 1 → 0, blob 202f3062ae59 → e8a66e1276cb). Predicted: both console-boot rows red, the three controls green. Observed at 409496f: Tests 2 failed | 3 passed (5), both failing on registering the placeholders raced the stub: expected [ Array(1) ] to deeply equal []. Restore proven: blob == HEAD (202f3062ae59), git diff HEAD empty.\n- Guard reverted, against the console full-load pin (same anchor, same tool). Predicted: red on the race assertion, naming both keys. Observed at bdcd1a9: Tests 1 failed (1) on a registration took a bare key another declaration holds: expected [ …(2) ] to deeply equal [], the two captured warnings being the view:calendar and view:timeline races, each now naming the stub's claim once (claims for \"view:calendar\"). Restore proven: blob == HEAD (202f3062ae59), git diff HEAD empty.\n- Doubling fix reverted (Registry.ts replaced by its 9ca3cac blob 69457e1f5bb2 under an EXIT/INT/TERM trap with an absolute path). Predicted: the four new Registry pins red, the rest green. Observed at 18b8a39: Tests 4 failed | 55 passed (59): claimant 'view:view:calendar' twice, expected [] to have a length of 1 for the silent guard, and unregister deleting the stub. Restore proven: blob == HEAD (c7d525e29bc9), git diff HEAD empty.\n- Resolution path: vitest aliases @object-ui/core, @object-ui/components and @object-ui/react to src, so no leg needed a rebuild of dist.\n\n## Gates\n\nThe branch merged origin/main at 89cc738 (merge bd6196d, no conflicts; that merge brings @object-ui/types and @object-ui/app-shell changes this diff does not touch). Head 14498a5 unless noted.\n\n| command | head | exit |\n|---|---|---|\n| pnpm --workspace-concurrency=2 --filter '@object-ui/console^...' build (the console's dependency closure, 35 of 47 workspace projects; needed because the console tsconfig resolves @object-ui/* through built dist) | bdcd1a9 | 0 |\n| pnpm --filter @object-ui/console type-check (tsc --noEmit && tsc -b tsconfig.node.json --force; --listFilesOnly shows the program reads the new pin) | bdcd1a9 | 0 |\n| pnpm --filter @object-ui/console lint | bdcd1a9 | 0 (0 errors; no warning in the new pin) |\n| pnpm --filter @object-ui/plugin-report type-check · lint · pnpm exec vitest run packages/plugin-report/ (26 files, 239 passed, 7 skipped) | 14498a5 | 0 · 0 (no new warning) · 0 |\n| pnpm exec vitest run over the 20 console suites that import register-plugins (the new pin among them), plus the components pin, Registry.test.ts and palette-placeholder-blocks | bdcd1a9 | 0: 23 files, 601 passed |\n| pnpm exec vitest run over the console full-load pin, ReportRenderer.test.tsx, the components pin and Registry.test.ts | 14498a5 | 0: 4 files, 68 passed |\n| pnpm check:vi-mock-specifiers · check:vi-mock-inherit · check:vi-mock-override-shape | 14498a5 | 0 · 0 · 0 |\n| pnpm check:new-line-citations · check:control-bytes · check:test-path-roots | 14498a5 | 0 (0 new citations) · 0 · 0 |\n| pnpm check:registry-bare-names · check:changeset-claims · changeset:check · node scripts/check-changeset-presence.mjs | 14498a5 | 0 · 0 · 0 · 0 |\n| node scripts/check-governed-queue-guard.mjs --test over every changed path | 14498a5 (round-1 paths), 409496f (round-0 paths) | 0, NOT GOVERNED |\n| pnpm --filter '@object-ui/components...' build; @object-ui/core + @object-ui/components type-check and lint; vitest run packages/core/ (197 files, 3915 passed) and packages/components/ (361 files, 3679 passed); the 22 round-0 registration and script pins (548 passed); check:doc-types · check:prompt-keys · check:pending-changeset-literals | 409496f (round 0; the build at 18b8a39) | 0 each |\n\nDeclared to CI, not run locally: the full pnpm test shards, repo-wide pnpm lint, and the remaining workflows. CI at 409496f was red on Test (shard 6/8) and the aggregate Test only, which is the ReportRenderer suite above.\n",
"replace_2": "4. A live full-load pin was added in patch round 1, after the seat widened the claim's surface to one console test file: apps/console/src/__tests__/full-load-placeholder-race-11680.test.ts (see Pins). The one-time measurement in the table above remains as the before/after reading on the real chunks.\n"
},
"mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "1 — this os-dev-report comment, POST /repos//issues/11680/comments via post-stamped, through the fleet-write relay as objectstack-fleet[bot]. Plus 1 git push (not REST) carrying bd6196d, bdcd1a9 and 14498a5 (409496f to 14498a5). No PR PATCH, no label or assignee write.",
"open_questions": [],
"out_of_scope_findings": [
"none new this round. The round-0 observations (placeholder-first order, the eager door clearing a full-key stub, getContractConfig re-deriving the canonical type, view:* not being DeclaredNode types) stand unchanged as carrier: 承接者:无, noted in the PR Acceptance notes, not filed."
]
}

---

Generated by Claude Code

[objectstack-fleet]

ACCEPT: PR objectui#11718 at 14498a5 (Fixes #11680). No contract review is owed; it lands when CI is green

domain:ui execution seat 2 · session_01FngvPpdrnhHMdHHq6vwwju (os-justin) · 2026-10-06T13:54Z. Dev reports: round 0 and patch round 1 os-dev-reports on this card.

• Shape: draft, base main (merged in at 89cc738, merge bd6196d, no conflicts), first line Fixes #11680, Clause-②: no at the start of a line. NOT GOVERNED.

• File surface amended: the claim (6016458678) gains two test files:

  • apps/console/src/__tests__/full-load-placeholder-race-11680.test.ts, the live full-load pin. This answers the dev's open question with A: the card's pin is "a full load logs no race warning", and only the real registration list keeps it true for the next console stub.
  • packages/plugin-report/src/__tests__/ReportRenderer.test.tsx, for the CI red at 409496f.

  No production file changed in patch round 1.

• Seat's own reading of the diff:

  • registerPlaceholder returns early on the existing public ComponentRegistry.hasLazy(type). The get() line stays verbatim, because check:registry-bare-names reads it.
  • LazyEntry records fullType where registerLazy keys it, and lazyStubFullType is deleted. The eager door's warning, the lazy door's warning and unregister's ownership test all read entry.fullType, so none of them doubles the namespace (view:view:calendar).
  • The core Registry.d.ts is byte-identical at base and head, and nothing is added to or removed from an entry, so Clause-②: no holds.

• The CI red at 409496f, and its fix:

  • Test (shard 6/8) failed one suite, ReportRenderer.test.tsx, with TypeError: ComponentRegistry.hasLazy is not a function, raised from the eager palette placeholder side effect at module load.
  • The suite hand-builds a partial ComponentRegistry inside vi.mock('@object-ui/core'). It is the only such stub in the tree: three files mock the module, and the other two spread the real one.
  • The stub gains hasLazy: vi.fn(() => false), because no stub is pending in that suite. The suite's assertions are unchanged.
  • The registrar does not tolerate a missing method, and no suite is skipped.
  • Red was reproduced locally, then green: 3 passed. The vi-mock gates exit 0.

• Changeset (@object-ui/components + @object-ui/core: patch), checked sentence by sentence against the diff:

  • the user-facing sentence: an authored view:calendar / view:timeline loads its plugin, and the boot warning is gone;
  • the mechanism paragraph: the registrar asked about loaded components only, and now hasLazy makes it skip pending stubs, while a key nothing declares still gets its placeholder;
  • the collision-warning paragraph: one spelling, and the ownership comparisons are fixed;
  • the Clause-②: no sentence.

  patch fits a fix that adds no surface.

• Pins and ablations (dev, measured):

  • The console full-load pin replays main.tsx's boot over ../register-plugins and then the real registerPlaceholders(). It asserts no race warning. Its positive control is that every pre-placeholder stub key, view:calendar and view:timeline among them, is still hasLazy afterwards.
  • The components pin uses gated chunks, both orders, and the real SchemaRenderer. The Registry pins check one spelling and that unregister keeps the console's stub.
  • Guard reverted: the components pin goes 2 red / 3 green, and the console pin goes red naming both keys.
  • The doubling fix reverted: the 4 Registry pins go red.
  • Each leg was restored by blob.

• Gates (dev, measured):

  • At 409496f: core + components suites (3915 and 3679 tests) and their type-checks and lints.
  • At bdcd1a9: console closure build, type-check and lint, plus 23 neighbour files.
  • At 14498a5: plugin-report type-check, lint and tests, the 4-file pin set, and every check:* gate.

• PR body: updated in this stroke from the dev's pr_body_patch. The Pins, Ablation and Gates sections and Acceptance note 4 are replaced; the rest is byte-for-byte.

• Out-of-scope, Acceptance notes, not filed:

  • a placeholder registered before a stub still takes the key (no producer on this tree);
  • the eager door's full-key stub clearing is dormant now;
  • getContractConfig re-derives a stub's canonical type (diverges for nothing declared).

• CI on 14498a5: in progress at this read (29 success, 3 skipped, 10 running, 0 failed).

---

Generated by Claude Code

[objectstack-fleet]

Landed: PR objectui#11718 → main e6dcd85, verified by content. Fixes closed this card

domain:ui execution seat 2 · session_01FngvPpdrnhHMdHHq6vwwju (os-justin) · 2026-10-06T14:20Z.

• The merge. Squash e6dcd85 went through the merge queue, with one parent, c910630, and is an ancestor of origin/main. 7 files, +350/−23. Its diff against the parent is byte-identical to the PR's net diff at 14498a5 (md5 c28d7f31… on both, same paths). Clause-②: no, so no contract review was owed (ACCEPT 6017796970).

• Content check against the first parent:

  reading	c910630	e6dcd85
  if (ComponentRegistry.hasLazy(type)) return; in placeholders.tsx	0	1
  function lazyStubFullType in core's Registry.ts	1	0
  LazyEntry records fullType at registerLazy	0	1
  the console full-load pin full-load-placeholder-race-11680.test.ts present	0	1
  hasLazy on ReportRenderer.test.tsx's ComponentRegistry stub	0	1

• Closures. Fixes #11680 closed this card as completed at the merge. No other issue closed in the landing window. pm:dispatched comes off in this stroke.

• Recorded, not filed (Acceptance notes on the PR):

  • a placeholder registered before a stub still takes the key (no producer on this tree);
  • the eager door's full-key stub clearing is dormant now;
  • getContractConfig re-derives a stub's canonical type (it diverges only for a shape nothing declares).

---

Generated by Claude Code