test(runtime): meta-write-org-scope's ADR-0131 D6 overlay case ran 5210 ms and red main once (1a95c746), green on the next heads

triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

Filed by triage from the main-push readings. ⛔ Not a claim. The case was red once and green on the next heads, so this is not a standing main red. Standing ruling: 「平台测试清单反复变红，肯定要修啊」. A test that reds main intermittently is fixed, never retried away.

Reading

Test	Red on	Run	Signature
`packages/runtime/src/meta-write-org-scope.test.ts` › "⭐ [ADR-0131 D6] an overlay-enabled type lands env-wide too, and the read serves it"	`1a95c746` · Test Core (3/8)	41514567522	case ran 5210 ms; file 20 tests, 1 failed

The main CI runs after it (e832ddb1, b16ee8ba, d3175d59, 0814b892) are green. Triage did not reproduce the failure locally.

Wanted

Find out what makes this one case cost about 5 s under shard load. It may boot a full kernel where a narrower harness would do, or wait on a fixed delay. Fix the cost, or the wait, so the case runs well inside the default timeout.
⛔ Do not raise the timeout without naming the cost it covers. No retry, no skip, no quarantine.
Report the case's duration before and after the fix.

Reader who acts: domain:server (packages/runtime).

Dedupe: searched meta-write-org-scope. The only hits are QA run records; no card names the case. New.

Generated by Claude Code

domain:server#2 · os-steve · session_01ADHkmQxwoGb5fSnrFkpxmt

Claim: PM loop round 1 · card thread · 2026-11-02T21:07Z
Session: session_01ADHkmQxwoGb5fSnrFkpxmt (the seat); the card thread is session_01ADHkmRczGnhouFCUs9DfQ2
Account: os-steve (the seat's linked user; the card's assignee)
Branch: claude/project-thread-2g17eb (the card thread's harness-pinned branch; it names no issue, so this claim is the identity record)
Worktree: the card thread's own container checkout
Domain: domain:server
Seat: domain:server#2
Ruling-ref: the standing ruling quoted on the card, 「平台测试清单反复变红，肯定要修啊」. Triage filed and graded it p2.
File surface: packages/runtime/src/meta-write-org-scope.test.ts, plus a shared test harness only if the cost lives there. Product code only if the cost is a needless wait there, with no behaviour moving. ⛔ No protocol.ts (the #25467 E2 freeze), no packages/spec/**, no ui/**.
Container & model: M, mode:thread, model: opus (as opened: default model and effort)
Clause-②: no
Metadata-protocol: none
Responsibility: the code is the ADR-0131 D6 overlay case of meta-write-org-scope.test.ts · the platform path is Test Core shard 3 on every PR and main push · reached by every seat whose PR lands while the case is slow under shard load
Thread-read: none
Premise: holds. The case ran 5210 ms and went red once on 1a95c746 (run 41514567522), then went green on later main heads. The thread measures first.
Serial constraints cleared: no open PR touches meta-write-org-scope.test.ts (read this act).
Resume-from: none — fresh from origin/main 2da42e74

State: pm:queue → pm:dispatched, assigned to os-steve, recorded by this claim.

Generated by Claude Code

test(runtime): alias the @objectstack/lint package root to source (#2…

domain:server#2 · os-steve · session_01ADHkmRczGnhouFCUs9DfQ2
os-dev-report

{
  "issue": 29603,
  "status": "done",
  "branch": "claude/project-thread-2g17eb",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/29622",
  "session": "session_01ADHkmRczGnhouFCUs9DfQ2",
  "premise_still_valid": true,
  "summary": "The red on 1a95c746 (run 41514567522, Test Core 3/8) was a timeout at 5000 ms, not an assertion. The cost is the first view read of the file: getMetaItem's withFormLayoutWarnings imports the lint package root lazily (loadFormLayoutLint, packages/metadata-protocol/src/metadata-diagnostics.ts:481), a cold vitest transform of the lint graph inside the clocked case (first GET 1500 ms, second 2 ms, PUT 19 ms). Test-only fix per the clocked-window rule: a module-top import of the lint package in meta-write-org-scope.test.ts, plus the anchored source alias check:test-source-alias demands for it. No timeout raised, no retry/skip, no assertion moved; product code untouched.",
  "tests": "Star case p50/max, 20 runs, maxWorkers=1, under the lock: before 2da42e74 1202/1740 ms, after 867f4777 20/32 ms. Under 4 concurrent copies on 4 cores: before 1463/1654 ms, after 20/30 ms. File total (sum of 20 cases) p50/max: before 1433/1939 ms, after 218/268 ms; slowest sibling before 130 ms, after 72 ms. pnpm --filter @objectstack/runtime test on 867f4777: Test Files 528 passed (528), Tests 8506 passed | 19 skipped, VERDICT command-exit 0 (also 528/528 on da6bf162). runtime typecheck exit 0. No ablation: no tested behaviour changes; the before/after timing is the reverse reading.",
  "gates": "dispatch-gates --commands on 867f4777: 50 commands, all run, each exit 0; --ran: 50 derived, 50 run, 0 NOT-MEASURED, 0 UNRUN. check:test-source-alias was red on da6bf162 (new unaliased @objectstack/lint import), green after the alias. Lint: root eslint over packages/runtime --no-inline-config exit 0. CI check-runs on #29622: not read (in_progress at report time).",
  "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm",
  "line_budget": "+19 / -0 over 2 files (test 14, vitest config 5)",
  "files_changed": ["packages/runtime/src/meta-write-org-scope.test.ts", "packages/runtime/vitest.config.ts"],
  "deviations": [
    "packages/runtime/vitest.config.ts is outside the delta's expected file surface: one alias line plus comment, forced by check:test-source-alias for the new import (its own remedy). It also moves the protocol's lazy view-read import of @objectstack/lint from dist to the lint source for every runtime test (the runtime suite is green on it).",
    "Branch merged with origin/main 81cd9eb0e before pr_create; main has moved since (#29456, #29528, neither touches these two files); not re-merged."
  ],
  "mcp_calls": "1 — mcp__github__pull_request_read (get, read-back of #29622); no write tools",
  "api_writes": "2 — fleet-write relay pr_create (#29622, label skip-changeset in the same stroke); post-stamped.mjs os-dev-report comment on #29603",
  "open_questions": [
    {
      "question": "check:test-source-alias cannot see a cold load that happens inside another package's code (here metadata-protocol's lazy lint import, reached by a view read). Other runtime tests whose first view read goes through the real protocol pay the same load inside a clocked case. Track it?",
      "options": ["A leave it: no other file has timed out on it; fix per file if one does", "B file a devx card to measure the population (first view read per test file) and preload at module top where it lands in a clocked case"],
      "recommendation": "A for now, because only this file has a red on record and the fix per file is one line; B if a second file reds the same way."
    }
  ],
  "security_findings": "none",
  "out_of_scope_findings": ["carrier: none · noted, not filed — same cold-load shape possible in other runtime test files whose first view read runs the real protocol (see open_questions)"]
}

Stamped 2026-11-02T22:20Z.

Generated by Claude Code

domain:server#2 · os-steve · session_01ADHkmQxwoGb5fSnrFkpxmt

Landing record · PR #29622

domain:server seat 2 · session_01ADHkmQxwoGb5fSnrFkpxmt · 2026-11-02T23:05Z · card #29603

Reading	Probe	Result
Merged	PR #29622 state	merged through the queue at 2026-11-02T22:55:03Z, merge commit `a9b959a0ef`
On main	`git merge-base --is-ancestor a9b959a0ef origin/main`	yes
Queue refs	`git ls-remote origin refs/heads/gh-readonly-queue/*` matching 29622	0
Content	the PR's two files on `origin/main`	the lint-root preload in `meta-write-org-scope.test.ts` and the anchored alias in `packages/runtime/vitest.config.ts`

Landed: the ADR-0131 D6 overlay case no longer pays the cold lint-root transform inside its 5 s clock. Its p50/max went from 1202/1740 ms to 20/32 ms, measured by the card thread. The test is unchanged: no timeout was raised, nothing is retried, and no assertion moved.
ACCEPT should-fix 1 (shard wall time): the Test Core job logs on head 867f4777 print no per-file duration for this file, only the affected-package selection lines. So the shard-side reading is not available, and the record rests on the card thread's local measurement.
Card: test(runtime): meta-write-org-scope's ADR-0131 D6 overlay case ran 5210 ms and red main once (1a95c746), green on the next heads #29603 is closed by the merge. pm:dispatched and the assignee come off in the same stroke as this record.
Follow-up: test(devx): a clocked test whose first served view read cold-loads the lint package root can time out under shard load; find and warm the rest of the population #29627 (domain:devx, p3) is the rest of the population of clocked cases whose first served view read cold-loads the lint root.

Generated by Claude Code

Navigation Menu

Repository navigation

test(runtime): meta-write-org-scope's ADR-0131 D6 overlay case ran 5210 ms and red main once (1a95c746), green on the next heads #29603

Reading

Wanted

Landing record · PR #29622

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

test(runtime): meta-write-org-scope's ADR-0131 D6 overlay case ran 5210 ms and red main once (1a95c746), green on the next heads #29603

Description

Reading

Wanted

Activity

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

Landing record · PR #29622

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions