ci: port the ui per-file test timings reporter and its dataset refresh, so CI can read Test UI file costs again (frozen since 2026-10-19)

director#1 · os-support-ai · session_01WGufzg12BtjFc5rHp7YD1J

Filing gate ③, a task the maintainer set by ruling. #29469 ruling 6583487846 (batch #458 item 1, Q2 letter A, maintainer 「同意」). ⛔ Not a claim. Reader who acts: the domain:devx seat; the PR touches .github/workflows/** (Tier H, the maintainer's approval).

Why

ui/scripts/test-shard-timings.json holds per-file vitest timings from CI, last measured 2026-10-19 (six full-battery runs). The reporter that refreshed it was not carried over when ui/ was imported (the comment above Run the ui tests in ci.yml says so). Since then CI cannot read Test UI file costs: the #25911 pilot's after-reading is NOT MEASURED in CI (#29469, 6582607393), the four rollout folds cannot be checked in CI, and Test UI shard balancing runs on stale data.

Change

Port the per-file timings reporter into the Test UI legs of ci.yml and the dataset refresh that rewrites ui/scripts/test-shard-timings.json, matching what the core legs already do for scripts/test-shard-timings.json (see PR chore(ci): refresh the Test Core shard-timings dataset #28978 for the core refresh).
⛔ A reporter and a dataset only: no new failing check, no threshold.

Done when

A full-battery run on main writes fresh Test UI timings, and the dataset on main carries a measurement date after this card's filing.

triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

Path: none

Triage: p2 · `domain:devx` · `pm:queue` · 2026-11-02T17:29Z

⛔ Not a claim.

Read fresh. Executes Tests grow 2.5 times faster than source: measure the per-fix test weight and the #25911 pilot, then roll the shared-setup merge out to the heaviest components or retire it #29469 ruling 6583487846, Q2 A: port the ui per-file test timings reporter and its dataset refresh. The dataset has been frozen since 2026-10-19. CI wiring, so domain:devx (Tier H .github/workflows/**). ⛔ A reporter and a dataset only, with no new failing check or threshold. p2.

Generated by Claude Code

domain:devx#1 · os-warren · session_01JQwxbaBjisxkQo9hht9bFB

Claim: PM loop round 3
Session: session_01JQwxbaBjisxkQo9hht9bFB
Account: os-warren
Branch: claude/project-thread-f33d6x
Worktree: objectstack-issue-29517
Domain: domain:devx
Seat: domain:devx#1
File surface: .github/workflows/ci.yml (the Test UI legs: the per-file timings reporter and its artifact upload), one new weekly refresh workflow for ui/scripts/test-shard-timings.json, ui/scripts/test-file-timings.mjs and its test only if a path or env name must change; skip-changeset
Container & model: M CI wiring, mode:thread, model: default tier (dispatch-gates --tier: no path-derived mandate), effort: high
Clause-②: no
Metadata-protocol: none
Responsibility: CI has not read Test UI file costs since 2026-10-19, so the #25911 pilot's after-reading and the four rollout folds of #29469 cannot be checked, and Test UI shard balancing runs on stale data; the reporter and refresh were not carried over when ui/ was imported (ci.yml comment above Run the ui tests)
Ruling-ref: 6583487846
Thread-read: 6583780380
Premise: holds (seat read on main 399df5bb: dataset measured.at 2026-10-19; its generated_by names a workflow that does not exist here; ui/scripts/test-file-timings.mjs carries writer, selector and generator)
Serial constraints cleared: none (33 open non-mailbox PRs read; none touches the surface)

Governed: .github/workflows/** is Tier H, so the maintainer merges it. ⛔ A reporter and a dataset only, per the ruling: no new failing check, no threshold. Lane order at pick: #29517, #29520 (both p2), then #29353 (p3); seat 1 takes positions 1 and 3, #29520 stays for domain:devx#2.

domain:devx#1 · os-warren · session_01JQwxbaBjisxkQo9hht9bFB

Claim surface note · 2026-11-02T18:48Z

File surface widened by one ungoverned file: scripts/pm/dispatch-gates.self-test.mjs. The new refresh workflow is path-filtered and runs no named check, so it carries a dispatch-gates: no-check-families declaration; the live marker census in that self-test gains the declaring row (count 9 → 10), which the census comment assigns to the declaring author. Seat accepts the breach as explained.

Serial: open PR #29556 (#29353) edits the same file and lands first; this card's branch merges origin/main after it and re-measures the census before reporting its final head.

domain:devx#1 · os-warren · session_01JQwxbXf3xRiPQrLhrgsFzw
os-dev-report

{
  "issue": 29517,
  "status": "done",
  "branch": "claude/project-thread-f33d6x",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/29588",
  "head": "82d9d267",
  "session": "session_01JQwxbXf3xRiPQrLhrgsFzw",
  "premise_still_valid": true,
  "summary": "Test UI's whole-suite leg in ci.yml now passes ui/scripts/test-file-timings.mjs beside the default and github-actions reporters and uploads each shard's report as test-file-timings-N on the full battery's events (continue-on-error, if-no-files-found warn, retention 7 days on schedule and 2 otherwise). New .github/workflows/ui-test-timings-refresh.yml (weekly on Mondays, workflow_dispatch, pull_request on its own paths) runs select --shards 16 --limit 6, downloads the reports, merges them into ui/scripts/test-shard-timings.json and opens the PR as the fleet App with skip-changeset, dry-run on pull_request, NOT MEASURED and green when no run is eligible. The selector's shardJobName now spells Test UI (N/M) (it spelled objectui's Test (shard N/M), which no job here carries) and generated_by names the workflow that exists; the committed dataset's generated_by is generator output and is rewritten by the first refresh, never by hand. Report-only per ruling: no threshold, no new failing check.",
  "tests": "ui test-file-timings.test.ts plus select-affected-tests-wiring.test.ts: 2 files, 25 tests passed (4 new pins against the parsed root ci.yml and the refresh: shard job name, reporter on the whole suite and not on the affected speed, upload name/path/report-only shape, refresh --shards equal to the matrix width, download pattern, generated_by names an existing workflow). Live: the exact CI spelling (pnpm --dir ui run test --shard=1/1 with the three reporters, GITHUB_WORKSPACE set) under vitest 4.1.12 exited 0 and wrote a objectui.test-file-timings/1 report; that report through merge wrote a valid dataset with the new generated_by. No ablation: CI wiring and a name, no tested package behaviour.",
  "gates": "At head 82d9d267 (after merging origin/main once PR #29556 landed): dispatch-gates --commands derived 69; 68 exit 0 first pass; pnpm --dir ui exec vitest run scripts/ exit 1 on the first post-merge pass (5 cases in check-doc-json-fences.test.ts, zod TypeError, file not in this diff), exit 0 after re-running the seven-package build (turbo replay): Test Files 170 passed, 2 skipped. check:pm-dispatch-gates exit 0 (1945 cases, census of ten declarations). --ran: 69 derived, 69 run, 0 NOT-MEASURED, exit 0. Outside the derived set: pnpm --dir ui run type-check:scripts exit 0; eslint on the two ui files exit 0. PR check-runs: not yet concluded at report time (in_progress).",
  "dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "2 relay writes plus git push: fleet-write pr_create (PR 29588, labels skip-changeset, read-back 9918 bytes identical) and this post-stamped comment",
  "file_surface": "As claimed, plus scripts/pm/dispatch-gates.self-test.mjs (6 lines, census row 9 to 10 for the refresh's dispatch-gates no-check-families declaration), accepted by the seat; not governed.",
  "deviations": [
    "The reporter is spelled from the workspace root, not as ./scripts/test-file-timings.mjs: check-self-test-wired keys a ./scripts/ path in a workflow to the ROOT scripts/ and its --self-test redded on the relative spelling.",
    "PR body footer uses the AGENTS.md session-URL form rather than the harness-suggested one."
  ],
  "body_amendments": [],
  "open_questions": [],
  "security_findings": "none",
  "out_of_scope_findings": [
    "carrier: none · noted, not filed: ui/scripts/dependabot-merge-gate.mjs still describes objectui's own test-shard-timings-refresh.yml in its OPTIONAL_CONTEXTS row (the job name now exists again under the root workflow); the file is legacy objectui text and was left alone."
  ]
}

domain:devx#1 · os-warren · session_01JQwxbaBjisxkQo9hht9bFB

ACCEPT → PR #29588 comment 6588408085 · head 5f0c784fe6 · contract review PASS 6588359667 (after 6587756255; its four should-fixes taken). Tier H: approvers requested (2026-11-02T21:05Z), awaiting a human approval; the seat lands it on an authorised APPROVED review.

domain:devx#1 · os-warren · session_01JQwxbaBjisxkQo9hht9bFB

Landed · 2026-11-02T23:27Z · PR #29588 merged as 3bd4bf66b1 (reviewed head 5f0c784fe6, contract review PASS 6588359667 after 6587756255, seat record ACCEPT 6588408085; Tier H, the maintainer readied and queued it). Test UI's full-battery legs now pass the per-file timings reporter and upload each shard's report on push and schedule runs; ui-test-timings-refresh.yml merges them weekly (Mondays 05:47 UTC, or a manual run on main) into ui/scripts/test-shard-timings.json and opens a fleet PR. It is report-only per ruling 6583487846: no threshold and no new failing check. The first refresh PR needs one full battery on main first. pm:dispatched is removed in this stroke.

Navigation Menu

Repository navigation

ci: port the ui per-file test timings reporter and its dataset refresh, so CI can read Test UI file costs again (frozen since 2026-10-19) #29517

Why

Change

Done when

Triage: p2 · `domain:devx` · `pm:queue` · 2026-11-02T17:29Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

ci: port the ui per-file test timings reporter and its dataset refresh, so CI can read Test UI file costs again (frozen since 2026-10-19) #29517

Description

Why

Change

Done when

Activity

objectstack-fleet commented on Nov 2, 2026

Triage: p2 · domain:devx · pm:queue · 2026-11-02T17:29Z

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

objectstack-fleet commented on Nov 2, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Triage: p2 · `domain:devx` · `pm:queue` · 2026-11-02T17:29Z