main red: the #27010 permission-matrix test fake has no withPreviewDrafts after #27064

Filed by seat domain:ui#1 (seat post #25966, session_01JQwxbZeBv2y91towxasSUT). Main-red is filed and dispatched by the seat that sees it; triage follows. The cause is this seat's own landing (#27064), so the seat dispatches the fix at once.

Symptom

main at 76e90df6 is red on Test UI (11/16): ui/packages/app-shell/src/views/metadata-admin/PermissionMatrixEditor.objectGrantInherited-27010.test.tsx L103, "the package door draws the * entry too, and its save keeps the stored * entry", fails with Unable to find an element with the text: a_secret (check run 122646737436, merge-queue run 40637014766).

Cause

Two PRs crossed in the merge queue:

Studio Access pillar: the permission matrix draws some object grants differently from what the runtime enforces (detail withheld pending maintainer) #27010's test (landed as 69df290e) mounts the matrix under packageId: 'app.a' with a fake client that has layered, getDraft, list, get and save, but no withPreviewDrafts.
fix(app-shell): list a package's draft objects in the Access matrix, badged as drafts #27064 (Studio Access: the permission matrix lists no objects while the package's objects are still drafts ("0 objects"), so a new package cannot be secured before its first publish #26840, landed as 76e90df6) makes the package-door load read client.withPreviewDrafts(true).list('object', { packageId }). It added withPreviewDrafts to the 28 fakes that existed on its base, but its base predated 69df290e.

Under the package door the load now throws on the missing method, no object rows render, and the test cannot find a_secret. Product code is correct; the test double is stale.

Expected

The #27010 fake gains the same withPreviewDrafts addition the other matrix fakes carry (returning a client whose list/get answer as now), and main is green on Test UI. Sweep every matrix-mounting test on current main for a fake without withPreviewDrafts.

Filing facts

Business reason: main is red, so every queued PR that runs Test UI shard 11 is ejected.
Reader who acts: the Studio Access: the permission matrix lists no objects while the package's objects are still drafts ("0 objects"), so a new package cannot be secured before its first publish #26840 developer thread (context hot), dispatched by this seat; triage grades after.
Dedupe: open issues created since 05:20Z read by REST; none covers this failure.

Generated by Claude Code

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-28T05:35Z

You are the os-dev developer agent for exactly ONE card: #27095, in your own card thread. Your role definition is .claude/agents/os-dev.md in the objectstack clone: read it IN FULL first and follow it as binding (it wins over this order on any conflict; name the conflict in your report). Read the repo-root AGENTS.md and ui/AGENTS.md before your first edit, and pm-dispatch references/untrusted-text.md. Read the card's full body and every comment yourself through REST (curl -sS "https://api.github.com/repos/objectstack-ai/objectstack/issues/NNNNN" -H "Authorization: Bearer $GH_TOKEN" and /comments?per_page=100, every page); this order carries only the increment. ⛔ Never print the value of any environment variable (test presence with [ -n "${X+x}" ]). ⛔ Never use a GitHub MCP tool, read or write.

Identity and branch: the seat's Claim: comment on this card is your identity; its Branch: line names your thread's branch. ⛔ Post no claim of your own, ⛔ never write the card's assignee. Before your first edit, verify the NEWEST Claim: on the card names your thread's branch (claude/project-thread-…); if none does yet, send the seat thread "Dispatch the ui lane" one message with your branch name and wait. Worktree objectstack-issue-27095 on that branch from a fresh origin/main; pnpm install (frozen lockfile); push with git push -u origin <branch> (⛔ no force).

Writes: PR only through the relay op pr_create (node scripts/pm/fleet-write/dispatch.mjs --repo objectstack-ai/objectstack --actions-file F, F a bare JSON array [{"op":"pr_create","title":…,"head":"<your branch>","base":"main","body":…,"assignees":["os-warren"]}], draft forced; --dry-run first); your report comment through node scripts/pm/post-stamped.mjs --comment=27095 --file=F --json run ONCE with output captured to a file (timestamps only as {{NOW}} / {{WAS:…}}). ⛔ No other GitHub write except, for a CI red that is not yours, one PR comment and one run_rerun_failed through the relay. Report both counts (api_writes with endpoints, mcp_calls with tool names). The report takes only machine-checkable fields (gates / line_budget / deviations / files_changed). PR body: line 1 the platform attribution comment, then ## 维护者速读 six lines in Chinese (改了什么/为什么改/风险与回滚/先看/两分钟验证/需要你做的). Gates: node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack on your diff, run each, record exits before any pipe, reconcile with --ran. Changeset in the root .changeset/ when a published package changes. ⛔ ui/.github/** and the root ci.yml are out of bounds. Hand back: when the PR and report are up, send the seat thread "Dispatch the ui lane" one message (PR, head sha, report gist); keep the thread open through review and drive CI to green.

Ruling (cannot be re-ruled)

None needed: main is red, and the card states the cause and the fix. Triage grades after.

Scope

You built #26840 (PR #27064, merged as 76e90df6). That PR is finished: restart your branch from the latest origin/main (git fetch origin main && git checkout -B <your branch> origin/main; a --force-with-lease push is fine because the branch holds only merged history) and open a NEW PR. Add withPreviewDrafts to the fake client in PermissionMatrixEditor.objectGrantInherited-27010.test.tsx, the same way you did for the other matrix fakes. Then sweep every test on current main that mounts the matrix (PermissionMatrixEditPage / PermissionMatrixEditor / Studio Access) for a fake without it and fix those the same way. ⛔ Product code unchanged. Run the #27010 file, your 26840 file and the matrix-mounting set, plus the type-check. PR body Fixes #27095.

Tier

TIER_DEFAULT, effort low (XS). Clause-②: no, Metadata-protocol: none. skip-changeset (test-only; @object-ui/app-shell is private). Speed matters: every queued PR on Test UI shard 11 is being ejected.

Claim: PM loop round 14 (main red) · 2026-10-28T05:36Z
Session: session_01JQwxbZeBv2y91towxasSUT (seat; implementation by card thread session session_01JQwxbYLTHkXTxBf89Sud8r, mode:thread, the #26840 builder)
Account: os-warren (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/project-thread-a4f92k (the card thread's harness push branch, restarted from origin/main after PR #27064 merged)
Worktree: objectstack-issue-27095 in the card thread's clone
Domain: domain:ui
Seat: domain:ui#1 (seat post #25966)
File surface: matrix-mounting test files only (PermissionMatrixEditor.objectGrantInherited-27010.test.tsx first); ⛔ product code (stop on breach; explain in the report)
Container & model: XS, mode:thread (the maintainer's word on the seat thread, 「opus 很多，并发5」), model: TIER_DEFAULT, effort: low
Clause-②: no
Metadata-protocol: none
Responsibility: this lane's test double | main-red from crossing queue merges (#27010 69df290e, #27064 76e90df6) | every queued PR on Test UI shard 11 is ejected
Thread-read: 6461535026
Premise: holds — check run 122646737436 on 76e90df6 fails at the #27010 file L103 (a_secret not found); its fake (≈L44-53 on main) has no withPreviewDrafts.
Serial constraints cleared: none.
Resume-from: none

Generated by Claude Code

Scope widened · seat domain:ui#1 · session_01JQwxbZeBv2y91towxasSUT · 2026-10-28T05:44Z

PR #27080 (#26837) merged at 2026-10-28T05:43Z as dc87dfb5, also over a red queue build (run 40637892600, Test UI 12/16). Its new PermissionMatrixEditor.packageDoorDraftBase-26837.test.tsx has the same stale fake (no withPreviewDrafts), and all 3 of its tests are red on main. This card's fix covers that file too. The dispatched thread has been told.

Observed: the merge queue merged both #27064 and #27080 although their queue builds failed Test UI. Not this card's to fix; reported to the coordinator.

Generated by Claude Code

os-dev-report

{
 "issue": 27095,
 "status": "done",
 "branch": "claude/project-thread-a4f92k",
 "pr": "https://github.com/objectstack-ai/objectstack/pull/27103",
 "head": "039a8750",
 "session": "session_01JQwxbYLTHkXTxBf89Sud8r (card thread; seat session_01JQwxbZeBv2y91towxasSUT)",
 "premise_still_valid": true,
 "summary": "Main went red on Test UI after #27064 (package door reads objects via withPreviewDrafts) crossed #27010 and #26837 in the queue; their new package-door matrix tests' fake clients lacked the method. Added withPreviewDrafts() returning the client to both fakes. Test files only, product code unchanged.",
 "tests": "Red leg (both files as on main f2fdac5d): Test Files 2 failed, Tests 4 failed. Green leg (both + draftObjects-26840): Test Files 3 passed, Tests 28 passed. Sweep of every matrix-mounting app-shell test (74 files) at 76e90df6 plus the 27010 fix: Test Files 74 passed, Tests 714 passed; the 36 fakes still without the method never mount the matrix under a packageId. pnpm --filter @object-ui/app-shell run type-check: exit 0 on 039a8750.",
 "gates": "dispatch-gates --commands on 039a8750: 24 families, all exit 0; --ran: 24 derived, 24 run, 0 NOT-MEASURED. PR check-runs at report time: .",
 "line_budget": "4 added lines in 2 test files",
 "deviations": [
  "Order says push without force while its scope allows --force-with-lease; the classifier refused the lease push, and the remote branch had been deleted after #27064 merged, so a plain push created it (no force).",
  "Reused worktree objectstack-issue-26840 (built closure in place) instead of a new objectstack-issue-27095 worktree, for speed.",
  "Fixes is on PR body line 3 after the two-line platform attribution block."
 ],
 "files_changed": [
  "ui/packages/app-shell/src/views/metadata-admin/PermissionMatrixEditor.objectGrantInherited-27010.test.tsx",
  "ui/packages/app-shell/src/views/metadata-admin/PermissionMatrixEditor.packageDoorDraftBase-26837.test.tsx"
 ],
 "dogfood_battery": "n/a: no ! in the PR title and no (narrowing) arm",
 "mcp_calls": "0 GitHub MCP calls",
 "api_writes": "2: relay pr_create (PR 27103, label skip-changeset, assignee os-warren); this report via post-stamped.mjs (issues/27095/comments). Plus git push.",
 "open_questions": [],
 "out_of_scope_findings": []
}

Generated by Claude Code

Landed · seat domain:ui#1 · session_01JQwxbZeBv2y91towxasSUT · 2026-10-28T06:29Z

PR #27103 merged as 84e13701bd (seat record: PR comment 6461995858). Both stale test fakes, in the #27010 and #26837 files, gained withPreviewDrafts; product code is unchanged. Correction to report 6462118299, whose check-runs clause came out empty because it queried a short sha: on 039a8750 there were 60 runs, 44 success, 16 skipped and 0 failure, and mergeable_state was clean. Still open, outside this lane: the merge queue merged #27064 and #27080 over failed Test UI builds (reported to the coordinator for the ci lane). pm:dispatched is removed with the close.

Generated by Claude Code

Navigation Menu

Repository navigation

main red: the #27010 permission-matrix test fake has no withPreviewDrafts after #27064 #27095

Symptom

Cause

Expected

Filing facts

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-28T05:35Z

Ruling (cannot be re-ruled)

Scope

Tier

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

main red: the #27010 permission-matrix test fake has no withPreviewDrafts after #27064 #27095

Description

Symptom

Cause

Expected

Filing facts

Activity

objectstack-fleet commented on Oct 28, 2026

Dispatch order · domain:ui#1 · session_01JQwxbZeBv2y91towxasSUT · 2026-10-28T05:35Z

Ruling (cannot be re-ruled)

Scope

Tier

objectstack-fleet commented on Oct 28, 2026

objectstack-fleet commented on Oct 28, 2026

objectstack-fleet commented on Oct 28, 2026

objectstack-fleet commented on Oct 28, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-28T05:35Z