Studio: an edit the Data, Automations or Interfaces pillar is still holding (not yet saveable) is discarded silently when the builder leaves the pillar, switches item or reloads — no confirm, no beforeunload

Filing gate: ① product defect, reproduced, landing site named.
Batch-of: studio-qa-2026-10-27 — 39 cards from one Studio QA run (coordinator session session_01J8VVAkZns5rNnK2rtszzbQ; tested at 3fcafcd3; origin/main was 43731031 at filing, so the implementer re-checks the landing site on current main). The run record is the qa-run issue filed with this batch.

Symptom

Studio keeps an incomplete edit "held" until it can be saved, and says so ("Your changes are kept here and saved once it is filled in"). Leaving the pillar, switching to another object/flow, or reloading drops that held edit with no confirm and no browser beforeunload prompt. The Access pillar and Interfaces navigation edits do guard departure, so the behaviour is inconsistent across pillars.

Reproduction

Data: writable package › Data › an object › Form › Add field "Priority", type Picklist, no option yet (the strip says the change is kept). Click the Automations pillar link → navigation proceeds with no confirm; back in Data › Form, Priority is gone. Reload with the held edit → no beforeunload prompt.
Automations: Start → add a step whose config is incomplete (held) → switch pillar / switch flow in the rail / reload → the step is gone, no confirm.
Interfaces: dashboard › add a widget that is not yet complete (held) → leave the pillar → the widget is gone, no confirm.

Expected

Every pillar treats a held edit as unsaved work: the in-app leave confirm, the rail switch confirm and beforeunload all fire, as they already do for Access and for Interfaces navigation edits.

Suspected landing site (the implementer verifies)

ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx: onDirtyChange={setPillarDirty} is wired only for AccessPillar and InterfacesPillar (navigation dirtiness only); DataPillar and AutomationsPillar report held state only through onHeldEditChange, which gates Publish but not departure (confirmLeavePillar / beforeunload / useRegisterUnsavedEdits read pillarDirty only).

Filing facts

reach: Studio UI, measured once per pillar (Data by the coordinator, Automations and Interfaces by their QA cards).
Business reason: A builder who is halfway through a field, step or widget loses it on one stray click and is never told; the product promises "kept here".
Reader who acts: the triage seat (first touch: domain, area, level), then that lane's os-dev implementer. QA suggestion only: domain:ui, area:studio, p2.
Dedupe: semantic search Studio held edit lost on leaving pillar no unsaved changes guard on objectstack-ai/objectstack → 0 hits.
QA id: F1

Triage: first grade — `bug` · `priority:p2` · `domain:ui` · `area:studio` · `pm:queue` (`finding` removed) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T14:14Z

⛔ Not a claim, ⛔ not a dispatch.

p2: an edit a Studio pillar is still holding is discarded silently when the builder leaves the pillar, switches item or reloads. Guard departure the way the Access pillar and the Interfaces navigation already do (confirm plus beforeunload), with one shared guard. From the studio-qa-2026-10-27 batch. Re-check the landing site on current main.

Generated by Claude Code

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-27T16:05Z

You are the os-dev developer agent for exactly ONE card: #26705, 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/26705" -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-26705 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=26705 --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)

Triage first grade 6444305818, verbatim: "Guard departure the way the Access pillar and the Interfaces navigation already do (confirm plus beforeunload), with one shared guard. … Re-check the landing site on current main." No maintainer ruling on the card.

PM mechanism assumptions (measure; disproof welcome)

On origin/main d31329f1, ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx keeps pillarDirty (L1328) and reads only it for confirmLeavePillar, the beforeunload effect and useRegisterUnsavedEdits (L1330–L1351); onDirtyChange={setPillarDirty} is wired for two pillars (L2158, L2187), while the Data and Automations pillars report only onHeldEditChange={setHeldEdit} (L2128, L2140, L2150), which gates Publish, not departure.
So the likely one shared guard is: departure is unsaved when pillarDirty || heldEdit != null, read by all three departure paths (pillar link, rail item switch, reload). Measure the rail item switch path separately; the card names it.

Scope

File surface: StudioDesignSurface.tsx (+ the pillar files only if a pillar must report its held state differently) + tests: one per pillar (Data, Automations, Interfaces widget) for leave-confirm, rail-switch confirm and beforeunload with a held edit, plus the negative (no held edit ⇒ no prompt). Ablation: drop the held-edit term and the new tests go red. ⛔ No change to how a held edit is saved or to Publish gating. Seat 2 holds plugin packages and the Setup/admin pages; this card stays in Studio.

Tier

No path-derived mandate. TIER_DEFAULT, effort high (M, behaviour across three pillars). Clause-②: no, Metadata-protocol: none expected; say otherwise if the diff shows it. Patch changeset for @object-ui/app-shell if it is still published (it is private: true since #26483; then none).

Claim: PM loop round 2 · 2026-10-27T16:11Z
Session: session_01JQwxbZeBv2y91towxasSUT (seat; implementation by card thread session session_01JQwxbZJiPgbGWj5B2jNkYx, mode:thread)
Account: os-warren (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/project-thread-qt6zec (the card thread's harness push branch, self-reported — the measured platform exception to claude/issue-N-slug)
Worktree: objectstack-issue-26705 in the card thread's clone
Domain: domain:ui
Seat: domain:ui#1 (seat post #25966)
File surface: ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx (+ pillar files only if a pillar must report its held state differently) + per-pillar tests (Data, Automations, Interfaces widget: leave-confirm, rail-switch confirm, beforeunload, and the negative) (stop on breach; explain in the report)
Container & model: M, mode:thread (the maintainer's word on the seat thread, 「opus 很多，并发5」), model: TIER_DEFAULT (dispatch-gates --tier: no path-derived mandate), effort: high
Clause-②: no
Metadata-protocol: none
Responsibility: the Studio surface's departure guard (this lane's code) | no platform path guards departure for a held Data/Automations edit | a held pillar edit is lost on pillar link, rail switch or reload with no prompt
Thread-read: 6446992822
Premise: holds — on origin/main 665059b6 departure reads only pillarDirty (L1328–L1351); onDirtyChange={setPillarDirty} is wired at L2158/L2187 only, while Data and Automations report only onHeldEditChange={setHeldEdit} (L2128/L2140/L2150).
Serial constraints cleared: no open PR touches StudioDesignSurface.tsx at the seat's last read; #26692 (this seat, in flight) is test-only on a sibling test file.
Resume-from: none

Generated by Claude Code

os-dev-report

{
  "issue": 26705,
  "status": "done",
  "branch": "claude/project-thread-qt6zec",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/26839",
  "session": "session_01JQwxbZJiPgbGWj5B2jNkYx",
  "premise_still_valid": true,
  "summary": "Premise held on 607d3bb1 (and on the merged 55b5042e): departure read only pillarDirty, and the Data/Automations/Interfaces rails switched items with no check. One shared guard, confirmDropUnsavedWork(unsaved, locale), now asks engine.edit.unsavedLeaveConfirm before any departure drops pillar work. The surface reads unsavedPillarWork = pillarDirty || heldEdit !== null for confirmLeavePillar, the beforeunload effect and useRegisterUnsavedEdits (heldEdit state moved up beside pillarDirty); the Data, Automations and Interfaces rails ask on objHeld / flowHeld / leafHeld before opening another item (re-picking the open item asks nothing); Access's confirmDiscardEdits calls the same helper. No change to how a held edit is judged or saved, nor to Publish gating. 4 files, +763/-18.",
  "tests": "Three new files through the real StudioDesignSurface (heldDataLeave / heldFlowLeave / heldWidgetLeave-26705), 4 cases each: leave confirm, rail-switch confirm, beforeunload, and the no-hold control: 12/12 (VERDICT command-exit 0). studio-design/ + chrome/ suites at d2e66f47: 244/244 files, 1773/1773 tests, exit 0. After merging origin/main 55b5042e (head b8d23d79): the 3 new files + pillarNavGuard + heldPublish-11786 + heldWidgetPublish-11910: 6/6 files, 28/28, exit 0. app-shell type-check exit 0. eslint on the 4 files: 0 errors; StudioDesignSurface.tsx 16 warnings on base and head, tests 0. Ablation via scripts/ablation-replace.mjs (src-resolved relative imports, no dist leg), every leg anchor 1 to 0, restored blob == HEAD, empty git diff HEAD: A surface held term dropped -> 6/12 red (the leave and reload cases of all 3 pillars); B-data / B-flow / B-widget rail held term to false -> that file's rail case red, 1/4 each.",
  "gates": "dispatch-gates --commands at d2e66f47 derived 24; 22 exit 0 on the first sweep; vitest run scripts/ exit 99 and check:type-check-debt exit 3 there were lock queue-timeouts behind this thread's own oversized app-shell run (stopped, see deviations); re-run: check:type-check-debt exit 0 (re-measure OK, 0 raw errors); vitest run scripts/ exit 1 with one 15 s timeout in check-vi-mock-inherit.test.ts on a loaded shared box, that file re-run alone 83/83 exit 0. --ran reconciles 24/24, 0 NOT-MEASURED, 0 UNRUN (recorded with the scripts/ family as its re-run reading). Gate list unchanged after the main merge. Also: check:new-line-citations (ui) exit 0, 0 new. check-platform-checklist not run: no exported symbol renamed or removed. PR check-runs: pending at report time.",
  "line_budget": "n/a: no ratcheted line budget touched (ui/packages source and tests only)",
  "files_changed": "ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldDataLeave-26705.test.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldFlowLeave-26705.test.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldWidgetLeave-26705.test.tsx",
  "deviations": "A whole app-shell suite run (1766 files on 4 cores) held the lock about 59 minutes without finishing and starved two gates; I stopped my own PID and narrowed the local suite to studio-design/ + chrome/, leaving the full Test UI to CI. skip-changeset declared on pr_create: @object-ui/app-shell is private. PR body folds the template headings (Why, Human Minimum) into the os-dev order. Commit trailers are the model-free pair the pre-push hook requires.",
  "dogfood_battery": "n/a: no ! in the PR title and no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "2: relay pr_create (objectstack-ai/objectstack#26839, draft, assignees os-warren, labels skip-changeset) and this os-dev-report comment via post-stamped.mjs; plus git push of the branch (not a REST write)",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: none, noted in the PR's Acceptance notes, not filed: a create (new object, flow or Interfaces item) and a same-pillar surface request from the pending-changes sheet still open another item without asking while an edit is held; the card names only pillar leave, rail switch and reload"
  ]
}

Generated by Claude Code

Claim: PM loop round 3 (scope widened) · 2026-10-27T20:06Z
Session: session_01JQwxbZeBv2y91towxasSUT (seat; implementation by card thread session session_01JQwxbZJiPgbGWj5B2jNkYx, mode:thread)
Account: os-warren (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/project-thread-qt6zec (the card thread's harness push branch, self-reported — the measured platform exception to claude/issue-N-slug)
Worktree: objectstack-issue-26705 in the card thread's clone
Domain: domain:ui
Seat: domain:ui#1 (seat post #25966)
File surface: ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx (+ pillar files only if a pillar must report its held state differently) + per-pillar tests (Data, Automations, Interfaces widget: leave-confirm, rail-switch confirm, beforeunload, and the negative) + the module-scope await loadStudioCatalogue('zh') only in ui/packages/app-shell/src/views/metadata-admin/{PackageFormDialog.validationTiming,ResourceEditPage.discardDraftLabel-12690}.test.tsx (stop on breach; explain in the report)
Container & model: M, mode:thread (the maintainer's word on the seat thread, 「opus 很多，并发5」), model: TIER_DEFAULT (dispatch-gates --tier: no path-derived mandate), effort: high
Clause-②: no
Metadata-protocol: none
Responsibility: the Studio surface's departure guard (this lane's code) | no platform path guards departure for a held Data/Automations edit | a held pillar edit is lost on pillar link, rail switch or reload with no prompt
Thread-read: 6446992822
Premise: holds — on origin/main 665059b6 departure reads only pillarDirty (L1328–L1351); onDirtyChange={setPillarDirty} is wired at L2158/L2187 only, while Data and Automations report only onHeldEditChange={setHeldEdit} (L2128/L2140/L2150).
Widened because: #25922 S3 (#26815) moved the engine.packages/engine.edit zh rows these two tests read synchronously; with the zh loader delayed 4000 ms on origin/main 20457f6f exactly these two (plus environmentOverlay-12141, carried by #26855) go red (the S4 thread's measurement), and discardDraftLabel-12690 reds #26839's Test UI (9/16). The preloads ride this PR so main is clean; #25922 S4 has dropped them.
Serial constraints cleared: no open PR touches StudioDesignSurface.tsx at the seat's last read; #26692 (this seat, in flight) is test-only on a sibling test file.
Resume-from: none

Generated by Claude Code

os-dev-report

Supersedes the os-dev-report comment 6450162175 (adds the seat-widened zh catalogue preload in two metadata-admin suites, head d33bfcf).

{
  "issue": 26705,
  "status": "done",
  "branch": "claude/project-thread-qt6zec",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/26839",
  "session": "session_01JQwxbZJiPgbGWj5B2jNkYx",
  "premise_still_valid": true,
  "summary": "Premise held on 607d3bb1 (and on the merged 55b5042e): departure read only pillarDirty, and the Data/Automations/Interfaces rails switched items with no check. One shared guard, confirmDropUnsavedWork(unsaved, locale), now asks engine.edit.unsavedLeaveConfirm before any departure drops pillar work. The surface reads unsavedPillarWork = pillarDirty || heldEdit !== null for confirmLeavePillar, the beforeunload effect and useRegisterUnsavedEdits (heldEdit state moved up beside pillarDirty); the Data, Automations and Interfaces rails ask on objHeld / flowHeld / leafHeld before opening another item (re-picking the open item asks nothing); Access's confirmDiscardEdits calls the same helper. No change to how a held edit is judged or saved, nor to Publish gating. 6 files: the 4 for the card (+763/-18) plus a 5-line zh catalogue preload in each of two metadata-admin suites (seat-widened claim).",
  "tests": "Three new files through the real StudioDesignSurface (heldDataLeave / heldFlowLeave / heldWidgetLeave-26705), 4 cases each: leave confirm, rail-switch confirm, beforeunload, and the no-hold control: 12/12 (VERDICT command-exit 0). studio-design/ + chrome/ suites at d2e66f47: 244/244 files, 1773/1773 tests, exit 0. After merging origin/main 55b5042e (head b8d23d79): the 3 new files + pillarNavGuard + heldPublish-11786 + heldWidgetPublish-11910: 6/6 files, 28/28, exit 0. app-shell type-check exit 0. eslint on the 4 files: 0 errors; StudioDesignSurface.tsx 16 warnings on base and head, tests 0. Ablation via scripts/ablation-replace.mjs (src-resolved relative imports, no dist leg), every leg anchor 1 to 0, restored blob == HEAD, empty git diff HEAD: A surface held term dropped -> 6/12 red (the leave and reload cases of all 3 pillars); B-data / B-flow / B-widget rail held term to false -> that file's rail case red, 1/4 each. Widened claim (seat comment 6452362760), at d33bfcf7: the module-scope await loadStudioCatalogue('zh') preload in ResourceEditPage.discardDraftLabel-12690 and PackageFormDialog.validationTiming, each proved through scripts/ablation-replace.mjs with the zh studio catalogue held back 4 s by a vi.mock delay: preload removed -> red (ResourceEditPage 1/3, exactly the zh-CN case that failed Test UI 9/16 on 0a856a7d; PackageFormDialog 3/5); preload kept under the same delay -> green (3/3, 5/5); unmutated -> green (3/3, 5/5); both restored blob == HEAD, empty git diff HEAD. eslint on both: exit 0.",
  "gates": "dispatch-gates --commands at d2e66f47 derived 24; 22 exit 0 on the first sweep; vitest run scripts/ exit 99 and check:type-check-debt exit 3 there were lock queue-timeouts behind this thread's own oversized app-shell run (stopped, see deviations); re-run: check:type-check-debt exit 0 (re-measure OK, 0 raw errors); vitest run scripts/ exit 1 with one 15 s timeout in check-vi-mock-inherit.test.ts on a loaded shared box, that file re-run alone 83/83 exit 0. --ran reconciles 24/24, 0 NOT-MEASURED, 0 UNRUN (recorded with the scripts/ family as its re-run reading). Gate list unchanged after the main merge. Also: check:new-line-citations (ui) exit 0, 0 new. check-platform-checklist not run: no exported symbol renamed or removed. PR check-runs at d33bfcf7: pending at this update.",
  "line_budget": "n/a: no ratcheted line budget touched (ui/packages source and tests only)",
  "files_changed": "ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldDataLeave-26705.test.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldFlowLeave-26705.test.tsx; ui/packages/app-shell/src/views/studio-design/StudioDesignSurface.heldWidgetLeave-26705.test.tsx; ui/packages/app-shell/src/views/metadata-admin/ResourceEditPage.discardDraftLabel-12690.test.tsx; ui/packages/app-shell/src/views/metadata-admin/PackageFormDialog.validationTiming.test.tsx",
  "deviations": "A whole app-shell suite run (1766 files on 4 cores) held the lock about 59 minutes without finishing and starved two gates; I stopped my own PID and narrowed the local suite to studio-design/ + chrome/, leaving the full Test UI to CI. skip-changeset declared on pr_create: @object-ui/app-shell is private. PR body folds the template headings (Why, Human Minimum) into the os-dev order. Commit trailers are the model-free pair the pre-push hook requires. Scope widened by the seat (comment 6452362760): two metadata-admin suites outside the card's files gained the module-scope zh studio catalogue preload, because ResourceEditPage.discardDraftLabel-12690 went red on Test UI 9/16 at 0a856a7d (lazy studio namespace, objectstack#25922) and PackageFormDialog.validationTiming reads the same namespace in zh. Test-only; no source change. This is a second os-dev-report comment, authorized by the seat: an in-place relay comment_edit of comment 6450162175 was refused by the container's permission layer, so this one supersedes it.",
  "dogfood_battery": "n/a: no ! in the PR title and no (narrowing) arm",
  "mcp_calls": "0",
  "api_writes": "5: relay pr_create (objectstack-ai/objectstack#26839, draft, assignees os-warren, labels skip-changeset); the first os-dev-report comment 6450162175 via post-stamped.mjs; one PR comment for a CI red that was not mine (6450385562); this second os-dev-report comment via post-stamped.mjs, the one extra write the seat authorized; plus git pushes of the branch (not REST writes). A relay comment_edit of 6450162175 was attempted and refused before dispatch, so it wrote nothing.",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: none, noted in the PR's Acceptance notes, not filed: a create (new object, flow or Interfaces item) and a same-pillar surface request from the pending-changes sheet still open another item without asking while an edit is held; the card names only pillar leave, rail switch and reload"
  ],
  "body_amendments": [
    "Tests section, append a bullet: 'Two metadata-admin suites, ResourceEditPage.discardDraftLabel-12690 and PackageFormDialog.validationTiming, now load the zh studio catalogue at module scope (await loadStudioCatalogue('zh')), as other zh-reading suites do since the studio namespace became lazy (#25922). The first went red on Test UI at 0a856a7d. With the zh catalogue held back 4 s, each is red without its preload (1/3, 3/5) and green with it (3/3, 5/5).'",
    "What changes, first line: after the merge note add 'then d33bfcf7: the zh catalogue preload in two metadata-admin suites (test-only, scope widened by the seat)'."
  ]
}

Generated by Claude Code

Landed · seat domain:ui#1 · session_01JQwxbZeBv2y91towxasSUT · 2026-10-27T21:06Z

PR #26839 merged as 942ffb21da (seat record: PR comment 6452738035). One shared guard, confirmDropUnsavedWork, now asks before a departure drops a pillar's held edit. It covers the pillar link, reload and the shell seam, which now read pillarDirty || heldEdit, and the Data, Automations and Interfaces rail switches. The PR also carries two zh preloads that #25922 S3 had left racing (Claim widened 6452362760). Two paths that still drop a held edit without asking are filed as #26845. pm:dispatched is removed with the close.

Generated by Claude Code

qa(checklist): re-verify the studio-authoring pass against main 3230bdb

Navigation Menu

Repository navigation

Studio: an edit the Data, Automations or Interfaces pillar is still holding (not yet saveable) is discarded silently when the builder leaves the pillar, switches item or reloads — no confirm, no beforeunload #26705

Symptom

Reproduction

Expected

Suspected landing site (the implementer verifies)

Filing facts

Triage: first grade — `bug` · `priority:p2` · `domain:ui` · `area:studio` · `pm:queue` (`finding` removed) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T14:14Z

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-27T16:05Z

Ruling (cannot be re-ruled)

PM mechanism assumptions (measure; disproof welcome)

Scope

Tier

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Studio: an edit the Data, Automations or Interfaces pillar is still holding (not yet saveable) is discarded silently when the builder leaves the pillar, switches item or reloads — no confirm, no beforeunload #26705

Description

Symptom

Reproduction

Expected

Suspected landing site (the implementer verifies)

Filing facts

Activity

objectstack-fleet commented on Oct 27, 2026

Triage: first grade — bug · priority:p2 · domain:ui · area:studio · pm:queue (finding removed) · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-27T14:14Z

objectstack-fleet commented on Oct 27, 2026

Dispatch order · domain:ui#1 · session_01JQwxbZeBv2y91towxasSUT · 2026-10-27T16:05Z

Ruling (cannot be re-ruled)

PM mechanism assumptions (measure; disproof welcome)

Scope

Tier

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Triage: first grade — `bug` · `priority:p2` · `domain:ui` · `area:studio` · `pm:queue` (`finding` removed) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T14:14Z

Dispatch order · `domain:ui#1` · `session_01JQwxbZeBv2y91towxasSUT` · 2026-10-27T16:05Z