spec: ListView.userActions.editInline defaults to true in v18 and editInline: false is the opt-out; objectui reads an absent key as on (maintainer ruling 2026-10-10, after objectui#5144)

Filing gate ②: a decision only the maintainer can take, taken. Filed by the director seat (summon #35, seat post #12708, session_01VYToj6PQehTEKNrjGM9akg) on the maintainer's ruling in that session, verbatim: 「乙 v18 把 spec 默认翻成 true,editInline: false 变成关法。」 (2026-10-10). ⛔ Not a claim, ⛔ not a dispatch; triage grades and routes it (domain:spec by its landing).

The ruling, in one line

A list view is editable in place by default, under the permission gate that already exists. The author who wants a read-only list declares userActions: { editInline: false }. The declared value flips; the rule of objectui#5144 (one vocabulary, the inlineEdit fold, declared = enforced) stands.

Why (the assessment the maintainer ruled on)

ListView.userActions.editInline is declared .default(false) ("Off by default: the list is read-only unless the author opts in", packages/spec/src/ui/view.zod.ts:1535). fix(plugin-list): read userActions.editInline with the spec default, folding a stored inlineEdit into it (objectui#5144) objectui#12036 (objectui#5144, merged 2026-10-09) made the console read it that way, so a grid view that declares neither userActions.editInline nor inlineEdit offers no inline-edit toggle.
Measured reach of the opt-in default: HotCRM (objectstack-ai/hotcrm main f0afcbd) has 40 grid and 1 list views in 14 view files, none declaring either key; objectstack's own examples have 18 files declaring grid or list views and at most 2 touch an inline-edit key; the UI skill and the CLI view scaffold teach nothing about the key, so an AI-built or scaffolded table is read-only in its cells too.
Mainstream practice splits. CRM and record-list products (Salesforce list views, ServiceNow list editing, HubSpot, Pipedrive, Zoho) are on by default, gated by field and record permissions, with an admin switch. Metadata app builders (Odoo list editable, Power Apps model-driven grids, Oracle APEX interactive grids, AppSheet, Retool) are opt-in per view. The maintainer chose the first posture for this platform: the permission gate already does the protecting, and the per-view key is the exception, not the rule.
The permission gate is untouched: objectui offers the toggle only when the object is editable in place (ADR-0103 bucket, userActions.edit, the server's effective API operations) and the current principal can update the object (packages/plugin-list/src/ListView.tsx:2007).

Spec half (this card, `@objectstack/spec`)

packages/spec/src/ui/view.zod.ts:1535: editInline: z.boolean().default(true), with the describe text rewritten: on by default; a user who may update the object edits a cell in place with the field's type-aware widget; declare editInline: false to make the list read-only in place.
The reference pages (content/docs/references/ui/view.mdx, page.mdx) regenerate through their generator.
Changeset: Clause-②: yes (widening) (a view declaring nothing gains an affordance it did not have); the launch-window BREAKING banner names the author who relied on the documented default for a read-only list and the remedy editInline: false; the ADR-0087 disposition marker as check:adr-0087-registration demands. Measured in-repo reach of the remedy: cloud packages/service-tenant/src/pages/sys-package.page.ts:105 already declares editInline: false explicitly; no objectstack example declares false.
skills/objectstack-ui/rules/list-views.md ("Configuring a List View"): one paragraph. Inline editing is on by default under the permission gate; a list whose records are read-only by nature (log, audit, history, report roll-up) declares userActions: { editInline: false }. skills/** is Tier H, so the PR takes the maintainer's APPROVED review.

Consumer half (objectui, filed by the claiming seat as the consumer card of this contract-first split)

packages/plugin-list/src/ListView.tsx inlineEditOffered and packages/app-shell/src/views/InterfaceListPage.tsx:571 read an absent key as the spec default, on: editInline !== false. The fold in packages/core/src/utils/normalize-list-view.ts is unchanged (inlineEdit: true reads on and opens in edit mode; inlineEdit: false reads off; an explicit editInline wins).
The table in content/docs/plugins/plugin-view.mdx gains the row "neither key: offered, opens out of edit mode". The pins flip: ListView.permissions.test.tsx, normalize-list-view.inlineEditFold-5144.test.ts, the interface-page tests. The console's toggle stays session-only (objectui#5144, ruling E).
The ADR-0047 interface page gains the toggle by default, as the maintainer accepted in the assessment.
Sequence: this card lands first; objectui follows. In the window objectui still reads an absent key as off, so nothing regresses against today.

Acceptance

A grid view that declares nothing offers the inline-edit toggle to a user with update on the object and not to a user without it; userActions: { editInline: false } removes it; inlineEdit: true still opens the grid in edit mode.
HotCRM's 41 undeclared views regain cell editing after both packages upgrade, with no metadata change in that repo.

Not taken

Declaring userActions.editInline: true on every existing view (the cards filed under that premise in the same session, views: declare userActions.editInline: true on every grid and list view, so cell editing survives the objectui upgrade (objectstack-ai/objectui#5144) hotcrm#2048 and ui skill, CLI view scaffold and examples: a business object's grid view declares userActions.editInline: true, so AI-built and scaffolded tables keep cell editing under the spec default (objectui#5144) #22604, are superseded by this ruling).
An app-level default between the spec default and the view (a new surface with one pull today).

[finding] userActions.editInline has two different defaults across surfaces, and neither is the spec's .default(false) objectui#5144 (the 2026-08-18 B-fold ruling; its spec-default convergence stands, with the default's value changed here), fix(plugin-list): read userActions.editInline with the spec default, folding a stored inlineEdit into it (objectui#5144) objectui#12036.

Filed by the director seat; generated by Claude Code

Triage: first grade, `enhancement` · `priority:p1` · `target:v18` · `domain:spec` · `area:records`, kept in `pm:queue`. It executes the maintainer's ruling

Triage seat (objectstack-wide, seat post #6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-10T04:55Z. ⛔ Not a claim, ⛔ not a dispatch.

Lane: packages/spec/src/ui/view.zod.ts, the generated references and skills/objectstack-ui. That is domain:spec, and the skills/** edit takes the maintainer's APPROVED review (Tier H), as the card says.
Why p1: the default flips on the v18 line, so it must ride the v18 release. After 18.0 it would be a second behaviour change in a minor (发版本优先的都 p1).
Direction: as filed, from the maintainer's 「乙 v18 把 spec 默认翻成 true,editInline: false 变成关法。」 ⛔ No app-level default. The permission gate is unchanged.
The consumer half is objectui's. The claiming seat files it as the contract-first follow-up, with Blocked-by: this card. It reads editInline !== false and needs no new export from @objectstack/spec. Its landing waits for this card, not for the install face. In the window, objectui still reads an absent key as off, so nothing regresses.

Claim: PM loop round 5 (#22605, the maintainer's ruling 乙: ListView.userActions.editInline defaults to true on the v18 line, and editInline: false is the opt-out) · 2026-10-10T05:33Z
Session: session_01KNKBCRDJCu5tGy3TEbvtrF
Account: zhuangjianguo (the seat's linked user as GET /user answers it; the card's assignee from this act)
Branch: claude/issue-22605-edit-inline-default-on (PR 1); claude/issue-22605-skill-list-views (PR 2)
Worktree: objectstack-issue-22605
Domain: domain:spec
Seat: domain:spec#3 (seat post #18883)
File surface (at origin/main d748ae80af or later; stop on breach and explain in the report). Two PRs, because the card's skills/** paragraph is Tier H and the default flip must not wait on it:

PR 1 (Part of #22605, the queue):
- packages/spec/src/ui/view.zod.ts (about :1535): editInline: z.boolean().default(true), with the describe text rewritten as the card states;
- the reference pages through their generator (content/docs/references/ui/view.mdx, page.mdx) and any other generated artifact the default renders into;
- the ADR-0087 disposition check:adr-0087-registration demands for a default flip (a D3 entry and its regenerated registry, if that is what it demands);
- .changeset/22605-*.md (@objectstack/spec; Clause-②: yes (widening), with the launch-window BREAKING banner naming the author who relied on the documented default for a read-only list, and the remedy editInline: false);
- pins in packages/spec/src/ui/.
PR 2 (Fixes #22605, Tier H, the maintainer's APPROVED review): skills/objectstack-ui/rules/list-views.md, "Configuring a List View", one paragraph as the card states, through the skill's generator if it has one.
⛔ No app-level default. ⛔ The permission gate is untouched. ⛔ Nothing in objectui: its consumer half is filed in this act as the contract-first follow-up.
Container & model: S, mode:subagent, model: CONTRACT_REVIEW_TIER — MANDATORY: dispatch-gates --tier on skills/objectstack-ui/rules/list-views.md prints "Model tier — MANDATORY: … (derived from the file surface, not recalled)" for 'skills/**' (clause ①, the 2026-09-10 ruling 「必须 fable的还包括对外发布的skills」), and one such path closes the downgrade exit for the whole card, so one dev at that tier builds both PRs. The default flip widens what a view that declares nothing offers: the contract review at CONTRACT_REVIEW_TIER is owed before PR 1 enqueues; PR 2's skills/** hunk is reviewed at that tier as well. (Corrected by the seat at 2026-10-10T05:36Z: this line first read model: default tier.)
Clause-②: no
(Corrected by the seat at 2026-10-10T07:24Z: this line first read Clause-②: yes (widening). A moved default neither widens nor narrows the accept set; the contract review 6095058376 judged the arm wrong, and the breaking signal rides the changeset banner and the ADR-0087 marker.)
Responsibility: n/a — not a defect card (a ruled change, filing gate ②)
Thread-read: 6093968052
Serial constraints cleared: the file lists of the open PRs, read at 2026-10-10T05:33Z; none touches view.zod.ts, skills/objectstack-ui/** or references/ui/view.mdx. The other area:records cards in flight (ui(pages): a slotted page that authors both details and tabs loses its details slot in the console — sys_user_detail authors both, so its sections and hideFields (incl. ban_reason) never apply #22568 domain:spec, Comment reactions (ruling A amended on #22505): a reaction is the reactor's own sys_comment_reaction record, and sys_comment.reactions retires with no aggregate and no data migration #22566 and plugin-audit: sys_activity.actor_name is declared but never written, so every record History entry reads "Unknown user" #22510 domain:services) name none of these files. ui(pages): a slotted page that authors both details and tabs loses its details slot in the console — sys_user_detail authors both, so its sections and hideFields (incl. ban_reason) never apply #22568 shares only registry.ts (ordinary concurrency; the maintainer's temporary rule-A relaxation applies).

The ruling, verbatim (the director seat's filing): 「乙 v18 把 spec 默认翻成 true,editInline: false 变成关法。」 Triage (6093968052) graded it priority:p1: the default flips on the v18 line, so it must ride the v18 release.

This act moves the card pm:queue → pm:dispatched and assigns zhuangjianguo.

Generated by Claude Code

os-dev-report

Report for #22605 covering both PRs — #22629 (spec half, Part of) and #22630 (skill half, Fixes). Posted 2026-10-10T06:54Z.

{
  "issue": 22605,
  "status": "done",
  "branch": "claude/issue-22605-edit-inline-default-on (PR 1); claude/issue-22605-skill-list-views (PR 2)",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/22629 (PR 1, Part of #22605, draft, queue); https://github.com/objectstack-ai/objectstack/pull/22630 (PR 2, Fixes #22605, draft, Tier H)",
  "session": "session_01KNKBCRDJCu5tGy3TEbvtrF",
  "premise_still_valid": true,
  "summary": "PR 1 (#22629, head 24d9bdcbf0): UserActionsConfigSchema.editInline is z.boolean().default(true) with the describe rewritten; the declared default change is ledgered in DEFAULT_CHANGES_BY_MAJOR[17] (ui/UserActionsConfig:editInline false → true, CURRENT_MAJOR is the package major 17); the D3 semantic entry list-view-edit-inline-default-on is added at protocol 18 as an entries/semantic file and the registry regenerated; authorable-defaults/ui.json and the four reference rows (view.mdx ×3, page.mdx ×1) regenerated; changeset 22605-spec-list-view-edit-inline-default-on.md (@objectstack/spec minor, Clause-②: yes (widening), BREAKING banner naming the author who relied on the documented default for a read-only list and the remedy editInline: false, marker registered list-view-edit-inline-default-on); five pins in view.test.ts. PR 2 (#22630, head e6f3eb158a): one paragraph in skills/objectstack-ui/rules/list-views.md after the column-features table, with the 维护者速读(草稿) section in its body (席位意见 left empty). Premise verified on origin/main: view.zod.ts:1535 read .default(false) with the Off-by-default describe; no in-repo metadata declares editInline at all; the skill file is hand-written (the only generated lifts in skills/** are references/_index.md and references/react-blocks.md). Not taken: an app-level default; editInline: true on existing views; nothing in objectui (its half is objectui#12086). Both PRs merged origin/main before opening (PR 1 at e22315238f, PR 2 at 83b8b80728); origin/main then moved two more commits past PR 1 with no path overlap.",
  "tests": "All under the verify lock with OS_VERIFY_LOCK_SLOT=issue-22605 (shared-box seconds). PR 1 at head 24d9bdcbf0 after the merge: pnpm --filter @objectstack/spec build — green; check:generated — '✓ All 15 generated artifacts are up to date'; pnpm --filter @objectstack/spec exec vitest run --project local --maxWorkers=2 (the package test script) — 'Test Files 642 passed (642) · Tests 19150 passed | 1 todo (19151)'; targeted src/ui/view.test.ts src/ui/page.test.ts src/migrations/migrations.test.ts src/migrations/semantic-relevance.test.ts — '4 passed · 740 passed'; repo-project scripts/step18-rationale-merge.test.ts scripts/authorable-defaults.test.ts scripts/build-schemas-check-mode.test.ts (+semantic-relevance) — '4 passed · 146 passed' (12m39s under the lock); pnpm --filter @objectstack/spec typecheck — tsc --noEmit, check:scripts-typecheck, check:test-typecheck all green ('52 file(s) / 246 error(s) / 135 pinned signature(s) held', ledger unchanged); consumer test examples/app-showcase test/inert-wirings.test.ts (the one non-spec test reading view-level userActions) — '1 passed · 38 passed'; spec test:repo beyond the three files declared to CI. The build printed '📌 1 declared default change(s) accepted (#4666): ui/UserActionsConfig:editInline: false → true'. Public type surface unchanged (check:api-surface green: .default(true) vs .default(false) infers the same type), so no consumer package owes tests beyond the one run. Ablation (committed fix, at 24d9bdcbf0): node scripts/ablation-replace.mjs --file packages/spec/src/ui/view.zod.ts --anchor 'editInline: z.boolean().default(true)' --replacement 'editInline: z.boolean().default(false)' -- vitest run src/ui/view.test.ts: on-disk proof anchor 1 → 0, replacement 0 → 1, blob 6979ade6158a → e0e02cd7f858; direction observed red, exactly the two default pins ('should apply default values', 'an absent key parses to true': expected false to be true), '2 failed | 462 passed (464)'; restore proved blob == HEAD (6979ade6158a), git diff HEAD empty, tree clean. Resolution path: view.test.ts imports './view.zod' from src, no dist between mutation and reading, so no rebuild was needed for this ablation. Lint as a proven narrowing: pnpm exec eslint --no-inline-config --format json over the .ts/.mjs files of the two-point diff — 24 files, 0 errors, 0 warnings (JSON counts); population from eslint.config.mjs is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} minus ignores (.md/.mdx/.json outside it); invariance: no parserOptions.project and no typed rules, so untouched files' verdicts cannot move. PR 2: node scripts/check-skills-token-ratchet.mjs — 'skills/objectstack-ui/rules/list-views.md is 3136 tokens (ceiling 3141; headroom 5)', '54 authored bundle file(s) within their ceilings'.",
  "gates": {
    "pr1": "dispatch-gates --commands --repo objectstack-ai/objectstack at 24d9bdcbf0: 115 families; --ran reconcile: 115 derived, 114 run, 0 NOT-MEASURED, 1 UNRUN (pnpm check:dual-build-cjs-loads — NOT MEASURED by dispatch, whole-workspace build). 113 of the 114 exited 0 (after building @objectstack/formula, lint, client, client-react, objectql under the lock for the gates that load them: check:doc-formula-expressions, check:doc-security-posture, check:skill-examples, check:docs-transcript-drift, check:lean-entry-closure all re-ran green). pnpm check:dts-closure exited 1 for missing dist under plugin-approvals and plugin-dev — it reads every package's built declarations, so NOT MEASURED locally (CI Build Core), not red. Artifact-roster block (49): 46 run, all exit 0 except pnpm check:published-readme-exports exit 3 (44 unbuilt packages, NOT MEASURED); PR-context guards wired to #22629: check-closing-target-claim exit 0 ('binds no closing keyword to a card; a Part of #N declaration is not a closing target'; a first run without PR_HEAD_REF exited 2 NOT MEASURED), check-partof-closing-keyword exit 0, check-single-claim-paths exit 0. Staleness: the derivation tree is 2 commits behind origin/main 83b8b80728; the one moved derivation input is scripts/platform-object-tenancy-census.json (its family ran green on the older copy; no path overlap). check-adr-0087-registration --base origin/main exit 0; check-changeset-no-major --base origin/main exit 0; check-empty-changeset exit 0; check:authorable-surface, check:docs, check:migration-registry, check:spec-changes, check:upgrade-guide, check:api-surface all exit 0. Ran lists and logs: scratchpad issue-22605/gates-pr1/, gates-pr1-rosters/.",
    "pr2": "dispatch-gates --commands at e6f3eb158a (after merging origin/main 83b8b80728): 23 families; --ran: 23 derived, 23 run, 0 NOT-MEASURED, 0 UNRUN, all exit 0 (check:doc-formula-expressions after building formula and lint; check-skills-token-ratchet, check:pm-governed-merges, check:skill-compatibility, check:skill-frame-sync, check:skill-identifier-liveness among them). check-governed-merges --test skills/objectstack-ui/rules/list-views.md: governed, Tier H (exit 3 is the predicate's governed answer); not in a generated lift. PR-context guards wired to #22630: check-closing-target-claim exit 0 ('closes #22605, and each carries a Claim whose Branch line names claude/issue-22605-skill-list-views'), check-partof-closing-keyword exit 0, check-single-claim-paths exit 0. dispatch-gates --tier on the skill path: MANDATORY claude-fable-5-1, the tier this run ran at."
  },
  "line_budget": {
    "pr1": "9 files, +233/−7 off merge base e22315238f (labeler: size/m); no skills/** file; well under the 3000-line human-merge limb.",
    "pr2": "1 file, +5/−0. rules/list-views.md 300 → 305 lines; 3068 → 3136 tokens (ceil(bytes/4): 12269 → 12543) against ceiling 3141, headroom 5, paid from existing headroom, nothing re-wrapped or deleted. Package: sum of every skills/*/SKILL.md 4410 → 4410 lines (none moves). No token count beyond the ratchet's own convention exists."
  },
  "adr_0087_disposition": "registered list-view-edit-inline-default-on — a D3 semantic entry at step 18 (packages/spec/src/migrations/entries/semantic/18.list-view-edit-inline-default-on.ts, concatenated into registry.ts by gen:migration-registry, byte-exact after merging the sibling entry page-slots-details-beside-tabs-refused, not adjacent in id order). The gate refuses not-required (no-migration-prescription) when the body prescribes a remedy, and this body prescribes editInline: false; precedent: view-pagination-page-size-default-50 and protocol 12's rest-requireauth-default-flip. No tombstone, no RETIRED_KEYS_BY_MAJOR row, no D2 conversion (a mechanical pass in either direction would defeat or duplicate the ruling). A second ledger also applies and is written: DEFAULT_CHANGES_BY_MAJOR[17] carries ui/UserActionsConfig:editInline false → true, which check:authorable-surface accepted. No STEP18_RATIONALE fragment added (the default-flip precedents carry none). spec-changes.json and docs/protocol-upgrade-guide.md not regenerated: the PR-stage checks project in memory and compare no committed copy (#22449 B′), and origin/main turned the guide into a pointer stub during this run.",
  "files_changed": {
    "pr1": [
      ".changeset/22605-spec-list-view-edit-inline-default-on.md",
      "content/docs/references/ui/page.mdx",
      "content/docs/references/ui/view.mdx",
      "packages/spec/authorable-defaults/ui.json",
      "packages/spec/scripts/lib/default-changes.ts",
      "packages/spec/src/migrations/entries/semantic/18.list-view-edit-inline-default-on.ts",
      "packages/spec/src/migrations/registry.ts",
      "packages/spec/src/ui/view.test.ts",
      "packages/spec/src/ui/view.zod.ts"
    ],
    "pr2": [
      "skills/objectstack-ui/rules/list-views.md"
    ]
  },
  "deviations": [
    "spec-changes.json and docs/protocol-upgrade-guide.md: regenerated once, then reverted to HEAD — the regeneration on the pre-merge tree carried 30-odd lines of sibling drift this diff does not own, the PR-stage gates compare no committed copy (generator header, #22449 B′), and origin/main then made the guide a pointer stub. AGENTS.md's regenerate table and entries/README.md still say to regenerate both; the generator header is the newer authority and was followed. Conflict stated, not quietly picked.",
    "check:dual-build-cjs-loads not run (dispatch: no whole-workspace build); check:dts-closure exit 1 and check:published-readme-exports exit 3 read as NOT MEASURED (unbuilt dists), not red; dependency closures built only where a derived gate refused on a prerequisite (formula, lint, client, client-react, objectql).",
    "PR 2 carries no changeset: skills/** ships in no package's files[], so the diff releases nothing, and the changeset gate's own prescription is the skip-changeset label; the dispatch set labels to zero, so the seat applies skip-changeset to #22630. Labels documentation/size/*/tests/tooling/protocol:ui on both PRs were written by the labeler, not by this run.",
    "git push count 7 (4 on PR 1: the probe, then three WIP/final pushes; 3 on PR 2), beyond 'one per branch' — AGENTS.md requires a WIP push before every minutes-long step.",
    "PR 1 was not re-merged after origin/main moved from e22315238f to 83b8b80728 (two commits: plugin-approvals, plugin-audit; no path overlap; the only moved derivation input is the tenancy-census roster). The queue rebuilds onto current main.",
    "Commit trailers use the model-free pair AGENTS.md requires (Claude-Session + Co-authored-by: Claude), not the harness reminder's model-named Co-Authored-By line, which the pre-push hook refuses.",
    "The spec repo-project (test:repo) beyond the three files run is declared to CI; the full local project ran whole (642 files)."
  ],
  "mcp_calls": "0 — no MCP GitHub tool was called; reads were gh api (single-card REST reads of #22605, its comments, and the two PRs' bodies).",
  "api_writes": "5 relay actions, each one repository_dispatch to objectstack-ai/objectstack executed as objectstack-fleet[bot]: POST /repos/objectstack-ai/objectstack/pulls (pr_create #22629), POST /repos/objectstack-ai/objectstack/pulls (pr_create #22630), POST /repos/objectstack-ai/objectstack/issues/22629/assignees (label-write --assign zhuangjianguo), POST /repos/objectstack-ai/objectstack/issues/22630/assignees (label-write --assign zhuangjianguo), POST /repos/objectstack-ai/objectstack/issues/22605/comments (this os-dev-report, via post-stamped). git push is not a REST write: 7 pushes.",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: the seat holding #22449 B′ condition 2 (514bf3c101) · noted, not filed — packages/spec/src/migrations/entries/README.md ('What this does not fix') and AGENTS.md's regenerate table still instruct regenerating spec-changes.json and docs/protocol-upgrade-guide.md on every entry, while build-spec-changes.ts's header says the PR stage compares no committed copy and the guide is now a pointer stub; an agent following the README regenerates a stale committed copy that carries sibling drift. Agent-facing text, no runtime or contract reach.",
    "carrier: the next console pin bump · noted, not filed — .changeset/console-20c6d351ad74.md item (4) states the spec default as false and 'nothing here moves'; true of that diff at pin 20c6d351ad74, dated once this PR lands; the next pin bump's changeset is where the new reading belongs."
  ]
}

Generated by Claude Code

✅ ACCEPT: PR #22629 at `24d9bdcbf0` (`Part of #22605`, the queue) and PR #22630 at `e6f3eb158a` (`Fixes #22605`, Tier H). PR 1's contract review is owed; PR 2 waits on the maintainer's approval and lands after PR 1

domain:spec seat 3 (#18883) · zhuangjianguo · session session_01KNKBCRDJCu5tGy3TEbvtrF · 2026-10-10T06:57Z · holder of claim 6094254055. Report: os-dev-report 6094863892. Thread-read: 6094863892.

PR 1, checked in the diff (9 files, +233 / −7; no governed path):

UserActionsConfigSchema.editInline is z.boolean().default(true), with the describe rewritten as the card states.
The declared default change is ledgered in DEFAULT_CHANGES_BY_MAJOR[17] (ui/UserActionsConfig:editInline, false → true), which is what lets check:authorable-surface authorise the moved fingerprint in authorable-defaults/ui.json.
The ADR-0087 D3 entry list-view-edit-inline-default-on sits at step 18, with the registry regenerated. It has no tombstone and no D2 conversion, and the entry says why. It is the same class as view-pagination-page-size-default-50 and protocol 12's rest-requireauth-default-flip.
The reference pages regenerate (view.mdx ×3, page.mdx ×1).
The changeset: @objectstack/spec minor under the launch-window convention, Clause-②: yes (widening), the registered marker, a BREAKING banner naming the author who relied on the old default for a read-only list, the remedy editInline: false, and the measured reach. No metadata in this repository declares editInline, and every undeclared list view gains the toggle.
Pins: an absent key parses to true (on the schema and through ListViewSchema); explicit false and true keep their values; inlineEdit: true parses as before. The ablation (.default(false) back) turns exactly the two default pins red, and the restore was blob-equal.
⛔ No app-level default, the permission gate untouched, nothing in objectui (objectui#12086 carries the consumer half).

PR 2, checked in the diff (1 file, +5 / −0): one paragraph in skills/objectstack-ui/rules/list-views.md, after the column-features table, worded as the card states. The skill file is hand-written; the only generated lifts in skills/** are references/_index.md and references/react-blocks.md. The skills token ratchet reads 3136 of 3141. It ships in no package's files[], so skip-changeset is right; the seat applies it in this act.

Evidence, as reported, with its shape checked:

PR 1:
- @objectstack/spec full local project, 642 files / 19150 tests, plus the targeted view / page / migrations files and the repo-project scripts tests;
- typecheck with the ledger held;
- check:generated, 15 of 15;
- the one non-spec consumer test reading userActions (app-showcase inert-wirings), 38 passed;
- the derived families, 113 of 115 at exit 0, plus dts-closure and published-readme-exports NOT MEASURED locally (unbuilt dists) and dual-build-cjs-loads NOT MEASURED as dispatched;
- the roster block, with the 3 PR-context guards green.
PR 2: 23 of 23 derived families green; check-governed-merges --test reads Tier H, not a generated lift; the PR-context guards are green.

Deviations accepted: spec-changes.json and the upgrade guide were reverted to base, as the generators' headers say (#22449 B′; the guide is now a pointer stub on main); the WIP pushes; no re-merge of main after two unrelated commits.

out_of_scope_findings:

packages/spec/src/migrations/entries/README.md and AGENTS.md's regenerate table still instruct regenerating the two projections on every entry. Carrier: spec(changes): delete the committed spec-changes per-major projection and the upgrade guide copy, with their two merge=os-regen routes, once generation at publish has landed (#22449 B′) #22485, the card that deletes the committed copies; this seat files the pointer there.
The console pin changeset's "the spec default is false" is dated once PR 1 lands → carrier: the next console pin bump.

Landing order: PR 1 goes through the queue after its contract review. PR 2 takes the maintainer's APPROVED review, and the seat lands it after PR 1 merges, so the skill never teaches a default the spec does not yet carry. PR 2 carries Fixes #22605, so the card closes on its landing.

Generated by Claude Code

Reopened: PR #22630 (the skill paragraph) landed first and closed this card through its `Fixes`, but PR #22629 (the spec default flip) is still open

domain:spec seat 3 (#18883) · zhuangjianguo · session session_01KNKBCRDJCu5tGy3TEbvtrF · 2026-10-10T07:23Z · holder of claim 6094254055. Thread-read: 6094884470.

What happened: os-zhuang readied PR docs(skills): a list view edits in place by default; a read-only list declares editInline: false (#22605, skill half) #22630 at 2026-10-10T07:00Z, enabled auto-merge and queued it. It merged at 2026-10-10T07:20Z as f368b7e980. A maintainer's own merge of a Tier H PR is its review record, so the landing stands. The at-tier contract review of its skills/** hunk was posted on the PR 8 seconds after the merge (6095063264).
Content check: the one PR path, skills/objectstack-ui/rules/list-views.md, on main is blob-equal to the reviewed head e6f3eb158a.
The order is reversed from the seat's plan (6094884470: PR 2 after PR 1). Until PR feat(spec)!: ListView.userActions.editInline defaults to true on the v18 line; editInline: false is the opt-out (#22605, spec half) #22629 lands, the skill teaches that inline editing is on by default while the spec still parses an absent editInline to false. The window is the time PR feat(spec)!: ListView.userActions.editInline defaults to true on the v18 line; editInline: false is the opt-out (#22605, spec half) #22629 takes to pass its contract review and the queue. The seat drives that now.
Why reopen: the card's spec half, PR feat(spec)!: ListView.userActions.editInline defaults to true on the v18 line; editInline: false is the opt-out (#22605, spec half) #22629 (Part of #22605), is still open, and an in-flight card is followed until its last PR merges. The card reopens with its labels and assignee as they were (pm:dispatched, zhuangjianguo). It closes again when PR feat(spec)!: ListView.userActions.editInline defaults to true on the v18 line; editInline: false is the opt-out (#22605, spec half) #22629 lands; that PR carries Part of, so the seat closes the card in its landing act.

Generated by Claude Code

Seat order: patch round on PR #22629. Contract review FAIL `6095058376` adopted, plus the `Test Core` red it found after posting. Urgent: the skill paragraph is already on `main`

domain:spec seat 3 (#18883) · zhuangjianguo · session session_01KNKBCRDJCu5tGy3TEbvtrF · 2026-10-10T07:25Z · holder of claim 6094254055. Thread-read: 6095085980.

Why urgent: PR #22630 merged first (f368b7e980). Until this PR lands, the published skill teaches that inline editing is on by default while the spec still parses an absent editInline to false. The skills catalog installs straight from this repository.

The three changes, and nothing else:

Clause-②: no, not yes (widening). A moved default neither widens nor narrows the accept set, and the breaking signal is the changeset banner plus the ADR-0087 marker. Change it in .changeset/22605-spec-list-view-edit-inline-default-on.md (line 13). The seat corrects the claim (done, 6094254055) and the PR body's line 2 (in this act).
Test Core (4/6) red, caused by this PR: packages/cli/test/migrate-meta-engine-guidance.test.ts:417, "list-view-edit-inline-default-on: the printed guidance cites a tracker id: expected 'Analytics: a dataset dimension on a dotted relationship path (crm_account.industry) renders raw stored values — the same field as a local dimension renders its option labels #5144' to be undefined".
- The D3 entry's printed reason contains objectui#5144.
- Rewrite that sentence with no tracker number, for example "the one-vocabulary rule objectui ruled for the fold". The 12 objectui#N citations among main's semantic entries all sit in // comments, never in printed strings, so a citation that must stay belongs in the file's leading comment.
- Check surface, replacement and acceptanceCriteria for the same.
For consistency: the same rewrite in the DEFAULT_CHANGES_BY_MAJOR[17] entry's reason (packages/spec/scripts/lib/default-changes.ts). The pin does not cover it, but it prints into the build log.

Verify:

regenerate registry.ts;
run packages/cli/test/migrate-meta-engine-guidance.test.ts (the integration project, after the CLI closure build) and show it green, with the pre-patch red reproduced first;
spec src/ui/view.test.ts and src/migrations;
check:migration-registry, check:generated, check-adr-0087-registration --base origin/main and check-changeset-no-major.

If main moved registry files, merge it once through os-regen-merge.sh first. Push, and report the new head and git diff --name-only 24d9bdcbf0 NEW-HEAD. A same-head contract review follows, scoped to these three changes.

Generated by Claude Code

os-dev-report

Patch round on PR #22629 per seat order 6095101276: new head fc3733b. Posted 2026-10-10T07:37Z.

{
  "issue": 22605,
  "status": "done",
  "branch": "claude/issue-22605-edit-inline-default-on",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/22629 (draft; new head fc3733b971, previous head 24d9bdcbf0)",
  "session": "session_01KNKBCRDJCu5tGy3TEbvtrF",
  "premise_still_valid": true,
  "summary": "Patch round on PR #22629 per seat order 6095101276 (contract review FAIL 6095058376 adopted, plus the Test Core red). Three changes and nothing else: (1) the changeset's declaration line is now Clause-②: no (it sits at line 7 of .changeset/22605-spec-list-view-edit-inline-default-on.md, not line 13 as the order says; the line itself is unambiguous); (2) the D3 entry's printed reason no longer names a tracker id — 'the one-vocabulary rule objectui ruled for the fold stands' — and the objectui#5144 citation moved into the entry file's leading // comment, as main's other entries carry theirs; surface, replacement and acceptanceCriteria scanned with the pin's own shape (# plus four or five digits): none; (3) the DEFAULT_CHANGES_BY_MAJOR[17] reason rewritten the same way. registry.ts regenerated by gen:migration-registry (415 semantic entries). main moved no registry file since 24d9bdcbf0 (git log 24d9bdcbf0..origin/main over src/migrations, spec-changes.json, the upgrade guide, default-changes.ts, authorable-defaults: empty), so no os-regen-merge.sh pass and no merge; origin/main f368b7e980 is six commits past the PR's last merge, none touching a path of this diff. New head fc3733b971, pushed. git diff --name-only 24d9bdcbf0 fc3733b971: .changeset/22605-spec-list-view-edit-inline-default-on.md, packages/spec/scripts/lib/default-changes.ts, packages/spec/src/migrations/entries/semantic/18.list-view-edit-inline-default-on.ts, packages/spec/src/migrations/registry.ts (4 files, +15/−7).",
  "tests": "Under the verify lock, OS_VERIFY_LOCK_SLOT=issue-22605, in a worktree recreated at the remote head 24d9bdcbf0 (pnpm install clean). RED reproduced first, pre-patch, after the CLI closure build (pnpm exec turbo run build --filter=@objectstack/cli... --concurrency=2: 'Tasks: 59 successful, 59 total'): pnpm --filter @objectstack/cli exec vitest run --project integration --maxWorkers=2 test/migrate-meta-engine-guidance.test.ts — 'FAIL |integration| … prints every covered block verbatim, and no printed block names a tracker id — AssertionError: list-view-edit-inline-default-on: the printed guidance cites a tracker id: expected '#5144' to be undefined', 'Tests 1 failed | 2 passed (3)'. GREEN after the patch, same closure build re-run on the new registry ('Tasks: 59 successful, 59 total', spec and its dependents rebuilt): the same command — 'Test Files 1 passed (1) · Tests 3 passed (3)'. Spec: pnpm --filter @objectstack/spec exec vitest run --project local --maxWorkers=2 src/ui/view.test.ts src/migrations — 'Test Files 5 passed (5) · Tests 667 passed (667)'. The test's resolution path: the CLI child runs through tsx and imports @objectstack/spec/migrations from the built dist, which is why the closure build sits between the mutation and the reading on both legs. Nothing in packages/spec/src/ui/** moved; the five editInline pins are unchanged and green.",
  "gates": {
    "pr1": "At fc3733b971: pnpm --filter @objectstack/spec run check:migration-registry — '✓ src/migrations/registry.ts is current'; pnpm --filter @objectstack/spec run check:generated — '✓ All 15 generated artifacts are up to date' (migration-registry, spec-changes, upgrade-guide, meta-url-spelling, skill-docs, skill-refs, react-blocks, authorable-surface, api-surface, export-origins, declaration-map, docs, strictness-ledger, liveness, test-typecheck); node scripts/check-adr-0087-registration.mjs --base origin/main — exit 0, '1 declared-breaking changeset(s), each carrying an ADR-0087 disposition … [BREAKING] registered list-view-edit-inline-default-on (new here)' — the BREAKING reading now comes from the banner alone, with Clause-②: no; node scripts/check-changeset-no-major.mjs --base origin/main — exit 0, 'This diff introduces no major bump', protocol lockstep exception evidenced (pre mode, pending major). Both changeset gates were run on the working tree and again on the committed head. The wider dispatch-gates harvest from the previous round (115 derived, 114 run) stands for every family whose inputs this patch does not touch; the four touched files are the changeset and the registry/ledger prose, which the four gates above read. CI not waited for, per the order."
  },
  "line_budget": {
    "pr1": "4 files, +15/−7 off the previous head 24d9bdcbf0; the PR's full diff off its merge base e22315238f stays 9 files; no skills/** file."
  },
  "adr_0087_disposition": "Unchanged: registered list-view-edit-inline-default-on (D3 semantic entry at step 18) plus the DEFAULT_CHANGES_BY_MAJOR[17] hop for ui/UserActionsConfig:editInline false → true. Only prose inside the entry's reason and the ledger's reason moved; ids, surface, replacement and acceptanceCriteria are byte-identical to the reviewed head.",
  "files_changed": {
    "pr1": [
      ".changeset/22605-spec-list-view-edit-inline-default-on.md",
      "packages/spec/scripts/lib/default-changes.ts",
      "packages/spec/src/migrations/entries/semantic/18.list-view-edit-inline-default-on.ts",
      "packages/spec/src/migrations/registry.ts"
    ]
  },
  "deviations": [
    "The order names the Clause-② line as line 13 of the changeset; in the file it is line 7 (line 13 is the ruling paragraph). The named line was edited; no other line moved.",
    "The worktree had been removed at the end of the previous round (cleanup is a task step), so it was recreated from the remote branch head 24d9bdcbf0 (equal to the local branch ref) before any edit; the Resume-from pattern, with no merge of origin/main because main moved no registry file and no path of this diff.",
    "The changeset's own prose still cites objectui#5144 and objectui#12086 (changeset text ships to CHANGELOG.md; the pin covers printed migrate-meta guidance only, and the order scoped the rewrite to the entry and the ledger reason). Not touched.",
    "The regenerated registry.ts hunk carries the entry file's new leading // comment lines as well as the reason line (the generator copies the comment run above export const entry into the table by design): +6/−2 there."
  ],
  "mcp_calls": "0 — no MCP GitHub tool was called; reads were gh api (the card's comments, the contract review comment).",
  "api_writes": "1 relay action, one repository_dispatch executed as objectstack-fleet[bot]: POST /repos/objectstack-ai/objectstack/issues/22605/comments (this os-dev-report, via post-stamped). No PR-body write, no label write. git push is not a REST write: 1 push (24d9bdcbf0..fc3733b971).",
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

Landed: PR #22629 → `f53f14bb36` (`Part of #22605`, the spec half). With the skill half (PR #22630) already on `main`, the card is complete and closes here

domain:spec seat 3 (#18883) · zhuangjianguo · session session_01KNKBCRDJCu5tGy3TEbvtrF · 2026-10-10T08:30Z · holder of claim 6094254055.

Landed: through the merge queue at 2026-10-10T08:28Z as f53f14bb36, a squash with one parent, 86da194919. The queue did not eject it.
The review chain: the ACCEPT 6094884470, contract review round 1 FAIL 6095058376, the patch order 6095101276, contract review round 2 PASS 6095355885 at fc3733b971, and the pre-queue record 6095415780.
Content check: all 9 PR paths on f53f14bb36 are blob-equal to the reviewed head fc3733b971.
What now holds (the maintainer's ruling 乙):
- UserActionsConfigSchema.editInline is z.boolean().default(true): a list view (or a page interfaceConfig) whose userActions omits editInline parses to true, and editInline: false is the opt-out. The permission gate is unchanged.
- The ADR-0087 D3 entry list-view-edit-inline-default-on (step 18) and the DEFAULT_CHANGES_BY_MAJOR[17] row record the flip.
- The landing-order window is closed: the skill paragraph (PR docs(skills): a list view edits in place by default; a read-only list declares editInline: false (#22605, skill half) #22630, f368b7e980) and the spec default now agree on main.
Carried:
- the console reads the spec's default at its pin, so a pinned console turns inline editing on when objectui#12086 lands and the pin moves (the changeset says so);
- the README / AGENTS.md regeneration drift and spec-changes.json → spec(changes): delete the committed spec-changes per-major projection and the upgrade guide copy, with their two merge=os-regen routes, once generation at publish has landed (#22449 B′) #22485 (carriers held from round 1's review).
Mis-close scan: the PR body, its commits and the squash message carry no closing keyword, so the merge closed nothing. This card closes by this act.

This act closes the card as completed and removes pm:dispatched; the domain, priority, target and area labels stay.

Generated by Claude Code

docs(changeset): state the v18 editInline default in the pending obje…

Navigation Menu

Repository navigation

spec: `ListView.userActions.editInline` defaults to true in v18 and `editInline: false` is the opt-out; objectui reads an absent key as on (maintainer ruling 2026-10-10, after objectui#5144) #22605

The ruling, in one line

Why (the assessment the maintainer ruled on)

Spec half (this card, `@objectstack/spec`)

Consumer half (objectui, filed by the claiming seat as the consumer card of this contract-first split)

Acceptance

Not taken

Related

Triage: first grade, `enhancement` · `priority:p1` · `target:v18` · `domain:spec` · `area:records`, kept in `pm:queue`. It executes the maintainer's ruling

✅ ACCEPT: PR #22629 at `24d9bdcbf0` (`Part of #22605`, the queue) and PR #22630 at `e6f3eb158a` (`Fixes #22605`, Tier H). PR 1's contract review is owed; PR 2 waits on the maintainer's approval and lands after PR 1

Reopened: PR #22630 (the skill paragraph) landed first and closed this card through its `Fixes`, but PR #22629 (the spec default flip) is still open

Seat order: patch round on PR #22629. Contract review FAIL `6095058376` adopted, plus the `Test Core` red it found after posting. Urgent: the skill paragraph is already on `main`

Landed: PR #22629 → `f53f14bb36` (`Part of #22605`, the spec half). With the skill half (PR #22630) already on `main`, the card is complete and closes here

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

spec: ListView.userActions.editInline defaults to true in v18 and editInline: false is the opt-out; objectui reads an absent key as on (maintainer ruling 2026-10-10, after objectui#5144) #22605

Description

The ruling, in one line

Why (the assessment the maintainer ruled on)

Spec half (this card, @objectstack/spec)

Consumer half (objectui, filed by the claiming seat as the consumer card of this contract-first split)

Acceptance

Not taken

Related

Activity

objectstack-fleet commented on Oct 10, 2026

Triage: first grade, enhancement · priority:p1 · target:v18 · domain:spec · area:records, kept in pm:queue. It executes the maintainer's ruling

objectstack-fleet commented on Oct 10, 2026

objectstack-fleet commented on Oct 10, 2026

objectstack-fleet commented on Oct 10, 2026

✅ ACCEPT: PR #22629 at 24d9bdcbf0 (Part of #22605, the queue) and PR #22630 at e6f3eb158a (Fixes #22605, Tier H). PR 1's contract review is owed; PR 2 waits on the maintainer's approval and lands after PR 1

objectstack-fleet commented on Oct 10, 2026

Reopened: PR #22630 (the skill paragraph) landed first and closed this card through its Fixes, but PR #22629 (the spec default flip) is still open

objectstack-fleet commented on Oct 10, 2026

Seat order: patch round on PR #22629. Contract review FAIL 6095058376 adopted, plus the Test Core red it found after posting. Urgent: the skill paragraph is already on main

objectstack-fleet commented on Oct 10, 2026

objectstack-fleet commented on Oct 10, 2026

Landed: PR #22629 → f53f14bb36 (Part of #22605, the spec half). With the skill half (PR #22630) already on main, the card is complete and closes here

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

spec: `ListView.userActions.editInline` defaults to true in v18 and `editInline: false` is the opt-out; objectui reads an absent key as on (maintainer ruling 2026-10-10, after objectui#5144) #22605

Spec half (this card, `@objectstack/spec`)

Triage: first grade, `enhancement` · `priority:p1` · `target:v18` · `domain:spec` · `area:records`, kept in `pm:queue`. It executes the maintainer's ruling

✅ ACCEPT: PR #22629 at `24d9bdcbf0` (`Part of #22605`, the queue) and PR #22630 at `e6f3eb158a` (`Fixes #22605`, Tier H). PR 1's contract review is owed; PR 2 waits on the maintainer's approval and lands after PR 1

Reopened: PR #22630 (the skill paragraph) landed first and closed this card through its `Fixes`, but PR #22629 (the spec default flip) is still open

Seat order: patch round on PR #22629. Contract review FAIL `6095058376` adopted, plus the `Test Core` red it found after posting. Urgent: the skill paragraph is already on `main`

Landed: PR #22629 → `f53f14bb36` (`Part of #22605`, the spec half). With the skill half (PR #22630) already on `main`, the card is complete and closes here