Skip to content

fix(plugin-view): the host delegation reads a named view's remaining protocol members off the named view (objectui#10758) - #10884

Merged
objectstack-fleet[bot] merged 3 commits into
mainfrom
claude/issue-10758-named-view-protocol-members
Sep 28, 2026
Merged

objectstack-fleet[bot] merged 3 commits into
mainfrom
claude/issue-10758-named-view-protocol-members

Conversation

@objectstack-fleet

Copy link
Copy Markdown
Contributor

Fixes #10758
Clause-②: no

Claim 5861950053, seat domain:ui#4, PM session session_01MCg3z5cGrV51xEddpGcxEY (https://claude.ai/code/session_01MCg3z5cGrV51xEddpGcxEY). This is bucket ① of the objectui#7924 ruling (5690906005). densityMode and allowExport stay excluded, per the re-ruling 5852199633 as amended by 5856690997.

Re-sample (step 1), on objectui origin/main de1b879

Each reading was taken off disk and off the installed @objectstack/spec 17.4.0. A scratch script walked the spec schema, parsed the NamedListView interface with the TypeScript parser, and parsed both host literals. The types census re-derives the same partition on every run.

  • The protocol's named view. ObjectListViewSchema declares 50 keys. Five of them are retirement tombstones, so 45 are live. NamedListView declares 64 members, and every one of the 45 live protocol keys is among them.
  • What the renderer read off a named view: 20 names. They are appearance, calendar, chart, columns, data, fieldOrder, filter, gallery, gantt, grouping, kanban, label, map, name, rowColor, sort, timeline, tree, type and userActions.
    • The card says 21. PR objectui#10821 removed the options read after the card was filed, and options is not a protocol key.
    • The delegation itself (the renderListView literal) read 10 of the 20. Eight are read inside the fence: label, columns, grouping, userActions, rowColor, data, fieldOrder and appearance. filter and sort are read upstream, in authoredFilters and mergedSort.
  • Bucket ①: 23 protocol members that no route read off a named view. These are the ruling's count and the ruling's names: addRecord, allowPrinting, aria, bulkActionDefs, bulkActions, compactToolbar, conditionalFormatting, description, emptyState, exportOptions, filterableFields, hiddenFields, inlineEdit, navigation, pagination, resizable, rowActions, rowHeight, searchableFields, selection, sharing, showRecordCount and userFilters.
    • tabs and pageName are also unread. They stay out because objectui#8980 ruled them declared-inert.
  • The delegation's rungs for these 23, before this PR.
    • 19 read the host views entry, then the node (activeView?.X ?? (schema as any).X).
    • rowHeight read the host entry through the density fold.
    • description, exportOptions and bulkActionDefs had no rung at all.
  • App-shell's object page relay drops none of the 23.
    • 22 have a rung that reads its active view (viewDef.X).
    • compactToolbar arrives through the plugin-view composition. ObjectView.relayRungCensus-7559.test.ts records it as relayed upstream.
    • So packages/app-shell/src/views/ObjectView.tsx is not touched (N4). App-shell hands its named view to plugin-view as the host views entry, not as listViews.
  • The consumer (N3). ListView reads all 23 off the list-view node it is handed.
    • Most are read directly.
    • Seven are forwarded to the object-grid that ListView renders, and ObjectGrid reads each one: pagination, selection, resizable, rowActions, bulkActions (as batchActions), bulkActionDefs and conditionalFormatting.
    • No member reaches the node without a reader, so N3 has no declared-not-enforced gap to list.

What changed

  • packages/plugin-view/src/ObjectView.tsx, in the renderListView({ schema: { … } }) literal inside the objectui#5097 HOST-COMPOSITION fence. Each of the 23 rungs now reads the active named view first (currentNamedViewConfig?.X). Then it reads the host views entry, and then the node, but only where the rung already read the node. This is the precedence grouping, rowColor and appearance use.
    • The three new rungs (description, exportOptions, bulkActionDefs) read the two views only. They never read the node.
    • A named view's rowHeight is read as itself: the strict record refuses densityMode, so there is nothing to fold. The host entry still goes through the fold (ruling A′).
    • ⛔ No rung gained a (schema as any) read, so the objectui#5097 exemption stays at 27 names. ⛔ No alias and no non-protocol key is read off the named view.
  • What moves for a host. Only a host that supplies both listViews and its own renderListView sees a change.
    • Where the active named view and the host views entry both carry one of these members, the named view's value now wins.
    • A named view that carries none of them hands down exactly what it did before.
    • The registered object-view renderer passes no renderListView, so an authored node is unaffected.
  • Changeset: 10758-named-view-protocol-members.md, patch for @object-ui/plugin-view.

Beyond the claimed file surface, and why

  1. packages/types/src/__tests__/object-view-unmirrored-keys-7779.test.ts, the types census. It pins every bucket-① member UNREAD by name, so the change turns it red by construction. Every pin that held one of these members unread is inverted. ⛔ None is deleted.
    • The census moves from 20 read / 44 unread to 43 read / 21 unread.
    • Bucket ① is asserted EMPTY, derived rather than listed. The 23 names are kept as the pinned reading of what moved, one READ assertion each.
    • The census's negative twin moves from rowHeight to allowExport, which is read off the host entry and the node and not off a named view.
    • The absence pin for bulkActionDefs / exportOptions becomes a read pin that also holds "no node read".
    • The densityMode pin and the two READ_TEXT rows follow the new rung text.
  2. packages/plugin-view/README.md and content/docs/plugins/plugin-view.mdx each gain one prose paragraph describing the host-path precedence. This follows the docs-driven rule (AGENTS.md section 5, Add automated testing infrastructure and CI/CD workflows #2). Neither file gains a code fence.

Pins, one named view per member family

  • ObjectView.namedViewProtocolKeys-8980.test.tsx, a new route-3 block. There are four families: list chrome (7), record actions (6), grid presentation (6), and search, filter and navigation (4). Each family runs four cases:
    • the named view parses under the strict ObjectListViewSchema, so every value is one a conforming author can write;
    • THE FIX: the named view's value reaches the list-view node the host renders, and wins over the host entry and the node;
    • CONTROL: with the family absent from the named view, the host entry supplies it;
    • CONTROL: with neither view declaring it, the node supplies it. The four rungs with no node read stay undefined.
    • The members already read (label, columns, fieldOrder, grouping) ride every fixture as the firing control. A last case holds rowHeight: the named value is read as itself, and the host entry's retired spelling still arrives through the fold.
  • objectViewHostSurface.test.tsx, a new section 1b, derived with the TypeScript parser over the fence only:
    • the named-view read set inside the fence, by name;
    • the property that every rung for a protocol member reads the named view, except type (node identity) and filter / sort (read upstream);
    • no named-view read of a key the protocol does not declare there, with allowExport as the control;
    • no new cast read.
  • "The rendered view" here is the list-view node that the host's renderListView receives. A plugin-view test cannot import ListView: @object-ui/plugin-list is not a plugin-view dependency, and check:phantom-deps refuses the import. The link from that node to the rendered list is ListView's own reads, listed under N3 above.

Verification at 5dc80e3

Every heavy run went through os-verify-lock.sh, and each exit code was captured redirect-first.

  • vitest packages/plugin-view/ plus the types census: 53 files, 745 tests passed, VERDICT command-exit 0.
  • vitest on the tests that read the three changed markdown files (from scripts/markdown-test-inputs.mjs --list), every app-shell ObjectView* test (the relay census included) and StudioDesignSurface.gridRefresh: 58 files, 1243 tests passed, VERDICT command-exit 0.
  • type-check, exit 0 for @object-ui/plugin-view (tsc --noEmit && tsc -p tsconfig.test.json) and for @object-ui/types (plus tsconfig.examples.json). --listFilesOnly confirms the three edited test files are in those programs.
  • ESLint --no-inline-config on the 4 edited TS/TSX files, which is the whole population: eslint.config.js covers them with files: ['**/*.{ts,tsx}'], and --format json reports 4 files.
    • 0 errors and 86 warnings, per file identical to the base (60 / 26 / 0 / 0).
    • Not type-aware: projectService / parserOptions count 0 against a languageOptions control of 1. So no untouched file's verdict can move.
  • Gates, exit 0 each:
    • check-changeset-presence (4 source files of 2 released packages, 1 changeset), check-changeset-no-major, check-changeset-fixed, check-changeset-overwrite, and check-changeset-claims (report-only; the six pending paragraphs it names were read and are still true);
    • check:pending-changeset-literals, check:new-line-citations (0 new), check:control-bytes, check:test-path-roots, and the three check:vi-mock-* gates;
    • check:spec-symbols and check:phantom-deps;
    • check:doc-types, check:doc-fences, check:doc-example-ids, check-doc-expression-carriage, check-doc-links and check:doc-example-readers;
    • check-governed-queue-guard --test on the 7 paths: NOT GOVERNED.
  • NOT MEASURED:
    • check:readme-exports: prerequisite not met (the workspace dists are not built). This diff adds no fenced block to any README.
    • check:doc-snippets / check:doc-examples: not run, because no fenced block was added.
    • The full @object-ui/types suite: only a test file changed there, and that file and the types tests that read the changed markdown ran.
    • The full app-shell suite and the Spec Main Shape Gate: CI runs them.

Reverse verification

Run once on the committed tree 59046a2 with ablation-replace.mjs. The whole HOST-COMPOSITION fence was replaced by its base text; the mutated blob e1e9557 equals the base blob. The directions were predicted before the run.

  • Result: 64 failed, 225 passed.
    • 8980 pin: 7 red. These are the four THE FIX cases, the host-entry CONTROL of the two families holding a new rung, and the rowHeight case.
    • Host-surface pin: 3 red.
    • Census: 54 red. These are the 23 READ rows, the 23 BUCKET ① → READ rows, the counts, the partition, bucket ① empty, the inverted absence pin, the densityMode pin and the two READ_TEXT rows.
  • The predicted-green controls stayed green: the host CONTROL of the other two families, all four node CONTROLs, the four protocol parses, the off-protocol pin, the 21 UNREAD rows, and the bucket ② / ③ rows.
  • One prediction missed. The four-way partition test stayed green. It partitions the pinned literal, and the link from literal to derivation is the "43 read + 21 unread" test, which went red.
  • The restore was proven: the blob equals HEAD (89b6c7e) and git diff HEAD is empty.
  • The later commits (2960ef6, 5dc80e3) touch only a test title and the types of the probe helper. No assertion moved.

Acceptance notes

  • Reach, measured on this tree. No in-tree host passes both listViews and renderListView. App-shell's object page passes its named view as the host views entry, and the Studio grid passes neither. So the new rungs serve an out-of-tree host that composes both, which is the same reach as the objectui#8980 rungs beside them. This is recorded, not claimed wider.
  • Route 2 (the registered renderer's grid path) still reads only columns, filter, sort, grouping and rowColor off a named view.
    • A scratch probe, not committed: ObjectView without renderListView, which is exactly what the registered renderer mounts. A named grid view authored pagination, selection, hiddenFields, rowHeight, conditionalFormatting, rowActions, bulkActions and resizable, and the object-grid schema carried columns only. ObjectGrid reads each of those keys.
    • This is outside the claimed literal: the ruling sends bucket ① to "the object-view delegation".
    • ⚠️ The types census counts a member as read when ANY route reads it, so bucket ① now reads empty while route 2 still ignores these members.
    • No real producer was found: no authored object-view node with listViews in either repository.
    • Reported to the PM, not filed.
  • The delegation does not relay a named view's eight view-kind blocks (kanban, calendar, gallery, timeline, gantt, map, chart, tree). Only options: activeView goes down, while ListView reads top-level schema.kanban and the others.
    • The same probe: a named type: 'kanban' view with kanban: { groupByField: 'stage' } handed the host viewType: 'kanban' and columns, and no kanban.
    • These members are counted read (route 1), so they are not bucket ①.
    • Same reach as above. Reported, not filed.

Generated by Claude Code

…protocol members off the named view (objectui#10758)

Bucket ① of the objectui#7924 ruling (comment 5690906005). Re-sampled on
origin/main de1b879: the protocol declares 45 live members on a named view,
NamedListView declares all 45, the renderer read 20 of them off a named
view, and 23 (the ruling's figure and names) were read off none.

The renderListView literal inside the objectui#5097 HOST-COMPOSITION fence
now reads each of the 23 off the active named view first, then the host
views entry, then the node only where the rung already did. Three rungs
are new (description, exportOptions, bulkActionDefs) and view-sourced
only; no rung gained a cast read, so the exemption stays at 27 names.

Pins: one named view per member family on the delegation (named wins over
host and node; host and node rungs unchanged as controls; the members
already read ride every fixture as the firing control); the fence's
named-view read set by name plus the property that every protocol rung
reads the named view; the types census inverted (bucket ① empty, 43 read
/ 21 unread), with no pin deleted.

Claude-Session: https://claude.ai/code/session_01MCg3z5cGrV51xEddpGcxEY
Co-authored-by: Claude <noreply@anthropic.com>
…le (objectui#10758)

Claude-Session: https://claude.ai/code/session_01MCg3z5cGrV51xEddpGcxEY
Co-authored-by: Claude <noreply@anthropic.com>
…objectui#10758)

The new route-3 helper read the handed-down node as `any` and cast the host
views entry, adding five no-explicit-any warnings to the file. Both are
typed now: the node as a string-keyed record, the host entry as the views
prop's own element.

Claude-Session: https://claude.ai/code/session_01MCg3z5cGrV51xEddpGcxEY
Co-authored-by: Claude <noreply@anthropic.com>
@github-actions github-actions Bot added documentation Improvements or additions to documentation package: types plugin tests labels Sep 28, 2026
@github-actions

Copy link
Copy Markdown
Contributor

changeset-claim-re-read

⚠️ 6 pending changeset(s) describe a file this change touches

Their bodies publish verbatim into the CHANGELOG at the next release, so this is a request to re-read them against your diff — addressed here because you are the one seat that can answer it without re-deriving anything.

⛔ Nothing here blocks, and nothing here is a verdict on your change. This gate exits 0, is not a required context, and judges name resolution, never meaning: it asked whether a pending body names a file you touched. "Is this sentence still true?" is the one question it will not answer, and the one you are being asked to answer.

.changeset/6726-find-envelope-records-arms.md

  • names plugin-view/src/ObjectView.tsx → packages/plugin-view/src/ObjectView.tsx — edited by this change

    | module | what it does | | --- | --- | | components/src/hooks/related-count-store.ts | related-list tab badge count | | components/src/renderers/basic/data-list.tsx | element:repeater rows | | components/src/renderers/basic/elements.tsx | element:number client-side aggregate | | components/src/renderers/basic/record-picker.tsx | element:record_picker options | | plugin-detail/src/renderers/record-activity.tsx | record:activity self-fetch | | plugin-detail/src/renderers/record-history.tsx | record:history self-fetch | | plugin-view/src/ObjectView.tsx | non-grid (kanban / calendar / gallery / timeline) fetch |

.changeset/7070-no-invented-gantt-date-fields.md

  • names plugin-view/src/ObjectView.tsx → packages/plugin-view/src/ObjectView.tsx — edited by this change

    • app-shell/src/views/ObjectView.tsx — the console object page. The inline branch becomes ganttViewOptions, the sibling of calendarViewOptions and timelineViewOptions: the declared block spread whole, title floored at 'name', no date field invented. - plugin-list/src/ListView.tsx — the render branch AND the capability gate. - plugin-view/src/ObjectView.tsx — generateViewSchema, the authored object-view element route, which bypasses ListView entirely.

.changeset/7499-gantt-non-axis-floors-omitted.md

  • names plugin-view/src/ObjectView.tsx → packages/plugin-view/src/ObjectView.tsx — edited by this change

    • plugin-list/src/ListView.tsx — the object-gantt render branch. - plugin-view/src/ObjectView.tsx — generateViewSchema, the authored object-view element route, which bypasses ListView entirely.

.changeset/8221-retire-legacy-string-sort.md

  • names content/docs/plugins/plugin-view.mdx → content/docs/plugins/plugin-view.mdx — edited by this change

    Docs teach the array only: content/docs/plugins/plugin-map.mdx, content/docs/plugins/plugin-view.mdx and packages/plugin-view/README.md.

  • names packages/plugin-view/README.md → packages/plugin-view/README.md — edited by this change

    Docs teach the array only: content/docs/plugins/plugin-map.mdx, content/docs/plugins/plugin-view.mdx and packages/plugin-view/README.md.

.changeset/8653-listview-title-retired-rowactiondefs-pinned.md

  • names packages/plugin-view/src/ObjectView.tsx → packages/plugin-view/src/ObjectView.tsx — edited by this change

    title — retired. ListView resolved its export filename through schema.label || (schema as any).title. @objectstack/spec/ui's ListViewSchema refuses title by name (unrecognized_keys: ['title']) while ObjectGridPropsSchema accepts it; packages/types mirrors the platform contract rather than ruling over it, so declaring title on ListViewSchema would have made this repo accept what the platform save gate rejects. That asymmetry is also why objectui#6639 could take the declare branch for ObjectGridSchema.title one package over and this site could not. A parse-based census of apps/ examples/ content/ and packages/ found zero list-view nodes authoring title, so the retirement costs no author a filename. Over that same corpus the instrument reports three object-grid nodes carrying the key: two authored ones, both in content/docs/api/schema-reference.md, plus one that is not authored at all — packages/plugin-view/src/ObjectView.tsx composes title: schema.table?.title onto a grid node it builds, so it is a producer writing the key rather than an author declaring it. ObjectGrid's own title reads are untouched — they remain declared, ruled and read.

.changeset/object-view-unmirrored-keys-7779.md

  • names packages/plugin-view/src/ObjectView.tsx → packages/plugin-view/src/ObjectView.tsx — edited by this change

    What was measured. Every reading was taken on the object-view node renderer (packages/plugin-view/src/ObjectView.tsx, registered by plugin-view/src/index.tsx) with schema.objectName / schema.layout as the positive controls of the same schema.KEY query, so each zero is a reading; the repo-wide census of viewTabBar finds the key in no source file outside @object-ui/types (two doc tables listed it as authorable and are corrected here). The spec side was read through the installed pin (@objectstack/spec@17.2.0, ui entry, 117 exported object schemas walked; control keys objectName / columns / navigation / listViews hit): the three spec-modelled keys are optional slots on ListViewSchema and ObjectListViewSchema; the six local keys have no spec slot anywhere.

Read the paragraph, not the line: both false halves of the objectui#8617 claim sat in one paragraph, and correcting either alone would have left it asserting the same wrong thing.

If a claim did go false, correct the body. That is precedented and prose-only, frontmatter untouched; check-changeset-overwrite.mjs will report the correction as its own case 2 ("correcting a declaration on purpose … legitimate"), which is the intended shape — one gate asks for the read, the other records the write.

Not covered, stated so nobody reads this as more: a born-false claim that spells no line address at all (objectui#9495 coordinated one by ORDINAL — "a grep finds that member first" — and deciding that means reading what the sentence means), a claim spelled as a symbol or a package rather than a backticked file name, and a file named ambiguously.

Compared the checked-out tree with c6678b1bd (merge-base with origin/main): 6 file(s) changed outside .changeset/, read against 1649 pending declaration(s) that publish a body (2240 pending in total). · run

@github-actions

Copy link
Copy Markdown
Contributor

✅ Console Performance Budget

Metric Value Budget
Eager closure (gzip, 329 chunks) 3087.5 KB 3104.5 KB
Main entry chunk (gzip) 148.4 KB 350 KB
Entry file index-DeV4epry.js —
Status PASS —

The eager closure is every chunk the entry reaches through static imports — what the browser fetches and parses before the app renders. The entry chunk on its own is a small fraction of it.


📦 Bundle Size Report

Package Size Gzipped
app-shell (consoleActionDispatch.js) 0.20KB 0.19KB
app-shell (index.js) 16.57KB 6.17KB
app-shell (runtime-config.js) 20.68KB 7.36KB
app-shell (types.js) 0.01KB 0.04KB
app-shell (urlParams.js) 10.06KB 3.86KB
auth (ActiveOrganizationStorage.js) 27.95KB 10.04KB
auth (AuthContext.js) 0.31KB 0.24KB
auth (AuthGuard.js) 2.07KB 1.00KB
auth (AuthProvider.js) 40.17KB 10.58KB
auth (AuthShell.js) 3.49KB 1.40KB
auth (ForgotPasswordForm.js) 12.21KB 3.45KB
auth (LoginForm.js) 18.15KB 5.39KB
auth (PreviewBanner.js) 0.90KB 0.50KB
auth (RegisterForm.js) 6.65KB 2.22KB
auth (SocialSignInButtons.js) 9.61KB 3.89KB
auth (UserMenu.js) 3.39KB 1.21KB
auth (auth-gate-events.js) 1.29KB 0.66KB
auth (authStyles.js) 5.04KB 1.72KB
auth (createAuthClient.js) 40.21KB 10.80KB
auth (createAuthenticatedFetch.js) 8.52KB 3.45KB
auth (index.js) 3.63KB 1.64KB
auth (invitation-status.js) 1.22KB 0.70KB
auth (org-roles.js) 6.66KB 2.78KB
auth (phone-identifier.js) 1.11KB 0.66KB
auth (types.js) 0.59KB 0.35KB
auth (useAuth.js) 5.30KB 1.02KB
auth (useWorkspaceAdminStatus.js) 11.08KB 4.58KB
collaboration (CommentThread.js) 27.13KB 7.95KB
collaboration (LiveCursors.js) 3.17KB 1.27KB
collaboration (PresenceAvatars.js) 6.49KB 2.64KB
collaboration (PresenceProvider.js) 2.79KB 1.13KB
collaboration (index.js) 1.68KB 0.73KB
collaboration (useCollaborationTranslation.js) 6.05KB 2.52KB
collaboration (useCommentSearch.js) 1.98KB 0.88KB
collaboration (useConflictResolution.js) 7.75KB 1.86KB
collaboration (useMentionNotifications.js) 1.81KB 0.68KB
collaboration (usePresence.js) 6.33KB 1.84KB
collaboration (useRealtimeSubscription.js) 7.91KB 2.01KB
components (index.js) 557.59KB 133.61KB
core (index.js) 9.93KB 3.94KB
create-plugin (index.js) 27.94KB 9.51KB
data-objectstack (index.js) 225.01KB 62.50KB
fields (index.js) 261.26KB 66.37KB
i18n (LocalizationContext.js) 1.76KB 0.96KB
i18n (builtinAggregateLabels.js) 0.86KB 0.49KB
i18n (currency.js) 2.59KB 1.22KB
i18n (fallbackInterpolation.js) 6.25KB 2.77KB
i18n (i18n.js) 8.87KB 3.64KB
i18n (index.js) 5.24KB 2.27KB
i18n (pickLocalized.js) 9.86KB 3.95KB
i18n (provider.js) 39.40KB 12.91KB
i18n (translateFn.js) 0.20KB 0.18KB
i18n (useDisplayLocale.js) 3.52KB 1.76KB
i18n (useObjectLabel.js) 34.34KB 9.17KB
i18n (useSafeTranslation.js) 5.60KB 2.33KB
layout (index.js) 39.28KB 11.09KB
mobile (MobileProvider.js) 0.92KB 0.49KB
mobile (ResponsiveContainer.js) 0.94KB 0.38KB
mobile (breakpoints.js) 1.51KB 0.70KB
mobile (createOfflineDataSource.js) 5.61KB 1.75KB
mobile (index.js) 1.99KB 0.87KB
mobile (offlineQueue.js) 3.91KB 1.35KB
mobile (pwa.js) 0.97KB 0.49KB
mobile (serviceWorker.js) 1.48KB 0.62KB
mobile (serviceWorkerSource.js) 3.41KB 1.48KB
mobile (useBreakpoint.js) 1.54KB 0.65KB
mobile (useGesture.js) 6.96KB 1.98KB
mobile (useOfflineSync.js) 1.99KB 0.72KB
mobile (usePullToRefresh.js) 6.62KB 2.45KB
mobile (useResponsive.js) 0.72KB 0.42KB
mobile (useSpecGesture.js) 5.52KB 2.10KB
mobile (useTouchTarget.js) 1.01KB 0.54KB
permissions (MePermissionsProvider.js) 13.52KB 4.88KB
permissions (PermissionContext.js) 0.31KB 0.25KB
permissions (PermissionGuard.js) 0.89KB 0.45KB
permissions (PermissionProvider.js) 6.24KB 2.16KB
permissions (discardProofCache.js) 1.04KB 0.55KB
permissions (evaluator.js) 8.33KB 3.07KB
permissions (index.js) 0.93KB 0.41KB
permissions (store.js) 0.91KB 0.42KB
permissions (useFieldPermissions.js) 1.28KB 0.53KB
permissions (usePermissions.js) 4.83KB 2.27KB
plugin-ai (index.js) 16.01KB 3.93KB
plugin-calendar (index.js) 51.73KB 14.72KB
plugin-charts (index.js) 83.99KB 22.86KB
plugin-chatbot (index.js) 197.67KB 46.90KB
plugin-dashboard (index.js) 136.82KB 36.44KB
plugin-designer (index.js) 216.42KB 44.47KB
plugin-detail (index.js) 233.48KB 61.79KB
plugin-editor (index.js) 2.23KB 1.05KB
plugin-form (index.js) 161.21KB 41.41KB
plugin-gantt (index.js) 169.83KB 41.99KB
plugin-grid (index.js) 218.66KB 59.90KB
plugin-kanban (index.js) 48.43KB 15.11KB
plugin-list (index.js) 114.78KB 28.44KB
plugin-map (index.js) 22.90KB 7.62KB
plugin-markdown (index.js) 13.88KB 4.80KB
plugin-report (index.js) 44.17KB 12.20KB
plugin-timeline (index.js) 31.00KB 9.09KB
plugin-tree (index.js) 11.21KB 3.89KB
plugin-view (index.js) 88.42KB 22.18KB
providers (DataSourceProvider.js) 0.75KB 0.39KB
providers (MetadataProvider.js) 1.37KB 0.59KB
providers (ThemeProvider.js) 1.90KB 0.85KB
providers (UploadProvider.js) 11.81KB 3.58KB
providers (index.js) 0.45KB 0.23KB
providers (types.js) 0.01KB 0.04KB
react-runtime (index.js) 5.62KB 2.34KB
react (LazyPluginLoader.js) 4.47KB 1.63KB
react (SchemaRenderer.js) 119.16KB 39.05KB
react (data-invalidation.js) 5.05KB 2.08KB
react (index.js) 4.03KB 1.86KB
react (schema-input.js) 4.25KB 2.04KB
react (spec-input.js) 0.20KB 0.18KB
sdui-parser (body-dialect.js) 4.78KB 2.09KB
sdui-parser (codegen.js) 7.50KB 3.05KB
sdui-parser (dashboard-widget-options.js) 3.08KB 1.30KB
sdui-parser (index.js) 6.16KB 2.71KB
sdui-parser (input-type.js) 2.84KB 1.40KB
sdui-parser (kanban-quick-add.js) 3.89KB 1.87KB
sdui-parser (parse.js) 25.28KB 7.80KB
sdui-parser (provenance.js) 3.84KB 1.90KB
sdui-parser (types.js) 0.28KB 0.23KB
sdui-parser (validate.js) 18.27KB 6.22KB
types (ai.js) 4.11KB 2.06KB
types (api-types.js) 0.20KB 0.18KB
types (app.js) 3.52KB 1.36KB
types (base.js) 0.20KB 0.18KB
types (blocks.js) 0.20KB 0.18KB
types (complex.js) 2.93KB 1.49KB
types (crud.js) 0.20KB 0.18KB
types (dashboard-filter-alias.js) 6.23KB 2.74KB
types (data-display.js) 3.75KB 1.85KB
types (data-protocol.js) 0.20KB 0.19KB
types (data.js) 0.20KB 0.18KB
types (designer.js) 1.85KB 0.85KB
types (disclosure.js) 0.20KB 0.18KB
types (error-code.js) 1.54KB 0.88KB
types (expression.js) 0.20KB 0.18KB
types (feedback.js) 0.20KB 0.18KB
types (field-types.js) 0.20KB 0.18KB
types (form.js) 0.20KB 0.18KB
types (http-inflight.js) 8.87KB 3.73KB
types (http-retry.js) 4.32KB 2.02KB
types (icon-key-migration.js) 4.26KB 1.63KB
types (index.js) 4.74KB 2.26KB
types (layout.js) 0.20KB 0.18KB
types (managed-by.js) 0.19KB 0.18KB
types (mobile.js) 5.00KB 2.39KB
types (navigation.js) 0.20KB 0.18KB
types (objectql.js) 0.20KB 0.18KB
types (overlay.js) 0.20KB 0.18KB
types (permissions.js) 2.52KB 1.31KB
types (plugin-scope.js) 0.20KB 0.18KB
types (record-components.js) 0.20KB 0.19KB
types (record-semantics.js) 1.28KB 0.67KB
types (registry.js) 0.20KB 0.18KB
types (reports.js) 0.20KB 0.18KB
types (select-option.js) 0.20KB 0.19KB
types (spec-report.js) 5.05KB 1.93KB
types (spec-ui-namespace.js) 0.20KB 0.19KB
types (strict-authoring-face.js) 17.15KB 6.32KB
types (system-fields.js) 3.33KB 1.54KB
types (theme.js) 6.27KB 2.87KB
types (ui-action.js) 8.11KB 3.32KB
types (views.js) 0.20KB 0.18KB
types (widget.js) 0.20KB 0.18KB

Size Limits

  • ✅ Core packages should be < 50KB gzipped
  • ✅ Component packages should be < 100KB gzipped
  • ⚠️ Plugin packages should be < 150KB gzipped

@objectstack-fleet

Copy link
Copy Markdown
Contributor Author

Contract review

Served-tier: 209/209 CONTRACT_REVIEW_TIER
Head-sha: 5dc80e39981133d955377a72544f9ce9be727916

Read: card objectui#10758 body and its four comments (seat notes 5856946147, 5861063883, claim 5861950053, and the os-dev-report 5862298983, treated as claims to test); the objectui#7924 ruling 5690906005, re-ruling 5852199633 and amendment 5856690997; merged PR objectui#10793 (66e8b2ab3) and objectui#10821 (b93e245) metadata and diffs; PR objectui#10884 body, file list and diff at head 5dc80e3 (fetched as refs/review/pr-10884; merge-base with origin/main is de1b879a6bdc7feb7b9f8a16fcb8abdca1004873, the dev's stated re-sample base; origin/main is now 1c5ee334); packages/plugin-view/src/ObjectView.tsx at base and head (whole file, the HOST-COMPOSITION fence at lines 2399–2584 of the head); the three edited test files and their diffs; the changeset, README and mdx diffs; the published @objectstack/spec 17.4.0 tarball (the lockfile's resolved version; objectui has no node_modules, so I pulled the registry tarball plus zod 4.4.3 into scratch and evaluated ObjectListViewSchema directly); packages/types/src/objectql.ts at base (NamedListView, ObjectViewSchema.listViews); packages/plugin-list/src/ListView.tsx, packages/plugin-grid/src/ObjectGrid.tsx, packages/app-shell/src/views/ObjectView.tsx and ObjectView.relayRungCensus-7559.test.ts at base; packages/core/src/utils/normalize-list-view.ts (the density fold); packages/plugin-view/src/index.tsx (the registered renderer); examples/byo-backend-console/src/App.tsx, packages/app-shell/src/views/metadata-admin/previews/ViewPreview.tsx, packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx; objectstack examples/**, apps/**, packages/** grepped for listViews; objectui AGENTS.md (§5 commandments, §9 changeset rule at line 168, 测试纪律, 版本号策略), .changeset/config.json, every pending .changeset/*.md grepped for claims the change could falsify; the head's check-runs at three polls. Not re-run: the dev's ablation and the three test files locally (a run would need a full pnpm install and a build of plugin-view's dependency closure; CI's eight test shards and Type Check on this exact sha are green, see ③).

① Derived judgments

1. The re-sample — RIGHT. The published spec 17.4.0 ObjectListViewSchema has 50 keys; 5 are never tombstones (bordered, performance, responsive, striped, virtualScroll), so 45 are live (my own evaluation of the dist, not the dev's). NamedListView at the merge-base declares every one of the 45 (0 missing). At the merge-base the whole of ObjectView.tsx reads 19 distinct currentNamedViewConfig?.KEY names plus view.name in the tab strip: appearance calendar chart columns data fieldOrder filter gallery gantt grouping kanban label map name rowColor sort timeline tree type userActions = 20, not 21; options is not among them (the base fence carries options: activeView, the host entry only, with the objectui#7928 comment, and PR objectui#10821's diff is what removed the named-view read). 45 live minus those 20 leaves 25; minus tabs and pageName (ruled declared-inert at objectui#8980, the census's BUCKET_DECLARED_INERT) leaves exactly the ruling's 23: addRecord allowPrinting aria bulkActionDefs bulkActions compactToolbar conditionalFormatting description emptyState exportOptions filterableFields hiddenFields inlineEdit navigation pagination resizable rowActions rowHeight searchableFields selection sharing showRecordCount userFilters. Every one is a live protocol key under its protocol name; no alias. densityMode and allowExport are outside the list, and the strict schema refuses both on a named view (my parse: REFUSED, REFUSED).

2. The rungs — RIGHT. The source diff touches only the renderListView({ schema: { … } }) literal inside the objectui#5097 fence. Nineteen existing rungs gained currentNamedViewConfig?.X ?? in front of an unchanged activeView?.X ?? (schema as any).X tail; rowHeight became currentNamedViewConfig?.rowHeight ?? (normalizeListViewSchema(activeView ?? {}) as { rowHeight?: string }).rowHeight, the named view read as itself before the host fold; the three rungs with no prior rung (description, exportOptions, bulkActionDefs) read currentNamedViewConfig?.X ?? activeView?.X and never the node. Counted: the set of (schema as any).K reads inside the fence is byte-identical at base and head — 31 real reads (plus the letter K in the header comment), which is OBJECT_VIEW_HOST_COMPOSITION_KEYS (27 names, unchanged) plus OBJECT_VIEW_DECLARED_FORWARDED_KEYS (4: data filterableFields navigation searchableFields, unchanged). The fence's named-view read set grew from 8 (appearance columns data fieldOrder grouping label rowColor userActions) to 31, the difference being exactly the 23. The eight pre-existing named-view rungs and the upstream filter / sort reads are untouched, so no already-read member changed precedence. Fall-through is safe: currentNamedViewConfig is the raw schema.listViews[active] entry typed NonNullable‹ObjectViewSchema['listViews']›[string] (the spec's z.input), nothing in core, react, plugin-view or app-shell parses or normalizes listViews entries before this read, and no bucket-① member carries a zod .default() (a minimal named view parses to columns label type only), so an unauthored member is undefined and ?? yields exactly the old chain. App-shell's objectViewSchema memo carries no listViews at all, so on the console's object page currentNamedViewConfig is null and every new rung falls through; the Studio surface passes renderListView without listViews. The only page whose render moves is an out-of-tree host that passes both listViews and renderListView and whose named view authors one of the 23 — which is the fix.

3. N3 and N4 — RIGHT. ListView at the merge-base has a schema.X read for 22 of the 23 (counts from 1 to 8 per member) and forwards bulkActionDefs into the object-grid node through (schema as any).bulkActionDefs; it maps bulkActions to the grid's batchActions. ObjectGrid reads pagination (6 sites), selection (3), resizable (1), rowActions (2), batchActions (1), bulkActionDefs (1), conditionalFormatting (4). So each of the 23 has a reader in the rendered list. App-shell's relay has a viewDef rung for 22 (rowHeight through normalizeListViewSchema(viewDef ?? {}), hiddenFields gated on activeViewDeclaresColumns, bulkActionDefs through a cast) and its census row for compactToolbar is relayed-upstream with upstreamReads: ['activeView']. The relay drops none; app-shell is untouched by the PR, correctly.

4. Pins — RIGHT. ObjectView.namedViewProtocolKeys-8980.test.tsx adds a route-3 block: four families (list chrome 7, record actions 6, grid presentation 6, search/filter/navigation 4), each with a strict-spec parse, THE FIX (named wins over host and node), a host CONTROL (family absent from the named view, host entry supplies) and a node CONTROL (neither view; node supplies, and the four no-node-rung members must arrive undefined); a coverage test asserts the families hold 23 distinct protocol names with allowExport as the refused control; a rowHeight case checks named 'tall' wins and a host densityMode: 'compact' still arrives as 'compact' through the fold (the fold maps compact to compact). The already-read label columns fieldOrder grouping ride every fixture through expectControl. I parsed all four family fixtures plus the control against the published schema myself: all OK, and the rowHeight enum is compact short medium tall extra_tall, covering every fixture value. objectViewHostSurface.test.tsx section 1b walks the fence with the TypeScript parser: the named-view read set pinned by name (31 entries, equal to my count), the property that every protocol-member rung reads the named view except filter / sort / type (with the upstream reads asserted by text), the converse (no off-protocol named-view read, allowExport as control), and the cast count equal to 27 + 4. typescript is a declared devDependency of plugin-view. The census object-view-unmirrored-keys-7779.test.ts is inverted honestly: 12 expect lines removed, each with an inverted counterpart (20 to 43, 44 to 21, 23 to 0, not.toContain(member) to toContain(member), the absence pin to a read-plus-no-node-read pin, the negative twin rowHeight to allowExport with its rung text asserted); bucket ① empty is derived from the live protocol keys minus reads with BUCKET_DECLARED_INERT as the expected remainder and a control on the same derivation; the 23 names are kept as BUCKET_ONE_READ_SINCE_10758 with a per-member READ assertion. The 21 remaining unread names contain none of the 43 the head file reads (checked against my whole-file grep). Each new pin fails on base by construction (the base fence has 8 named reads, not 31; the base census partition is 20/44). No timeout raised, no skip/only/todo, no heavyDomTests entry (0 hits in the added lines). The dev's ablation report is structurally consistent with this (the one missed prediction, the partition test staying green because it partitions the pinned literal, is disclosed); I did not re-run it.

5. Prose and shape — RIGHT. Every sentence of .changeset/10758-named-view-protocol-members.md holds at the head: the 23 never came off the named view on that node (base fence); named first, then host, then the node where the rung already read it (diff); the four families list all 23; the three new rungs read the two views only (no (schema as any).description|exportOptions|bulkActionDefs at head) and the exemption keeps 27 names (counted); rowHeight read as itself with the host entry through the fold (diff); named wins where both carry, a named view carrying none hands down exactly what it did (no defaults, no upstream normalization); the registered renderer passes no renderListView (index.tsx mounts ‹ObjectView schema dataSource› only). The README and mdx paragraphs say the same and name the pin file. patch for @object-ui/plugin-view is the right level: a renderer starts honouring members its contract already declares, no export, type or accept set moves. The types change is test-only; AGENTS.md's rule (line 168, enforced by check-changeset-presence) is "declare once" for any src/ change in a released package, satisfied by the one changeset, and the fixed group of 40 publishes together anyway; Changeset Declaration, Changeset Bump Policy and Changeset Fixed Group Check are green. No major. The PR body's first line is Fixes #10758 and its second Clause-②: no, which holds: NamedListView, ObjectViewSchema, OBJECT_VIEW_HOST_COMPOSITION_KEYS (27) and OBJECT_VIEW_DECLARED_FORWARDED_KEYS (4) are unchanged, and the strict named-view schema is the spec's own.

6. The dev's open question — facts verified, with one correction. Route 2 is real: the registered renderer mounts ObjectView without renderListView, and off a named view that path reads only columns, filter, sort, grouping, rowColor and the eight kind blocks (kanban calendar gallery timeline gantt map chart tree), while the new rungs sit inside if (renderListView). The delegation does not relay the eight kind blocks (no such rung in the fence; only options: activeView goes down). Producers: the dev's "no authored object-view node carries listViews in either repo" is WRONG for objectui. (i) examples/byo-backend-console/src/App.tsx (ObjectPage) authors listViews: { default: { label: 'All', type: 'grid', columns } } on an object-view node rendered through ‹ObjectView› without renderListView — route 2 — though that named view authors no bucket-① member. (ii) packages/app-shell/src/views/metadata-admin/previews/ViewPreview.tsx spreads the whole stored ViewItem body (...foldStoredListOptions(body)) into listViews[id] and renders through SchemaRenderer, i.e. the registered renderer, route 2; a stored view carrying rowHeight, hiddenFields, inlineEdit or pagination (keys the console persists per view, by app-shell's own comments) is therefore an in-tree producer whose bucket-① grid members route 2 ignores. In objectstack, listViews appears only on config.views[] containers (app-crm, app-showcase, app-todo view files, the CLI); those reach objectui as objectDef.listViews, then app-shell's mergedViews, then the host views prop, whose relay covers all 23 — no object-view node with listViews exists there. No in-tree host passes both listViews and renderListView (app-shell: views plus renderListView; Studio: renderListView only; byo and ViewPreview: listViews only), so the dev's reach statement for the new rungs holds. The seat decides; option A's stated premise (zero producers on route 2) does not survive ViewPreview.

② Semver level

patch on @object-ui/plugin-view (fixed group of 40, so the group bumps together). No major, no minor needed: nothing declared, exported or accepted moves; a host composing listViews with its own renderListView gets the named view's authored value where it previously got the host entry's, which is the contract the spec already declared.

③ Boundary flags

  • CI on 5dc80e3: poll 1 at 2026-09-28T02:42:24Z — 42 runs: 29 success, 10 in progress (Spec Main Shape Gate, Test shards 1–8, Type Check), 3 skipped, 0 failure. Poll 2 at 02:53:53Z — 43 runs: 40 success, 3 skipped, 0 in progress, 0 failure (the Test rollup appeared, success 02:49:30Z; Type Check success 02:47:15Z; Spec Main Shape Gate success 02:44:30Z; shards 1–8 all success by 02:49:17Z). Poll 3 at 02:55:08Z — unchanged. The 3 skipped are the coverage-shard matrix placeholder, Test (coverage) and dependabot. Green includes Lint, Build & E2E, Build Docs, Doc Snippet Type Check, README Export Check, Line Citation Gate, all five changeset gates, Governed Surface Queue Guard, Pre-Install Import Graph Check.
  • Not re-run here: the three edited test files (no node_modules in the checkout; would need a full install plus dependency build) and the dev's ablation. CI's shards and Type Check on this exact sha stand in.
  • The PR is behind origin/main (1c5ee334), but git diff de1b879 origin/main over the seven touched files is empty and git merge-tree is clean.
  • Pending changesets that publish verbatim carry census figures this PR inverts: 7924-named-list-view-liveness-census.md ("21 read and 43 unread of 64 today"), 7924-named-list-view-disposition-buckets.md ("43 unread … 23 + 8 + 10 + 2"), object-view-unmirrored-keys-7779.md ("renderer reads 21 off a named view", "43 unread members", and "listViews stays unmirrored", the last already false since objectui#10821). All three were stale before this PR (20/44 at the merge-base) and check-changeset-overwrite forbids amending a pre-existing changeset, so this is not a defect of the change; but the dev's report that the 7779 paragraphs were "read, still true" is inaccurate for those figures. The seat may want a carrier for the CHANGELOG-time correction.
  • Pin boundary, disclosed by the dev: "carried through to the rendered view" is pinned at the list-view node renderListView receives, not at a mounted ListView (@object-ui/plugin-list is not a plugin-view dependency and check:phantom-deps would refuse it). The link is N3's reads, verified above by reading.
  • The mdx paragraph cites src/__tests__/ObjectView.namedViewProtocolKeys-8980.test.tsx without a package prefix; in content/docs that path is relative to nothing. Cosmetic; the link gate is green.
  • The types census counts a member read when any route reads it, so bucket ① reads empty while route 2 (with ViewPreview as a live in-tree producer) still ignores the grid members — disclosed by the dev under "open questions", corrected here on the producer fact (item 6).
  • Item 6's kind-block gap (the delegation relays no kanban etc.) has the same reach limit: no in-tree host passes both listViews and renderListView.
  • No model identifier in the diff (0 hits); commit trailers carry Claude-Session and Co-authored-by: Claude. No worktree was created by this review; nothing to remove.

Implemented-by: claude/issue-10758-named-view-protocol-members
Reviewed-by: session_01MCg3z5cGrV51xEddpGcxEY

VERDICT: PASS

@objectstack-fleet

Copy link
Copy Markdown
Contributor Author

Merge-queue ejection: not this PR's. domain:ui seat 4 · session_01MCg3z5cGrV51xEddpGcxEY · 2026-09-28T04:04Z.

  • Failing check: Test (shard 7/8) on the queue commit cc7098f9b (origin/main b2683a2 plus this PR). The failing case is packages/plugin-detail/src/__tests__/RelatedList.unmaterializedSort.test.tsx › "keeps that button in client mode": Unable to find an accessible element with the role "button" and name /Total/.
  • Why it is not this PR's: this PR changes only plugin-view's ObjectView.tsx, its tests and docs, and packages/plugin-detail/src has no path to them. The file passed 12 of 12 runs across the merge with current main, the exact queue tree, plain main and this head.
  • Root cause, reproduced deterministically: the case does not wait for the object schema, and the sort button exists only after it lands (since objectui#10728). Delaying getObjectSchema by 50 ms gives exactly the CI signature; adding await settleObjectSchema(dataSource) makes it pass.
  • Fix: none exists yet on main. It is filed with the ready patch as objectui#10895. It is not ported here, because it is test code this PR does not touch.
  • Re-run: the one re-run allowed for a failure established as not this PR's; auto-merge is re-enabled in this act, so the PR re-enters the queue.

os-tesla pushed a commit that referenced this pull request Sep 28, 2026
…9 dated note

Patch round 2. The note stated the census split as 20 read and 44 unread,
which objectui#10758 (PR #10884) moved on main after this branch's
measurement. The note now says the renderer reads some of the declared
members off a named view and leaves the rest unread, names the two changes
that moved the split away from the entry's own 21 and 43, and points at the
census in object-view-unmirrored-keys-7779.test.ts for the current figures
instead of restating them. No other note, sentence or frontmatter moves.

Co-Authored-By: Claude <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014fWVhLzhxR8qrFsJ5o8TYW
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
… accepts — a separator without label, branding kept on edit, no Layout control (objectui#10867) (objectstack-ai#10889)

Fixes objectstack-ai#10867
Clause-②: yes (narrowing). Published `@object-ui/types` members narrow:
the separator arm of `NavigationItem` loses `label`, and
`AppWizardDraft` loses `layout`. That makes this `minor` plus a BREAKING
(authoring) banner, and never `major`.

Dispatched by the `domain:ui` seat 2 PM (claim comment 5861866471,
triage 5861615446), session
`https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN`
(`mode:subagent`, so that id is also this run's). Branch
`claude/issue-10867-app-wizard-saves-spec-app`, cut from `origin/main`
`b93e245f9` (BASE); head `2c19f3047` after fix round 1 (round-0 head
`e2493c796`).

## What changed

**Member 1 — a separator carries only `type`, `id` and `order` (the
p1).**
- `@object-ui/types`: `NavigationItem` is now a union of two arms,
discriminated by `type`. `NavigationEntryItem` is the former interface
with `type` narrowed to every spec nav type except `'separator'`, and it
keeps its required `label`. `NavigationSeparatorItem` admits `type`,
`id` and `order`, the spec separator's keys, and every other entry key
is `?: never` on it. Those `never` keys are derived from
`NavigationEntryItem`, so a key added there is refused on the separator
without an edit. Both arms are exported from the package root.
`menuItemToNavigationItem` maps a legacy separator to `{ id, type:
'separator' }` and drops its label.
- `spec-derived-unions.test.ts`: the blocker-3 pin
(`_specSeparatorStillHasNoLabel`) is lifted. It is replaced by three
agreement pins. `_separatorAdmitsTheSpecKeys` and
`_separatorAdmitsTheSpecInputKeys` say the separator arm admits exactly
the spec separator's keys at the output and input tiers.
`_localSeparatorIsSpecValid` says the value types of the keys the arm
admits are ones the spec's separator accepts at both tiers. `extends`
ignores extra keys, so the key agreement is carried by the first two:
the third alone passes a labelled arm, and it is vacuous on BASE.
"Admits" means a key whose type is not `never`. The two remaining
blockers (`visible: boolean`, and `pinned` / `defaultOpen`) are
unchanged, and so is the umbrella verdict.
- `@object-ui/types` zod mirror (fix round 1): `NavigationItemSchema`'s
refinement has an explicit separator branch, and it used to return
early, so `objectui validate` passed a separator `label` that the save
door refuses with 422. That branch now refuses every key the spec's
separator does not declare, each at its own path, with a message that
names the key and the fix ("a separator carries only `type`, `id` and
`order`; drop `label`"). The allowed set is read off the spec, not
restated: the spec does not export `SeparatorNavItemSchema`, so it is
taken from the separator arm of the spec `AppSchema`'s own `navigation`
union. That schema already crosses the mirror's import boundary, so the
objectui#8317 population is unchanged. The read throws if the arm is
ever missing, instead of silently allowing everything.
- `@object-ui/plugin-designer`: `AppCreationWizard`'s "Add separator"
and `NavigationDesigner`'s two add paths (now one `newNavItem` helper)
write `{ id, type }`. `NavigationDesigner` no longer writes a `label`
onto a new separator, and its label, icon and visibility patchers skip
one.
- The union makes code that writes an entry-only key onto an unnarrowed
item fail to compile, and each such site now narrows first. None of
these edits changes what renders for valid data.
- `@object-ui/layout`: `resolveNavItemLabel` answers `''` for every
separator. A stored separator carrying a non-empty `label`, which
`menuItemToNavigationItem` produced before this change, used to resolve
to that label. No renderer asks it for a separator's label. The mobile
bottom nav's leaf list, which already skipped separators, is typed as
entries. `collectPinnedItems` tests the type before `pinned`.
- `@object-ui/app-shell`: `useNavPins` registers a favorite only for an
entry and leaves a separator as it is. The Studio sidebar's navigation
walk passes a separator through unchanged.
- Test fixtures that spread an item and add `visible` or
`requiredPermissions` are typed as `NavigationEntryItem`. These are in
layout's `AppSchemaRenderer` and `resolveHref` tests and app-shell's
`UnifiedSidebar.derivedAreaVisibility` test.

**Member 2 — an edit keeps the stored branding.** `EditAppPage` now
merges `branding` one level down, by the objectui#10842 rule. It keeps
every stored key the spec's `AppBrandingSchema` declares, read from
`AppBrandingSchema.shape`, and the wizard's values win for the keys it
maintains. A stored `accentColor` survives the save. A stored key the
spec does not declare is still left out.

**Member 3 — the Layout control is gone.** `AppWizardDraft.layout` is
removed. The Basic Info step's Layout fieldset is removed, with its
three icons and the draft default. `EditAppPage` no longer pre-fills
`layout`. The step description now reads "Name, title, and icon". The
four `appDesigner` layout keys are removed from
`DESIGNER_DEFAULT_TRANSLATIONS` and from all ten locale packs, and
`appDesigner.stepBasicDesc` is retranslated in all ten.

**Text that the change made false:**
- the `app.ts` module header and the `NavigationArea` docblock;
- the `wizardDraftToAppSchema` docblock;
- the `@object-ui/types:NavigationItem` and
`@object-ui/types:NavigationItemSchema` ledger reasons in
`scripts/check-spec-symbol-derivation.mjs` (the second now counts four
divergences, not five);
- the mirror's own `NavigationAreaSchema` docblock, which listed a
separator `label` among the keys only the spec refuses;
- the "separator label" row in `navigation-spec-parity.test.ts`: the
faces now agree, so it left the divergence list and is asserted as an
agreement (both refuse it);
- the ROADMAP line for the wizard's Basic Info step.

**Changesets.**
- New `.changeset/10867-app-wizard-saves-spec-app.md`:
`@object-ui/types` and `@object-ui/i18n` minor (removing published pack
keys is a narrowing), and `@object-ui/plugin-designer`,
`@object-ui/layout` and `@object-ui/app-shell` patch, with the
`Clause-②: yes` line. The BREAKING (authoring) banner names every
consequence, and each was checked against the code at TypeScript 6.0.3
by a throwaway probe:
  - a read of `label` into a `string` slot fails (TS2322);
  - a spread of an entry-only key fails (TS2322);
  - `interface … extends NavigationItem` fails (TS2312);
  - an augmentation of `NavigationItem` fails (TS2300);
  - `objectui validate` refuses a separator `label`;
- `t()` on a removed key renders the raw key unless the call passes a
`defaultValue`. The provider-less fallback is the defaults table, then
`defaultValue`, then the key, and i18next is configured with
`returnNull: false` and no missing-key parser.
- Dated notes in the objectui#10533 form on two pending entries.
- `10842-wizard-app-declared-keys.md`, titled "the wizard has no layout
choice": its "the wizard's layout choice is not saved" has no choice
left. Frontmatter md5 `5b8e68791c8bc51ef72179570fd15a1c` before and
after.
- `3162-types-ledger-batch8-verdicts.md`: its "three semantic blockers"
count. Frontmatter md5 `2ee28f1e1b449ccc2dd97fdcc6c2bc1f` before and
after.

**Fixture triage:**
- The `layout` key was dropped from the `AppWizardDraft` fixtures in
`app-creation-types`, `app-declared-keys-10842` and
`app-logo-one-spelling-10827`. These are spelling fixes: the key was
there only because the type required it.
- The vacuous `'layout' in schema` assertion in `app-creation-types`
"should handle draft with empty navigation" was removed.
- `navigation-model`'s "should convert a separator" pinned the removed
`label: ''` branch, so it was replaced by an exact-shape assertion.
- The `label: ''` was dropped from separator fixtures in
`navigation-model`, layout's `AppSchemaRenderer.test` and app-shell's
`phase1-page-blocks.render.test`.

## H1 to H3 (measured on BASE `b93e245f9`, installed `@objectstack/spec`
17.4.0)

**H1 — reproduced.** A probe run through `wizardDraftToAppSchema`, then
the spec's `AppSchema.safeParse`:
- a draft with one object entry plus the separator the wizard wrote, `{
id, type: 'separator', label: '' }`, gave
`[{"code":"unrecognized_keys","path":"navigation.1","keys":["label"]}]`;
- control, the same draft without the separator: parses green;
- control, the same separator without `label`: parses green;
- `menuItemToNavigationItem({ type: 'separator', label: 'Section' })`
gave `{"id":"migrated_0","type":"separator","label":"Section"}`.

The new DOM pin gives the same reading through the real wizard and
`CreateAppPage` on the base source: "the saved document parses green
through the spec `AppSchema`" fails with `unrecognized_keys` `["label"]`
at `navigation.1`.

**H2 — reproduced.** The new pin "a stored `accentColor` survives the
edit round trip" drives the real wizard through `EditAppPage`. On the
base source it fails: the saved `branding` is `{ favicon, logo,
primaryColor }`, and `"accentColor": "#f59e0b"` is missing.

**H3 — census** (`git grep` over packages, apps, docs, examples, skills
and content). Each hit goes or is corrected:
- **`AppWizardDraft.layout` and the draft default:**
  - the member in `app.ts` — removed;
  - `DEFAULT_DRAFT.layout` in `AppCreationWizard` — removed;
- `EditAppPage`'s `layout: appToEdit.layout || 'sidebar'` pre-fill —
removed;
  - four draft fixtures — the key dropped (see the triage above).
- **The step's UI:** the Layout fieldset (`app-layout-*` test ids,
`draft.layout` reads and writes) and its three icon imports — removed.
- **The step's strings:**
- `appDesigner.layout`, `layoutSidebar`, `layoutHeader` and
`layoutEmpty` in `DESIGNER_DEFAULT_TRANSLATIONS` and in ten locale packs
— removed;
- `appDesigner.stepBasicDesc` ("Name, title, and layout") in eleven
places — corrected.
- **Docs:** the ROADMAP Basic Info line — corrected.
`content/docs/guide/console.md` and `console-architecture.md` describe
the wizard's four steps and never the Layout control, so they are
unchanged.
- **Not this member:** the `layout: 'sidebar'` hits in
`content/docs/core/app-schema.mdx`,
`content/docs/guide/schema-overview.md`, the runner tests and the
`AppComponentSchema` fixtures are `AppComponentSchema.layout`, the
renderer node's own layout strategy. They are untouched, and the
changeset says so.

## Pins and ablation

New pins:
-
`packages/types/src/__tests__/app-wizard-separator-layout-10867.test.ts`:
vitest halves, and `@ts-expect-error` / `satisfies` halves run under
`tsc -p tsconfig.test.json`.
-
`packages/plugin-designer/src/__tests__/AppWizard.specDocument-10867.test.tsx`:
drives the real `AppCreationWizard` through `CreateAppPage` and
`EditAppPage` and reads what reaches `saveItem`.
- The `spec-derived-unions.test.ts` agreement pins.
- Fix round 1:
`packages/types/src/__tests__/nav-separator-mirror-refusal-10867.test.ts`.
`safeValidateSchema`, the strict authoring face
(`StrictAnyComponentSchema`) and the mirror itself each refuse a
separator `label`, each at its own path; the mirror also refuses `icon`
and `pinned` on a separator. Controls: a bare `{ id, type: 'separator'
}` (and one with `order`) is accepted by both faces and the spec, the
spec refuses the same `label`, and an entry keeps its `label`.
- Fix round 1: a new `EditAppPage` row, "a logo the author clears saves
as cleared, not as the stored value". It is a guard, not a red-on-BASE
pin: green on BASE and on HEAD, it fails if a later tidy-up drops empty
strings from the merge.

**Red on BASE.** Run from a committed tree: the source files were
checked out at BASE, the new pins kept, and everything was restored from
`HEAD` under a trap. The restore was proven by blob hash equal to HEAD
and an empty `git diff HEAD`.
- vitest `Tests 8 failed | 6 passed (14)`. The 8 failures are every
non-control row:
  - `menuItemToNavigationItem`;
  - the two create-with-separator rows;
  - the three branding rows;
  - the two Layout rows.
- The 6 passes are the controls and compile-only rows.
- `tsc -p tsconfig.test.json` in `packages/types` gives the new pins':
  - TS2741 ×2 (a bare separator is missing `label`);
- TS2578 ×2 (unused `@ts-expect-error` for a labelled separator and for
a draft with `layout`);
- TS1360 ×3 (the no-`layout` `satisfies` and the two separator-agreement
pins).

**Red before the mirror fix (round 1).** From the committed round-1
head, only `app.zod.ts` was checked out at the round-0 head `e2493c796`,
under a trap; the restore was proven blob == HEAD and `git diff HEAD`
empty. The new mirror pins plus `navigation-spec-parity` gave vitest
`Tests 4 failed | 17 passed (21)`. The 4 failures are the refusal rows
and the new agreement row; the controls passed.

**Green on HEAD.** Round 0, at `e2493c796`: `Test Files 284 passed
(284)`, `Tests 5676 passed (5676)`. Round 1, at `2c19f3047`: see Gates.
The `packages/types` `tsc -p tsconfig.test.json` has 0 errors at both
heads.

**Ablation: the separator branch is reverted, and only the separator
rows redden.** Three anchored mutations went through
`ablation-replace.mjs`, each proven landed and restored by blob hash:
- the `NavigationItem` alias gives its separator arm a required `label`
again;
- `menuItemToNavigationItem` writes `label` again;
- the wizard's add-separator writes `label: ''` again.

Result:
- vitest `Tests 3 failed | 11 passed (14)`. The failures are
`menuItemToNavigationItem`, "the saved document parses green" and "the
saved separator carries only `type` and `id`". The branding and Layout
rows stay green.
- `tsc -p tsconfig.test.json`: 5 errors, all on separator lines. TS2322
×2 (bare separators), TS2578 ×1 (labelled separator) and TS1360 ×2 (the
two agreement pins). The Layout type pin stays green.
- Restore: `blob == HEAD` and `git diff HEAD` empty for both files.

**Cross-package reverse check.** The downstream type-checks resolve
`@object-ui/types` through its rebuilt `dist`. Their first runs went red
only on sites the new union refuses: layout 3, plugin-designer 4,
app-shell 3 in source plus 4 in tests. They were green after narrowing.
That red could only come from the rebuilt declaration.

## Gates

Fix round 1, at `2c19f3047`, all under the verify lock:
- vitest `packages/types/`, `packages/plugin-designer/` and
`packages/i18n/`: `Test Files 356 passed (356)`, `Tests 6861 passed
(6861)`.
- vitest over the 18 app-shell navigation, pins and sidebar files,
`examples/schema-catalog/test/`, the two cli `validate` suites and the
three spec-symbol script suites: `Test Files 57 passed (57)`, `Tests
2487 passed (2487)`.
- **Arm-list consumer sweep** for the mirror change. Every test that
imports `NavigationItemSchema` or the app zod mirror, or runs
`safeValidateSchema` / `StrictAnyComponentSchema` over an app or
navigation document, was found by `git grep` and run with per-file
verdicts: 34 files, 1251 tests, 0 failed. They include
`zod-mirror-parity` 37/37, `navigation-spec-parity` 15/15,
`imported-defaults-8317` 124/124, `nav-target-exclusivity-8563` 15/15,
`navigation-model` 28/28, `zod-lazy-getter-identity-7918` 31/31,
`page-nav-misc-spec-parity` 44/44, `handler-keys-json-refusal-6124`
336/336, the two app-shell studio-design nav suites 22/22 and 3/3, cli
`validate-root-path-line` 6/6 and the catalog
`tree-view-nodes-mirror-6939` 22/22.
- `tsc -p packages/types/tsconfig.test.json`: 0 errors.
`@object-ui/types` `type-check`: 0 `error TS`.
`@object-ui/plugin-designer` `type-check`, after rebuilding its closure:
clean.
- Changeset gates:
  - `check-changeset-presence` ✅;
  - `changeset:check` ✅ (no `major`);
- `check:changeset-claims` (report-only) lists 15 pending entries naming
a touched file; the six new ones (4895, 6335, 7113, 7735, 7918 and 8871)
were read, and none claims anything a separator refusal falsifies;
  - `check-changeset-overwrite` (report-only): the two dated-note edits;
  - `check:pending-changeset-literals` ✅.
- Other gates: `check:control-bytes` OK, `check:new-line-citations` 0
new, `check:spec-symbols` ✅, the three i18n gates ✅,
`type-check:coverage` ✅.
- Round 1 changed no exported type and no layout, app-shell or console
source, so the round-0 downstream type-checks below still describe this
head.

Round 0, at `e2493c796` unless noted:
- `pnpm exec vitest run packages/types/ packages/plugin-designer/`:
`Test Files 284 passed (284)`, `Tests 5676 passed (5676)`.
- `pnpm --filter` `@object-ui/types`, `@object-ui/layout`,
`@object-ui/plugin-designer` and `@object-ui/app-shell` `run type-check`
(each `tsc --noEmit && tsc -p tsconfig.test.json`; types also runs
`tsconfig.examples.json`): verify-lock `VERDICT command-exit 0`.
- `@object-ui/console` `type-check` after rebuilding its dependency
closure: exit 0, 0 `error TS`.
- vitest `packages/layout/` and `packages/plugin-designer/` (after the
narrowing): `Tests 436 passed (436)`.
- vitest `packages/i18n/`, app-shell `defaults-maps-mirror-en-pack`,
`phase1-page-blocks.render` and layout `AppSchemaRenderer`, plus the
`check-spec-symbol-derivation`,
`spec-symbol-ledger-anchor-liveness-9537` and
`spec-symbol-ledger-types-7265` script suites: `Tests 1354 passed
(1354)`.
- 17 app-shell files covering pins, favorites, the sidebar and nav
fixtures: `Tests 160 passed (160)`.
- Changeset gates:
  - `check-changeset-presence` ✅ (5 released packages, 1 changeset);
  - `changeset:check` ✅ (no `major`);
- `check:changeset-claims` (report-only): 9 pending entries name a
touched file, and each was read — none is falsified;
- `check-changeset-overwrite` (report-only): lists the two dated-note
edits;
  - `check:pending-changeset-literals` ✅.
- i18n gates: `check:i18n-keys` ✅, `check:i18n-drift` ✅ ("Every changed
en value was followed by all nine translation packs"),
`check:i18n-designer-parity` ✅.
- Other gates:
  - `check:spec-symbols` ✅;
  - `check:control-bytes` OK;
- `check:new-line-citations`: `VERDICT new-cross-file-line-citations: 0
new citation(s)`;
  - `type-check:coverage` ✅ (43/43);
  - `check:test-path-roots` OK;
- `check:vi-mock-specifiers`, `check:vi-mock-inherit` and
`check:vi-mock-override-shape` OK;
  - `check:metadata-write-doors` OK;
  - `check:unreferenced-sources` OK;
  - `check:handler-key-reads` OK.
- ESLint on the 34 changed `.ts` / `.tsx` / `.mjs` files (repo config,
inline config honoured): 0 errors and 47 warnings. Each warned file
carries the same warning count at BASE, measured by linting the BASE
blob through `--stdin`, so no warning is new. The repo-wide `pnpm lint`
is CI's.
- NOT MEASURED:
- the full app-shell suite, because it runs past the foreground cap;
declared to CI;
- `check:readme-exports` for the unbuilt `@object-ui/cli` and
`@object-ui/plugin-ai` READMEs, which are untouched here. Its judged
population read 548 self-imports and 0 fabricated.
- Scheduling: two light runs (the red-on-BASE pair) went outside the
verify lock. The lock was held for more than 20 minutes by a sibling's
full app-shell suite, and this is declared as a narrowing. Every build,
suite and type-check above ran under the lock.

## Serial

- Open PRs read at 2026-09-28T03:3xZ: objectstack-ai#10888, objectstack-ai#10886, objectstack-ai#10884, objectstack-ai#10852,
objectstack-ai#10780, objectstack-ai#10777, objectstack-ai#10714 and objectstack-ai#10278, plus the release PR. None of their
file lists intersects this diff.
- `origin/main` has moved to `b2683a2c0` since BASE. `git merge-tree
--write-tree` against it is clean. Of the files main moved, only `en.ts`
overlaps, on other keys. No file main moved mentions `NavigationItem`,
`AppWizardDraft` or the removed keys.

## Acceptance notes

- **The zod twin, corrected in fix round 1.** Round 0 read the claim's
"if the mirror carries the separator branch" as not met, because the
mirror is one flat object. The contract review measured that the
refinement does have an explicit separator branch, so the condition
held, and the mirror now refuses what the TypeScript face refuses (see
What changed).
- **Boundary note, not taken: pre-17 tombstones.** Dropping the spec's
tombstone keys from `EditAppPage`'s kept set would make the save succeed
silently and lose the author's value. The spec's current loud refusal
carries a prescription, so this is left as it is.
- **Boundary note, not taken: `BrandingConfig.fontFamily`.**
`BrandingEditor` reads and writes it, and removing it is not a cheap
change inside these files.
- **`AppComponentSchema.layout` is untouched.** It is a different
member, the renderer node's layout strategy.
- **An edit through the wizard drops stored navigation** (found while
pinning member 2; not fixed here; the seat is filing it as its own card,
and the review confirmed this PR does not make it worse). Leaving the
Objects step regenerates `navigation` from the selected objects
unconditionally. In edit mode that replaces the stored tree, so
separators, groups, URL entries and dashboards are lost on every save.
Measured with a throwaway probe through `EditAppPage` and the real
wizard: a stored `[object, separator, group with a dashboard child,
url]` saved as `[object]` only.

---

_Generated by [Claude
Code](https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
…ts name 41 objectui issues that answer 404 (objectui#10803, batch 5) (objectstack-ai#10892)

Part of objectstack-ai#10803
Clause-②: no

Dispatched implementation of the `domain:ui` seat 2 claim (comment
`5862515185`) on objectui#10803, batch 5, session
`https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN`. Citations
only: no sentence's claim moves, and every edited pending changeset's
frontmatter is byte-identical. The runtime text that moves loses its
dead pointer and nothing else, all in `@object-ui/types`: the six
`ChatbotSchema` tombstone guidance strings open `RETIRED (ADR-0049) —`
(rulings `5861129870` and `5861311219`), the `page` node's `actions`
refusal, an ADR-0049 tombstone too, cites `(ADR-0049)` in place of its
dead card (ruling `5861129870`, as measured in patch round 1), and the
`ChartDataSeriesSchema.variant` description drops its pointer (amendment
`5860244997`, Q1 = A). Two tests that pin changed text follow it, anchor
only (ruling `5861900779`, route a).

**Patch round 1** answers the contract review's one FAIL item on
`afed8aa6b`: Special cases 7 had read the `actions` refusal as not
ADR-0049 from a proxy (the `breadcrumbs` changeset, which only says the
2026-09-09 RULING is not borrowed). Measured from the retirement itself,
it is an ADR-0049 tombstone, so under the seat's option (a) its string
now reads "never was (ADR-0049): no renderer reads it". The
runtime-strings changeset follows, and the sweep changeset notes that
`@object-ui/types`' runtime text is declared separately. One text-only
commit, `8255f14f4`; no merge, no force-push.

## Why `Part of`, not a closing line

The card stays open for the other 23 numbers of the family remainder
(seat amendment `5860244997`, Q2 = A), listed in **Acceptance notes** 1,
and for the bare sister-repo numbers the C0 census below lists.

## Premise, re-measured

- REST `GET /repos/objectstack-ai/objectui/issues/N` for each of the 41:
all 41 answer 404, and a second read of each answers 404 again (41 of
41). `GET .../pulls/N` answers 404 for all 41 too. Lit controls:
objectui#10533 and objectui#7714 answer 200. No number was dropped from
the batch.
- The same 41 in objectstack all answer 200. Three of them are the
sites' real subject, read not guessed, and are re-qualified (Zone 1 item
5):
- objectstack#3720 "列表行级 edit/delete 与批量删除未接入服务端 effective 操作集": the
five `[objectstack-ai#3720]` sites in `ObjectGrid.tsx`, `rowCrudAffordances.ts` and
`ListView.tsx` describe that gate, and their landing `ba4514511` is
titled "gate list row Edit/Delete and bulk delete on the effective
operation set (objectstack#3720)".
- objectstack#5506, the "five more hardcoded English chrome strings"
card: `useDetailTranslation.ts` wrote "the SAME key objectstack-ai#5506 gave
NavigationOverlay's identical handle", and nine sibling in-scope lines
already cite `objectstack#5506` for those strings.
- objectstack#6936, the metadata-admin predicate fail-CLOSED card:
`predicate.ts` wrote "objectstack-ai#6936's `true` verdict" one line below
"objectstack#6936's unresolved-path ruling".
- Every other site names an objectui card or pull request, confirmed by
its landing commit below (for example objectstack#7853 is a
`driver-memory` finding, while objectui#7853 is the cite-by-content
ruling AGENTS.md objectstack-ai#11 names).
- Every edited changeset is pending: it is present in `.changeset/` on
`main`.
- Branch point: `b2683a2c0`. `main` is now `4f8b7f86a`; its census still
reads 215, and the merge of this head with it reads 0. Not a shallow
checkout.

## C0: the bare-number remainder (measurement only)

The instrument of PR objectui#10875, at the branch point:

```
git grep -hoP '(?:^|(?<=[^\w#&/.\-]))#\d+(?![0-9A-Za-z_])' b2683a2 -- '.changeset/*.md' 'packages/*/src/**' ':!**/__tests__/**' ':!**/*.test.*' ':!**/*.spec.*' ':!**/test/**' ':!**/tests/**' | sort -u | wc -l
```

- **998 distinct bare numbers** at `b2683a2c0` (1002 at batch 4's
`e32dae160`): the eight batch 4 re-pointed are gone, and four are new
since: objectstack-ai#10321, objectstack-ai#8365 and objectstack-ai#9242 (at or below 10900, each read once: 200)
and #20051.
- **Above 10900: 23**, batch 4's 22 plus #20051. Each was read once with
REST `GET .../issues/N`: **23 answer 404**, and each answered 404 again
on a second read (23 of 23); `GET .../pulls/N` answers 404 for the 18
that are not colours.
- None of them is an objectui citation, and none sits in a sentence or
hunk this batch edits, so **nothing is folded here**:
  - **5 are CSS colours:** #334155, #336699, #475569, #737373, #953800.
- **18 are objectstack numbers**, 16 of which answer 200 there with the
sites' own subject. #13033 and #13413 answer 404 in objectstack too.
Sites (lines / files, at `b2683a2c0`) are listed in **Acceptance notes**
2 for later batches.

## Census (the enumeration pin for this batch)

The instrument PR objectui#10854 printed and PRs objectui#10869 / objectstack-ai#10875
reused, with this batch's 41 numbers substituted (REF = a commit or
tree):

```
git grep -nE '(objectui#|#|issues/)(7091|7097|7108|7177|7612|7620|7623|7658|7666|7667|7669|7678|7681|7682|7703|7704|7708|7804|7844|7853|7869|7874|7877|7926|7959|7967|7979|7980|8058|8060|3720|5420|5503|5506|5737|6467|6936|6945|7622|7662|7684)([^0-9]|$)' REF -- '.changeset/*.md' 'packages/*/src/**' ':!**/__tests__/**' ':!**/*.test.*' ':!**/*.spec.*' ':!**/test/**' ':!**/tests/**' | grep -v 'objectstack#' | wc -l
```

- REF = `b2683a2c0` (branch point, and `main` now): **215** lines, 74
changeset lines in 55 files and 141 src lines in 47 files.
- REF = `8255f14f4` (this head): **0**. Round 0's head `afed8aa6b` also
read 0.
- This head merged with a fresh `main` (`4f8b7f86a`, `git merge-tree
--write-tree`, clean, tree `08f80168a`): **0**. REF = `4f8b7f86a` alone:
215.
- **The filter hid nothing this time.** The 18 branch-point lines the
`grep -v 'objectstack#'` drops name only objectstack#6936 or
objectstack#5506; with those tokens deleted, none of them matches the
instrument (0). Unfiltered, this head reads **26**, all live
`objectstack#` lines: the 18, the seven re-qualified here, and the same
spelling in the new sweep changeset.
- Lit control, the same printed instrument over live objectui#7714 at
this head: 17 lines. Hex-colour false positives (a number followed by a
hex letter) at the branch point: 0.
- **Out of scope, as it stands** (the instrument with the 41 numbers,
unfiltered, whole tree at this head): 367 test lines in 98 files, 23
scripts lines in 5 files, 0 `.github`, 5 governed lines in 4 files
(`AGENTS.md` and three `skills/objectui` files), and 51 other lines in
18 files: published `CHANGELOG.md` history (21 lines in 8 files),
`content/docs` (11 in 4), `apps/console` (7 in 3), `eslint-rules` (8 in
2) and `eslint.config.js` (4). See **Acceptance notes** 4.

## Citation form

- The 9-character backticked sha of the commit on `main` that landed the
change the sentence rests on, as in PRs objectui#10707 / objectstack-ai#10766 / objectstack-ai#10797
/ objectstack-ai#10854 / objectstack-ai#10869 / objectstack-ai#10875. All 40 distinct shas in the mapping are
ancestors of `main`: `git merge-base --is-ancestor`, exit 0 each.
Control leg in the same checkout: the head of PR objectui#10714
(`d2afdb6bf`) answers exit 1, and a known ancestor, `5f789538d`, answers
exit 0. `git rev-parse --short=9` returns the same 9 characters for
each.
- A changeset's pointer to its OWN card is dropped, not replaced:
`@changesets/changelog-git` prefixes each released entry with the hash
of the commit that added the file, which is that landing. Where dropping
it left a dangling word, the clause was minimally repaired (**Special
cases** 5).
- Where the sentence rests on a RULING, the ruling is cited by its date
(**Special cases** 4).
- Where a dead card sits beside a live pull request that IS its landing,
the dead card is dropped and the live pointer stays (**Special cases**
1).
- Where the claim lived only on the dead card (a finding, a ruling, a
scope, a comment), the sha only locates it: "the card behind SHA", "the
ruling behind SHA" (**Special cases** 3). Review `5861306588`'s lesson:
a sha must carry what the sentence cites it for.
- Where nothing answers, the pointer is dropped and the sentence names
the card by role (objectstack-ai#7612, objectstack-ai#7666, objectstack-ai#7682, objectstack-ai#7844, and the open objectstack-ai#7804
parent).
- Runtime text carries no sha: see **Special cases** 7.

## Mapping (number to resolution)

Lines / files are the branch-point census for that number. "Method" is
how the landing was found. The last column is why that commit carries
what the sentence cites it for.

| dead number | resolution | method | lines / files | why the commit
carries it |
|:--|:--|:--|:--|:--|
| objectstack-ai#7091 (a pull request) | `969ba84f4` | `git log --grep`: its squash
subject ends "(objectstack-ai#7091)" | 1 / 1 | it registers the console `app:launcher`
renderer, which is what made the "not a page" clause false |
| objectstack-ai#7097 | `3cab5703b`; own-card pointer dropped | changeset's adding
commit | 3 / 3 | it adds the `2xl` read arm and the literal
`GRID_COLS_2XL` class map, the value-level fix both sentences name |
| objectstack-ai#7108 | `1267508ea`; own-card pointer dropped | changeset's adding
commit | 3 / 2 | it tombstones a deleted record reference in the
approvals inbox, the platform-flagged class the sentences name |
| objectstack-ai#7177 | own-card pointer dropped (landing `ebc05b4d6`) | changeset's
adding commit | 1 / 1 | the only site is its own changeset's pointer |
| objectstack-ai#7612 | "the 2026-09-20 ruling"; the card by role | `git log --grep`
finds only `2ca238075`, the objectui#10150 record of the ruling; the
switch, PR objectui#10054, closed unmerged | 3 / 2 | nothing landed: the
ruling was "no". See **Special cases** 3 and 4 |
| objectstack-ai#7620 | `c15d7eca6`; "ruling A behind `c15d7eca6`" where the sentence
cites the ruling; own-card pointer dropped | changeset's adding commit,
`git log -S` | 13 / 10 | it moves `data-table`'s date-only fallback onto
`formatDate`'s default face, and its message records the two-faces fact
pattern the sentences cite |
| objectstack-ai#7623 | `5d0876c5c`; own-card pointer dropped | changeset's adding
commit | 3 / 3 | it retires the inert dashboard-root `title` declaration
|
| objectstack-ai#7658 | `90c6d090d`; own-card pointer dropped | changeset's adding
commit | 2 / 2 | it keeps code-span content out of `extractToc`'s HTML
strip |
| objectstack-ai#7666 | nothing answers: the card by role ("already-filed defect") |
`git log -S` finds only `a472b0716`, which records it as an
already-filed defect | 1 / 1 | see **Special cases** 3 |
| objectstack-ai#7667 | `a472b0716`; own-card pointer dropped | changeset's adding
commit | 2 / 2 | it gives `extractToc`'s emphasis rules the flanking
rule that keeps `SCREAMING_SNAKE` underscores |
| objectstack-ai#7669 (a pull request) | `a3eb5d07a` | `git log --grep`: its squash
subject ends "(objectstack-ai#7669)" | 2 / 2 | it retires
`FloatingChatbotConfig.triggerIcon` as the tombstone both sentences name
|
| objectstack-ai#7678 | `5f8190c8c` ("in its amended form"); own-card pointers dropped
| changeset's adding commit; its body opens "Part of objectstack-ai#7678" | 21 / 14 |
it brings every statement of the retire-vs-remove discriminator to the
amended form, the surviving-carrier precondition included. See **Special
cases** 6 |
| objectstack-ai#7681 | `01c27c431`; own-card pointer dropped | changeset's adding
commit | 6 / 4 | it routes a `dataKey`-shaped series through
`normalizeSeries`, removing the fast path, so its `type` override is
honoured |
| objectstack-ai#7682 | nothing answers: "a separate card"; runtime pointer dropped |
`git log -S` finds only `8fe8e5c16`, which defers that decision to it |
3 / 3 | see **Special cases** 3 |
| objectstack-ai#7703 | `a4611b3e2`; runtime text `RETIRED (ADR-0049) —`; own-card
pointer dropped | changeset's adding commit | 15 / 3 | it retires the
six dark `ChatbotSchema` keys as ADR-0049 tombstones |
| objectstack-ai#7704 | `20316bac3`; "the card behind `20316bac3`" for the scope note;
own-card pointer dropped | changeset's adding commit | 3 / 2 | it adds
`chatbot-enhanced` and `chatbot-floating` to `SchemaRegistry` |
| objectstack-ai#7708 | `d3499b315`; PR objectstack-ai#8077 kept where it stands beside the dead
card; "the card behind `d3499b315`" for the ruling; own-card and runtime
pointers dropped | changeset's adding commit, `git log --grep` | 13 / 6
| it fences `chatbot-floating`'s raw props spread through `toDomProps`
at the head of the element |
| objectstack-ai#7804 (the handler-key parent) | per slice, by `git blame`:
`75fca9669` (data-table), `7ca6ddd4b` (detail), `f1cd29032` (list-view),
`5a41ce733` (object-kanban), `8d50bc2bf` (objectql.ts), `604476d97`
(tree-view); own-card pointers dropped; "the handler-key parent card"
where the sentence means the open parent | each slice changeset's adding
commit, and `git blame` per line | 49 / 17 | each commit declares
exactly the handler keys the site's arm or face names |
| objectstack-ai#7844 | nothing answers: "their own card" | `git log -S` finds only
`c4a1d39fe`, which leaves both sub-shapes dark | 1 / 1 | see **Special
cases** 3 |
| objectstack-ai#7853 | "the card behind `fa7d66c45`" | AGENTS.md objectstack-ai#11 names
`fa7d66c45` as its landing | 3 / 3 | the ruling lived on the card;
`fa7d66c45` is its one-site landing |
| objectstack-ai#7869 | "the finding objectui#8344 discharged" / "the finding this
card discharges" (landing `841dd2b7e`) | `841dd2b7e`'s message; the live
objectui#8344's body, "Discharges the finding objectstack-ai#7869" | 2 / 2 | the
measurement lived on the dead finding, and the live card that discharged
it restates it |
| objectstack-ai#7874 | `d9580f464`; PR objectstack-ai#7887 kept beside the dead card; own-card
pointer dropped | changeset's adding commit | 3 / 3 | it retires the
five dead `timeline.relative.*` rows |
| objectstack-ai#7877 | `7dcda9cd4` | `git log -S` on the heading | 1 / 1 | it moves
the hand-rolled-table data into `hand-rolled-tables.json`, the heading's
subject |
| objectstack-ai#7926 | `12b599219`; "the 2026-09-09 ruling"; "the `actions` refusal"
by role beside PR objectstack-ai#8870; runtime text `(ADR-0049)`; own-card pointer
dropped | changeset's adding commit | 15 / 6 | it refuses `actions` by
name on the `page` node, measures `breadcrumbs` at 1 site and leaves it
parsing, and pins that |
| objectstack-ai#7959 | `36fc74629`; "the card behind `36fc74629`" for its scope
restraint | the tests it added are headed "objectui#7959" | 8 / 6 | it
adds `readEnvelopeFailureText`, the userMessage-first rule on
`PackagesPage`, and the load-failure testid |
| objectstack-ai#7967 | own-card pointer dropped (landing `0544ae958`) | changeset's
adding commit | 1 / 1 | the only site is its own changeset's pointer |
| objectstack-ai#7979 | own-card pointer dropped (landing `bfa4fe728`) | changeset's
adding commit | 1 / 1 | the only site is its own changeset's pointer |
| objectstack-ai#7980 | `a810bb2ae`; "the card behind `a810bb2ae`" for its surface and
its comment; own-card pointer dropped | changeset's adding commit | 4 /
3 | it publishes `readEnvelopeFailureText` from the package entry |
| objectstack-ai#8058 | `3f4b45886` | `git log -S storedRelationshipTarget` | 5 / 2 |
it adds `storedRelationshipTarget`, the read door under both spellings,
and rewrites the "costs nothing" claim |
| objectstack-ai#8060 | `7dc31bb8d`; "step 1 of the card behind `7dc31bb8d`" | `git
log --grep`: its subject names "(objectstack-ai#8060)" | 8 / 3 | it carries a stored
type the designer cannot author through verbatim
(`partitionStoredFields`) instead of flattening it to `text` |
| objectstack-ai#3720 (C0) | re-qualified `objectstack#3720` | `ba4514511`'s subject;
objectstack#3720's subject | 5 / 3 | see **Premise** |
| objectstack-ai#5420 (C0) | own-card pointer dropped (landing `b470e9150`) | the
adding commit's subject names "(objectstack-ai#5420)" | 1 / 1 | the only site is its
own changeset's heading |
| objectstack-ai#5503 (C0, a pull request) | `6f017e99c` | its squash subject ends
"(objectstack-ai#5503)" | 1 / 1 | it puts the chart min-height floor on the element
Recharts measures |
| objectstack-ai#5506 (C0) | re-qualified `objectstack#5506` | nine sibling in-scope
lines already cite it for the same strings | 1 / 1 | see **Premise** |
| objectstack-ai#5737 (C0, "PR objectstack-ai#5737") | `d1ab06f0f` | its squash subject ends
"(objectstack-ai#5737)" | 4 / 4 | it declares `record.*` the row-predicate canon and
warns on the two deprecated spellings, which is Phase 1 |
| objectstack-ai#6467 (C0) | own-card pointer dropped (landing `5ad0641e0`) | its
subject names "(objectstack-ai#6467)" before its own "(objectstack-ai#7332)" | 1 / 1 | the only site
is its own changeset's pointer |
| objectstack-ai#6936 (C0) | re-qualified `objectstack#6936` | the line above already
cites it | 1 / 1 | see **Premise** |
| objectstack-ai#6945 (C0, "PR objectstack-ai#6945") | `2c45966ff` | its squash subject ends
"(objectstack-ai#6945)" | 1 / 1 | it declares the 13 renderer-read keys the sentence
says objectui#6150 declared first |
| objectstack-ai#7622 (C0, "PR objectstack-ai#7622") | `1cca678ba` | its squash subject ends
"(objectstack-ai#7622)" | 1 / 1 | it retires the dashboard-root `title` read on all
five surfaces |
| objectstack-ai#7662 (C0, "PR objectstack-ai#7662") | `bc640ec56` | its squash subject ends
"(objectstack-ai#7662)" | 1 / 1 | it is the change to `SchemaRegistry`'s kanban entry
the sentence ascribes to objectui#7645 |
| objectstack-ai#7684 (C0, "PR objectstack-ai#7684") | "the contract review of `8fe8e5c16`'s pull
request" | its squash subject ends "(objectstack-ai#7684)" | 2 / 2 | the review lived
on the dead pull request; the sha only locates it |

## Special cases (the judgement calls)

1. **A dead card beside its own live landing.**
- `defaults-table-scan.ts`: "rows of objectui#7874 at all (retired in
objectstack-ai#7887)" becomes "rows at all (retired in objectstack-ai#7887)". PR objectstack-ai#7887 answers 200
and is `d9580f464`'s pull request.
- `7703-chatbot-dark-keys-retired.md`: "**after** objectui#7708's fence
landed as PR objectstack-ai#8077" becomes "**after** the fence landed as PR objectstack-ai#8077". In
the same file "until objectui#7708 the `chatbot-floating` registration
ended…" becomes "until PR objectstack-ai#8077 the…", and the next sentence's "That
card was ruled **fence**, not declare, and landed as PR objectstack-ai#8077" becomes
"Its card was ruled **fence**, not declare, and landed as that PR", so
it keeps an antecedent.
- `8871-page-node-refuses-breadcrumbs.md` and `layout.zod.ts`: "corrects
objectui#7926's "1 site"", beside "(PR objectstack-ai#8870)", becomes "corrects the
`actions` refusal's "1 site"". The changeset's "objectui#7926's `patch`
does not transfer here" becomes "The `actions` refusal's `patch` does
not transfer here".
2. **The sister-repo numbers.** `[objectstack-ai#3720]` becomes `[objectstack#3720]`
at five sites; "objectstack-ai#5506" and "objectstack-ai#6936's" become `objectstack#5506` and
`objectstack#6936`'s. See **Premise**.
3. **The card named by role.**
- **objectstack-ai#7612**, `clientValidation.ts`: "The switch was then put to the
maintainer on objectui#7612" becomes "…on its own card". Nothing landed:
the ruling said no, and the switch PR closed unmerged. The verbatim
ruling quote 「7612 只需要服务端校验」 stays as the maintainer wrote it; it
carries no `#`.
- **objectstack-ai#7666**, `7667-toc-emphasis-flanking.md`: "(objectui#7666, a heading
`extractToc` lists…)" becomes "(a heading `extractToc` lists…)". The
sentence already calls it "a different, already-filed defect".
- **objectstack-ai#7682**: "objectui#7682 owns that decision" becomes "a separate card
owns that decision" (`7546-chart-series-keys-declared.md`), and
"objectui#7682's decision" becomes "a separate card's decision"
(`data-display.ts`).
- **objectstack-ai#7844**, `8754-…` changeset: "the two further sub-shapes
objectui#7844 records" becomes "the two further sub-shapes recorded on
their own card".
- **objectstack-ai#7804 as the open parent**: "whose disposition is still open on
objectui#7804" becomes "…on the handler-key parent card" (`9447-…`);
"both are ledgered to objectui#7804, which owns it" becomes "…to the
handler-key parent card" (`handler-key-reads-follow-cast-receivers.md`);
`crud.zod.ts`'s "objectui#7804's rows are the handler keys" becomes "the
handler-key parent card's rows…".
- **"The card behind SHA"**, where the sentence rests on something that
lived only on the card:
- objectstack-ai#7620's ruling A: "following the maintainer's ruling A on
objectui#7620" becomes "…ruling A behind `c15d7eca6`" in four
`@object-ui/fields` widgets, and `GridField`'s "the split objectstack-ai#7620 ruled
on" becomes "the split ruling A settled". `RelatedList.tsx`:
"objectui#7443 and objectui#7620 spent two cards" becomes "objectui#7443
and the card behind `c15d7eca6` spent two cards".
- objectstack-ai#7708's ruling: "objectui#7708 ruled FENCE" becomes "The card behind
`d3499b315` ruled FENCE", and "That ruling was FENCE" after it becomes
"The ruling behind it was FENCE" (`complex.ts`).
- objectstack-ai#7853: "objectui#7853 ruled…" becomes "The card behind `fa7d66c45`
ruled…" in both objectstack-ai#8875 changesets and in `defaults-table-scan.ts`'s list.
- objectstack-ai#7959's scope restraint: "objectui#7959 kept/left it off the public
entry as scope restraint — that card's file surface was…" becomes "The
card behind `36fc74629` …" (`7980-…` changeset, `app-shell` `index.ts`).
- objectstack-ai#7980: "see objectui#7980 comment 5583988475" becomes "as the card
behind `a810bb2ae` records" (a comment on a dead card resolves nowhere);
`PackageFormDialog.tsx`'s "is objectui#7980's surface" becomes "is the
surface of the card behind `a810bb2ae`".
- objectstack-ai#7704's scope note (`registry.ts`) and objectstack-ai#8060's "step 1"
(`MetadataFieldsPage.tsx`) take the same form.
- objectstack-ai#7869: the finding card is dead, and the live objectui#8344 discharged
it and restates the measurement, so `base.zod.ts` reads "That is the
finding objectui#8344 discharged: an off-spec `size`…" and objectstack-ai#8344's own
changeset reads "the finding this card discharges measured…".
4. **Rulings by date.**
- objectstack-ai#7612: the `AUTHOR_SHAPE_ONLY_TYPES` heading ends "(the 2026-09-20
ruling)"; `10150-author-shape-header-ruling.md` reads "the switch was
put to the maintainer, and the 2026-09-20 ruling was option A". Its last
sentence, "requires the docblock to name the ruling's card", becomes
"…to name the ruling", because the pin it describes now anchors on the
ruling (**Special cases** 8).
- objectstack-ai#7926: "objectui#7926's maintainer ruling covers `actions`" becomes
"The 2026-09-09 maintainer ruling covers `actions`", and "objectui#7926
does not rule on it" becomes "the 2026-09-09 ruling does not cover it"
(`layout.zod.ts`).
5. **Own-card pointers that needed a word.**
- `7097-grid-2xl-breakpoint.md`: the pointer sat on its own line, so the
period moved up with it.
- `7678-tombstone-discriminator-agreement.md`: "the site objectui#7678
was filed about" becomes "the site this card was filed about", the
file's own wording two lines earlier.
- `7804-object-kanban-handler-keys-judged.md`: "its
`KNOWN_UNDECLARED_READS` row naming objectui#7804, which stays open and
stays the parent" becomes "its `KNOWN_UNDECLARED_READS` row, which names
the parent card".
- The six `7804-*` slice changesets keep "(the … slice)" once the
pointer goes, as batch 4 kept "(slice 2)".
6. **The discriminator form (objectstack-ai#7678).** "in the form objectui#7678
amended it to" becomes "in its amended form (`5f8190c8c`)" at every
site; "as amended by objectui#7678" and "amended by objectui#7678" take
the same wording; a bare "objectstack-ai#7678" in a precedent list becomes
`5f8190c8c`. `mobile.ts`'s "without the precondition objectui#7678
added" becomes "without the precondition the amendment added", whose
antecedent is the preamble's "amended form" two sentences up.
`9684-discriminator-cited-once.md`: "objectui#7678's amendment had to be
applied by hand" becomes "the amendment (`5f8190c8c`) had to be applied
by hand", which is what that commit did.
7. **The runtime strings.** Only the listed text moves in each.

| file | member | before | after |
|:--|:--|:--|:--|
| `types/src/zod/complex.zod.ts` | `ChatbotSchema.loading`,
`.showAvatars`, `.userAvatar`, `.assistantAvatar`, `.markdown`,
`.height` (tombstones) | "RETIRED (objectui#7703, ADR-0049) — …" |
"RETIRED (ADR-0049) — …" |
| `types/src/zod/complex.zod.ts` | `ChatbotSchema.showAvatars`
(tombstone) | "… unfiltered props spread — was fenced by objectui#7708.
Delete the key: …" | "… unfiltered props spread — was fenced. Delete the
key: …" |
| `types/src/zod/layout.zod.ts` | `PageNodeSchema.actions` refusal |
"`actions` is not a key of the `page` node and never was
(objectui#7926): no renderer reads it, …" | "`actions` is not a key of
the `page` node and never was (ADR-0049): no renderer reads it, …" |
| `types/src/zod/data-display.zod.ts` | `ChartDataSeriesSchema.variant`
`.describe()` | "… the renderer-internal `current` spelling is not a
member (objectui#7682)" | "… the renderer-internal `current` spelling is
not a member" |

**The ADR-0049 condition, measured from the retirement itself** (Zone 1
item 4): the six tombstones already cited `ADR-0049` beside the dead
card; the zod comment above them opens "ADR-0049 retirement tombstones";
their landing `a4611b3e2` is titled "retire six dark ChatbotSchema keys
as ADR-0049 tombstones"; and the pending
`7703-chatbot-dark-keys-retired.md` says "ADR-0049 retirement
tombstones". The `actions` refusal is an ADR-0049 tombstone too,
measured the same way: its arm is
`retirementTombstone(PAGE_ACTIONS_REFUSAL)`, its landing `12b599219`
calls it "an ADR-0049 `retirementTombstone` arm", and
`7926-page-node-refuses-actions.md` calls it "an ADR-0049 refusal arm".
Its string has no `RETIRED` opener, so its own parenthetical becomes
`(ADR-0049)`, the spelling its `breadcrumbs` sibling already uses. The
TypeScript twins in `complex.ts` read "ADR-0049 RETIREMENT TOMBSTONE —
`loading` (`a4611b3e2`)", the comment form.
8. **What follows the text, and nothing else in it** (ruling
`5861900779`, route a; ruling `5861129870`).
- `chatbot-dark-keys-retired-7703.test.ts` pins the six guidance strings
as full literals. Their openings follow to `RETIRED (ADR-0049) — `, the
`showAvatars` literal follows "was fenced.", its `prescriptive` anchor
`'was fenced by objectui#7708'` becomes `'was fenced.'`, and
`toContain('RETIRED (objectui#7703')` becomes `toContain('RETIRED
(ADR-0049) —')`. Its test names and `@ts-expect-error` comments still
read "RETIRED (objectui#7703)"; they describe the `?: never` TypeScript
face and are out of class (**Acceptance notes** 6).
- `clientValidation.authorShapeRuling-10150.test.ts` pins that the
`AUTHOR_SHAPE_ONLY_TYPES` header names the ruling and
`extractDraftBody`. Its subject `'objectui#7612'` becomes `'the
2026-09-20 ruling'`; its message and title are unchanged.
- No test, doc or changeset quotes the `actions` refusal or the
`variant` description verbatim: a whole-tree search for the old strings
returns only the source lines.
9. **The anchor sweep (PR objectui#10875 Special cases 10 shape).**
- **Instrument:** every string, template, numeric and regex literal in
all 3822 test and script files under `packages/` and `scripts/`,
enumerated with `git ls-tree` at head and parsed with TypeScript (113091
distinct literals).
- **Test applied to each literal:** does its occurrence count DROP
between `b2683a2c0` and head in any file this PR changes, in raw text or
in a comment-flattened form?
- **Run 1, on `8c267051f`** (the sweep commit, before any test moved):
336 distinct literals drop. Read against the files they sit in, the
specific anchors are the two pins in item 8 (the six literals, the
`prescriptive` anchor and the `toContain` in the first;
`'objectui#7612'` in the second). Also listed and read: the
`'objectui#7804'` values of `check-handler-key-read-sites`'s
`KNOWN_UNDECLARED_READS` ledger and its test, which assert over the
script's own data and never read a changed file, and
`tombstone-discriminator-agreement-9684`'s `/RETIRED \(objectui#/`,
which it applies to `mobile.ts` alone, untouched here.
- **Everything else** is a generic token (whitespace, punctuation,
single letters and digits, `objectui#`, character-class regexes of
whole-tree scanners), none an assertion over text this PR changed.
- **Run 2, on round 0's head `afed8aa6b`:** 327 distinct literals; the
nine literals the two pins moved are gone from the drop set, nothing new
appears, and the only one left naming a batch number is that ledger
value.
- **Run 3, patch round 1, over the delta `afed8aa6b..8255f14`:** 18
(literal, file) drops, every one a generic token (`' is'`, `'dropped'`,
a `/(\w+):/g` scanner); no test pins the changed refusal text.

## Held

**By the serial rule: nothing.** Patch round 1 re-read: 8 open besides
this PR (objectui#10896, objectstack-ai#10891, objectstack-ai#10890, objectstack-ai#10889, objectstack-ai#10888, objectstack-ai#10777, objectstack-ai#10278
and the release PR objectstack-ai#5400; objectstack-ai#10886 and objectstack-ai#10714 have left the open list).
New since round 0, objectui#10896 shares
`plugin-grid/src/ObjectGrid.tsx` and `plugin-list/src/ListView.tsx`;
trial merges of this head with every open head are clean except objectstack-ai#10278
(conflicts identical against `main` alone) and the release PR objectstack-ai#5400,
which regenerates. The round's one source edit, in `layout.zod.ts`, is
held by no open PR.

Round 0, re-read before its push: 11 open besides this PR. They are
objectui#10890, objectstack-ai#10889, objectstack-ai#10888, objectstack-ai#10886, objectstack-ai#10884, objectstack-ai#10852, objectstack-ai#10780, objectstack-ai#10777,
objectstack-ai#10714, objectstack-ai#10278 and the release PR objectstack-ai#5400.
- Five of them hold files this PR edits:
  - objectui#10890: `types/src/ai.ts`;
  - objectui#10888: `.changeset/8344-node-recursion-point-redirect.md`;
- objectui#10886: `types/src/objectql.ts` and
`types/src/zod/objectql.zod.ts`;
- objectui#10714: `types/src/data-display.ts`, `types/src/layout.ts`,
`types/src/zod/data-display.zod.ts` and `types/src/zod/layout.zod.ts`;
  - objectui#10278: `plugin-grid/src/ObjectGrid.tsx`.
- Each one's hunks were read against its merge-base, with 3 lines of
context. **None of the 47 lines this PR changes in those files appears
inside or beside any of their hunks.**
- Trial merges of each open PR head with this head (`git merge-tree
--write-tree`) are clean for objectstack-ai#10890, objectstack-ai#10889, objectstack-ai#10888, objectstack-ai#10886, objectstack-ai#10884,
objectstack-ai#10852, objectstack-ai#10780, objectstack-ai#10777 and objectstack-ai#10714.
- objectstack-ai#10278 (`eab4c8e52`) conflicts in `ObjectGrid.tsx`,
`plugin-grid/README.md` and `content/docs/plugins/plugin-grid.mdx`, and
conflicts identically against `main` alone.
- PR objectstack-ai#5400 (Version Packages) regenerates and is not a hold.

## Changesets

- `.changeset/10803-dead-citation-sweep-fifth-batch.md`, EMPTY
frontmatter: the comment-only edits in 15 released packages and the
private `@object-ui/test-support`. No published behaviour changes
through them. It marks `@object-ui/types` as the package "whose runtime
text is declared separately below" (patch round 1) and points at the
second file for it.
- `.changeset/10803-fifth-batch-runtime-strings.md`,
`'@object-ui/types': patch`: the six tombstones open `RETIRED (ADR-0049)
—` with everything after the dash unchanged except the `showAvatars`
pointer; the `actions` tombstone cites `ADR-0049` in place of its dead
card; and the `variant` description loses its pointer. No key, path,
issue code, accept set, refusal or severity moves.

## Proof of prose-only (C4), against `b2683a2c0`, on this head
`8255f14f4`

- **Source.** Each of the 49 touched `.ts` / `.tsx` files (47 non-test
sources and the two pins) was parsed at `b2683a2c0` and at this head
with TypeScript's `createSourceFile` and re-printed by `createPrinter({
removeComments: true })`.
  - 44 of 49 prints are identical.
- The other 5 are equal once exactly the 19 listed substitutions are
applied to the base print, each matched once: `complex.zod.ts` 7 (six
openings and the fence pointer), `layout.zod.ts` 1 (the pointer becomes
`(ADR-0049)`), `data-display.zod.ts` 1,
`chatbot-dark-keys-retired-7703.test.ts` 9 (the same seven, the
`prescriptive` anchor and the `toContain` anchor),
`clientValidation.authorShapeRuling-10150.test.ts` 1.
  - 0 parse diagnostics.
- Lit controls on the same instrument: dropping "(ADR-0049)" from the
`loading` string moves the print; re-spacing the "ADR-0049 retirement
tombstones" comment does not.
- **Changesets.** The frontmatter block of every one of the 55 edited
changesets is byte-identical at `b2683a2c0` and this head (55 of 55, by
md5). The overwrite gate below agrees.
- **Scope of the diff:** 106 files: 55 edited and 2 new changesets, 47
non-test source files, and 2 tests.

## Gates, on this head `8255f14f4`

Each line is the gate's own verdict and exit code, captured by
redirect-then-`$?`.

- `node scripts/check-changeset-presence.mjs`, exit 0: "48 source
file(s) of 15 released package(s) changed, and this change declares 2
changeset(s): .changeset/10803-dead-citation-sweep-fifth-batch.md,
.changeset/10803-fifth-batch-runtime-strings.md." (48 = the 46 released
sources and the 2 tests under released packages' `src/`; `test-support`
is under a package changesets ignores.)
- `pnpm changeset:check`, exit 0: "All workspace packages are in the
changeset fixed group." / "No changeset declares a `major` bump."
- `node scripts/check-changeset-overwrite.mjs` (report-only), exit 0: "2
changeset(s) added, 55 modified, 0 deleted". `declared at base` equals
`declares now` for 55 of 55.
- `pnpm check:changeset-claims` (report-only), exit 0:
- born-false: "Every one of those 2 address(es) either names the tree it
was read from, or points at a line this change does not move";
- self-contradiction: "Every package declared across those 45 body(ies)
is either not negated in its own prose, or negated only in an ASPECT of
it that may legitimately hold still";
- the standing notice "103 pending changeset(s) describe a file this
change touches".
- Read against the diff: a pending changeset quoting a replaced pointer
would itself carry the dead number and so sit in the census, which reads
0. No pending changeset quotes a changed runtime string.
- `pnpm check:control-bytes`, exit 0: "check-control-bytes: OK (scanned
9145 tracked text file(s); skipped 85 binary)."
- `pnpm check:new-line-citations`, exit 0: "VERDICT
new-cross-file-line-citations: 0 new citation(s), enforcement
report-only -> exit 0".
- Also run: `pnpm check:pending-changeset-literals`, exit 0: "No test
source names a pending changeset."; the governed-surface predicate over
the 106 paths, exit 0: "NOT GOVERNED — 106 path(s) checked against 5
governed surface(s); none matched." (lit control `AGENTS.md`: exit 3).

**Tests**, through the shared verify lock.
- **Red, reproduced first.** The branch-point copies of the two pins
were written over the committed sweep tree and run: `Test Files 2 failed
(2)`, `Tests 13 failed | 29 passed (42)`, for example "expected 'RETIRED
(ADR-0049) — never read: chat…' to contain 'RETIRED (objectui#7703'".
Both files were then restored with `git checkout HEAD --`, each checked
by blob hash against `HEAD`, and `git diff HEAD` came back empty.
- **Patch round 1, on this head `8255f14f4`:**
`page-actions-refusal-7926` (it pins the message's remedy words
`actions`, `children` and `page:header` and that it is not zod's generic
text, never the pointer; verified by reading it),
`page-breadcrumbs-refusal-8871`, `guide-layout-page-buttons-7926`,
`zod-mirror-parity`, `chatbot-dark-keys-retired-7703`,
`clientValidation.authorShapeRuling-10150` and
`tombstone-discriminator-agreement-9684` give `Test Files 7 passed (7)`,
`Tests 134 passed (134)`, under `VERDICT command-exit 0`. A whole-tree
search for "never was: no renderer" finds no test or doc quoting it
(only the source line and the runtime-strings changeset, both updated);
an anchor sweep over the round's delta finds only generic tokens.
- **Green, round 0, on `afed8aa6b`:** `chatbot-dark-keys-retired-7703`,
`clientValidation.authorShapeRuling-10150`,
`tombstone-discriminator-agreement-9684`, `page-actions-refusal-7926`,
`page-breadcrumbs-refusal-8871`, `chart-series-keys-7546`,
`zod-mirror-parity`, `chatbot-display-mode-retired` and
`schema-registry-chatbot-keys-7704` give `Test Files 9 passed (9)`,
`Tests 173 passed (173)`, under `VERDICT command-exit 0`.

**Declared narrowing (per the dispatch: the named checks, not wider
sweeps).** NOT MEASURED locally: the full suites and type-check of the
15 touched packages, and eslint. Reason: the comment-stripped syntax
tree of 44 of the 49 touched files is identical to `main`, and the other
5 differ only by the listed literal edits. Every test that anchors on a
changed string or comment was found by the anchor sweep (**Special
cases** 9) and run above. CI runs the full farm.

## Acceptance notes

1. **The remainder rides this card, unchanged (seat amendment
`5860244997`, Q2 = A).** Of the 53 numbers PR objectui#10875's
Acceptance notes 1 lists, this PR carries the first 30 in ascending
order, plus the 11 C0 404s of its Acceptance notes 2. The other **23**
are: objectstack-ai#8072 objectstack-ai#8127 objectstack-ai#8137 objectstack-ai#8204 objectstack-ai#8229 objectstack-ai#8248 objectstack-ai#8307 objectstack-ai#8408 objectstack-ai#9231 objectstack-ai#9241 objectstack-ai#9244
objectstack-ai#9365 objectstack-ai#9373 objectstack-ai#9375 objectstack-ai#9542 objectstack-ai#9553 objectstack-ai#9585 objectstack-ai#10117 objectstack-ai#10119 objectstack-ai#10120 objectstack-ai#10129 objectstack-ai#10132
#14026. None of them occurs in any hunk of this diff.
2. **The C0 bare numbers above 10900, for later batches** (sites at
`b2683a2c0`, lines / files). Each answers 404 as an objectui issue, and
none is an objectui citation, so the fix is a re-qualification to
objectstack, not a landing sha:
- written with a word that names the repository but not in `repo#N` form
("objectstack PR #N", "upstream #N", "framework #N", "spec #N"): objectstack-ai#11289
(2 / 2), objectstack-ai#11662 (1 / 1), #12616 (1 / 1), #12718 (1 / 1), #13033 (1 / 1),
#13413 (1 / 1), #13733 (1 / 1), #13906 (6 / 3), #14274 (1 / 1), #15469
(2 / 2), #20160 (2 / 2);
- written bare: #13337 (1 / 1), #13632 (3 / 3), #13855 (1 / 1), #14945
(1 / 1), #15948 (1 / 1), #17493 (1 / 1), #20051 (1 / 1);
   - #13033 and #13413 answer 404 in objectstack as well;
- not citations: #334155, #336699, #475569, #737373, #953800 (CSS
colours).
3. **A live-but-wrong bare number beside a re-qualified one.**
`rowCrudAffordances.ts` cites "(`/me/permissions` `apiOperations`,
objectstack-ai#3391)" at the two sites that now read `[objectstack#3720]`.
objectui#3391 answers 200 with an unrelated subject (a record-header api
action placeholder), while objectstack#3391 is the apiMethods whitelist
contract card that objectstack#3720's own title names. It is not a 404,
so it is outside the family pin and was not touched. Carrier: none.
4. **Published or shipped text outside the two in-scope classes still
names some of these numbers** (counts in **Census**). It is untouched by
scope, as another class the triage split did not scope here: published
`CHANGELOG.md` history (never re-addressed), `content/docs` (for example
the chatbot and layout guides), `apps/console` (the approvals inbox and
the agent-connect section), `eslint-rules` and `eslint.config.js`, three
`skills/objectui` files and `AGENTS.md` (governed). Carrier: none.
5. **Filenames are not citations.** Pending changeset and test FILENAMES
carry several of these numbers. They stay, as in PRs objectui#10707,
objectstack-ai#10797, objectstack-ai#10854, objectstack-ai#10869 and objectstack-ai#10875.
6. **Test comments left as they are.**
`chatbot-dark-keys-retired-7703.test.ts`'s test names and
`@ts-expect-error` comments still read "RETIRED (objectui#7703)", and
`clientValidation.authorShapeRuling-10150.test.ts`'s header, title and
message still name objectui#7612. Only anchors moved (ruling
`5861900779`).

---
_Generated by [Claude
Code](https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
…tchet counts the namespaced half (objectui#10872, batch 1) (objectstack-ai#10888)

Part of objectstack-ai#10872
Clause-②: yes

Batch 1 of the namespaced slice of the objectui#10859 family. **20 of
the 28 namespaced ADR-0080 public blocks** now validate — every one
whose props `@objectstack/spec` declares as a `ComponentPropsMap` row,
less `element:number`. The other 8 are held on measured readings, and
the seat's ruling on objectui#10872 keeps them held (Acceptance notes).
Measured split posted on the card before this PR: comment 5862647089.
Head `0212421eda`: fix round 1 (`42f1cd1d2c`) fixed the two CI shards
red at `4679a3120` (block-config node oracle, handler-key alias census)
and swept the repo for suites that introspect the arm list; fix round 2
(`0212421eda`) is text only — the contract review's wording corrections
to the changeset, the module docblock and one checker comment.

## What changed

- **`packages/types/src/zod/public-blocks.zod.ts` (new).** One arm per
armed block — `PageHeaderBlockSchema`, `PageTabsBlockSchema`,
`PageCardBlockSchema`, `PageAccordionBlockSchema`,
`PageSectionBlockSchema`, `PageFooterBlockSchema`,
`PageSidebarBlockSchema`, `RecordDetailsBlockSchema`,
`RecordHighlightsBlockSchema`, `RecordRelatedListBlockSchema`,
`RecordPathBlockSchema`, `RecordActivityBlockSchema`,
`RecordDiscussionBlockSchema`, `RecordHistoryBlockSchema`,
`RecordQuickActionsBlockSchema`, `RecordReferenceRailBlockSchema`,
`RecordAlertBlockSchema`, `ElementTextBlockSchema`,
`ElementButtonBlockSchema`, `ElementDividerBlockSchema` — and the
category union `PublicBlockComponentSchema`.
- **`packages/types/src/zod/index.zod.ts`.** The 21 names on the `./zod`
barrel; `PublicBlockComponentSchema` joins `AnyComponentSchema`'s arm
list.
- **`packages/types/src/zod/README.md`.** An "ADR-0080 Public Blocks
(20)" group.
- **Pins:**
`packages/types/src/__tests__/public-block-arms-10872.test.ts` (new),
and the objectui#10859 ratchet file
`packages/cli/src/__tests__/registered-types-validate-ratchet-10859.test.ts`
extended with the namespaced count and an in-process `objectui validate`
run.
- **Ledgers:** 21 `EXCLUSIONS` rows in `zod-mirror-parity.test.ts`; the
17 `ComponentPropsMap` row symbols the arms read added to
`imported-defaults-8317.test.ts`'s `IMPORTED` differential.
- **Fix round 1 (`42f1cd1d2c`):**
`block-config-schema-parity-8216.test.ts` (app-shell previews) — the
NODE oracle judges an arm whose props live in a `properties` object
member AT that member, derived from the arm (`propsLevelOf`), plus a
control row; `check-handler-key-read-sites.test.ts` — `page:tabs` leaves
the objectui#9573 alias population and its `onTabChange` read is pinned
as judged on its own arm; `check-handler-key-read-sites.mjs` — two prose
paragraphs that named `page:tabs` as an alias, amended.
- **Fix round 2 (`0212421eda`, text only):** the changeset's lead
sentence names twenty blocks; its flat-props bullet separates keys
`BaseSchema` does not declare from the ones it does; the `action:*`
reason in the changeset and in `public-blocks.zod.ts`'s docblock is the
seat's (no spec row yet, measured upstream on
objectstack-ai/objectstack#20371), and the docblock's flat-key paragraph
carries the same `BaseSchema` distinction;
`check-handler-key-read-sites.mjs`'s amendment points at boundary 3.
- **Changesets:** `.changeset/10872-public-block-zod-arms.md`
(`'@object-ui/types': minor`, the `Clause-②: yes` line, no BREAKING
banner); one dated note on
`.changeset/8344-node-recursion-point-redirect.md` (objectui#10533 form,
frontmatter byte-identical) — see Acceptance notes.

## M1 — reproduced on `main` (`c6678b1bd`)

- `safeValidateSchema({ type: KEY })` over the 419 namespaced keys of
`KNOWN_SCHEMA_TYPES`: **418 refused** with `invalid_union` at `type`;
`ui:calendar` accepted. Bare: 76 (the objectui#10859 pin).
- `objectui validate` (the CLI's `validate`, in-process on source) on a
page whose children are `page:header` (`properties.title`),
`action:button` (`label`, `actionType: "url"`, `target`) and
`record:details`: exit 1, `Schema validation failed`, one refusal per
block — `No arm accepts type "page:header" at children → 0.`, `…
"action:button" at children → 1.`, `… "record:details" at children → 2.`
- At head the same page fails at `action:button` alone; the page of
`page:header` + `record:details` + `element:text` — batch 1's page pin
as the seat's ruling amends it — prints `✓ Schema is valid!`, `Type:
page`, exit 0.

### The α / β split (419 namespaced keys)

Twin = text after the first colon. "Same registration" = the namespaced
key and its bare twin come out of the same registration call (read from
`deriveRegistryKeys`, the derivation `KNOWN_SCHEMA_TYPES` is generated
from).

| class | keys | refused at `type` |
|---|---|---|
| (α) same registration, bare twin validates | 149 | 149 |
| (α′) same registration, bare twin itself refused — objectui#10859's
bare half | 77 | 77 |
| (β) bare spelling belongs to a DIFFERENT registration | 44 | 43
(`ui:calendar` accepted) |
| (β) no bare spelling | 149 | 149 |

By spelling alone α would read 189; 40 of those are false twins, the
triage's own example among them (`view:grid` is plugin-grid's
registration, bare `grid` the layout primitive). All 28 public blocks
are β (15 with no bare spelling, 13 whose bare spelling is another
component).

## M2 — the 28 public blocks, by declaration source

| source | count | blocks |
|---|---|---|
| spec `ComponentPropsMap` row (installed 17.4.0) | 21 | `page:header`
`page:tabs` `page:card` `page:accordion` `page:section` `page:footer`
`page:sidebar` · `record:details` `record:highlights`
`record:related_list` `record:path` `record:activity`
`record:discussion` `record:history` `record:quick_actions`
`record:reference_rail` `record:alert` · `element:text` `element:number`
`element:button` `element:divider` |
| no spec row; registration `inputs` only | 3 | `record:line_items`
`element:definition-list` `element:repeater` |
| no spec row; TS schema types in `@object-ui/components` + registration
`inputs` | 4 | `action:button` `action:group` `action:menu`
`action:icon` |
| no declaration anywhere | 0 | — |
| bare-key twin arm | 0 | — |

Disagreements, spec wins: `element:definition-list` / `element:repeater`
are in the spec's reserved `element:` namespace but not its vocabulary
(`isKnownComponentType` false, `hasReservedComponentNamespace` true), so
the spec refuses the TYPE at author time; `record:line_items` is in the
spec's `STRING_ARM_REGISTERED_TYPES` ledger as a row still to be
measured.

## M3 — homes, Serial

New module `public-blocks.zod.ts` (no open PR can hold it) plus
`index.zod.ts`. ⛔ Not `blocks.zod.ts`: it is the objectui#4895
tombstone, pinned empty by `block-family-retired-4895.test.ts`. Re-read
at fix round 2. `origin/main` is `45dbf1cbae`: since `b2683a2c0e` it has
landed objectui#10714, objectstack-ai#10780 and objectstack-ai#10852. None of them touches a file of
this diff, a `type` literal, `BaseSchema` (`base.zod.ts` / `base.ts`),
`known-schema-types.ts` or `public-blocks.ts` — objectstack-ai#10714's zod edits are
the list-item and tabs-item `body` refusals in `data-display.zod.ts` /
`layout.zod.ts`. `git merge-tree --write-tree` of this branch
(`0212421eda`) against `45dbf1cbae` is clean (exit 0), so main was not
merged in. CI on `42f1cd1d2c` ran on a merge ref over `b2683a2c0e` —
every workflow run for that head reports PR base `b2683a2c0e`, all 26
completed `success` or `skipped`. Open PRs on these surfaces:
objectui#10890 edits `zod-mirror-parity.test.ts` in a different hunk
(merge-tree of the two heads, `0212421eda` against `f6891cde45`, exit
0); objectui#10889 edits other `packages/types` files and none of this
diff's; objectui#10892 (a citation sweep, 106 files) edits
`.changeset/8344-node-recursion-point-redirect.md`, which this diff also
touches, in a different hunk (merge-tree against `afed8aa6b4`, exit 0).
The other open PRs (objectui#10891, objectstack-ai#10886, objectstack-ai#10884, objectstack-ai#10777, objectstack-ai#10278)
touch no file of this diff.

## M4 — the strict face

Confirmed derived, no edit: `StrictAnyComponentSchema` walks
`AnyComponentSchema`, and the pins read it accepting every minimal and
every spec-valid document of the 20, and refusing an undeclared node key
by name (`unrecognized_keys`).

## The arms

- **Where the props live.** A `ComponentPropsMap` row is the spec's
declaration of the props BAG — `PageComponentSchema.properties` — so
each arm is `BaseSchema` + the `type` literal + `properties`, whose
value IS the row, by reference through `stripImportedDefaults`
(objectui#8317). Members, value types, `.describe()` text, strictness,
refinements and the spec's retired-key tombstones all arrive from the
spec; nothing is restated. The bag is what every one of these renderers
reads (`SchemaRenderer` hoists it onto the node; the `element:*`
renderers read nothing else), and what the synthesizer and page designer
write.
- **Optional bag.** As on `PageComponentSchema`; the spec's props gate
(`validateComponentProps`) judges a bag only when present, and so do the
arms.
- **One restatement.** `element:divider`'s row is
`emptyProps('element:divider')`, exported under no name, so it cannot
cross the objectui#8317 boundary (its census admits only a spec binding
passed straight to `stripImportedDefaults`). The arm restates it as a
member-less strict object; the pin holds its key set and accept set
equal to the row's.
- **One `on*` member.** `page:tabs`' renderer calls a host-injected
`onTabChange` (app-shell's `withPageTabsUrlSync`), so
`check:handler-key-reads` required it declared:
`handlerKeyRefusal('onTabChange', 'runtime-slot', …)`. No other arm's
renderer reads an `on*` key off the node (the gate: 128 arms, 50 reads,
all judged).
- **Flat keys** are not judged against the block's row. A key
`BaseSchema` does not declare (`{ "type": "record:details", "columns":
"2" }`) passes the tolerant face, as an undeclared key does on any arm,
and the strict face refuses it. A key the base does declare is judged by
the base instead: `children` (`page:card`, `page:section`,
`page:footer`, `page:sidebar`), `label` and `disabled`
(`element:button`) and `visible` (`record:alert`) against the base's own
types, and `body` (`page:card`, `record:alert`) is refused on both faces
by the base's objectui#6771 retirement — a `record:alert` banner's
`body` goes inside `properties`. Measured by probe on source: flat
`columns` tolerant-accepted and strict-refused; flat `body` on
`record:alert` and `page:card` refused on both; flat `label: 5` and
`visible: 5` refused on both; the same `body` inside `record:alert`'s
`properties` accepted on both.

## Census rows

`zod-mirror-parity.test.ts` `EXCLUSIONS`, 21 rows: 19 "spec-owned BY
REFERENCE" (`page:tabs`' names its `onTabChange` refusal),
`ElementDividerBlockSchema` "spec-owned, RESTATED" pointing at its pin,
`PublicBlockComponentSchema` "a union OVER the public-block arms". No
`MIRRORS` pair: no TS declaration in `@object-ui/types` restates any of
these nodes. `record-components.ts`'s `Record*ComponentProps` restate
six of these bags (not the nodes) on the TS face and are pinned to the
spec by their own tests, not to these arms. One measured difference:
`RecordRelatedListComponentProps.title` is `string` where the row takes
`I18nLabel` (the row accepts `{ en: 'Tasks', 'zh-CN': '任务' }` as
`title`). No `KnownDrift` or other ledger row moved.

## Pins and ablation

`public-block-arms-10872.test.ts` — per armed block: minimal document
and a spec-valid bag accepted by `safeValidateSchema` and by
`StrictAnyComponentSchema` (each fixture first parsed by the spec's own
row); bag key set equal to the row's, and the bag IS the row object
wherever the boundary strips nothing. Plus: an undeclared member inside
the bag refused on the TOLERANT face by name; an undeclared node key
refused on the STRICT face by name (control: tolerant passes it); a
spec-refused value refused at `properties.variant`; a spec-retired key
(`page:card` `actions`) refused at the member; `onTabChange` refused
`custom`; a nested block judged by its own arm; `element:divider` pinned
to its row both ways.

| run | result |
|---|---|
| base — detached worktree at `c6678b1bd`, the two pin files copied in |
`Tests 94 failed / 6 passed (100)`. The 6 green: the four pre-existing
bare-ratchet rows, the objectui#10859 README row, the namespaced
non-vacuity control |
| head `4679a3120` | `100 passed` — both files, inside the full `5967
passed` run under Gates |
| ablation: `PublicBlockComponentSchema` dropped from
`AnyComponentSchema`'s list | `Tests 51 failed / 49 passed (100)` |

Ablation, observed direction: red — all 40 acceptance rows, both page
documents, the namespaced ratchet (`refuses 418 … the pin is 398.` + the
MORE-than-the-pin guidance), the armed-list row, and every refusal row
(with no arm they refuse at `type`, not by name). Green — the 43 rows
that read the arms directly off `PublicBlockComponentSchema` (literal
list, barrel, key sets, identity, divider) plus the 6 above; those rows
are introspection, not validation. Mutation and restore through
objectstack's `scripts/ablation-replace.mjs` in WRAP mode inside a
script with its own EXIT/INT/TERM trap: anchor x1 → x0, blob
`0a4466287501` → `c53fe2dc222a`, restored to the HEAD blob with `git
diff HEAD` empty. Vitest aliases `@object-ui/types/zod` to `src/`, so no
build sits between mutation and reading.

## The ratchet

Same file, same style, a second pin: `NAMESPACED_REFUSED_AT_TYPE = 398`
(418 on `main`, less the 20), lower-it / never-raise messages, refused
keys printed either way; non-vacuity (population over 400, `ui:calendar`
lit, an invented key refused); a row naming the 20 armed keys. Counted
apart from the bare pin because it is another card's batches. **The bare
pin stays 76**: no bare literal was added. The file also runs `objectui
validate` on the three-block page (exit 0, `Schema is valid`) and on the
same page with an invented bag key (exit 1, the key named).

## Gates

**Fix round 2, at head `0212421eda` (text only), under the verify
lock:** the tests that pin the changed texts —
`check-changeset-claims.test.ts`, `changeset-polarity-census.test.ts`,
`render-changeset-claims-comment.test.ts`,
`markdown-test-inputs.test.ts`, `check-handler-key-read-sites.test.ts`,
`public-block-arms-10872.test.ts`, `zod-mirror-parity.test.ts` and the
ratchet file → `Test Files 8 passed (8)`, `Tests 421 passed (421)`. No
test or script quotes the rewritten prose of `public-blocks.zod.ts` (a
`git grep` of that file name in tests finds the census reason keys and
one header path only). exit 0: `check-changeset-presence` (`8 source
file(s) of 3 released package(s) changed, and this change declares 1
changeset(s)`), `pnpm changeset:check` (`No changeset declares a major
bump`), `check-changeset-overwrite` (report-only: the 8344 note,
frontmatter unchanged), `check:changeset-claims` (report-only;
self-contradiction corpus 2 bodies, `✅ Every package declared … is
either not negated`; 20 paragraphs named, the one new name —
`9447-detail-view-handler-key-refusal.md`, on
`check-handler-key-read-sites.mjs` — describes the gate's transitive hop
and is untouched by a comment edit), `check:pending-changeset-literals`,
`census:changeset-polarity`, `markdown-test-inputs.mjs --audit` (`105
candidate test files, all adjudicated`), `check:handler-key-reads` (`OK
128 arm(s)`), `check:control-bytes`, `check:new-line-citations` (`0 new
citation(s)`).

At head `42f1cd1d2c` (fix round 1), all under the verify lock where they
run tests or builds:

- `pnpm exec vitest run packages/types/ packages/cli/
scripts/__tests__/markdown-test-inputs.test.ts` plus the checker's other
test readers (`detail-handler-slots-7804`,
`detail-view-handler-slots-9447`, `handlerKeyDispositionsMeasured-7804`,
`check-component-surface-parity`) → `Test Files 285 passed (285)`,
`Tests 6049 passed (6049)` — **0 failed**.
- **CI shard 6/8 — `block-config-schema-parity-8216.test.ts`.** Red at
`4679a3120`, reproduced locally by running that revision's copy of the
file against the arms: "unledgered BLOCK_CONFIG field names refused by
their schema", **29 rows**, all `@node` (28 MISSING such as
`page:header::title@node`, `element:text::content@node`,
`record:related_list::objectName@node`, and `record:alert::body@node`
RETIRED — `BaseSchema`'s node-level `body` refusal, not the row's
`body`). The oracle was reading the node's top level while the designer
edits `properties`. Green at the fix: the node oracle descends into a
`properties` member that unwraps to an object (derived from the arm, no
type list), and **0 rows stay red** — every designer field of the 20
blocks is a member of its spec row, so there is no real designer-to-row
mismatch to report. No ledger or exemption row was added. Non-vacuity at
the fix: `no node arms resolved` (more than 20) passes; the
`object-kanban` `groupField` RETIRED / `groupBy` accepted pair passes; a
new control row pins the descent both ways (`page:header` judged at its
bag: `title` accepted, `icon` RETIRED, `properties` not a listed key;
`object-kanban` still judged at its top level).
- **CI shard 2/8 — `check-handler-key-read-sites.test.ts`.** Red at
`4679a3120` (same probe method): the objectui#9344 leg (`page:tabs`
`unmirroredAlias` read `undefined`) and the objectui#9573 population (5
aliases, not 6). Green at the fix: the population is the five remaining
aliases, and the `page:tabs` row asserts the new state — no
`unmirroredAlias`, `schema` `PageTabsBlockSchema`, `declared` true,
disposition `runtime-slot`, the arm table's own `onTabChange` member
`runtime-slot`, no re-keying onto `tabs`. Ablation, WRAP mode inside a
script with its own EXIT/INT/TERM trap: `type: z.literal('page:tabs'),`
→ `…'page:tabs-ablated-10872'…` (anchor x1 → x0, blob `33a2d8a26927` →
`4f91055807d6`) → `Tests 2 failed | 47 passed (49)`: the new row
(`page:tabs is an armed type, not an alias: expected { bareType: 'tabs',
… } to be undefined`) and the population (6 vs 5); restored, blob ==
HEAD, `git diff HEAD` empty.
- **The sweep round 0 missed.** Every test file in the repo matching
`AnyComponentSchema`, `StrictAnyComponentSchema`, `safeValidateSchema`,
`nodeOracles`, `KNOWN_SCHEMA_TYPES`, `PUBLIC_BLOCKS`, `unmirroredAlias`,
`collectArms`, `getPublicConfigs` or `ComponentPropsMap`, plus every
test importing `@object-ui/types/zod` that walks `Object.values` /
`.entries` / `.keys` or `.options`: **204 files**. The 88 under
`packages/types` and `packages/cli` ran in the suite above; the other
**116** ran together → `Test Files 116 passed (116)`, `Tests 2044 passed
(2044)`. By package: `examples/schema-catalog` 15, `packages/components`
15, `packages/plugin-calendar` 10, `packages/plugin-kanban` 8,
`apps/console` 7 (incl. `registry-inputs-spec-parity`,
`public-contract`, `public-block-binding-reach`,
`record-block-record-reach`, `html-tier-manifest`), `scripts/__tests__`
6, `packages/app-shell` 8 (incl. the four previews suites
`block-config-schema-parity-8216`, `block-config`,
`definition-list-item-keys-8279`, `page-tabs-item-value-8278`),
`packages/plugin-view` 5, `packages/plugin-grid` 5,
`packages/plugin-map` 4, `packages/plugin-list` 4,
`packages/plugin-detail` 7, `packages/core` 6, `packages/plugin-tree` 3,
`packages/plugin-form` 4, `packages/plugin-dashboard` 3,
`packages/sdui-parser` 2, and one each in `test-support`,
`plugin-gantt`, `plugin-chatbot`, `plugin-charts`. Every one passed; the
per-file list is on the card's report.
- `pnpm --filter @object-ui/types type-check` (all three programs) →
exit 0 and `pnpm --filter @object-ui/cli type-check` → exit 0 (both at
`4679a3120`; round 1 changed neither package); `pnpm type-check:scripts`
→ exit 0 at `42f1cd1d2c`.
- exit 0 at `42f1cd1d2c`: `check:handler-key-reads` ("OK 128 arm(s)"),
`check:control-bytes`, `check:new-line-citations` ("0 new citation(s)"),
`check-changeset-presence`, `check:pending-changeset-literals`,
`check:test-path-roots`. Earlier at `4679a3120` and untouched by round
1: `pnpm changeset:check`, `check-changeset-overwrite` (report-only; the
8344 note), `check:changeset-claims` (report-only; read below),
`check:spec-symbols`, `check-type-check-coverage`,
`check:unreferenced-sources`, `check:esm-specifiers`, `check:doc-types`,
`check:prompt-keys`, `check:registry-bare-names`,
`check:side-effects-array`, `check:self-import`;
`markdown-test-inputs.mjs --audit` → "105 candidate test files, all
adjudicated".
- eslint, a proven narrowing, `--no-inline-config --format json`: round
0's six `.ts` files → 0 errors, 2 warnings; round 1's three files → 0
errors, 4 warnings. All six warnings are `no-explicit-any` on lines this
diff did not write, and the same files at `c6678b1bd` / `4679a3120`
carry the same counts. Population: `@object-ui/types` lints `.`,
`@object-ui/cli` lints `src`, `@object-ui/app-shell` and the root
`scripts/` sit under the root `eslint .`; `eslint.config.js` sets no
`projectService` / `parserOptions.project`, so no type-aware rule can
move an untouched file's verdict. The full `pnpm lint` is CI's.
- NOT MEASURED: `@object-ui/app-shell` `type-check` (its
`tsconfig.test.json` leg covers the edited previews test) — reason: it
needs the app-shell dependency closure built, the cost the order's scope
does not name; the edit is two `unknown`-typed helpers. CI's `Type
Check` measures it. Also `check:readme-exports`, `check:doc-snippets`,
`check:doc-examples` — they need the full package build; no README or
content doc changed.

## Acceptance notes

- **Held — with a declaration, not armed:**
- `element:number` — the row requires `object`, which the spec's props
gate waives when the node binds through `dataSource.object`
(`DATASOURCE_SUPPLIED_PROP`, `@objectstack/lint`); the row alone would
refuse that spec-valid node. Arming needs that carve-out, and
`dataSource` itself is an envelope key no arm declares yet.
- `record:line_items` — no spec row on purpose (the
`STRING_ARM_REGISTERED_TYPES` ledger: to be measured from the renderer's
read points). Its five registration `inputs` undercount real authoring:
the showcase's project-detail page authors `title` in the bag.
- `element:definition-list`, `element:repeater` — the spec's closed
`element:` vocabulary does not declare them, so the spec refuses the
type at author time while `PUBLIC_BLOCKS` curates both. Per the seat's
ruling on objectui#10872 (option A), they get spec rows upstream first
and stay held here until then; arming them now would make `objectui
validate` accept what `os validate` refuses.
- `action:button`, `action:group`, `action:menu`, `action:icon` — no
spec row. Per the seat's ruling on objectui#10872 they stay held (option
D now, route C next): no spec row yet; one is being measured from the
renderers' read points upstream (objectstack-ai/objectstack#20371). The
ruling's reading: `action-button.tsx` reads `schema.name ??
schema.label`, so `name` is optional where the renderer reads it, and
AGENTS.md objectstack-ai#4's taught `action:button` carries none; the spec `Action` is
a different declaration (object-metadata actions keyed by `name`,
recorded as structurally unrelated in `ui-action.ts`), so restating it
would refuse the governed instructions file's canonical example.
- **The page pin:** the seat's ruling on objectui#10872 amends batch 1's
page pin to `page:header` + `record:details` + `element:text`, which is
what ships; the `action:button` page pin moves to this card's `action:*`
batch.
- **No declaration anywhere:** none of the 28.
- **α batch plan (not this batch):** pair by registration, never
spelling — 149 α keys reuse their bare arm under a second literal, one
schema, two literals, ⛔ no copy; the 77 α′ keys follow as objectui#10859
arms each bare key. The 40 false spelling-twins (`view:grid` first among
them) are β and need their own declarations.
- **Flat props channel:** whether `page:` / `record:` blocks may also
author props flat on the node (their renderers read the hoisted keys) is
not decided here; declaring it later is additive, and never for
`element:*`, whose renderers do not read a flat key.
- **Envelope keys** the spec's `PageComponentSchema` declares beside
`properties` and `BaseSchema` does not — `events`, `responsiveStyles`,
`dataSource`, `aria`, `visibility` (the spec's deprecated alias of
`visibleWhen`) and `responsive` (which the spec itself retired) — are
not on these arms: tolerant passes them, strict refuses them. No taught
public-block document carries one (the 14 teaching sites in `content`,
`skills`, `docs`, `examples` and the READMEs that name an armed type,
read 12 lines past each), and the strict face has no in-repo consumer,
so nothing taught is newly refused.
- **Nested blocks inside a bag** (`properties.children`,
`items[].children` on `page:tabs` / `page:accordion`) are
`z.array(z.unknown())` in the spec rows, so a child block there is not
judged by this validator — measured: an invented child type with an
invented key inside `page:section`'s `properties.children` parses. That
is correct by reference: the row is the spec's, and the spec does not
judge it either.
- **Content channels:** objectui#9256's per-channel narrowing is not
applied to the new arms (not measured for these renderers in this
batch).
- **Pending changesets read.** `check:changeset-claims` named the
entries that cite a file this diff touches. Falsified by this change:
`8344-node-recursion-point-redirect.md` — its objectui#10859 dated note
says both module counts are "one higher"; with `public-blocks.zod.ts`
they are two higher. Dated note added, frontmatter byte-identical. Read
and left, true of this diff: `5903`, `5928`, `6051`, `6150`, `6175`,
`6639`, `7344`, `7654`, `7694`, `7804`, `8338`, `8498` and `8784` (their
objectui#10859 notes describe that change's addition; no `type` literal
collides), `8499` (every new arm is named on the barrel, so its figure
does not move), `8572`, `8598`, `9067`, `9511`, `9659`,
`object-view-unmirrored-keys-7779`, `10859-ai-node-zod-arms`.
`7917-export-breadcrumb-object-tree-zod-schemas.md` ("107 node component
types") was already stale before this diff. A scan of pending entries
naming a public-block literal beside `safeValidateSchema` / `objectui
validate` / `AnyComponentSchema` found none claiming a refusal.

- **Fix round 1's census follows the change, anchor only** (ruling
5861900779 route a): the objectui#9573 alias population loses
`page:tabs` because it gained an arm, not because a read was exempted,
and `check-handler-key-read-sites.mjs`'s two prose paragraphs that used
`page:tabs` as the alias example carry an objectui#10872 amendment.

Attribution: the batch-1 dev of `domain:ui` seat 2, session
`https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN`.

---
_Generated by [Claude
Code](https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
…bject views, detail-view, email, password) and repair the shorthand Omit erasure (objectui#9256) (objectstack-ai#10897)

Part of objectstack-ai#9256
Clause-②: yes

**Clause-② `yes`, as the claim declared:** twelve published arms stop
accepting an authored `children`, and two published TypeScript faces
stop erasing the members they inherit. Both narrow an accept set; a
contract review is owed before landing.

The E3 residual that the maintainer's ruling on objectui#8284 (comment
`5861449497`, Q2 A) moved onto this card: keys with a published arm
whose renderer reads **neither** content channel. It also carries the
ruling's execution parameter 3, the `InputShorthandSchema` /
`UiCalendarSchema` TypeScript `Omit` erasure. objectui#9256 stays open
for the rest of family D.

## What changed

- **Narrowed (12 keys):** `object-grid`, `object-form`, `object-kanban`,
`object-map`, `object-tree`, `object-view`, `object-gantt`,
`object-calendar`, `object-chart`, `detail-view`, `email`, `password`.
- TypeScript face: `body?: never` and `children?: never`, each with a
docblock naming what the node renders instead.
- zod mirror: `body` and `children` are two `retirementTombstone`
members fed ONE string from the new `neitherContentChannelGuidance`
builder in `zod/tombstone.zod.ts`. Both stay MEMBERS, so
`zod-mirror-parity`'s key sets stay equal.
- `body` was already refused on these arms (by `BaseSchema`,
objectui#6771), except on the `email` / `password` TypeScript face,
which the `Omit` erased. It is restated because `BaseSchema`'s refusal
names `children` as the remedy, and `children` is dead here too.
- **`Omit` erasure repaired:** `InputShorthandSchema` and
`UiCalendarSchema` now inherit through `OmitDeclared` (`form.ts`, not
exported). It is a key-remapping mapped type: it drops exactly the named
keys and keeps every other named member and the index signature.
- `ui:calendar` gains no refusal of its own. Its face now re-exposes
what `CalendarSchema` declares, so the published TypeScript face refuses
`body`, `children` and a `mode` outside the declared enum, which the
erased face accepted. Its mirror (a `.extend()` of that one) already
refused them.
- I measured the `ui:calendar` renderer anyway
(`renderers/form/calendar.tsx`, typed `CalendarSchema`). It reads
`defaultValue`, `mode` and `value`, and no channel.
- **Held (3 keys): `br`, `hr`, `img`.** See "Held, and the question it
leaves" below.

## Per-key measurement, taken on `origin/main` `0c50f1863` before any
edit

**Instrument.** The TypeScript compiler API, ⛔ not grep. It builds one
program per workspace package from that package's own `tsconfig.json`:
40 programs over 1573 source files, on a BUILT tree (turbo build 42/42
first), with `@object-ui/types` resolving to `packages/types/dist`. It
files every `.body` / `.children` read (property access, string element
access, object destructuring) under the declared type of its receiver.

- 547 channel reads were filed.
- **Lit controls fire:** `ButtonSchema` 1, `DivSchema` 1, `CardSchema`
1, `ContainerSchema` 1 (each `children`).
- **Receiver-reachability control:** the same program sees every
narrowed declaration as a receiver of the keys its renderer reads. The
zeros below are readings, not blindness.

| key | registration(s) claiming it | registration hop → reader, and the
type it reads the node as | `body` / `children` reads under that type |
receiver reads (reachability) | arm, TS face: before → after | arm, zod
face: before → after | disposition |
|---|---|---|---|---|---|---|---|
| `object-grid` | `plugin-grid:object-grid` (+ console `registerLazy` of
the same plugin) | `ObjectGridRenderer` (`any`) → `ObjectGrid` ·
`ObjectGridSchema` | 0 / 0 | 158 reads, 38 keys | body refused
(BaseSchema), children accepted → both refused | same → both refused by
name | narrowed |
| `object-form` | `plugin-form:object-form` | `ObjectFormRenderer`
(`any`) → `ObjectForm` · `ObjectFormSchema` | 0 / 0 | 200 / 44 | same |
same | narrowed |
| `object-kanban` | `plugin-kanban:object-kanban` (+ lazy) |
`ObjectKanbanRenderer` (`any`) → `ObjectKanban` · `ObjectKanbanSchema` |
0 / 0 | 67 / 13 | same | same | narrowed |
| `object-map` | `plugin-map:object-map` (+ lazy) | `ObjectMapRenderer`
(`any`) → `ObjectMap` · `ObjectMapSchema` | 0 / 0 | 30 / 16 | same |
same | narrowed |
| `object-tree` | `plugin-tree:object-tree` (+ lazy) |
`ObjectTreeRenderer` (`any`) → `ObjectTree` · `ObjectTreeSchema` | 0 / 0
| 13 / 8 | same | same | narrowed |
| `object-view` | `plugin-view:object-view` (and the `view` alias) |
`ObjectViewRenderer` (`any`) → `ObjectView` · `ObjectViewSchema` | 0 / 0
| 112 / 19 | same | same | narrowed |
| `object-gantt` | `plugin-gantt:object-gantt` (+ lazy) |
`ObjectGanttRenderer` (`any`) → `ObjectGantt` · `ObjectGanttSchema` | 0
/ 0 | 61 / 48 | same | same | narrowed |
| `object-calendar` | `plugin-calendar:object-calendar` (+ lazy) |
`ObjectCalendarRenderer` (`any`) → `ObjectCalendar` ·
`ObjectCalendarSchema` | 0 / 0 | 40 / 12 | same | same | narrowed |
| `object-chart` | `plugin-charts:object-chart` (+ lazy) |
`ObjectChartBlock` (`any`) → `ObjectChart` · `ObjectChartSchema` | 0 / 0
| 123 / 15 | same | same | narrowed |
| `detail-view` | `plugin-detail:detail-view` | `DetailViewRenderer`
(`any`) → `DetailView` · `DetailViewSchema` | 0 / 0 | 195 / 31 | same |
same | narrowed |
| `email` | `ui:email` (`field:email` stands down) | `any` wrapper
spreads the node into `InputRenderer` · `InputSchema` | 0 / 0 | 23 / 17
| BOTH accepted (erased `Omit`) → both refused | body refused, children
accepted → both refused by name | narrowed |
| `password` | `ui:password` (`field:password` stands down) | same as
`email` | 0 / 0 | 23 / 17 | same as `email` | same as `email` | narrowed
|
| `br` | `ui:br` | html-elements factory (`any`), void tag: passes
`undefined` as the child | reads neither (the one `schema?.children`
read is behind the void-tag guard) | n/a (`any`) | body refused,
children accepted (family arm) → unchanged | same → unchanged | **held**
|
| `hr` | `ui:hr` | same | same | n/a | same | same | **held** |
| `img` | `ui:img` | same | same | n/a | same | same | **held** |

The residual count stays 15: no key reads `children` today. 12 are
narrowed and 3 are held.

**`any`-typed receivers, attributed rather than assumed.**

- Every hop above is `any`-typed. Each hop body was read: none contains
a channel read.
- The any-typed reads in the eleven hosting files and packages were
attributed one by one:
- React `children` destructured off internal components' own props
(`SchemaRenderer` passes no node child as a React child);
  - `document.body` and `err.body`;
  - `TreeNode.children` on the record hierarchy `ObjectTree` builds;
  - `FeedItem.body`;
  - `record:alert`'s own `props.body`, which is another registration.
- The html factory's `schema?.children` read sits behind `isVoid ?
undefined : …`.

**Delegation hosts, checked for pass-through.** Every `SchemaRenderer`
call in the ten object/detail packages was traced. The node it renders
is built one of three ways:

- key by key — `ObjectForm` / `ModalForm` / `DrawerForm` / `TabbedForm`
/ `SplitForm` / `WizardForm` all build their `form` node that way, and
the `form` renderer DOES draw `children`, which is why this was checked;
- from a sub-key — `DetailView`'s `header` / `footer` / tab `content`,
`ObjectChart`'s drill-down table;
- or from a generated `object-*` node — `ObjectView` assembles
`object-kanban`, `object-calendar`, `object-gallery`, `object-timeline`,
`object-gantt`, `object-map`, `object-tree` and `object-chart` from
named keys, and every one of those reads neither channel (for gallery
and timeline: `packages/plugin-list` and `packages/plugin-timeline` hold
no node-channel read).

No host spreads its own node into a node that reads `children`.

**Generic traversers** (`validateChildren` in core, `sdui-parser`
`validateTree`, the metadata-admin preview canvases, the vscode
validator) walk children to validate or preview. They render nothing and
are node-agnostic, as for every row family D already holds.

**objectui#9910 cross-check.** None of the 12 narrowed registrations
declares a `children` slot input: a grep over their registration files
finds 0 hits, and the lit control is `html-elements.tsx` / `div.tsx`,
which do.

**Producers.** `pnpm census:body-dialect --keys` over the 15, plus
`ui:calendar` and `input`, with `div` / `card` / `page` / `button` / `p`
/ `h1` as lit controls in the same pass: 8981 files, 2744 candidate
nodes.

- **0** nodes author `children` on any of the 15.
- 1 node authors `body` on `ui:calendar`: it was the family-D TRIPWIRE
test line (test bucket), moved by this change.
- The controls read `div` 170 / `card` 183 / `page` 47 `children`.
- Producers derived: none. Nothing was migrated.

## The `Omit` repair — measured on the BUILT d.ts

A probe program under `packages/components` resolved `@object-ui/types`
to `packages/types/dist/index.d.ts` (`--traceResolution`). It carried 15
`@ts-expect-error` pins (the 10 object/detail `children`, `email`
`children`, `password` `body`, `email` `label: 42`, and the
`ui:calendar` `children` and `mode: 'default'`), a lit control (`const
n: number = 'x'`), and 12 channel-free controls.

- **On this branch:** exit 0 — every directive used, every control
compiles, the lit control fires.
- **On the base dist:** exit 2 with 15 × TS2578, and the lit control is
still used.
- The emitted `dist/form.d.ts` declares `OmitDeclared` locally and both
heritage clauses use it.

## Red on base, then green — one-shot, the restore proven by state

The predictions were written to a file BEFORE the mutation. The mutation
was `git checkout 0c50f18 -- THE-SEVEN-DECLARATION-FILES`, with every
test file at HEAD. It was proven by 7/7 blob hashes equal to BASE plus
anchor counts (`neitherContentChannel` 0, `OmitDeclared` 0). The run sat
under `trap … EXIT INT TERM` with absolute paths.

| reader | mutated (BASE arms) | restored (HEAD `14fd13c51`) |
|---|---|---|
| vitest, the new pin file | **RED** exit 1 — 73 failed / 90 passed;
exactly the predicted 12 × 5 `children` families + 12 `body`-message
rows + 1 nested row | **GREEN** 163 / 163 |
| `tsc -p packages/types/tsconfig.test.json` | **RED** exit 2 — 18 ×
TS2578 and 3 × TS2322 (the `Eq` pins) in the new file, plus 3 in
`zod-mirror-parity` naming the two re-collapsed pairs and the stale
`KnownDrift` row; 0 elsewhere | **GREEN** exit 0 |
| built-d.ts probe (types rebuilt; dist markers 0 → 29 / 6) | **RED**
exit 2, 15 × TS2578 | **GREEN** exit 0 |

The restore was proven by 7/7 blob hashes equal to HEAD, `git diff HEAD`
empty and `git status` clean, both after the restore and at the end.

⚠️ **The `body` rows on the ten object/detail arms are GREEN on base, by
design:** `BaseSchema` already refused `body` there. Only their MESSAGE
changes, and the message rows are the ones that turn red. `?: never` is
erased before vitest runs, so the TypeScript half is read by `tsc`
alone. The vitest run cannot see it.

## Serial constraints

- **Mapped at branch time (`0c50f1863`).** PRs objectui#10714, objectstack-ai#10883
and objectstack-ai#10884 were open then; all three have since merged, and PR
objectui#10875 was the base.
- `layout.ts` (TabItem) and `layout.zod.ts` (TabItemSchema): untouched,
because `br` / `hr` / `img` are held and neither file is edited.
- `content-channel-family-d-9256.test.ts`: the only edits are forced
ones, clear of PR objectui#10714's hunks, which sit in the ListItem and
Tabs sections:
- The file's own `ui:calendar` TRIPWIRE said "when the collapse is
repaired this line goes RED … turn it back into a `@ts-expect-error`
pair". The repair fired it.
- The pair moved to the new file. That file lost the tripwire line, its
`uiCalendarBody` array entry (32 → 31) and the now-unused
`UiCalendarSchema` type import.
    - No pin was added there.
- `objectql.ts`: the insertion after `ObjectViewSchema`'s last member is
clear of PR objectui#10883's `NamedListView` docblock edit.
- `object-view-unmirrored-keys-7779.test.ts` (PR objectui#10884):
untouched.
- **Open PRs the branch-time map missed** (re-mapped by contract review
`5863631182`):
- PR objectui#10890 rewrote the same two `KnownDrift` header figures in
`zod-mirror-parity.test.ts` and landed first (`c30c8dd4c`). This branch
merged `main` at that commit (merge commit `31bb63bcb`, no rebase, no
force-push) and recounted the figures from the merged table; see "Merge
round" under Gates.
- PRs objectui#10888, objectstack-ai#10889 and objectstack-ai#10892 are textually disjoint, and all
three have since merged. `git merge-tree` of the merge head against
`main` at `6a7f24e92`, which carries all of them, reports no conflict.
- ⛔ `zod/index.zod.ts` and the `packages/cli` ratchet (objectui#10872)
are untouched.

## Collateral that the repair and narrowing made true or false — each
measured, none widened

- **`zod-mirror-parity.test.ts`:**
- The two `Omit` COLLAPSE rows (32 + 26 keys) leave
`MirroredUndeclared`. Its figures move from 4 / 60 to 2 / 2, and the
class note records the emptying.
- Restoring the named members NARROWED those faces: a collapsed face
accepted every key at `any`. So this is not the "declare the keys"
remedy that ledger refuses.
- The repaired shorthand face now carries `InputSchema`'s `onChange`
runtime slot. It enters `KnownDrift` as inherited drift: one entry, one
key. After the merge with PR objectui#10890 (which took
`ai.zod.ts#AIFormAssistSchema` out, one entry and two keys), the table
holds 49 entries / 88 keys, and both header sites say so.
- The union-arm non-vacuity leg had no multi-arm entry left in this
ledger. It now takes the same reader's arm-walking on `KnownDrift` and
re-arms here when a multi-arm entry lands.
- **`object-view-spec-parity.test.ts`:** `body` / `children` are
redeclared envelope keys on `object-view` now. They are triaged into
`SANCTIONED_LOCAL` as refusal-only surface, like `viewTabBar`, and the
declared-surface figure moves 24 → 26.
- **`CalendarSchema`'s texts on both faces said the TypeScript `Omit`
"collapses … and carries no member".** That became false; they are
corrected.
- **Two comments** that spelled the old plain-`Omit` heritage
(`index.ts`, `form-barrel-mirror-9406.test.ts`) are corrected.
- **`scripts/check-handler-key-read-sites.mjs`** caught my first
spelling. It was a computed spread (`...helper(…)`) that the gate cannot
resolve, so it stopped judging all eleven arms and its `detail-view`
`onTabChange` waiver read stale. The fix, the second commit, declares
both members explicitly. Now 49 / 49 reads are judged and 0 are left
unjudged.

## Held, and the question it leaves

`br` / `hr` / `img` read neither channel, but their only published arm
is `HtmlElementSchema`: ONE object over 38 tags, whose renderer draws
`children` for the other 35 (`p`, `h1`, …). Refusing the key on three
tags means splitting that family arm, or refining it per tag. Both
change a published contract's shape.

- The option the ruling adopted (option A in measurement `5859897857`)
reads: "leaving the three void tags to the tier because their arm is the
shared HtmlElementSchema family whose per-tag split was ruled out
(objectui#9067)".
- The parser tier already names them: the registration declares no
`children` slot on a void tag, so `validateTree` reports
`not-a-container`.
- A HELD control in the new pin file records the shared-arm fact, so the
hold is visible.

That the brief listed them among the keys to narrow is recorded as a
conflict in the report, ⛔ not resolved here by guessing.

## Gates (at HEAD `14fd13c51`, exit codes captured by
redirect-then-capture; heavy runs through the shared verify lock)

| gate | result |
|---|---|
| `@object-ui/types` build (tsc + vite + dist completeness) | exit 0 —
132 emitted files verified |
| `@object-ui/types` `type-check` (`tsc --noEmit` + examples +
`tsconfig.test.json`) | exit 0 |
| `pnpm exec vitest run packages/types/` (parity, node-slot, arm,
mirror, family-D and the new pins) | 259 files / 5670 tests, exit 0 |
| `pnpm exec vitest run scripts/` | 177 passed + 2 skipped of 179 files
/ 5317 tests, exit 0 |
| consumer type-check — every package naming a narrowed type (28
packages: the ten object plugins, app-shell, components, core, fields,
cli, react, runner, console, schema-catalog, vscode-extension …) | exit
0, 28 × `type-check: Done` |
| `pnpm --filter @object-ui/types lint` | 0 errors; 0 findings on any
added line (eslint JSON over the 13 touched files) |
| `check:control-bytes` · `check:new-line-citations` (0 new) ·
`changeset:check` · `check-changeset-presence` ·
`check:changeset-claims` · `check:pending-changeset-literals` | exit 0
each |
| `check:doc-types` · `check:doc-snippets` · `check:doc-examples` ·
`check:skill-examples` · `check:doc-fences` · `check:doc-example-ids`
(against the built types) | exit 0 each |
| `check:handler-key-reads` · `check:component-surface-parity` ·
`check:prompt-keys` · `check:registry-bare-names` · `check:spec-symbols`
· `check:readme-exports` · `pnpm check` (CLI self-check) | exit 0 each |
| governed guard `--test` over the 14 paths | NOT GOVERNED |

NOT MEASURED:

- `check:sdui-registration-pins` exits 2 without a console build. The
diff touches no registration and no `sideEffects` array. Left to CI.
- The 4-way `pnpm test` shards beyond `packages/types` and `scripts/`,
plus `test:dist` and E2E: left to CI.

### Merge round — head `31bb63bcb` (merges `main` at `c30c8dd4c`)

The one conflict was the two `KnownDrift` header figures in
`zod-mirror-parity.test.ts`. It was resolved by RECOUNTING the merged
table with the test's own AST reader, not by picking a side:

| tree | `KnownDrift` |
|---|---|
| base `0c50f1863` | 49 entries / 89 keys |
| PR objectui#10890 removes `ai.zod.ts#AIFormAssistSchema` | −1 entry /
−2 keys (48 / 87 on `c30c8dd4c`) |
| this branch adds `form.zod.ts#InputShorthandSchema` | +1 entry / +1
key (50 / 90 on `14fd13c51`) |
| merged `31bb63bcb` | **49 entries / 88 keys** |

The population bullet and the KNOWN_DRIFT restatement both read 49 / 88,
and the bullet keeps both histories in landing order. The other ledgers
read the same as before: `MirroredUndeclared` 2 / 2,
`UnmirroredDeclared` 12 / 84, `RuntimeOnlyDeclared` 3 / 7,
`WiderThanDeclared` 6 / 7. The merge head differs from `main` by exactly
this PR's 14 files.

| gate at `31bb63bcb` | result |
|---|---|
| `zod-mirror-parity` + `content-channel-e3-residual-9256` | 2 files /
200 tests, exit 0 |
| `@object-ui/types` build · `type-check` (incl. `tsconfig.test.json`) |
exit 0 · exit 0 |
| `pnpm exec vitest run packages/types/` (incl. PR objectui#10890's
`ai-zod-arms-10859` / `ai-node-faces-agree-10874` and the family-D file
PR objectui#10714 edited) | 261 files / 5696 tests, exit 0 |
| `pnpm exec vitest run scripts/` | 177 passed + 2 skipped of 179 files
/ 5322 tests, exit 0 |
| `check:control-bytes` · `check:new-line-citations` · `changeset:check`
· `check-changeset-presence` · `check:changeset-claims` ·
`check:pending-changeset-literals` · `check:handler-key-reads` | exit 0
each |

## Acceptance notes — out of scope, not fixed here

- **`input` reads neither channel as well.** `InputSchema`: 0 channel
reads, 17 keys read; the registration declares no `children` slot.
- It is this card's remaining family-D name, and it was not in this
claim.
- Its old hold-out ground ("a TS `Omit` carries nothing") is gone with
this repair: a tombstone on `InputSchema` now propagates on both faces.
  - Carrier: the next objectui#9256 slice.
- **`PartialSchema`'s docblock in `index.ts`** says every generic
re-spelling of its `Omit` collapses for the same `keyof T` reason.
`OmitDeclared` is a generic spelling that does not: this PR measures it
on two faces.
- Carrier: the next objectui#9256 slice, which touches the same `Omit`
subject. objectui#6397, that alias's card, is closed.
  - Noted, not filed.
- **The family-D test header** still records `InputSchema` as "family E,
frozen" and `DetailViewSchema` as unattributable. Both readings are
superseded by this change.
  - Carrier: the next objectui#9256 slice.
- Noted, not edited: PR objectui#10714 held that file while this PR was
written. It has since merged, so the next slice edits the header freely.

---
_Generated by [Claude
Code](https://claude.ai/code/session_01DuWo5bdP9SdVebamn99GGk)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
… dashboard filter's optionsFrom options re-read on the data-invalidation bus (objectui#10887) (objectstack-ai#10898)

Part of objectstack-ai#10887
Clause-②: no — members 1 and 2 add a re-read on the canonical bus to two
existing readers and change no declared contract. Member 3 stops at
`needs_decision` before any scope widening: this PR touches neither the
react-page scope nor its guide.

`Part of`, not a closing keyword: member 3 (the `kind: 'react'` page
scope) measured shipped readers and stops at `needs_decision`, as the
dispatch order rules for that branch. Members 1 and 2 land here, and the
card stays open for member 3.

## What changed

- **Member 1, `@object-ui/plugin-view`.** `ObjectView`'s non-grid fetch
effect (the one commented "Fetch data for non-grid view types") names
the `useDataInvalidation` nonce for `schema.objectName`. The hook sits
directly above that effect, and the nonce is the last entry of the
effect's dependency list. No second channel, no remount, no key change.
- **Member 2, `@object-ui/plugin-dashboard`.** `SelectFilter` in
`DashboardFilterBar` names the nonce for `optionsFrom.object` in the
options effect, on both reads (the dataset GROUP BY and the client-side
`find` fallback). The effect never resets `dynamicOptions` and never
writes the dashboard variable, so the options stay on screen until the
re-read answers and the selected value is kept.
- One `patch` changeset per package:
`10887-object-view-non-grid-bus-reader` and
`10887-dashboard-filter-options-bus-reader`.
- Not touched: `renderContent` (PR objectstack-ai#10884's region), `PageView` (PR
objectstack-ai#10777), the react-page scope and `react-pages` guide.

## H1 — reproduced on base `35d68c4ce5`, through the real
`SchemaRenderer`

Each block is rendered in the page-region shape (`{ type, properties }`
for `object-view`, a `dashboard` node with `globalFilters`) under
`SchemaRendererProvider`, over a data source that counts reads and
implements no `onMutation` (a page action over raw HTTP announces none).

| block | mount | after `'*'` | after an unrelated object | after its
own object |
| --- | --- | --- | --- | --- |
| `object-view` kanban, calendar, gallery, timeline, map, gantt — base |
1 `find` | 1 | 1 | 1 |
| same — head | 1 | 2 | 1 | 2 |
| `dashboard` `optionsFrom` filter, dataset read — base | 1
`queryDataset` | 1 | 1 | 1 |
| same, `find` fallback — base | 1 `find` | 1 | 1 | 1 |
| both reads — head | 1 | 2 | 1 | 2 |

Base failure text, quoted: `the rows never re-read after the bus
reported a change: expected "vi.fn()" to be called 2 times, but got 1
times` (every `object-view` type) and `the options never re-read after
the bus reported a change: expected "vi.fn()" to be called 2 times, but
got 1 times` (both option reads).

Lit controls, green on base and head in the same tree: a bare
`useDataInvalidation` reader moves to 1 on every event (asserted before
the count), a grid `object-view` re-reads once through `ObjectGrid`'s
own reader, and an `object-metric` block beside the dashboard re-reads
once through its own.

## H2 — seam

- **Member 1.** Hook: `useDataInvalidation(fetchDrawsView && dataSource
? schema.objectName || undefined : undefined)`, where `fetchDrawsView`
is "no host `renderListView`, and the view type is none of `grid`,
`tree`, `chart`". Effect: the non-grid fetch, dependency list extended
by `invalidationNonce`.
- `dataChangeMatches(change, schema.objectName)` with no record id: true
for `'*'`, and for any change whose `objectName` equals the view's,
record-scoped or not. False for another object.
- No object or no adapter: the hook does not subscribe, the nonce stays
0, and the effect's own early return reads nothing (pinned).
- `renderListView` and `grid` are excluded because `ListView` and
`ObjectGrid` read the bus themselves, and the effect body returns early
for both anyway.
- `tree` and `chart` (host-only, reachable only through a `views` prop)
are excluded because their renderers query for themselves and read the
bus themselves: `ObjectTree` runs its own query ahead of the rows handed
to it (objectui#10778) and re-queries whenever that array changes, and
`ObjectChart` never reads the rows. Subscribing there would add a read
per event beside theirs (ablation 3 shows `find` 1 → 2 for each). This
narrows the card's parenthetical list, which names `tree` among the
fetch-driven views: the fetch does not drive the tree's rows once a data
source is present.
- **Member 2.** Hook: `useDataInvalidation(from && dataSource &&
(queryDataset or find is a function) ? from.object || undefined :
undefined)`. Effect: the `optionsFrom` options effect, dependency list
extended by `invalidationNonce`.
- `dataChangeMatches(change, optionsFrom.object)`: true for `'*'` and
for any change to that object. False for another object.
- A filter with authored `options` and no `optionsFrom` does not
subscribe and reads nothing (pinned).

## H3 — member 3 census (the `kind: 'react'` page scope)

Enumerated from the committed trees, not a working tree: objectui
`origin/main` `35d68c4ce5` and objectstack `origin/main` `862b6ce869`,
over examples, docs, skills and templates.

- **Population:** `kind: 'react'` pages that ship: 3, all in
objectstack's `examples/app-showcase` (`showcase_crm_workbench`,
`showcase_renewals_pipeline`, `showcase_task_desk`). objectui ships none
in `examples/`; its react pages live in the `react-pages` guide and the
console's SDUI preview harnesses.
- **Shipped pages that read data through `useAdapter`: 2 of 3.**
- `showcase_crm_workbench`: the KPI strip reads
`adapter.find('showcase_project', { $top: 200 })` in an effect keyed on
the page's own `reloadKey`, bumped only by its own `ObjectForm`
`onSuccess`.
- `showcase_renewals_pipeline`: two `adapter.find` reads
(`showcase_project`, `showcase_invoice`) in an effect keyed on the
selected account.
  - `showcase_task_desk`: no `useAdapter` read.
- **Taught:** the objectui `react-pages` guide ("Live data" teaches
`adapter.find(...)` in an effect keyed on `[adapter]`, and the
record-family table teaches `useAdapter().findOne`); objectstack
`content/docs/ui/react-pages.mdx`; objectstack
`skills/objectstack-ui/rules/pages.md`.
- **Lit control:** the same `useAdapter` pattern hits the scope's
injection site in `@object-ui/components`' react-page renderer 7 times.

Verdict: a shipped and taught react page reads data through the injected
`useAdapter`, so injecting `useDataInvalidation` into the author scope
widens a public surface. Member 3 stops at `needs_decision`, with the
four-axis analysis in the dev report on the card. Nothing is injected
and the guide is not edited here.

## Pins, red on base and green on head

`ObjectView.busReread-10887.test.tsx` (plugin-view) and
`DashboardFilterBar.busReread-10887.test.tsx` (plugin-dashboard), both
through the real `SchemaRenderer` and each package's own registration.
The inner non-grid views of `object-view` are stand-ins registered in
the real registry (plugin-view does not depend on the plugins that
register them, and the rows are `ObjectView`'s read); each carries an
instance id, so a remount would show.

- Per member, per view type or read: one read on mount, a second after
`'*'`, a second after its own object, none after an unrelated object;
for `object-view`, the re-read rows reach the view and the view keeps
its instance.
- Member 2: the value chosen through the real select survives the
re-read on the same trigger node, and the re-read options reach the
list.
- Controls: an `object-view` with no object and a filter with no
`optionsFrom` read nothing on mount or on `'*'`; host-composed `tree`
and `chart` views are not re-read by this fetch; the two lit controls
above.

Final head `56354a1c38`: `Tests 23 passed (23)`. The same committed pins
over base's two source files (restored from `HEAD` afterwards, blob
equal and `git diff HEAD` empty): `Tests 17 failed | 6 passed (23)`, the
6 being the controls.

## Ablations (each through `ablation-replace`: anchor hit once, blob
changed, restored to the `HEAD` blob, `git diff HEAD` empty)

| ablation | anchor | result |
| --- | --- | --- |
| 1: drop the nonce from `ObjectView`'s fetch dependency list |
`invalidationNonce,` | `Tests 12 failed \| 11 passed (23)`: only member
1's rows (6 types x 2) |
| 2: drop the nonce from `SelectFilter`'s dependency list | `,
invalidationNonce]);` | `Tests 5 failed \| 18 passed (23)`: only member
2's rows |
| 3: drop the `tree` / `chart` exclusion | the two `currentViewType`
comparisons | `Tests 2 failed \| 21 passed (23)`: only the two host-only
controls, `find` 1 → 2 each |

The pins import their subjects by relative path and `@object-ui/react`
through the root vitest alias, so no `dist/` sits between the mutation
and the run.

## Consumer sweep

- `plugin-view` suite (under the verify lock): `Test Files 53 passed
(53)`, `Tests 527 passed (527)`.
- `plugin-dashboard` suite (under the verify lock): `Test Files 147
passed (147)`, `Tests 1339 passed | 6 skipped (1345)`.
- Render-level consumers outside the two packages that mount
`object-view` or a `dashboard` with `globalFilters`, 9 files in one run:
`Test Files 9 passed (9)`, `Tests 657 passed (657)`:
`timelineAxisRefusalReach-7070` (console), `catalog-gallery-render` and
`plugin-dashboard-gallery-render` (schema-catalog),
`rowClickModifierPayloadForward-9462`,
`ObjectView.densityModeThroughFold-7924`,
`ViewPreview.optionsFold-7928`, `StudioDesignSurface.gridRefresh`,
`DashboardWidgetInspector` (app-shell),
`DashboardDesignPage.presetAlias` (plugin-designer).
- objectui#10519 census test on `main`: none (0 test files under `apps/`
or `packages/app-shell` name 10519, at base and at `origin/main`
`6a7f24e92c`; the same grep over `packages/` test files at base finds
the 11 earlier reader pins, so the pattern is live).
- The suites ran at `a70c7607c4`. Every later commit touches only the
two pin files, which were re-run at `56354a1c38`.

## Gates (head `56354a1c38`)

- `type-check` for both packages (echoed as `tsc --noEmit && tsc -p
tsconfig.test.json`, the second program includes the pins): Done, Done.
- `check:control-bytes`, `check:new-line-citations` (`0 new
citation(s)`), `check:phantom-deps`, `check:unused-deps`,
`check:test-path-roots`, `check:changeset-claims`,
`check:pending-changeset-literals`, `check:shell-escape-residue`,
`changeset:check`, `check-changeset-presence` (`4 source file(s) of 2
released package(s) changed, and this change declares 2 changeset(s)`):
all exit 0.
- `check-governed-queue-guard --test` over the 6 changed paths: `NOT
GOVERNED`.
- ESLint on the 4 changed files, `--no-inline-config --format json`: 4
files, 0 errors; the pins carry 0 warnings. The two source files carry
the same warnings per rule on base and head (`ObjectView.tsx` 60,
`DashboardFilterBar.tsx` 18, read by linting each file's base blob
through `--stdin-filename`). A narrowed run, not the repo lint: the root
config enables no type-aware linting, so this diff cannot move a verdict
on an untouched file. `pnpm lint` is CI's.
- Dependency closure built first (`turbo run build` over the `^...`
closures of both packages, 16 tasks), because `type-check` reads the
built `.d.ts`.

## Serial

- PR objectstack-ai#10884 is merged (squash `35d68c4ce5`, which is this branch's
base). `git merge-tree --write-tree` of this head against that PR's head
`5dc80e3998` and against `origin/main` `6a7f24e92c`: both clean.
- This change's three `ObjectView` hunks are the `@object-ui/react`
import, the hook above the non-grid fetch effect, and that effect's
dependency list, all above `renderContent`.
- No open PR named `DashboardFilterBar` at claim time (claim record).

## Acceptance notes

- **What this unblocks for objectui#10519 (round 4 of PR objectstack-ai#10777):**
census rows for `object-view` (kanban, plus calendar, gallery, timeline,
map, gantt, all measured here) and for a dashboard with an `optionsFrom`
filter can now be green without the remount. The react-page row waits on
the member 3 decision.
- The `renderContent` comment on `identityKey` still says the data-fed
views are re-read "when `refreshKey` moves". That is now incomplete
(they also re-read on the bus) but not false, and it sits in the region
this PR was told not to touch. Carrier: none.
- Observation, not filed: for the host-only `tree` and `chart`,
`ObjectView` still runs its non-grid fetch on mount and on `refreshKey`
although neither renderer draws those rows. That is a read nobody uses;
it predates this change and this change adds no read to it. Carrier:
none.

---
_Generated by [Claude
Code](https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
akarma-synetal pushed a commit to akarma-synetal/objectui that referenced this pull request Sep 28, 2026
…i#10793, objectstack-ai#10802 and objectstack-ai#10821 made false (objectui#10877) (objectstack-ai#10891)

Fixes objectstack-ai#10877
Clause-②: no. It is release-text hygiene, and no published type or
behaviour moves.

## What this does

Eight pending changesets each state a contract fact that a later PR in
this same unreleased batch made false: PR objectstack-ai#10793 (objectui#7924,
`densityMode` retired), PR objectstack-ai#10821 (objectui#7928,
`ObjectViewSchema.listViews` mirrored by reference, a named view's
`options` bag refused whole) and PR objectstack-ai#10802 (objectui#10770,
`ObjectChartSchema.series` became two arms). A pending changeset
publishes verbatim into the CHANGELOG at the next release, and the
release PR objectui#5400 is still open. So each false sentence now has a
dated supersession note beside it.

The note form is copied from the landed precedents, PR objectstack-ai#10533 (the
`6235-mergedsort-wrap.md` note) and PR objectstack-ai#10828 (objectui#10801): a bold
headline naming the date and the card whose change supersedes the
sentence, then what holds now, then the superseding entry by file and
PR, then "the text above is kept as the reading of this change".

- Zero lines deleted: `git diff --numstat` against the merge-base
`9f0c84a448` is 168 insertions, 0 deletions, over 8 files. The first
round added 10 notes in 6 files. Patch round 1 added 4 notes in 2 files
(below). Patch round 2 edited the 7779 note and added none, so there are
14 notes.
- The frontmatter is byte-identical in all nine entries the claim and
its addendum `5862987983` name. The md5 of the frontmatter block is
equal at the merge-base `9f0c84a448` and the head `50deccd71a` for each:
7924-retire `ecedb10c`, 8355 `8f322642`, 10321 `564c93d5`, 10584
`84f96a9a`, 8365 `379475b3`, 9242 `acef2a2c`, 8086 `e927372d`, 7779
`ecedb10c`, 7946 `2ffd8635`.
- The seventh entry, `8086-chart-series-type-arm.md`, is untouched
(whole-file md5 `7ea6b544` at base and head). Every sentence in it
measured STILL-TRUE (see the lit control below).
- No new changeset: the diff is `.changeset/` only, and
`check-changeset-presence` says none is owed.

| Entry | FALSE-NOW sentences, quoted | Superseding entry (PR) | Notes |
| --- | --- | --- | --- |
| `7924-named-list-view-retire.md` | "`NamedListView` (one entry of
`ObjectViewSchema.listViews`)"; "Two bucket-③ members are NOT retired"
and "They stay declared as they were." |
`7928-listviews-by-reference-fold.md` (objectstack-ai#10821);
`7924-density-mode-through-fold.md` (objectstack-ai#10793) | 2 |
| `8355-retire-calendar-date-aliases.md` | "it now names the key and the
remedy on all five authoring surfaces", for "a named view's two nestings
under `listViews`"; "plus a `.check()` on `ObjectViewSchema` for a named
view's two nestings"; "`ObjectViewSchema.listViews` stays UNMIRRORED"
and the rest of that paragraph; "`timeline.dateField` stays a live,
accepted alias", on a named view; "the structural guard that `listViews`
stays out of the object's `shape`" |
`7928-listviews-by-reference-fold.md` (objectstack-ai#10821) | 3 |
| `10321-named-view-kanban-groupby-refused.md` |
"`ObjectViewSchema.listViews` is unmirrored"; "and in its legacy
`options.kanban` bag" with the issue path under `options.kanban`;
"reports one issue per written key, naming the key and the replacement";
"Not changed, deliberately: `listViews` stays unmirrored" and the rest
of that paragraph | `7928-listviews-by-reference-fold.md` (objectstack-ai#10821) | 2 |
| `10584-object-chart-series-describe.md` | "It now says the copy is
that `{ dataKey }` arm minus its per-series `type`, and that the
per-series family override on an `object-chart` node is `chartType`";
"taking `type` up would widen a published accept set" |
`10770-object-chart-react-tier-node.md` (objectstack-ai#10802) | 1 |
| `8365-stray-kanban-groupby-refused.md` | "objectui#10321 gives that
route its own door, which refuses the key in either nesting with the
same message" | `7928-listviews-by-reference-fold.md` (objectstack-ai#10821) | 1 |
| `9242-stray-kanban-groupby-lane-second-route.md` | "which refuses the
key in either nesting with the same message"; "(a named view's `kanban`
is typed as the `list-view`'s), but not the key inside the untyped
`options.kanban` bag on either" | `7928-listviews-by-reference-fold.md`
(objectstack-ai#10821) | 1 |
| `object-view-unmirrored-keys-7779.md` (patch round 1) | the table's
"still unmirrored"; "`listViews` stays unmirrored, on the ruling's own
fallback clause."; "The declaration's value is the local
`NamedListView`"; the 21-read / 43-unread figures; "The key therefore
stays in the parity ledger … until the maintainer decides its value
type" | `7928-listviews-by-reference-fold.md` (objectstack-ai#10821) | 1 |
| `objectchart-schema-anchor-7946.md` (patch round 1) | the `series`
half of "value types taken from their READ sites (`ChartRendererProps`
for `xAxisKey` / `series`)" and of "two are INTERNAL, relay-composed
(`xAxisKey`, `series` …)"; "`series` entries must be `dataKey`-shaped.";
"`aggregate` is the exception, and only because the spec's own object is
strict." | `10770-object-chart-react-tier-node.md` (objectstack-ai#10802) | 3 |

**Placement.** Each note sits directly after the paragraph that holds
the false sentence. Where two false paragraphs are adjacent, one note
covers both: the first 10321 note, the 10584 note and the last 8355
note. The 8355 note after the UNMIRRORED paragraph also covers the
matching clause in the `@object-ui/types` bullet above it, because both
describe the same named-view check. No note cites a line address; source
facts are cited by symbol and test name, and `check:new-line-citations`
reads 0 new citations.

## Measured, not inherited (one-time probes, not committed)

The probes ran the worktree sources at base `b2683a2c0e`, with
`@objectstack/spec` 17.4.0 resolved from `packages/types`. The zod face
ran through `tsx`, and the TS face through `tsc --noEmit` against
`packages/types/src/index.ts`. Each probe sits beside a control that
differs only by the key. A named view is `listViews.v` with `columns`
set.

Zod face (`ObjectViewSchema.safeParse`; `safeValidateSchema` gives the
same issues on the two `kanban` rows):

| Probe | Result |
| --- | --- |
| control `kanban: { groupByField, columns }` | parses |
| `kanban.groupBy` | two issues: `unrecognized_keys` at
`listViews.v.kanban` naming `groupBy`, and `custom` at
`listViews.v.kanban.groupBy` |
| `options: { kanban: { groupBy } }` | one issue: `unrecognized_keys` at
`listViews.v` naming `options`; nothing at the `options.kanban.groupBy`
path |
| `calendar.dateField` | `unrecognized_keys` at `listViews.v.calendar`,
and `custom` at `listViews.v.calendar.dateField` |
| `options: { calendar: { dateField } }` | one issue: `options` refused
|
| `timeline.dateField` | `unrecognized_keys` at `listViews.v.timeline`;
control `timeline: { startDateField, titleField }` parses |
| each of the 16 retired `NamedListView` spellings, and `densityMode` |
`unrecognized_keys` at `listViews.v` naming the key (17 of 17); control
`userActions: { search: false }` parses |
| `listViews` in `ObjectViewSchema.shape` | present |
| list-view route: `kanban.groupBy`, `options.kanban.groupBy`,
`calendar.dateField`, `options.calendar.dateField`, `timeline.dateField`
| `invalid_type`, `custom`, `invalid_type`, `custom`, parses: unchanged
|

TS face (`tsc`):

| Probe | Result |
| --- | --- |
| named view `kanban.groupBy` in an `ObjectViewSchema` literal | TS2353,
excess property on the spec's kanban block |
| named view `options` | TS2353, `options` does not exist |
| named view canonical control | compiles |
| `list-view` `kanban.groupBy` | TS2322 (`never`): unchanged |
| `list-view` `options.kanban.groupBy` | compiles: unchanged |
| a `NamedListView` value as a `listViews` entry | TS2322, not
assignable |
| `NamedListView` with `densityMode`; with `allowExport` | TS2322;
compiles |

Chart series (`ObjectChartSchema.safeParse`): `{ name, type: 'line' }`
parses with `type` kept; `{ dataKey, type: 'line' }` parses with `type`
stripped; `{ name, chartType }` is refused (`invalid_union`). The series
`.describe()` string now names both arms.

The runtime fold was read, not probed: `normalizeListViewSchema` still
folds `densityMode` and the eight `show*` flags, and `ListView` still
merges `options.kanban` under `kanban`.

## Lit control: STILL-TRUE, left byte-identical

- `8086-chart-series-type-arm.md`, the whole entry. PR objectstack-ai#10802 changed no
`plugin-charts` source (its one file there is a test).
`ChartRendererProps.schema.series` still declares `type?: string` on
both arms. The `ObjectChartSchema.series` docblock still says the copy
of the internal arm omits `type`, and that internal arm's type is the
same literal as before PR objectstack-ai#10802.
- 8365, "takes the identical message through a check on that untyped
bag", and the distinguishing `options.kanban` fixture. Both are the
`list-view` route, measured unchanged.
- 9242, "a `list-view` document carrying the key, under `kanban` or
under the legacy `options.kanban` bag, is refused by
`safeValidateSchema`". Measured unchanged.
- 8355, the `@object-ui/types` bullet's "on the legacy
`options.calendar` nesting (a check, since an open record declares no
member". This is the list-view surface, measured unchanged.
- 7924-retire, "`densityMode` is folded onto `rowHeight` by
`normalizeListViewSchema`" and "Only the authoring face changes". The
fold still does both.
- 10584, "Wording only. The accepted shape does not move on either
face". This is a statement about that change's own delta, and it is
still true of that change.
- 7779 (patch round 1): `viewTabBar` retired; the eight keys enforced;
"64 declared top-level members"; the spec record requires `columns` and
refuses `options`, tuple filters and `default`; "not papered over with
`z.any()`"; "absent stays valid on all nine"; the ledger move 14/96 →
14/87 (that change's own delta). Each measured true at `0ea55c3701`.
- 7946 (patch round 1): the `xAxisKey` half of the INTERNAL sentence
(its `.describe()` still says all five producers compute it); `type:
'object-chart'` required; `chartType` must be the declared union;
`normalizeChartSchema` translates the `{ name }` arm; `aggregate`
`function` / `groupBy` required with unknown members refused; a
misspelled key on the node itself is not refused; `filter` keeps both
arms. Each measured true at `0ea55c3701`.

## Gates (re-taken at head `50deccd71a`; hand-derived from
`package.json` and `.github/workflows/`; each exit captured by
redirect-then-status)

| Command | Verdict line | Exit |
| --- | --- | --- |
| `pnpm check:pending-changeset-literals` | ✅ No test source names a
pending changeset. | 0 |
| `node scripts/check-changeset-claims.mjs` | ✅ No pending changeset
names a file this change touches. (Self-contradiction reading: every
declared package across the 8 bodies is not negated.) | 0 |
| `node scripts/check-changeset-presence.mjs` | ✅ No source or published
contract of a released package changed in this range, so no changeset is
owed. | 0 |
| `node scripts/check-changeset-no-major.mjs` | ✅ No changeset declares
a `major` bump. | 0 |
| `node scripts/check-changeset-overwrite.mjs` | report-only: this
change touches 8 changesets it did not add. The declared packages are
identical at base and head for all 8, which is the prose-correction
case. | 0 |
| `node scripts/check-changeset-fixed.mjs` | ✅ All workspace packages
are in the changeset fixed group. | 0 |
| `pnpm check:new-line-citations` | VERDICT
new-cross-file-line-citations: 0 new citation(s) | 0 |
| `pnpm check:control-bytes` | ✅ check-control-bytes: OK | 0 |
| `pnpm check:spec-symbols` | ✅ spec member citations: nothing cites a
key its spec symbol does not declare. | 0 |
| `pnpm check:test-path-roots` | ✅ check-test-path-roots: OK | 0 |
| `node scripts/check-installed-spec-pin-claims.mjs` | OK (it excludes
`.changeset/`, so it was run for completeness only) | 0 |
| `node scripts/check-governed-queue-guard.mjs --test` (the 8 paths) | ✅
NOT GOVERNED | 0 |

Tests, from the repo root under the shared verify lock: every
`scripts/__tests__` file that names `.changeset` (derived with `git grep
-l`: 24 test files, plus one fixture that is not a test), and the five
`@object-ui/types` pins that re-derive what the notes state
(`calendar-date-alias-refusal-8355`,
`named-view-kanban-stray-group-by-10321`,
`object-view-listviews-by-reference-7928`,
`object-chart-react-tier-node-10770`,
`object-view-unmirrored-keys-7779`). Result at `f5e0123f4b`: **Test
Files 29 passed (29), Tests 1193 passed (1193)**, lock `VERDICT
command-exit 0`. Re-taken at `0ea55c3701` with the 24
`scripts/__tests__` files and the four pins the new notes rest on
(`object-view-unmirrored-keys-7779`,
`object-chart-react-tier-node-10770`,
`object-view-listviews-by-reference-7928`, `zod-mirror-parity`): **Test
Files 28 passed (28), Tests 1200 passed (1200)**, lock `VERDICT
command-exit 0`. Re-taken at `50deccd71a` (patch round 2): the 24
`scripts/__tests__` files plus `object-view-unmirrored-keys-7779` gave
**25 files / 1146 tests**, and the four other pins
(`calendar-date-alias-refusal-8355`,
`named-view-kanban-stray-group-by-10321`,
`object-view-listviews-by-reference-7928`,
`object-chart-react-tier-node-10770`) gave **4 files / 48 tests**. Both
runs gave lock `VERDICT command-exit 0`.

Not applicable, with the reason:
- `@object-ui/types` type-check, package vitest runs and consumer
sweeps: the diff touches no package source, only `.changeset/`.
- The Spec Main Shape Gate: no touched type.
- ESLint `--no-inline-config` on the touched files: `eslint.config.js`
has no configuration that matches `.md`. All 8 files report "File
ignored because no matching configuration was supplied", so it judged
nothing (NOT MEASURED).

## Acceptance notes

- **Residue in the same family: closed by patch round 1.** The first
round stopped on two files outside the claimed surface,
`object-view-unmirrored-keys-7779.md` and
`objectchart-schema-anchor-7946.md`. Each is named by a superseding PR
itself. The seat extended the file surface (addendum `5862987983`,
option A), and patch round 1 (`80ee15eecb`) added one note to 7779 and
three to 7946, in the same form and measured the same way. `Fixes
objectstack-ai#10877` stays.
- **For the at-tier review.** The last 7946 note also names two changes
outside the three PRs: `drillDown` (objectui#8885, PR objectstack-ai#8895) and `xAxis`
/ `yAxis` (objectui#10518, PR objectstack-ai#10601). Both landed after the 7946 entry,
and both also make its "`aggregate` is the exception" sentence false.
That sentence of the note can be dropped if the review holds it out of
scope; the rest of the note stands without it.
- Not measured: the wider pending population that names `listViews` /
`NamedListView` or `ObjectChartSchema`. Only the entries the superseding
PRs name were read.
- **Source prose. These are observations and were not filed; carrier:
none.**
- The `ObjectChartSchema.series` TS docblock
(`packages/types/src/objectql.ts`) says "A `type` written on this arm
anyway is an excess property on a literal typed by this interface."
Since PR objectstack-ai#10802 that is false. `{ dataKey: 'amount', type: 'line' }`
compiles in an `ObjectChartSchema` literal, because the union's `{ name
}` arm declares `type` (tsc probe above). The zod mirror still strips
it.
- The zod docblock headed "WHERE THIS ARM IS INSTALLED — TWO ROUTES, TWO
NESTINGS EACH, ONE STRING" (`packages/types/src/zod/objectql.zod.ts`)
still says a named view's `listViews` "is unmirrored", and it still
names a `custom` issue under `options.kanban` on that route. The
docblocks of `checkNamedViewKanbanStrayGroupBy` and
`namedViewKindRefusals` are current.
- **A dispatch premise, corrected.** "Each one already states in its own
body that it supersedes the older entry" holds for
`7924-density-mode-through-fold.md` and
`7928-listviews-by-reference-fold.md`.
`10770-object-chart-react-tier-node.md` names no superseded entry. The
supersession is stated in PR objectstack-ai#10802's body ("Superseded pending
changesets"), which names the objectui#10584, objectui#7946-anchor and
objectui#8086 entries.
- **`main` merged, twice.** `origin/main` `4f38f39286` was merged as
`0ea55c3701`: the two merged commits touch none of the probed sources,
and the probes re-ran byte-identical. `origin/main` `9f0c84a448` was
merged as `bade176fff` (patch round 2), which carries `35d68c4c` (PR
objectstack-ai#10884, objectui#10758). That merge moved probed sources
(`objectql.zod.ts`, `objectql.ts`, `ChartRenderer.tsx`); the probes
re-ran at `50deccd71a`, byte-identical to round 1's head outputs.
Neither merge used a rebase or a force-push.
- **Patch round 2 (at-tier record `5864532186`, FAIL on one note).** The
7779 note said the renderer reads 20 members and leaves 44 unread. That
was true at `4f38f39286`, but PR objectstack-ai#10884 re-took the census to 43 / 21 on
`main`. Per the seat's decision (REWORK `5864543090`), the note now
states no figure. It says the split is no longer the 21 / 43 the entry
counted, names both changes that moved it (objectui#7928 and
objectui#10758), and cites the census pin in
`object-view-unmirrored-keys-7779.test.ts`, which re-derives the figures
on every run (`43 read + 21 unread = 64` at the head). No other note
moved.
- **The headline card.** Each note's headline names the card whose
change supersedes the sentence, as the PR objectstack-ai#10533 notes did. PR objectstack-ai#10828's
notes named their own card instead, because the cause there was a spec
release that has no pending entry.
- **Commit trailers are model-free** (`Co-Authored-By: Claude`), per the
dispatch. The harness offered a trailer that carries a model name, and
it was not used.

Implemented by a `domain:spec#1` dispatch (dev subagent). Session:
`https://claude.ai/code/session_014fWVhLzhxR8qrFsJ5o8TYW`.

---
_Generated by [Claude
Code](https://claude.ai/code/session_014fWVhLzhxR8qrFsJ5o8TYW)_

---------

Co-authored-by: Claude <noreply@anthropic.com>
os-elon-musk pushed a commit that referenced this pull request Sep 28, 2026
…e false (objectui#10885)

Append a dated note, in the objectui#10533 form, to the two pending
changesets whose census figures no longer hold for the release:

- 7924-named-list-view-liveness-census.md: "21 read and 43 unread of 64
  today" is 43 read / 21 unread of 64 since objectui#7928 and
  objectui#10758; the disposition no longer "stays with objectui#7928".
- 7924-named-list-view-disposition-buckets.md: "43 unread ... 23 + 8 +
  10 + 2" is 21 unread, 0 + 8 + 11 + 2.

Frontmatter byte-identical; zero lines deleted.

Claude-Session: https://claude.ai/code/session_014mXUNuFomfj24w7s1pZzhN
Co-authored-by: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

documentation Improvements or additions to documentation package: types plugin tests

Projects

None yet

Development

Successfully merging this pull request may close these issues.

object-view: a named view's protocol-declared members are not read by the delegation — bucket ① of objectui#7924's ruling, split off as its own card

2 participants