docs: cite landing commits where pending changesets and source comments name fourteen objectui issues that answer 404 (objectui#10803)

.changeset/10803-dead-citation-sweep-second-batch.md

@@ -0,0 +1,20 @@
+---
+---
+
+Comment-only in `@object-ui/app-shell`, `@object-ui/components`, `@object-ui/core`,
+`@object-ui/fields`, `@object-ui/plugin-calendar`, `@object-ui/plugin-chatbot`,
+`@object-ui/plugin-dashboard`, `@object-ui/plugin-gantt`, `@object-ui/plugin-grid`,
+`@object-ui/plugin-list`, `@object-ui/plugin-map`, `@object-ui/plugin-timeline`,
+`@object-ui/plugin-tree`, `@object-ui/plugin-view`, `@object-ui/react`,
+`@object-ui/sdui-parser` and `@object-ui/types`: docblocks and code comments that cited
+objectui issues which answer 404 now cite the commit that landed each change, as a
+9-character sha, or the ruling by its date where no single commit answers (objectui#10803,
+the second batch after objectui#10771). Some of these docblocks reach the emitted `.d.ts`
+and `.js`; none of them moves a claim, and none of these comment edits changes a code or
+type token. No published behaviour changes through them, so this declares no release.
+The one runtime text the sweep touches, eleven zod `.describe()` strings in
+`@object-ui/types`, is declared on its own as a patch in
+`10803-describe-string-citations.md`.
+
+The same repair in the pending changesets that carried these citations is prose-only, and
+their frontmatter is byte-identical.

.changeset/10803-describe-string-citations.md

@@ -0,0 +1,19 @@
+---
+'@object-ui/types': patch
+---
+
+fix(types): eleven zod `.describe()` strings no longer point at an objectui issue that answers 404
+
+The runtime descriptions of eleven members of the published zod mirrors ended with, or
+carried, a pointer to an objectui issue that answers 404: `CheckboxSchema.wrapperClass`,
+`TreeViewSchema.nodes`, `ObjectMapSchema.objectName`, `ObjectGanttSchema.objectName`,
+`ObjectGanttSchema.data`, `TooltipSchema.trigger` / `content` / `children`, and
+`ContextMenuSchema.triggerClassName` / `contentClassName` / `modal`. A reader of a zod
+`description` has no repository to resolve a commit against, so each dead pointer is dropped
+rather than replaced (objectui#10803). On `ObjectGanttSchema.data` the clause that carried it,
+"undeclared on either face until" that issue, goes with it, leaving
+"Data source configuration — read FIRST by resolveRecordSourceConfig": the same shape as its
+`ObjectMapSchema.data` twin, which names `getDataConfig`.
+
+Nothing else in any string moves, and no key, path, accept set or refusal changes: every
+document that parsed before parses the same way.

.changeset/5853-tablecolumn-type-canonical-union.md

@@ -6,7 +6,7 @@
 ---
 
 `TableColumn.type` now has ONE canonical value set across all three ends that disagreed
-(objectui#5853, maintainer ruling 2026-08-25, Option B: the 8-literal interface union is
+(maintainer ruling 2026-08-25, Option B: the 8-literal interface union is
 canonical). The interface declared `'text' | 'number' | 'date' | 'datetime' | 'currency' |
 'percent' | 'boolean' | 'action'`; the zod mirror declared `z.string()` and accepted
 anything; the renderer branched on a third set and could only read the key through an

.changeset/6004-grid-column-emit-boundary.md

@@ -26,6 +26,6 @@ Key verdicts: `headerIcon`, `pinned` and `wrap` are HELD and now declared at the
 seam; `options` is RETIRED — nothing on either side of the seam read it, and
 every value it carried still reaches its consumer through the field metadata the
 cell closure captures and the object schema the inline editor reads. `type` stays
-objectui#5853's and `name` is not emitted here at all.
+under `fc62bb490`'s fold and `name` is not emitted here at all.
 
 No rendering change.

.changeset/6150-undeclared-but-consumed-keys.md

@@ -86,7 +86,7 @@ decisions publish in this same release.** This paragraph used to end "Declaring
 `data` stays required on both faces." Neither half of that is true any more, on
 either face:
 
-- **A `nodes`-only `tree-view` document IS legal.** objectui#6939 made `data`
+- **A `nodes`-only `tree-view` document IS legal.** `777e5c6f4` made `data`
   optional, so the `nodes` spelling the renderer reads FIRST stands on its own
   (`6939-tree-view-nodes-mirror.md`, published beside this note). `bind` is read
   before either and is unchanged.
@@ -95,7 +95,7 @@ either face:
   and the zod arm is a `retirementTombstone(...)` whose guidance points the author
   at `nodes` (`6951-tree-view-data-retired.md`, also published beside this note).
   `nodes` is `z.array(TreeNodeSchema).optional()`, and its own describe text
-  records that a `nodes`-only document became legal at objectui#6939.
+  records that a `nodes`-only document became legal (at `777e5c6f4`).
 
 ⛔ Nothing on the branch carrying this amendment falsified that sentence: it was
 already untrue at that branch's base, and both cards that made it untrue are

.changeset/6373-datatable-emit-boundary.md

@@ -17,7 +17,7 @@ closure captures is what this widget's type-aware rendering has always read, and
 untouched. Authored spellings still pass through, so a column the author wrote as
 `{ format: '$0,0' }` keeps its `format` exactly as before.
 
-`type` is unchanged — objectui#5853's fold at this seam still applies. `name` is unchanged
+`type` is unchanged — the fold `fc62bb490` put at this seam still applies. `name` is unchanged
 and still written: `data-table` reads `col.accessorKey || col.name` and objectui#5120 holds
 that alias while two published skill guides still teach a `{ name, label }` column. The hold
 is now declared at the seam instead of arriving anonymously inside a spread.

.changeset/6594-headercolor-mirror-enum.md

@@ -48,13 +48,13 @@ against synthetic inputs so the guard has been shown to fail rather than only to
 
 ## Shape, and where it departs from the nearest precedent
 
-The nearest precedent is objectui#5853 (`.changeset/5853-tablecolumn-type-canonical-union.md`),
+The nearest precedent is `fc62bb490` (`.changeset/5853-tablecolumn-type-canonical-union.md`),
 which narrowed `TableColumn.type` on the same three-ends pattern and **exported** a
 `TABLE_COLUMN_TYPES` tuple for the zod mirror to build its enum from. That shape is not
 available here and the difference is structural, not a preference: `packages/types/src/views.ts`
 is a **type-only** module, so a tuple there would add a runtime export to the package barrel
 (a value export cannot ride the barrel's `export type` block) and a runtime import edge from
-the zod entry into `views.js`. #5853 had a second reason to export — producers needed its
+the zod entry into `views.js`. `fc62bb490` had a second reason to export — producers needed its
 `normalizeTableColumnType()` at their emit seam — and `headerColor` has no producer that needs
 a runtime value. The literals are therefore written on each half and the anti-drift guarantee
 is carried by the pin above, which also covers the third end a shared tuple could not reach:

.changeset/6938-checkbox-wrapper-class.md

@@ -2,8 +2,8 @@
 '@object-ui/types': patch
 ---
 
-Declare `wrapperClass` on `CheckboxSchema`, on both faces (objectui#6938 — the
-residue of that card; its `context-menu` half landed with objectui#6939 group 1).
+Declare `wrapperClass` on `CheckboxSchema`, on both faces (the residue of its
+card; that card's `context-menu` half landed with `bfaa1589c`).
 
 `packages/components/src/renderers/form/checkbox.tsx:36` reads
 `cn("flex items-center space-x-2", schema.wrapperClass)` — classes on the wrapper

.changeset/6939-filter-builder-catalog-operator-spellings.md

@@ -1,7 +1,7 @@
 ---
 ---
 
-Corpus + test-only (objectui#6939): the three `components-complex-filter-builder`
+Corpus + test-only: the three `components-complex-filter-builder`
 schema-catalog entries now spell `conditions[].operator` with the members
 `FilterOperatorSchema` declares (`equals` / `less_than` / `greater_than`) instead
 of the spec's alias table (`eq` / `lt` / `gt`), so they pass

.changeset/6939-filter-builder-mirror.md

@@ -4,8 +4,8 @@
 
 Repair the `filter-builder` mirror: the field key is `value`, the type
 vocabulary is the value families the component actually folds a column into,
-and a filter group is `{ id, logic, conditions }` (objectui#6939, maintainer
-ruling recorded 2026-09-02 — one of the eight groups on that card, dispatched
+and a filter group is `{ id, logic, conditions }` (maintainer ruling
+recorded 2026-09-02 — one of the eight groups under that ruling, dispatched
 as its own PR per the ruling).
 
 Three independent mis-declarations sat in one member, and each is a key-name or
@@ -77,7 +77,7 @@ and both flagged for contract review rather than made quietly:**
   column draws the option-driven Select and no `<input>` at all, against a text
   box for an unrecognised spelling. Dropping it would refuse a spelling this
   mirror accepts *today* and the renderer draws distinctly, which is a fresh
-  instance of the class objectui#6939 exists to close.
+  instance of the class that ruling's eight groups exist to close.
 - **The group's `id` is declared OPTIONAL.** `isValidGroup` never consults it
   and nothing reads `filterGroup.id`; deleting it from an authored group renders
   byte-identically. Requiring it would invent a refusal the renderer does not
@@ -100,11 +100,11 @@ inside this same release by objectui#9559: `FilterOperatorSchema` became the
 spec rule's own operator member, so the alias and dropdown spellings are now
 accepted and normalised to canonical on parse), while the
 canonical spellings it accepts render a **blank** operator trigger — and it is
-reported on objectui#6939 rather than decided here. Seven further live field
+reported on its own card, objectui#7561, rather than decided here. Seven further live field
 types (`status`, `currency`, `percent`, `rating`, `lookup`, `master_detail`,
 `user`) each have their own bucket and control and are still refused; they were
 refused before this change as well, so that gap is pre-existing rather than a
-regression introduced here, and it is reported on the same card. The published
+regression introduced here, and it is reported on its own card, objectui#7562. The published
 doc for this component already offers all fourteen spellings and already marks
 `type` OPTIONAL, which the mirror still does not — a third declaration that
 agrees with the renderer, pinned here so the gap is measured rather than

.changeset/6939-kanban-column-cards.md

@@ -3,7 +3,7 @@
 ---
 
 Rename `KanbanColumn.items` to `cards`, in both halves of the published surface
-(objectui#6939, maintainer ruling 2026-09-02).
+(maintainer ruling 2026-09-02).
 
 **Breaking, deliberately.** `KanbanColumn` declared its card list as `items` in
 `complex.ts` and in the zod mirror `complex.zod.ts`. Every board reads `cards`.

.changeset/6939-objectql-record-source-refinement.md

@@ -4,8 +4,8 @@
 
 Repair the `object-map` and `object-gantt` mirrors: `objectName` is optional,
 and a refinement requires that at least one of `data`, `staticData`,
-`objectName` is present (objectui#6939, maintainer ruling recorded 2026-09-02 —
-this is one of the eight groups on that card, dispatched as its own PR per the
+`objectName` is present (maintainer ruling recorded 2026-09-02 — this is one
+of the eight groups under that ruling, dispatched as its own PR per the
 ruling).
 
 Both renderers resolve their records from one of three keys, in this order —

.changeset/6939-overlay-trigger-mirror.md

@@ -4,8 +4,8 @@
 
 Repair the `tooltip` and `context-menu` mirrors: declare the keys their renderers
 actually read, and stop requiring the `children` neither of them reads
-(objectui#6939, maintainer ruling recorded 2026-09-02 — this is one of the eight
-groups on that card, dispatched as its own PR per the ruling).
+(maintainer ruling recorded 2026-09-02 — this is one of the eight groups under
+that ruling, dispatched as its own PR per the ruling).
 
 Both members demanded `children` and omitted keys the renderer reads first, so
 `safeValidateSchema` refused two catalog entries that draw correctly:

.changeset/6939-tree-view-nodes-mirror.md

@@ -3,8 +3,8 @@
 ---
 
 Repair the `tree-view` mirror: `data` is optional, so the `nodes` spelling the
-renderer reads FIRST is a legal document on its own (objectui#6939, maintainer
-ruling recorded 2026-09-02 — this is one of the eight groups on that card,
+renderer reads FIRST is a legal document on its own (maintainer ruling
+recorded 2026-09-02 — this is one of the eight groups under that ruling,
 dispatched as its own PR per the ruling).
 
 `TreeViewSchema` REQUIRED `data`, the limb the renderer reads THIRD:

.changeset/6950-binding-framework-injected-input.md

@@ -4,7 +4,7 @@
 '@object-ui/core': patch
 ---
 
-`binding` on a component input is framework-set, not author-declared: `@object-ui/types` gains `InjectedComponentInput`, the `'field'` binding arm is retired from `@object-ui/sdui-parser`'s `RegistryConfigLike`, and the `as ComponentMeta` cast at the injection seam in `@object-ui/core` is gone (objectui#6950; maintainer ruling of 2026-09-07, director decision batch #69; ADR-0049 enforce-or-remove).
+`binding` on a component input is framework-set, not author-declared: `@object-ui/types` gains `InjectedComponentInput`, the `'field'` binding arm is retired from `@object-ui/sdui-parser`'s `RegistryConfigLike`, and the `as ComponentMeta` cast at the injection seam in `@object-ui/core` is gone (maintainer ruling of 2026-09-07, director decision batch #69; ADR-0049 enforce-or-remove).
 
 **What was measured.** `binding` was published — the manifest serializer forwards it — and read — `validateTree` records a binding site for it — while `ComponentInput`, the authoring type every registration writes against, did not declare it. The one writer in the tree, `ELEMENT_DATA_SOURCE_INPUT`, therefore carried a hand-written inline type and reached a registration's `inputs` through `as ComponentMeta` in `Registry.register`. Declared narrower than enforced, on a published type — and `ComponentInput`'s own docblock listed `binding` among the forwarded per-input keys.
 

.changeset/6951-tree-view-data-retired.md

@@ -15,7 +15,7 @@ unchanged and still read first.
 spellings for its one inline-nodes slot — `nodes` (read second) and `data` (read
 third: `boundData || schema.nodes || schema.data || []` at
 `renderers/data-display/tree-view.tsx:105`), both declared by objectui#6150.
-`data` had been REQUIRED until objectui#6939 / PR #7533 made it optional, so
+`data` had been REQUIRED until `777e5c6f4` (PR #7533) made it optional, so
 this retirement starts from a declared-and-optional member on both faces. The
 in-repo corpus at the retirement: seven `tree-view` nodes under
 `examples/schema-catalog` and `packages/types/examples` plus one `content/docs`

.changeset/7014-select-option-spec-claim-correction.md

@@ -15,7 +15,7 @@ Option-A ruling on objectui#6140 / objectui#6153 — and this repo's pin has mov
 Everything below was measured correctly against 17.2.0 and is kept as the dated record of
 why the comments were rewritten; ⛔ do not read it as the contract you are authoring
 against today. `rows` and `options[].description` are **authorable now**, and the doc
-comments were corrected again in objectui#7635, which releases alongside this entry.
+comments were corrected again in `544ecba84`, which releases alongside this entry.
 
 Three doc comments claimed the installed `@objectstack/spec` DECLARES a key that it in
 fact **refused by name**. Measured on `@objectstack/spec@17.2.0`, each paired with a

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

@@ -4,7 +4,7 @@
 '@object-ui/plugin-view': minor
 ---
 
-Gantt views no longer render on invented date field names (objectui#7070).
+Gantt views no longer render on invented date field names.
 
 The half PR #7062 fenced out and reported separately. A view that carried no
 `gantt:` block used to have a complete-looking date axis synthesized for it:
@@ -45,7 +45,7 @@ Also corrected: the objectui#3129 note at the top of `app-shell/ObjectView.tsx`
 certified the gantt branch below it as already using the safe two-rung shape.
 It did not. The note now states each sibling branch as measured, and says
 explicitly which fabrication REMAINS — the timeline `'created_at'` floor at the
-two plugin faces, which objectui#7070 routes to a ruling rather than settling
+two plugin faces, which is routed to a ruling rather than settled
 per-face.
 
 Deliberately out of scope, and left in place: `progressField` / `dependenciesField`

.changeset/7070-timeline-date-axis-floors-retired.md

@@ -4,7 +4,7 @@
 ---
 
 Retire the `'created_at'` timeline date-axis floors at both plugin faces
-(objectui#7070 step ③, maintainer ruling 2026-09-01, 总监批 #28).
+(step ③ of the maintainer ruling 2026-09-01, 总监批 #28).
 
 **Breaking, deliberately.** A timeline view that declares **no** date axis anywhere no
 longer renders. `ListView`'s and `ObjectView`'s timeline branches used to hand

.changeset/7083-richtext-field-metadata.md

@@ -5,7 +5,7 @@
 `RichtextFieldMetadata` — the third registry key of `RichTextField` becomes declarable
 (objectui#7083, maintainer ruling 2026-09-07, director decision batch #71).
 
-`markdown`, `html` and `richtext` are one widget (objectui#5498). Two of the three
+`markdown`, `html` and `richtext` are one widget (`4bb940b6e`). Two of the three
 already had an exported metadata type; `richtext` had none, so the runtime served it by
 structure while an author could not write its metadata under an annotation at all. The
 only way to write one was `as unknown as MarkdownFieldMetadata`, and that deliberate

.changeset/7113-chart-data-model.md

@@ -5,11 +5,11 @@
 `ChartSchema` declares the data model it renders — chart-level `data` and `xAxisKey`, with
 the bare-string `xAxis` folded onto the latter — and `ChartDataSeries` accepts both binding
 dialects (objectui#7113 option B, 项目总监席 总监批 #28 2026-09-01 「同意」; and
-objectui#6939's `chart` row, maintainer ruling 2026-09-02 「同意」 — both rulings
+the `chart` row of the eight-group maintainer ruling of 2026-09-02 「同意」 — both rulings
 independently instructed declaring these two keys, so they land as one change).
 
 ⚠️ Shipped as `minor`, not `patch`, because two document classes that validated before now
-REFUSE. objectui#6939 grades this class "patch where the accept set only widens toward what
+REFUSE. The 2026-09-02 ruling grades this class "patch where the accept set only widens toward what
 already renders"; this change is not a pure widening, so it takes the level objectui#6896
 set for the same transition in this same file — the mirror starting to refuse — and for the
 same reason: this repository's `major` is a cross-repo pin to `@objectstack`'s major rather
@@ -40,7 +40,7 @@ documents whose chart was already broken. Class 3 is not: at base it parsed to
 `normalizeChartSchema` renders it — `str(123)` is `undefined`, so the read falls back to
 `name` and yields `series: [{ dataKey: 'a' }]` (`normalizeChartSchema.ts:239`). So this is a
 narrowing away from a document that **renders today**, which is precisely the distinction
-objectui#6939's grading language turns on. `dataKey: null` behaves identically. Measured on
+the 2026-09-02 ruling's grading language turns on. `dataKey: null` behaves identically. Measured on
 both states; the declaration itself is right, and this note is the disclosure it was owed.
 
 ## Corrected: what class 2 actually did
@@ -95,7 +95,7 @@ would discard them.
 
 ## Not done, deliberately
 
-objectui#6939's `chart` row also says "`series[].data` stops being required". On this base
+The 2026-09-02 ruling's `chart` row also says "`series[].data` stops being required". On this base
 it already is not: objectui#6896 replaced it with `retirementTombstone(...)` —
 `z.never({ error }).optional()` — which is optional AND refuses any authored value by name.
 Implementing the clause literally would re-widen a retired key and reverse a landed ruling,

.changeset/7122-spec-17-3-0-parity-reconciliation.md

@@ -80,10 +80,10 @@ and `@object-ui/plugin-detail` `^17.1.0`, so a floor above the family minimum is
 this repo's normal state, not an exception.
 
 ⚠️ **Measured on both sides, because it is bump-caused rather than pre-existing
-and objectui#7688 records the opposite.** The gate is a scheduled / push-to-main
+and the card that filed the `[floor-too-low]` finding recorded the opposite.** The gate is a scheduled / push-to-main
 workflow that cannot red a pull request, and `main` is green on it — the last
 eight runs, most recently at `c2e3cee2c`. On this branch's built tree it exits 1
 with CI's own `--cross-check` invocation, and exits 0 with this raise, judging
 278 (subpath, symbol) pairs across 19 published packages either way. Its blocking
 copy runs on the publish path, so leaving it would have surfaced as a cancelled
-release rather than as a red check. The correction is recorded on objectui#7688.
+release rather than as a red check. The correction is recorded in `639114c4d`, the commit that landed this entry.