hygiene(changesets): pending entries that PRs #10793, #10802 and #10821 made false still publish verbatim — add dated supersession notes before Version Packages #10877

Activity

1. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   Path: business objects, records and views | 缺项 (no gate reads a pending changeset against a later entry in the same unreleased batch) | P2

   Triage: first grade — bug · documentation · priority:p2 · domain:spec · area:devpath · pm:queue

   Triage: lands in .changeset/*.md only. The four sentences the card quotes were re-read on objectui origin/main e2b3826 at 2026-09-28T02:53Z, and each is still there byte for byte: 7924-named-list-view-retire.md:43, 8355-retire-calendar-date-aliases.md:129, 10321-named-view-kanban-groupby-refused.md:10-12 and 10584-object-chart-series-describe.md:10-12. The three superseding entries (7924-density-mode-through-fold.md, 7928-listviews-by-reference-fold.md, 10770-object-chart-react-tier-node.md) exist on the same tip.

   Rationale:

   • Grade. A pending changeset publishes verbatim into the CHANGELOG at the next release. That is release-frozen text, the reach: exception, graded p2 by the same rule as objectui#10801, objectui#10771 and objectui#10606.
   • Filing-gate category. The body says ②. It is ① under the release-frozen-text exception, because no maintainer decision is asked. The card is kept.
   • Lane: domain:spec, not domain:ui. objectui#10801 went to domain:ui; this card does not follow it. Docs go with the surface they document. Every false sentence here states a packages/types contract fact: the NamedListView tombstone, the ObjectViewSchema.listViews mirror, and the ObjectChartSchema.series union. That is the spec lane's objectui contract surface (lanes/spec.md:13). All three superseding PRs, feat(types,plugin-view,app-shell)!: retire NamedListView.densityMode; both relays read the density through the fold (objectui#7924) #10793, feat(types,plugin-view,app-shell)!: ObjectViewSchema.listViews mirrors the spec's ObjectListViewSchema by reference; ViewPreview folds a stored options bag (objectui#7928) #10821 and feat(types): ObjectChartSchema accepts the react tier's object-chart node — the spec's { name } series arm, family as specType (objectui#10770) #10802, landed domain:spec cards (objectui#7924, objectui#7928, objectui#10770).

   Triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-09-28T02:53Z. ⛔ Not a claim, ⛔ not a dispatch. Read: this card (no comments), its four cited source cards, and the open objectui PR list.

   Duplicate check (the body searched open cards only; this adds closed ones). Every objectui issue created since 2026-09-25T10:09Z (125 issues, all states), matched by title on changeset|supersed|hygiene, gives 7 hits. None is a duplicate:

   • objectui#10868 (in flight) carries the code leftovers of the same retirements and does not touch .changeset/.
   • objectui#10803 (in flight) is the dead-citation family.
   • objectui#10801, objectui#10771 and objectui#10606 (closed) are other stale claims in other files.

   No serial constraint. At 2026-09-28T02:53Z, none of the 12 open non-release objectui PRs touches any of the seven files the card names. That includes PR objectui#10875 (objectui#10803) and PR objectui#10883 (objectui#10868).

   Deadline. The release PR objectui#5400 is open. If it publishes before this lands, the text is frozen and the card is moot. Close it then; don't re-point it at the CHANGELOG.

   Execution notes.

   1. The claimant re-reads each named entry on origin/main. It edits only the sentences that are false now, and leaves every true sentence byte-identical as the lit control. The options.KIND sentences and 8086-chart-series-type-arm.md are unmeasured; test them, don't assume them.
   2. Add a dated supersession note beside each false sentence, in the objectui#10533 / objectui#10801 form, naming the superseding entry. The frontmatter stays byte-identical. ⛔ Don't delete the history the note corrects. No new changeset.
   3. Run the pending-changeset gates the card names (check:pending-changeset-literals, check-changeset-claims) before pushing.

   Size/model suggestion: S · a docs-only .changeset/*.md edit. Changeset prose is a contract-review surface, so it needs an at-tier review.
2. objectstack-fleet

   added

   area:devpathThe road — create, dev, verify, publish/install, connect an agent, iterateThe road — create, dev, verify, publish/install, connect an agent, iterate

   bugSomething isn't workingSomething isn't working

   documentationImprovements or additions to documentationImprovements or additions to documentation

   domain:specobjectui spec stream: fix lands on packages/types, schema corpus or spec pin coupling — spec laneobjectui spec stream: fix lands on packages/types, schema corpus or spec pin coupling — spec lane

   pm:queue

   priority:p2

   on Sep 28, 2026
3. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   Claim: PM loop, serial — objectui#10877: each pending changeset sentence that PRs #10793, #10802 and #10821 made false gets a dated supersession note that names the superseding entry. True sentences stay byte-identical.
   Session: session_014fWVhLzhxR8qrFsJ5o8TYW
   Account: os-tesla
   Branch: claude/issue-10877-changeset-supersession-notes
   Worktree: objectui-issue-10877
   Domain: domain:spec
   Seat: domain:spec#1
   File surface: .changeset/*.md only, among the entries the card names: 7924-named-list-view-retire, 8355-retire-calendar-date-aliases, 10321-named-view-kanban-groupby-refused, 10584-object-chart-series-describe, and, measured first, 8365-stray-kanban-groupby-refused, 9242-stray-kanban-groupby-lane-second-route and 8086-chart-series-type-arm. Frontmatter stays byte-identical. No new changeset. Stop on breach and explain in the report.
   Container & model: S, mode:subagent, model: opus (TIER_DEFAULT; --tier refuses objectui paths)
   Clause-②: no. It is release-text hygiene, and no published type or behaviour moves.
   Thread-read: 5862427904
   Serial constraints cleared: re-measured on objectui origin/main b2683a2c0e at 2026-09-28T03:26Z. None of the 9 open PRs (#10888, #10886, #10884, #10852, #10780, #10777, #10714, #10278 and the release PR #5400) lists any of the seven files. The release PR objectui#5400 is the deadline: if it publishes first, this card is moot (triage 5862427904).

   _{domain:spec#1 · session_014fWVhLzhxR8qrFsJ5o8TYW · claim · 2026-09-28T03:26Z}

   ---

   Generated by Claude Code
4. objectstack-fleet

   added

   pm:dispatched

   and removed

   pm:queue

   on Sep 28, 2026
5. objectstack-fleet

   assigned

   os-tesla

   on Sep 28, 2026
6. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   os-dev-report
   {
   "issue": 10877,
   "status": "done",
   "branch": "claude/issue-10877-changeset-supersession-notes",
   "pr": "#10891",
   "session": "session_014fWVhLzhxR8qrFsJ5o8TYW (dev subagent; the parent's harness-stamped id)",
   "premise_still_valid": true,
   "summary": "The card's premise holds. Every one of the four quoted sentences is still false on base b2683a2. Of the unmeasured ones, some turned out false and some still true. Ten dated supersession notes were added across six pending changesets, in the PR #10533 / PR #10828 (objectui#10801) form, and each note names the superseding entry by file and PR. The notes cover 7924-named-list-view-retire (2), 8355 (3), 10321 (2), 10584 (1), 8365 (1) and 9242 (1). The diff is 115 insertions and 0 deletions, all frontmatter is byte-identical, there is no new changeset, and 8086-chart-series-type-arm is untouched: every sentence in it measured STILL-TRUE. Draft PR #10891 is open. Two things are left for the seat. First, the label-write that assigns os-tesla on the PR was refused by the auto-mode classifier. Second, two more same-family entries that the superseding PRs themselves name are false and sit outside the claimed file surface: object-view-unmirrored-keys-7779.md and objectchart-schema-anchor-7946.md. They were not edited, because the claim says to stop on a breach; see open_questions.",
   "tests": "gates (head f5e0123, exit captured by redirect-then-status): pnpm check:pending-changeset-literals exit 0 'No test source names a pending changeset.'; node scripts/check-changeset-claims.mjs exit 0 'No pending changeset names a file this change touches.' (self-contradiction reading PASS); node scripts/check-changeset-presence.mjs exit 0 'no changeset is owed'; node scripts/check-changeset-no-major.mjs exit 0; node scripts/check-changeset-overwrite.mjs exit 0 (report-only: 6 touched, declared packages identical at base and head); node scripts/check-changeset-fixed.mjs exit 0; pnpm check:new-line-citations exit 0 'VERDICT new-cross-file-line-citations: 0 new citation(s)'; pnpm check:control-bytes exit 0; pnpm check:spec-symbols exit 0; pnpm check:test-path-roots exit 0; node scripts/check-installed-spec-pin-claims.mjs exit 0 (excludes .changeset, run for completeness); node scripts/check-governed-queue-guard.mjs --test (6 paths) exit 0 NOT GOVERNED. vitest under os-verify-lock: the 24 scripts/tests test files that name .changeset (derived with git grep -l; one further hit is a fixture .md, not a test) plus 5 @object-ui/types pins the notes rely on (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): 'Test Files 29 passed (29), Tests 1193 passed (1193)', 'VERDICT command-exit 0' at f5e0123. line_budget: not applicable (no skills/** file). files_changed: 6, all .changeset/*.md (7924-named-list-view-retire, 8355-retire-calendar-date-aliases, 10321-named-view-kanban-groupby-refused, 10584-object-chart-series-describe, 8365-stray-kanban-groupby-refused, 9242-stray-kanban-groupby-lane-second-route). Measurement (one-time probes, not committed): zod through tsx and TS through tsc --noEmit against the worktree sources at spec 17.4.0; the probe tables are in the PR body. Lit control: 8086 whole entry; 8365's list-view options.kanban check; 9242's list-view sentence; 8355's list-view options.calendar arm; 7924's fold sentences; 10584's 'Wording only' delta. Ablation: none (docs-only). deviations: (1) label-write --assign os-tesla on PR 10891 was refused by the auto-mode classifier [External System Writes] before any request, so 0 writes; it was not retried by another route, and the seat does it. (2) A post-PR read-only 'git status' plus 'git ls-remote' was also refused by the classifier and was not retried. The branch tip is evidenced by the push output (b2683a2..f5e0123) and by the PR head sha f5e0123. The clean tree is evidenced by 'git worktree remove' succeeding without --force. (3) Not applicable, with the reason: the type-check, package vitest and consumer leads (the diff is .changeset/ only), the Spec Main Shape Gate (no touched type), and ESLint --no-inline-config (all 6 .md files report 'File ignored because no matching configuration was supplied', so NOT MEASURED). (4) The commit trailers are model-free per the dispatch; the harness offered a trailer carrying a model name, and it was not used. (5) A dispatch premise was corrected: 10770-object-chart-react-tier-node.md does not state which entries it supersedes. That statement is in PR #10802's body.",
   "mcp_calls": "0",
   "api_writes": "2 attempted, 1 relayed so far before this comment: (1) pr_create through fleet-write/dispatch.mjs, which sent POST /repos/objectstack-ai/objectstack/dispatches (204) and was executed as POST /repos/objectstack-ai/objectui/pulls, giving draft PR 10891 (run 36374916724 success); the body read back byte-identical (14235 bytes). (2) This os-dev-report comment through post-stamped.mjs: POST /repos//issues/10877/comments. Not sent: label-write --assign (refused by the classifier before any request). git push is not a REST write.",
   "open_questions": [
   {
   "question": "Same-family residue outside the claimed file surface. Two pending entries, each named by a superseding PR itself, carry a sentence the same PRs made false. The first is .changeset/object-view-unmirrored-keys-7779.md, whose paragraph opens 'listViews stays unmirrored, on the ruling's own fallback clause.'; the 7928 entry names the objectui#7779 entry as superseded. The second is .changeset/objectchart-schema-anchor-7946.md, with 'two are INTERNAL, relay-composed (xAxisKey, series ...)' and the migration bullet 'series entries must be dataKey-shaped.'; PR #10802's body names the 7946 anchor entry. Neither was edited, because the claim says to stop on a breach. How should they land before objectui#5400 merges?",
   "options": [
   "A: the seat amends the claim's file surface and runs a patch round on PR 10891. Two more dated notes in the same form, measured the same way. The PR keeps 'Fixes #10877'.",
   "B: land PR 10891 as it is and route the two files as a follow-up under this card. The PR body's first line then needs 'Part of #10877', and the seat writes that body edit.",
   "C: leave them. They would publish false in the CHANGELOG."
   ],
   "recommendation": "A. It is the same defect class, the same form and the same gate family, and the release-PR deadline favours one PR over two. A second PR also means a second at-tier review of the same surface."
   }
   ],
   "out_of_scope_findings": [
   "class: b · Seam: spec:ObjectViewSchema.listViews → consumer: none (release text) · reach: exception: release-text (a pending changeset publishes verbatim at objectui#5400) · evidence: .changeset/object-view-unmirrored-keys-7779.md says the value is the local NamedListView and that the key 'stays in the parity ledger ... until the maintainer decides its value type'. On base b2683a2, 'listViews' is in ObjectViewSchema.shape, and a NamedListView is not assignable to a listViews entry (tsc TS2322). Route: residue of this card (#10877), see open_questions. · dedupe words: listViews stays unmirrored, 7779 changeset, supersession note, pending changeset",
   "class: b · Seam: spec:ChartSeriesSchema → consumer: none (release text) · reach: exception: release-text · evidence: .changeset/objectchart-schema-anchor-7946.md says 'series entries must be dataKey-shaped' and calls series INTERNAL. On base, ObjectChartSchema.safeParse accepts a { name, type: 'line' } entry, and the series .describe() names two arms. Route: residue of this card (#10877), see open_questions. · dedupe words: 7946 anchor changeset, series dataKey-shaped, object-chart series union",
   "carrier: 承接者:无 · noted, not filed. This is source prose (an observation, not class a/b/c). The ObjectChartSchema.series TS docblock in packages/types/src/objectql.ts says 'A type written on this arm anyway is an excess property on a literal typed by this interface'. It has been false since PR #10802: { dataKey: 'amount', type: 'line' } compiles in an ObjectChartSchema literal, because the union's { name } arm declares type (tsc probe). The zod mirror still strips it.",
   "carrier: 承接者:无 · noted, not filed. This is source prose (an observation). The zod docblock headed 'WHERE THIS ARM IS INSTALLED — TWO ROUTES, TWO NESTINGS EACH, ONE STRING' in packages/types/src/zod/objectql.zod.ts still says a named view's listViews 'is unmirrored', and still names a custom issue under options.kanban on that route. The docblocks of checkNamedViewKanbanStrayGroupBy and namedViewKindRefusals are current.",
   "Not measured: the wider pending population that names listViews / NamedListView (about 30 entries) or ObjectChartSchema (12 entries). Only the entries this card names, plus the two that the superseding PRs name, were read."
   ]
   }

   ---

   Generated by Claude Code
7. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   File-surface addendum to claim 5862760726 (patch round 1). domain:spec#1 @ objectui · session_014fWVhLzhxR8qrFsJ5o8TYW · 2026-09-28T03:48Z. The claim continues; ⛔ this is not a new claim.

   The dev report (the newest os-dev-report above, PR #10891 at f5e0123f4b) stopped on a breach, correctly. Two more pending entries of the same family carry a sentence that the same superseding PRs made false, and each of those PRs names its entry. The seat re-read both on origin/main:

   • .changeset/object-view-unmirrored-keys-7779.md:47: 「listViews stays unmirrored, on the ruling's own fallback clause.」, followed by 「The declaration's value is the local NamedListView」. The 7928-listviews-by-reference-fold entry (PR feat(types,plugin-view,app-shell)!: ObjectViewSchema.listViews mirrors the spec's ObjectListViewSchema by reference; ViewPreview folds a stored options bag (objectui#7928) #10821) names the objectui#7779 entry as superseded.
   • .changeset/objectchart-schema-anchor-7946.md:34: 「two are INTERNAL, relay-composed (xAxisKey, series …」, plus the migration bullet that series entries must be dataKey-shaped. PR feat(types): ObjectChartSchema accepts the react tier's object-chart node — the spec's { name } series arm, family as specType (objectui#10770) #10802's body names the objectui#7946 anchor entry.

   Decision (execution, within the card): option A. The claim's file surface now also covers these two files, on the same terms:

   • measure-first;
   • a dated note beside each false sentence only;
   • frontmatter and every true sentence byte-identical;
   • no new changeset.

   The reasons: it is the same defect class and the same form, the release PR objectui#5400 is the deadline, and one PR needs one at-tier review. No open PR touches either file (re-measured at 2026-09-28T03:48Z across the 11 non-self open PRs).

   Patch round 1 runs on PR #10891. The PR keeps Fixes #10877.

   ---

   Generated by Claude Code
8. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   os-dev-report
   {
   "issue": 10877,
   "status": "done",
   "branch": "claude/issue-10877-changeset-supersession-notes",
   "pr": "#10891",
   "session": "session_014fWVhLzhxR8qrFsJ5o8TYW (patch-round-1 dev subagent; the parent's harness-stamped id)",
   "premise_still_valid": true,
   "summary": "Patch round 1 on PR #10891, under claim 5862760726 and the file-surface addendum 5862987983. Both addendum leads measured FALSE-NOW, and so did two sentences the leads did not name: the 7946 provenance bullet about aggregate being the only by-reference key, and the closing 'aggregate is the exception' sentence. Four dated notes were added in the form of the ten already on the branch, each naming the superseding entry by file and PR: one in object-view-unmirrored-keys-7779.md, placed after the paragraph at :47, and three in objectchart-schema-anchor-7946.md, placed after :38, after :77 and at the end. The round adds 51 lines and deletes none. Frontmatter is byte-identical, there is no new changeset, and no other file was touched. main had moved from b2683a2 to 4f38f39, so it was merged in with a merge commit (no rebase, no force-push). The new head is 0ea55c3, pushed. The release PR objectui#5400 is still open. The PR body is stale and belongs to the seat; deviations (1) lists what it needs.",
   "tests": "GATES at head 0ea55c3 (merge-base with origin/main 4f38f39). Each exit code was captured by redirect-then-status. pnpm check:pending-changeset-literals: exit 0, 'No test source names a pending changeset.' node scripts/check-changeset-claims.mjs: exit 0, 'No pending changeset names a file this change touches.' (8 bodies; self-contradiction reading PASS). node scripts/check-changeset-presence.mjs: exit 0, '8 file(s) changed, 0 of them published source ... no changeset is owed.' node scripts/check-changeset-no-major.mjs: exit 0. node scripts/check-changeset-overwrite.mjs: exit 0 (report-only; 8 modified; declared packages identical at base and head for all 8). node scripts/check-changeset-fixed.mjs: exit 0. pnpm check:new-line-citations: exit 0, 'VERDICT new-cross-file-line-citations: 0 new citation(s)'. pnpm check:control-bytes: exit 0, 'OK (scanned 9147 tracked text file(s)...)'. The self-scan grep -naP over the control-byte class on both files found no match (exit 1). pnpm check:spec-symbols: exit 0. pnpm check:test-path-roots: exit 0. node scripts/check-governed-queue-guard.mjs --test on the 8 branch paths: exit 0, 'NOT GOVERNED'. VITEST under os-verify-lock (slot objectui-issue-10877-r1), from the repo root: the 24 scripts/tests test files that name .changeset (git grep -l -F .changeset -- scripts/tests gives 25 hits: 24 tests plus 1 fixture .md), plus 4 @object-ui/types pins that 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). Result: 'Test Files 28 passed (28)', 'Tests 1200 passed (1200)', 'VERDICT command-exit 0' at 0ea55c3. MEASUREMENT (one-time probes, not committed): the zod face ran through tsx on packages/types/src/zod/objectql.zod.ts, and the TS face through tsc --noEmit on packages/types/src/index.ts, with @objectstack/spec 17.4.0. The probes ran on the pre-merge tree and were re-run at 0ea55c3 with byte-identical output. The two merged main commits (4128188, 4f38f39) touch none of the probed sources and change no lockfile or manifest. 7779 FALSE-NOW, covered by one note after :68: (a) the table's ':33 still unmirrored' and ':47 listViews stays unmirrored, on the ruling's own fallback clause.' The deciding source is ObjectViewSchema (zod), which declares listViews as stripImportedDefaults(SpecViewSchema).shape.listViews; probe: 'listViews in ObjectViewSchema.shape' is true. (b) ':47-48 The declaration's value is the local NamedListView'. The deciding source is the TS member ObjectViewSchema.listViews, now typed off SpecObjectListViewSchema's input; tsc gives TS2322 for a NamedListView value as a listViews entry. (c) The 21-read / 43-unread figures at :50-51 and :64. The deciding source is NAMED_VIEW_READ_DECLARED and the census it() '20 read + 44 unread = 64' in object-view-unmirrored-keys-7779.test.ts, green at head. (d) ':66-68 The key therefore stays in the parity ledger ... until the maintainer decides its value type.' The deciding source is the UnmirroredDeclared interface in zod-mirror-parity.test.ts, which has 12 entries and none for ObjectViewSchema, plus ruling A in 7928-listviews-by-reference-fold.md. 7779 STILL-TRUE, left byte-identical (lit controls): ':5-12 viewTabBar retired' (zod invalid_type at viewTabBar; tsc TS2322). ':14-31 eight keys enforced' (defaultViewType 'tree': invalid_value; navigation 'page': invalid_type; navigation {mode:'page'}: parses; searchableFields 'name' and viewActions 'share': invalid_type). ':48 64 declared top-level members' (pinned toBe(64)). ':55-61 the spec record requires columns and refuses options, tuple filters and default', with its two examples ({label}: invalid_union at listViews.v.columns; options: unrecognized_keys; filter.0: invalid_type; default: unrecognized_keys). ':68 It is not papered over with z.any()' (value 42: invalid_type; key bogus: unrecognized_keys). ':79 absent stays valid on all nine' (parses). ':80-83 ledger move 14/96 to 14/87': that change's own delta, still recorded in the zod-mirror-parity.test.ts header. ':62-63 Mirroring the spec value would have lost documented behaviour': a reading of its own time; the note adds that the docs have taught the spec shape since objectui#8255. 7946 FALSE-NOW, covered by three notes: (e) ':21-27 value types taken from their READ sites (ChartRendererProps for xAxisKey / series) ... except where @objectstack/spec already owns the shape, which is aggregate' and ':32-36 two are INTERNAL, relay-composed (xAxisKey, series ...)'. The series half is false and the xAxisKey half is true. The deciding source is ObjectChartSchema.series: on the TS face a SpecChartSeries arm in the element union, on the zod face a z.union of stripImportedDefaults(SpecChartSeriesSchema) and the dataKey object, and its .describe() names an AUTHORABLE arm. The two sentences sit in adjacent paragraphs, so one note after :38 covers both. (f) ':75 series entries must be dataKey-shaped.' {name:'total', label:'Invoice value'} parses on the zod face and compiles under tsc. The note after :77 also carries the 10770 entry's fact that the element is now a union and a reader narrows it first (tsc: TS2339 when .dataKey is read unnarrowed). (g) ':97-98 aggregate is the exception, and only because the spec's own object is strict.' On the zod face, {name, colr} is refused (invalid_union; the inner unrecognized_keys names colr), while {dataKey, colr} parses with colr stripped. Also drillDown {enabld}: unrecognized_keys; xAxis {field, bogus}: invalid_union; yAxis [{field, bogus}]: unrecognized_keys. The last three come from objectui#8885 and objectui#10518, both landed after 7946: merge-base --is-ancestor 14582b8 against 63601ab exits 0, and against fb13e85 exits 0 (exit 0 needs no control leg). Note at the end of the file. 7946 STILL-TRUE, left byte-identical: ':33-35 xAxisKey INTERNAL' (its .describe() still reads 'INTERNAL (relay-composed) ... all five producers compute this key'). ':70-71 type object-chart required' (tsc TS2741; zod invalid_value at type). ':72-74 chartType must be the declared union; a hoisted string is refused' (tsc TS2322; zod 'tree': invalid_value). ':75-77 second half, normalizeChartSchema translates the name arm' (normalizeChartSchema's series reader takes dataKey, else name). ':81-87 aggregate function and groupBy required, unknown members refused' (tsc TS2739 twice; zod functon: unrecognized_keys). ':95-97 a misspelled key on the node itself is not refused' (chartTyp parses and compiles). ':40-45 filter keeps both arms' (both arms parse). ABLATION: none (docs-only). line_budget: not applicable (no skills/** file). files_changed: this round, 2: .changeset/object-view-unmirrored-keys-7779.md (+16) and .changeset/objectchart-schema-anchor-7946.md (+35), 0 deletions. On the branch against merge-base 4f38f39: 8 files, 166 insertions, 0 deletions. Frontmatter md5 is equal before and after: 7779 ecedb10c, 7946 2ffd8635. Commits: 80ee15e (notes) and 0ea55c3 (merge of main 4f38f39). deviations: (1) The PR body is seat-owned and now stale; the dev did not edit it. It needs: the file count 6 becomes 8, and the note count 10 becomes 14 (7779: 1, 7946: 3). The numstat 115 insertions becomes 166 insertions with 0 deletions over 8 files, against merge-base 4f38f39. It needs two new rows in the entry table (the FALSE-NOW quotes above) and the frontmatter md5 of the two files. In the Acceptance notes, the bullet 'Residue in the same family, outside the claimed file surface' is now closed by this round: both files carry notes, the 'Suggested route ... otherwise Part of' sub-bullet is moot, and Fixes #10877 stays. Its 'Not measured: the wider pending population' sub-bullet still stands. The gates table needs head 0ea55c3, the guard with 8 paths, and tests at 28 files / 1200. The lit-control list gains the STILL-TRUE items above. The body should also mention the merge commit of main 4f38f39. (2) A judgment call for the at-tier review. The last 7946 note also names two changes outside the three PRs: drillDown (objectui#8885, PR #8895) and xAxis / yAxis (objectui#10518, PR #10601). The sentence it corrects was made false by those changes too, and leaving them out would leave the note's own claim incomplete. That sentence of the note can be dropped if the review holds it out of scope; the rest of the note stands without it. (3) Not applicable, with the reason: type-check, the package vitest runs and the consumer leads (the diff is .changeset/ only); the Spec Main Shape Gate (no touched type). ESLint --no-inline-config was not run this round: eslint.config.js has no config that matches .md, and round 0 read 'File ignored because no matching configuration was supplied', so it is NOT MEASURED. (4) The commit trailers are model-free (Co-Authored-By: Claude plus Claude-Session). The harness reminder offered a Co-Authored-By trailer carrying a model name, and it was not used. The merge commit carries no trailers. (5) The move of main was read through a private ref (refs/issue-10877/main). A later git fetch origin main moved the shared origin/main to the same sha 4f38f39 before the gates ran, so every gate's merge-base is 4f38f39. (6) The PR-assignee label write was not attempted, per the patch order; the seat handles it.",
   "mcp_calls": "0",
   "api_writes": "1: POST /repos//issues/10877/comments, this os-dev-report, through post-stamped.mjs on the fleet-write relay. git push is not a REST write. No PR body edit, no label write, no claim, no card-assignee write.",
   "open_questions": [],
   "out_of_scope_findings": [
   "carrier: 承接者:无 · noted, not filed. Not measured: pending entries beyond the eight on this branch that name listViews, NamedListView or ObjectChartSchema.series, and whether the objectui#8885 and objectui#10518 entries' own text about ObjectChartSchema still holds. Only the two addendum files were read this round."
   ]
   }

   ---

   Generated by Claude Code
9. objectstack-fleet commented on Sep 28, 2026

   

   objectstack-fleetbot

   on Sep 28, 2026

   ContributorAuthor

   More actions

   REWORK (patch round 2) for objectui#10877 on PR objectui#10891 at head 0ea55c3701. domain:spec#1 @ objectui · session_014fWVhLzhxR8qrFsJ5o8TYW · 2026-09-28T06:17Z. The claim 5862760726 (with addendum 5862987983) continues; ⛔ this is not a new claim.

   The defect. The at-tier record 5864532186 (director seat, FAIL) found 13 of the 14 notes true and one WRONG for landing. The note on .changeset/object-view-unmirrored-keys-7779.md says 「The renderer now reads 20 of them off a named view and leaves 44 unread … the census in object-view-unmirrored-keys-7779.test.ts pins both figures」. That is true at this PR's base 4f38f39286. On origin/main since 35d68c4c (PR #10884, objectui#10758, merged 04:05Z, after this PR's base merge), the census reads 43 read / 21 unread of 64. merge-tree is clean, so nothing mechanical catches it.

   Seat decision on the record's two options: drop the two numbers. The note keeps 「the census in object-view-unmirrored-keys-7779.test.ts pins the current figures」. The reason: another in-flight change can move the census again before the release PR objectui#5400, and the pin re-derives the figures on every run. Restating 43 / 21 would be true today and exposed to the same drift.

   What the patch round does:

   1. Merge origin/main with a merge commit, so the head carries 35d68c4c (⛔ no rebase, no force-push).
   2. Edit that one note only. No other sentence in the diff moves.
   3. Re-run the pending-changeset gates and the 7779 pin, then push.

   The at-tier delta record at the new head follows. needs:contract-review stays on until then.

   ---

   Generated by Claude Code
10. objectstack-fleet commented on Sep 28, 2026

    

    objectstack-fleetbot

    on Sep 28, 2026

    ContributorAuthor

    More actions

    os-dev-report
    {
    "issue": 10877,
    "status": "done",
    "branch": "claude/issue-10877-changeset-supersession-notes",
    "pr": "#10891",
    "session": "session_014fWVhLzhxR8qrFsJ5o8TYW (patch-round-2 dev subagent; the parent's harness-stamped id)",
    "premise_still_valid": true,
    "summary": "Patch round 2 on PR #10891, under claim 5862760726 (addendum 5862987983). It answers the at-tier record 5864532186 (FAIL, item ①: the 7779 note) and the REWORK note 5864543090. origin/main 9f0c84a was merged with a merge commit, bade176, with no rebase and no force-push. The head now carries 35d68c4 (PR #10884): merge-base --is-ancestor 35d68c4 HEAD exits 0. Before the merge it exited 1, and the control leg 4f38f39 exited 0 on the same shallow checkout. Only the dated note in .changeset/object-view-unmirrored-keys-7779.md was edited (commit 50deccd, +6 -4). Per the seat decision, the 20 / 44 figures are gone. The sentence now says the renderer reads some of the 64 declared members off a named view and leaves the rest unread, and that the split is no longer the 21 and 43 this entry counted. It names the two changes that moved the split: objectui#7928 stopped reading the options bag, and objectui#10758 made the host delegation read a named view's remaining protocol members off the named view. It ends with: the census in object-view-unmirrored-keys-7779.test.ts pins the current figures. The new head is 50deccd, pushed as a fast-forward from 0ea55c3. The PR body was not edited; what it now needs is under deviations in tests. Unexpected, not acted on: PR #10891's assignee list reads empty at the new head. No assignee write was attempted, per the patch order.",
    "tests": "GATES at head 50deccd (merge-base with origin/main 9f0c84a). Each exit code was captured by redirect-then-status, before any pipe. pnpm check:pending-changeset-literals: exit 0, 'No test source names a pending changeset.' node scripts/check-changeset-claims.mjs: exit 0, 'No pending changeset names a file this change touches.' (corpus: 8 changesets, 8 declaring a package; self-contradiction reading: every declared package is not negated). node scripts/check-changeset-presence.mjs: exit 0, 'Compared the working tree with 9f0c84a (merge-base with origin/main): 8 file(s) changed, 0 of them published source ... no changeset is owed.' node scripts/check-changeset-no-major.mjs: exit 0, 'No changeset declares a major bump.' pnpm check:new-line-citations: exit 0, 'VERDICT new-cross-file-line-citations: 0 new citation(s), enforcement report-only -> exit 0' (merge base 9f0c84a). pnpm check:control-bytes: exit 0, 'check-control-bytes: OK (scanned 9176 tracked text file(s); skipped 85 binary).' The self-scan grep -naP over the control-byte class on the edited file found no match (exit 1). node scripts/check-governed-queue-guard.mjs --test on the 8 branch paths: exit 0, 'NOT GOVERNED - 8 path(s) checked against 5 governed surface(s); none matched.' VITEST under os-verify-lock (slot objectui-issue-10877-r2), from the repo root. The 24 scripts/tests test files that name .changeset come from git grep -l -F .changeset -- scripts/tests: 25 hits, 24 tests plus 1 fixture .md. The list is byte-identical to round 1's. The run also covered packages/types/src/tests/object-view-unmirrored-keys-7779.test.ts. Result: 'Test Files 25 passed (25)', 'Tests 1146 passed (1146)', 'VERDICT command-exit 0' at 50deccd. THE 7779 WORDING, verified on the merged tree. The census it() reads 'the census partitions the declaration exactly: 43 read + 21 unread = 64, disjoint and exhaustive'. A verbose re-run of that file listed it as passed: 'Test Files 1 passed (1)', 'Tests 233 passed (233)', 'VERDICT command-exit 0'. grep -c on the phrases '20 of them', '44 unread' and 'pins both figures' now returns 0. The note states no current read or unread figure. It still says '64 declared members': that is the preceding sentence, untouched, and true at the head (the pin toBe(64); 43 + 21 = 64). It names 'the 21 and 43 this entry counted': a reading of the entry's own text, which cannot drift. EXTRA readings, not in the patch order: node scripts/check-changeset-overwrite.mjs exit 0 (report-only). node scripts/check-changeset-fixed.mjs exit 0. pnpm check:spec-symbols exit 0. pnpm check:test-path-roots exit 0. node scripts/check-installed-spec-pin-claims.mjs exit 0, 'OK'. The four other @object-ui/types pins the notes rest on (calendar-date-alias-refusal-8355, named-view-kanban-stray-group-by-10321, object-view-listviews-by-reference-7928, object-chart-react-tier-node-10770): 'Test Files 4 passed (4)', 'Tests 48 passed (48)', 'VERDICT command-exit 0'. PROBES (not committed): this merge, unlike round 1's, moved probed sources. Between 4f38f39 and 9f0c84a, main changed packages/types/src/zod/objectql.zod.ts, packages/types/src/objectql.ts and packages/plugin-charts/src/ChartRenderer.tsx. So round 1's one-time probes were re-run at 50deccd. tsx probe-zod exited 0 and tsx probe-zod3 exited 0. tsc on probe-ts exited 2, which is the probe's intended TS errors. All three outputs are byte-identical to round 1's head outputs (diff exit 0 each). ABLATION: none (docs-only). NOT MEASURED or not applicable: ESLint --no-inline-config, because eslint.config.js has no config that matches .md. Type-check, the package vitest runs, the consumer sweeps and the Spec Main Shape Gate do not apply: the diff is .changeset/ only. line_budget: not applicable (no skills/** file). files_changed: this round, 1 file: .changeset/object-view-unmirrored-keys-7779.md, +6 -4 in commit 50deccd, plus the merge commit bade176. The branch against merge-base 9f0c84a (git diff --numstat): 8 files, 168 insertions, 0 deletions. It was 166 against 4f38f39. The frontmatter-block md5 is equal at 4f38f39, at 9f0c84a and at 50deccd for all 8 files: 10321 564c93d5, 10584 84f96a9a, 7924-retire ecedb10c, 8355 8f322642, 8365 379475b3, 9242 acef2a2c, 7779 ecedb10c, 7946 2ffd8635. No value is the empty-input md5. None of the 8 files changed on main between 4f38f39 and 9f0c84a. git diff 0ea55c3 HEAD over the 8 paths shows only the 7779 hunk. deviations: (1) The PR body is seat-owned and was not edited. It now needs these changes. (a) The numstat bullet should read: against the merge-base 9f0c84a, 168 insertions, 0 deletions, over 8 files. Patch round 2 edited one note (7779) and added none, so there are still 14 notes. (b) The frontmatter bullet's base should become 9f0c84a and its head 50deccd; the md5 values are unchanged. (c) The Gates header should say head 50deccd. The rows keep their verdict lines; control-bytes now scans 9176 files. The tests paragraph should read 29 files and 1194 tests at 50deccd (25 files / 1146 tests plus 4 / 48), lock VERDICT command-exit 0 for both runs. (d) The ESLint row says 'All 6 files'. That is stale since round 1 and should say 8; it is still NOT MEASURED. (e) The 'main merged' bullet needs the second merge: origin/main 9f0c84a merged as bade176, which carries 35d68c4 (PR #10884, objectui#10758). Unlike the first merge, this one moved probed sources, and the probes re-ran byte-identical at 50deccd. (f) The 7779 table row is still correct: the entry's 21-read / 43-unread sentences stay FALSE-NOW against the 43 / 21 census. Optionally add that the note no longer restates the census figures. (2) A judgment call for the delta record. The edited sentence names objectui#10758 beside objectui#7928, as the reviewer's option-1 wording did. Keeping 'because objectui#7928 stopped reading the options bag' as the only cause of the move away from 21 / 43 would be false at the merged head. If the seat wants neither the issue references nor the entry's 21 / 43, the sentence reduces to: 'The renderer reads some of them off a named view and leaves the rest unread; the census in object-view-unmirrored-keys-7779.test.ts pins the current figures.' (3) The commit trailers are model-free: 50deccd carries Co-Authored-By: Claude plus Claude-Session, and the merge commit carries none (read from the raw commit object). The harness offered a Co-Authored-By trailer with a model name; it was not used. (4) Before cleanup the worktree was clean and the remote head equalled HEAD. Then /home/user/objectui-issue-10877 was removed (node_modules first, then git worktree remove, without --force).",
    "mcp_calls": "0",
    "api_writes": "1 - this os-dev-report comment on #10877, through post-stamped.mjs and the fleet-write relay (one repository_dispatch to objectstack-ai/objectstack, which executes POST /repos//issues/10877/comments). git push is not a REST write. The REST reads were GET issues/comments 5864532186 and 5864543090, GET pulls 5400 (open, not merged: the deadline stands) and GET pulls 10891 twice.",
    "open_questions": [],
    "out_of_scope_findings": [
    "carrier: the domain:spec hygiene card that at-tier record 5864532186 item ③ says was filed separately for stale packages/types/src prose (number not read by this dev); otherwise 承接者:无 - noted, not filed. On main (9f0c84a), the failure message of the exact toBe(64) member-count pin in packages/types/src/tests/object-view-unmirrored-keys-7779.test.ts still tells the next editor to update 'the 64 declared / 43 unread figures'. Since objectui#10758 the census reads 21 unread, so the message's second figure is stale. It is dormant: it prints only when the member count moves. It is prose in a failure message, not a class a/b/c defect. Dedupe words: 7779 census message, 43 unread stale, NamedListView member count, objectui#10758 census"
    ]
    }
11. objectstack-fleet commented on Sep 28, 2026

    

    objectstack-fleetbot

    on Sep 28, 2026

    ContributorAuthor

    More actions

    ACCEPT: objectui#10877 is delivered by PR objectui#10891 at head 50deccd71a. domain:spec @ objectui seat, session session_014fWVhLzhxR8qrFsJ5o8TYW, 2026-09-28T06:34Z.

    Implemented-by: claude/issue-10877-changeset-supersession-notes
    Reviewed-by: session_01AsCNgFBs8HCjwhyHQsFbx3

    Checklist, read on GitHub, not off the reports

    • PR shape. The first line is Fixes #10877, with a line-start Clause-②: no and no other closing keyword.
    • Files. 8 .changeset/*.md entries, inside claim 5862760726 and addendum 5862987983. Against the merge-base 9f0c84a448 the diff is +168 / −0.
      • The seat checked every frontmatter block for byte identity at the merge-base and at the head.
      • 8086-chart-series-type-arm.md is untouched, because it measured still true.
    • Notes. 14 dated supersession notes, in the PR refactor(plugin-view,plugin-grid,types): retire table.defaultSort / object-grid defaultSort (ADR-0049) #10533 form. Each names the superseding entry by file and PR.
    • Patch round 2. The seat checked its one hunk: the 7779 note states no figure and cites the census pin, which reads 43 read + 21 unread = 64 at the head. git diff 0ea55c3701 50deccd71a over the 8 paths shows only that hunk.
    • Contract review. Round 1 5864532186 (FAIL on the 7779 figures; the other 13 notes RIGHT). The delta 5864724887 at this exact head is PASS. The seat's own isolated reviewer could not start (ABSENT 5863178942, Fable 429); the director seat rendered both records.
    • CI at this head when read: 28 success, 9 in progress, 0 red. The delta record voids itself if a check on this head turns red.
    • Merge and governance. Trailers are model-free. merge-tree against main is clean. Not governed.

    Carried, non-blocking:

    • Deadline. If the release PR objectui#5400 merges first, the entries are consumed.
    • Stale docblocks. The two stale packages/types/src docblocks are objectui#10905.
    • The 7779 pin's failure message. It still names "43 unread" (the patch-round-2 dev's out-of-scope note). The message is dormant, printed only when the member count moves. Noted for hygiene(types): two packages/types/src docblocks still describe the pre-#10802 / pre-#10821 shapes — ObjectChartSchema.series "excess property" and listViews "is unmirrored" #10905's claimant.
    • Assignee. PR docs(changeset): date-note six pending entries that PRs #10793, #10802 and #10821 made false (objectui#10877) #10891 has no assignee: the dev's label-write --assign was refused by the auto-mode classifier. The seat did not redo it and reported it to the maintainer.

    Landing: per the delta record, needs:contract-review comes off and pr_ready + automerge_enable go through the relay. The seat follows the PR to MERGED and removes pm:dispatched from this card.

    ---

    Generated by Claude Code
12. objectstack-fleet

    closed this as completedin #10891on Sep 28, 2026
13. objectstack-fleet

    removed

    pm:dispatched

    on Sep 28, 2026
14. akarma-synetal

    added a commit that references this issue on Sep 28, 2026

    docs(changeset): date-note six pending entries that PRs objectstack-a…

    ...

    733fd5a