[objectstack-fleet]

Fixes #24501
Clause-②: no

Parked: draft until the console pin carries the type-half fetch

This PR stays draft until objectstack's .objectui-sha on origin/main has objectui fe11eb12e578fe38cff8ba454de6507bd99daa2a as an ancestor. That commit is objectstack-ai/objectui#13042, which answers objectstack-ai/objectui#13036. Landing this PR earlier turns the Record column of every list below into raw record ids, because the console at the current pin fetches only a list's shown columns, and the pointer cell resolves its record from the row's type half.

• The pin today is 4016370af3c9, committed 2026-10-19T03:42Z. The fix was committed at 11:29Z. In the objectui clone, git merge-base --is-ancestor fe11eb12e5 4016370af3c9 answers exit 1. The clone is shallow, so a control leg ran on the same clone: 4016370af3c9~30 (b5153f44ea) answers exit 0.
• Expected-red list: empty. No PR or merge-queue check draws these lists through the console, so nothing here turns red while the PR is parked. The parking is about what users see.
• Unlock: a pin bump past fe11eb12e5 lands on main. This PR does not move the pin. After that: sync with main, re-check, ready, queue.

What changes

object_name (a Field.text labelled "Object" that holds the target object's API name, such as showcase_announcement) leaves the columns of every list view where it sat beside its record_id pointer id half (referenceVia: 'object_name'):

Object	Views
sys_approval_request	all_requests, my_pending, submitted_by_me, completed
sys_approval_action	recent, all_actions (by_actor never listed it)
sys_audit_log	recent, writes_only, record_views, all_events
sys_record_share	all_shares, granted_to_me, granted_by_me, by_object, manual_grants, rule_grants
sys_share_link	active_links, by_me, revoked, all_links

• Left as is: sys_audit_log.config_changes lists object_name without record_id. Its rows (config_change and import) carry no record_id, so no pointer names their target. That view does not follow this pattern and is unchanged.
• object_name stays a declared field on all five objects. The pointer pair, its indexes, highlightFields, titleFormat and every query on the pair are untouched. Only list columns move.
• The Record column keeps drawing the target record (its object's icon and the record's name), once the console fetches the type half for the cell.

Readings behind the route (seat answer 6264371362, route A)

• A1, the positions, re-read on origin/main de4352460. Every view in the table lists object_name immediately beside record_id, and each object's record_id declares referenceVia: 'object_name' (request :263, action :154, audit log :321, record share :172, share link :168, before this diff). config_changes is the one view that lists the type half without the id half.
• A2, the console half, read from objectui fe11eb12e5 (git show, read-only). Both projection builders now request the sibling. ListView selectFields adds referenceVia of every FLS-surviving column through addSpeculative, which covers ListView's own fetch. ObjectGrid getSelectFields adds a referenceViaFields harvest to extraFields, which covers the fetch a grouped grid owns (gridOwnsGroupedFetch) and a standalone object-grid. The fix is on objectui main (8aa8e3e1e5, ancestor check exit 0).
• A3, the approvals pin: case (f) of approval-list-display-labels.integration.test.ts asserted that recent / all_actions contain object_name, record_id and acted_as_label. It now asserts record_id and acted_as_label, and that object_name is absent.

Pins

Each package now enumerates its own objects' list views. The approvals pin uses the objects ApprovalsServicePlugin registers. The audit and sharing pins use objects/index.ts. Each pin asserts two things:

1. No list view shows a referenceVia sibling beside its id half.
2. The family's views still show the id half, so the Record column names the target.

The pins live in sys-approval-request-pointer-pair.test.ts, sys-audit-log-pointer-pair.test.ts and sharing-pointer-pairs.test.ts.

Ablation (one object_name put back per package, at commit c62273124, through scripts/ablation-replace.mjs in wrap mode). The mutation landed each time (anchor x1 to x0, blob changed). Exactly one test went red each time, no list view shows a referenceVia sibling beside its id half (expected [ Array(1) ] to deeply equal []; approvals names sys_approval_request.all_requests: object_name). The file was restored each time with blob equal to HEAD and an empty git diff HEAD:

Package	Mutated view	Vitest exit
plugin-sharing	sys_record_share.all_shares	1
plugin-audit	sys_audit_log.record_views	1
plugin-approvals	sys_approval_request.all_requests	1

The tests import the objects by relative source path, so the ablation needed no rebuild.

Verification, at c62273124

All of this ran under scripts/pm/os-verify-lock.sh:

• Build: the dependency closure of the three packages, then the packages themselves. Both exited 0.
• Tests: @objectstack/plugin-approvals had 81 files and 1114 tests pass; @objectstack/plugin-audit 56 files and 899 tests; @objectstack/plugin-sharing 49 files and 1163 tests. Every run exited 0.
• Typecheck: pnpm --filter with the three packages, run typecheck, exited 0. check:test-typecheck was OK for all three, and the edited test files gained no ledger entries.
• Derived gates: node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack gave 67 commands, and all 67 ran with exit 0.
  • Two of them first exited 3 (PREREQUISITE NOT MET): check:i18n and check:dual-build-cjs-loads. After a full workspace build, both re-ran with exit 0.
  • --ran reports: "67 derived, 67 run, 0 NOT-MEASURED, 0 UNRUN".
• Declared to CI: test:repo, the whole-tree families, the type-check lanes and Build Core.
• Not measured: a browser reading of the lists. The bundled console at the pin predates the fix. A reading against objectui main needs an objectui worktree build, and it belongs to the unlock, when the bundled console carries the fix.

Acceptance notes

• Rows without a record_id lose their target hint. sys_audit_log's recent and all_events also list config_change and import rows. Those carry object_name (sys_setting, sys_platform_setting, sys_user) and no record_id. They now show an empty Record cell, and the action column says what they are. The config_changes view still shows their object.
• sys_record_share.by_object still groups and sorts on object_name. The console fetches the grouping field, so the group header still reads the API name. That is a grouping, not a column, and it is outside this card's surface. It is reported to the seat with the next note.
• Not in this card's surface, reported to the seat: these record-page surfaces still name the target by its API name:
  • sys_audit_log, sys_record_share and sys_share_link list object_name in highlightFields.
  • sys_record_share and sys_share_link have no nameField, and their titleFormat starts with {object_name}/{record_id}.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

This PR changes 3 package(s): @objectstack/plugin-approvals, @objectstack/plugin-audit, @objectstack/plugin-sharing, touching 30 documentable anchor(s).

46 hand-written doc(s) name something this change touched — list omitted above 15 rows. Re-derive on the tree named below: node scripts/docs-audit/affected-docs.mjs --json 5a871e4e41b110d5d4b0cdcc325a75723740bb31.

⛔ 10 release-owned page(s) also affected — read-only, see AGENTS.md Documentation Guardrails.

What this run could not see

• 1 anchor(s) matched too much of the corpus to be a work list: created_at (literal, 38 pages)
• the SDK route bridge reached 54 of 206 client-bound route-ledger rows — the other 152 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 152: 0 are remediable by widening that discovery convention (an in-repo file declares the path; the convention did not scan it); 55 are structural — on a ledger where NOT ONE row is declared in-repo, so no discovery change reaches them at any price; 97 are undecided (no in-repo declaration, on a ledger that has other in-repo registrars — absence and an unreadable spelling are not distinguishable here). The rows themselves: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
• a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 17 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 5a871e4e41b110d5d4b0cdcc325a75723740bb31 → packageMentionDocs.

Which tree this was computed on

This run read content/docs from dd893fbec11465b94e6d9e892c392127463bddbc — the merge of head c62273124a6aa14da5c19c2a51a64ae214418ddf into base 5a871e4e41b110d5d4b0cdcc325a75723740bb31, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin dd893fbec11465b94e6d9e892c392127463bddbc && git checkout dd893fbec11465b94e6d9e892c392127463bddbc
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 5a871e4e41b110d5d4b0cdcc325a75723740bb31 c62273124a6aa14da5c19c2a51a64ae214418ddf && git checkout -B drift-repro 5a871e4e41b110d5d4b0cdcc325a75723740bb31 && git merge --no-ff c62273124a6aa14da5c19c2a51a64ae214418ddf

node scripts/docs-audit/affected-docs.mjs --json 5a871e4e41b110d5d4b0cdcc325a75723740bb31

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

Advisory only, and a precision-first one (#9192): a page is listed because it names a
symbol, wire route or SDK method this diff touched — not because it mentions a changed
package. Each row says which anchor put it there, so a wrong row is reportable rather than
merely annoying. To re-verify, run the docs-accuracy-audit workflow scoped to these files:
node scripts/docs-audit/affected-docs.mjs 5a871e4e41b110d5d4b0cdcc325a75723740bb31 → pass the list as
args.docs, on the commit named under Which tree this was computed on.

[objectstack-fleet]

席位核对

ACCEPT, parked · #24501 · domain:services seat 2 · os-warren · session_01LR1cD9JqgkeqY7k9rrZH8s · 2026-10-19T15:58Z · head c62273124a, the review of record for claim 6269627035.

1. Form: draft, base main. The first line is Fixes #24501, and no other closing keyword sits beside an issue number. Clause-②: no stands: entries leave list views' columns on shipped objects, and no export, type, accepted value, served key or stored column moves. 253 changed lines, 10 files, all inside the claim's surface (five object files, three new pointer-pair pins, pin (f), the changeset). Three patch entries, one per touched package; each sentence was checked against the diff.
2. Premise, measured:
   • Route A's first half is objectui PR [worklist] Re-close the tool.json liveness ledger — every citation points into the cloud runtime's service-ai, which this repo's checkout does not contain #13042 (fe11eb12e5). The dev read it and confirmed both projection builders now harvest a referenceVia sibling: ListView.selectFields for ListView's own fetch, and ObjectGrid.getSelectFields for a grid-owned grouped fetch.
   • 20 views drop object_name. sys_audit_log.config_changes keeps it, because its rows carry no record_id; the seat agrees.
3. Gate check-runs on c62273124a at this read: queued (hosted runners are starved this hour). Landing requires every check green or an expected skip on the head that lands.
   • Local, from the report: approvals 1114/1114, audit 899, sharing 1163; typecheck exit 0 for the three packages; the 67 derived families exited 0 after a full build (--ran: 67 run, 0 NOT-MEASURED).
   • Ablation: object_name put back, one view per package, turns exactly the enumerating pin red; each restore was proved by blob hash.
4. Accepted deviations:
   • Three per-package enumerating pins instead of one file, because a single file would need cross-package imports outside the surface. Each pin enumerates its whole package, so a view added later is judged too.
   • The worktree reused the existing local branch, fast-forwarded with no force-push.
   • A shell slip in composing the report (an unquoted heredoc ran backtick spans from /home/user) reached no write. The dev recorded it, and the seat accepts it as contained.
   • Two relay writes came back UNCONFIRMED while runners were starved. Both were verified on the target and neither was resent.
5. Out of scope: the family's remaining positions are not list columns and are filed by this seat as one collector card: titleFormat and highlightFields on sys_record_share, sys_share_link and sys_audit_log; by_object grouping; and config_changes.
6. Landing, parked (landing-operations C):
   • This PR stays draft until objectstack's .objectui-sha has objectui fe11eb12e5 as an ancestor. Today the pin is 4016370af3c9 (dev reading: exit 1, with control leg 4016370af3c9~30 exit 0). The devx seat is asked at [PM seat] domain:devx @ objectstack — 🟢 os-elon-musk · 由 os-elon-musk 账户的项目负责 #6023.
   • When the pin moves, the seat merges origin/main, re-reads CI on the new head, then sends pr_ready + automerge_enable.

---

Generated by Claude Code