Repository navigation
app-shell: count an object's served listViews as served views — objectstack now translates them server-side, so the console's second viewLabel pass is redundant #11336
Description
Activity
Triage: first grade —
enhancement·priority:p3·domain:ui·area:i18n·pm:on-hold. Count embeddedlistViewsas served, at the bump that carries the server's translationTriage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) ·
session_01AavokzJ5DndAwitDXvKy4U· 2026-10-01T06:14Z. ⛔ Not a claim, ⛔ not a dispatch.Restart-when: an
@objectstack/specrelease that carriesc27404f0a9is published on npm, and objectui resolves it (17.5.0 does not; read at this write).Why p3. The console's second pass is redundant once the server translates. It is not a wrong label today, because the console path still translates.
Routing.
app-shellisdomain:ui.Direction. It is the card's own scope, confirmed:
isServedViewanswers true for an embedded view key, the tab and the crumb draw the served label as given, and the stale docblock bullet is corrected. Pins: an embedded view'szh-CNtab label is the served string; a view with no catalogue entry keeps its authored label.
Generated by Claude Code
Hold released:
pm:on-hold→pm:queue. TheRestart-when:condition holds on the installed@objectstack/spec17.6.0domain:uiexecution seat 1 @ objectui (seat post #5560) ·session_015W8GBu6sBiqus2L2xjMsAL(os-steve) · 2026-10-04T14:07Z. ⛔ Not a claim, ⛔ not a dispatch.The condition is the
Restart-when:line of the latest transition comment,5925839788.Readings at this write:
npm view @objectstack/spec versionanswers17.6.0.- objectui
origin/mainfd060f0767:pnpm-lock.yamlresolves@objectstack/speconly at17.6.0(2 entries, no other version). - REST
compare/c27404f0a9...617f25f8a(617f25f8ais the commit tag@objectstack/spec@17.6.0points at):ahead 124, behind 0, soc27404f0a9is in the 17.6.0 release. - Install-face probe, with
@objectstack/specinstalled from npm into a scratch directory at each version. 17.5.0 is the control, and shows the probe can fire:system.translateObject(doc, bundle, { locale: 'zh-CN' })over an object whoselistViews.mine.labelisMy Links, with a bundle carryingobjects.sys_account._views.mine.label: 我的链接: 17.5.0 servesMy Links; 17.6.0 serves我的链接.- Control in both versions: a sibling view with no catalog entry keeps its authored label (
Other).
- Release double-check ②: the card's timeline shows no merged PR referencing it after
5925839788.
The card returns to the
domain:uiqueue under triage's direction in5925839788.
Generated by Claude Code
Claim: PM loop round 1
Session:session_015W8GBu6sBiqus2L2xjMsAL
Account:os-steve
Branch:claude/issue-11336-served-listviews
Worktree:objectui-issue-11336
Domain:domain:ui
Seat:domain:ui#1
File surface:packages/app-shell/src/hooks/useServedViewItems.ts(isServedViewand its docblock), its label sitespackages/app-shell/src/views/ObjectView.tsx(viewTabLabel) andpackages/app-shell/src/layout/AppHeader.tsx(the crumb), the tests beside them, and.changeset/11336-*.md(stop on breach; explain in the report)
Container & model:M,mode:subagent,model: default tier(dispatch-gates --tier --repo objectstack-ai/objectui: no path-derived mandate; floor sonnet · default opus · ceiling fable)
Clause-②: no
Thread-read: 5980864937
Serial constraints cleared: none, read 2026-10-04T14:15Z: the four open objectui PRs (objectui#11614, #11607, #11600 the release PR, #11069) were read in full (29, 202, 101 and 15 files; the release PR touches only changesets, CHANGELOGs, manifests andQUICK_REFERENCE.md) and the threepm:dispatchedcards (objectui#11610, #10202, #8347) by their claims. objectui#11607 editspackages/plugin-view/src/ObjectView.tsx, a different file from this surface'spackages/app-shell/src/views/ObjectView.tsx; objectui#10202 declarespackages/app-shell/src/views/metadata-admin/**, disjoint.Why
Clause-②: no:useServedViewItemsandisServedVieware not reachable from@object-ui/app-shell's entry (0 hits insrc/index.tsandsrc/hooks/index.tsatfd060f0767; controluseRecentItems2 hits insrc/index.ts). The change moves which labels are drawn as served; it changes no accept set. The dispatch fences new exports and new i18n keys.
Generated by Claude Code
os-dev-report
{
"issue": 11336,
"status": "done",
"branch": "claude/issue-11336-served-listviews",
"pr": "#11620",
"session": "session_015W8GBu6sBiqus2L2xjMsAL — subagent; the parent seat's harness-stamped id",
"premise_still_valid": true,
"summary": "isServedView(served, objectName, viewId) answers true for a /meta/view document under that name, as before, and now also when viewId is a key of the served /meta/object document's own listViews. That document is read pre-merge from useMetadata().getItemsByType('object'), only the canonical listViews key is read, and the entry must be a record. The predicate is in packages/app-shell/src/hooks/useServedViewItems.ts, and the stale docblock bullet is corrected there. useServedViewItems() now returns both reads as one record held in a module-level WeakMap keyed on the two cache arrays (AGENTS.md #10). viewTabLabel and viewOriginLabel (ObjectView.tsx) and the AppHeader crumb pass the object name, so the tab and the crumb draw an embedded view's served label as given. The Zone 1 pins are on both surfaces: the zh-CN served string, and a no-catalogue view keeping its authored label. A published-edit cell discriminates, and a control shows a stack container's expansion is still bundle-named. Changeset: @object-ui/app-shell patch. Zone 2 assumption 3 was measured true. The label sites' objects come from useMetadata().objects, which merges readType('object') and readType('view'). getItemsByType('object') is the cache's own /meta/object array, mutated only by normalizeSchemaReferenceKeys. Embedded keys pass through the merge verbatim and win collisions. Merged origin/main 2abec3a (it contains b403bb3 / objectui#11607) at 444300b with no conflicts and no overlap with this surface, then reran type-check and the full app-shell suite on final head 0a2c2af.",
"tests": "All on head 0a2c2af (merged) unless noted. (1) pnpm exec vitest run packages/app-shell/ --maxWorkers=2 --reporter=dot, under os-verify-lock: 'Test Files 1000 passed | 1 skipped (1001)', 'Tests 9930 passed | 9 skipped (9939)', VERDICT command-exit 0. (2) pnpm --filter @object-ui/app-shell type-check (the script is spelled type-check; it echoed 'tsc --noEmit && tsc -p tsconfig.test.json'): exit 0, on 444300b and again on 0a2c2af. Run after building the @object-ui/app-shell^... closure (28 packages, exit 0). tsc -p tsconfig.test.json --noEmit --listFiles lists all 3 touched test files (count 3) and the hook source (count 1). (3) The 3 touched test files at 5b212d6, before the merge: 'Test Files 3 passed (3)', 'Tests 35 passed (35)'. (4) Reverse verification at 5b212d6 (implementation committed first), via objectstack scripts/ablation-replace.mjs. Anchor ' if (!objectName) return false;' hit x1 and went x1 to x0; replacement 'return false; // ablation-11336' went x0 to x1; blob 3eb10f0c9cff to a98e6cf57893. Result: 'Tests 7 failed | 28 passed (35)', exactly the predicted reds. They are the 2 ObjectView and 2 AppHeader 'published edit the object read served' cells (en, zh-CN) plus 3 unit cells (embedded key answers; key answers only under its own object; identity record answers after a refetch). Restore proof: blob after restore 3eb10f0c9cff == HEAD, git diff HEAD empty, marker grep count 0. Direction note: the zh-CN served-string cells are green on both sides, because the catalogue and the server say the same string; the discriminating cell is the published edit. (5) Package lint: pnpm --filter @object-ui/app-shell lint ('eslint .') exit 0, '3486 problems (0 errors, 3486 warnings)'. Per-file eslint --format json, base fd060f0 vs now: ObjectView.tsx 174w/174w, AppHeader.tsx 22w/22w, the ObjectView #11295 suite 3w/3w, the AppHeader #11295 suite 17w/17w, the hook 0w/0w, the new unit file 0w. (6) Root gates on 0a2c2af, each exit 0: check-changeset-presence ('6 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)'), check-changeset-no-major, check:changeset-claims, check:pending-changeset-literals, check:new-line-citations ('0 new citation(s)'), check:control-bytes, check:vi-mock-specifiers, check:vi-mock-inherit, check:vi-mock-override-shape, check:test-path-roots, check:unreferenced-sources. (7) Install-face probe of the resolved @objectstack/spec 17.6.0 (system.translateObject, zh-CN): listViews.mine gives 我的链接; a sibling with no catalogue entry gives Other; an org edit with packagedBase is served as edited (the catalogue loses); legacy list_views.mine gives My Links, untranslated. NOT MEASURED: the pre-merge full app-shell run, stopped by me with exit 143 so the run would be on the merged head; superseded by (1).",
"mcp_calls": "0 — no MCP GitHub calls",
"api_writes": "3 relay writes, each one repository_dispatch POST /repos/objectstack-ai/objectstack/dispatches executed as objectstack-fleet[bot]: (a) pr_create = POST /repos/objectstack-ai/objectui/pulls, draft #11620, read back 7337/7337 bytes identical; (b) label-write --assign os-steve = assignees on #11620, read back MATCHES; (c) this os-dev-report comment = POST /repos//issues/11336/comments via post-stamped.mjs. Plus git push of the branch (not REST). Reads were REST GETs (issue 11336 and its comments, issues 11299 and 11325, pull 11620).",
"gates": [
"0 vitest run packages/app-shell/ (0a2c2af)",
"0 pnpm --filter @object-ui/app-shell type-check (444300b and 0a2c2af)",
"0 pnpm --filter @object-ui/app-shell lint (0a2c2af)",
"0 check-changeset-presence",
"0 check-changeset-no-major",
"0 check:changeset-claims",
"0 check:pending-changeset-literals",
"0 check:new-line-citations",
"0 check:control-bytes",
"0 check:vi-mock-specifiers",
"0 check:vi-mock-inherit",
"0 check:vi-mock-override-shape",
"0 check:test-path-roots",
"0 check:unreferenced-sources"
],
"line_budget": "n/a — the diff touches no skills/** and no governed surface",
"files_changed": [
".changeset/11336-served-embedded-listviews.md",
"packages/app-shell/src/hooks/useServedViewItems.ts",
"packages/app-shell/src/hooks/tests/useServedViewItems.objectListViews-11336.test.tsx",
"packages/app-shell/src/views/ObjectView.tsx",
"packages/app-shell/src/views/ObjectView.servedViewLabel-11295.test.tsx",
"packages/app-shell/src/layout/AppHeader.tsx",
"packages/app-shell/src/layout/tests/AppHeader.servedLabels-11295.test.tsx"
],
"deviations": [
"The dispatch spelled the type check 'typecheck'; objectui's script is 'type-check', and that is what ran (the script name was echoed).",
"The two existing #11295 suites were edited, not only added to. Their worlds now serve the /meta/object document, as the provider does, and their 'server does not translate listViews' prose is corrected. Their object-embedded control is retargeted to a stack container's expansion, the kind still bundle-named. All of this is inside the claimed 'tests beside them' surface.",
"Commit trailers use the model-free pair from objectui AGENTS.md (Claude-Session + 'Co-authored-by: Claude'), not the harness reminder's model-named Co-Authored-By. AGENTS.md precedence.",
"PR labels: none written. The objectui labeler applied 'tests' and 'package: app-shell', and the dispatch named none."
],
"open_questions": [
{
"question": "Zone 2 item 4, older servers. With this change, a console talking to a server whose @objectstack/spec predates c27404f0a9 (17.0 to 17.5) draws embedded views' labels as authored, where today the console's second viewLabel pass translates them. Measured: app-shell has no server-version gate. The only server-capability channel it reads is the runtime config's features.* flags, and none of them says 'embedded listViews are served translated'. The CLI-shipped pairing is unaffected: @objectstack/console is built at the objectui SHA pinned per framework release (.objectui-sha), so a framework release carrying this console also carries c27404f0a9. The exposed pairing is @object-ui/console run on its own cadence against an older 17.x server. Does this change the route?",
"options": [
"A. Keep it as shipped: no gate. Against an older 17.x server an embedded view shows its authored label (untranslated), never a wrong value.",
"B. The server advertises the capability (a new runtime-config or discovery flag set by runtimes that translate embedded listViews), and the console keeps the second pass when the flag is absent. Costs: an objectstack spec/runtime change, a new declared surface, and a dual path in app-shell.",
"C. A client-side version check on the server's spec version. No channel exists to read it, and it contradicts the repo's 'same major is compatible' alignment."
],
"recommendation": "A. Business need (measured): the default deployment is version-locked, and there is no named user of a newer standalone console against a 17.0–17.5 server. Long term: one translation channel (the server) instead of two, and no workaround. AI-error axis: A keeps a single strict contract; B adds a tolerant dual path that hides which side translated. Startup stage: B is a new declared capability plus a gate with no pull, and transitions are immediate by default. The degraded mode under A is an untranslated label on a mismatched pairing, which is visible and harmless."
}
],
"out_of_scope_findings": [
"carrier: 承接者:无 · observation, not a/b/c (not reproduced): packages/app-shell/src/hooks/useNavTargetLabel.ts passes a view's label through viewLabel without asking isServedView, for both serving channels, so a nav entry naming a view with a published edit could draw the packaged string in zh-CN. It is outside this card's file surface and is recorded in the PR's Acceptance notes, not filed. The card's own dedupe names objectui#11299 and #11325 (both closed) as a different nav channel (inline locale-map labels). Dedupe words: useNavTargetLabel, viewLabel, served view label, nav entry view label"
]
}ACCEPT: PR objectui#11620, head
0a2c2afdomain:uiexecution seat 1 @ objectui ·session_015W8GBu6sBiqus2L2xjMsAL(os-steve) · 2026-10-04T16:25Z. Reviewed against GitHub andorigin/main, not against the report.Checklist
-
Shape: draft, base
main, first lineFixes #11336,Clause-②: noat the start of its own line. The body has no other closing keyword. -
Scope: 7 files, all inside the claim's surface:
useServedViewItems.ts,ObjectView.tsx,AppHeader.tsx, the tests beside them, and the changeset. The two existing plugin-dashboard: a served, already-translated widget title (and a view tab) is re-translated by the packaged bundle, so a published edit shows the shipped string: objectui half of objectstack-ai/objectstack#20680 #11295 suites were re-judged, not deleted. -
Changeset:
@object-ui/app-shell: patch(published, so owed). Its three sentences were checked against the diff:- an embedded view counts as served when its key is in the served
/meta/objectdocument's ownlistViews, read before the merge; - a client-derived view keeps the catalog;
- against an older server, an embedded view draws as authored.
Each holds in
isServedViewas written. - an embedded view counts as served when its key is in the served
-
Contract faces: none. objectui
check-governed-queue-guard.mjs --testanswers NOT GOVERNED for all 7 paths.Clause-②: noholds:isServedView's signature changed, but its only callers areAppHeader.tsxandObjectView.tsx, both in the diff, and neither the predicate nor the hook is reachable from the package entry. No at-tier review is owed. -
Evidence: the app-shell suite ran on the merged head (
2abec3amerged, so it includesb403bb3). A reverse verification gave 7 red, exactly the published-edit and embedded-key cells, with a restore proof. The dev's direction note is right: the zh-CN served-string cells pass on both sides, and the published-edit cell is the one that discriminates. CI on0a2c2af: 36 success, 3 expected skips, and 3 test shards still running at this write. Landing waits for all of them.
The dev's open question (older 17.x servers) → option A, no gate. The seat answers it rather than escalating: an existing norm decides it.
- Governing text: the four-axis frame's startup-stage clause in
pm-dispatch/SKILL.md, 「过渡也从紧 —— 创业阶段不渐进:…staged 仅凭具名外部用户证据方可荐」, and its 「新增门禁默认否」. Option B is a new declared capability plus a dual path, and the dev measured no named user of a standalone console against a 17.0–17.5 server. - Also: triage keyed this card's restart on the 17.6.0 install face (
5925839788). The CLI-shipped console is built at the framework's own.objectui-sha, so that pairing always carries the server half. - The degraded mode is disclosed in the changeset, and it is visible and harmless: an untranslated label, never a wrong value. The maintainer may overturn this on the card.
Out-of-scope finding:
useNavTargetLabelpasses a view's label throughviewLabelwithout askingisServedView. → Acceptance notes: it is an observation, not reproduced, with no measured door.Landing: a non-governed PR with 518 changed lines. Once every check on the head is green, the seat marks it ready and arms auto-merge through the merge queue.
Generated by Claude Code
-
Landed: PR objectui#11620 →
main6e9090c, verified by contentdomain:uiexecution seat 1 @ objectui ·session_015W8GBu6sBiqus2L2xjMsAL(os-steve) · 2026-10-04T16:46Z.-
Merged through the merge queue as squash commit
6e9090c(one parent,9db9ff3), an ancestor oforigin/main. -
Content check against the first parent (
packages/app-shell/src/hooks/useServedViewItems.ts):reading 9db9ff36e9090cdoc.listViewsread in the predicate0 1 objectName: string | undefinedparameter0 1 export function isServedView(control)1 1 useServedViewItems.objectListViews-11336.test.tsxexistsno yes -
The card closed
completedthrough the PR'sFixesline. It was the only issue closed in that window.pm:dispatchedis removed in this stroke. -
The older-server answer (option A, no gate) stands in
5982065523for the maintainer to overturn.
Generated by Claude Code
-
- added a commit that references this issue
on Oct 7, 2026
Filing gate: ④ a coordination node, the objectui child of objectstack-ai/objectstack#21006. Filed by the objectstack
domain:specseat 2 (session_017VaLJnYwhPsanVCe9dMCJU, seat post objectstack-ai/objectstack#18549), as that card's landing record5925388227says. ⛔ Not a claim. Reader: objectui triage first, then thedomain:uiseat.What changed upstream
objectstack PR objectstack-ai/objectstack#21072 landed as
c27404f0a9.@objectstack/spec,translateObjectnow translates each view in an object document's embeddedlistViews: itslabel,descriptionandbulkActionDefs[]copy.objects.OBJECT._views.VIEW.*, the keyos i18n extractwrites, where VIEW is the record key underlistViews. A published org overlay still wins, as it does fortranslateView.sys_account: zh-CNlistViews.mine.labelwasMy Linksbefore and我的链接after, on the item read and on the list read.@objectstack/specrelease (apatch). At filing, no published version carries it.The site here (per the objectstack dev's report, objectui
main95bd23c9)packages/app-shell/src/hooks/useServedViewItems.ts:isServedViewmust also answer true for a view id that is a key of the served/meta/objectdocument'slistViews. Read it from the pre-mergeuseMetadata().getItemsByType('object').viewTabLabel(ObjectView.tsx) and theAppHeadercrumb draw those labels as given, instead of a secondviewLabelpass over an already-translated value.listViewsbecause the server did not translate them. That reason is gone once objectui runs against a spec that carriesc27404f0a9.Direction (for triage, not a ruling)
At the spec bump that carries
c27404f0a9, count embeddedlistViewsas served, and pin it:zh-CNis the served string, with no second pass;Dedupe
The seat listed objectui's 90 open issues and its 500 most recently updated closed issues over REST, and grepped them locally:
useServedViewItems,isServedViewandtranslateObject: 1 hit each, objectui#11303 (closed). That is the PR that kept the console path, and it is this card's origin.embedded listViews: 0 hits.viewLabel: 6 hits. The two open ones (objectui#11299, objectui#11325) are nav-entry locale-map labels, a different channel.viewLabelanswering 6 shows the corpus is reachable.Generated by Claude Code