types: chain the spec's exported navItemVisibleDataNeedsDataSource in the hand-written NavigationItemSchema superRefine (objectui half of objectstack-ai/objectstack#24534)

Blocked-by: objectstack-ai/objectstack#24534

Follow-up filed under the card's Done-when 2 ("a follow-up on objectui attaches it to its flat NavigationItemSchema at the next pin bump (filed there when this lands)"), by the seat that accepted the spec PR: objectstack domain:spec seat 2 (objectstack-ai/objectstack#18549), session session_019oFY6Hi35J7atvgVFnmoaK. ⛔ Not a claim.

Unlock takes two readings:

The blocker spec(ui): export navItemVisibleDataNeedsDataSource, so mirrors of NavigationItemSchema can attach it by identifier — objectui's NavigationItemSchema accepts a nav entry whose visible reads data with no dataSource, which the spec refuses objectstack#24534 closes. feat(spec/ui): export navItemVisibleDataNeedsDataSource, so a mirror of NavigationItemSchema attaches it by identifier objectstack#24628 merges with Fixes #24534.
objectui's .objectstack-sha is a descendant of that merge commit. Check it with REST compare, not a local merge-base on a shallow clone.

objectui builds @objectstack/* from that pin's source, so the export is installable once the pin covers the merge. No npm release is involved.

Measured

At objectui origin/main fb85637e1 (.objectstack-sha f4755a68a7):

packages/types/src/zod/app.zod.ts:256 declares visible: z.union([z.boolean(), z.string()]).optional().
Its superRefine (:279) chains the spec's objectNavTargetExclusivity(item, ctx) at :335, under if (item.type === 'object'). That is the objectui#8563 precedent.
navItemVisibleDataNeedsDataSource occurs nowhere under packages/.

Consequence: a navigation entry with no dataSource whose visible reads data, such as the spec card's { id: 'nav_envs', type: 'object', objectName: 'sys_environment', visible: 'data.total == 1' }, passes objectui validate. The spec refuses it with one custom issue at visible.

What the spec now exports (objectstack-ai/objectstack#24628)

navItemVisibleDataNeedsDataSource(item: unknown, ctx: z.RefinementCtx): void, from @objectstack/spec/ui:

It is the refinement the spec's NavigationItemSchema runs, on every branch, attached by identifier.
It reads both spellings of visible: the authored bare string, normalised through the spec's own EvaluatedExpressionInputSchema, and the parsed { dialect, source } envelope.
So this mirror, whose visible stays the authored string, chains it as it is.
A boolean or absent visible judges nothing.

What to do

Confirm the unlock readings above.
Chain navItemVisibleDataNeedsDataSource(item, ctx) in the superRefine for every item type. The spec mounts it on the whole navigation union, not on the object branch alone. Import it by identifier; ⛔ do not restate the rule.
Pin it in packages/types/src/__tests__/navigation-spec-parity.test.ts:
- the card's entry is refused by both doors with the same issue at visible;
- the control (the same entry with a dataSource) is accepted by both;
- one non-object branch (for example url) is refused too.
Clause-②: objectui's validator starts refusing an entry the spec already refuses. That is a narrowing toward the spec, so follow objectui's changeset rules for it.

Related, not this card

objectui#13075: this mirror's visible arm refuses the spec's { dialect, source } envelope. That is an unpinned divergence, filed separately. This card does not depend on it.

Dedupe

The semantic issue search on objectstack-ai/objectui that #13075 records found no attach card for this export. objectui#8563 (closed) is the objectNavTargetExclusivity precedent.

Triage: p1 + `target:v18` · lane `domain:spec` → `domain:ui` · `area:api` · `pm:blocked` kept · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-19T19:54Z

Triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015). ⛔ Not a claim, ⛔ not a dispatch.

The lane moves to domain:ui, for the same reason as objectui#13075: objectui's packages/types, and no objectui seat reads domain:spec.
p1 + target:v18, kept with its spec half. The maintainer pulled spec(ui): export navItemVisibleDataNeedsDataSource, so mirrors of NavigationItemSchema can attach it by identifier — objectui's NavigationItemSchema accepts a nav entry whose visible reads data with no dataSource, which the spec refuses objectstack#24534 into 18.0 (「24534 同意提前」), and this consumer face is what that export exists for: objectui validate refuses what the server refuses.
The unlock reading is the card's own, and it is right. First, objectstack#24534 closes through PR #24628. Second, objectui's .objectstack-sha descends from that merge commit (REST compare). This seat applies exactly that pin-ancestry reading. Closing the upstream alone does not unlock it.

Generated by Claude Code

Unlocked by pin ancestry. objectstack triage, R542 · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-20T16:55Z

The pin: objectui's .objectstack-sha is now 06cab4849a8a616487d468a11dd370e929137990, objectstack 18.0.0's version commit. objectui PR chore(deps): move .objectstack-sha to objectstack 18.0.0 version commit 06cab4849a (objectui#13261) #13266 (objectui#13261) moved it.
The reading: REST compare/35450f0732...06cab4849a answers ahead with behind_by: 0. Both of this card's readings hold: spec(ui): export navItemVisibleDataNeedsDataSource, so mirrors of NavigationItemSchema can attach it by identifier — objectui's NavigationItemSchema accepts a nav entry whose visible reads data with no dataSource, which the spec refuses objectstack#24534 is closed completed (PR #24628), and the pin descends from that merge, so the pin carries the commit this card waited on.
State: pm:blocked → pm:queue. The shape is unchanged.

Generated by Claude Code

Claim: PM loop round 18 (priority:p1)
Session: session_01XVp19E1MYUTFM3dnkwz8gB
Account: baozhoutao (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-13076-nav-visible-data
Worktree: objectui-issue-13076
Domain: domain:ui
Seat: domain:ui#2
File surface:

packages/types/src/zod/app.zod.ts (the NavigationItemSchema superRefine ~279, chaining the spec's navItemVisibleDataNeedsDataSource for every item type);
packages/types/src/__tests__/navigation-spec-parity.test.ts;
one .changeset/13076-nav-visible-data.md (@object-ui/types, level per objectui's rules for a narrowing toward the spec).

⛔ Import the rule by identifier; do not restate it. Stop on breach; explain in the report.
Container & model: S, mode:subagent, model: opus (default tier; no path-derived mandate)
Clause-②: yes (the published NavigationItemSchema starts refusing an entry the spec already refuses; a contract narrowing toward the spec, contract review owed)
Responsibility: n/a — not a defect card (spec mirror attach)
Thread-read: none
Serial constraints cleared:

objectui#13261 (2a15427, pin 06cab4849a) landed; the unlock reading (objectstack#24628's merge is an ancestor of the pin) is the dev's first step.
No open objectui PR touches packages/types/src/zod/app.zod.ts (read at 17:30Z).
Claimed at: 2026-10-20T17:28Z

Generated by Claude Code

os-dev-report

{
  "issue": 13076,
  "status": "blocked",
  "branch": "claude/issue-13076-nav-visible-data",
  "pr": null,
  "session": "session_01XVp19E1MYUTFM3dnkwz8gB",
  "premise_still_valid": true,
  "summary": "BLOCKED on the file surface, a one-file amendment only the PM can make. Unlock readings both hold: objectstack#24534 closed completed; REST compare 35450f0732 (merge of #24628)...06cab4849a = ahead, ahead_by 116, behind_by 0; navItemVisibleDataNeedsDataSource is a function on the installed @objectstack/spec/ui. Implemented within the claimed surface and pushed as Resume-from: claude/issue-13076-nav-visible-data@3d9e53d637 (base 8cd4db4974): app.zod.ts imports the spec function by identifier and chains it in the flat arm's superRefine for every type (outside the object branch) AND in the doc arm's superRefine, because the doc arm's judge is the spec's DocNavItemSchema member, which does not carry the rule the spec mounts on its whole union; pins (a)-(d) plus doc and nested-group rows in navigation-spec-parity.test.ts; changeset .changeset/13076-nav-visible-data.md at minor. BREACH: packages/types/src/__tests__/imported-defaults-8317.test.ts goes red (2 tests) because its raw-import ledger does not list the new binding; the fix is two lines in that file (add navItemVisibleDataNeedsDataSource to its @objectstack/spec/ui import and a REFINEMENT_EXCEPTIONS entry beside objectNavTargetExclusivity), the same ledger every chained spec check (objectui#8563, #7715, #11073, ...) was entered into. That file is not in the claim's file surface, and the order says stop and report, so I stopped. Ask: amend the claim surface with that file and resume from the sha above; no draft PR opened.",
  "tests": "Probe (spec vs mirror, after chaining): card entry custom@visible on both; same + dataSource OK on both; url data.on custom@visible both; url envelope {cel,data.on} custom@visible both; doc data.on custom@visible both; group visible data.on custom@visible both; group child url custom@children.0.visible both; url visible true: spec invalid_union@visible, mirror OK (the existing boolean divergence); url without visible OK both. Locked run at 3d9e53d637 content: pnpm exec vitest run --maxWorkers=1 navigation-spec-parity + imported-defaults-8317 + nav-target-exclusivity-8563 = VERDICT command-exit 1, Test Files 1 failed | 2 passed (3), Tests 2 failed | 347 passed (349); the 2 failures are imported-defaults-8317 'no value read bypasses stripImportedDefaults' and 'every symbol the mirrors import is covered by the differential above' (expected [ 'navItemVisibleDataNeedsDataSource' ] to deeply equal []); navigation-spec-parity and nav-target-exclusivity-8563 green. pnpm --filter @object-ui/types type-check (tsc src + examples + test configs) = VERDICT command-exit 0. pnpm check:control-bytes = exit 0. node scripts/check-changeset-presence.mjs = exit 0 (1 changeset declared). NOT MEASURED: ablation (drop chain, (a) and (c) red on objectui's door, blob-hash restore), narrowed lint, the pre-tsc battery and check-changeset-fixed; reason: stopped at the file-surface breach before the gate pass, all owed on resume.",
  "mcp_calls": "0",
  "api_writes": "1 - POST issues/13076/comments (this os-dev-report, via the fleet-write relay); git push x2 to the branch (empty probe, then 3d9e53d637), not REST. Reads via gh api: issues/13076, its comments, objectstack issues/24534, pulls/24628, compare.",
  "open_questions": [
    { "question": "Amend the claim's file surface to include packages/types/src/__tests__/imported-defaults-8317.test.ts (two-line ledger entry for the new raw spec refinement import)?", "options": ["A: amend and re-dispatch from claude/issue-13076-nav-visible-data@3d9e53d637", "B: keep the surface and find another mount - none exists without restating the rule or routing the import around the ledger, which is gaming the gate"], "recommendation": "A, because the ledger exists precisely to list chained spec refinements and every precedent (objectNavTargetExclusivity, the check* exports) was entered there; the edit is mechanical and the claim surface simply missed it." },
    { "question": "Changeset level for the narrowing", "options": ["minor (chosen)", "patch"], "recommendation": "minor: the claim marks Clause-2 yes (published NavigationItemSchema starts refusing input), objectui AGENTS.md versioning says objectui's own breaking or narrowing changes are marked minor and never major (check-changeset-no-major), and @object-ui/types precedents that change the strict authoring face (11070 pair) are minor." }
  ],
  "out_of_scope_findings": [
    "carrier: the publish-time floor step (scripts/objectstack-floor.mjs) - noted, not filed: @object-ui/types declares @objectstack/spec ^18.0.0-next.5, and the export first ships in 18.0.0 (git show of the next.5 tag's ui/app.zod.ts has 0 declarations, 18.0.0 has 1; control leg next.3 ancestor-of next.5 = exit 0). Per check-spec-range-floors.mjs's header the floor is computed from .objectstack-sha at publish, so this is expected to resolve itself; check:spec-floors may read it red on the branch."
  ]
}

Generated by Claude Code

os-dev-report

{
  "issue": 13076,
  "status": "done",
  "branch": "claude/issue-13076-nav-visible-data",
  "pr": "https://github.com/objectstack-ai/objectui/pull/13275",
  "session": "session_01XVp19E1MYUTFM3dnkwz8gB",
  "premise_still_valid": true,
  "summary": "Addendum after the PM's Q1=A ruling. This report supersedes the earlier blocked report (comment 6297236278). I resumed from 3d9e53d637 (base 8cd4db4974; merging origin/main was a no-op). app.zod.ts chains @objectstack/spec/ui's navItemVisibleDataNeedsDataSource by identifier in two places: the flat arm's superRefine, outside the object branch, and the doc arm's superRefine. The second call is needed because that arm's judge is the spec's DocNavItemSchema member, which does not carry the union-level rule. Pins (a) to (d), plus a doc row and a nested-group row, are in navigation-spec-parity.test.ts. Under the amended surface, the ledger entry went into imported-defaults-8317.test.ts: the import plus a REFINEMENT_EXCEPTIONS row beside objectNavTargetExclusivity. The changeset is @object-ui/types minor. Draft PR #13275 is open at head 6931b9c4de, assigned to baozhoutao. Its body's first line is the Fixes line for this card; its second line is Clause-②: yes, with a one-line description of the narrowing.",
  "tests": "Everything below ran at head 6931b9c4de. (1) The types family under the lock: the 23 packages/types test files that name zod/app.zod or NavigationItemSchema, run with --maxWorkers=1, gave Test Files 23 passed (23), Tests 765 passed (765), VERDICT command-exit 0. The full package directory is declared to CI. (2) Ablation under the lock: ablation-replace.mjs --delete --expect 2 on the anchor 'navItemVisibleDataNeedsDataSource(item, ctx);'. Anchor 2 went to 0 and the blob went from 31d00ce63f50 to dd5095036d88. The run was red as expected: the card object entry, url, doc and nested-group rows failed, Tests 4 failed | 53 passed (57), and both controls stayed green. Restore: the blob after restore is 31d00ce63f50, equal to the HEAD blob, and git diff HEAD is empty; an independent git hash-object afterwards also read 31d00ce63f50. Source is imported directly by the test file (../zod/app.zod.js), so no dist build was involved. (3) pnpm --filter @object-ui/types type-check exited 0; the tsconfig.test --listFiles output includes both edited test files. (4) Pre-tsc battery, all exit 0: check:phantom-deps, self-import, unreferenced-sources, doc-example-readers, handler-key-reads, side-effects-array, element-data-source-declaration, spec-symbols, action-forward-parity, designer-field-key-parity, icon-record-names, i18n-keys, i18n-drift, i18n-designer-parity, control-bytes, and check-hand-rolled-comment-mask. check-changeset-fixed and check-changeset-presence also exit 0. (5) Narrowed lint: eslint --no-inline-config --format json on the 3 changed .ts files. The checked population is those 3 files, all linted under eslint.config.js. The JSON counts are 3 files, 0 errors and 1 warning; the warning is the existing no-explicit-any on the NavigationItemSchema export, outside this diff. Invariance: the config sets no parserOptions.project or projectService, so the linting is not type-aware and this diff cannot move any untouched file's verdict. (6) check:spec-floors exit 1, no-artifact (it needs a workspace build); recorded, not fixed, as ordered.",
  "mcp_calls": "0",
  "api_writes": "3 relay writes in total. (1) POST issues/13076/comments: the first, blocked report, comment 6297236278. (2) pr_create, draft forced, PR #13275 with assignees baozhoutao; read-back byte-identical at 4221 bytes. (3) POST issues/13076/comments: this addendum. Git pushes to the branch were not REST writes. Labels: zero writes, because no objectui gate reads a dev-applied label and the order named none.",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: the publish-time floor step (scripts/objectstack-floor.mjs) · noted, not filed (the PM ruled it noted). @object-ui/types declares @objectstack/spec ^18.0.0-next.5, but the export first ships in 18.0.0."
  ]
}

Generated by Claude Code

ACCEPT — PR #13275 @ 6931b9c4dec54819a86a2645ada173e2ff3aad0a · domain:ui seat 2 (session_01XVp19E1MYUTFM3dnkwz8gB)

Clause-②: yes. The published NavigationItemSchema narrows toward the spec. The contract review PASSED on 6931b9c4de (record posted on this PR).
Seat review: navItemVisibleDataNeedsDataSource from @objectstack/spec/ui is chained by identifier, never restated, in the flat arm's superRefine for every type and in the doc arm's. The doc-arm call is necessary: the spec mounts the rule on its union, not on DocNavItemSchema. It cannot double-report, because the arms are discriminated on type. A navigation entry at any depth with no dataSource whose visible reads root data is refused at visible with the spec's issue; everything else is unchanged.
Unlock confirmed: spec(ui): export navItemVisibleDataNeedsDataSource, so mirrors of NavigationItemSchema can attach it by identifier — objectui's NavigationItemSchema accepts a nav entry whose visible reads data with no dataSource, which the spec refuses objectstack#24534 is closed, and compare 35450f0732...06cab4849a answers ahead.
Surface amended by the seat: imported-defaults-8317.test.ts gains the raw-import ledger row beside objectNavTargetExclusivity (ruled Q1 = A on the dev's blocked report).
Pins: the card's object entry, a url entry, a doc entry and a nested group are each refused by both doors with the same issue; the dataSource and absent-visible controls are accepted. The ablation (both chain calls deleted) turned the four refusal rows red with the controls green, and the restore was proven by blob. The types family (23 files, 765 tests) is green.
Changeset: @object-ui/types minor (objectui's narrowing rule; check-changeset-no-major). The ^18.0.0-next.5 range is lifted to the pin's floor (18.0.0, whose tag is the pin commit) by objectstack-floor.mjs at publish.
CI: all 36 check-runs completed on the head, 33 passed and 3 skipped.
Path fork: no governed path. Seat landing.
Findings: filed docs: the console guide Expression Visibility row teaches a navigation visible that reads data with no dataSource, which both validators now refuse #13276, the console guide's "Expression Visibility" row, which teaches a spelling both doors now refuse (escalated by the contract review).

Next: ready + auto-merge; landing is verified by content on main.

Generated by Claude Code

Landed · `domain:ui` seat 2 · `session_01XVp19E1MYUTFM3dnkwz8gB`

Merged: PR feat(types): chain the spec's navItemVisibleDataNeedsDataSource on every navigation arm #13275 through the merge queue as main 8556295f04.
Verified by content: each of the 4 files the PR changes has the same blob at 8556295f04 as at the accepted head 6931b9c4de. The merge message carries no model identifier.
Contract review: PASS, on this PR.
Landed: @object-ui/types' NavigationItemSchema chains @objectstack/spec/ui's navItemVisibleDataNeedsDataSource by identifier, on the flat arm for every type and on the doc arm. A navigation entry with no dataSource whose visible reads data is now refused by objectui validate, as the spec refuses it. Changeset minor.
Follow-up: objectui#13276 (in flight with this seat): the console guide's "Expression Visibility" row taught the now-refused spelling.

pm:dispatched is removed.

Generated by Claude Code

feat(types): chain the spec's navItemVisibleDataNeedsDataSource on ev…

Navigation Menu

Repository navigation

types: chain the spec's exported `navItemVisibleDataNeedsDataSource` in the hand-written `NavigationItemSchema` superRefine (objectui half of objectstack-ai/objectstack#24534) #13076

Measured

What the spec now exports (objectstack-ai/objectstack#24628)

What to do

Related, not this card

Dedupe

Triage: p1 + `target:v18` · lane `domain:spec` → `domain:ui` · `area:api` · `pm:blocked` kept · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-19T19:54Z

Landed · `domain:ui` seat 2 · `session_01XVp19E1MYUTFM3dnkwz8gB`

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

types: chain the spec's exported navItemVisibleDataNeedsDataSource in the hand-written NavigationItemSchema superRefine (objectui half of objectstack-ai/objectstack#24534) #13076

Description

Measured

What the spec now exports (objectstack-ai/objectstack#24628)

What to do

Related, not this card

Dedupe

Activity

objectstack-fleet commented on Oct 19, 2026

Triage: p1 + target:v18 · lane domain:spec → domain:ui · area:api · pm:blocked kept · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-19T19:54Z

objectstack-fleet commented on Oct 20, 2026

objectstack-fleet commented on Oct 20, 2026

objectstack-fleet commented on Oct 20, 2026

objectstack-fleet commented on Oct 20, 2026

objectstack-fleet commented on Oct 20, 2026

objectstack-fleet commented on Oct 20, 2026

Landed · domain:ui seat 2 · session_01XVp19E1MYUTFM3dnkwz8gB

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

types: chain the spec's exported `navItemVisibleDataNeedsDataSource` in the hand-written `NavigationItemSchema` superRefine (objectui half of objectstack-ai/objectstack#24534) #13076

Triage: p1 + `target:v18` · lane `domain:spec` → `domain:ui` · `area:api` · `pm:blocked` kept · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-19T19:54Z

Landed · `domain:ui` seat 2 · `session_01XVp19E1MYUTFM3dnkwz8gB`