feat(spec)!: object-kanban conditionalFormatting takes the list view's own rules, and its hold exits (#21464, S-kanban-cf)

[objectstack-fleet]

Part of #21464
Clause-②: yes (narrowing)

What this does

The S-kanban-cf stage of the ComponentPropsMap z.unknown() close-out: the exit of the object-kanban conditionalFormatting hold, per the claim 5977940136, the landing note 5977935364, triage's direction 5961300594 and the seat answer 5963787404 (item 2). The census ran first. It found no working writer that the list view's member refuses, and the reader's declared dialect is exactly the list view's rule. So the member is typed by reference (def identity) to ListViewSchema.shape.conditionalFormatting, as object-grid's is. The enumeration pin's held-for-decision line leaves the ledger, and so does the emptied held-for-decision stage.

member	was	now	read point at the .objectui-sha pin ab1879721595
object-kanban conditionalFormatting	z.unknown()	ListViewSchema.shape.conditionalFormatting, by reference: [{ condition, style }], a non-blank CEL condition (string or { dialect, source } envelope) and a style map of strings	KanbanBoardCore.tsx:114 hands schema.conditionalFormatting to the board; getCardStyles (KanbanImpl.tsx:182) evaluates it through resolveConditionalFormatting (core/src/evaluator/listConditional.ts:532), the evaluator the grid's rows use (ObjectGrid.tsx:2970)

No new export and no new schema: the row reads the list view's member, so one rule is judged the same way on the list view, object-grid and object-kanban.

The hold's exit

• Why it was held (5963787404 item 2): objectui's kanban authored a second rule dialect, the native { field, operator, value, backgroundColor } rule and the flat colour rule, which the list view's member refuses.
• What changed in objectui: objectui#11522's state is completed, landed by objectui#11532 (c73cdb5695). object-kanban's conditionalFormatting takes the spec list view's { condition, style } rule only, and the native and flat-colour dialects are refused by name. git merge-base --is-ancestor c73cdb5695 ab1879721595 exits 0, and so does the pin against objectui main 2e818d0b51. Exit 0 proves itself in any checkout, so no control leg is owed.
• The reader's declaration at the pin: KanbanConditionalFormattingRuleSchema extends the spec list view's rule element, read by reference (types/src/zod/objectql.zod.ts:337, :3074), with tombstones for field, operator, value, backgroundColor, borderColor and textColor. Its TS twin KanbanConditionalFormattingRule extends SpecConditionalFormattingRule (types/src/objectql.ts:5166). The registration's description reads the same (plugin-kanban/src/index.tsx:676, recorded in this repo's sdui.manifest.json).
• Two differences remain. Neither is a working rule:
  1. The shared evaluator still reads the native, expression and top-level colour arms, for a rule a relay hands the board in a stored dialect (KanbanImpl.tsx:160-166, listConditional.ts:495-499). It is the grid's evaluator, unchanged. No authored object-kanban writer uses those arms (census below), and objectui's own faces refuse them.
  2. objectui's condition takes a string first, so its mirror still admits a blank condition (objectql.zod.ts:363-366, pinned by objectui's spec-expression-wire-slots-10946.test.ts:151-155). The evaluator answers a blank predicate with the caller's fallback (listConditional.ts:252-260, single-eval route :339-340), and the board passes fallback: false (:546), so such a rule paints no card. It is refused here, as on the list view and on object-grid. This was read, not run.
• Pin to objectui main 2e818d0b51: KanbanBoardCore.tsx, KanbanImpl.tsx, ObjectKanban.tsx, plugin-kanban/src/index.tsx, listConditional.ts and ObjectGrid.tsx are byte-identical (git diff --quiet, exit 0 each). The two @object-ui/types files changed only around grouping: zero changed lines mention conditional, and the rule schema's line 3074 did not move.

The census

A writer is a value written on the block: a page-component node (an object literal naming object-kanban, flat or in its properties bag, or a literal asserted as one), a direct parse through the row, the block's React component inside its schema prop, or the argument of a same-file helper that mounts one. Values resolve through same-file constants and spreads, and parameters resolve at every same-file call site and it.each table. Instrument 1 is a TypeScript-AST walk over every tracked file naming conditionalFormatting (code, .json, and fenced code in .md / .mdx). It records each conditionalFormatting member with its static value and block context, and parses the value through the list view's member. Instrument 2 parses every rule-shaped object within 400 characters after a conditionalFormatting token, in any syntax (tuples, key: / value: rows), as a one-rule list. Each hit outside a block context was read by hand. Controls: instrument 1 finds the showcase's list-view writer and the object-grid pin's writers in this repo.

corpus	object-kanban writers	parse	refused
objectstack 16d241a6af (every tracked file; 33 name conditionalFormatting)	1	0	1 · packages/spec/src/ui/component.test.ts:3947
objectui pin ab1879721595 (94 files name the member, 61 of them also kanban)	9 working, 10 probe	9	10 · all probes
objectui main 2e818d0b51	the same 19	9	10
hotcrm 4054ec2680	0
cloud 2205b53010	0

• objectstack's one writer is this package's own test that the key survived the quickAdd retirement: [{ field: 'priority', value: 'high' }]. It has no operator, so the evaluator built no predicate from it (listConditional.ts:474) and painted nothing, and objectui's own faces refuse it. Its assertion is about the key, not the value, so it is respelled to { condition: "record.priority == 'high'", style: { backgroundColor: '#fee2e2' } } and still asserts success: true.
• objectui's nine working writers, every one a test fixture, all { condition, style }: the board test's three rules through its mount helper (ObjectKanban.structuredMembersReachTheirSinks-8313.test.tsx:472, :492, :509), the asserted node in objectFieldsIsAPropNotASchemaKey-7742.test.tsx:97, the declared-keys row in structuredKeysAreDeclaredAndHonoured-8313.test.ts:109 (which parses through this very row on the installed spec), the live-member row in object-kanban-allow-collapse-retired-8801.test.ts:198, the dialect test's control (kanban-conditional-formatting.test.ts:48), and the wire-slot test's string and envelope conditions (spec-expression-wire-slots-10946.test.ts:123).
• objectui's ten refused values are refusal probes. Nine are refused by objectui's own faces too: the native rule (kanban-conditional-formatting.test.ts:88), the flat colour rule (:103), a colour beside style (:118, three keys), an undeclared label (:139), and three malformed conditions (spec-expression-wire-slots-10946.test.ts:158). The tenth is that file's probe that its mirror still admits a blank condition (:152), which the board answers with no paint (above).
• Not writers, and all of them parse: eight rules mount the runtime KanbanBoard directly (cardPredicateScope.test.tsx). Nine are view-face relays of a list view's rules (ObjectView.kanbanConditionalFormatting.test.tsx, objectViewHostSurface.test.tsx). The other members in those 61 files are object-grid or list-view writers, which the list view's member already judges. The non-test hits are run-time hand-offs and the schema declarations. objectui main adds one non-writer (InterfaceListPage.relayCensus-11572.test.ts, a relay census record).
• hotcrm and cloud have no conditionalFormatting at all, case-insensitive and snake case included. Controls: kanban in 48 and 35 files, and objectName in 115 and 145.

Changes

• packages/spec/src/ui/component.zod.ts: the object-kanban row's conditionalFormatting is ListViewSchema.shape.conditionalFormatting with a card-level .describe(), and its docblock records the read points, objectui's declaration, the two differences and the census. The import comment names the kanban member beside the grid's.
• packages/spec/src/ui/component-props-unknown-members.pin.test.ts: the held-for-decision line and stage leave, along with the now-unused staged() helper. The kanban rule's conditionalFormatting[].condition.ast joins the expression-AST lines, exactly as the grid's does. The new §5 pins the def identity, three accepted values parsing to what the list view's member answers, eight refusals by code and path, the absent case, and the D3 id. The header records the exit.
• packages/spec/src/ui/component.test.ts: the one census writer, respelled (above). This file is outside the claim's file list. It is the member's own fixture, and the change reddens it otherwise.
• packages/spec/src/migrations/entries/semantic/18.ui-object-kanban-conditional-formatting-typed.ts (new) and packages/spec/src/migrations/registry.ts: one step-18 D3 entry (the semantic region regenerated by gen:migration-registry), and one rationale fragment at order 77, the next free order on main at the base (73 to 76 are taken, 74 twice), inserted where its id sorts. [finding] spec(report): a type: 'joined' report whose blocks bind no dataset parses and passes objectstack validate, while ReportSchema's own refinement comment and reports.mdx say "each block dataset-bound" #21702 also adds a fragment, so whichever lands later re-reads the order and merges origin/main through scripts/pm/os-regen-merge.sh. No D2 conversion and no RETIRED_KEYS_BY_MAJOR row: a page component's properties is not parsed on the save or load path, and the refused values are nested member values.
• content/docs/references/ui/component.mdx: regenerated by check:generated --fix, which proved only check:docs stale.
• .changeset/21464-component-props-kanban-conditional-formatting-typed.md: @objectstack/spec minor, BREAKING banner, the Clause-② line, FROM → TO, the census, and the ADR-0087 marker registering ui-object-kanban-conditional-formatting-typed.

Measurements

Head 011e53fa2a (this branch merged with main 7e0066af7a, a comment-only re-anchor of spec test citations that touches none of these files). Base 16d241a6af. Heavy runs went through scripts/pm/os-verify-lock.sh, and every exit code was captured before any pipe.

• Red first, on the published @objectstack/spec@17.6.0 (the npm tarball, ComponentPropsMap['object-kanban'].safeParse) and on main 16d241a6af (built): each of these is ACCEPTED on both: conditionalFormatting: 42, 'red', the native rule, the flat colour rule, an expression rule, a rule with no style, a blank condition, and a style value of 5. Control: navigation: 42 is REFUSED there (invalid_type at navigation). Each junk value is refused on this branch, pinned in §5 by code and path.
• pnpm --filter @objectstack/spec test at 011e53fa2a: exit 0, Test Files 612 passed (612), Tests 18199 passed, 1 todo, nothing skipped. The same at 27a0e12a09, before the merge.
• pnpm --filter @objectstack/spec typecheck at 011e53fa2a: exit 0, check:test-typecheck: OK (52 files / 246 errors / 135 signatures held). tsc -p tsconfig.test.json --listFilesOnly names both touched tests and the new D3 entry.
• Build and generated artifacts: the spec build is green. check:generated proved only check:docs stale, --fix regenerated that one, and the re-check at 011e53fa2a reads "All 15 generated artifacts are up to date". check:api-surface reads "public API surface + factory signatures unchanged". turbo run build --filter=!@objectstack/docs --concurrency=2 built 72 of 72 tasks.
• Consumers: pnpm --filter @objectstack/lint test (after building lint's dependency closure, --filter '@objectstack/lint^...'): 119 files passed, 5622 tests passed, 5 skipped. The 5 read lint's own dist; after pnpm --filter @objectstack/lint build, validate-component-props, lazy-deps and runtime-lazy-deps passed 61 of 61, nothing skipped. The first lint run before that closure build failed 56 files on Failed to resolve entry for package "@objectstack/formula" / "@objectstack/sdui-parser": unbuilt dependencies, not a measurement.
• The public door, both ways (validateComponentProps from lint src over the built spec, one object-kanban node per page). These report nothing: a { condition, style } rule, and an envelope condition. These are reported: 42 as component-props-invalid at properties.conditionalFormatting; a rule with no style as component-props-invalid at …conditionalFormatting.0.style; the native rule as component-props-invalid at .0.condition and .0.style plus component-props-unknown-key at .0.field, .0.operator, .0.value and .0.backgroundColor; an expression rule at .0.condition plus component-props-unknown-key at .0.expression; and a blank condition as component-props-invalid at .0.condition.
• Ablation, one leg, on the committed tree (27a0e12a09; the merge did not touch these files). The driver used node scripts/ablation-replace.mjs in wrap mode, inside its own EXIT INT TERM restore trap on the absolute path, with an empty hash read as failure, against HEAD blob 830a02f619. The kanban member went back to z.unknown().optional(): anchor x1 → x0, replacement x0 → x1, blob → 4b67209318, re-counted in the leg (ListViewSchema.shape.conditionalFormatting 2 → 1). Red: 12 failed, 421 passed. §1 received exactly [ "object-kanban conditionalFormatting" ] as unlisted, and [ "object-kanban conditionalFormatting[].condition.ast" ] as an outliving line. §5 failed its identity, the string-condition parse and all eight refusals. Still green, as expected under z.unknown(): the envelope and empty-list parses, the absent case, the D3 id, and §1's size control (one member in, one out). Restore: blob equals HEAD (830a02f619), git diff HEAD is empty, and git status is clean. The pins import ./component.zod from source, so no build or dist preflight sits between mutation and run.
• Derived gates at 011e53fa2a: node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands (no paths; merge base 7e0066af7, 7 paths, 276 changed lines) derived 114 commands, all exit 0. --ran printed "114 derived, 114 run, 0 NOT-MEASURED, 0 UNRUN". At 27a0e12a09 the first pass recorded six exit-3 PREREQUISITE NOT MET refusals for unbuilt packages. After the builds above they re-ran exit 0, and the pass at 011e53fa2a ran all 114 clean.
• check-widening-tells over the branch diff: --declaration no exits 0. That is no tell: the row takes an existing member by reference and adds no key, arm or export. --declaration yes exits 0.
• Narrowed lint: eslint --no-inline-config --format json over the 5 changed TypeScript files reports 5 files, 0 errors and 0 warnings. The population comes from eslint's own --print-config: those 5 resolve a config, and the changeset and component.mdx resolve none. Invariance: eslint.config.mjs never enables type-aware linting (its own text, about line 327), so this diff cannot move a verdict on an untouched file. The full pnpm lint is CI's.
• NOT MEASURED: the Console Pin Gate, the Dogfood Regression Gate and the full pnpm lint are CI-owned. objectui was read at the pin and at main, not built against this spec. The board's paint for a blank condition was read, not run. The showcase validate was not run: no example in this repository authors an object-kanban node. CI on this PR was not waited on.
• Rationale order: [finding] spec(report): a type: 'joined' report whose blocks bind no dataset parses and passes objectstack validate, while ReportSchema's own refinement comment and reports.mdx say "each block dataset-bound" #21702's branch (c4e3ab9631) also takes order 77 for ui-report-joined-block-dataset-required. That is not refused, because joinRationale breaks a tie by id. Whichever lands later re-reads the next free order and merges origin/main through scripts/pm/os-regen-merge.sh.

Acceptance notes

• objectui's next @objectstack/spec bump reads this row in one place only, structuredKeysAreDeclaredAndHonoured-8313.test.ts, whose value parses. objectui's object-kanban node is its own flat ObjectKanbanSchema and takes no propsBag over this row. The console's member-pin prose for object-kanban.conditionalFormatting (apps/console/src/__tests__/registry-inputs-spec-parity.test.ts:3003) says "The spec row is z.unknown()", which goes stale with that bump. Carrier: objectui's next spec bump. Noted, not filed.
• The blank condition is the one value objectui's mirror admits and this row refuses. The same split already exists on the list view and on object-grid, and the board paints nothing for it. Noted, not filed.
• The lint door's nested unknown-key message reads "field is not a prop object-kanban declares" for a key inside a rule. The path (properties.conditionalFormatting.0.field) and the zod text ("on this conditional formatting rule") are right. The wording predates this change and reads the same on object-grid. Noted, not filed.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

3 anchor(s) derived from 1 changed package(s); no hand-written page names any of them, so this run has nothing to list — not a clean bill of health. This check sees only pages that NAME a derived anchor: one that documents this change in prose, or enumerates it in an authoring dialect, names none and stays invisible to it on every run.

What this run could not see

• 6 name(s) were too generic to anchor anything (single lowercase words)
• 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 — 138 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 7e0066af7a8e05709dc60096c69bc85e299d0331 → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 44900e55d7fed22a383fd0578499fdd2a0037467 — the merge of head 011e53fa2ad81dbddbfad0e5ac57c05274f196ec into base 7e0066af7a8e05709dc60096c69bc85e299d0331, 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 44900e55d7fed22a383fd0578499fdd2a0037467 && git checkout 44900e55d7fed22a383fd0578499fdd2a0037467
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 7e0066af7a8e05709dc60096c69bc85e299d0331 011e53fa2ad81dbddbfad0e5ac57c05274f196ec && git checkout -B drift-repro 7e0066af7a8e05709dc60096c69bc85e299d0331 && git merge --no-ff 011e53fa2ad81dbddbfad0e5ac57c05274f196ec

node scripts/docs-audit/affected-docs.mjs --json 7e0066af7a8e05709dc60096c69bc85e299d0331

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

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 37193945077 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集），
所以失败的测试可能在本 PR 没碰过的包里 —— 那不是重排能修的。每次盲目重排都会让排在后面的所有 PR 重建一轮。

失败的 job（日志抽取，best effort）：

• Console Pin Gate — 失败步骤: Build the Console SPA at the pinned objectui SHA

  ✗ Built console still carries the PUBLISHED @objectstack/spec.
  

↳ 失败原因 是判读的关键：超时（Test timed out in … / Hook timed out in …）多半是负载/时序，不是本 PR 的回归；
断言（AssertionError: …）才指向真实的行为改变。两者的 FAIL 行长得一模一样，只有这一行能区分。

⚠️ 断言这一侧有一类例外，判据是断言在测什么，不是它是不是 AssertionError。 断言的对象是产品行为（一个值、一个形状、一次拒收）⇒ 照上面读：真实的行为改变，去查，⛔ 不要重排掉；
断言的对象是这次实验自身的有效性前提（跑完的耗时、负载下的先后、任何只在时间预算内才成立的条件）⇒ 它跟超时是同一类，同样对负载敏感，重排一次是合法的判别手段。
识别是机械的：断言的消息或它比较的值本身点名了一段时长、一个时间戳、一个耗时计数。实测过的一对 —— AssertionError: SecurityPlugin.init() ran: expected false to be true 测的是产品行为（真回归）；
AssertionError: this run took over a second, so second-precision stamps could have differed too: expected 1006 to be less than 1000 测的是实验前提：它守护的那条不变式当时是绿的，同一个 head 原样重排一次即成功。
穿着 AssertionError 外衣的时间测量，仍然是时间测量。（⛔ 这只改「怎么读一次红」，不改「哪些测试可以重排」——后者由别处管。）

跨 PR 相同签名（24h，按失败测试文件聚合）：

• ⚠️ 本次没有可用的聚合签名（日志里没有能解析出测试文件名的 FAIL 行）—— 这不是「没有同签名的其他 PR」，是这一轮没测到。跨 PR 聚合本次不可用，请手工比对其他 PR 的同类评论。
• ⚠️ 24h 评论账本没读完（超过 5 页仍未读到窗口尽头），所以上面的「不同 PR 数」是下界，不是全量。

历史信号：

• 本 PR 过去 24h 无队列失败记录（首次）。
• 过去 24h 队列共有 10 个失败构建（不含本次）。

分诊清单：

1. 失败测试在本 PR 改动的包里 → 真回归，修 PR。
2. 失败测试与本 PR 无关 → 看上面的「跨 PR 相同签名」；已有汇总 issue ⇒ flaky/环境问题实锤，去那张 issue 上谈，修好前重排只会再烧一轮全队列。
3. 两者都不是 → 可能与同组 PR 语义冲突；等前面的 PR 落地或失败出队后再重排一次即可，不要连续重排。

---

Generated by Claude Code · merge-queue-triage workflow (#4859)