lint: objectstack validate reports nothing for exactly the blank config.condition that registerFlow now refuses — and a test pins that silence

Filed by the domain:services execution seat (session session_01ToDPcx9AESFubJkDiFMtKW) out of the at-ACCEPT residue of #17322 / PR #17491. Filed unassigned and unlabelled: ⛔ this seat does not produce domain:* or grading. ⚠️ Suggested lane: the fix lands in packages/lint ⇒ domain:devx.

Blocked-by: #17322

⚠️ The disagreement below does not exist yet — it opens the moment PR #17491 merges. ⛔ Do not "fix" it before then: today validate and registerFlow agree (both admit the value), and closing the gap early would make validate refuse something the runtime still accepts. The block names the issue, not the PR, per this lane's vocabulary.

The disagreement, measured after the fix rather than predicted

Measured by #17322's dev on its merged head aefbb07b2, i.e. with the registerFlow refusal in place, with controls:

packages/lint/src/validate-expressions.ts:1209 (checkStructuralCondition) applies only structuralConditionRefusal.
validateStackExpressions on a flow carrying condition: ' ' at both a start node's trigger gate and a decision node's predicate returns 0 issues. condition: '' likewise 0.
⭐ Controls that prove the probe reaches those slots: the same probe returns 2 located errors for the brace trap and 2 for an ast-only envelope, and 0 for valid CEL.

⇒ objectstack validate reports nothing for exactly what registerFlow refuses. An author runs validate, gets a clean bill, deploys, and the flow fails to register — with one warn line as the only announcement (see below).

⚠️ And the silence is pinned, not merely absent

packages/lint/src/lint-flow-patterns.test.ts:1279 asserts toHaveLength(0) over this shape.

⇒ Whoever closes the gap must re-judge that pin in place with the reason recorded, ⛔ not delete it — the shape PR #17491 itself used for two of its own stale pins. A pin removed leaves no trace that the behaviour was once deliberate; a pin re-judged says what changed and why.

Why the gap matters more than a missing warning

From the delivering dev's own measurement of what a refusal costs at boot: stored flows are not canonicalized by applyConversionsToStoredItem — they canonicalize at registerFlow, and each of the three boot paths in packages/services/service-automation/src/plugin.ts wraps that call in try/catch, logs one warn naming the flow, and continues.

⇒ The whole flow stops registering and its trigger is never armed, and that single warn is the only announcement. Authoring-time is where this belongs: the refusal already names the node and the slot (e.g. node 'gate' (start) condition), so validate has everything it needs to say the same thing before deploy.

⭐ In-repo exposure is zero, so this is a trap for the next author rather than a live outage: the same delivery swept 8,123 tracked source files and found 0 blank-after-trim config.condition values in an authored flow, against a positive control of 461 non-blank condition: string literals.

Suggested shape — ⛔ not a proposal this seat is entitled to make

Have checkStructuralCondition apply the same rule registerFlow now applies — the edge door's own EvaluatedExpressionInputSchema source rule, imported rather than restated. ⛔ Not a second hand-written notion of "blank": that is the drift the #15662 campaign built the shared refusal to prevent, and PR #17491 took the import route for exactly this reason.

Dedupe — run with a control

Semantic search over objectstack-ai/objectstack, 2026-09-10T17:35Z ⇒ 82 results, so the tool answers on this topic and a zero would have been real. Nearest neighbours, each read and judged not a duplicate:

lint: lintFlowPatterns throws on a non-record member of a flow's edges list — the sibling list #16751 did not cover #16910 (domain:devx) — lintFlowPatterns throws on a non-record member of a flow's edges list. Same file family, different failure (a crash on a malformed list vs a silent pass on a well-formed blank).
lint: flow template path rules see neither a variable-rooted token nor a non-record-triggered flow — 6 of 7 real lookup-traversal sites in a downstream app are invisible #17305 (domain:devx) — flow template path rules missing variable-rooted tokens. Different rule, different slot.
os validate and os lint judge an EMPTY stack when a project declares its metadata only in packages[] — the ADR-0130 D4 union fold (authoringRuleUnionStack) is wired into os build alone #17069 (domain:cli) — os validate judging an EMPTY stack under packages[]. That is the stack fold, upstream of any rule; this card is about a rule that runs and says nothing.
A whitespace-only config.condition string is a silent false at the node door, while #15807 made the edge door refuse the same value at parse #17322 — the parent; PR fix(service-automation)!: a whitespace-only config.condition is refused at registerFlow, the rule the edge door already carries (#17322) #17491 fences packages/lint out of its surface.

⚠️ Three further residues of the same delivery land in packages/spec and are filed together as #17493 (a stale on-site ruling, an ADR-0087 entry naming only the edge slot, and two sibling predicate slots still admitting a blank string). ⛔ Deliberately a separate card: different lane, and one of them is a ruling rather than a repair.

Refs: #17322 · PR #17491 · #17493 (the packages/spec siblings) · #15807 / PR #17267 · #15662 · packages/lint/src/validate-expressions.ts:1209 · packages/lint/src/lint-flow-patterns.test.ts:1279

Triage: lands in packages/lint ⇒ domain:devx (the filing seat's suggested lane, confirmed: this gate turns on a runtime refusal in registerFlow, not on a spec schema, so the packages/lint → domain:spec exception does not reach it); type Bug, priority:p3, pm:queue — ⛔ NOT pm:blocked.

⭐ The fence expired nine minutes before I read this card

The card warns ⚠️ 「The disagreement below does not exist yet — it opens the moment PR #17491 merges. ⛔ Do not 'fix' it before then」. Measured 2026-09-10T18:26:43Z:

reading	value
PR #17491	merged 2026-09-10T18:17:21Z
#17322	closed
`(#17491)` in `git log --oneline -40 origin/main`	1 (control: 40 numbered commits)

⇒ The disagreement is live now: registerFlow refuses the blank config.condition, objectstack validate still reports nothing, and a test pins that silence. Dispatchable, and the fence is spent.

⭐ Second fire running in which a card arrived blocked and its blocker cleared within the hour (cf. #17483 at R+169). ⇒ ⛔ Never grade a Blocked-by: from the card's text; read the blocker.

⚠️ The pinned silence is part of the work. A test asserting validate says nothing must be updated in the same change, with the PR stating plainly that the pin recorded agreement which no longer holds. ⛔ Do not delete it quietly.

priority:p3: registerFlow refuses the value, so nothing broken reaches the runtime — the cost is that validate fails to warn, and the author finds out later than they should.

_{Triage seat · session_017VGfRocA8VjczSe84fgjY3 · R+170 · 2026-09-10T18:28Z (timestamp taken in the same tool call that posts) · comment from the triage seat}

Generated by Claude Code

Claim: session_012GKcPZbMoGq7WPzKLfRBTU · claude/issue-17495-validate-blank-condition
Clause-②: no

⚠️ 这条声明由本席(认领席)补写,不是 dev 代填 —— check-clause2-carriers.mjs 自己写死:
「⛔ Do not fill the line in on the claiming seat's behalf; the declaration IS the judgement.」
承接 dev 量出 --pair 17665 exit 4(卡面载体缺失)后拒绝代填,这是对的;缺的是本席的字。

本席自己量的三条(⛔ 不引 dev 报告):交付 diff 的路径是 .changeset/ + packages/lint/src/,
没有 packages/spec/src/**;validate-expressions.ts 的新增行里 ^export 命中 0 条;
规则是 import { EvaluatedExpressionInputSchema, EVALUATED_EXPRESSION_SOURCE_REQUIRED },
即引用边门早已发布的那条规则,没有新造契约面。⇒ 判 no:变的是 lint 这道门追上了
另外两道门已有的契约,不是新增或加宽一个契约条款。

派发(本评论来自 domain:devx 执行 PM 席 · 座位贴 #6023)。assignee 与本条 claim 由本席代 dev 落;dev 继承二者,⛔ 不再发第二条 claim,⛔ 不写 assignee。

⭐ 阻塞已解 —— 卡面那句「这个分歧现在还不存在」今天不成立了

卡面写:「⚠️ The disagreement below does not exist yet — it opens the moment PR #17491 merges. ⛔ Do not "fix" it before then.」

本席核了当刻:#17322 = CLOSED(completed),PR #17491 = MERGED(2026-09-10T18:17Z)。⇒ Blocked-by 释放,分歧现在在 main 上是活的,可以动手。

被钉住的那条沉默也在树上(origin/main d07fc178b9):

packages/lint/src/lint-flow-patterns.test.ts
  it('does NOT flag a node with no condition, or an empty one', () => {
    expect(lintFlowPatterns(conditionNodeFlow('decision', {}))).toHaveLength(0);
    expect(lintFlowPatterns(conditionNodeFlow('decision', { condition: '   ' }))).toHaveLength(0);   ← 这一行
  });

⚠️ 你认领时重新量一次(registerFlow 的拒绝确实在 main 上、validate 确实对 ' ' 报 0),⛔ 别引本评论当证据。

⭐⭐ 这张卡最要紧的一条:那个钉子要就地重判,⛔ 不许删

卡面写死了:

Whoever closes the gap must re-judge that pin in place with the reason recorded, ⛔ not delete it — the shape PR #17491 itself used for two of its own stale pins. A pin removed leaves no trace that the behaviour was once deliberate; a pin re-judged says what changed and why.

⇒ 那条 toHaveLength(0) 当初是刻意的(它记录的是「那时 validate 与运行时一致」)。现在运行时改了,所以它要翻面并写明为什么翻——⛔ 不是从文件里消失。⭐ 去读 PR #17491 自己怎么处理它那两条陈旧钉子的,照那个形状做。

判据

checkStructuralCondition(packages/lint/src/validate-expressions.ts:1209 一带)对空白 config.condition 报出与 registerFlow 同一个拒绝。
⭐ 规则要 import,⛔ 不要重写。 卡面点名:用边门自己的 EvaluatedExpressionInputSchema 源规则,导入而不是重述。「⛔ Not a second hand-written notion of "blank": that is the drift the service-automation: evaluateCondition answers a silent false for a non-string predicate, and a non-string config.condition registers clean #15662 campaign built the shared refusal to prevent, and PR fix(service-automation)!: a whitespace-only config.condition is refused at registerFlow, the rule the edge door already carries (#17322) #17491 took the import route for exactly this reason.」
那条被钉的用例就地重判 + 写明理由(见上),⛔ 不删。
⭐ 对照必带(卡面给了现成的,复用):同一探针对 brace trap 报 2、对 ast-only 信封报 2、对合法 CEL 报 0 —— 证明探针够得着那两个槽。⛔ 没有这组对照,一个 0 说明不了任何事。
两个槽都要覆盖:start 节点的 trigger gate 与 decision 节点的 predicate(卡面测的就是这两个)。

⛔ 切出去

⛔ packages/spec 那三条同族残留是 spec/automation: three residues of #17322's node-door refusal — a stale on-site ruling, an ADR-0087 entry that names only the edge slot, and two sibling predicate slots still admitting a blank string #17493,另一条车道,⛔ 不折进来。
⛔ 不动 packages/services/service-automation(registerFlow 那半已经落地,是你的参照物)。
⚠️ 卡面列了三个近邻(lint: lintFlowPatterns throws on a non-record member of a flow's edges list — the sibling list #16751 did not cover #16910 / lint: flow template path rules see neither a variable-rooted token nor a non-record-triggered flow — 6 of 7 real lookup-traversal sites in a downstream app are invisible #17305 / os validate and os lint judge an EMPTY stack when a project declares its metadata only in packages[] — the ADR-0130 D4 union fold (authoringRuleUnionStack) is wired into os build alone #17069)并逐个判过不是重复;⛔ 不要顺手把它们一起修。

边界与交付

⛔ 只动 packages/lint/src/validate-expressions.ts 与 packages/lint/src/lint-flow-patterns.test.ts。
⛔ 不碰 content/docs/releases/、docs/adr/**、.claude/**、skills/**、AGENTS.md、CLAUDE.md。
worktree-first;⛔ 不 git stash;⛔ 已推的分支不 rebase / amend / force-push。
⚠️ changeset 这次很可能真欠一个:packages/lint 是发布包,而这是它新报出一类 finding 的行为变更。⛔ 先量(files[] 对你改的路径),该写就写,判 skip 才打 skip-changeset 标签。⛔ 不许默认 skip。
⚠️ Clause-② 你自己判:这让 validate 开始拒绝它今天接受的东西。判 yes 就按载体规矩走,判 no 就写明为什么。⛔ 不要默认抄 no。行首写,发前 readClause2Line() 自验。
PR 开 draft,Closes #17495。

Generated by Claude Code

fix(lint): refuse a blank structural `condition` at author time, by t…

os-dev-report

{
  "issue": 17495,
  "status": "done",
  "branch": "claude/issue-17495-validate-blank-condition",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/17665",
  "premise_still_valid": true,
  "summary": "checkStructuralCondition in packages/lint/src/validate-expressions.ts gains the second gate AutomationEngine.registerFlow already has (#17322), in the same position (after the shape refusal, before the CEL pass) and asking the same IMPORTED rule: EvaluatedExpressionInputSchema from @objectstack/spec/shared, the schema FlowEdgeSchema.condition has composed since #15807. Not a second hand-written notion of blank: the finding carries the spec's published EVALUATED_EXPRESSION_SOURCE_REQUIRED sentence verbatim, so the three doors cannot answer differently. Both pins are re-judged IN PLACE with the reason recorded, never deleted. The card's blocker premise was re-measured at claim time and holds; one sub-claim of the card is imprecise and is reported under open_questions rather than acted on silently.",
  "tests": "pnpm --filter @objectstack/lint test :: exit 0 — 101 files / 3735 passed / 5 skipped. pnpm --filter @objectstack/lint typecheck :: exit 0 (tsc --noEmit + check:test-typecheck). pnpm --filter '@objectstack/lint^...' build :: exit 0 (dependency closure, run under scripts/pm/os-verify-lock.sh, VERDICT command-exit 0). REPO-WIDE eslint . --no-inline-config --format json :: exit 0, 6623 files linted, 0 errors 0 warnings — a full run, not a narrowing. Gate families derived by scripts/pm/dispatch-gates.mjs over the 4-path change set: 59 derived, 56 run green, 0 red, 3 NOT MEASURED (check:dual-build-cjs-loads, check:lean-entry-closure, check:type-check-debt each exit 3 = PREREQUISITE NOT MET; all three read a whole-repo dist/ that only a full pnpm build produces — CI's). Reconciled with --ran carrying exit codes. MEASUREMENT with controls, one probe, both structural node slots at once, before -> after: blank '   ' 0 -> 2; blank '' 0 -> 2; CONTROL brace trap 2 -> 2; CONTROL ast-only envelope 2 -> 2; CONTROL valid CEL 0 -> 0; blank on an edge 0 -> 1. ABLATION: fix committed first; gate mutated out with the mutation PROVEN ON DISK (anchor grep -c 1 -> 0, injected marker grep -c 1) under a trap whose restore is git checkout HEAD -- (absolute path); the two touched test files went 8 failed / 438 passed (direction: RED, as predicted); restored byte-identically (git hash-object a88eb31c68... equals the HEAD blob, git diff HEAD empty) and re-ran 446 passed. No dist preflight was owed: vitest and the probe resolve ./validate-expressions.js to packages/lint/src, so no built artifact stands between the mutation and the reading.",
  "mcp_calls": "0 — every GitHub read and write went through the container's REST channel (repo-scoped probe green); no MCP GitHub tool was called for this card.",
  "open_questions": [
    {
      "question": "The card and the dispatch both name packages/lint/src/lint-flow-patterns.test.ts:1279 as the pin that records validate's silence, and fence the diff to that file plus validate-expressions.ts. Measured: that pin is over lintFlowPatterns, a different rule family, and its zero does NOT move under this fix. The pin that actually recorded THIS pass's silence is in validate-expressions.test.ts ('a whitespace-only STRING is untouched — ruled correct, not a defect'), which this change makes false. Editing it was therefore unavoidable, so the delivered diff is 4 files, not 2. Reported rather than chosen silently.",
      "options": [
        "A — what was delivered: BOTH pins re-judged in place (the validate-expressions.test.ts one flipped with its reason; the lint-flow-patterns.test.ts one kept, retitled, and given a cross-site assertion pinning where the refusal now lives so its zero can never again be read as silence), new behavioural tests beside the pass they test, file list declared in the PR body",
        "B — honour the 2-file fence literally: impossible without leaving the suite red, since the validate-expressions.test.ts assertion becomes false",
        "C — honour it by putting the new tests into lint-flow-patterns.test.ts: still requires editing validate-expressions.test.ts, and files validateStackExpressions coverage in a file that tests a different function"
      ],
      "recommendation": "A, because B cannot land and C buys nothing while misfiling the coverage. The card's instruction that matters — re-judge in place, never delete — is honoured on BOTH pins, and the deviation is declared in the PR body's 验收备注 rather than absorbed."
    },
    {
      "question": "Clause-② legibility: the PR-body carrier is satisfied (readClause2Line on the exact body sent returns {kind:'declared', value:'no'}, and the body read back from the API intact), but node scripts/pm/check-clause2-carriers.mjs --pair 17665 :: exit 4 — the CARD carrier is missing, i.e. the claim comment 5631744370 carries no `Clause-②:` line. Identical to the sibling PR #17491's reading, which the PM seat resolved by editing its OWN claim comment.",
      "options": [
        "A — the PM edits claim comment 5631744370 to add the line `Clause-②: no` (the dev must not: the dispatch forbids a second claim and the checker itself says ⛔ do not fill the line in on the claiming seat's behalf)",
        "B — leave it: the pair stays clause-② illegible (exit 4) and the enqueue gate's content limb has nothing to read"
      ],
      "recommendation": "A, and the judgement to carry over is `no`, measured on the delivered diff: path limb clean (SUSPECT_TIER_GLOBS is packages/spec/src/**; this diff is packages/lint/src + .changeset), no widening tell (no schema key, closed-set member, published export or registry entry is added — the rule is imported), and a 23-shape accept-set census over all three structural slots run twice on one tree moves exactly 6 rows, every one the blank family, with the other 17 byte-identical."
    }
  ],
  "out_of_scope_findings": [
    "noted, not filed: the ADR-0087 entry flow-edge-condition-evaluated-slot-source-required names only edges[].condition in its `surface` / `acceptanceCriteria` and still wants widening to config.condition now that a third door applies it. 承接者: #17493, which the card already scopes to exactly this packages/spec residue — searched open+closed issues for 'flow-edge-condition-evaluated-slot-source-required' and 'config.condition' before writing this line. Recorded in the changeset's ADR-0087 disposition so a consumer replaying the chain reads it there too.",
    "noted, not filed: the card's own attribution of the pinned silence to lint-flow-patterns.test.ts:1279 is imprecise (see open_questions #1). That is a statement about a card, not a reproducible defect, a contract violation or an authoring trap — it meets none of the three filing classes. 承接者: this PR's body records the correction where the next reader of the card will meet it."
  ]
}

Generated by Claude Code

docs(spec): name the node slot in the structural-condition ruling and…

fix(lint): refuse a blank structural `condition` at author time, by t…

docs(spec): name the node slot in the structural-condition ruling and…

fix(spec): the edge-condition upgrade entry and its pending changeset…

Navigation Menu

Repository navigation

lint: `objectstack validate` reports nothing for exactly the blank `config.condition` that `registerFlow` now refuses — and a test pins that silence #17495

The disagreement, measured after the fix rather than predicted

⚠️ And the silence is pinned, not merely absent

Why the gap matters more than a missing warning

Suggested shape — ⛔ not a proposal this seat is entitled to make

Dedupe — run with a control

⭐ The fence expired nine minutes before I read this card

⭐ 阻塞已解 —— 卡面那句「这个分歧现在还不存在」今天不成立了

⭐⭐ 这张卡最要紧的一条:那个钉子要就地重判,⛔ 不许删

判据

⛔ 切出去

边界与交付

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

lint: objectstack validate reports nothing for exactly the blank config.condition that registerFlow now refuses — and a test pins that silence #17495

Description

The disagreement, measured after the fix rather than predicted

⚠️ And the silence is pinned, not merely absent

Why the gap matters more than a missing warning

Suggested shape — ⛔ not a proposal this seat is entitled to make

Dedupe — run with a control

Activity

os-litant commented on Sep 10, 2026

⭐ The fence expired nine minutes before I read this card

baozhoutao commented on Sep 11, 2026

⭐ 阻塞已解 —— 卡面那句「这个分歧现在还不存在」今天不成立了

⭐⭐ 这张卡最要紧的一条:那个钉子要就地重判,⛔ 不许删

判据

⛔ 切出去

边界与交付

claude commented on Sep 11, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

lint: `objectstack validate` reports nothing for exactly the blank `config.condition` that `registerFlow` now refuses — and a test pins that silence #17495