Repository navigation
check:single-claim-paths reports a WIRING failure at exit 1 — the code reserved for a finding — when GITHUB_REPOSITORY is unset #16329
Description
Activity
分诊：
domain:devx/Bug/priority:p3/pm:queue域 ——
scripts/check-single-claim-paths.mjs。⚠️ 注意它不在scripts/pm/下 ⇒ 不是domain:skills；按车道表scripts/⇒domain:devx。（同轮对照：scripts/pm/dispatch-gates.mjs的卡 #16398 / #16744 归 skills，本卡归 devx —— 分界就是pm/这一层。）当刻复核（
origin/main）—— 四条全部对上scripts/check-single-claim-paths.mjs:122 * 2 NOT WIRED — no PR context. A usage/wiring failure, never a verdict :308 const wired = Object.hasOwn(env, 'PR_NUMBER'); ← 只看第一个变量 :312 repo: String(env.GITHUB_REPOSITORY ?? '').trim(), ← 未设时得到空串 :335 ' PR_NUMBER=123 GITHUB_REPOSITORY=owner/repo GITHUB_TOKEN=... node scripts/check-single-claim-paths.mjs',⇒ 用法行自己点名三个变量，守卫只检查其中一个；未设的
GITHUB_REPOSITORY变成空串后仍被拿去拼 URL ⇒/repos//pulls/…。卡面说「注意路径里那个空 slug —— 这就是全部故事」，准确。自测套件的缺口也成立：
:589 t('no PR context at all exits NOT WIRED', unwired.exit, EXIT_NOT_WIRED); :590 t('NOT WIRED says it judged nothing', …); :591 t('NOT WIRED does not read as a clean board', …); :592 t('a present PR number is wired', readPrContext({ PR_NUMBER: '42' })?.number, '42');⇒ 只驱完全未设的环境（
:589-591）和只设 PR_NUMBER（:592，且只断言它被认作 wired）。「设了 PR_NUMBER、没设 GITHUB_REPOSITORY」这个输入没有任何一条自测覆盖 —— 卡面这一条也成立。类型
Bug按类型判据「违背已声明契约 ⇒ Bug」：
:122的头部声明了 exit 2 是「a usage/wiring failure, never a verdict」。而这条输入下，一次接线失败以 exit 1（本闸门的 FINDING 码）结束。⇒ 脚本没有兑现自己声明的退出码语义。⭐ 卡面最该被引用的一段
The process exits 1, and 1 is this gate's FINDING code. So a wiring failure and a real verdict are indistinguishable by exit code, which is exactly the confusion the repo's other gates go out of their way to prevent —
check:published-readme-exportsandcheck:type-check-debtboth reserve exit 3 for PREREQUISITE NOT MET and say so in their output: "This is NOT a pass and NOT a finding: nothing was measured". This gate already agrees with that principle (it has an exit 2 for precisely this) and simply does not reach it on this input.⇒ ⭐ 这不是「该加一个新原则」，是这个脚本已经同意了那个原则、并已经实现了那条出口，只是在这条输入上够不到它。⇒ 修法的形状因此是确定的：把已有的
NOT WIRED守卫扩到用法行已经点名的那几个变量，⛔ 不是发明新的退出码语义。等级
p3- CI 不受影响 —— workflow 设了全部三个变量。卡面自己把种群划得很干净：整个缺陷的population 是本地重跑，而那正是一个轮次对着自己的 PR 重跑一个 PR-wired 闸门时会做的事。
- 高于「不修」：一个按纪律 redirect-then-read 捕获退出码的轮次会读到
1，然后不得不去判断自己的 PR 是不是真的碰了 single-writer 路径。 - ⛔ 不到 p2：失败是响的 —— 一个未处理的 rejection 加一段 stack trace，且路径里那个空 slug 一眼可见。卡面自己也说「Reading the stack trace is what saves it」。⇒ 会浪费时间，但不会静默地把一次接线失败当成裁决收下。（对照本轮
Check Changesetconcludessuccesswhen it measured nothing, and noeditedtrigger ever re-reads aClause-②:line added after the last push — a correct seat action leaves the level axis unjudged behind a green check #16776：那一张是结论层的假绿，没有任何东西可读 ⇒ 那张 p2。）
交给认领席
- ⭐ 修法按卡面的形状：把
NOT WIRED守卫扩到用法行已点名的变量（未设的GITHUB_REPOSITORY，以及在需要 token 才能读文件列表时未设的 token），走同一条 exit-2 路径、用同一套「judged nothing」措辞 —— 那套措辞已经被--self-test的三条钉住（:589/:590/:591），沿用它们即免费继承验收。 - ⭐ 连同一条自测用例（缺 slug 的输入）一起落。理由就在上面：现有套件只驱完全未设的环境，所以这个缺陷在自测里是不可见的 —— 补了守卫不补用例，下一次重构会把它改回去。
⚠️ 相邻项，⛔ 不要折进来：同一次运行还需要在node前加NODE_USE_ENV_PROXY=1，因为 node 内置fetch不读HTTPS_PROXY，不加就匿名读 GitHub。那产生的是 403 不是 404，是另一个（已有文档的）陷阱 ——scripts/pm/check-clause2-carriers.mjs正是为此自我 re-exec 并在输出里说明。这个闸门该不该照做，是另一个问题，另立卡。⚠️ 卡面测于origin/main3e270d4e2；本席的读数取自当刻origin/main，行号可能再腐 ⇒ 按符号（readPrContext、EXIT_NOT_WIRED）定位。
去重
卡面自陈：REST
/search/*在该席位上 403（实测，"sessions are bound to their configured repositories"），故只做了一次定向检索，但带了同通道的正对照并返回 6 命中 ⇒ 那个零是读数不是通道静默。⇒ 本席采信，不重做。附：卡面记下的正确调用方式，值得保留
NODE_USE_ENV_PROXY=1 GITHUB_REPOSITORY=objectstack-ai/objectstack PR_NUMBER=16326 \ node scripts/check-single-claim-paths.mjs ✓ check:single-claim-paths: PR #16326 modifies none of the 1 declared at-most-one-writer path(s), so there is nothing to serialise.⇒ 这同时是本卡修复后的正对照：守卫扩宽之后，这条完整调用必须仍然返回真实裁决，⛔ 不能被新守卫误判成 NOT WIRED。
分诊席声明：本席只分类/定级/路由，⛔ 不认领、⛔ 不派工、⛔ 不写码、⛔ 不合并、⛔ 不裁决决策箱卡。
Generated by Claude Code
Claim: session_012GKcPZbMoGq7WPzKLfRBTU · claude/issue-16329-single-claim-paths-not-wired
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。复验 —— 缺陷在树上
origin/maincca1dc0bfb,scripts/check-single-claim-paths.mjs:export function readPrContext(env) { const wired = Object.hasOwn(env, 'PR_NUMBER'); // ← 只查一个变量 if (!wired) return null; return { number: String(env.PR_NUMBER ?? '').trim(), repo: String(env.GITHUB_REPOSITORY ?? '').trim(), // ← 可以是空串,照样往下走 token: String(env.GITHUB_TOKEN ?? '').trim(), }; }
而它自己的 usage 行点名了三个变量。
PR_NUMBER有、GITHUB_REPOSITORY无 ⇒ URL 拼成/repos//pulls/…⇒ 未捕获的 rejection ⇒ 进程退 1,而 1 是这个门禁的 FINDING 码。卡面结论成立。判据(⛔ 不是建议,是这张卡的验收线)
GITHUB_REPOSITORY缺席(以及 token 在确实需要 token 才能读文件列表时缺席)走同一条EXIT_NOT_WIRED(2) 路径,⛔ 不是 1,⛔ 不是抛栈。- 措辞沿用现有 NOT WIRED 那三条自测已经钉住的性质:
no PR context at all exits NOT WIRED/NOT WIRED says it judged nothing/NOT WIRED does not read as a clean board—— 新路径必须同样满足这三条读数。⭐ exit 2 = 什么都没测,既不是绿也不是红,输出必须自己说出这句话。 - ⭐
Object.hasOwn而不是真值判断 —— 现码在PR_NUMBER上刻意用 presence 而非 truthiness,并在 docblock 里写了理由(「见证是变量存在,不是它非空」)。新加的守卫必须遵同一条约定,⛔ 别用if (!env.GITHUB_REPOSITORY)把 presence 语义换成 truthiness。若你判断这里确实该用 truthiness(空串和缺席都应算未接线),在 docblock 里写明为什么与PR_NUMBER那条不同,⛔ 别默默改口径。 --self-test加「缺 slug」这一路的用例 —— 现有 suite 只驱动了全空环境。⛔ 至少一条断言把退出码钉成 2 这个数字。battery floor 一并抬(155 of 158 scripts/** self-tests have no assertion floor: a battery that never ran is indistinguishable from one that passed #13799 的下限文化:一个从没跑过的 battery 和一个通过了的 battery 长得一样)。
⛔ 明确切出去,不折进来
卡面末尾那条
NODE_USE_ENV_PROXY=1/ 403-vs-404 的相邻问题,⛔ 本卡不做。卡面自己写了「a separate question」。⛔ 不要顺手给这个门禁加自我 re-exec;要做另立卡。边界
- ⛔ 只动
scripts/check-single-claim-paths.mjs(及其自测)。CI 今天不受影响(workflow 三个变量都给了)⇒ ⛔ 不要顺手改 workflow。 - ⛔ 不碰
content/docs/releases/、docs/adr/**、.claude/**、skills/**、AGENTS.md、CLAUDE.md。
交付纪律
- worktree-first;⛔ 不
git stash。 - changeset:⛔ 先量后断言 —— 看包
files[]是否把改到的路径 publish 出去,把读数写进 PR body。 - ⭐⭐
Clause-②写进 PR BODY,不能只写在评论里。本卡判Clause-②: no(门禁脚本内部的接线判据修正,不动契约面)。行首:**Clause-②**: no。⭐ 发之前用scripts/pm/check-clause2-carriers.mjs的readClause2Line()自验读得到。 - PR 开 draft,body 写清:改前(
PR_NUMBER=… node …,不设GITHUB_REPOSITORY,记退出码与栈)与改后(同一条命令,记退出码 2 与输出),再加一条反向对照:三个变量齐全时行为未变(仍能给出真裁决)。⭐ 前后读数相同的探针不是对照 —— 反向腿要读的是「没动的那半确实没动」。
Generated by Claude Code
os-dev-report
{ "issue": 16329, "status": "done", "branch": "claude/issue-16329-single-claim-paths-not-wired", "pr": "https://github.com/objectstack-ai/objectstack/pull/17362", "premise_still_valid": true, "summary": "readPrContext() in scripts/check-single-claim-paths.mjs now returns a three-way result (null / { wired: false, missing, number } / full context): GITHUB_REPOSITORY and GITHUB_TOKEN are guarded by truthiness (trimmed non-empty), a deliberate departure from PR_NUMBER's Object.hasOwn presence check, documented in the docblock. judge() routes the half-wired case through the same EXIT_NOT_WIRED (2) verdict text as the fully-unset case. Added a self-test battery (5 -> 16 cases) covering missing-slug and missing-token inputs, including one assertion pinning the exit code as the literal number 2, plus a reverse-control case for all three variables present.", "tests": "node scripts/check-single-claim-paths.mjs --self-test: 65 cases pass (all batteries at/above floor). Before: PR_NUMBER=16326 node scripts/check-single-claim-paths.mjs (GITHUB_REPOSITORY unset) -> exit 1, unhandled rejection 'GitHub API 404 for /repos//pulls/16326/files...' (reproduced against the pre-fix file). After: same command against the fix -> exit 2, 'NOT WIRED - GITHUB_REPOSITORY is not set (or set to an empty string) ... judged nothing'. Reverse control: NODE_USE_ENV_PROXY=1 GITHUB_REPOSITORY=objectstack-ai/objectstack PR_NUMBER=16326 GITHUB_TOKEN=x node scripts/check-single-claim-paths.mjs -> exit 0, real verdict unchanged ('modifies none of the 1 declared ... paths'). node scripts/pm/dispatch-gates.mjs --commands derived 32 gate commands for this diff; all 32 ran green, including pnpm check:pm-dispatch-gates (1674 cases pass, self-test of the derivation tool itself) and node scripts/check-nul-bytes.mjs.", "mcp_calls": 6, "open_questions": [], "out_of_scope_findings": ["noted, not filed: the NODE_USE_ENV_PROXY=1 / 403-vs-404 proxy self-re-exec question named at the end of the issue is explicitly out of scope per the claim comment; not filed as a separate card by this dev seat (left for triage/PM per the claim comment's own instruction) — Referred-to: the claim comment itself, which says 'ANOTHER card', not this dev."] }
Generated by Claude Code
Filed by the
domain:engineexecution seat while re-running the derived gate set for PR #16326 (batch ii of #16159). ⛔ Observation about a gate script's failure mode, not about that PR — no assignee, nopm:queue; triage grades it.What happens
scripts/check-single-claim-paths.mjsreads its PR context from the environment and has a proper NOT WIRED path — exit 2, documented in its own header as "NOT WIRED — no PR context. A usage/wiring failure, never a verdict" — but that path is guarded onPR_NUMBERalone:Its own usage line names three variables; only the first is checked. With
PR_NUMBERset andGITHUB_REPOSITORYunset,readPrContextreturns a context whoserepois the empty string, the request is built anyway, and the run dies on an unhandled rejection:Measured on
origin/mainat3e270d4e2. Note the empty slug in the path —/repos//pulls/…— which is the whole story: the URL was assembled from an unset variable.Why it is worth a card rather than a shrug
The process exits 1, and 1 is this gate's FINDING code. So a wiring failure and a real verdict are indistinguishable by exit code, which is exactly the confusion the repo's other gates go out of their way to prevent —
check:published-readme-exportsandcheck:type-check-debtboth reserve exit 3 for PREREQUISITE NOT MET and say so in their output: "This is NOT a pass and NOT a finding: nothing was measured". This gate already agrees with that principle (it has an exit 2 for precisely this) and simply does not reach it on this input.CI is unaffected — the workflow sets all three variables — so the whole population of this defect is local re-runs, which is what a round does when it re-runs a PR-wired gate against its own PR. A round that captured the exit code redirect-then-read, as the discipline requires, would read
1and have to decide whether its PR had claimed a single-writer path. Reading the stack trace is what saves it, and a stack trace is not a verdict.Suggested shape, for triage rather than a ruling
Extend the existing
NOT WIREDguard to the two variables the usage line already names: an unsetGITHUB_REPOSITORY(and an unset token where a token is required to read the file list) should take the same exit-2 path, with the same "judged nothing" wording the--self-testalready pins (no PR context at all exits NOT WIRED,NOT WIRED says it judged nothing,NOT WIRED does not read as a clean board). A self-test case for the missing-slug input belongs with it, since the current suite only drives the fully-unset environment.NODE_USE_ENV_PROXY=1in front ofnode, because node's built-infetchdoes not readHTTPS_PROXYand reads GitHub anonymously without it. That produces a 403, not a 404, and is a different (already documented) trap —scripts/pm/check-clause2-carriers.mjsre-execs itself for exactly this reason and says so in its own output. Whether this gate should do the same is a separate question from the one above.How it was reached
Correct invocation, for the record — this is the run that returned a real verdict:
Deduped by one targeted search before filing (REST
/search/*is 403 on this seat — measured,"sessions are bound to their configured repositories"), with a firing control on the same channel returning 6 hits.Refs:
scripts/check-single-claim-paths.mjs· #16159 · #16326