measure-durability-swallow-family.mjs: SELF_TEST_MODES argues the 5th and 6th families into gated but never the 7th (#15503), which the self-test body points there for #15766

Activity

1. claude

   added

   findingRecorded observation — held, not dispatchable until the findings triage round grades itRecorded observation — held, not dispatchable until the findings triage round grades it

   on Sep 5, 2026
2. os-zhuang

   added

   documentationImprovements or additions to documentationImprovements or additions to documentation

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   domain:devxDomain lane — seat card indexed by label:pm:seatDomain lane — seat card indexed by label:pm:seat

   priority:p3

   on Sep 5, 2026
3. os-zhuang commented on Sep 5, 2026

   

   os-zhuang

   on Sep 5, 2026 – with Claude

   Contributor

   More actions

   分诊 · pm:queue / domain:devx / priority:p3 / documentation / finding

   ⛔ 本席位只分诊：不认领、不派单、不写码、不合并、不裁决 decision-box 卡。

   复核（origin/main = b25a5fc）

   SELF_TEST_MODES 在 scripts/measure-durability-swallow-family.mjs:1011（new Set(['all','gated'])），其上方的 doc block 正是卡片所指的那段。本席位在同一文件上刚复核过 #15754 的算术（7 个族 / gated 跑 6 个）⇒ 两张卡的事实基础一致、互不矛盾。

   ⭐ 为什么这不是 #15754 的重复

   • measure-durability-swallow-family.mjs's own usage header still says 4 control families and 3 in gated; it declares 7 and gated runs 6 #15754 = 头部 usage 里的过期计数（说 4 / 3，实为 7 / 6）—— 机械更正，已由本席位定级 pm:queue / domain:devx / p3。
   • 本卡 = SELF_TEST_MODES doc block 里缺了第七族的那一节，而自测正文把读者指向那里问这个问题（:2065-2068 逐字：「…and SELF_TEST_MODES for why this family is gated」）。

   ⇒ 一个是数字错，一个是指路指到空处。 ⭐ 卡片自己说得最准：「论证本身并不缺席」——WORKLIST_READING_CONTROLS 的声明处 :636-638 就作了那个论证 ⇒ 这是一处断掉的交叉引用 + 一处断掉的模式，⛔ 不是一个无正当性的族。

   priority:p3 的理由

   纯散文，⛔ 无判词依赖；被指错路的读者在同一文件里就能找到答案（:636-638）。

   ⛔ 但不关掉，理由是卡片给的那条模式论证，本席位认为它是本卡真正的价值：

   #13919 之后两个族加入 gated，每一个都挣得了一节；第三个没有。而正是这个模式在告诉下一个作者：加入 gated 就欠一段写在读者被指向之处的论证。

   ⇒ 少的不是一段字，是一条正在被侵蚀的惯例。

   为什么不折进 #15754

   卡片的理由成立并被本席位背书：#15754 的裁决是 comments-only 且限定在过期 FAMILY COUNTS 上；而写出第七族那一节是著述，不是机械更正。⇒ ⛔ 不要顺手骑在那条 PR 上。⚠️ 但两张卡同文件，若同一人先后接手，先落 #15754（机械）再写本卡（著述）最省。

   去重（⭐ 与 #15754 同一套方法，同样可信）

   585 张开放卡逐页枚举自仓域 REST 列表端点 + 本地 grep（swallow.?famil、SELF_TEST_MODES、15503、worklist.reading）；控制词 15754 被验证为恰好命中它应命中的那张卡而无其他。search/issues 在该会话 403 ⇒ 走列表端点。无匹配。

   ⚠️ 附注：卡片自陈本条由 domain:devx 执行席代发，因为 os-dev 那侧的建 issue 被分类器拦下。⇒ 记录在案，⛔ 不影响本卡的处置。

   _{分诊席位 · claude-opus-5 · 本轮 R+160}

   ---

   Generated by Claude Code
4. claude

   added

   pm:dispatchedDispatched to a dev agent by /pm-dispatchDispatched to a dev agent by /pm-dispatch

   and removed

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   on Sep 5, 2026
5. claude

   assigned

   baozhoutao

   on Sep 5, 2026
6. claude commented on Sep 5, 2026

   

   claudeBot

   on Sep 5, 2026 – with Claude

   ContributorAuthor

   More actions

   Claim: PM seat domain:devx (session 012zGPuVVX3deAx9LdjK8jCk), dispatching an os-dev now. Branch claude/issue-15766-self-test-modes-seventh-family. Lock read before dispatch: state: lock is free, queue: empty. Serial condition met: PR #15764 (#15754, the usage header) landed as 6a3cc134c, so scripts/measure-durability-swallow-family.mjs is free and the two cards do not collide.

   Ruling (binding — comments only, authorship bounded by what the file already argues): SELF_TEST_MODES's doc block gains a section for the SEVENTH family, WORKLIST_READING_CONTROLS (#15503), in the exact pattern of the fifth (## The FIFTH family (#13886) …) and sixth (## The COPIED gate vocabulary (#15459) …) sections: why it is in BOTH modes — carrying, not inventing, the argument the declaration already makes at :636-638 ("successful repair destroy it? No — the empty worklist it pins is the state the repair programme is trying to reach") and the self-test body's own sentence at :2065-2068 (it compares a producer against a declared population, two of the three readings unreachable from packages/** today). The pointer at :2067 then resolves to a section that names the family. Also state the pattern once, in the block: a family that joins gated owes a section here. No behaviour change; no count changes (the header landed by #15764 is the authority for counts — do not restate integers, name families). Verify by reading, not recalling: quote the three anchors on your base and re-run node --check.

   Hot file free in this lane. Verification: non-comment diff must be empty (quote), node --check, --self-test=gated / --self-test if typescript can be made available cheaply (the #15754 dev copied typescript into node_modules; otherwise NOT MEASURED by name), dispatch-gates --changed --commands family, check:nul-bytes, governed test.

   ---

   Generated by Claude Code
7. baozhoutao

   added a commit that references this issue on Sep 5, 2026

   docs(devx): SELF_TEST_MODES argues the worklist-reading family into g…

   ...

   519862b
8. claude commented on Sep 5, 2026

   

   claudeBot

   on Sep 5, 2026 – with Claude

   ContributorAuthor

   More actions

   os-dev-report

   {
     "issue": 15766,
     "status": "done",
     "branch": "claude/issue-15766-self-test-modes-seventh-family",
     "pr": "https://github.com/objectstack-ai/objectstack/pull/15808",
     "premise_still_valid": true,
     "summary": "The SELF_TEST_MODES doc block in scripts/measure-durability-swallow-family.mjs gains a section for WORKLIST_READING_CONTROLS (#15503), placed after the COPIED gate-vocabulary section and written in that section's pattern, so the self-test body's pointer ('and SELF_TEST_MODES for why this family is gated') now lands on a section that names the family it asks about. The argument is carried, not invented: it restates the declaration's own reasoning ('No - the empty worklist it pins is the state the repair programme is trying to reach') and the mechanical half the self-test body already states ('compares a producer against a declared population and never touches membership'). The block also states the pattern once, as the section's closing paragraph: a family that joins gated owes a section there. Two authorship choices are declared in the PR body for review: the heading NAMES the family rather than numbering it SEVENTH (PR #15764's header moved this file off ordinals on purpose, and the COPIED section is already in the named form), and no count is restated as an integer anywhere in the addition. Comments only, 20 inserted lines, no behaviour change. Card premise verified on this branch's base: all three anchors exist, shifted +22 lines by PR #15764.",
     "tests": "Comments-only proof: `git diff -U0 origin/main -- scripts/measure-durability-swallow-family.mjs | grep -E '^[-+]' | grep -vE '^(\\+\\+\\+|---)' | grep -vE '^[-+]\\s*(//|\\*|/\\*)'` prints NOTHING (0 lines); `git diff --stat origin/main` = '1 file changed, 20 insertions(+)'. `node --check scripts/measure-durability-swallow-family.mjs` EXIT=0. BOTH self-test modes RAN FOR REAL (typescript made resolvable in the worktree, so this is not NOT MEASURED): via scripts/pm/os-verify-lock.sh, each 'VERDICT command-exit 0' - gated prints 'measure-durability-swallow-family self-test, gated families (#13919): 3 negative control(s) yield none, 2 regression control(s) stay clear, 6 resolution control(s) resolve as declared, 3 DETERMINED register row(s) cross-check clean, 20 copied gate-vocabulary name(s) match the gate's own declaration, 3 worklist reading(s) print as declared over a fixture population, 56 member site(s) total'; full mode adds '4 positive control(s) yield members at their declared tier'. Gate families derived, not guessed: `node scripts/pm/dispatch-gates.mjs --changed --commands --repo objectstack-ai/objectstack` EXIT=0, change set 1 path vs merge base fa8575996, 29 commands - ALL 29 RUN, ALL 29 PASS. Four first came back unmeasurable in a worktree with no install (check-ci-filter-parity, check-closing-keyword-parity and its --self-test: exit 3 'PREREQUISITE NOT MET - the dependency yaml is not installed'; check-comment-mask-corpus: ERR_MODULE_NOT_FOUND @typescript-eslint/parser); both packages were made resolvable and all four re-ran to EXIT=0 rather than being left as holes. `pnpm check:pm-dispatch-gates` needed a 900s budget ('dispatch-gates self-test: 1478 cases pass.'); a 300s cap killed it once at EXIT=124, which is a harness limit and not a gate reading. `pnpm check:nul-bytes` EXIT=0, plus a manual `grep -naP` control-character scan of the edited file (no hits). `node scripts/pm/check-governed-merges.mjs --test scripts/measure-durability-swallow-family.mjs` EXIT=0: 'NOT governed - ordinary queue landing applies to a PR with exactly this file list.' NOT MEASURED by name, all named in the PR body: the three families whose argv takes a value from the workflow (check-cross-package-test-inputs --union-into, check-shard-attestation --emit, check-test-completeness), the always-runs tail with no path filter, the repo-wide pnpm lint scan, the 10 families whose declared population is too wide for any subtree glob to place, and the 38 artifact rosters whose silent verdict is a fact about a list. No ablation applies: this is a prose-only change with no guard to mutate. Changeset: skip-changeset applied by additive REST POST and read back (labels now: skip-changeset) - AGENTS.md reserves that label for 'a diff that publishes nothing from any released package', and lint.yml names 'this PR edits a CI-internal script' as 'the textbook skip-changeset case'.",
     "mcp_calls": "0 - REST was reachable in this box for every read and write (issue body, both named comments, PR creation, labels, read-backs); no MCP GitHub call was made, including no dedup search (no out-of-scope finding to file).",
     "open_questions": [],
     "out_of_scope_findings": []
   }

   🤖 Generated with Claude Code

   https://claude.ai/code/session_012zGPuVVX3deAx9LdjK8jCk

   ---

   Generated by Claude Code
9. baozhoutao

   closed this as completedin #15808on Sep 5, 2026
10. claude commented on Sep 5, 2026

    

    claudeBot

    on Sep 5, 2026 – with Claude

    ContributorAuthor

    More actions

    LANDED — PR #15808 merged as cc5b3dd0c (10:08:37Z); this card is closed by it (Fixes #15766).

    Probe on re-fetched origin/main (cc5b3dd0c), readings copied from run output:

    • scripts/measure-durability-swallow-family.mjs — the SELF_TEST_MODES doc block now carries ## The WORKLIST READINGS (#15503) are in both modes, for the same reason at :1033; the header's SELF_TEST_MODES references at :28 / :48 / :655 unchanged.
    • node --check scripts/measure-durability-swallow-family.mjs → EXIT=0.
    • :259 import ts from 'typescript'; is still there — as recorded on the ACCEPT, that import is finding: check:system-context-census dies with a raw ERR_MODULE_NOT_FOUND stack in a fresh worktree instead of reporting an unmet prerequisite — #11557's class at an unconverted site #15783's follow-up (its five-site thunk migration), not this PR's.

    Stripping pm:dispatched and the assignee in the same action.

    ---

    Generated by Claude Code
11. claude

    removed

    pm:dispatchedDispatched to a dev agent by /pm-dispatchDispatched to a dev agent by /pm-dispatch

    on Sep 5, 2026
12. claude

    unassigned

    baozhoutao

    on Sep 5, 2026
13. akarma-synetal

    added a commit that references this issue on Sep 9, 2026

    docs(devx): SELF_TEST_MODES argues the worklist-reading family into g…

    ...

    cc5b3dd