Uh oh!
There was an error while loading. Please reload this page.
fix(pm): complete pm label vocabulary (pm:seat, priority:p0) and add desc-cap guard - #10116
Merged
Conversation
Both are rows of the pm state model (SKILL.md "State model") with no `gh label create` entry anywhere in ensure-pm-labels.sh, so on a repo where the label object does not exist yet the first application has to remember to create it and a rerun of the script can never repair the gap. Placement is measured, not assumed: - `priority:p0` -> the FOUR-REPO loop. Its consumers are the lane pull order and the half-state sweep's H10/H11, and that sweep is repo-parameterized (PM_SWEEP_REPO). Measured 2026-08-20: objectui carries a live priority:p0 card whose label object was auto-created by that first application, with GitHub's default colour and an empty description. - `pm:seat` -> the MAIN-REPO-ONLY section. There is exactly one seat post per `domain:*` / `repo:*` lane and both families are main-repo-only, so the `label:pm:seat` index can only be read here. Measured 2026-08-20: zero pm:seat cards in objectui and objectos. Neither label is named in the header's deliberate-absence records (the retired lanes; the objectos lane/target exclusions), so the absence was drift. Both descriptions are within the 100-character cap the header warns about (95 and 92 characters). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AeA3nU1B5Q2pgxqxgUrexd
scripts/pm/ensure-pm-labels.sh documents a silent failure mode that nothing enforced: GitHub hard-caps label descriptions at 100 characters, `gh label create` 422s above that, and the `|| true` that makes the script rerunnable swallows the 422 — so an over-long -d means the label is NEVER created on a repo where it does not exist yet, and a rerun can never repair it. The only enforcement was the author counting characters by hand, which has already been paid once by a PR that trimmed the descriptions that had drifted over. Headroom is thin, so this is not theoretical: measured on this branch, the longest live description (repo:objectui) is exactly 100 characters. The measure is CHARACTERS, never bytes: the live needs:contract-review description is 97 characters / 101 bytes and is accepted by GitHub, so a naive byte-length guard would red a label that works. Length is counted in Unicode code points rather than UTF-16 code units, so an astral character counts once. Shell expansion is part of the measurement: a description interpolating a loop variable is measured after substituting that variable's longest value, because what GitHub receives is the expanded string. A variable the checker cannot resolve is red, not skipped. Two defences against a parser that has silently stopped matching: a LOWER bound on the number of descriptions found (not a ratchet — a "the parse still finds the file" assertion), and a self-test that puts a fake `gh` on PATH, runs the real script, and checks the descriptions it was actually invoked with against the ones parsed statically. The script has no dry-run mode of its own, which is the other half of why nothing measured this before. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AeA3nU1B5Q2pgxqxgUrexd
This was referenced Aug 20, 2026
os-warren
marked this pull request as ready for review
August 20, 2026 04:57
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32333782612 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32333809661 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
github-merge-queueBot
removed this pull request from the merge queue due to failed status checks
Aug 20, 2026
Uh oh!
There was an error while loading. Please reload this page.
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32334243432 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32334639902 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32335265453 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
github-merge-queueBot
removed this pull request from the merge queue due to failed status checks
Aug 20, 2026
Uh oh!
There was an error while loading. Please reload this page.
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32335487463 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
Contributor
⛔ merge queue 构建失败 — 先分诊，再决定要不要重排队列构建 32335889263 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集）， 失败的 job（日志抽取，best effort）：
历史信号：
分诊清单：
Generated by Claude Code · merge-queue-triage workflow (#4859) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes#10098
Scope was expanded by consolidation. The gate half's spec authority is #10099 — a consolidated duplicate of this card, no longer open. Two halves, one commit each.
1. Vocabulary lines (
3ca1b8a)pm:seatandpriority:p0are rows of the pm state model (SKILL.md "State model") with nogh label createentry anywhere inscripts/pm/ensure-pm-labels.sh. On a repo where the label object does not exist yet, the first application has to remember to create it, and a rerun of the script can never repair the gap.Placement was measured, not assumed — the card left this explicitly unpinned, so both readings were checked against live consumers rather than argued from shape:
priority:p0scripts/pm/check-half-states.mjs), and that sweep is repo-parameterized (PM_SWEEP_REPO). Measured 2026-08-20: objectui carries a livepriority:p0card (objectui#5304).pm:seatdomain:*/repo:*lane, and both families are main-repo-only, so thelabel:pm:seatindex can only be read here. Measured 2026-08-20: zeropm:seatcards in objectui and objectos.The objectui
priority:p0label object exists but was auto-created by that first application: GitHub's default colouredededand an empty description. That is exactly the drift this row closes for the next repo. Creation stays create-if-missing, so objects that already exist keep their current colour and description; aligning those is a separate deliberate PM action, stated in the comment block the same way the retired label objects are.Neither label is named in the header's deliberate-absence records (the retired lanes; the objectos lane/target exclusions), so the absence was drift rather than intent — the card's second open question, answered.
Both descriptions are within the cap: 95 and 92 characters.
Boundary: objectstack-ai/cloud is not reachable from this session, so the cloud half of each measurement is unverified. It does not change either placement —
priority:p0goes in the loop that already includes cloud, andpm:seat's main-repo-only argument rests on the lane families, not on a cloud card count.2. Desc-cap guard gate (
0efb508)New
scripts/pm/check-label-desc-cap.mjs, wired ascheck:pm-label-desc-capand run unconditionally inlint.ymlbeside the other self-tested pm gates.GitHub hard-caps label descriptions at 100 characters;
gh label create422s above it, and the|| truethat makes the vocabulary script rerunnable swallows the 422. The script header documented all of this and nothing enforced it. Headroom is thin: the longest live description is exactly 100 characters (repo:objectui).Design points that are load-bearing:
needs:contract-reviewdescription is 97 characters / 101 bytes and is accepted by GitHub. Measured leg: pointing the same parse at a byte length reds exactly that one live, working label, while the shipped character guard is clean. Counted in Unicode code points, not UTF-16 code units, so an astral character counts once.ghon PATH, runs the real script, and checks the descriptions it was actually invoked with against the ones parsed statically. That dry-run idea comes from [finding] ensure-pm-labels.sh documents a silent 100-char label-description failure mode with no mechanical guard behind it #10099's verification note; the script has no dry-run mode of its own, which is the other half of why nothing measured this before.The gate self-routes:
dispatch-gates.mjsmatches it for a card touchingscripts/pm/ensure-pm-labels.shthrough its own module-body path constant.Reverse verification
Each leg's direction was predicted in writing first, then observed.
repo:objectui's description, 100 to 101. Gate exit 1:repo:objectui — description is 101 characters, over GitHub's 100-character cap. The self-test went red independently on two cases, including the fake-ghleg (no description GitHub would 422 reaches the fake gh) — the runtime path, not only the static parse. Restored byte-identical:git hash-objectequals the HEAD blob shaec076de35b768d434c7e5c5efb1c225dce11e8c5.needs:contract-review (97 chars / 101 bytes); the character guard reds nothing.ghon PATH, real script:priority:p0invoked 4 times, one per repo;pm:seatinvoked once, main repo only. Baseline leg on theorigin/maincopy of the script: 48 create calls, 0 and 0. After: 53.One self-test expectation of mine was wrong on the first run and the self-test caught it (
v17/v18are the same length, so the longest-value tie keeps the first) — the case now pins the tie-break behaviour instead.Gates
Union derived from the real diff with
node scripts/pm/dispatch-gates.mjs(no paths — the script takes its own change set off the merge base), at0efb508, after the final commit:Green:
check:pm-label-desc-cap,check:node-version,check:required-contexts,check:workflow-status-functions,check:shard-attestation,check:cross-package-test-inputs,check:type-check-coverage,check:pm-dispatch-gates,check:nul-bytes.bash -non the shell file: clean. Control-byte scan of every edited file: clean.check:type-check-debtexits 1 in this worktree on a build-state precondition, not a verdict:--re-measure cannot run: 55 workspace dependenc(ies) of the ledgered packages have no built type entry point on disk. Control run: revertinglint.ymlandpackage.jsontoorigin/mainin place reproduces the identical error, so it is the unbuilt worktree, not this diff. CI builds before it runs.No changeset:
scripts/**,.github/workflows/**and a rootpackage.jsonscript entry publish nothing.Generated by Claude Code