Repository navigation
Conversation
…the stable gate The stable publication gate needs every commit and tag, but on the CodeBuild fleet a full fetch-depth: 0 clone takes 8.6-10+ minutes and the 10-minute job timeout cancels the gate inside the checkout step. Add filter: blob:none to the checkout (commits and tags all arrive, the working tree is still populated at the tagged commit) and raise timeout-minutes to 20 as headroom. A new test pins the filter and the budget.
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS The blobless checkout targets the measured cause, a full clone that overruns the budget on CodeBuild. The timeout bump is reversible headroom. Notes
What this change shipsInventory (3 items) — 2 justifiedIntent: let the v0.11.2 stable release get past its publication gate instead of being cancelled during checkout. This is a FIX.
[FIRST-PRINCIPLES-REVIEWED] 1038169 Description read[DESCRIPTION-READ] 9346a6d3405181289590ac955c23165cfd20da52b6edd483831b48b3c7bbe97c That sha256 names the PR description this verdict read. Recompute it with |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the gate only runs [DESIGN-REVIEWED] 1038169 Description read[DESCRIPTION-READ] 4552758b997ba800315a21f13dc759295713491c0022b53fa77ca3587a1e8c0b That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
|
Superseded by #29325. |
Problem / Motivation
Goal: The v0.11.2 stable release run on
release/0.11.0gets past the Stable publication gate instead of being cancelled inside its checkout.Backport of #29301 (gate part only) to
release/0.11.0. On this branch the gate runs on the CodeBuild fleet for tag pushes, withfetch-depth: 0and a 10-minute budget. Every failed v0.11.2 gate attempt in run 41703229734 ran on CodeBuild (jobs 126151838848 and 126179898554, labelscodebuild-kirocrew-gha-linux-41703229734-1/-2) and was cancelled at 10m insideactions/checkout.Why it matters
The stable tag v0.11.2 lives on this branch, so its release run uses this branch's workflow. Until the gate fits its budget, v0.11.2 cannot publish.
Not a goal
runs-onon this branch.ubuntu-latest.test/test_workflow_bounded_checkout.py, which does not exist on this branch.What changed (motivation → approach → change)
Symptom: gate cancelled at 10m inside
actions/checkout. Root cause: a full clone (every blob of every commit) is slower than the budget on the CodeBuild fleet. The gate needs every commit and tag and the working tree at the tagged commit, but never reads older blobs.Change in
.github/workflows/release.ymlstable-gate(hand-applied;runs-ondiffers from main):filter: blob:noneon the checkout, keepingfetch-depth: 0.timeout-minutes: 10→20as headroom.Backwards compatibility
Compatible: the gate reads the same refs and files; only the clone shape and the budget change.
Removes nothing: the job keeps full commit and tag history and every check it ran.
Tests
test/test_stable_release_gate.py::test_gate_full_history_checkout_is_blobless_and_has_headroom(new): pinsfilter: blob:noneon the gate checkout andtimeout-minutes >= 20.test_gate_checks_out_full_historystill pinsfetch-depth: 0.Proof: not provable -- docs or config only
Manual verification
Required after merge: rerun the v0.11.2 release run; the Stable publication gate should finish well under 20 minutes.
Related Issues
N/A — backport of #29301.
Sibling sweep
Swept: grep -n "fetch-depth: 0" -B20 .github/workflows/*.yml on release/0.11.0 for jobs on the CodeBuild fleet with timeout-minutes of 10 or less
Sibling: .github/workflows/release.yml -- changed here
Sibling: .github/workflows/internal-content-scan.yml -- not affected -- runs on ubuntu-latest on this branch
Pattern harvest
Not generalizable: main already ratchets this class in
test/test_workflow_bounded_checkout.py; this is a backport of that fix to a release branch.Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.