Skip to content

ci: backport Issue Gate workflow to release/0.8.0 - #16869

Merged
bolichen97 merged 1 commit into
release/0.8.0from
ci/0.8.0-issue-gate
Oct 5, 2026
Merged

bolichen97 merged 1 commit into
release/0.8.0from
ci/0.8.0-issue-gate

Conversation

@bolichen97

@bolichen97 bolichen97 commented Oct 5, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: a pull request into release/0.8.0 gets an Issue Gate check, so PR Readiness can finish instead of waiting forever.

Every PR into release/0.8.0 now sits on PR Readiness: pending — 1 readiness check(s) still pending; waiting on Issue Gate (not started), even with every other check green. #16388 shows it on head 2d90fd42a0: 23 checks green, readiness pending for over an hour, and an edited event re-ran the other pull_request lanes but started no Issue Gate run.

Why it matters

The 0.8 line is what users install. Every backport into it is blocked on a required status that can never pass, so each one needs a maintainer to merge past PR Readiness.

Not a goal

  • Changing main's pr-readiness.yml lane list or the gate's rule.
  • Turning the gate on: it stays paused (GATE_ENFORCED: "false"), exactly as on main.
  • Backporting .github/scripts/issue_gate_refs.py or the prepare-pr scripts it imports (see below).

What changed (motivation → approach → change)

Only the workflow file is needed. Its job checks out ${{ github.event.repository.default_branch }} into base/ and runs base/.github/scripts/issue_gate_refs.py from there, so the grammar script and its pr_status.py import are always read from main, never from the release branch. A copy of them here would be dead code.

Backwards compatibility

Compatible: this adds one CI check that is paused (passes without reading the PR). Nothing that works on release/0.8.0 today stops working.

Tests

No new tests: workflow file only. The 60 test files under test/ that read .github/workflows pass on this tree (6382 passed, 2 skipped).

Manual verification

This PR's own Issue Gate run is the check: it must start and pass, and PR Readiness must stop naming it as "not started".

Related Issues

Fixes #16868

Unblocks #16388 and every other PR into release/0.8.0.

Pattern harvest

Not generalizable: a one-off gap from a release branch cut before a new pull_request lane landed on main.

Checklist

  • At most two commits (one here), with a Conventional Commits title
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable) — N/A, the file is unchanged from main
  • No secrets, credentials, or internal references in the diff

PR Readiness reads main's lane list, which waits on an Issue Gate check.
Issue Gate is a pull_request workflow, so GitHub reads it from the PR's
merge ref, and release/0.8.0 has no issue-gate.yml. Every PR into the
release branch therefore waits on a check that never starts.

Copy .github/workflows/issue-gate.yml from main unchanged. Its grammar
script is read from a checkout of the default branch at run time, so the
workflow file is the only piece the release branch needs.

Source: origin/main 2359397 (file last changed there in #16740;
introduced by #16241, #16634, #16697).

Fixes #16868
@bolichen97
bolichen97 requested a review from a team as a code owner October 5, 2026 00:37
@bolichen97
bolichen97 requested review from iamwhatever and removed request for a team October 5, 2026 00:37
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 5, 2026
@github-actions

github-actions Bot commented Oct 5, 2026 •

Copy link
Copy Markdown
Contributor

GPT 5.6 Review — ✅ no blocking findings

GPT 5.6 completed its review of f09c1ec07dead1f243e3bf0fada86d19957f42e6 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] f09c1ec

False positive or not applicable? A repository writer can comment:
/ai-review override gpt f09c1ec07dead1f243e3bf0fada86d19957f42e6: <one-sentence reason>

@github-actions

github-actions Bot commented Oct 5, 2026

Copy link
Copy Markdown
Contributor

Design Review (Fable 5) — ✅ PASS

Design-level review of f09c1ec07dead1f243e3bf0fada86d19957f42e6 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Verify this PR's own Issue Gate run started and PR Readiness stopped reporting "not started" — the fix's only proof is its own checks tab.

Suggestions

  • The gap recurs at every release cut that predates a new pull_request lane on main; a follow-up on main (pr-readiness tolerating a lane absent from the base branch, or a release-cut checklist item) retires this class of backport.
  • When maintainers flip GATE_ENFORCED on main, this copy stays silently paused; the flip PR should sweep live release branches.

[DESIGN-REVIEWED] f09c1ec

@github-actions

github-actions Bot commented Oct 5, 2026

Copy link
Copy Markdown
Contributor

First Principles Review (Fable 5) — ✅ PASS

Premise-level review of f09c1ec07dead1f243e3bf0fada86d19957f42e6 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

First-Principles-Verdict: PASS

Confirm the file is byte-identical to main @ 2359397574 — the dormant 400-line rule body earns its place only as an unmodified copy.

Not justified as shipped

  • Item 2 — rides along: the full enforcement rule and its GATE_ENFORCED switch ship dormant; declared ("it stays paused … exactly as on main"), byte-for-byte, inert while paused.
  • Item 3 — rides along: edited/labeled/unlabeled triggers now fire an extra instant-pass run on every release-branch PR edit; copied with the file, harm-free.

What this change ships

Inventory (3 items) — 1 justified

Intent: give PRs into release/0.8.0 the Issue Gate check main's readiness verdict requires, so PR Readiness can finish — FIX.

  1. PRs into release/0.8.0 now get an Issue Gate check that starts and passes, unblocking PR Readiness — justified
  2. The paused gate rule and GATE_ENFORCED switch now exist on the branch, one line from enforcement — rides along: declared, byte-identical copy, inert
  3. Body edits and label changes on release PRs now trigger an extra instant-pass run — rides along: trigger list came with the file

Defect provenance is linked (#16868, #16388) and self-verifying on this PR's own checks. The cause (readiness verdict computed from main's lane list, lanes read from the base branch) is named in the description and its general fix is excluded by the frozen Goal; every other lane main's release-cut list names exists on this branch (grepped .yml| specs in pr-readiness.yml against .github/workflows/ — 0 missing siblings). No existing mechanism on the branch provides this check.

[FIRST-PRINCIPLES-REVIEWED] f09c1ec

@github-actions

github-actions Bot commented Oct 5, 2026

Copy link
Copy Markdown
Contributor

Opus 5 Review — ✅ no blocking findings

Reviewed f09c1ec07dead1f243e3bf0fada86d19957f42e6 — this comment is updated in place on each push.

Review details

Advisory only: the backport ships the gate file without the readiness enrolment and doc section its own header claims exist, so the advertised one-line enablement would produce a gate that cannot block.

FINDING — .github/workflows/issue-gate.yml:29 — # HOW IT IS ENFORCED: \pr-readiness.yml` lists this workflow in its lane specs is false on this branch (issue-gate.ymlis absent fromMONITORED_LANESatpr-readiness.yml:213, from the workflow_run: workflows:allowlist, and from everyworkflow_specs+=(...)), and line 24's see docs/ci/ci-and-reviews.mdpoints at a doc with no Issue Gate text anywhere underdocs/, so a maintainer performing the header's "one-line change: GATE_ENFORCED: "true"" gets a lane that reds while PR Readiness— the statuspull_request_targetevaluates from *this* base branch — still publishes success → Fix: backport the threepr-readiness.ymlenrolment lines and theci-and-reviews.mdsection (untouched files, so not merge-gating), or reword lines 12/24/29 to state that onrelease/0.8.0` the lane is advisory until readiness enrols it.

[OPUS-REVIEWED] f09c1ec

Verdict parsed from the review's SHA-scoped output markers for commit f09c1ec07dead1f243e3bf0fada86d19957f42e6.

False positive or not applicable? A repository writer can comment:
/ai-review override fable f09c1ec07dead1f243e3bf0fada86d19957f42e6: <one-sentence reason>

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=d4538c96263a issue-gate.yml:29 "pr-readiness.yml lists this workflow" is false on this branch: rebutted (true for the file that runs)

Outcome: rebutted, no change.
PR Readiness runs on pull_request_target, so GitHub reads pr-readiness.yml from the default branch, never from release/0.8.0. Main's copy lists issue-gate.yml|Issue Gate in workflow_specs and in MONITORED_LANES, and that is the copy that enforces this lane on a 0.8 PR. #16388 shows it: its readiness status names "Issue Gate (not started)" although 0.8's pr-readiness.yml never mentions the lane. The file is copied byte-for-byte from main so the two copies cannot drift; editing the comment here would fork it for a sentence that is accurate about the running gate.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=a8ee8a53ab84 the gap recurs at every release cut that predates a new pull_request lane: deferred, out of this PR's goal

Outcome: deferred. This PR's goal is to unblock 0.8; changing main's readiness rules is listed under "Not a goal", and the conductor ruled main's rules stay unchanged here. A release-cut checklist item or a lane-absent tolerance on main is a main-first design change; recorded on #16868 for whoever owns the next release cut.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=a131a417cbbf when GATE_ENFORCED flips on main, this copy stays paused: accepted-and-deferred

Outcome: accepted as a known follow-up, no change here. Enforcement is main's decision and is out of this PR's goal. Recorded on #16868 so the PR that flips GATE_ENFORCED on main also sweeps live release/* branches.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=62eae48524b8 full enforcement rule and GATE_ENFORCED switch ship dormant: kept, by design

Outcome: kept, no change. The file is copied unchanged from main on purpose: a trimmed 0.8 variant would drift from main and would need its own review each time main's rule changes. While paused, the rule costs one instant-pass job and reads nothing.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=d00c12571321 edited/labeled/unlabeled triggers add an instant-pass run per PR edit: kept, by design

Outcome: kept, no change. Those triggers are what lets a body edit or a label change re-judge the gate once it is enforced, the same as on main. While paused each run exits in seconds with no API read, and cancel-in-progress collapses bursts.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Oct 5, 2026
@bolichen97
bolichen97 merged commit aa37ae7 into release/0.8.0 Oct 5, 2026
32 of 33 checks passed
@bolichen97
bolichen97 deleted the ci/0.8.0-issue-gate branch October 5, 2026 01:57
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 5, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant