Release pipeline: pull_request:closed event for PR #190 did not reach Actions (no auto-version run) #191

Closed
opened 2026-09-06 15:35:51 +01:00 by conductor · 2 comments
Owner

Plain summary

Merging PR #190 (retroactive semver 0.54.0) produced NO auto-version run at all — the pull_request: closed webhook never reached Actions. The v0.54.0 release was created manually via the releases API as an idempotent fallback. PR #189's merge (v0.51.0) fired the workflow normally, so the gap is intermittent.

Evidence

  • PR #190 merged=true, closed_at 2026-09-06T15:01:44+01:00, merge commit 2db0c3d.
  • Searched all runs after the merge: only the push-triggered CI run (828) exists; zero runs carry a release job. A 90-second wait confirmed nothing queued either.
  • Control: PR #189's merge fired release run 817 minutes earlier. Run 828 (push of the same merge commit) succeeded, so the runner and webhooks were alive.

Suspected causes (to investigate)

  1. Gitea 1.27.2 event-delivery gap for pull_request: closed when the merge is performed via API (Do: merge) immediately followed by branch deletion (the API DELETE of the head branch may race the closed-event dispatch).
  2. Actions event queue drop under concurrent scheduled runs (hourly pipeline-gate at :23 was near).

Acceptance

  • Reproduce: merge two consecutive no-bump PRs via API with immediate head-branch deletion; both must produce release runs.
  • Fix the root cause (or add a retry/detection step in the workflow: e.g., a push-triggered release check that runs when master's version != latest tag).
  • No release should ever require manual API fallback again.

How Nick can verify this

Merge any no-bump PR; within ~1 minute an auto-version run appears and a release/tag is created without manual steps.

## Plain summary Merging PR #190 (retroactive semver 0.54.0) produced NO auto-version run at all — the `pull_request: closed` webhook never reached Actions. The v0.54.0 release was created manually via the releases API as an idempotent fallback. PR #189's merge (v0.51.0) fired the workflow normally, so the gap is intermittent. ## Evidence - PR #190 merged=true, closed_at 2026-09-06T15:01:44+01:00, merge commit 2db0c3d. - Searched all runs after the merge: only the push-triggered CI run (828) exists; zero runs carry a release job. A 90-second wait confirmed nothing queued either. - Control: PR #189's merge fired release run 817 minutes earlier. Run 828 (push of the same merge commit) succeeded, so the runner and webhooks were alive. ## Suspected causes (to investigate) 1. Gitea 1.27.2 event-delivery gap for `pull_request: closed` when the merge is performed via API (`Do: merge`) immediately followed by branch deletion (the API DELETE of the head branch may race the closed-event dispatch). 2. Actions event queue drop under concurrent scheduled runs (hourly pipeline-gate at :23 was near). ## Acceptance - Reproduce: merge two consecutive no-bump PRs via API with immediate head-branch deletion; both must produce release runs. - Fix the root cause (or add a retry/detection step in the workflow: e.g., a push-triggered release check that runs when master's __version__ != latest tag). - No release should ever require manual API fallback again. ## How Nick can verify this Merge any no-bump PR; within ~1 minute an auto-version run appears and a release/tag is created without manual steps.

⚠️ Priority gate: this issue has no priority. Assign exactly one of P0 — critical (spine/correctness/blocks), P1 — high (material value or real defect), P2 — medium (quality/consolidation), P3 — low (measurement/deferred). An issue without a priority is not filed.

⚠️ **Priority gate:** this issue has no priority. Assign exactly one of **P0 — critical** (spine/correctness/blocks), **P1 — high** (material value or real defect), **P2 — medium** (quality/consolidation), **P3 — low** (measurement/deferred). An issue without a priority is not filed.
gitea-actions Bot added the bugcustodianinfrastructure labels 2026-09-06 15:35:56 +01:00
Author
Owner

Duplicate of #192 (both created due to an ambiguous first response — the issue actually landed). Keeping #192.

Duplicate of #192 (both created due to an ambiguous first response — the issue actually landed). Keeping #192.
Sign in to join this conversation.
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: conductor/rune#191