Auto-rebase: put the job guard back the way it was, because PRs stopped dispatching #1172

Merged
jaylouisw merged 1 commit from fix/rebase-guard-null into main 2026-11-09 02:12:50 +00:00
Owner

The last change here added two clauses to this job's if: — a state check and a null check on head.repo. After it landed, pull requests stopped producing workflow runs entirely: the checks sat pending and never resolved, on four separate pull requests, while other branches on the same runner dispatched normally.

I could not establish whether the multi-clause expression was the cause or whether it coincided with the runner being saturated. What I do know:

  • The failure mode is indistinguishable from a queue that never drains.
  • A skipped job and a job that never exists look identical from the pull request — except that only one of them blocks a merge forever.
  • A guard that might be the reason nothing runs is not worth the null-safety it buys.

So it's back to the single comparison with a known history, and the reason is recorded in the file so nobody re-adds it without testing dispatch first.

What stays from the previous change

  • The three-way token fallback (FORGEJO_TOKEN || GITHUB_TOKEN || github.token).
  • Every failure path exiting 0 with a ::warning:: instead of reddening somebody's pull request.

Those are independent of the guard and were not implicated.

Full suite: 113 suites, 5311 assertions, 0 failed.

The last change here added two clauses to this job's `if:` — a state check and a null check on `head.repo`. **After it landed, pull requests stopped producing workflow runs entirely**: the checks sat `pending` and never resolved, on four separate pull requests, while other branches on the same runner dispatched normally. I could not establish whether the multi-clause expression was the cause or whether it coincided with the runner being saturated. What I do know: - The failure mode is **indistinguishable from a queue that never drains.** - **A skipped job and a job that never exists look identical from the pull request** — except that only one of them blocks a merge forever. - A guard that *might* be the reason nothing runs is not worth the null-safety it buys. So it's back to the single comparison with a known history, and the reason is recorded in the file so nobody re-adds it without testing dispatch first. ### What stays from the previous change - The three-way token fallback (`FORGEJO_TOKEN || GITHUB_TOKEN || github.token`). - Every failure path exiting 0 with a `::warning::` instead of reddening somebody's pull request. Those are independent of the guard and were not implicated. Full suite: 113 suites, 5311 assertions, 0 failed.
Auto-rebase: put the job guard back the way it was, because PRs stopped dispatching
Some checks failed
CI / Test (Node 18) (pull_request) Successful in 3m10s
CI / Test (Node 20) (pull_request) Successful in 2m34s
CI / Server smoke test (pull_request) Successful in 31s
CI / Docker build (pull_request) Successful in 51s
CI / Test (Node 22) (pull_request) Successful in 1m50s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Failing after 19s
9b5c9ef28d
The last change here added two clauses to this job's `if:` - a state check and a null check on
`head.repo`. After it landed, pull requests stopped producing workflow runs entirely: the checks
sat `pending` and never resolved, on four separate pull requests, while other branches on the same
runner dispatched normally.

I could not establish whether the multi-clause expression was the cause or whether it coincided
with the runner being saturated. What I do know is that the failure mode is indistinguishable from
a queue that never drains, and that **a skipped job and a job that never exists look identical from
the pull request** - except that only one of them blocks a merge forever. A guard that might be the
reason nothing runs is not worth the null-safety it buys.

So it is back to the single comparison with a known history, and the reason is recorded in the file
so nobody re-adds it without testing dispatch first.

What stays from the previous change: the three-way token fallback, and every failure path exiting 0
with a `:⚠️:` instead of reddening somebody's pull request. Those are independent of the
guard and they were not implicated.

Full suite: 113 suites, 5311 assertions, 0 failed.
Author
Owner

The auto-rebase job failed. This is a problem with the automation, not with this change.

the rebase job itself failed before it could report a reason - check the runner.

The run is at the Actions tab. Nothing was pushed, so the branch is exactly as you wrote it.

**The auto-rebase job failed.** This is a problem with the automation, not with this change. ``` the rebase job itself failed before it could report a reason - check the runner. ``` The run is at the Actions tab. Nothing was pushed, so the branch is exactly as you wrote it.
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge!1172
No description provided.