Open missymessa opened 2 months ago
/cc @AlitzelMendez @agocke
I think this sequence summarizes the situation pretty well:
So it seems like there's a problem in 3-4.
There seems to be a delay (~10s) between pushing a new commit and start of the runtime pipelines. See where the merge button is green even though the pipelines haven't started yet.
Does the "auto-merge" feature treat neutral check statuses as "passing"?
https://docs.github.com/en/rest/checks/runs?apiVersion=2022-11-28#update-a-check-run seems to indicate merely providing a conclusion (including "neutral") is considered as a "completed" check.
A "completed" check does not mean "passing". There's a difference between check status and state. If a check is completed, it could also have "failed" or something other than "passing".
@agocke what are the branch protections for runtime? Does it include "Require status checks to pass before merging"?
If that's the case, then this is likely a bug on GitHub's side: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/incorporating-changes-from-a-pull-request/automatically-merging-a-pull-request#enabling-auto-merge
Yup we have that enabled. Sorry, it’s not clear where the bug on the GitHub side is. Is there some documentation that says that “neutral” checks should block merging?
https://github.com/octokit/octokit.net/blob/main/Octokit/Models/Common/CheckStatus.cs#L18
So, the Checks have a Status and Conclusion. A "neutral" check is not the same as a "success" (passing). If GitHub's documentation says that it observes the branch protection of only merging when the checks "pass" then either 1) GitHub treats "neutral" the same as "success" or 2) this is a bug and it shouldn't be doing it this way.
Same issue reported a while ago about neutral counting as a passing check: https://github.com/orgs/community/discussions/45229.
I assume this is by design so you can have status checks that no-op but it'd be good if GitHub clarified this.
Release Note Category
Release Note Description