You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I noticed that you have an all-successful job in the workflows. The idea is great but the way it's implemented may let broken changes slip through. There are corner cases when this job would be marked as skipped which branch protection would interpret as "not red, good to go". To solve this properly, it'd have to run with if: always() and enumerate+check statuses of each dependency-job.
This is a lot more work so I've generalized all necessary in an action @ https://github.com/marketplace/actions/alls-green#why.
Another problem I foresee is that there' multiple workflows reporting the same job name all-successful so if it happens that some workflows succeed but others fail (or would be slow) to post a status (hello, eventual consistency of GH), it'd also allow merging broken changes. I normally recommend having unique check/gate job names per workflow because of this.
The text was updated successfully, but these errors were encountered:
This issue has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.
[a drive-by advice]
Hi, I noticed that you have an
all-successful
job in the workflows. The idea is great but the way it's implemented may let broken changes slip through. There are corner cases when this job would be marked asskipped
which branch protection would interpret as "not red, good to go". To solve this properly, it'd have to run withif: always()
and enumerate+check statuses of each dependency-job.This is a lot more work so I've generalized all necessary in an action @ https://github.com/marketplace/actions/alls-green#why.
Another problem I foresee is that there' multiple workflows reporting the same job name
all-successful
so if it happens that some workflows succeed but others fail (or would be slow) to post a status (hello, eventual consistency of GH), it'd also allow merging broken changes. I normally recommend having unique check/gate job names per workflow because of this.The text was updated successfully, but these errors were encountered: