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
CI has a concept of merge candidate to track which PRs are worth re-testing even if their original test set has passed, so that we are only merging PRs which have been tested against the head of main.
CI needs updating to reflect that we now expect 2 github approvals per review.
Version
CI (current)
Relevant log output
No response
The text was updated successfully, but these errors were encountered:
Fixes#14660 by using the graphQL API to query github directly. Replaces
our current parallel interpretation of reviews into a review decision,
which is brittle if we ever change review requirements in github again.
Tested by manually updating the live CI to use the test batch generated
image. Results:
- Review decisions correctly fetched from github, not based on CI's
parallel interpretation of individual reviews:
![image](https://github.com/user-attachments/assets/67c03aa9-000a-44e7-91aa-3a42d04238dc)
- No merge candidate was being incorrectly nominated (in particular,
#14645 is now considered pending, rather than approved, which is what we
are currently, incorrectly, calculating)
What happened?
CI has a concept of merge candidate to track which PRs are worth re-testing even if their original test set has passed, so that we are only merging PRs which have been tested against the head of
main
.CI needs updating to reflect that we now expect 2 github approvals per review.
Version
CI (current)
Relevant log output
No response
The text was updated successfully, but these errors were encountered: