I noticed this as part of #130 and #142 , even if it's not really related to the changes.
The previous configuration would trigger two runs when changes were made to PRs. That's hugely wasteful, increases the time-to-feedback, and simply not useful in the slightest.
Update: It actually only triggers the duplicate runs when the PRs are merged (on the merge commit), but the same reasoning holds.
I noticed this as part of #130 and #142 , even if it's not really related to the changes.
The previous configuration would trigger two runs when changes were made to PRs. That's hugely wasteful, increases the time-to-feedback, and simply not useful in the slightest.
Update: It actually only triggers the duplicate runs when the PRs are merged (on the merge commit), but the same reasoning holds.