Close stale PRs with merge conflicts #218
Conversation
|
looks extra close for another attempt to sieze power Eh, it's too late in the day for that shit. 👍 |
|
@reddraggone9 I've considered various means of obfuscating a power seize, but especially with a min vote of 3, it would require sweeping changes. It's simpler just to sneak it into an existing PR un-obfuscated. |
|
Well, I'll be going to sleep in short order, so you've got about eight hours to push your change and get it approved with my name on it. As is usually the case, we have chosen convenience over security 🤦♂️ |
|
At this rate I doubt I'll be able to collect the votes in 8 hours even without a malicious change. |
|
But it's already approved... It's just waiting on the window to close. |
|
Oops. Apparently GitHub will load new comments without a refresh, but not new reactions. |
|
🙆♀️ PR passed with a vote of 6 for and 0 against, with a weighted total of 6.0 and a threshold of 3.6. See merge-commit 43050ac for more details. |
|
It looks like this won't actually work until we stop setting labels every time we poll: |
|
@reddraggone9 I think it works, some PRs have been closed. If you look at those PRs something happened, like being mentioned in another PR. |
|
@PlasmaPower Can you point to a "This PR has merge conflicts, and hasn't been touched in {hours} hours. Closing." message that @chaosbot has left? |
|
@reddraggone9 Oops, I was just checking that, and I think you're right. Changing to |
#265: Fix stale PR closing mechanism Description: See discussion in #218. I think applying the conflicts label was updating `updated_at`, so the condition would never be met. I've switched it to only looking at `pushed_at`. 24 hours should still be good enough for that. :ok_woman: PR passed with a vote of 8 for and 0 against, with a weighted total of 8.0 and a threshold of 6.1. Vote record: @Germanaz0: 1 @PlasmaPower: 1 @abrad1212: 1 @amoffat: 1 @e-beach: 1 @ozyx: 1 @reddraggone9: 1 @rhengles: 1
This should take care of some of the PRs rotting at the bottom of the list.
I'm looking at
updated_atso that comments will refresh the stale timer. That way, if people are deciding how to approach the merge conflicts, the PR won't close on them.