Overall I’m supportive. It looks like there is also prior art for an ASF project to do something like this (although it’s not clear how that bot is run). There is also some mention of preserving provenance on the ASF mail archives (e.g. suppose the bot squashes and merges a branch of commits, but not all commits in the branch were authored or committed by the GH contributor).
There is also a question in my mind about who should be allowed to command the bot to merge. The prior art restricts this capability to committers only. I realize this seems like a bit of an annoyance…there are a couple reasons why I could see this being useful:
- Suppose multiple committers have reviewed and one wants to approve the PR but not merge it until the others have looked the PR over. Traditionally we’d do this with a PR comment, but a merge bot couldn’t distinguish the difference in this scenario.
- Suppose we want to approve but hold a PR until another PR lands. In this scenario I usually just comment LGTM but don’t explicitly approve, so people could just get used to that. But it means people do need to be a bit more intentional about approving things.
Would be great to get some other feedback from the community about this.
cc @junrushao @tqchen @jroesch @kparzysz @manupa-arm @ramana-arm @Mousius