ACP process for libs changes?


I resolved a baseline conflict in my PR today and somehow triggered rustbot to comment it as a new PR. I'm OK with that, but curiously I noticed that rustbot comment is different from the original comment when I opened the PR.

Specifically, rustbot mentions something new:

If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already.

Is the ACP process required for any PR that proposes new API in std libs? For some reason, I was not aware of ACP before, should I create one and go through the process now even after the PR was opened?

Btw, I'm asking here instead of in the PR hoping to have a better chance to get help. If that's not the proper way to do it, my apologies.


The ACP process is new, thus not seeing it before. Existing PRs probably don't need to make an ACP issue unless requested.

Exception: PRs for API additions currently using issue="none" and "will add a tracking issue if this sounds reasonable" should get an ACP; the ACP is basically for the "does this seem reasonable" approval for tracking issue by T-libs-api.

This is due to the migration from highfive to rustbot. PRs triaged by highfive will be retriggered on new content by rustbot; this is a known limitation, but not one considered worth fixing due to the fixed number of impacted PRs.

A small note not directly related to the question but to the PR and conflict resolution. If I am not mistaken it looks like you added a merge commit to resolve to conflict, but PR's should not contain merge commits based on the projects workflow.

Thanks for letting me know. The merge commit appeared as I used GitHub's online tool to resolve the conflict. The tool seems to be convenient so I tried it. Won't use it again.

What's the impact of the merge commit? If I do a force-push again, will that solve the issue?

I believe it is the project's police to not merge PRs which contain merge commits. If there is no merge commit between you branch of point from master and the commit you are force-pushing then yes that should solve the problem.

Thanks for the info. I think my PR probably falls in the "Exception" you mentioned as it does not have an issue number yet. I will try to ask the reviewers to confirm that's the case.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.