Fix commitlint configuration that considers too many commits
Fails because of a non-conventional commit in !58 (merged).
Chat from IRC/Matrix:
│11:35:48 rgdd | nielsm[m]: do you understand why https://git.glasklar.is/system-transparency/core/stprov/-/merge_requests/56 is failing? Linus broke pipeline on main, and i'm getting that │
│ | error in my MR as well. I would expect commitlint to only lint commits in my branch based on our setup -- do you understand what's wrong / have an idea of how to fix it? │
│12:15:00 nielsm[m] | No good clue, on MR runs I think it's supposed to only check the commits that are new. Maybe that rev-list command includes the base commit too, typically last commit on main? │
│ | For runs on main it makes sense to me to skip all ci lint tests (it's too late, since we don't want force pushes just to fix commit messages). │
│13:50:58 rgdd | the rev-list command works as expected locally, which is part of why i'm so clueless why it's behaving differently in the CI │
│13:51:44 rgdd | i also tried now creating a new branch that was not force-pushed to bring in the latest ("non-convential commit"), just to see if maybe there was some GL magic not beting │
│ | updated correctly on force push │
│13:51:47 rgdd | but that didn't work either │
│13:51:50 rgdd | still same error │
│13:53:25 rgdd | i have two pending MRs, both failing on this commit lint issue. I'll try merging one of them, rebase the next one, and if that one still fails (aka issue doesn't just │
│ | magically dissapear now that there's something valid on main again) i will disable commit-lint before merging. Because it's not worth the time trying to debug this more imo │
│14:28:18 nielsm[m] | Some ideas: The fetch --deepen loop ends with the output c4407a60eda6709029439c8970a7c97f545e8d76, which is a merge commit from back in October (looking at the failed job │
│ | https://git.glasklar.is/system-transparency/core/stprov/-/jobs/14157. So that's something not working as intended. │
│14:28:18 nielsm[m] | And then your MR includes unusually many commits. So my best guess is that you need to add HEAD (if that works, alternatively, name of current branch) to the fetch command, to │
│ | deepen *both* HEAD and $CI_MERGE_REQUEST_TARGET_BRANCH_NAME. │
│14:28:18 nielsm[m] | I think this is fixable, but if you want to postpone it, maybe mark the job as `allow_failure: true` for now.