https://delta.io logo
r

Robert

05/12/2023, 6:10 PM
quick question, I noticed that we have a bunch of commits on main that do not comply with conventional commits. i think this is due to us sometimes forgetting to squash commits in a pr and using another merge method. should we configure the repo to enforce this? i guess otherwise setting up generating changelogs from our commits may not work as we are hoping to 🙂
👀 1
r

rtyler

05/12/2023, 6:11 PM
it me. I have only been doing the conventional commit thing on the pull request, not writing my commit messages with that weird syntax. 😛
🤣 1
I didn't think commits were useful for changelog consideration
r

Robert

05/12/2023, 6:13 PM
there are github actions around that will do automatic semantic versioning based on the commit messages. i think in some other thread somewhere we were hoping to introduce these once we have the pr messages enforced …
“weird syntax” - not your favourite? :D
r

rtyler

05/12/2023, 6:15 PM
not everything is a fixfeatchore to me, no 🙂
r

Robert

05/12/2023, 6:17 PM
these are also ci, style, test, and build for more flexibility 😄
😀 1
w

Will Jones

05/12/2023, 6:23 PM
Oh my expectation is we should always be squashing, so that the commits in main are the conventional commit titles from the PRs
Then we can automate releases
I haven’t looked at how this is working in practice so far
r

rtyler

05/12/2023, 6:24 PM
I'm not the biggest fan of commit squashing from pull requests because it is a lossy compression of useful information for
git-blame(1)
and other tools like that (git-bisect too)
w

Will Jones

05/12/2023, 6:28 PM
If you are disciplined about making individual commits meaningful within PRs that’s awesome. I don’t think that’s typical (I certainly don’t conform to that), so I see them mostly as noise.
I guess the question is, if we don’t squash and we don’t enforce conventional commits for individual commits, can we still get the automated releases with semantic release?
r

rtyler

05/12/2023, 6:31 PM
because the checker complains about the pull request title, I honestly just thought that was all we were subjecting to the convention. I strongly agree that pull requests should turn into changelog entries
w

Will Jones

05/12/2023, 6:32 PM
Well we can decide this once we start implementing automated releases
i

Ian Joiner

05/12/2023, 8:21 PM
This does take a while to get used to but as someone who has been using it for a while I think its benefits outweigh its cost.
r

rtyler

05/12/2023, 8:55 PM
fwiw I think it would be reasonable to keep the enforcement on pull requests, encourage the use for commits, but I worry that if we enforce for each commit that we'll ward off potential contributors (how many people really know and love rebasing) For any particularly messy pull requests we always have that squash button