Towncrier is used by NumPy and we are considering it for scikit-image

Had looked at that one previously as well. 👍 All the better if your experiences have been good.

We have a draft PR and were planning to start using it in the near future, so I don't have much experience with it yet. I was not aware of the ansible one you linked above, but it sounds like something similar.

Another approach that SciPy has used in the past is having contributors update a wiki page that the release manager can then pull from when making the release notes. I'm not sure how consistently it is applied, though.

Towncrier or antsibull will solve the conflicts problem, but there is still the separate problem of prompting contributors to create the fragments needed!

I'd hope it'd be easy enough to add an action that says, "must have fragment". Worst case: we can push it to the PRs manually. 😉

Another possible option: https://sphinx-github-changelog.readthedocs.io/en/latest/ (...as I struggle with the anchors for pre-releases ...)

cc: @MSanKeys963

• @jakirkham suggests https://github.com/regro/rever
• @alimanfoo suggests looking at how dask does it.

But generally, 👍 for:

• keeping it simple
• having a check on all PRs
• ability to break the current formatting & move to something new (e.g. with zarr-python 3.x)

We've used rever in conda-smithy

Josh mentioned this changelog proposal.

Another option somewhat similar to towncrier is scriv

@MSanKeys963: this is something to review before we (potentially) launch the outreachy project.

Sure. Adding to the list.
We can let the Outreachy intern experiment with multiple tools before we finalise a single one for Zarr.

+1 though it would be good to make sure that those experimenting get credit for that (even if it doesn't wind up merged). The experimenting is itself a useful step 🙂