Could you also bump cython in the ci/deps files?

Bumped. The pin there was cython>=0.29.33 (except freethreading which doesn't have any pin), but I do think having some upper bound is preferable.

From the 2 to 3 transition, I was cautious about accepting new versions and I think now overly so. What do you think about cython<4 instead?

From the 2 to 3 transition, I was cautious about accepting new versions and I think now overly so. What do you think about cython<4 instead?

Yeah I would be OK with that (modulo pinning below alpha/rc release versions since Cython often to pre-releases before major version bumps)

Owee, I'm MrMeeseeks, Look at me.

There seem to be a conflict, please backport manually. Here are approximate instructions:

1. Checkout backport branch and update it.

git checkout 2.3.x
git pull

2. Cherry pick the first parent branch of the this PR on top of the older branch:

git cherry-pick -x -m1 5c286b7ef9f63487169bed5a8fb5d31c00b9b1c3

3. You will likely have some merge/cherry-pick conflict here, fix them and commit:

git commit -am 'Backport PR #61519: BUILD: Bump Cython to 3.1'

4. Push to a named branch:

git push YOURFORK 2.3.x:auto-backport-of-pr-61519-on-2.3.x

5. Create a PR against branch 2.3.x, I would have named this PR:

"Backport PR #61519 on branch 2.3.x (BUILD: Bump Cython to 3.1)"

And apply the correct labels and milestones.

Congratulations — you did some good work! Hopefully your backport PR will be tested by the continuous integration and merged soon!

Remember to remove the Still Needs Manual Backport label once the PR gets merged.

If these instructions are inaccurate, feel free to suggest an improvement.

Thanks @rhshadrach

Why was the lower bound removed? I just tried building with cython 0.29.36 and the build failed. What is the correct lower bound?

It seemed unnecessary, our CI will pull in 3.x regardless of the presence of a lower bound or not.

I would imagine you'll need at least 3.0. If you have a use case for pandas guaranteeing support for older versions of Cython, can you open an issue.

I definitely don't care about supporting older cython, I'm just worried that not specifying a lower bound will lead to issues. My specific issue occurred with the Spack package manager, not with pip. Not all package managers will always use a newer version if an old version is already installed. Conda may or may not have similar issues, I don't know enough about its internals.

Thanks; I can update this to include >= 3.0.