Im Trying to solve this ''BUG: np.astype(copy=False) copies for scalars #29455''

Thank you, but I think we should probably consider this a bug, although I could imagine it being common enough that I am not 100% sure about it (and it might even be good to go via a deprecation, even though copy=False is very new).

I wonder what others think about this?

Thanks for the feedback! I see your point about this being more of a bug than just a documentation issue. I’m happy to adjust the PR depending on what the maintainers decide (e.g., updating docs, working on a deprecation path, or helping with a fix). Please let me know the preferred next step.

I agree it doesn't make much sense to document this bad behavior. We should just fix it.

Also for future documentation-only PRs, consider adding [skip cirrus] [skip actions] [skip azp] to your commit message. That will only run the CI on circleci, where we do the doc builds and other doctests.

Thanks for the clarification! I understand now that it makes more sense to fix the behavior instead of documenting it.

I have a branch that addresses part of the issue main...eendebakpt:numpy:astype. It relaxes the documentation a bit and (for the common case) avoids the copy.

@meet-vasita If you are interested, feel free to convert the branch into a PR.

Thanks for sharing this! 🙏 I’d be happy to pick up your branch and convert it into a PR. I’ll test it locally and update the docs/tests as needed