P. L. Lim

Please add change log. I think this is performance bug, so can be under bugfix. Thanks!

Why does bottleneck brings back bad vague memories... Maybe @larrybradley remembers...

> Does GH Actions honour `[ci skip]`? Yes, they do now. They skip more things than I like but something is better than nothing.

Re: TESS -- Timing aside, are the answers at least more correct with `astropy` v5? I would say slightly slower run for better result is a good trade-off?

Sounds like a reasonable request, especially given you were the one who added it in https://github.com/astropy/astropy/pull/11444 . 😸

I converted this to draft so this won't accidentally go into v5.2. Thanks for your understanding!

Hello! Can you please provide a minimally reproducible example so others can reproduce your problem, and also tell us what the correct answer should look like, if possible? Thanks!

p.s. Also the version of astropy and numpy that you use.

@mcara , have you memorized the standards by now? 😆

Interestingly, this is not caught by existing tests! https://github.com/astropy/astropy/blob/e183e398d8bdbbe5c7ccec7a793afed2ecdac4e1/astropy/io/fits/tests/test_diff.py#L239-L257