New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
TST: migrate macOS CI to arm64, move x86 to cron #16321
base: main
Are you sure you want to change the base?
Conversation
Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.
|
👋 Thank you for your draft pull request! Do you know that you can use |
We were not even able to test OSX ARM64 wheels until recently, but now that we can, I guess does not hurt. But I would want @astrofrog and/or @saimn's inputs on whether a new CI job is really necessary here. As it stands, I think we would have caught this in the nightly wheel build (the scheduled ones, e.g., https://github.com/astropy/astropy/actions/runs/8778293486) but given transient nature, maybe that is not often enough, but then again, is scipy BLAS issues really our problem? |
.github/workflows/ci_workflows.yml
Outdated
|
||
- name: Python 3.11 with all optional dependencies (MacOS X, x86) | ||
macos: py311-test-alldeps | ||
posargs: --durations=50 --run-slow |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We probably don't need --durations=50 --run-slow
on both ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good point. I'll update it when we know wether your idea of moving x86 to cron gets traction. Thanks !
Moving to the arm64 job is a good thing I think since that's what osx users are more and more using, and it's a different architecture. Not sure if we need to keep the x86 one here though, maybe it could be moved to cron jobs ? |
Yes and no:
sounds reasonable to me. |
Swapping out x86 with arm64 without changing the test name would LGTM. If we change the name, then I have to go update the branch protection rule and all PRs have to rebase (again). Thanks! |
good point, I didn't think of that. Thanks ! |
Looks like we just went through the migration process, so |
I merged #16330 which would have to be reverted here along with fixes for if still necessary:
I think we should wait for scipy to sort it out first. Also see #16329 for an unmerged attempt. |
For now I'm working on producing a useful reproducer for scipy folks (the full-fledge astropy test involves several thousands lines of code so it's basically impossible to debug). I'm getting there ! |
c14c813
to
146acb2
Compare
I rebased to fix conflicts and resolve existing suggestions. CI is still expected to fail. |
146acb2
to
4253dd0
Compare
Unless the fix in scipy is backported, this will wait for scipy 1.14.0. See https://discuss.scientific-python.org/t/proposed-release-schedule-for-scipy-1-14-0/ for the proposed release schedule |
Description
In the wake of #16319 and #16320, it seems we can no longer assume that both macOS' taget archs (x86, arm64) behave similarly enough that we can just test one.
arm64 is the only arch targeted by macOS these days, but it still makes sense to test x86 which wan discountinued in Jan 2023 according to Wikipedia, so I'm adding an env rather than changing the one we have.
Also note that jobs relying on the moving tag
macos-latest
are being migrated to amr64, so I'm settingmacos-12
andmacos-14
explicitly (which is the documented way to request x86 and amr64 runners, respectively).