-
Notifications
You must be signed in to change notification settings - Fork 49
Build Wheels with ABI3 #627
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
Conversation
|
I don't know how tox works to fix the failures here, but a |
|
I cannot figure out what is wrong with the CIs. In a different env, where the wheels are built and tested, everything passes. I'm guessing it is the lockfiles, maybe the cython version there is different? |
|
Regarding Tox: we've been worried about it for a while, since tox-conda is unsupported and out-of-date. Unfortunately it has now stopped working. We've been meaning to adopt Pixi anyway so I've spent this afternoon writing a PR for that: Hopefully I can get that merged soon and then we can update your PR. |
ba039fd to
41ddd6d
Compare
|
I cannot reproduce that failure running |
Hey, this is still progress - we have new information! But yeah that's frustrating if it only errors in GHA. |
I'll remove the ABI3 from the specialized tests. As long as they build and pass all the tests in the cibuildwheel, that is what really matters. Also, we do need to keep the specialized tests in case folks are building it that way themselves. There is one expected failure and one, regarding the configuration tests, that is eluding me: |
Yeah the output is unclear - I was in a rush! As you might have guessed, this test is to make sure that we change ALL the appropriate files when we advance our supported Python versions; there are so many and it's easy to miss one. In this case it is talking about You just need to change this line: cf-units/.github/workflows/ci-locks.yml Line 28 in 19e1055
|
|
Ugh apparently there are more Python version alignments to make! This is the offending test:
Any ideas for making it less rubbish are welcomed! ... or you can just read the test and then check all the files manually. |
I don't want to tackle it in the PR, but I can give it a go later. IMO, we can try to make these a "default" latest and forget about them. Some of those can be replaced by pre-commits that updates the |
2150a52 to
49932f1
Compare
trexfeathers
left a comment
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.
OK I'm comfortable I understand this now. Just a couple of minor things @ocefpaf. Thanks!
|
Cool! Another conflict 😭 |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #627 +/- ##
=========================================
- Coverage 57.61% 0 -57.62%
=========================================
Files 61 0 -61
Lines 6293 0 -6293
Branches 1150 0 -1150
=========================================
- Hits 3626 0 -3626
+ Misses 2376 0 -2376
+ Partials 291 0 -291 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
This PR builds a single wheel for each platform using Python 3.11 ABI3. That means we drop support for 3.10 and 3.9, I can try to re-add 3.10 later but I don't think 3.9 is worth it. I would stick with SPEC 0 though, that way we can remove 3.10 now and make our lives a bit easier.
Note that, with ABI3, we don't need to rebuild the extension every new Python release. I just tested this wheel against 3.14 and it works as expected.
Closes #626
@scitools-templating: please no share prompt