Skip to content
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

ENH: Add Alpine to CI #50511

Closed
3 tasks done
WillAyd opened this issue Dec 30, 2022 · 14 comments
Closed
3 tasks done

ENH: Add Alpine to CI #50511

WillAyd opened this issue Dec 30, 2022 · 14 comments
Labels
CI Continuous Integration Enhancement

Comments

@WillAyd
Copy link
Member

WillAyd commented Dec 30, 2022

Feature Type

  • Adding new functionality to pandas

  • Changing existing functionality in pandas

  • Removing existing functionality in pandas

Problem Description

We have historically had some bug reports come from Alpine users but have never tested that platform in the past. Now that we have an open source Docker account we should be able to create an Alpine image for testing use

Related to #50046

Feature Description

see description

Alternative Solutions

I don't believe GH actions has an alpine image, but if it does that could work too

Additional Context

No response

@WillAyd WillAyd added Enhancement Needs Triage Issue that has not been reviewed by a pandas team member CI Continuous Integration and removed Needs Triage Issue that has not been reviewed by a pandas team member labels Dec 30, 2022
@lithomas1
Copy link
Member

I don't think we should commit to this until numpy tests Alpine in their CI.

IIRC, we do have a good chunk of users on Alpine. I think I ran the stats a couple of months back, and they were one of the most common platforms installing from sdist.

@WillAyd
Copy link
Member Author

WillAyd commented Dec 31, 2022

Yea my guess is Alpine will be the container of choice for any deployments of pandas within lambda functions

@kostyafarber
Copy link
Contributor

Should we open an issue in nummy asking them to add Alpine to their CI?

@WillAyd
Copy link
Member Author

WillAyd commented Jan 11, 2023

Sure that would be awesome @kostyafarber

@kostyafarber
Copy link
Contributor

@WillAyd numpy have closed the issue so I guess we can start making tracks on this now

@WillAyd
Copy link
Member Author

WillAyd commented Jan 20, 2023

Awesome!

@kostyafarber
Copy link
Contributor

Happy to assist on this if needed. Pointing towards any resources/instructions would be helpful.

@WillAyd
Copy link
Member Author

WillAyd commented Jan 29, 2023

Awesome! Yea the pandas CI is housed in the .github folder - you might be particularly interested in the Workflows therein to extend or model this after.

We need a source image to start from. I don't know if GitHub actions has a musl image to build off of. If not I would see what numpy is using as their foundation, or could also use the alpine image from #50046

@sanmai-NL
Copy link

@WillAyd You can just use the Alpine Linux image, which is the main Linux distribution using musl libc right now.

@lithomas1
Copy link
Member

Closing as this is completed.

@MrMino
Copy link

MrMino commented Aug 2, 2023

@lithomas1 @WillAyd do you have any plans for providing musllinux wheels (for PyPI)?

@WillAyd
Copy link
Member Author

WillAyd commented Aug 2, 2023

I think would be open to it but needs a community PR - would you be interested in trying that @MrMino ?

Note that we just recently approved PDEP-10 #52711 so we probably need pyarrow to distribute a musllinux wheel first.

@lithomas1
Copy link
Member

I think Matt enabled musllinux wheels already for 2.1, but the point about pyarrow not having musllinux is a good point.

@MrMino
Copy link

MrMino commented Aug 2, 2023

Yup, seems like the relevant CI stuff is already there.

I also nudged the issue on the pyarrow side, it has been discussed before PEP-656 happened.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CI Continuous Integration Enhancement
Projects
None yet
Development

No branches or pull requests

5 participants