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

Test nightly dependencies when possible in CI ? #37031

Open
1 task done
Carreau opened this issue Jan 8, 2024 · 3 comments
Open
1 task done

Test nightly dependencies when possible in CI ? #37031

Carreau opened this issue Jan 8, 2024 · 3 comments

Comments

@Carreau
Copy link

Carreau commented Jan 8, 2024

Problem Description

in #36975 IPython 8.19 broke sage, and I did a 8.20 release.
It would be great to test pre-release of dependencies to possible warn me ahead of time.

Proposed Solution

A number of nightly wheel are available here:

https://anaconda.org/scientific-python-nightly-wheels/repo (see Spec 4) using --pre --upgrade --extra-index-url https://pypi.anaconda.org/scientific-python-nightly-wheels/simple in the right place should be enough.

I guess weekly cron that open an issue on failure would be enough.

Alternatives Considered

Do nothing, or have upstream test this repo.

Additional Information

I guess it might also catch early issue in SciPy/Numpy...

It is of course an extra burden as nightly is more likely to fail, and maintainer need to decide if it's true failures.

Is there an existing issue for this?

  • I have searched the existing issues for a bug report that matches the one I want to file, without success.
@mkoeppe
Copy link
Member

mkoeppe commented Jan 8, 2024

In the direction of the alternative "have upstream test this repo":

@tobiasdiez
Copy link
Contributor

What about having one central place (scientific ecosystem ci), that runs the current branches of registered projects against the latest versions of their dependencies (or against a certain PR). Something similar to https://github.com/vitejs/vite-ecosystem-ci.

@Carreau
Copy link
Author

Carreau commented Jan 10, 2024

IMHO a central place to test all does not work well as each project need to be tested slightly differently.

In addition if you run the main branches, then they need to be built, which is always a bit different for each project.

Then when there is a failure, one need to decide who is responsible for investigating, and in which project the failure lies.

I don't think the scientifc Python ecosystem is standardized enough yet for that.

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

No branches or pull requests

3 participants