Test warnings with setuptools-based get_info

conda-forge / aesara-feedstock

A conda-smithy repository for aesara.

BSD 3-Clause "New" or "Revised" License

4 stars 9 forks source link

Test warnings with setuptools-based get_info #92

Closed maresb closed 2 years ago

maresb commented 2 years ago

Check for warnings as per https://github.com/aesara-devs/aesara/pull/980#issuecomment-1182502124

The warnings are already in allowed-warnings-*.txt, so we have to check the Windows CI logs.

Checklist

[X] Used a personal fork of the feedstock to propose changes
[ ] Bumped the build number (if the version is unchanged)
[ ] Reset the build number to 0 (if the version changed)
[X] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
[X] Ensured the license file is being packaged.

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

brandonwillard commented 2 years ago

I didn't see any WARN: * messages in the Windows CI logs I looked at, so perhaps that issue is solved; however, one of the Windows tests failed, but I'm not sure about exactly what that failure is indicating.

maresb commented 2 years ago

It does seem to me like the issue has been solved! :sweat_smile:

But in order to properly check for warnings regarding the loading of BLAS, it looks like I need to also import aesara.tensor which imports aesara.tensor.blas which imports aesara.tensor.blas_headers which on pypy 3.7 on Windows in the main package emits the disallowed WARNING (aesara.tensor.blas): Using NumPy C-API based implementation for BLAS functions.

I just added that import so that we can check if the warning is triggered on other platforms. (Nope, all good :heavy_check_mark:)

Next I think we can remove couldn't find executable from the allowed warnings. (All good :heavy_check_mark:)

Then we need to make sure we get to the root cause of that warning... or simply discontinue pypy 3.7 support, since pypy 3.8 and 3.9 are now supported on Conda-Forge. (I could start that migration if you'd like, but one thing at a time...)

maresb commented 2 years ago

This looks really good to me from the perspective of the warnings. I think the problem is solved.

Regarding Using NumPy C-API based implementation for BLAS functions on pypy 3.7 on Windows, I'm inclined to drop support for pypy 3.7. I don't think it's being actively maintained on Conda-Forge now that 3.8 and 3.9 are out.

maresb commented 2 years ago

This has been cherry-picked into #95.

rzyu45 commented 2 years ago

Thanks to the contributors for their efforts in addressing the issue. Now it works well on my windows.