search

scipy / oldest-supported-numpy

Meta-package providing the oldest supported Numpy for a given Python version and platform
BSD 2-Clause "Simplified" License
55 stars 33 forks source link

support the IBMi platform #66

Closed zheddie closed 1 year ago

zheddie commented 2 years ago

I am adding the oldest Numpy version for IBM i , which include our fixes for IBM i platform. Please help to review. thanks.

@rgommers need your help again. thx.

zheddie commented 2 years ago

if some tags needed to be bundled into the next release/tag, please also help to add it. thanks.

zheddie commented 2 years ago

Curious is that SciPy can be successfully installed on IBM i now. I am assuming it is using the latest Numpy (> 1.23.3) in it's sandbox during install. @rgommers

zheddie commented 2 years ago

@rgommers , any more comments ? Thanks.

zheddie commented 2 years ago

@rgommers can you help to review it again? thanks.

rgommers commented 2 years ago

The test_has_at_most_one_pinned_dependency[CPython-3.9-OS400-s390x] failure is kinda what I expected. Let me have a quick go at fixing that up.

zheddie commented 1 year ago

@rgommers , Hi Ralf, I refined the code. Wish this could help to fix the auto-checkk failure. How can I trigger the auto-check again? thanks.

rgommers commented 1 year ago

Okay, I pushed some tweaks - this LGTM now, in it goes. Thanks @zheddie

rgommers commented 1 year ago

New version released: https://pypi.org/project/oldest-supported-numpy/2022.11.19