Numpy conda package broken inside tox on Windows, Python 3.7/3.8

[goerz]

The conda-package for numpy does not work on Windows inside a tox environment. See cross-reference numpy/numpy#15537

[jan-janssen]

I tried to trigger a new build - hopefully that is sufficient #224

[jan-janssen]

For me the issue seems to be fixed now.

[jakirkham]

Thanks for the update. Let’s go ahead and close then

[basnijholt]

I still see the issue described by @goerz in the upstream issue numpy/numpy#15537 for v1.20.1:

E   ImportError: 
E   
E   IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE!
E   
E   Importing the numpy C-extensions failed. This error can happen for
E   many reasons, often due to issues with your setup or how NumPy was
E   installed.
E   
E   We have compiled some common reasons and troubleshooting tips at:
E   
E       https://numpy.org/devdocs/user/troubleshooting-importerror.html
E   
E   Please note and check the following:
E   
E     * The Python version is: Python3.8 from "d:\a\1\s\.tox\py38\python.exe"
E     * The NumPy version is: "1.20.1"
E   
E   and make sure that they are the versions you expect.
E   Please carefully study the documentation linked above for further help.
E   
E   Original error was: DLL load failed while importing _multiarray_umath: The specified module could not be found.

[rgommers]

So did you activate your conda environment, like the solution numpy/numpy#15537 (comment) says?

[basnijholt]

Unfortunately, I am not in control of that. It happens for me when using Tox with tox-conda in our CI.

The related issue is tox-dev/tox-conda#37.

[jakirkham]

It would be good if someone could reproduce the issue without tox. If that's not possible, it may simply be a tox issue

[basnijholt]

I don't have access to a Windows machine :(

However, the following might give someone with more knowledge than me some clues.

Using set CONDA_DLL_SEARCH_MODIFICATION_ENABLE=1 while using numpy from the defaults channel fixed the issue. The conda docs describe what this does, however, there it's the solution to a (seemly?) different problem.