Closed aulemahal closed 1 year ago
Well... It works on my machine though...
Looks like an issue with ESMF?
Oh I see, I thought the test failures were because of my PR, but your PRs to main fix them, nice!
The RTD error seems to be related to this issue : conda-forge/esmf-feedstock#91 and the solution is to explicitly activate the env, which I don't think RTD does...
Related RTD issue is : readthedocs/readthedocs.org#4067. We will need to patch this environment variable in the code, I'll push something.
I'm just curious, but is there any way for us to manually restart the environment using the .readthedocs.yml
? I.e.:
build:
jobs:
pre-build:
- conda deactivate
- conda activate xscen-env
Well, I thought of that and decided not to do it for 2 reasons:
conda activate
if you haven't called conda init
? Does RTD call the latter? I'm suspecting it doesn't and this is why the conventional activation mechanism isn't used.That makes sense. Was just wondering. Your fix is clear enough for me. :+1:
Pull Request Checklist:
number
) and pull request (:pull:number
) has been addedWhat kind of change does this PR introduce?
environment
files to fit what's insetup.py
(an alphabetical order)pip install -e .
on xscennumba
for the envs, it is not a direct dep and the bug related to version 0.55 is caduc with 0.56.Does this PR introduce a breaking change?
No.
Other information:
This still install
netCDF4
1.6.2, I'm not sure if this version has been fixed or not?