-
Notifications
You must be signed in to change notification settings - Fork 2
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
Python / Numpy compilation and cross compatibility #13
Comments
What's the error message when it fails ? Compiling main is PIA (sha256 will change multiple time a day), I would stick to released version with patch. To generate a patch, you can
|
Mainly during one of the API steps, i.e.
It can be during the wrap, the first or second step, it depends.. I haven't printed out the whole compilation process, just the outcome of it.
Thanks that's a great hack! I had seen in one of your
I will finish the first matrix build I started and add a second table with the patch fix |
I opened conda-forge/staged-recipes#24529 |
With patch in PR, il was able to build on py3.11 and numpy 1.25 locally but it fails with same recipe on azure pipeline |
what is the error on azure?
|
I have this error, would you know why? Using :
|
Check you have an empty line at end of patch file |
It worked! I'll matrix build overnight and update the table tomorrow. This will give us insight for the constraints to put on numpy on conda-forge. |
@sebourban @nicogodet full linux versions (MPI/MUMPS/API/AED/GOTM) are available on my anaconda channel for the |
You can add a |
EDIT (after applying patch provided by @nicogodet, see details below):
Recapitulative table:
❌ Compiling fails
🔶 API fails
✔️ everything works
TODO:
mtri.triangulation.Triangulation/mtri.Triangulation
to apply depending on the numpy versionNote:
for the API test, a simple test was done :
@nicogodet
validate_telemac.py
does not work because package does not contain the examples. I f you have a fix for this, I'm keen to hear.The text was updated successfully, but these errors were encountered: