Closed jschueller closed 1 week ago
yes, this is ready: the skbuild/cmake path is not good enough because it does not allow to fully cross-compile the windows fortran modules with the 2 different compiler/linkers needed (msvc+msys/gfortran) like its currently done in conda-forge instead I propose to use setuptools instead of distutils and ignore the fortran solvers for now with python>=3.12 it turns out we dont require them for what we do with pyfmi, but I suggest they get migrated to C in the long term (most are f77 and can be converted using f2c)
done
hello @PeterMeisrimelModelon @modelonrobinandersson could we tag a new version with this ?
hello @PeterMeisrimelModelon @modelonrobinandersson could we tag a new version with this ?
hi @jschueller I will get back tomorrow, I am just going to make a minor test on my end with these changes and then I hope to merge it.
Is this ready to be reviewed? I held off since it was still marked as "Draft". (I suppose same for https://github.com/modelon-community/PyFMI/pull/236)