try to add ifx support for fortran win support

[Krande]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

[Krande]

@conda-forge-admin, please rerender

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-souschef (grayskull). This parser is not currently used by conda-forge, but may be in the future. We are collecting information to see which recipes are compatible with grayskull.
• ℹ️ The recipe is not parsable by parser conda-recipe-manager. The recipe can only be automatically migrated to the new v1 format if it is parseable by conda-recipe-manager.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/15496672277. Examine the logs at this URL for more detail.}

[Krande]

@conda-forge-admin, please rerender

[Krande]

Looks like it is missing mpi.mod?

[440/3489] Building Fortran object fortran\src\CMakeFiles\hdf5_fortran-shared.dir\H5Pff.F90.obj
FAILED: fortran/src/CMakeFiles/hdf5_fortran-shared.dir/H5Pff.F90.obj mod/shared/h5p.mod 
%BUILD_PREFIX%\Library\bin\ifx.exe  /nologo /fpp -DBUILD_HDF5_DLL -DH5_BUILT_AS_DYNAMIC_LIB -DHDF5F90_WINDOWS -D_BIND_TO_CURRENT_VCLIBS_VERSION=1 -D_CONSOLE -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -Dhdf5_fortran_shared_EXPORTS -I%SRC_DIR%\fortran\src -I%SRC_DIR%\fortran\src -I%SRC_DIR%\build\mod\shared -I%SRC_DIR%\build\fortran -I%SRC_DIR%\build\fortran\src -I%SRC_DIR%\src -I%SRC_DIR%\src\H5FDsubfiling -I%SRC_DIR%\build\src -I%BUILD_PREFIX%\opt\compiler\include -I%BUILD_PREFIX%\opt\compiler\include\intel64 -I%PREFIX%\Library\include /nologo /fpp /O2 /DNDEBUG -threads -libs:dll -module:mod\shared /Fofortran\src\CMakeFiles\hdf5_fortran-shared.dir\H5Pff.F90.obj /Fdfortran\src\CMakeFiles\hdf5_fortran-shared.dir\ -c fortran\src\CMakeFiles\hdf5_fortran-shared.dir\H5Pff.F90-pp.f90
%SRC_DIR%\fortran\src\H5Pff.F90(46): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [MPI]
  USE MPI, ONLY : MPI_INTEGER_KIND
------^

https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=1229639&view=logs&j=9f906e47-be60-57ed-17d2-d9dc2d3ee176&t=a5b8a820-c99c-5475-014a-49a7b5a57d3b&l=2250

By looking for mpi.mod using https://conda-metadata-app.streamlit.app/Search_by_file_path?path=include%2Fmpi.mod, I can see that other mpi packages like mvapich and mpich, they both have a include/mpi.mod.

[Krande]

@hmaarrfk I think I have tracked down the root issue with impi support for windows using Intel Fortran IFX. There seems to be missing files in the impi-devel package. I have raised the issue in their issue-tracker.

I have a few packages relying on HDF5 fortran support on windows that could really benefit from merging only the "nompi" version. So if we can merge only the nompi variant if the impi-support takes too long, that would help a lot.

I'll keep you updated on the progress

[hmaarrfk]

I have a few packages relying on HDF5 fortran support on windows that could really benefit from merging only the "nompi" version.

Can you point me to the packages that could benefit from this on conda-forge?

I think the question is:

Can you simply release this version on your own channel giving time for the issue to be resolved in the repack repo?

The reason I ask this is that conda-forge has "grown up" a little bit, and creating incompatibilities means that now users that build with nompi + fortran need a way to "declare that" and make themselves incompatible with the impi version. This is not something we easily do.

We could mark this with an additional HDF5 output, like hdf5-fortran or something, but this is also "more work" compared to "waiting for the fix upstream and then rebuilding when done".

[Krande]

The reason I ask this is that conda-forge has "grown up" a little bit, and creating incompatibilities means that now users that build with nompi + fortran need a way to "declare that" and make themselves incompatible with the impi version. This is not something we easily do.

That's a fair point. I am fine with waiting to fix impi support.

Can you point me to the packages that could benefit from this on conda-forge?

Sure. The packages that would directly benefit are windows variants of; libmed, medcoupling and code-aster. Libmed and medcoupling can be used individually for purposes where fortran support is not necessary. However, Code_Aster requires fortran support in HDF5 to work. So It is really Code_Aster that is triggering the real need.

Can you simply release this version on your own channel giving time for the issue to be resolved in the repack repo?

Yup, I already have a separate channel dedicated to this -> https://prefix.dev/channels/code-aster. My motivation for moving away from the code-aster channel to conda-forge is primarily because packages that are intended to be used together with code-aster also rely directly/indirectly on HDF5. So in total there are quite a lot of packages I've had to re-compile.

So let's wait for the intel_repack team to reply on my issue before doing anything else

[hmaarrfk]

The packages that would directly benefit

Great yeah i remember helping with the libmed packages!

Yup, I already have a separate channel dedicated

awesome.

Hopefully intel repack helps.

[Krande]

@conda-forge-admin, please rerender