You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There's a check in ompi_configure_options.m4 for python 3.6 or higher to support generation of the top level c and fortran f08 bindings from the code in ompi/mpi/bindings. That's fine, but the file being used as a sentinel to tell whether or not the bindings have already been generated is not correct. the names of the generated files has changed and ompi_send.c is not being generated, rather a differently named file.
The text was updated successfully, but these errors were encountered:
I also think we probably need an explicit check in OMPI's configure for Python. At the moment, we have a Sphinx setup macro which requires python, that's only if you're building the docs (which obviously is optional).
Also right now, the only python required for a tarball build is for the Fortran MPI bindings (per #13231). So we could protect the python check behind a few things:
if you're building the fortran MPI bindings, or
if you're in a git clone that doesn't have various sentinel files, etc., or
There's a check in
ompi_configure_options.m4
for python 3.6 or higher to support generation of the top level c and fortran f08 bindings from the code in ompi/mpi/bindings. That's fine, but the file being used as a sentinel to tell whether or not the bindings have already been generated is not correct. the names of the generated files has changed and ompi_send.c is not being generated, rather a differently named file.The text was updated successfully, but these errors were encountered: