Closed kgerheiser closed 2 years ago
@climbfuji
Have you ever seen this error before error: invalid argument '-std=c++11' not allowed with 'C'
when building ESMF with Clang? I don't normally use Clang. The clang + gfortran
test build fails because of it.
@climbfuji
Have you ever seen this error before
error: invalid argument '-std=c++11' not allowed with 'C'
when building ESMF with Clang? I don't normally use Clang. Theclang + gfortran
test build fails because of it.
I think yes, but I don't remember ... is it possible that it doesn't find the correct CXX compiler?
@kgerheiser what should we do with this old PR?
Whoops. I should not have approved this...
@kgerheiser @edwardhartnett What should we do with this PR?
I don't know. I don't really want to develop NCEPLIBS-external anymore if we don't have to. If it's pertinent I can try to complete this PR and get develop updated with the ufs-v2.0.0 changes.
I don't know. I don't really want to develop NCEPLIBS-external anymore if we don't have to. If it's pertinent I can try to complete this PR and get develop updated with the ufs-v2.0.0 changes.
I guess this is up to @edwardhartnett to decide. I doubt we'll ever need it again. With the two upcoming MRW and SRW releases, the previous ones won't be supported any longer and NCEPLIBS-external won't be needed.
We certainly want to retire NCEPLIBS-external as soon as feasible.
Merge release/public-v2 into develop.
Fix failing Github workflow by building MPI on macOS. Some Homebrew update causes MPI not to be found. It's good to test it both ways anyway.
Make HDF5 depend on MPI to ensure correct build order and
add_custom_target(mpi)
earlier in the script.Set
MPI_<lang>_COMPILER
when building MPI so that variable is set.Fixes #93
Fixes #92 (It builds on the runner and on my personal machine, with and without building MPI)
Fixes #85
Might need to cleanup the documentation some. I didn't handle the conflicts in too much depth.