Include documentation for building/running SRW App on Mac

[gspetro-NOAA]

DESCRIPTION OF CHANGES:

• Updated hpc-stack submodule to reflect new instructions for building the hpc-stack on MacOS
• Added section to BuildRunSRW.rst on configuring the build on MacOS

TESTS CONDUCTED:

None required. @natalie-perlin and @laurenfrederick have built the hpc-stack on MacOS using the submodule info; Natalie has build the SRW on MacOS following the updated procedure in BuildRunSRW.rst.

DEPENDENCIES:

None.

DOCUMENTATION:

If this PR is contributing new capabilities that need to be documented, please also include updates to the RST files (docs/UsersGuide/source) as supporting material.

ISSUE:

Resolves Issue #222 .

CONTRIBUTORS (optional):

@natalie-perlin @laurenfrederick

[natalie-perlin]

The syntax above and below is for the csh. Instructions for the MacOS followed use of bash shell, as is stated in the line 197 above! and the build_macos_gnu needs to use bash syntax to be consistent. The changes for the bash shell syntax to be made for the lines 204, 207, 214-216, 219-221, are shown after each corresponding line.

[danielabdi-noaa]

One of the motivation for using modulefiles is to set the environment irrespective of the login shell used (bash or tcsh).

  module load build_macos_gnu

should work for bash even though it is written in "csh-like" syntax. The only place where login shell dependent code is needed is when loading Lmod itself i.e.

  source etc/lmod-setup.sh macos
             or
  source etc/lmod-setup.csh macos

[gspetro-NOAA]

@natalie-perlin This code is already in the modulefile, which is written in csh-like syntax. In reality, the user is only changing the path. If we change this to bash, we have to redo the entire modulefile, which I don't think is worth it since, as @danielabdi-noaa pointed out, the modulefile should work regardless of the user's shell.

[natalie-perlin]

export HPCstack=/Users/username/hpc-stack/install

[danielabdi-noaa]

TCL modulefiles use setenv instead of export so I think it is fine as it is.

[natalie-perlin]

export SRW=/Users/username/ufs-srweather-app

[natalie-perlin]

export CC=/opt/homebrew/bin/gcc
export CXX=/opt/homebrew/bin/g++
export FC=/opt/homebrew/bin/gfortran

[natalie-perlin]

#export CC=/usr/local/bin/gcc
#export CXX=/usr/local/bin/g++
#export FC=/usr/local/bin/gfortran

[natalie-perlin]

Also, the following were set for the in build_macos_gnu after lines 222 (again, for bash shell syntax):

export MPI_CC=mpicc
export MPI_CXX=mpicxx
export MPI_FC=mpif90

export CMAKE_C_COMPILER=${MPI_CC}
export CMAKE_CXX_COMPILER=${MPI_CXX}
export CMAKE_Fortran_COMPILER=${MPI_FC}
export CMAKE_Platform=macos.gnu
export CMAKE_Fortran_COMPILER_ID="GNU"
export LDFLAGS="-L${MPI_ROOT}/lib"
export FFLAGS="-DNO_QUAD_PRECISION -fallow-argument-mismatch "

[gspetro-NOAA]

@natalie-perlin All of this is in the modulefile already, so I did not type it out in the instructions because the user should not have to adjust the code in any way. As mentioned above, it doesn't need to get changed to bash syntax because (if I understood @danielabdi-noaa correctly), the modulefile will work regardless of the user's shell.

[danielabdi-noaa]

TCL modulefiles use setenv instead of export so I think it is fine as it is.