I modified setup.py to look for the netcdf4 directory in the environment variable $NETCDF_HOME. If the environment variable does not exist, it will default back to the original hardcoded paths.
I did this for both the "Caltech" and "Linux" config sections, because when I configure pycles on my linux machines, they always use the "Caltech" config section because the check for that machine is pretty general ('x86_64' and 'LD_LIBRARY_PATH' are pretty general, and would exist on most compute clusters).
I modified
setup.pyto look for the netcdf4 directory in the environment variable$NETCDF_HOME. If the environment variable does not exist, it will default back to the original hardcoded paths.I did this for both the "Caltech" and "Linux" config sections, because when I configure pycles on my linux machines, they always use the "Caltech" config section because the check for that machine is pretty general ('x86_64' and 'LD_LIBRARY_PATH' are pretty general, and would exist on most compute clusters).