Improvements to affinity.c

[ngs333]

This PR improves affinity.c by:

1. replacing affinity.c with one copied from FMS/affinity;
2. modifies configure.ac to check for sched_getaffinity;
3. modifies the new affinity.c to #include configure.h.

This has been tested with gcc on analysis (an008) with and without the --with-mpi compilation flag using the canonical compilation method. It has also bee tested on an ubuntu laptop.

I think it would be prudent to build and test on Apple OSX.

FYI, The new affinity.c #includes the generated configure.h file. Somehow FMS is able to compile affinity.c without this change - to my surprise.

[ngs333]

@ceblanton @underwoo
In this PR I was merely following Seths suggestion (in #17 (comment) ). Now the differences between the NCTools affinity.c of this PR and the FMS affinity.c is that a) #include "config.h" was also added to the NCTools one as its needed to compile; and b) some comments that were added for the FMS doxygen system were removed.

[ngs333]

@ilaflott , @ceblanton
Ian and Chris, this is the PR which I mentioned could use compiling and running on a Mac. Thanks.

[ilaflott]

Build on Macbook Air, 202 13inch, macOS 13.3.1 (ventura), intel chip failed. Though to be fair, it seems to fail buildingNOAA-GFDL's master branch with an identical complaint. Of course, I attempted to rebuild/configure etc. following the suggestion in the error message, but to no avail.

checking netCDF Fortran 90 binding available... no
configure: WARNING: Unable to locate a working Fortran netCDF library.  Please specify --with-netcdf-fortran=<LOCATION> where <LOCATION> is the full path to where netCDF is installed.
configure: error: Unable to find NetCDF Fortran library.
make: *** No targets specified and no makefile found.  Stop.

Of course, I made sure i had this library:

[build $] nf-config

This  4.6.0-development has been built with the following features:

  --cc        -> clang
  --cflags    -> -I/usr/local/Cellar/netcdf-fortran/4.6.0_1/include  -g -Wall -Wno-unused-variable -Wno-unused-parameter -O2

  --fc        -> /usr/local/bin/gfortran
  --fflags    -> -I/usr/local/Cellar/netcdf-fortran/4.6.0_1/include -I/usr/local/Cellar/netcdf-fortran/4.6.0_1/include
  --flibs     -> -L/usr/local/Cellar/netcdf-fortran/4.6.0_1/lib -lnetcdff -lnetcdf -lnetcdf
  --has-f90   -> TRUE
  --has-f03   -> yes

  --has-nc2   -> yes
  --has-nc4   -> yes

  --prefix    -> /usr/local/Cellar/netcdf-fortran/4.6.0_1
  --includedir-> /usr/local/Cellar/netcdf-fortran/4.6.0_1/include
  --version   ->  4.6.0-development

Though, for what it's worth, I don't think this means the branch is necessarily at fault. Maybe my env is bunked up because i basically lean on homebrew to manage most of it. But maybe this is enough to say it doesn't pass the "scientist can easily use it with a mac" test.