fix PhysiCell constants for more compilers

[rheiland]

This may be premature, but would like to at least suggest it get merged in this release. This addresses compiler errors with PhysiCell_constants{.h,.cpp} for certain compilers (I had mentioned clang++ when I need to do serious debugging, and issue #236 mentioned the error on nvc++, as did @clasqui to @vincent-noel for PhysiBoSS).

I've only tested with the "compile all the sample projects" script. Obviously, it would be good to test with actual simulations (EDIT: see next comment).

[rheiland]

I used the 3 scripts discussed in #247 to verify that the results generated using the modified versions of core/PhysiCell_constants.{h,cpp} in this PR's repo matched (bitwise) the simulation results using the core/PhysiCell_constants.{h,cpp} in 1.13.1

Specifically, I ran this 3rd script (after running its 2 predecessor scripts to generate simulation results):
M1P~/git/new-build-test$ python beta/test_diffs_svg.py ~/git/new-build-test/samples_test ~/git/fix-constants/samples_test
and it confirmed that the (early) .svg results were bitwise reproducible for these sample models:

# note that we omit 'output_physimess' only because, currently, it fails the recursive copy of
# additional files needed in config/subdirs (until the Makefiles and "make load PROJ" is updated)

for out_dir in ['output_template','output_biorobots','output_cancer_biorobots','output_celltypes3',
'output_heterog','output_interaction','output_mechano','output_pred_prey','output_virus_mac',
'output_worm','output_rules','output_cancer_immune']:

[rheiland]

I also confirmed that nvc++ compiles without any errors (on Ubuntu). It's actually enlightening to see the warnings that it returns which hint at some code cleanup that could be done.

[vincent-noel]

Works on my ubuntu too