Modernize fortran

[tacaswell]

While looking at the flang issues on windows I tried to things to compile with std=f2018 which did not go super well. I mechanically fixed some issues, but am now stuck on a combination missing intrinsic (functions around complex numbers) and go to to other blocks.

Opening this up as a draft with out the expectation that it will be merged, but so that it is not only on my computer.

[briantoby]

My experience with moving between fortran compilers with GSAS, was that every time I did (for example when moving from the VAX to Fortran compiler to the SGI), one would find dodgy bits of code and it took a lot of work to create code that would run the same from both the original and new compiler. It took extensive testing. Fast forward by a decade or two and GSAS standardized on g77. FWIW, if compiled w/gfortran, the code runs but produces wrong results. That has never been fixed (which is why GSAS now only runs on Win32).

For the Fortran codes used in GSAS-II, we have used gfortran, since it is available everywhere. I would be very reluctant to trust that the code compiled with any other compiler is producing the same results, with the one exception of pyspg, where I do have a pretty comprehensive test suite.

I'm not against some code cleanup -- I'd really love to get rid of the reams of compilation warnings, but in the end, please let's use gfortran on Windows.