H5fortran Versions Save

Previously, fixed/variable character were only scalar reads. Now, scalar through 7D character arrays can be read for datasets and attributes, fixed/variable. Writing was already enabled.

Add bounds checks and unify with h5fortran-mpi.

Add ability to cast on read from real to int and int to real

also work with CMake 3.11+

• remove little-used %open(verbose=) which is subsumed into %open(debug=)
• default %open(action=) was undocumented but was long default to 'rw'. As a better practice in accord with h5py, default %open(action='r') is now default and documented

UTF8 pass-through for character Attributes and Datasets. We don't use the character(kind=UCS4) because of Intel oneAPI still lacking support for this.

Attribute read/write completely overhauled: overwrite, %delete_attr, %exist_attr for 0d...7d and now can do int64 also

more robust is_hdf5 check for r,r+ mode fletcher32, shuffle optional check for dangling handles at file close add %filesize method add %deflate() method read: NDarray scalar read add %create(..., fill_value) GCC 8.3.0 bug workaround readd %ndims alias to %ndim

• rearrange directories for FPM de facto practice
• add Fortran Package Manager (FPM)
• instead of generated source code, use Fortran standard "include"

Thanks @milancurcic for idea to implement FPM

General cleanup of CMake scripts.

Intel compiler: don't specify xHost flag as some systems shouldn't use them. Let the user decide.

HDF5 1.12.2 is the default version built by the script if the user chooses.

Specify script HDF5 version to build by Git tag for simplicity.