read: add input dataset validation and buffering

with GCC-10 and -O3 on Windows, under certain conditions it was observed that reading datasets with implied type coercion would fail by giving random data, changing run to run of the same program/data. This was observed on scalars only.
To solve this issue and add general robustness to reading, we now get the dataset type from disk and compare it with the variable datatype, raising an error if the variable type and size do not match the dataset type and size.

add CMake 3.13 CI job

v3.5.1

add cmake 3.13 test ci

add COMPACT dataset option. improve packaging and ABI check

• COMPACT dataset is an option with h%write(..., compact=.true.) This allows up to several times faaster read/write for < 64 kByte variables, which can be noticable if your program uses many read/write of small variables.
• further improve packaging and ABI check.
• remove Intel rank() workaround by just specifying scalar
• bugfix; GLOBAL imported target h5fortran::h5fortran

install/package enhancements

CI auto uploads binaries too

improve FindHDF5 / autobuild

FindHDF5 doesn't search for Zlib till all components found. This helps avoid conflicts with ZLIB::ZLIB with the self-built ZLIB.

less chatty library dialog

v3.4.8.1

less chatty libs

CMake >= 3.13 compatible

Although we recommend CMake >= 3.14 in general, to make easier access, we allow CMake >= 3.13

examples use tagged version

We switched default branch to main and made examples used tagged version for best practices.

Improve Zlib and HDF5 autobuild

Use CMake native features better to make the Zlib and HDF5 autobuilds more rock solid.

Improve compiler detections and compatibility checks

bugfixes

• Intel oneAPI: correctly detect zlib name
• szip: make optional even when requirement because it may be statically linked