More robust hdf5, along with more refined I/O from cmake

[mauneyc-LANL]

PR Summary

Addressing issues @jdolence encountered with HDF5 resolution. Some small rearranging of dependency logic, and a less chatty cmake (by default, can be adjusted)

I had attempted to use module mode to resolve find_package(HDF5) with a distributed FindHDF5.cmake, but I think I made a mess of it and Josh hit an issue that arose from that. The large change here is to revert to a more "standard" approach that relies on the implementation of "official" Kitware CMake module (https://gitlab.kitware.com/cmake/cmake/-/blob/master/Modules/FindHDF5.cmake).

For reference, it can be informative to expose some trace information on find_package when configuring

cmake --debug-find-pkg=HDF5 .. -DSINGULARITY_USE_HDF5=ON #...

Other notable changes:

• Output has been reduced in the default log level. A more chatty output can be restored using --log-level=VERBOSE option: cmake --log-level=VERBOSE .. #...
• some cleanup of singularity_import_* macros

@jonahm-LANL @dholladay00 @jhp-lanl

PR Checklist

• Adds a test for any bugs fixed. Adds tests for new features.
• Format your changes by using the make format command after configuring with cmake.
• Document any new features, update documentation for changes made.
• Make sure the copyright notice on any files you modified is up to date.
• After creating a pull request, note it in the CHANGELOG.md file

[Yurlungur]

Don't forget to update changelog

[Yurlungur]

Looks good to me. @jdolence can you double check this resolves the issues you were seeing?

[dholladay00]

@jdolence what is your timeline on testing this? I'd like to get this in soon.

[dholladay00]

@mauneyc-LANL can you resolve all incomplete items in the checklist? When completed, I'll merge.

I suspect downstream testing will be necessary.

[Yurlungur]

Yeah I think @jdolence is not going to test this. Assuming tests pass I'm fine with merging without waiting for him.

[junghans]

@mauneyc-LANL We can drop this for cmake>=3.26, see: https://gitlab.kitware.com/cmake/cmake/-/merge_requests/8015

[mauneyc-LANL]

yes; Indeed, we don't require to do this for spiner, and I'm working on a PR for singularity to follow that approach.