Fix container builds, merge/rename spack-stack extension repos, add Intel oneAPI 2024.0.2 compiler/MPI on AWS ParallelCluster

[climbfuji]

Summary

Several changes to fix the broken container builds and reorganize our repos:

• Update spack submodule pointer to include pflogger updates (to fix broken clang/mpich container builds of pflogger)
• Merge spack-stack extension repos in spack-ext/repos: combine jcsda-emc and jcsda-emc-bundles into spack-stack (we also have EPIC and may have more partners in the future, so this makes a lot of sense)
• Add logic to copy the spack-stack extension repos into the container: this requires recording the spack-stack hash as part of running setup.sh (so that docker can recognize updates and doesn't use the cache if spack-stack has changed) and physically copying the repos directory into the container environment directory (because the docker build runs from that directory and everything above is not in its context)

Unrelated, and commented out by default:

• Addition of Intel oneAPI compiler/MPI for aws-pcluster site for testing

Testing

• CI container builds this can only be run after the spack PR is merged and the submodule ptr is updated in this PR
  • jedi-clang-mpich: https://github.com/JCSDA/spack-stack/actions/runs/7663766335/job/20887072436
  • jedi-gcc-openmpi: https://github.com/JCSDA/spack-stack/actions/runs/7668917120/job/20901724670
  • jedi-intel-impi - will be run later
• General CI builds
• CI build on AWS R&D ParallelCluster: https://github.com/JCSDA/spack-stack/actions/runs/7663464277/job/20886322600

Applications affected

n/a

Systems affected

spack-stack containers

Dependencies

• waiting on Update pflogger from spack/develop (add [email protected]) spack#397

Issue(s) addressed

Resolves #965
Resolves #960

Checklist

• This PR addresses one issue/problem/enhancement, or has a very good reason for not doing so.
• These changes have been tested on the affected systems and applications.
• All dependency PRs/issues have been resolved and this PR can be merged.

[climbfuji]

We haven't used this and I don't think we need it?

[AlexanderRichert-NOAA]

I guess I can go either way. I'm okay with removing it for now, but if we run into issues again like we had with the ESMF recipe where it was using the calling environment (incl. PATH) rather than the Spack build environment in places, then we should revisit and try to make it work on all systems (I don't remember how far I got with testing, I think it worked on more than just Acorn and Orion but those were the ones that needed added customization).

[climbfuji]

Ok, thanks. I'll take it out for now. I never had to do anything particular about the environment except for Cray systems where you need to have a separate set of modules loaded, depending on the compiler you build for (see Narwhal instructions). Since we've agreed today to move forward with separate environments for different compilers, that will be even less of an issue (again, essentially doing what we already do for Narwhal).

[AlexanderRichert-NOAA]

Looks good (apart from .gitmodules)