Drop OCCA backends

[jeremylt]

Fixes #816 Fixes #678

This backend compiles but does not run, and there are several known bugs in the implementation. Since this backend has not had a maintainer in 533 days, we don't have an avenue to fix it. The design of this backend is radically different that the rest of the backends, so our current maintainers would need extensive time to first learn OCCA and the design of the backend before starting to fix all of the bugs in this backend.

At this time, I am aware of no current users for this backend (makes sense, as this backend doesn't even pass the test suite anymore).

I think we should drop this backend. If we get support from an OCCA developer in the future who wants to assume responsibility for maintaining this backend, they can revert this commit if they wish to attempt repairing this version of the backend.

[jeremylt]

Updated for changes in main.

Since the OCCA backend does not function and we do not have anyone who is able to fix it, the question really is - why would we not remove this broken and unmaintained code?

[jeremylt]

@jedbrown @tzanio, since we have the CEED6AM meeting this week, I think we have a chance to finally make a decision here. I think that if we can't find someone who can do the massive fixes + ongoing basic maintance that the OCCA backend requires, I think we should go ahead with this PR and remove the OCCA backend. We can always revive it from Git history if someone wants to revive the backend in the future.