feat: user error for hetpar start and end core

[pgierz]

No description provided.

[mandresm]

Hey @pgierz! I might be wrong, but I think start_core and end_core are parameters defined on the backend that depend on the number of nprocs of all components. The problem is that we need to catch missing nproc for components that need an nproc so that start_core and end_core values are computed.

[pgierz]

whoops, yeah...I'll take care of that, sorry.

[mandresm]

yes, think about oasis or recom for fesom2: they are components for ESM-Tools, but they are libraries, meaning they don't need cores themselves.

[pgierz]

Alright. Then I think we should write something to that effect:

Suggested change

	# FIXME(PG): ...what? Just continue???
	# NOTE: For components like oasis or recom, no "own" cores are needed

[mandresm]

I'm almost certain that removing this will kill the het_par_wrappers functionality.

[pgierz]

One would think so. However, the start_core and end_core variables are not used anywhere in that function. Note that other functions further down might need those keys, but they aren't needed here specifically. To avoid getting errors in the wrong place, I removed them.