Improve handling of mesh stitching for non-conforming meshes

[pmoseley]

This adds a new flag "merge_boundary_nodes_all_or_nothing" (open to better names?) to stitch_meshes and stitch_surfaces, which can help to avoid scenarios where you would otherwise create a mix of merged and unmerged nodes at a stitched interface.

I've also add a boolean return value to help identify if nodes were actually stitched; I think this should help with #3580 .

[moosebuild]

Job Coverage on f8713a2 wanted to post the following:

Coverage

	b06cac	#3599 f8713a
	Total	Total	+/-	New
Rate	61.88%	61.88%	-0.01%	31.58%
Hits	66607	66607	-	6
Misses	41026	41037	+11	13

Diff coverage report

Full coverage report

Warnings

• New new line coverage rate 31.58% is less than the suggested 90.0%

This comment will be updated on new commits.

[roystgnr]

I like this the way it is ... but as long as we're changing the return value anyway, would it make sense to return the number of matching nodes, rather than just a boolean for whether that number is non-zero? map::size() is supposed to be O(1) (and likewise if we switch to unordered_map later).

[pmoseley]

It's a good comment; there's not much reason not to return the number.

I think it will only be useful in some rare cases, because if the user doesn't already know the total number of nodes on the boundary they'll be missing some context. But I'm sure there are some cases where it could be useful.