Test for out-of-mesh-value on bad MeshFunction variable indices

[roystgnr]

The initial suggestion ChatGPT gave me here was broken, but I added the (simple) fix, as well as an analogous test for the second derivative API.

[jwpeterson]

!!!FAILURES!!!
Test Results:
Run:  4985   Failures: 1   Errors: 0


1) test: MeshFunctionTest::test_bad_hessian_var_with_out_of_mesh_value (F) line: 271 ../../tests/mesh/mesh_function.C
double equality assertion failed
- Expected: 150
- Actual  : 150.000000000002
- Delta   : 1e-12

[roystgnr]

Of course ChatGPT wouldn't think about compatibility with parallel partitioning; it's more embarrassing that I didn't either.

[roystgnr]

Well, I was hoping I could fix things blind, but no such luck. I'll try to have a proper parallel fix tonight when I can plug in my dying Macbook.

[roystgnr]

I set auto-merge since this definitely adds some feature coverage, but I definitely need to remember to look at the line coverage changes after CI finishes, to see just how bad the ratio there is...

[moosebuild]

Job Coverage, step Generate coverage on 5244b48 wanted to post the following:

Coverage

	e2aad5	#4391 5244b4
	Total	Total	+/-	New
Rate	65.34%	65.36%	+0.02%	-
Hits	77753	77772	+19	0
Misses	41246	41227	-19	0

Diff coverage report

Full coverage report

This comment will be updated on new commits.

[roystgnr]

161 lines of tests for 19 lines of (new) coverage. 😆