Just a few things I found on a stroll through the code.
The docs entry for MaximumGravityRefinementLevel refers to behavior that is currently inside an #ifdef UNUSED and wasn't correct anyway.
LGTM. I just wanted to make a note to others that the test failure is coming from the lack of old answers to the new MHD metal advection test.
You're right, John. I will update the tag for the test results and either update this PR or decline and resubmit, depending on what's easier.
Looks good. I'm pretty sure RefineByJeansLengthUnits was to get that refinement criterion working properly with EquationOfState=1 which doesn't have an energy term in BaryonField. I think it still doesn't do it right, but that parameter can go.