…e right.

Also some fixes to two tests that failed the updated logic because they were themselves inconsistent.

…schur_tests

…xistence) preconditioner_matrix.

Also allow correct velocity to take in a option path as it also has assumed locations for velocity solver options.

Add a test that appears to pass that uses Schur complement solenoidal projection.  Need to look at this more.

…e necessary but unrelated to other changes in this branch.

…ions.

diag and lump appear to be best with tolerances increased for no and mass due to poor convergence

…c numbering so that we get a second order halo when it's required.

…it appropriately.

for more information, see https://pre-commit.ci