Describe the proposal
We should add a CI test run that compiles and runs the 1D hydro test problems with the HLLD solver (with B=0). This will ensure that the MHD solver is at least as well tested as the hydro solver, and that we do not regress on any of the existing hydro test problems.

Describe alternatives you've considered
The alternative would be to create MHD versions of all of the existing hydro test problems, but that would be laborious and unnecessary if we could were able to recompile and run the hydro problems with the HLLD solver.

Additional context
There are currently subtle differences between the HLLC and HLLD solvers involving the wavespeeds as well as the carbuncle fix. For the latter, see #528.