add LLF fallback solver

[BenWibking]

Uses the Local-Lax Friedrichs (Rusanov) solver as a "fallback" Riemann solver.

Closes #379.

[BenWibking]

/azp run

[BenWibking]

/azp run

[markkrumholz]

The change to how Riemann solvers are referenced is going to cause conflicts with the MHD branch, since the same changes will need to applied to the HLLD solver in that branch. I'm not sure where @AstroKriel's MHD branch is with respect to the repo, but we might want to have a look just to make sure that there won't be any pain points when we need to integrate these changes with the MHD HLLD solver.

[BenWibking]

The change to how Riemann solvers are referenced is going to cause conflicts with the MHD branch, since the same changes will need to applied to the HLLD solver in that branch. I'm not sure where @AstroKriel's MHD branch is with respect to the repo, but we might want to have a look just to make sure that there won't be any pain points when we need to integrate these changes with the MHD HLLD solver.

The trivial solution is to just add another conditional case to the if constexpr statement in HydroSystem::ComputeFluxes. I don't think it should be more involved than that.

[markkrumholz]

The change to how Riemann solvers are referenced is going to cause conflicts with the MHD branch, since the same changes will need to applied to the HLLD solver in that branch. I'm not sure where @AstroKriel's MHD branch is with respect to the repo, but we might want to have a look just to make sure that there won't be any pain points when we need to integrate these changes with the MHD HLLD solver.

The trivial solution is to just add another conditional case to the if constexpr statement in HydroSystem::ComputeFluxes. I don't think it should be more involved than that.

OK. I guess that should be done in a separate PR, which will have to happen before @AstroKriel's code can be merged into the main branch.