Create a CI job that assures that our Euler operator exactly matches the book codes

[inducer]

Using inducer/meshmode#55

I.e. make sure that (at least for some doucmented set of settings) our Euler RHS matches the output of

[rhsQ] = EulerRHS2D(Q,time, ref_solution);

Note that ref_solution should not just be the exact solution, since that will not have jumps and will thus not meaningfully exercise the numerical fluxes.