project_velocity_to_continuous and superparametric issues

When using project_velocity_to_continuous a mesh has to be specified onto which the velocity should be projected. In normal P1DG velocity runs this will be the CoordinateMesh. However when using a higher order coordinate mesh this is not desired as the P1DG velocity would be projected to P2 continuous. In particular it appears there are bugs in momentum_dg when ele_loc(u,ele)/=ele_loc(u_nl,ele) - so the order of the nonlinear velocity has to be the same as that of the velocity that is solved for (this is in general also the desirable configuration for P1DG).

My suggestion therefore:
To avoid likely user errors we should add a check that errors when a mesh of a different degree is chosen under project_velocity_to_continuous. If for some reason people want to fix the mentioned bugs this could change to a warning.