Factorize repeated code to select Stokes degrees of freedom.

[bangerth]

We have at least 40 places where we have the following code:

      // First loop over all dofs and find those that are in the Stokes system
      // save the component (pressure and dim velocities) each belongs to.
      for (unsigned int i = 0, i_stokes = 0; i_stokes < stokes_dofs_per_cell; /*increment at end of loop*/)
        {
          if (introspection.is_stokes_component(fe.system_to_component_index(i).first))
            {
              scratch.dof_component_indices[i_stokes] = fe.system_to_component_index(i).first;
              ++i_stokes;
            }
          ++i;
        }

This could probably be better handled by a helper function. In fact, this is something that might even be calculated once and for all (in Introspection?) so that we don't have to do it in all of these places.

[gassmoeller]

Yes I think this makes a lot of sense and would clean up the code considerably. I think there is similar code for the advection assembly that is maybe not duplicated quite as often, but is worth checking as well.

[bangerth]

I'll add that we run this code in many functions that are called once per cell. So the code is run often. Pre-computing this information would be useful.