[dune-pdelab] PDELab Stokes

[Christian Engwer]

Dear all,

during the pdelab review we decided to seperate the interfaces of
constraints and boundary-conditions. There are still some issues with
the transition to the new interface. I started fixing parts of the
issues by unifying the parameter interface for the
stokes/navier-stokes discretizations. Currently I only updated the
classes to use the same interface. We might still want to clean up the
parameter interface.

I discussed with Steffen and we think we ought to fixes this kind of
problems better sooner than later, as any kind of multi-domain code
may run into hard-to-track down problems, unless we finish the
seperation of constraints and boundary-conditions. I'm not going into
further details, it would be rather lengthly... but I wanted to give a
short note about why I commited the previous set of patches.

Cheers
Christian