[dune-pdelab] adaptivity timedependent problem

[karpinsk at ma.tum.de]

Hello everybody,

I recently tried to extend the adaptivity example of the pdelab-howto
(course-examples/example07) to time-dependent problems.
Unfortunately I had to realize that the extension is not as straight
forward as I expected.

I tried to add the timedependent parts to the code of example07 (Time LOP,
timestepping,..) and after each timestep I called the adaptivity routine. 
I also thought of updating the uold vector (with respect to the new grid).
Unfortunately this does not work out as I'm getting segmentaion faults. 
My guess is that somewhere in the pdesolver routine in onestep-method the
program tries to access an old copy or something (I checked the Line in
which the code is breaking in onestep-method.)

Now, before i dig into the depth of the problem, I wanted to ask, if
anybody has already used adaptive in space FEM methods for instationary
problems in PdeLab respectively whether there is possibility of using
those?

Thank you very much in advance.

Best Regards,
Stefan