The fvGeometryVector builds all sub control volumes, sub control
volume faces and fvGeometries; where the fvGeometries are merely
providing iterators over scvs and scvfs of an element.

* multidomain models now also work for cases with pm on top

* functions bBoxMaxIsWall() and bBoxMaxIsWall() in the problem are now used instead of a property
* wall array is now a std::vector

Thanks to Martin for fixing a c&p bug.

interaction -> openPlotWindow

[plotmateriallaws] Outsource plot command, so that all plot options can be defined at a problem-level

The test seems to be very unstable. This has to be investiagted further.
Please follow issue #335

Update of reference vtu files since parameters are defined now on
elements instead of boxes. Furthermore the adaption routine has changed.

The tests have been adapted such that the total mass is calculated before
and after the adaption prodecure. Furthermore, the material parameters
are set element-wise.

* Cause for deviation form original solution cound not be found
* Going back to older commits (release 2.9) still made the test fail
* Maybe some subtle changes in dune?

Updated thresholds for test_stokes, test_2cstokes2p2c, test_2cnistokes2p2cni
and test_2cnistokes2p2cni_boundarylayer to make them pass.

The tests have been adapted such that the total mass is calculated before
and after the adaption prodecure. Furthermore, the material parameters
are set element-wise.

for diffusion instead of concentration gradients.
This change was introduced in revision 0dc99b97
but the reference solution was not updated then.

The Dune range generators can be restricted to a partion. This leads to cleaner
and possibly more efficient code.

Patch has been applied upstream

*Get rid of lowerLeftX, etc

*Get rid of lowerLeftX, etc

* Get rid of lowerLeftX, etc

*Remove unused parameters
*Clean-up

* Reduce number of template arguments
* Rename property from tau to TauTortuosity