- direct works
- ichol has problems for higher orders; maybe this is related to
inherent problems in the testfile?

There is still some error in the higher order version! The solver passes
the tests only if there is no smoothing on intermediate levels!

- Test if one step decreases energy norm
- Remove dead code based on coverage report

- Delete previous example files for MG and PCG
- Add DS Store files to gitignore
- Leave goafem examples for now

The merging branch contains a higher order PCG solver as well as the 1pp
variant of the multigrid solver.

Add new solver structure, lowest order yields same results as existing p1 PCG solver, higher order to be corrected