Removed duplicate code in DG conv. tests, made 3D test a little less demanding by lowering max level.

* fixes in macro-boundary detection
* 3D VolumeDoF PackInfo (direct only)
* DG 3D now also seems to converge with multiple macros

DG Poisson works on single macro-cell with hom. Dirichlet BCs. Penalty parameters can now be set from HyTeG.

Fixing bug in macro-boundary detection for DG in 3D. Diffusion system matrix is now symmetric. Still WIP - penalty parameters likely still off - to be fixed in HFG.

Fixes bad_cast issue in P2P1ElementwiseBlendingStokesOperator and adds PrintInfoTest

See merge request !500

The test checks that we get neither compile- nor run-time errors, when
trying to print meta information on the build using the functions from
Git.hpp and BuildInfo.hpp.

The sub-operators of P2ElementwiseBlendingVectorLaplaceOperator are of type
P2ElementwiseBlendingLaplaceOperator and not P2ElementwiseLaplaceOperator.

Maybe it would be wiser to cast to
P2ElementwiseBlendingVectorLaplaceOperator::scalarOpType

Updates BuildInfo

See merge request !498

- Commit adds a new function compilerInfo() that will return the
  compiler ID and version as obtained from CMake.
- printBuildInfo() will now include this information.
- We rename buildFlavour() to buildType() for naming consistency with cmake.

In some cases we need all neighboring volumes that share the same macro-vertices (e.g. MMOC) and in others we only need those that are interfaced through facets (e.g. DG communication).
Now we are storing both.

WIP: Petsc compare graph

See merge request !396

VectorToVectorOperators, Smoothability and Code Cleanup

See merge request !497