By using the same mechanism as in the shell script:  piping everything through
the 'tee' program.  This may not be the most elegant way from a Python standpoint,
but it seems to work.

Since now my Python testsuite seems to do everything the shell version did,
this commit also removes the shell version.

Having it in bash was starting to be a pain.  This particular testsuite uses
a bit of arithmetic, which was difficult to do in bash, but trivial in Python.

Such measurements are now centralized in the program compute-disc-error.cc.

This is planned to be used to start a fine-grid computation on the coarse-grid
solution, or to start a high-order computation from a low-order one.  This is
helpful, in particular, for measurements of the discretization error, which
take a lot of time.

It still tested for the dune-fufem wrapper of a first-order basis, while nowadays
the RiemannianTRSolver class is instantiated with dune-functions bases.

It is still no function in the dune-functions sense, but at least the
basis parameter needs to be a dune-functions basis now, rather than
a deprecated dune-fufem basis.

This is needed to be able to multiply FieldMatrices with Quaternions.

For those, it is necessary to have separate measurement for the displacement and orientation components.

Also, check whether the input files have the correct length

This creates a field with topological charge 2.

This is a preparation to allow to select the TargetSpace at run-time.

Needed for gradient-flow applications.

You never know, we may be doing dim!=dimworld grids eventually.

Of course we always need unitvector.hh.  vtkadapter.hh, on the other hand, only
exists in dune-functions 2.4.

According to Hanne, this is needed to turn 1d harmonic map heat flow into true
curve-shortening flow.

Hardwiring the weighting here is very hacky.  Something more flexible and
general will need to be implemented eventually.