CHOLMOD supports this, it just wasn't exposed in the interface.
It is needed for large problems.

The CMAKE_GUARD for multiple guarding variables requires a
porper boolean statement.

And fix all spelling errors that codespell currently finds.

Unused variable warning indicated this oversight

See https://reuse.software/ for a description.

add test for block size test for spqr

deactivate std::complex tests of spqr

Both behave like std::tuple, and currently even inherit from std::tuple.
According to

  https://en.cppreference.com/w/cpp/utility/tuple_element

we are allowed to specify std::tuple_element for our own types.

The note

/dune-istl/dune/istl/paamg/smoother.hh:130:26: note: the ABI of
passing structure with complex float member has changed in GCC 4.4

for DefaultConstructionArgs::getArgs got me thinking. I just could not
imagine why there should be comlex value involved. Turned out we
somehow assumed that for complex the relaxation factore would be
complex, too. IMHO a relaxation factor should scale and hence be real.
This MR make ssure that the relaxation factor is never complex.

Unforntunately, then a note for std::make_pair with a
complex in repartition.hh appeared:
/usr/include/c++/8/bits/stl_pair.h:524:5: note: the ABI of passing
structure with complex float member has changed in GCC 4.4
     make_pair(_T1&& __x, _T2&& __y)

and I just resorted to constructing Pair directly. Not sure whether
this is really needed but it should not hurt and ensures more silence.

This is achieved by using the flat matrix and vector forEach implementations.
The unit test is extended by a multi type example.

The routines traverse the container recursively and call a functor
at each scalar entry passing also the flat index.

This makes the naming consistent again.

The loaction of the macro guard lead to the method `test`
not returning a value if neiter SUperLU nore SuiteSparse
are found. While the test is not executed in this case due
to a CMake guard, these macro guards should be placed
in the correct place anyways for reasons of code hygiene.

This test should not be necessary anymore, since commit
05f80ffa fixed the CMAKE_GUARD argument
for this test, so that it is not compiled anymore if either UMFPACK or
SuperLU are missing.

That line had to be removed as it did not compile if either UMFPACK or
SuperLU were missing. For example while I had UMFPACK installed but not
SuperLU, gcc gave me the following error message for the removed line:

…/dune/istl/test/overlappingschwarztest.cc:260:17: error: ‘HAVE_SUPERLU’ was not declared in this scope

When MPI is not available or explicitly disabled with the CMake build
option -DCMAKE_IDSABLE_FIND_PACKAGE_MPI=TRUE, some tests were unable to
build.

The tests created from solverfactorytest.cc.in and part of
scalarproductstest.cc use Dune::OwnerOverlapCopyCommunication which is
defined behind `#if HAVE_MPI` and is thus not available in a non-MPI
build. I've thus disabled those tests when MPI is unavailable.

The matrixmarkettest did not work without MPI, as it contained some code
using the wrong template parameters when HAVE_MPI was not set. Those
template paramters have been fixed now.

I've confirmed, that after my changes `make build_tests` succeeds to
build all tests and that those tests run without failure.