In particular, getting rid of the German comment is necessary because my venerable ddd
debugger frontend chokes on umlauts.

The BCRSMatrix code clearly shows that you have to call endrowsizes before calling
addindex.  Therefore the third way to build a BCRSMatrix in the 'random' way is wrong
and should be removed from the test.

Use constexpr instead of DUNE_CONSTEXPR



See merge request !18

Feature/smaller cleanups

as the title suggests, these are a couple of smaller fixes, which are not directly related and aren't important enough to manage them in separate branches.

See merge request !13

we introduced a template Parameter for the type counting the current
iteration. In general this is an int, but e.g. the BiCGStab print also
the intermediate steps, thus it uses double as CountType. As Markus
pointed out, the original name Int was not very intuitive.

The information has been provided by the MatrixDimensions struct since roughly the end of 2010.
Correspondingly, the methods rowdim and coldim have been removed from the istl matrix implementations.
Apparently, we forgot to remove them from the Matrix class.

MultiTypeBlockVector: Fix NaN behaviour of infinity_norm()

As a follow-up to flyspray/FS#1147, in response to dd7e8ed3, this makes `MultiTypeBlockVector::infinity_norm()` `NaN`-aware: In analogy to previous patches for other matrix/vector classes, this patch turns `infinity_norm` into two functions; a straight-forward version for `field_type`s that do not have `NaN` values and a less efficient version for ones that do.

It should be noted, however, that because `field_type` is currently hard-coded to be `double` in `MultiTypeBlockVector`, the `NaN`-aware version will, as long as this is so, always be used.

See merge request !15

[bugfix] fix broken assertion.

This patch fixes the errornous commits
e8d16384
1a2f4d0f

The introduced range checks were plain wring, as they neglected the nested structure of the
BlockVector. We now use `b.dim()` to get the flat size.

This fixes flyspray/FS#1724.

See merge request !12

this feature is necessary to keep readable output if we use funny value types, like Vc::double_v

This patch fixes the errornous commits
e8d16384
1a2f4d0f

The introduced range checks were plain wring, as they neglected the nested structure of the
BlockVector. We now use `b.dim()` to get the flat size.

This fixes flyspray/FS#1724.

Test complex<T> for floating point T only

The issue that I originally filed for dune-common as flyspray/FS#1726 affects dune-istl, too.

See merge request !11

The standard leaves (the effect of instantiating) complex<T> for more
general T unspecified.

flyspray/FS#1147: Fix NaN behaviour of infinity_norm() for istl classes

This is an extension of the work related to flyspray/FS#1147 to dune-istl.

See merge request !10

This includes the behaviour of its rows, thus the
CompressedBlockVectorWindow class

[bugfix] Produce a linker error when pardiso lib is not found

If somebody tries to use pardiso despite it not being found,
he should ge t a linker error instead of an exception.
All other bindings handle it that way.

Fixes flyspray/FS#351

See merge request !4

[cleanup] Use template aliases where they beautify code.

I already wanted to use those when implementing the UMFPack
integration into the overlapping schwarz, but our compiler
requirements were not sufficient at the time.

See merge request !9

I already wanted to use those when implementing the UMFPack
integration into the overlapping schwarz, but our compiler
requirements were not sufficient at the time.

Feature/fix deprecation warnings amgtest

please see flyspray/FS#1526

@markus.blatt can you have a look (somehow I can't assign this request to you)

See merge request !7

Apparently, this has been rotting away for quite some time. Adapt the
calls to the Pardiso library. Use the C instead of the Fortran interface.
Use correct method for determining the block size of the matrix.