I took the formula from

 http://en.wikipedia.org/wiki/Eigenvalue_algorithm  (retrieved lated August 2014).

Wikipedia claims to have taken it from

 Smith, Oliver K. (April 1961), Eigenvalues of a symmetric 3 × 3 matrix,
 Communications of the ACM 4 (4): 168, doi:10.1145/355578.366316

I did not check stability of efficiency of this, but no problems related to
either are mentioned in the literature.  The main reason I need the closed-form
expression is to be able to pipe it into an algorithmic differentiation system.

Rip out the complete test for eigenValuesNonSym and replace it by the Rosser matrix test from fieldmatrixtest.cc

The old test was complete nonsense: an implementation of some sort of strange
undocumented shape function set that happened to need eigenvalues for something.
Lots of undocumented code, and no checks at all whether the values obtained
by eigenValuesNonSym actually make sense.

Only, for the particular choice of parameters used, eigenValuesNonSym was never
called.  Yes, you read this correctly: these hundreds of lines of code didn't
actually test anything.

The new test is copied form fieldmatrixtest.cc and adapted for DynamicMatrix
types.  It computes eigenvalues of a given 8x8 matrix and compares with the known
results.  Unfortunately, the test matrix is only symmetric.  Tests with more
general matrices are left for another day.

Rather than assuming the caller has set it to the correct size already.
Setting the size in the method eigenValuesNonSym itself makes code
more robust.  Plus, the time needed for that should be negligible
compared to the actual eigenvalue computations.

The AmiraMesh format is so obscure that we should not clutter the code we
generate by references to it.  If people really want AmiraMesh support they
can still add the flags, just like they have to do for many other things.

It is not actually needed, because the code also works in other cases.
A comment next the assertion even explicitly stated that it is not needed!

We decided to have all stl like stuff in Dune::Std:: and not
in Std:: Besides our decision with the previous version tupleutility.hh
did not compile if a header refering to Dune::Std:: was included.

This uses perfect forwarding (following Herb Sutter) and
SFINAE (following Stephan T. Lavavej) to get a standard
comforming implementation also for array.

Once CMake will be default, people can deactivate it.

These classes are deprecated, and some of the tests cause
problems with ctest due to a destructor throwing an exception.
It seems easier to remove the tests directly, as the tested code
is deprecated anyway.

See FS1508. C++11 initializer lists should be used instead.

testdebugallocator_fail{1..5} are a very special sort of tests,
as in order to pass they are expected to fail with a signal like
SIGABRT or SIGSEGV. While autotools, dont distinguish much
about how a test failed, if it is supposed to fail, ctest
handles signals from the test and distinguishes this from
normal termination with nonzero exit code.

This commit introduces a signal handling in the test itself,
that causes the program to terminate "normally" with a nonzero exit
code instead of passing the signal to the fathering process
(ctest). The signal, which is expected to be received can be set
from CMakeLists.txt via a preprocessor variable.

This basically wraps the raw snprintf in order to provide
overflow-save printf functionality while hiding the details
like buffers handling, multiple calls to snprintf, ...

Note that snprintf is a c++11 feature that is, however,
present in gcc-4.4.

Conflicts:
	cmake/modules/CheckCXX11Features.cmake
	dune/common/std/CMakeLists.txt
	dune/common/std/Makefile.am
	m4/Makefile.am
	m4/dune_common.m4

Dominic Kempf authored 10 years ago and Dominic Kempf committed 10 years ago

Autotools users kept complaining, that it is conceptionally
disappointing that targets cannot be build with flags set upon
entering make (e.g. make CXXFLAGS+="-myflag" mytarget). This
patch introduces a small hack, that allows for such feature
in cmake. It is, however, (at least) questionable whether we
want to include it. Still, for your information and convenience:

To use the feature you must:
- be using a CMake generator that is based on Makefiles (defaults to true)
- set the define ALLOW_EXTRA_CXXFLAGS to true, e.g. by including
  CMAKE_FLAGS="-DALLOW_EXTRA_CXXFLAGS=1" in your optsfile.
- reconfigure your project

cmake will then wrap a small shell script around your compiler
and allow you to append the content of an environment variable
to yout compiler flags. Run for above example:
EXTRA_CXXFLAGS="-myflag" make mytarget

The additional flags will be visually confirmed if and only if
you set the verbosity (as always):
EXTRA_CXXFLAGS="-myflag" make VERBOSE=1 mytarget

This very old gcc bug https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10200 is fixed in clang but not in icc or gcc, see FS#1505.

In order to simplify fallback handling between code with and without MPI
support, this patch makes the specialization of CollectiveCommunication
for MPI communicators default-constructible by defaulting the underlying
MPI communicator to MPI_COMM_WORLD, which should be reasonable for most
scenarios.

As the non-specialized version for sequential code is already
default-constructible, this allows grid constructors to accept a
CollectiveCommunication object regardless of whether MPI support is
enabled or not and to provide a default value for this parameter.

This patch resolves FS#1497.

This branch contains required infrastructure for the new iterator ranges
in dune-grid. In particular, it adds a configuration test that checks
whether the compiler supports range-based for loops and a small utility
class that serves as a container for a pair of begin and end iterators.

* feature/FS1042-range-based-for-loops-test-and-iteratorrange:
  [buildsystem] Check for range-based for.
  [Iterators] Add IteratorRange helper class

CollectiveCommunication serves as a fallback implementation, when
code that is written for parallel computations is executed in sequential.
The template parameters BinaryFunctions are no-op then, but they are
still there to define the interface. However, in a generic (parallel)
implementation these will be given. Including binaryfunctions.hh here
avoids error messages due to no-op template parameters whose names
could not be found by the compiler.

C++14 will introduce integer/index sequences in the header <utility>.
This patch introduces a backwards compatible implementation similar in
use. However, the implementation differs in the following technical
details:

  - index_sequence is not a template alias for integer_sequence but
  inherits from it

  - make_{integer, index}_sequence are not a template aliases but free
  standing functions

Both differences stem from the absence of template aliases introduced in
GCC aa late as gcc-4.7. This implementation will work in gcc-4.4 and
later.

Macro to use C++11 feature noexcept specifier.
Add test for CMake and autotools.