This enables users to have several build directories like
build-clang, build-gcc5, build-debug, and still get all
of them ignored.

Nothing needs to be included with proper C++11 compilers.

Old feature support macros which were tested until 2.4, kept around
for one more release. As these are now always supported due to the
new compiler requirements, they are directly defined without an
explicit test.

Conflicts:
	cmake/modules/CMakeLists.txt
	cmake/scripts/CMakeLists.txt

It is outdated.

We aim to document the entire relevant user API. Some functions
from CMake are not worth wrapping, so we need to add them to our
documentation.

Broken/missing references, wrong directives etc.

As documented, tranformTuple should support AddRefTypeEvaluator. As
std::make_tuple will remove references (using std::decay) from the
funtor's return type, the returned tuple would contain copies of the
values. This is not the intended behavior.

This patch adds AddRefTypeEvaluator to the utilitytest and fixes the
return type of transformTuple.

These commits broke compiling / linking if one package was found
but the includes / libraries were not added. We need a different
approach to hide the ENABLE_* trick from the user and keep full
testing capabilities.
This reverts commits 0f12caff
and d75c3a46.

We do not want one such documentation per module, but instead
we would allow people to use it from their module and get a
documentation generated from all the dependent modules. This
needs some changes to the machine.

This includes the following workflow:

* The cmake module path is globbed for any modules matching *.cmake
* The opening comment is parsed for the following rst directives:
  cmake_function, cmake_module, cmake_variables
* From those, a browsable documentation is generated.

To run properly, a python interpreter and sphinx are needed.

If you rely on it, add the deleted CMake files to your module.

This reverts commit 4441ec26.
The minimal required is finally GCC 4.7 or newer (it is actuall 4.9).
Original commit message for d26e5ed3 was
    [C++11] Follow the C++11 specification of operator new.

    The signature changed in C++11 which we require now.
    To get more details see the mailinglost in October 2012.

Thanks to Marian for the heads-up and the testing.

Use ALLOW_CXXFLAGS_OVERWRITE instead.

Drop support for std=c++0x.

They were tested until 2.3 and are now always supported due to
the new compiler requirements.
Removed macros: HAVE_INTEGRAL_CONSTANT, HAVE_DUNE_HASH,
HAVE_STD_HASH, HAVE_TYPE_TRAITS, HAVE_VARIADIC_TEMPLATES,
HAVE_VARIADIC_CONSTRUCTOR_SFINAE, HAVE_RVALUE_REFERENCES

Otherwise FieldVector<K,1> is not literal and its size()
method can never be used as constexpr.

This makes the code more robust and (a bit) more robust.  Additionally,
it repairs the recent breakage that I introduced by modifying promotiontraits.hh.

Otherwise FieldVector is not a literal type and
calling its size() method is not a const-expression
although size() itself is declared as constexpr.

The previous implementation consisted of a list of promotion rules for
various types.  The new implementation is much shorter and more general:
it will also work for types not on the old list.  The promoted type is
now defined as the result type of the addition of the two input types.
There is no special reason to use addition here, but it does reproduce
the precise behavior of the old implementation.

The 'macroname' variable is still used by the code, but its definition
was accidentally removed together with the AutoTools-generating code.
Therefore, duneproject would fail with

/home/sander/dune/dune-common/bin/duneproject: Zeile 695: dune-foo/cmake/modules/: Ist ein Verzeichnis

This patch puts the missing definition back in.

Document an incompatibility issue of index_sequence
and add a reminder to fix this as soon as we allow
template aliases.
This can only be fixed