The majority of developers favor individual headers for (emulation of) C++11
features over collecting them in a single header cxx11.hh. This patch
renames cxx11.hh to constexpr.hh and moves it into the subdirectory std.
This is the place most developers would like these headers in (see
http://users.dune-project.org/doodles/4).

As decided on the developer meeting in Aachen, we support (but do not
rely on) the C++11 keyword constexpr. This patch adds an m4-check for
constexpr and a header (cxx11.hh) defining a macro DUNE_CONSTEXPR either
as constexpr or empty (depending on the compiler support for it). It is
then used in FieldVector and FieldMatrix on size, rows, and cols.

Previously, a variable (found) was only initialized in one branch
of an if statement but used later on in any case. This patch makes
sure that is initialized in both branches.

Steffen Müthing authored 11 years ago and Christoph Grüninger committed 11 years ago

The current algorithm for hash combining behaves very poorly when hashing
multiple short streams of small integers like (0,1,1,0,x) that only differ
by a single large number x. These objects appear frequently in PDELab's ordering
framework, and the resulting bad hash distribution really hurts the performance
of the unordered containers in PDELab.

This patch switches to an algorithm that is inspired by CityHash (see the source
code comment) and has proven to be much more robust in testing.

Because C++11 implements this differently, and we now have an
stl-compliant implementation to switch to.

I would deprecate isPointer and isReference too, but I can't find
a way to do it.

These are replacements for TypeTraits::isPointer, TypeTraits::isReference,
TypeTraits::PointeeType.  They reimplement what is available in the
C++11 standard library, and are expected to supersede the TypeTraits
class eventually.

Note: this patch does not pull in the corresponding stl implementation.

That's easier to maintain.

If not found, our own fallback implementation will be used.

A previous commit introduced a check whether the element to-be-touched
is actually contained in the container.  That patch caused the test
to fail.  The new patch fixes this.

The two classes PushBackTuple and PushFrontTuple are both implemented
using a long list of template specializations.  This makes for a lot
of code, and it also limits the maximum tuple size that can be used.
Using variadic templates (FYI: available since gcc-4.3) instead
allows a generic implementation in only a few lines of code, and
only the compiler being the upper limit on tuple size.

An implementation using variadic templates has already been given
in the documentation of PushBackTuple/PushFrontTuple.  This patch
starts to actually use that implementation, provided that
HAVE_VARIADIC_TEMPLATES is set.  Otherwise the old implementation
is used.

Because the latter uses the C++11 standard library naming.

The C++11 standard library offers the functionality of SelectType,
it is, however, called 'conditional'.  This patch adds a copy
of our SelectType implementation with the class names changed
to match the ones used in C++11.  The old SelectType implementation
gets deprecated.

The default constructor of the FieldVector has intentionally avoided calling a
constructor for the contained data to eliminate the potential overhead of that
initialization.
During implementation of a new creation mode of the BCRSMatrix, it was discovered
that this behavior is problematic due to its inconsistency: If the data type
contained in the FieldVector is a non-trivial type, a default-constructed
FieldVector will be fully initialized. On the other hand, with a primitive data
type it will be in an uninitialized state, potentially causing undefined behavior.

As this behavior might be very unintuitive for users and made implementing the new,
more efficient creation method for the BCRS matrix very difficult to implement, this
patch changes the FieldVector default constructor to always default-construct its
contents regardless of its type. While this change does introduce a possible
performance regression (for primitive types, the compiler now has to zero out the
memory), the overhead has been measured to be minimal.

The actual implementation prefers to use the new C++11 unified initialization
syntax and falls back to std::fill for older compilers. The distinction was made
because newer compilers generate much more efficient code when using unified
initialization, and for user-defined types, this approach avoids first calling the
default constructors and then the copy assignment operator.

- update documentation
- add paranoia check to touch()

update documentation