GitLab

* 'subDiff' is now called 'subDifferential'
* It returns its result in the return value, not in its second argument
* The method 'domain' also returns its result in the return value

And get rid of the file newsumfunctional.hh, which was never intended to
be a permanent solution.  With this patch, the file sumfunctional.hh
contains both the new and the old implementation, and you can switch
between the two with USE_OLD_TNNMG.

In the new tnnmg implementation, the calling method is expected to know
whether a functional is quadratic or not.  In case of a quadratic
functional, the caller will not call bisection anyway, but it will
rather call a dedicated solver for quadratic problems.

This patch also introduces the preprocessor variable USE_OLD_TNNMG.
When this variable is defined, the old tnnmg implementation is
reenabled. This will simplify running old and new codes side by side
for a while, which I find helpful for debugging.

Previously, if the subdifferential at the initial iterate would contain zero,
no message would be written at all.  This made it more difficult to follow
the programm output, because it was impossible to see how many times
the bisection method was called.

It is not actually used.

First, the functional may not know the type.  Secondly, the type is not used anyway.

This is not true anymore: support for more general problems has been added.

Remove origin_ and direction_ members and accessors method
because the are not used anywhere. Also add const to arguments.

Since the last change to the functional interface the storage
type is determined via the 'is_reference' property of the passed
type.

If Matrix/Vector are passed as reference types, then they are
stored by reference. Otherwise they are stored by value.
To get the old behaviour you will now have to add '&'
to the passed types in your user code.

We need this to adjust the stored pointer

This is necessary to also accomodate MultiTypeBlockVectors.

This is WIP and not tested yet.

Expressions in the false branch that cannot be instantiated
must be guared by the (template) identity. Suprisingly gcc
accepted this code (clang did not).

Make indentation consistent with editor hints

Now BoxConstrainedQuadraticFunctionalConstrainedLinearization
can be used with nested problems with block size >1.