[[Imported from SVN: r5746]]

credits for finding this issue go to the Freiburg nightly builds

[[Imported from SVN: r5745]]

[[Imported from SVN: r5744]]

sync with my version.

[[Imported from SVN: r5743]]

[[Imported from SVN: r5742]]

[[Imported from SVN: r5741]]

[[Imported from SVN: r5740]]

[[Imported from SVN: r5739]]

[[Imported from SVN: r5738]]

[[Imported from SVN: r5737]]

[[Imported from SVN: r5736]]

Substituted exit with retun in main functions. 

This prevents clever people, like myself, from complaining about memory-leak after
perfectly OK changes by fellow developers. ;)

[[Imported from SVN: r5735]]

A proper cast-to-bool is still missing.

[[Imported from SVN: r5734]]

[[Imported from SVN: r5733]]

Adhere to shared_ptr: empty constructor constructs null pointer.

[[Imported from SVN: r5732]]

Removed exit of the test program. Now the destructors are called and
there are no leaks any more.

[[Imported from SVN: r5731]]

Damn it. I forgot half the check-in.

New version without reinterpret casts. Please test!

[[Imported from SVN: r5730]]

[[Imported from SVN: r5729]]

[[Imported from SVN: r5728]]

[[Imported from SVN: r5727]]

[[Imported from SVN: r5726]]

[[Imported from SVN: r5725]]

* fix fs#652

[[Imported from SVN: r5724]]

[[Imported from SVN: r5723]]

[[Imported from SVN: r5722]]

[[Imported from SVN: r5721]]

[[Imported from SVN: r5720]]

[[Imported from SVN: r5719]]

[[Imported from SVN: r5718]]

thanks to Bård Skaflestad for the patch

[[Imported from SVN: r5717]]

[[Imported from SVN: r5716]]

to check and then (if one does not remember all FS tasks) one
starts to play around with configure flags and wasts a lot of
time.
For testing sllisttest one can always write
   make sllisttest
(at least I hope that is what EXTRA_PROGRAMS does)


[[Imported from SVN: r5715]]

[[Imported from SVN: r5714]]

since the shared_ptr change?)

[[Imported from SVN: r5713]]

[[Imported from SVN: r5712]]

[[Imported from SVN: r5710]]

for a program run via make check this takes to long (7min) and nothing additional seems to be checked through more iterations - if this check did something more which I do not understand please revert and comment

[[Imported from SVN: r5705]]

*Add const to *multiplyany()
*Implement *multiplyany() for FieldMatrix<k,1,1>
Thanks to Atgeirr Flø Rasmussen!

[[Imported from SVN: r5704]]

[[Imported from SVN: r5703]]

[[Imported from SVN: r5702]]