(best solution would be not to allocate empty steps at all -- todo...)

Caveats:

- We don't support new MED3 features (evolving meshes, multi-profile per field step (?))
- I still get crashes on files that used to work with MED2. However the mdump code
  shipped with MED3 also crashes on these files, so the bug might not be with Gmsh...

speed-up reading post-pro files in .msh with huge number of time steps (usually sparse) - for gauthier

- fixed MED node ordering
- polished 64 bit Cocoa Mac OS X FLTK version
- updated copyright

trying to fix loading of view without partition, if view with identical name without partition already exists

No commit message

Reverted all fixes for Post/PViewDataGModelIO.cpp, I was just generating
inconsistent post-processing data, and geuzaine's fix on fix created
an offset between vertix indices saved by GModel::WriteMSH and
PViewDataGModel::writeMSH. 

AND a hotfix to Post/PViewDataGModelIO.cpp which was saving NodeData with
negative MVertex indices in .msh files.

but stored globally in the model)

- Plugin(Skin) now works on adapted datasets
- addData() should now also work for high order datasets [we still need to
  add a new type+ref to FS in PViewDataGModel]

Add FullDg formulation
Add Stress and InternalPoint (development)

separate multistep, multipartition datasets)

Changed homology generator chain mesh element orientations (determined by vertex order) so that
post-processing data isn't needed to save the generators, instead all the
elements have coefficient 1 now in the chain. Due to this also reverted
changes made to GModel writeMSH method in my earlier commit (they became obsolete).

entities.

Added option to save MSH without renumbering the mesh elements (required to
save post-processing $ElementData right!).

added GModel::getMeshElementByTag (uses a dynamically created cache, like
what we do for mesh vertices)

- moved GUI to FlGui + added experimental QGui to test Qt

- fixed most MSVC and gcc compiler warnings

- converted some remaining tabs to whitespace (that's why the patch is
  so large; please try to remember to indent with 2 spaces and not
  tabs when you commit stuff to CVS)

- merged modified patch from Konstantinos Poulios to allow saving
  physical/partition as element tags in mesh formats that don't
  support them natively

- unstructured mesh algos (+ refs)
- interpolation matrices in .pos format