prudently put between #if 0 #endif Those should allow
to produce better quality quad meshes and more robust
triangular meshes (random factor has disappeared) : 
Technically, the divide and conquer algo has been 
replaced by a bowyer-watson with robust predicates.

boundary layers in 3D are advancing (filtering intersecting layers using rtrees)
randomness of 2D meshes is now close to be ok (we still have problems with OCC ...)

TODO : avoid topological unconsistencies for (e.g.) cones when
delaunay meshing is involved.

IMPORTANT : 3D Delaunay Meshing now acts on connected volumes
	  separately, allowing to mesh regions that are not connected
	  topologically but that may intersect geometrically

Removed the -q option in initial mesh generation (generates meshes that do not respect size field)

bug fixed in delaunay surface meshing with embedded edges
bug fixed in cross filed computation for non conformal mappings 

natural convection added in the incompressible code

No commit message

not yet finished ... mesh smoothing has to be adapted 

enhanced bl capabilities (avoid self intersection of layers)

     boundary layers in 2D 
     periodi bug fixed
     many other stuff

JF    

The delquad algo is now much more stable and nice

Two 3D meshers have been added. 

One is an embryo of the 3D delhex mesh which is an extension of delquad.

The other one is MMG, the mobile mesh generator of CECILE DOBRZYNSKI (Bordeaux I) and PASCAL FREY (PARIS VI). We are very happy that this new anisotropic 3D mesher is now available as an open source !!

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

- fix bb computation to avoid model jumping around when adding points in z=0 plane
- pass interpol matrices as ref

untabify

- temp fix I/O to allow skipping $xx fields (this will be common
  with new post-pro)
- prune MElement (move mesh-only stuff in mesh algos)