CGAL users discussion list

[Mariette Yvinec <>]

too difficult ...

on the theoretical side, our meshing algorithm requires the contrained 
edges to be refined
in subedges having the Gabriel property, meaning that their diametral 
ball is empty of mesh vertices.
This is important because it guarantees that Steiner vertices are ony 
inserted inside
the domain which is crucial fr termination. Your will have to drop that 
requirement at least fro some edges.
Therefore if you want to avoid that the meshing algorithm goes on 
inserting Steiner vertices for ever
furthter and further away, make sure that the domain is surrounded by a 
bounding box
the edges of which  will be refined  on Gabriel edges

on the code sides, you will have to change several things.
If I am right, these things will be merely
in the class Refine_edge.h and/or  Refine_edge_base.h
all related to test_confilct_point***** :
no longer derive from  No_private_test_point_conflict
add a private_test_point_conflict
test_point_conflict_from_superior_impl



M. Hazegh wrote:
> I wonder if my question was too naive or too difficult?
>
> On Fri, Oct 10, 2008 at 2:21 PM, M. Hazegh 
> <>
>  wrote:
>   
> > Anybody can comment on this please?
> >
> >     

--
Mariette Yvinec
Geometrica project team
INRIA  Sophia-Antipolis