CGAL users discussion list

[Olivier Devillers <>]

Le 11/15/13 4:14 PM, Martin Uhrin a écrit :
> Thanks Monique.
>
> I tried this approach.  Or one that I think is equivalent:
>
> delaunay.incident_edges(delaunay.infinite_vertex());
>
> but as you said, this produces the convex hull of the triangulation.  This
> would give me edge 1-3, however I believe the dual of the ray that passes
> through the label b is in fact edge 1-2 which is the one I want, followed by
> edge 2-3.

The dual of b is not a ray but a finite Voronoi edge.
The fact that on your drawing you clip your Voronoi diagram
and the center of the circle circumscribnig triangle 123
is outside the screen does not mean it does not exist.
It exists, has been computed, and is finite.