CGAL users discussion list

["Laurent Rineau (CGAL/GeometryFactory)" <>]

Le samedi 30 juin 2012 00:55:11 Zohar a écrit :
> What do you mean fuzzy? I copy-pasted the exact error.

I know the error. As one of the author of the Mesh_3 code, I can tell that 
there most probable error is in the definition of the mesh domain. What is 
missing in you report is the definition of the domain. See below.
> 
> My inside is negative, and the outside is positive, and it does work, but
> not on all meshes, i.e. buggy.
> Even if I confused between the two, then since I used a sphere domain:
> 
> Mesh_domain domain(cgf, GT::Sphere_3(inner_pt, bounding_sphere_radius));

What do you call a sphere domain? What is the type 'Mesh_domain' in your 
code? 
My guess is that it is:
  CGAL::Implicit_mesh_domain_3<Function, BGT>
but I would need to know which Function and BGT parameters you have used in 
your code.

What is more, what is 'cgf' in that line? Probably the function. Are you sure 
that the Sphere(inner_pt, bounding_sphere_radius) defines a sphere that 
actually bounds the implicit domain? Remember that the sphere radius should 
be 
squared (see the documentation of CGAL::Sphere_3).

In case the sphere given as second parameter of the 

> I should have gotten a sphere of the size of my domain, with a hole in the
> middle in the shape of my object.
> 
> In any case I can't see why the mesh generation should crash on any input.
> Are there any restrictions on the input that I should have followed?

The documentation of the Implicit_mesh_domain says that the bounding sphere 
must bound the domain.

The issue with an implicit domain is that one has to know a bounding sphere 
of 
it, and be careful with the definition of that bounding sphere. Otherwise the 
boundary of the mesh domain is clipped by the sphere. In case the sphere is 
too small, that can gives an open boundary. And then probably the assertions 
you have.

But anyway that is only a guess. Please copy paste the definition of the 
types 
GT, Function, and Mesh_domain, so that it can be confirmed.

-- 
Laurent Rineau, PhD
R&D Engineer at GeometryFactory           http://www.geometryfactory.com/
Release Manager of the CGAL Project       http://www.cgal.org/