CGAL users discussion list

["Laurent Rineau (GeometryFactory)" <>]

On Friday 21 May 2010 09:27:10 fp2008 wrote:
> Hi Andreas,
> 
> I attached three OFF files. The remeshing tool crashes quickly when
> processing box.off and glass-mug.off, though it gets stuck instead with
> flash.off.

There are several issues.

** First, there is a non-exact use of the AABB_tree class, in the Polyhedron 
demo. Please apply the following patch:

--- demo/Polyhedron/Polyhedron_demo_remeshing_plugin_cgal_code.cpp.~BASE~     
  2010-05-25 11:24:12.000000000 +0200
+++ demo/Polyhedron/Polyhedron_demo_remeshing_plugin_cgal_code.cpp      
2010-05-21 15:12:43.000000000 +0200
@@ -110,10 +110,10 @@
 //
 // Types for meshing
 //
-typedef CGAL::Simple_cartesian<double> Simple_cartesian_kernel;
+// typedef CGAL::Simple_cartesian<double> Simple_cartesian_kernel;
 // input surface
-// typedef CGAL::Mesh_3::Robust_intersection_traits_3<Kernel> IGT;
-typedef 
CGAL::AABB_polyhedral_oracle<Polyhedron,Kernel,Simple_cartesian_kernel> 
Input_surface;
+typedef CGAL::Mesh_3::Robust_intersection_traits_3<Kernel> IGT;
+typedef CGAL::AABB_polyhedral_oracle<Polyhedron,Kernel,IGT> Input_surface;
 
 
 // A base non-templated class, to allow 


That will prevent the Surface_mesher from crashing on a simple cube surface. 
Actually, a cube a something particularly difficult to mesh:
  - there are a lot of coplanar points,
  - there are sharp edges (that the Surface_mesher package cannot mesh 
properly).


** Second, your file glass-mug.off is not a triangulated surface. Please use 
the menu entry Operations->Triangulate facets before launching the remeshing 
function.

** Third, your file flash.off is triangulated, but quite difficult to mesh: 
it has two cylindrical patches of surface that are very close from each 
other. It means that the least feature size of this surface 
is very small. You need to decrease the sizing field by a magnitude. However, 
I agree there is a problem in the surface mesher, because I have not been 
able to mesh that object without the assertion. Again, I 
know that the problematic feature is the coplanar facets at the bottom of the 
object. We have solutions for that, but the merge with the branch of CGAL-3.6 
is not easy.

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