CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

make_triangle creates 3 new vertices each time, even if a vertex with 
the same point exists.

The same goes for the incremental builder, each time you call
add_vertex, a new vertex is created. You need a side data-structure
to check if you already created a given vertex and get its index.

If you can't make everything in one builder call, you'll need to look
at the indexing mode documented here:

http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Polyhedron_ref/Class_Polyhedron_incremental_builder_3.html

Sebastien.


On 02/20/2013 11:47 AM, Chris Theis wrote:
> Hello,
>
> I've been trying to do some boolean ops on polyhedra and encountered a
> number of spurious asserts which I could trace back to problems in my
> polyhedron definition. However, I can't really see the problem and maybe
> there is some fundamental misunderstanding on my side.
>
> I'm trying to create for example a simple triangulated box. For this purpose
> I have written a builder function:
>
> ---------------------------------------------------------------------------------------------------------
>    void operator()( HDS&  HalfEdgeDS ) {
>      CGAL::Polyhedron_incremental_builder_3<HDS>  BuilderObj( HalfEdgeDS,
> true);
>
>      assert( !(m_Vertices.size() % 3) );      // m_Vertices = vector keeping
> all vertices of the polyhedron
>
>      for( size_t i = 0; i<  m_Vertices.size(); i += 3 ) {
>        BuilderObj.begin_surface(  3, 1 );
>
>          BuilderObj.add_vertex( Point( m_Vertices[i].X, m_Vertices[i].Y,
> m_Vertices[i].Z ) );
>          BuilderObj.add_vertex( Point( m_Vertices[i+1].X, m_Vertices[i+1].Y,
> m_Vertices[i+1].Z ) );
>          BuilderObj.add_vertex( Point( m_Vertices[i+2].X, m_Vertices[i+2].Y,
> m_Vertices[i+2].Z ) );
>
>          BuilderObj.begin_facet();
>            BuilderObj.add_vertex_to_facet( 0 );
>            BuilderObj.add_vertex_to_facet( 1 );
>            BuilderObj.add_vertex_to_facet( 2 );
>          BuilderObj.end_facet();
>
>        BuilderObj.end_surface();
>      }
>
>    }
> ---------------------------------------------------------------------------------------------------------
>
> When I then create a box I can export it to an OFF file and view it without
> any problem. However, if I call the is_valid() function I'm told that my
> structure is not valid. It contains 72 half-edges (I would have expected 6
> faces * 2 triangles = 36 half-edges) and every second HE is reported as a
> border half edge. Consequently, the solid does not classify as closed and
> boolean ops with nef-polyhedra fail.
>
> As a test I tried to build the solid with the make_triangle function like
> this:
> ---------------------------------------------------------------------------------------------------------
>    Poly.make_triangle( CGPoint( 0.0, 0.0, 0.0 ), CGPoint( 0.0, 1.0, 0.0 ),
> CGPoint( 1.0, 0.0, 0.0 ) );
>    Poly.make_triangle( CGPoint( 1.0, 0.0, 0.0 ), CGPoint( 0.0, 1.0, 0.0 ),
> CGPoint( 1.0, 1.0, 0.0 ) );
>
>    Poly.make_triangle( CGPoint( 0.0, 0.0, 0.0 ), CGPoint( 1.0, 0.0, 0.0 ),
> CGPoint( 0.0, 0.0, 1.0 ) );
>    Poly.make_triangle( CGPoint( 0.0, 0.0, 1.0 ), CGPoint( 1.0, 0.0, 0.0 ),
> CGPoint( 1.0, 0.0, 1.0 ) );
>
>    Poly.make_triangle( CGPoint( 0.0, 0.0, 0.0 ), CGPoint( 0.0, 0.0, 1.0 ),
> CGPoint( 0.0, 1.0, 0.0 ) );
>    Poly.make_triangle( CGPoint( 0.0, 1.0, 0.0 ), CGPoint( 0.0, 0.0, 1.0 ),
> CGPoint( 0.0, 1.0, 1.0 ) );
>
>    Poly.make_triangle( CGPoint( 0.0, 0.0, 1.0 ), CGPoint( 1.0, 0.0, 1.0 ),
> CGPoint( 0.0, 1.0, 1.0 ) );
>    Poly.make_triangle( CGPoint( 0.0, 1.0, 1.0 ), CGPoint( 1.0, 0.0, 1.0 ),
> CGPoint( 1.0, 1.0, 1.0 ) );
>
>    Poly.make_triangle( CGPoint( 1.0, 0.0, 0.0 ), CGPoint( 1.0, 1.0, 0.0 ),
> CGPoint( 1.0, 0.0, 1.0 ) );
>    Poly.make_triangle( CGPoint( 1.0, 0.0, 1.0 ), CGPoint( 1.0, 1.0, 0.0 ),
> CGPoint( 1.0, 1.0, 1.0 ) );
>
>    Poly.make_triangle( CGPoint( 0.0, 1.0, 0.0 ), CGPoint( 0.0, 1.0, 1.0 ),
> CGPoint( 1.0, 1.0, 0.0 ) );
>    Poly.make_triangle( CGPoint( 1.0, 1.0, 0.0 ), CGPoint( 0.0, 1.0, 1.0 ),
> CGPoint( 1.0, 1.0, 1.0 ) );
>
>
> ---------------------------------------------------------------------------------------------------------
>
> However, the result was exactly the same.
>
> I would appreciate if somebody could shed some light on this issue because I
> do not really understand where the border half-edges come from. In my
> current understanding border half edges are adjacent to holes and thus, have
> no face connected to them. However, I do not have any holes in my polyhedron
> or do I?
>
>
> Thanks a lot
> Chris
>
>
>
>
>
> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Polyhedron-validity-and-closure-tp4656773.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.