CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

If you are simply doing B1 vs {b1,..,bn} then it might be interesting to
build an AABB_tree with {b1,..,bn} and then use
all_intersected_primitives with B1,B2.

There is no primitive class handling BBox but defining one is not hard
(copy-paste the one for triangle and adapt it to bbox).

The primitive concept is here:
http://www.cgal.org/Manual/beta/doc_html/cgal_manual/AABB_tree_ref/Concept_AABBPrimitive.html
The AABB_tree doc is here:
http://www.cgal.org/Manual/beta/doc_html/cgal_manual/AABB_tree_ref/Class_AABB_tree.html

Sebastien.


Bruno Chareyre wrote:
>   Something comes to mind... would it work to define B2 within the 
> callback function, or would it fool the sorting algorithm?
> void callback(...) { define (*B2); }
> box_intersection_d(B1,B2,b1,bN,callback);
>
>
> On 13/09/11 18:56, Bruno Chareyre wrote:
> > Hello,
> >
> > I was wondering if there was a smart way to use box_intersection_d() 
> > in this sort of task:
> > Say, I want to find intersections between box B1 and a set {b1,... 
> > bN}, then depending on the result I will define a new box B2 to check 
> > again.
> >
> > The trivial way would be:
> >
> > box_intersection_d(B1,B1,b1,bN);
> > Box B2(some result of the callback mechanism)
> > box_intersection_d(B2,B2,b1,bN);
> >
> > However, if I understand the algorithm correctly, this will trigger 
> > the sorting of {b1,... bN} twice. Correct?
> > Is there a way to avoid?
> >
> > Thanks in advance.
> >
> > Bruno
> > //
> > //
>
>
> --
> _______________
> Bruno Chareyre
> Associate Professor
> ENSE³ - Grenoble INP
> Lab. 3SR
> BP 53 - 38041, Grenoble cedex 9 - France
> Tél : +33 4 56 52 86 21
> Fax : +33 4 76 82 70 43
> ________________