CGAL users discussion list

[Andreas Fabri <>]

I should add a precision. The callback is not passed as a reference, but
as a copy, so if your callback is not lightweight enough you need a
handle.

andreas



Andreas Fabri wrote:
> Hello,
>
> Shouldn't  adding
>
> operator()( const Box &box_a, const Box &box_b )
>
> to the class CallbackAid do it ?
>
> andreas
>
>
> Matthijs Sypkens Smit wrote:
> > Hi,
> >
> > I'm wondering if there is a clean way to use a class-method as a 
> > callback when intersecting boxes. I solved the problem by using a 
> > static method, but this feels really dirty. With the current 
> > implementation the class may only be instantiated once, or else things 
> > might break.
> >
> > Depending on the intersection of the boxes I want to access and modify 
> > additional data structures. Of course I would like this data to reside 
> > in a class rather than being global.
> > To make it more clear, here is the call I make:
> > CGAL::box_intersection_d( triangle_boxes.begin(), 
> > triangle_boxes.end(), query.begin(), query.end(), cb_aid.callback);
> >
> > cb_aid is of object of type CallbackAid, that looks roughly like this:
> > class CallbackAid;
> >
> > class CallbackAid
> > {
> > private:
> >     static CallbackAid *pObject;
> >     Point p;
> >
> > public:
> >     std::map<Point,double>  mydata;
> >
> >     CallbackAid() { pObject = this; }
> >     ~CallbackAid() {}
> >
> >     static void CallbackAid::callback( const Box &box_a, const Box 
> > &box_b )
> >     {
> >         CallbackAid *me = (CallbackAid *)pObject;
> >         [...]
> >         me->mydata[q] = value;
> >     }
> > }
> >
> > further I need to define the value of the static class variable:
> > CallbackAid * CallbackAid::pObject = (CallbackAid *)NULL;
> >
> > Depending on the result of the intersections I can thus modify 
> > 'mydata' which can be accessed by yet other routines. Is there a 
> > cleaner way to do this, preferably using a normal class method as a 
> > callback instead of the static one? I looked into this a while ago and 
> > remember then coming to the conclusion that is was not possible, but I 
> > would be happy to be told otherwise.
> >
> > If the problem is not clear, please say so and I'll try to clarify.