CGAL users discussion list

Mahmood Naderan wrote:
> > You have to give src+tgt. And three other points. A third point defined 
> > only a unique circle. In addition, not every combination of points 
> > defines a connected arc.
>
> Sorry I did not understand. What do you mean by src+tgt?

source+target
As Eric recalled, you need _5_ points to uniquely define a conic, 
moreover, the 5 points do not always lie on the same arc of the conic.

> Also will "CGAL::do_inersect" help me? according to manual, these types:
> Polygon 2 
> Polygon with holes 2 
> General polygon 2
> General polygon with holes 2
>  
> are supported by that function. So does this mean conic arc is not supported? (I think it is an abvious question)

right, more generally there are no conics nor conic arcs supported in 
the CGAL Kernels.
As Eric said, you may find some functionality in Arr_conic_traits_2.h (a 
traits class for CGAL arrangements), but _not_ in the kernel.

Circular arcs (restricted case of conic arcs) are supported by the 2D 
Circular Kernel. A circular arc can be uniquely defined by 3 points.

        Monique Teillaud

> ----- Original Message ----
> From: Eric Berberich 
> <>
> To: 
>
> Sent: Thursday, February 28, 2008 7:26:55 PM
> Subject: Re: [cgal-discuss] line and curve intersection
>
> Mahmood Naderan wrote:
>
> > > Can you specify "curve" more detailed? 
> >
> > Hi,
> > I mean a "conic arc". I think a conic arc is defined by three point. Am I 
> > right?
>
>
> You have to give src+tgt. And three other points. A third point defined 
> only a unique circle. In addition, not every combination of points 
> defines a connected arc.
>
> In CGAL, there is Arr_conic_traits_2.h, which also offers an 
> intersection predicate. I am not sure, whether it offers line/conic or 
> whether it line can be input as "degenerated" conic.
>
> eriC