CGAL users discussion list

["T.vanLankveld" <>]

I'm afraid I don't have a concrete example. The incident in which I 
experienced this problem used a 3D data set projected onto a plane. 
Saving the new 2D data points and retrying using those didn't give the 
same exception, probably because of roundoff when saving.

However, I can imagine the same problem will turn up using a 3 point 
data set with the coordinates (0, 0),  (-DBL_MAX, 0), and (DBL_MAX, 
DBL_MIN), but I don't know if this would already cause the Delaunay 
triangulation to throw an exception.


Sebastien Loriot (GeometryFactory) [via cgal-discuss] wrote:

> T.vanLankveld wrote:
>
> > As an additional note, I did some tests on my data and found
> something I
> > thought was weird, but this may seem like a beginners mistake to the
> more
> > veteran CGAL-programmers:
> >
> > I did the following tests on each triangle in the Delaunay
> triangulation
> > before using the points to construct the alpha-shape:
> >
> > - triangle.is_degenerate()
> > - CGAL::collinear(p0, p1, p2)
> > - Rational rdx1 = Rational(p1.x())-Rational(p0.x()), rdy1 =
> > Rational(p1.y())-Rational(p0.y());
> >   Rational rdx2 = Rational(p2.x())-Rational(p0.x()), rdy2 =
> > Rational(p2.y())-Rational(p0.y());
> >   rdx2*rdy1 == rdx1*rdy2
> > - NumberType dx1 = p1.x()-p0.x(), dy1 = p1.y()-p0.y();
> >   NumberType dx2 = p2.x()-p0.x(), dy2 = p2.y()-p0.y();
> >   CGAL::compare(dx2*dy1, dx1*dy2) == CGAL::EQUAL
> > - dx2*dy1 == dx1*dy2
> >
> > Only the last two checks returned true when using a
> > Filtered_kernel<Cartesian<double>>.
> > Rational here is Quotient<MP_Float> and NumberType is double.
> >
> Do you have a minimal example?
>
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss
>
>
>
> ------------------------------------------------------------------------
> View message @
> http://cgal-discuss.949826.n4.nabble.com/CGAL-NTS-is-zero-den-with-Delaunay-Alpha-shapes-tp952983p2262829.html 
>
> To unsubscribe from Re: CGAL_NTS is_zero(den) with Delaunay
> Alpha-shapes, click here
> <<Link Removed>>.
>
>

-- 
Drs. Thijs van Lankveld
+31(0)30-2534686
[hidden email]

---

View this message in context: Re: CGAL_NTS is_zero(den) with Delaunay Alpha-shapes
Sent from the cgal-discuss mailing list archive at Nabble.com.