CGAL users discussion list

[GSBicalho <>]

What I want to do seems simple, just create 1000 points in a 4D sphere and
then triangulate them. Yet, for some reason, the code below only works
aproximately 1 in 20 times, perhaps less.

---CODE---

const int D = 4; // we work in Euclidean 4-space
typedef CGAL::Dimension_tag< D > Dim_tag;
typedef CGAL::Epick_d< Dim_tag >  Kernel;
typedef CGAL::Triangulation< Kernel > Triangulation;
typedef Triangulation::Point Point_4;
typedef CGAL::Delaunay_triangulation<Kernel> DenTriangulation;

int main() {
        CGAL::Random_points_in_ball_d<Point_4> rand_it(D, 1.0);
        std::vector<Point_4> points;
        CGAL::cpp11::copy_n(rand_it, 1000, std::back_inserter(points));

        Triangulation t(D);                      // create triangulation
        CGAL_assertion(t.empty());
        t.insert(points.begin(), points.end());  // compute triangulation
        CGAL_assertion(t.is_valid());
}

----------

Whenever it doesn't work, it throws the exception:

Exception thrown at 0x6ACD835A (libgmp-10.dll) in array_convex_hull_2.exe:
0xC0000005: Access violation reading location 0x011C3000. occurred

It looks to me that whenever the code does work, it was lucky that the
points fell in places that allowed /something/ specific to happen. I just
don't know what.

Can anyone confirm that this also happens to them?
Do the points of a triangulation require some arcane property to them that I
don't know about?




--
View this message in context: 
http://cgal-discuss.949826.n4.nabble.com/Triangulation-of-points-in-a-4D-Sphere-tp4662712.html
Sent from the cgal-discuss mailing list archive at Nabble.com.