CGAL users discussion list

[Benjamin Kehlet <>]

Hello again!

Here is a simpler program that (if I understand the snap rounding
package correctly) reveals a bug

---
#include <CGAL/Cartesian.h>
#include <CGAL/Quotient.h>
#include <CGAL/MP_Float.h>
#include <CGAL/Snap_rounding_traits_2.h>
#include <CGAL/Snap_rounding_2.h>

typedef CGAL::Quotient<CGAL::MP_Float>           Number_type;
typedef CGAL::Cartesian<Number_type>             Kernel;
typedef CGAL::Snap_rounding_traits_2<Kernel>     Traits;
typedef Kernel::Segment_2                        Segment_2;
typedef Kernel::Point_2                          Point_2;

int main(int argc, char** argv)
{
  Point_2 p1(1, 0);
  Point_2 p2(0, 0);
  Number_type pixel_width(0.1);

  CGAL::Hot_pixel<Traits> hp(p1, pixel_width);
  CGAL::Segment_data<Traits> sd(p1, p2);

  CGAL::SEG_Direction dir;
  sd.determine_direction(dir);

  if (!hp.intersect(sd, dir))
    std::cout << "Error! Segment did not intersect hot pixel of
endpoint" << std::endl;
}
---

This hits the bug, but only with a debug build. However, my guess is
that this is not related to code that is removed when doing a release
build. I think it is more likely that something in the creation of the
hot pixel or in the intersection test is somehow prone to roundoff
errors.

I will try to investigate more. I would be nice if someone who know
the package better than me could confirm that I'm not doing something
completely wrong.

Best regards

Benjamin Kehlet

2014-06-27 1:58 GMT+02:00 Benjamin Kehlet 
<>:
> Hello!
>
> I am attempting to use the 2D snap rounding package to safely get from
> exact numbers to float in our mesh generation tool in FEniCS. However,
> I get some (to me) unexpected results. Here is a simple example:
> When running this simple modification of
> examples/Snap_rounding_2/snap_rounding.cpp:
> ---
> #include <CGAL/Cartesian.h>
> #include <CGAL/Quotient.h>
> #include <CGAL/MP_Float.h>
> #include <CGAL/Snap_rounding_traits_2.h>
> #include <CGAL/Snap_rounding_2.h>
>
> typedef CGAL::Quotient<CGAL::MP_Float>           Number_type;
> typedef CGAL::Cartesian<Number_type>             Kernel;
> typedef CGAL::Snap_rounding_traits_2<Kernel>     Traits;
> typedef Kernel::Segment_2                        Segment_2;
> typedef Kernel::Point_2                          Point_2;
> typedef std::list<Segment_2>                     Segment_list_2;
> typedef std::list<Point_2>                       Polyline_2;
> typedef std::list<Polyline_2>                    Polyline_list_2;
>
> int main()
> {
>   Segment_list_2 seg_list;
>   Polyline_list_2 output_list;
>
>   seg_list.push_back(Segment_2(Point_2(1, 0), Point_2(0, 0)));
>
>   // Generate an iterated snap-rounding representation, where the centers of
>   // the hot pixels bear their original coordinates, using 5 kd trees:
>   
> CGAL::snap_rounding_2<Traits,Segment_list_2::const_iterator,Polyline_list_2>
>     (seg_list.begin(), seg_list.end(), output_list, .1, true, false, 5);
>
>   int counter = 0;
>   Polyline_list_2::const_iterator iter1;
>   for (iter1 = output_list.begin(); iter1 != output_list.end(); ++iter1) {
>     std::cout << "Polyline number " << ++counter << ":\n";
>     Polyline_2::const_iterator iter2;
>     for (iter2 = iter1->begin(); iter2 != iter1->end(); ++iter2)
>       std::cout << "    (" << iter2->x() << ":" << iter2->y() << ")\n";
>   }
>
>   return(0);
> }
> ---
>
> I get this result:
> benjamik@benjamik-ThinkPad-X300:~/software/CGAL-4.4/examples/Snap_rounding_2/build$
> ./snap_rounding
> Polyline number 1:
>     (0.1/2:0.1/2)
> benjamik@benjamik-ThinkPad-X300:~/software/CGAL-4.4/examples/Snap_rounding_2/build$
>
> So the one segment in the input seems to be "reduced" to a single
> point. Is this expected? If the pixel size argument in the call to
> CGAL::snap_rounding_2() is changed to "Number_type(1, 10)", then I get
> the output that I would expect:
>
> benjamik@benjamik-ThinkPad-X300:~/software/CGAL-4.4/examples/Snap_rounding_2/build$
> ./snap_rounding
> Polyline number 1:
>     (210/200:10/200)
>     (10/200:10/200)
> benjamik@benjamik-ThinkPad-X300:~/software/CGAL-4.4/examples/Snap_rounding_2/build$
>
> This may be due to my limited knowledge of the
> Quotient<CGAL::MP_float> number type, but from the reference manual it
> seems that it should be constructable from a double.
>
> This occurs with a debug build of CGAL 4.4. I get the equal results on
> a 32-bit Ubuntu 13.10 box with gcc 4.8.1 and with a 64-bit debian
> Jessie box with gcc 4.9.0.
>
> Best regards
>
> Benjamin Kehlet