CGAL users discussion list

["Oleksiy Busaryev" <>]

Hello,

 

I noticed very strange things in the streamlines computation code.

 

Looks like float/double input values are truncated to int's in a lot of places;

as a result, the streamlines computation works fine for the CGAL example Stream_lines_2

(because the data file datap.tri.cin for this example contains points

with huge coordinate values, e.g. 20000-60000 etc.)

and fails when run on various vector fields with coordinates < 1 (everything is truncated to 0).

 

For instance, here is the code from CGAL/stream_lines_2.h header:

 

for (int i=(int) (min_x-separating_distance);i<max_x+(int)
    separating_distance;i=i+(int) (fSepStl_seed))

--
Best regards,
Oleksiy Busaryev.