CGAL users discussion list

["Laurent Rineau (GeometryFactory)" <>]

On Wednesday 14 October 2009 14:58:54 Marco Aurelio Sterpa wrote:
> Sorry I didn't read that page, anyway I'm using CGAL 3.5 installed with
> cmake on Ubuntu 8.04 with kernel 2.6.24-24-generic; the compiler version is
> gcc version 4.2.4. I got a "floating point exception" with this line:
> 
> while(Q2.top().point().x() <= x_u && !Q2.empty())
> 
> where NT x_u = -1/v_ij.x();
> 
> and v_ij is a Point_2 object given by the intersection of two lines. I hope
> this informations are enough to get into the problem. Thanks in advance,
> best regards.

You were talking about the script cgal_create_makefile. That script is no 
longer shipped within CGAL>=3.4. I guess you are mixing two different 
versions 
of CGAL, on that was installed by a Ubuntu package (or that you have 
installed 
by hand in the past), and one that you installed recently by hand. What is 
the 
value of your environment variable $CGAL_MAKEFILE?

Now that CGAL has switch to CMake, you should consider using the script 
cgal_create_cmake_script, that does the same job in the CMake world: it 
creates a file named CMakeLists.txt, which can be used by CMake. With CMake, 
to add -g, you just have to modify the CMake variable named CMAKE_BUILD_TYPE, 
in the CMake GUI (command line: "cmake-gui $PWD"), so that its value is 
"DEBUG".


If you are unwilling to use CMake, in your Makefile, add "-g" in the line 
that 
starts with:
  CXXFLAGS = 
so that it starts with:
  CXXFLAGS = -g
And make sure you have at least the flag -frounding-math in your calls to 
g++, 
during the compilation, otherwise the filtered arithmetic of CGAL kernels 
cannot run correctly.

-- 
Laurent Rineau, PhD
R&D Engineer at GeometryFactory           http://www.geometryfactory.com/
Release Manager of the CGAL Project       http://www.cgal.org/