CGAL users discussion list

["Laurent Rineau (CGAL/GeometryFactory)" <>]

Le mardi 02 octobre 2012 11:31:33 Amir Vaxman a écrit :
> Hello,
> 
> I have recently converted my religion from Windows to Linux Mint, and
> am trying to use CGAL copiled by g++ 4.6. I have configured cmake
> using:
> 
> amir@amir-office ~/workspace/CGAL-4.0.2 $ cmake -fp-model strict .

Not that cmake has silently ignored the options "-fp-model strict".

> -- Build CGAL from release in directory CGAL-4.0.2
> -- Packagenames: CGAL-4.0.2
> -- CGAL_MAJOR_VERSION=4
> -- CGAL_MINOR_VERSION=0
> -- CGAL_BUGFIX_VERSION=2
> -- CGAL_SONAME_VERSION=9
> -- CGAL_SOVERSION     =9.0.0
> -- CGAL_REFERENCE_CACHE_DIR=
> -- Building shared libraries
> -- Targetting Unix Makefiles
> -- Using /usr/bin/c++ compiler.
> -- USING CMake version: 2.8.9
> -- System: Linux
> -- USING GCC_VERSION = '4.6.3'
> -- Using gcc version 4 or later. Adding -frounding-math
> -- Build type: Release
> -- USING CXXFLAGS = ' -frounding-math -O3 -DNDEBUG'
> -- USING LDFLAGS = '  '

We can see that CGAL libraries are configured to be compiled with the -
frounding-math option.

> the building went ok, but I get the following assertion when running:
> 
> CGAL_assertion_msg(-CGAL_IA_MUL(-1.1, 10.1) != CGAL_IA_MUL(1.1, 10.1),
>                          "Wrong rounding: did you forget the
> -frounding-math  option if you use GCC (or  -fp-model strict  for
> Intel)?");
> 
> It seems as if this option was added, so what is the problem?

When running what? Between the configuration+compilation of the CGAL 
libraries, and the display of that error, you must have compiled an 
executable? How have you configured it? What compilation options? Every 
compilation unit that use CGAL filtered kernels must be compiled with the -
frounding-math option, and not only the CGAL libraries.

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