CGAL users discussion list

["peter heppel" <>]

Team.

 

I’ve been getting some failures in function precondition_fail ( assertions.cpp  CGAL 3.3)

The following change seems to fix it.

If I knew how to make a patch I’d have done that instead!!

 

void

precondition_fail( const char* expr,

                        const char* file,

                        int         line,

                        const char* msg )

{

    (*_error_handler)("precondition", expr, file, line, msg);

    switch (_error_behaviour) {

    case ABORT:

        CGAL_CLIB_STD::abort();

    case EXIT:

        CGAL_CLIB_STD::exit(1);  // EXIT_FAILURE

    case EXIT_WITH_SUCCESS:

        CGAL_CLIB_STD::exit(0);  // EXIT_SUCCESS

    case THROW_EXCEPTION:

            if(msg)

                  throw Precondition_exception("CGAL", expr, file, line, msg);

            else

                  throw Precondition_exception("CGAL", expr, file, line, "with no Message");

    case CONTINUE:

        ;

    }

}

 

And the question…

Is it OK to mix kernel types in a single application?  

 

The reason I ask is that I’m using the Exact_predicates_inexact_constructions_kernel because the exact kernel is too slow.

 

Not surprisingly I get quite a lot of precondition exceptions, for instance when trying to intersect nearly-contiguous polygons.

 

It would be nice to catch these exceptions and then recover by trying the same operation with the exact kernel. Is this workable?

 

 

 

Peter
email:  
(SUI)     +41  218 621 015

(mob)   +33   681307652

Home   +33 2 97 82 19 30

            Peter Heppel Associates

            8 rue du Faubourg Poissonnière

            75010 Paris – France

www.peterheppel.com