CGAL users discussion list

["Garth N. Wells" <>]

On 13 February 2012 19:32, Laurent Rineau (GeometryFactory)
<>
 wrote:
> Le lundi 13 février 2012 19:12:38 Garth N. Wells a écrit :
>> On 13 February 2012 19:08, Laurent Rineau (GeometryFactory)
>>
>> <>
>>  wrote:
>> > Le lundi 13 février 2012 19:03:21 Garth N. Wells a écrit :
>> >> On 8 February 2012 19:17, Garth N. Wells 
>> >> <>
>> >>  wrote:
>> >> > On 8 February 2012 10:43, Laurent Rineau (GeometryFactory)
>> >> >
>> >> > <>
>> >> >  wrote:
>> >> >> Le mardi 07 février 2012 12:22:52 Garth Wells a écrit :
>> >> >>> I've been running into some issues with mesh generation in 3D. I'm
>> >> >>> creating
>> >> >>> meshes of rather simple polyhedra (tetrahedra, cubes), but have been
>> >> >>> finding that the generation randomly fails at times. I'm also seeing
>> >> >>> a
>> >> >>> dependency on the gcc version and optimisation level. The code at 
>> >> >>> the
>> >> >>> end of this message, taken from
>> >> >>>
>> >> >>>
>> >> >>> http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Mesh_3/Chapter
>> >> >>> _ma
>> >> >>> in.h tml#Subsection_50.3.5
>> >> >>>
>> >> >>> runs without error with g++ 4.6.1, but with g++ 4.4.6 and the -O2 or
>> >> >>> -O3
>> >> >>> flag I get the error
>> >> >>>
>> >> >>>     terminate called after throwing an instance of
>> >> >>> 'CGAL::Precondition_exception'
>> >> >>>       what():  CGAL ERROR: precondition violation!
>> >> >>>     Expr: index != Subdomain_index()
>> >> >>>     File:
>> >> >>> /usr/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h
>> >> >>>     Line: 160
>> >> >>>     Aborted
>> >> >>>
>> >> >>> I've tested with CGAL 3.8 and 3.9. The code is compiled using
>> >> >>>
>> >> >>>     g++ -O2 -frounding-math  test.cpp -lCGAL -lboost_thread -lmpfr
>> >> >>> -lgmp
>> >> >>>
>> >> >>> Any help with this would be greatly appreciated.
>> >> >>
>> >> >> Can you please add:
>> >> >>
>> >> >> #define CGAL_MESH_3_VERBOSE 1
>> >> >>
>> >> >> as the first line of your cpp file, recompile, and give us the output
>> >> >> of
>> >> >> the run, when it fails?
>> >> >
>> >> > It get:
>> >> >
>> >> > Inserting protection balls...
>> >> > insert_corners() done. Nb of points in triangulation: 25
>> >> > insert_balls_on_edges() done. Nb of points in triangulation: 610
>> >> > refine_balls() done. Nb of points in triangulation: 616
>> >> >
>> >> > Start surface scan...end scan. [Bad facets:1202]
>> >> >
>> >> > Refining Surface...
>> >> > Legende of the following line: (#vertices,#steps,#facets to
>> >> > refine,#tets to refine)
>> >> > (1741,1124,0,0) (328.6 vertices/s)))
>> >> > Total refining surface time: 3.42021s
>> >> >
>> >> > Start volume scan...end scan. [Bad tets:2643]
>> >> >
>> >> > Refining...
>> >> > Legende of the following line: (#vertices,#steps,#facets to
>> >> > refine,#tets to refine)
>> >> > (1742,1,0,2642) (250.0 vertices/s)terminate called after throwing an
>> >> > instance of 'CGAL::Precondition_exception'
>> >> >  what():  CGAL ERROR: precondition violation!
>> >> > Expr: index != Subdomain_index()
>> >> > File: 
>> >> > /usr/include/CGAL/Mesh_3/Mesh_complex_3_in_triangulation_3_base.h
>> >> > Line: 160
>> >> > Aborted
>> >>
>> >> Considering that the failing example code comes from the CGAL
>> >> examples, I suppose that it should be considered a CGAL bug. Is there
>> >> a bug tracker where I can report it as a bug?
>> >
>> > Sorry, I forgot your mail.
>> >
>> > It seems you have modified the example. What off file have you used?
>>
>> The example file has two examples and some ifdefs - my post has just
>> the simple part.
>>
>> The off file is the fandisk.off file that is in
>>
>>     CGAL-3.9/examples/Mesh_3/data/fandisk.off
>
> Thanks. That looks like a bug. Can you give me the result of `uname -a`? I
> will try to reproduce that bug on one of my machines, and fix it.
>

Linux gnw-t410i 3.0.0-16-generic #28-Ubuntu SMP Fri Jan 27 17:44:39
UTC 2012 x86_64 x86_64 x86_64 GNU/Linux


> Does the bug occur at every run?
>

Yes.

Garth

> --
> Laurent Rineau, PhD
> R&D Engineer at GeometryFactory           http://www.geometryfactory.com/
> Release Manager of the CGAL Project       http://www.cgal.org/
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss
>