CGAL users discussion list

[王进 <>]

Hi,Philipp Moeller

you are right,Simple_cartesian quite faster than Cartesian,but crashed yet

I provide a piece of test data and code in attachment,please tell me the problem and give me some suggestions.

thanks very much.

WangJin

在 2012-03-18 21:21:02，"Philipp Moeller" <> 写道：
>王进 <> writes:
>
>> Hi,CGAL Developers:
>> I use Lindstrom-Turk edge-collapse algorithm for mesh simplification,but CGALcrashed at Handle_for.h  
>>
>> bool
>> is_shared() const
>>    {
>>         return ptr_->count > 1;   
>>     }
>>
>> because of ptr_ is invalid pointer.
>> the call stack show as below
>>
>> CGAL::Surface_mesh_simplification::EdgeCollapse
>> ->boost::put
>> ->CGAL::Point_3<CGAL::Cartesian<double> >::operator=
>> ->CGAL::PointC3<CGAL::Cartesian<double> >::operator=
>> ->CGAL::Vector_3<CGAL::Cartesian<double> >::operator=
>> ->CGAL::VectorC3<CGAL::Cartesian<double> >::operator=
>> ->CGAL::Handle_for<boost::array<double,3>,std::allocator<boost::array<double,3> > >::operator=
>> ->CGAL::Handle_for<boost::array<double,3>,std::allocator<boost::array<double,3> > >::~Handle_for
>> ->CGAL::Handle_for<boost::array<double,3>,std::allocator<boost::array<double,3> > >::is_shared() <---crashed
>>
>> for help!
>>
>> Wang 
>
>A hint first: Don't use Cartesian<double>, performance is probably going
>to be a lot better with Simple_cartesian<double>. Simple always means
>better, right? ;)
>
>Can you provide a complete, minimal test-case that shows the problem? I
>have used Mesh simplification extensively throughout last week and
>nothing blew up in my face, but then again I was using Simple_cartesian.
>
>Cheers,
>Philipp Moeller
>GeometryFactory
>
>-- 
>You are currently subscribed to cgal-discuss.
>To unsubscribe or access the archives, go to
>https://lists-sop.inria.fr/wws/info/cgal-discuss
>

Attachment: test.7z
Description: Binary data

Attachment: TestSurf.7z
Description: Binary data