Re: [Nel] GNU/Linux ... trying to load landscape

[Vincent Caron]

On Mon, 2002-02-11 at 18:49, Loic Dachary wrote:
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 1024 (LWP 2365)]
> 0x0807eb21 in _STL::_Rb_global<bool>::_M_increment (__it=0xbffff74c)
>     at /usr/local/src/STLport-4.0/stlport/stl/_tree.c:295
> 295       if (_M_node->_M_right != 0) {
> (gdb) bt
> #0  0x0807eb21 in _STL::_Rb_global<bool>::_M_increment (__it=0xbffff74c)
>     at /usr/local/src/STLport-4.0/stlport/stl/_tree.c:295
> #1  0x40dec4ac in _STL::_Rb_tree_iterator<_STL::pair<_STL::basic_string<char, 
> _STL::char_traits<char>, _STL::allocator<char> > const, 
> NL3D::CMeshBase::CLightInfoMapList>, 
> _STL::_Nonconst_traits<_STL::pair<_STL::basic_string<char, 
> _STL::char_traits<char>,
> _STL::allocator<char> > const, NL3D::CMeshBase::CLightInfoMapList> > 
> >::operator++ (
>     this=0xbffff74c) at /usr/local/src/STLport-4.0/stlport/stl/_tree.h:182
> #2  0x40cb9bfc in NL3D::CScene::createInstance (this=0x87115d8,
>     address@hidden) at scene.cpp:484
> #3  0x40cbc574 in NL3D::CInstanceGroup::addToScene (this=0x8e43564,
>     address@hidden, driver=0x0) at scene_group.cpp:414
> #4  0x40c0fd5c in NL3D::CInstanceGroupUser::addToScene (this=0x8e43560,
>     address@hidden, driver=0x0) at instance_group_user.cpp:145
> #5  0x40c0fd26 in NL3D::CInstanceGroupUser::addToScene (this=0x8e43560,
>     address@hidden, driver=0x0) at instance_group_user.cpp:139
> #6  0x08070ed2 in initLandscape () at landscape.cpp:161
> #7  0x08063dd7 in main (argc=1, argv=0xbffffcd4) at client.cpp:194
> #8  0x419376cf in __libc_start_main () from /lib/libc.so.6
> 
>       _M_node is null, hence the core. 

You're running into a bug that was previously reported, but you're the
first one to reach it with NeL compiled in debug mode (I didn't have
enough HD space !). We run the demo by commenting the faulty line, but
we never actually managed to find out the real bug in there ... It seems
that this bug does not appear under Win32, so we're not sure if it's a
STL bug or something more subtle. Sorry, no more info :)