Re: [osg-users] osgDB::Registry::instance(true) sometimes segfaults

2013-08-09 Thread Chris Stankevitz
On Fri, Aug 9, 2013 at 11:02 AM, Chris Stankevitz wrote: > Interestingly, the segfault only happens when I link to osgEarth, > eventhough I am not using osgEarth. Solved, although I don't understand the technical mechanism by which the failure happened. osgEarth static objects were being initial

Re: [osg-users] osgDB::Registry::instance(true) sometimes segfaults

2013-08-09 Thread Chris Stankevitz
On Wed, Aug 7, 2013 at 12:27 PM, Chris Stankevitz wrote: > #include > > int main() > { > osgDB::Registry::instance(true); > > return 0; > } Here is a better callstack: #0 0x74cd56e0 in malloc_consolidate () from /lib64/libc.so.6 #1 0x74cd6148 in _int_free () from /lib64/li

Re: [osg-users] osgDB::Registry::instance(true) sometimes segfaults

2013-08-07 Thread Chris Stankevitz
On Wed, Aug 7, 2013 at 12:27 PM, Chris Stankevitz wrote: > The first program does not segfault. The second does segfault: I neglected to mention these important details: - Loading the node file but not clearing the registry singleton does not segfault. However it does report "Inconsistency det

[osg-users] osgDB::Registry::instance(true) sometimes segfaults

2013-08-07 Thread Chris Stankevitz
Hello, The first program does not segfault. The second does segfault: === #include int main() { osgDB::Registry::instance(true); return 0; } === #include int main() { osgDB::readNodeFile("/home/cstankevitz/Work/2013-07-02_osgearth/osgearth/tests/crs.earth"); osgDB::Registry::in