Here's an updated back trace with debugging symbols for plib ... the
segfault is on "typeinfo for puDial()" which is strange. Is this RTTI
related? I observe that the crash happens when the dialog box is being
closed, so could we have something out of order or a race condition between
the thread that creates/deletes dialog boxes and the render thread, where
the dialog box is destroyed while the render thread is in the middle of
drawing the scene? The random nature of the crash could fit in with the
theory that it's a race condition with the render thread.
How about we get some people opening and closing a lot of dialog boxes with
the latest cvs version of everything to see I'm the only one seeing this?
(I'm on fedora 11 with osg-2.9.5 which is the latest developer release as of
a few minutes ago.)
We really need to stomp out all the hangs and crashes before we can even
think about another public release!
Thanks,
Curt.
(gdb) where
#0 0x086a9620 in typeinfo for puDial ()
#1 0x0862b68b in puGroup::draw (this=0x17d49d88, dx=0, dy=0)
at puGroup.cxx:314
#2 0x0862b68b in puGroup::draw (this=0xb736e4d8, dx=0, dy=0)
at puGroup.cxx:314
#3 0x086287ed in puDisplay () at pu.cxx:304
#4 0x08070371 in SGPuDrawable::drawImplementation(osg::RenderInfo&) const
()
#5 0x007c61a7 in osgUtil::RenderLeaf::render(osg::RenderInfo&,
osgUtil::RenderLeaf*) () from /usr/local/lib/libosgUtil.so.60
#6 0x007bf7b2 in osgUtil::RenderBin::drawImplementation(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
#7 0x007bf43e in osgUtil::RenderBin::draw(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
#8 0x007bf7f4 in osgUtil::RenderBin::drawImplementation(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
#9 0x007c884e in osgUtil::RenderStage::drawImplementation(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
#10 0x007bf43e in osgUtil::RenderBin::draw(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
#11 0x007ca3a0 in osgUtil::RenderStage::drawInner(osg::RenderInfo&,
osgUtil::RenderLeaf*&, bool&) () from /usr/local/lib/libosgUtil.so.60
#12 0x007cdb93 in osgUtil::RenderStage::draw(osg::RenderInfo&,
osgUtil::RenderLeaf*&) () from /usr/local/lib/libosgUtil.so.60
---Type <return> to continue, or q <return> to quit---
#13 0x007d56c2 in osgUtil::SceneView::draw() ()
from /usr/local/lib/libosgUtil.so.60
#14 0x00afcfed in osgViewer::Renderer::draw() ()
from /usr/local/lib/libosgViewer.so.60
#15 0x00afa697 in osgViewer::Renderer::operator()(osg::GraphicsContext*) ()
from /usr/local/lib/libosgViewer.so.60
#16 0x00fa8963 in osg::GraphicsContext::runOperations() ()
from /usr/local/lib/libosg.so.60
#17 0x00fae1dd in osg::RunOperations::operator()(osg::GraphicsContext*) ()
from /usr/local/lib/libosg.so.60
#18 0x00fae467 in osg::GraphicsOperation::operator()(osg::Object*) ()
from /usr/local/lib/libosg.so.60
#19 0x00ffeedd in osg::OperationThread::run() ()
from /usr/local/lib/libosg.so.60
#20 0x00fae4f9 in osg::GraphicsThread::run() ()
from /usr/local/lib/libosg.so.60
#21 0x00523e89 in OpenThreads::ThreadPrivateActions::StartThread(void*) ()
from /usr/local/lib/libOpenThreads.so.11
#22 0x006a5935 in start_thread () from /lib/libpthread.so.0
#23 0x02cb894e in clone () from /lib/libc.so.6
(gdb)
On Tue, Dec 1, 2009 at 11:24 AM, Curtis Olson wrote:
> On Tue, Dec 1, 2009 at 11:21 AM, Curtis Olson wrote:
>
>> On Tue, Dec 1, 2009 at 11:06 AM, James Turner wrote:
>>
>>>
>>> On 1 Dec 2009, at 16:59, Curtis Olson wrote:
>>>
>>> > I've been observer quite frequent segfaults in the most recent version
>>> of FlightGear. I've started running under gdb and so far the trend seems to
>>> be that the crash is inside the pui code and as a result of clicking "ok" or
>>> "cancel" on a dialog box. Here's an example backtrace (the build is without
>>> -g though). Anyone else seeing this?
>>>
>>> Not seeing that here - I assume if it was any particular dialog box, you
>>> would have mentioned that.
>>>
>>
> Sorry for the multiple emails ... ok it just happened again to me with the
> autopilot dialog box. When I clicked on the "close" button, I got a
> segmentation fault ... with a backtrace *very* similar to what I posted
> before.
>
> Curt.
> --
> Curtis Olson:
> http://baron.flightgear.org/~curt/<http://baron.flightgear.org/%7Ecurt/>
>
--
Curtis Olson: http://baron.flightgear.org/~curt/
------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
