Menu
▾
▴
opensg-users
|
From: Rasmus A. <ra...@vr...> - 2002-11-27 13:42:26
Attachments:
transparency.wrl
|
Hi With a particular (very simple) wrl model I have a cluster pronlem: On the client side in the display function the StreamSockConnection::readBuffer() throws a ReadError exception with the message "peek got 0 bytes!". Right after this, one server segfaults and then the client side says "Broken pipe" (probably an OS statement?) and exits. Any idea what causes this? PS: I attached the wrl file (my transparency test, which you may have seen before :-)) -- Rasmus Agerholm Research Assistant VR Media Lab (+45) 9635 8792 Niels Jernes Vej 14 DK-9220 Aalborg Phone: (+45) 9635 8784 Fax: (+45) 9815 2444 http://www.vrmedialab.dk/~rasta |
|
From: Marcus R. <mar...@ig...> - 2002-11-27 14:52:50
|
Hi, I'm able to view your wrl file with testClusterClient and testClusterServer. How old is your OpenSG-Build? Did the server crash or the client? could you start server and client with gdb? Marcus Rasmus Agerholm wrote: > Hi > > With a particular (very simple) wrl model I have a cluster pronlem: > > On the client side in the display function the > StreamSockConnection::readBuffer() throws a ReadError exception with the > message "peek got 0 bytes!". Right after this, one server segfaults and > then the client side says "Broken pipe" (probably an OS statement?) and > exits. > > Any idea what causes this? > > PS: I attached the wrl file (my transparency test, which you may have > seen before :-)) > |
|
From: Rasmus A. <ra...@vr...> - 2002-11-28 09:07:36
|
Hi,
On Wed, 2002-11-27 at 15:52, Marcus Roth wrote:
> Hi,
>
> I'm able to view your wrl file with testClusterClient and
> testClusterServer.
Hmmm, that's strange. Of course I have modified the client... but I have
the same problem with the testClusterClient/Server from OpenSG :-(
> How old is your OpenSG-Build?
I updated my CVS version on tuesday this week. So, pretty new!
> Did the server crash or the client?
One of the servers crashes. Then the client exits.
> could you start server and client with gdb?
Yes... the testClusterServer segfaults and here is the gdb output:
Program received signal SIGSEGV, Segmentation fault.
0x40477dc9 in osg::Node::getToWorld(osg::TransformationMatrix<float>&) (
this=0x81621e8, result=@0xbfffe860)
at
/pack-linux/opensg-cvs/src/opensg-cvs/Source/System/FieldContainer/OSGNode.cpp:391
391 getCore()->accumulateMatrix(result);
Current language: auto; currently c++
/Rasmus
--
Rasmus Agerholm
Research Assistant
VR Media Lab (+45) 9635 8792
Niels Jernes Vej 14
DK-9220 Aalborg
Phone: (+45) 9635 8784
Fax: (+45) 9815 2444
http://www.vrmedialab.dk/~rasta
|
|
From: Marcus R. <mar...@ig...> - 2002-11-28 12:17:29
|
Hi, I have no idea, what's going wrong. When the server crashes in getCore()->... it looks like the client as called node->setCore without calling beginEdit, endEdit for node. If you are using the testClusterClient, this looks like a bug in the loader. But as it works fine with my OpenSG, I don't know to search. Does the server and the client use the same OpenSG-Version? Marcus Rasmus Agerholm wrote: > Hi, > > > On Wed, 2002-11-27 at 15:52, Marcus Roth wrote: > >>Hi, >> >>I'm able to view your wrl file with testClusterClient and >>testClusterServer. > > > Hmmm, that's strange. Of course I have modified the client... but I have > the same problem with the testClusterClient/Server from OpenSG :-( > > > >>How old is your OpenSG-Build? > > > I updated my CVS version on tuesday this week. So, pretty new! > > > >>Did the server crash or the client? > > > One of the servers crashes. Then the client exits. > > > >>could you start server and client with gdb? > > > Yes... the testClusterServer segfaults and here is the gdb output: > > Program received signal SIGSEGV, Segmentation fault. > 0x40477dc9 in osg::Node::getToWorld(osg::TransformationMatrix<float>&) ( > this=0x81621e8, result=@0xbfffe860) > at > /pack-linux/opensg-cvs/src/opensg-cvs/Source/System/FieldContainer/OSGNode.cpp:391 > 391 getCore()->accumulateMatrix(result); > Current language: auto; currently c++ > > > /Rasmus > |
|
From: Rasmus A. <ra...@vr...> - 2002-11-28 13:03:47
|
Hi, On Thu, 2002-11-28 at 13:17, Marcus Roth wrote: > Hi, > > I have no idea, what's going wrong. When the server crashes in > getCore()->... it looks like the client as called node->setCore without > calling beginEdit, endEdit for node. If you are using the > testClusterClient, this looks like a bug in the loader. But as it works > fine with my OpenSG, I don't know to search. I am using Gerrits exprimental VRML loader. If you are using the same VRML loader, it is really strange. Are you? > Does the server and the client use the same OpenSG-Version? Yes, I just tried to recompile them both and the result is the same :-( Thanks /Rasmus -- Rasmus Agerholm Research Assistant VR Media Lab (+45) 9635 8792 Niels Jernes Vej 14 DK-9220 Aalborg Phone: (+45) 9635 8784 Fax: (+45) 9815 2444 http://www.vrmedialab.dk/~rasta |
|
From: Marcus R. <mar...@ig...> - 2002-11-28 14:35:19
|
Hi,
Rasmus Agerholm wrote:
> Hi,
>
> On Thu, 2002-11-28 at 13:17, Marcus Roth wrote:
>
>>Hi,
>>
>>I have no idea, what's going wrong. When the server crashes in
>>getCore()->... it looks like the client as called node->setCore without
>>calling beginEdit, endEdit for node. If you are using the
>>testClusterClient, this looks like a bug in the loader. But as it works
>>fine with my OpenSG, I don't know to search.
>
>
> I am using Gerrits exprimental VRML loader. If you are using the same
> VRML loader, it is really strange. Are you?
>
No, I'm not using Gerrits VRML-Loader. Do you need the experimental version?
OSGVRMLDirectionalLight.cpp
VRMLDirectionalLightBinder::finish(VRMLToOSGAction *pAction)
beginEditCP(pBeaconNode, Node::CoreFieldMask);
{
pBeaconNode->setCore(pBeaconCore);
}
endEditCP (pBeaconCore, Node::CoreFieldMask);
I found one bug. I dont't whant to change gerrits experimental code.
Could you fix this in your code? the last endEditCP is wrong
endEditCP (pBeaconNode, Node::CoreFieldMask);
>
>
>>Does the server and the client use the same OpenSG-Version?
>
>
> Yes, I just tried to recompile them both and the result is the same :-(
>
>
> Thanks
> /Rasmus
>
Marcus
|
|
From: Rasmus A. <ra...@vr...> - 2002-11-29 08:58:37
|
Hi, Marcus
On Thu, 2002-11-28 at 15:35, Marcus Roth wrote:
> No, I'm not using Gerrits VRML-Loader. Do you need the experimental version?
>
> OSGVRMLDirectionalLight.cpp
>
> VRMLDirectionalLightBinder::finish(VRMLToOSGAction *pAction)
>
> beginEditCP(pBeaconNode, Node::CoreFieldMask);
> {
> pBeaconNode->setCore(pBeaconCore);
> }
> endEditCP (pBeaconCore, Node::CoreFieldMask);
>
> I found one bug. I dont't whant to change gerrits experimental code.
> Could you fix this in your code? the last endEditCP is wrong
>
> endEditCP (pBeaconNode, Node::CoreFieldMask);
Thanks, that helped. Now I don't experience a crash :-)
So now that the model doesn't crash the app I noticed another strange
thing: Only the servers have textures. On the client textured surfaces
are just the material color. That seems pretty strange since it is the
client that loads the model?
/Rasmus
--
Rasmus Agerholm
Research Assistant
VR Media Lab (+45) 9635 8792
Niels Jernes Vej 14
DK-9220 Aalborg
Phone: (+45) 9635 8784
Fax: (+45) 9815 2444
http://www.vrmedialab.dk/~rasta
|
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X