Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Sébastien Ramage]

Maybe it can help...

sram...@japc17:~/Bureau/pythonOCC-0.4-samples/Level1/Animation$ 
LIBGL_DEBUG=verbose python animation.py

Display3d class initialization starting ...
libGL: OpenDriver: trying /usr/lib/dri/tls/i965_dri.so
libGL: OpenDriver: trying /usr/lib/dri/i965_dri.so
Graphic device created.
Xw_Window created.
Viewer created.
libGL: Can't open configuration file /etc/drirc: Aucun fichier ou 
dossier de ce type.

Erreur de segmentation



Le 02/11/2010 09:45, Sébastien Ramage a écrit :

Hi Thomas,

I have try your patch, but it doesn't change anything...



Hi all,

I'm not sure whether this Intel problem comes from OCC or pythonOCC. I
suspect a part of the initialization code of pythonOCC 0.4 (file
Display3d.cpp) to be not really optimized. After diving into the Salome
source code, I recently committed changes to this file (
http://code.google.com/p/pythonocc/source/diff?spec=svn1055&r=1039&format=side&path=/trunk/src/wrapper/Visualization/Display3d.cpp&old_path=/trunk/src/wrapper/Visualization/Display3d.cpp&old=1025 
) 


and moved the lights initialization (and made them optional) to the
OCCViewer.py script (
http://code.google.com/p/pythonocc/source/diff?spec=svn1055&r=1039&format=side&path=/trunk/src/addons/Display/OCCViewer.py&old_path=/trunk/src/addons/Display/OCCViewer.py&old=909 
 


).

I guess it may fix this issue but it has to be tested on Linux platforms
with Intel graphic driver.

Best Regards,




___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Thomas Paviot]

Hi guys,

Many thanks for this interesting discussion and your feedbacks. I still have
difficulties to check whether the display issues come from OCC or pythonOCC.
It's hard to compare the behaviour of pythonOCC and Salome since Salome
relies on a patched release of OpenCascade: I checked the OCC6.3Sp8 source
code, and it appears that there are a few fixes in the OpenGl code, but I
didn't diff the files. Maybe these changes will be included in the upcoming
6.4 OCC release.

I thought the code I recently committed to the trunk would solve the display
issue and I'm a bit confused with the tweak related to the X server
configuration. Anyway, I do not have any machine to drive any tests with the
Intel graphic driver, so I trust your experiments. So just let me know if I
can be of any help.

Best Regards,

Thomas

2010/7/15 Oliver Borm 

>  Hello Marcos,
>
> well now it is also running at my laptop. I've added the following lines to
> the /etc/X11/xorg.conf (if this file is not present, one has to create it):
>
> Section "Device"
> Identifier  "Configured Video Device"
> Driver  "intel"
> Option  "DRI" "off"
> EndSection
>
>
> After that I have to restart my X-Server, in my case with kdm running, with
> the following command:
>
> sudo service kdm restart
>
> and then I can import the python-occ simple GUI. More information can be
> found at:
> https://wiki.ubuntu.com/X/Troubleshooting/IntelPerformance
> https://wiki.kubuntu.org/X/Bugs/Lucidi8xxFreezes
>
> The debian package was done in a native environment. If you build a deb
> package of python-occ from the svn sources you are free to contribute to the
> launchpad platform.
>
> Best regards,
> Oliver
>
>   Hi Oliver,
>
> I will explain you what I have done:
>
> 1. I Installed the Ubuntu package that you contribute to python-occ.
> 2. I got the "segmentation fault".
> 3. I deactivated the graphic acceleration.
> 4. I ran the HelloWorld.py example and works.
>
> Because of necessary deactivation of graphic acceleration, I compilate
> python-occ from svn (including mesh and geom modules).
>
> Previously I installed Salome and works with out graphic problem
> (open-cascade of course was installed from Ubuntu10.04 packages)
>
> So, because of Salome installation I redefine the CASROOT to opencascade
> libs in to Salome installation folder.
>
> I would like to contribute with python-occ and Ubuntu users developing an
> Ubuntu package in the native platform. And said good bye to segmentation
> fault.
>
> I will appreciate if you would like to help me to do so.
>
> ¿Did you develop the python-occ package in native Ubuntu 10.04 environment
> or from virtual machine running over different native environment?
>
> Best regards,
>
> Marcos.
>
>
>
>
> ___
> Pythonocc-users mailing 
> listpythonocc-us...@gna.orghttps://mail.gna.org/listinfo/pythonocc-users
>
>
> ___
> Pythonocc-users mailing list
> Pythonocc-users@gna.org
> https://mail.gna.org/listinfo/pythonocc-users
>
>
___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Oliver Borm]

Hello Marcos,

well now it is also running at my laptop. I've added the following lines
to the /etc/X11/xorg.conf (if this file is not present, one has to
create it):

Section "Device"
Identifier  "Configured Video Device"
Driver  "intel"
Option  "DRI" "off"
EndSection


After that I have to restart my X-Server, in my case with kdm running,
with the following command:

sudo service kdm restart

and then I can import the python-occ simple GUI. More information can be
found at:
https://wiki.ubuntu.com/X/Troubleshooting/IntelPerformance
https://wiki.kubuntu.org/X/Bugs/Lucidi8xxFreezes

The debian package was done in a native environment. If you build a deb
package of python-occ from the svn sources you are free to contribute to
the launchpad platform.

Best regards,
Oliver

>  Hi Oliver,
>
> I will explain you what I have done:
>
> 1. I Installed the Ubuntu package that you contribute to python-occ.
> 2. I got the "segmentation fault".
> 3. I deactivated the graphic acceleration.
> 4. I ran the HelloWorld.py example and works.
>
> Because of necessary deactivation of graphic acceleration, I compilate
> python-occ from svn (including mesh and geom modules).
>
> Previously I installed Salome and works with out graphic problem
> (open-cascade of course was installed from Ubuntu10.04 packages)
>
> So, because of Salome installation I redefine the CASROOT to
> opencascade libs in to Salome installation folder.
>
> I would like to contribute with python-occ and Ubuntu users developing
> an Ubuntu package in the native platform. And said good bye to
> segmentation fault.
>
> I will appreciate if you would like to help me to do so.
>
> ¿Did you develop the python-occ package in native Ubuntu 10.04
> environment or from virtual machine running over different native
> environment?
>
> Best regards,
>
> Marcos.
>
>  
>
>
> ___
> Pythonocc-users mailing list
> Pythonocc-users@gna.org
> https://mail.gna.org/listinfo/pythonocc-users
>   
___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Oliver Borm]

Hello Marcos,

and you got the python-occ GUI working from that svn compilation, when
you restart the X-Server before every use?

Well I tried your supposed solution, but I have had no success. To which
"hello.py" script are you referring in your description? If I want to
start the python-occ GUI in console mode I come to the following error:

Error: couldn't connect to X Server

Which is absolutely clear, because no X-Server is running. And if I
neglect this magic "hello.py" I just end up in  restarting the X-Server
and that also solves not my problem. Maybe you have some more
information for me ... ?

I tested the python-occ also on the same laptop with the Intel graphic
card but with a gentoo Linux installation, and it just worked. So I
suppose it is just a very specific ubuntu (10.04) problem, which may be
solved with the right setting.

Thanks,
Oliver

> Hi Oliver,
>
> I suppose that work in virtual machine with Ubuntu 10.04, but it is
> not the idea.
>
> In native environment my partner indicate me that is necessary to stop
> the X-Server
>
> /*sudo service gdm stop (caution: your x-server will stop and terminal
> will appear) */
> /*
> ALT + F1 (or F2, F3 ...) to change between different terminals.
> *
> /insert your user name and pass.
>
> and execute the hello.py sample.
>
> /*sudo service gdm start (To x-server start again)*/
>  
> Regards,
>
> Marcos.
>
> P.S.: with support of my partner that have more experience with Linux,
> I installed python-occ from (svn) source code in Ubuntu 10.04.
>
>
> ___
> Pythonocc-users mailing list
> Pythonocc-users@gna.org
> https://mail.gna.org/listinfo/pythonocc-users
>   
___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Oliver Borm]

Hi Marcos,

thank's for your input. Can you exactly explain how do you have
deactivated the graphic acceleration? I installed the driconf utility,
and after deactivating the 3D acceleration at the tab Debugging, even
the glxgears app crashed and python-occ was also not able to start. So I
was not yet able to get this workaround to work.

Was this Ubuntu 10.04 only running in a virtual machine or does this run
also in a native environment?

Oliver

Am 12.07.2010 08:08, schrieb Marcos Elgueta Soulat:
> Oliver,
>
> With a friend we already detect the problem.
>
> We start under hypothesis that was compiled on a Virtual Machine, but
> by default all virtual machine have unable the "Graphic Acceleration".
> Due to, we deactivate the Graphic Acceleration on my laptop. After
> that python-occ GUI on a laptop with Intel graphic card and Ubuntu
> 10.04, starting to work!.
>
> Regards,
>
> Marcos.

___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Marcos Elgueta Soulat]

Oliver,

With a friend we already detect the problem.

We start under hypothesis that was compiled on a Virtual Machine, but by
default all virtual machine have unable the "Graphic Acceleration". Due to,
we deactivate the Graphic Acceleration on my laptop. After that python-occ
GUI on a laptop with Intel graphic card and Ubuntu 10.04, starting to work!.

Regards,

Marcos.

On 8 July 2010 08:13, Marcos  wrote:

>
> El 08-07-2010, a las 6:00, pythonocc-users-requ...@gna.org
>
>> Message: 2
>> Date: Wed, 07 Jul 2010 19:08:29 +0200
>> From: Oliver Borm 
>> Subject: [Pythonocc-users] python-occ GUI segmentation faults on intel
>>   graphic cards
>> To: pythonocc-users@gna.org
>> Message-ID: <4c34b48d.4010...@web.de>
>> Content-Type: text/plain; charset="utf-8"
>>
>> Hello,
>>
>> at the moment I have problems to start the python-occ GUI on a laptop
>> with Intel graphic card, as I get a segmentation fault. Attached you
>> will find my small test file and the gdb backtrace.
>>
>> I'm using ubuntu 10.04 and the paython-occ 0.4 debian package from the
>> CAE-Team PPA. Any suggestions how to get it working on intel graphic
>> cards? On nvidia and SiS graphic cards it is working.
>>
>> Thanks,
>> Oliver
>>
>
> Replay to this question will be useful for me also.
>
> Thanks,
> Marcos.
>
>  -- next part --
>> An embedded and charset-unspecified text was scrubbed...
>> Name: python-occ_backtrace.txt
>> Url: /public/pythonocc-users/attachments/20100707/6afc3c96/attachment.txt
>> -- next part --
>> A non-text attachment was scrubbed...
>> Name: test.py
>> Type: text/x-python
>> Size: 186 bytes
>> Desc: not available
>> Url : /public/pythonocc-users/attachments/20100707/6afc3c96/attachment.py
>>
>> --
>>
>> ___
>> Pythonocc-users mailing list
>> Pythonocc-users@gna.org
>> https://mail.gna.org/listinfo/pythonocc-users
>>
>>
>> End of Pythonocc-users Digest, Vol 20, Issue 12
>> ***
>>
>


-- 
Marcos A. Elgueta Soulat
Student of Aerospace Engineering,
Department of Mechanics Engineering,
University of Concepción, Chile.

mail: m.a.elguetasou...@gmail.com
mobile(1): +56 9 82208407
mobile(2): +56 9 83036927
skype: marcos.elgueta.soulat
___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Oliver Borm]

Hello,

yes glxgears is running without any problems. I've also tested the
following variable:

export LIBGL_ALWAYS_INDIRECT=yes

but that let's my whole desktop crashing, when starting the test python
script.

Best,
Oliver


Am 09.07.2010 10:23, schrieb M. Nawijn:
> Hello All,
>
> I have had some problems previously with getting the graphics working
> properly (with
> NVidia card),
> one of the reasons was that after a package update, some of the symbolic links
> that link to the proper libGL* classes were set wrong.
>
> Before diving into the problem, can you confirm that the glxgears
> utility found on most
> of the linux systems is running without problems? If not, it is not
> directly an OCC problem.
>
> Regards,
>
> Marco
>   


___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Thomas Paviot]

Hi all,

I'm not sure whether this Intel problem comes from OCC or pythonOCC. I
suspect a part of the initialization code of pythonOCC 0.4 (file
Display3d.cpp) to be not really optimized. After diving into the Salome
source code, I recently committed changes to this file (
http://code.google.com/p/pythonocc/source/diff?spec=svn1055&r=1039&format=side&path=/trunk/src/wrapper/Visualization/Display3d.cpp&old_path=/trunk/src/wrapper/Visualization/Display3d.cpp&old=1025)
and moved the lights initialization (and made them optional) to the
OCCViewer.py script (
http://code.google.com/p/pythonocc/source/diff?spec=svn1055&r=1039&format=side&path=/trunk/src/addons/Display/OCCViewer.py&old_path=/trunk/src/addons/Display/OCCViewer.py&old=909
).

I guess it may fix this issue but it has to be tested on Linux platforms
with Intel graphic driver.

Best Regards,

Thomas

2010/7/9 M. Nawijn 

> Hello All,
>
> I have had some problems previously with getting the graphics working
> properly (with
> NVidia card),
> one of the reasons was that after a package update, some of the symbolic
> links
> that link to the proper libGL* classes were set wrong.
>
> Before diving into the problem, can you confirm that the glxgears
> utility found on most
> of the linux systems is running without problems? If not, it is not
> directly an OCC problem.
>
> Regards,
>
> Marco
>
> On Thu, Jul 8, 2010 at 3:28 PM, Oliver Borm  wrote:
> > Just checked the OCC forum, but the only comment für WinXP was to use
> > the mesa libraries. Unfortunately I'm using these libraries already with
> > Intel graphic cards under kubuntu. Checked the linked libs of
> > libTKOpenGl-6.3.0.so (as I thought this is the one which makes trouble -
> > am I right?)
> >
> > ~$ ldd /usr/lib/libTKOpenGl-6.3.0.so
> >linux-gate.so.1 =>  (0x00f4c000)
> >libTKService-6.3.0.so => /usr/lib/libTKService-6.3.0.so(0x003d2000)
> >libTKernel-6.3.0.so => /usr/lib/libTKernel-6.3.0.so (0x0058a000)
> >libTKV3d-6.3.0.so => /usr/lib/libTKV3d-6.3.0.so (0x00afa000)
> >libXt.so.6 => /usr/lib/libXt.so.6 (0x0011)
> >libX11.so.6 => /usr/lib/libX11.so.6 (0x00163000)
> >libGLU.so.1 => /usr/lib/libGLU.so.1 (0x0028)
> >libGL.so.1 => /usr/lib/mesa/libGL.so.1 (0x00327000)
> >libpthread.so.0 => /lib/tls/i686/cmov/libpthread.so.0 (0x002f1000)
> >libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x007c5000)
> >libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0x0038c000)
> >libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0x00f4d000)
> >libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x003b2000)
> >libTKMath-6.3.0.so => /usr/lib/libTKMath-6.3.0.so (0x11fcd000)
> >libXmu.so.6 => /usr/lib/libXmu.so.6 (0x008bb000)
> >libXext.so.6 => /usr/lib/libXext.so.6 (0x00ee)
> >libdl.so.2 => /lib/tls/i686/cmov/libdl.so.2 (0x00a28000)
> >libTKG3d-6.3.0.so => /usr/lib/libTKG3d-6.3.0.so (0x041c7000)
> >libTKTopAlgo-6.3.0.so => /usr/lib/libTKTopAlgo-6.3.0.so(0x083fe000)
> >libTKGeomBase-6.3.0.so => /usr/lib/libTKGeomBase-6.3.0.so
> > (0x17b1)
> >libTKMesh-6.3.0.so => /usr/lib/libTKMesh-6.3.0.so (0x00a2c000)
> >libTKGeomAlgo-6.3.0.so => /usr/lib/libTKGeomAlgo-6.3.0.so
> > (0x0905)
> >libTKHLR-6.3.0.so => /usr/lib/libTKHLR-6.3.0.so (0x0fa09000)
> >libTKBRep-6.3.0.so => /usr/lib/libTKBRep-6.3.0.so (0x03432000)
> >libTKG2d-6.3.0.so => /usr/lib/libTKG2d-6.3.0.so (0x009ae000)
> >libTKV2d-6.3.0.so => /usr/lib/libTKV2d-6.3.0.so (0x0244f000)
> >libSM.so.6 => /usr/lib/libSM.so.6 (0x008d2000)
> >libICE.so.6 => /usr/lib/libICE.so.6 (0x008db000)
> >libxcb.so.1 => /usr/lib/libxcb.so.1 (0x008f4000)
> >libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0x0090e000)
> >libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0x00914000)
> >libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0x00918000)
> >libdrm.so.2 => /lib/libdrm.so.2 (0x0091e000)
> >/lib/ld-linux.so.2 (0x0030a000)
> >libuuid.so.1 => /lib/libuuid.so.1 (0x00929000)
> >libXau.so.6 => /usr/lib/libXau.so.6 (0x0092e000)
> >libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00a1a000)
> >librt.so.1 => /lib/tls/i686/cmov/librt.so.1 (0x00f1c000)
> >
> > And the interesting part is:
> >
> > /usr/lib/dri/i965_dri.sothis is the lib which seg faults and part of
> > libgl1-mesa-dri
> > /usr/lib/mesa/libGL.so.1 used lib is part of libgl1-mesa-glx
> > /usr/lib/libGLU.so.1 used lib is part of libglu1-mesa
> >
> > Well if it is a OCC problem, does anybody has a small and simple test
> > binary or know where to get one?
> >
> > Best,
> > Oliver
> >
> >
> > Am 08.07.2010 14:17, schrieb Jelle Feringa:
> >> Interesting indeed.
> >> I think this is a OCC, not so much =Python=OCC problem.
> >> Have you checked the OCC forum?
> >>
> >> -jelle
> >>
> >> On Jul 8, 2010, at 2:13 PM, Marcos wrote:
> >>
> >>
> >
> >
> >
> > _

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[M. Nawijn]

Hello All,

I have had some problems previously with getting the graphics working
properly (with
NVidia card),
one of the reasons was that after a package update, some of the symbolic links
that link to the proper libGL* classes were set wrong.

Before diving into the problem, can you confirm that the glxgears
utility found on most
of the linux systems is running without problems? If not, it is not
directly an OCC problem.

Regards,

Marco

On Thu, Jul 8, 2010 at 3:28 PM, Oliver Borm  wrote:
> Just checked the OCC forum, but the only comment für WinXP was to use
> the mesa libraries. Unfortunately I'm using these libraries already with
> Intel graphic cards under kubuntu. Checked the linked libs of
> libTKOpenGl-6.3.0.so (as I thought this is the one which makes trouble -
> am I right?)
>
> ~$ ldd /usr/lib/libTKOpenGl-6.3.0.so
>        linux-gate.so.1 =>  (0x00f4c000)
>        libTKService-6.3.0.so => /usr/lib/libTKService-6.3.0.so (0x003d2000)
>        libTKernel-6.3.0.so => /usr/lib/libTKernel-6.3.0.so (0x0058a000)
>        libTKV3d-6.3.0.so => /usr/lib/libTKV3d-6.3.0.so (0x00afa000)
>        libXt.so.6 => /usr/lib/libXt.so.6 (0x0011)
>        libX11.so.6 => /usr/lib/libX11.so.6 (0x00163000)
>        libGLU.so.1 => /usr/lib/libGLU.so.1 (0x0028)
>        libGL.so.1 => /usr/lib/mesa/libGL.so.1 (0x00327000)
>        libpthread.so.0 => /lib/tls/i686/cmov/libpthread.so.0 (0x002f1000)
>        libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x007c5000)
>        libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0x0038c000)
>        libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0x00f4d000)
>        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x003b2000)
>        libTKMath-6.3.0.so => /usr/lib/libTKMath-6.3.0.so (0x11fcd000)
>        libXmu.so.6 => /usr/lib/libXmu.so.6 (0x008bb000)
>        libXext.so.6 => /usr/lib/libXext.so.6 (0x00ee)
>        libdl.so.2 => /lib/tls/i686/cmov/libdl.so.2 (0x00a28000)
>        libTKG3d-6.3.0.so => /usr/lib/libTKG3d-6.3.0.so (0x041c7000)
>        libTKTopAlgo-6.3.0.so => /usr/lib/libTKTopAlgo-6.3.0.so (0x083fe000)
>        libTKGeomBase-6.3.0.so => /usr/lib/libTKGeomBase-6.3.0.so
> (0x17b1)
>        libTKMesh-6.3.0.so => /usr/lib/libTKMesh-6.3.0.so (0x00a2c000)
>        libTKGeomAlgo-6.3.0.so => /usr/lib/libTKGeomAlgo-6.3.0.so
> (0x0905)
>        libTKHLR-6.3.0.so => /usr/lib/libTKHLR-6.3.0.so (0x0fa09000)
>        libTKBRep-6.3.0.so => /usr/lib/libTKBRep-6.3.0.so (0x03432000)
>        libTKG2d-6.3.0.so => /usr/lib/libTKG2d-6.3.0.so (0x009ae000)
>        libTKV2d-6.3.0.so => /usr/lib/libTKV2d-6.3.0.so (0x0244f000)
>        libSM.so.6 => /usr/lib/libSM.so.6 (0x008d2000)
>        libICE.so.6 => /usr/lib/libICE.so.6 (0x008db000)
>        libxcb.so.1 => /usr/lib/libxcb.so.1 (0x008f4000)
>        libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0x0090e000)
>        libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0x00914000)
>        libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0x00918000)
>        libdrm.so.2 => /lib/libdrm.so.2 (0x0091e000)
>        /lib/ld-linux.so.2 (0x0030a000)
>        libuuid.so.1 => /lib/libuuid.so.1 (0x00929000)
>        libXau.so.6 => /usr/lib/libXau.so.6 (0x0092e000)
>        libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00a1a000)
>        librt.so.1 => /lib/tls/i686/cmov/librt.so.1 (0x00f1c000)
>
> And the interesting part is:
>
> /usr/lib/dri/i965_dri.so    this is the lib which seg faults and part of
> libgl1-mesa-dri
> /usr/lib/mesa/libGL.so.1 used lib is part of libgl1-mesa-glx
> /usr/lib/libGLU.so.1         used lib is part of libglu1-mesa
>
> Well if it is a OCC problem, does anybody has a small and simple test
> binary or know where to get one?
>
> Best,
> Oliver
>
>
> Am 08.07.2010 14:17, schrieb Jelle Feringa:
>> Interesting indeed.
>> I think this is a OCC, not so much =Python=OCC problem.
>> Have you checked the OCC forum?
>>
>> -jelle
>>
>> On Jul 8, 2010, at 2:13 PM, Marcos wrote:
>>
>>
>
>
>
> ___
> Pythonocc-users mailing list
> Pythonocc-users@gna.org
> https://mail.gna.org/listinfo/pythonocc-users
>

___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Oliver Borm]

Just checked the OCC forum, but the only comment für WinXP was to use
the mesa libraries. Unfortunately I'm using these libraries already with
Intel graphic cards under kubuntu. Checked the linked libs of
libTKOpenGl-6.3.0.so (as I thought this is the one which makes trouble -
am I right?)

~$ ldd /usr/lib/libTKOpenGl-6.3.0.so
linux-gate.so.1 =>  (0x00f4c000)
libTKService-6.3.0.so => /usr/lib/libTKService-6.3.0.so (0x003d2000)
libTKernel-6.3.0.so => /usr/lib/libTKernel-6.3.0.so (0x0058a000)
libTKV3d-6.3.0.so => /usr/lib/libTKV3d-6.3.0.so (0x00afa000)
libXt.so.6 => /usr/lib/libXt.so.6 (0x0011)
libX11.so.6 => /usr/lib/libX11.so.6 (0x00163000)
libGLU.so.1 => /usr/lib/libGLU.so.1 (0x0028)
libGL.so.1 => /usr/lib/mesa/libGL.so.1 (0x00327000)
libpthread.so.0 => /lib/tls/i686/cmov/libpthread.so.0 (0x002f1000)
libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x007c5000)
libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0x0038c000)
libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0x00f4d000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x003b2000)
libTKMath-6.3.0.so => /usr/lib/libTKMath-6.3.0.so (0x11fcd000)
libXmu.so.6 => /usr/lib/libXmu.so.6 (0x008bb000)
libXext.so.6 => /usr/lib/libXext.so.6 (0x00ee)
libdl.so.2 => /lib/tls/i686/cmov/libdl.so.2 (0x00a28000)
libTKG3d-6.3.0.so => /usr/lib/libTKG3d-6.3.0.so (0x041c7000)
libTKTopAlgo-6.3.0.so => /usr/lib/libTKTopAlgo-6.3.0.so (0x083fe000)
libTKGeomBase-6.3.0.so => /usr/lib/libTKGeomBase-6.3.0.so
(0x17b1)
libTKMesh-6.3.0.so => /usr/lib/libTKMesh-6.3.0.so (0x00a2c000)
libTKGeomAlgo-6.3.0.so => /usr/lib/libTKGeomAlgo-6.3.0.so
(0x0905)
libTKHLR-6.3.0.so => /usr/lib/libTKHLR-6.3.0.so (0x0fa09000)
libTKBRep-6.3.0.so => /usr/lib/libTKBRep-6.3.0.so (0x03432000)
libTKG2d-6.3.0.so => /usr/lib/libTKG2d-6.3.0.so (0x009ae000)
libTKV2d-6.3.0.so => /usr/lib/libTKV2d-6.3.0.so (0x0244f000)
libSM.so.6 => /usr/lib/libSM.so.6 (0x008d2000)
libICE.so.6 => /usr/lib/libICE.so.6 (0x008db000)
libxcb.so.1 => /usr/lib/libxcb.so.1 (0x008f4000)
libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0x0090e000)
libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0x00914000)
libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0x00918000)
libdrm.so.2 => /lib/libdrm.so.2 (0x0091e000)
/lib/ld-linux.so.2 (0x0030a000)
libuuid.so.1 => /lib/libuuid.so.1 (0x00929000)
libXau.so.6 => /usr/lib/libXau.so.6 (0x0092e000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00a1a000)
librt.so.1 => /lib/tls/i686/cmov/librt.so.1 (0x00f1c000)

And the interesting part is:

/usr/lib/dri/i965_dri.sothis is the lib which seg faults and part of
libgl1-mesa-dri
/usr/lib/mesa/libGL.so.1 used lib is part of libgl1-mesa-glx
/usr/lib/libGLU.so.1 used lib is part of libglu1-mesa

Well if it is a OCC problem, does anybody has a small and simple test
binary or know where to get one?

Best,
Oliver


Am 08.07.2010 14:17, schrieb Jelle Feringa:
> Interesting indeed. 
> I think this is a OCC, not so much =Python=OCC problem.
> Have you checked the OCC forum?
>
> -jelle
>
> On Jul 8, 2010, at 2:13 PM, Marcos wrote:
>
>   



___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users

Re: [Pythonocc-users] python-occ GUI segmentation faults on intel graphic cards

[Jelle Feringa]

Interesting indeed. 
I think this is a OCC, not so much =Python=OCC problem.
Have you checked the OCC forum?

-jelle

On Jul 8, 2010, at 2:13 PM, Marcos wrote:

> 
> El 08-07-2010, a las 6:00, pythonocc-users-requ...@gna.org
>> Message: 2
>> Date: Wed, 07 Jul 2010 19:08:29 +0200
>> From: Oliver Borm 
>> Subject: [Pythonocc-users] python-occ GUI segmentation faults on intel
>>   graphic cards
>> To: pythonocc-users@gna.org
>> Message-ID: <4c34b48d.4010...@web.de>
>> Content-Type: text/plain; charset="utf-8"
>> 
>> Hello,
>> 
>> at the moment I have problems to start the python-occ GUI on a laptop
>> with Intel graphic card, as I get a segmentation fault. Attached you
>> will find my small test file and the gdb backtrace.
>> 
>> I'm using ubuntu 10.04 and the paython-occ 0.4 debian package from the
>> CAE-Team PPA. Any suggestions how to get it working on intel graphic
>> cards? On nvidia and SiS graphic cards it is working.
>> 
>> Thanks,
>> Oliver
> 
> Replay to this question will be useful for me also.
> 
> Thanks,
> Marcos.


___
Pythonocc-users mailing list
Pythonocc-users@gna.org
https://mail.gna.org/listinfo/pythonocc-users