visualpython-users

[Visualpython-users] Segmentation Fault

[Nicolas B. <nbi...@gm...>]

Hi all,

I've ported one of my script from PyMol to vpython and it's pretty neat.
But, I cannot use it on 2 of my machine as vpython is segfaulting. Just
importing visual is enough to segfault. Here is a gdb stacktrace:
http://pastebin.com/6vLzqjfb I couldn't see anything from it...

I tried recompiling visual-5.32_release.tar.bz2 without success.

I'm using ArchLinux x86_64
python 2.6.5-3
python-numpy 1.4.1-1
pkg-config 0.25-2
boost 1.43.0-1
gtkglextmm 1.2.0-4
libglademm 2.6.7-1
boost-libs 1.43.0-1
tk 8.5.8-2

Thanx for any help!

Re: [Visualpython-users] Segmentation Fault

[Nicolas B. <nbi...@gm...>]

Just a followup of my own message... I've recompiled python and visual with
debug symbols. Here is the new stack trace:
http://pastebin.com/zihsxktb

Thanx

2010/8/17 Nicolas Bigaouette <nbi...@gm...>

> Hi all,
>
> I've ported one of my script from PyMol to vpython and it's pretty neat.
> But, I cannot use it on 2 of my machine as vpython is segfaulting. Just
> importing visual is enough to segfault. Here is a gdb stacktrace:
> http://pastebin.com/6vLzqjfb I couldn't see anything from it...
>
> I tried recompiling visual-5.32_release.tar.bz2 without success.
>
> I'm using ArchLinux x86_64
> python 2.6.5-3
> python-numpy 1.4.1-1
> pkg-config 0.25-2
> boost 1.43.0-1
> gtkglextmm 1.2.0-4
> libglademm 2.6.7-1
> boost-libs 1.43.0-1
> tk 8.5.8-2
>
> Thanx for any help!
>

Re: [Visualpython-users] Segmentation Fault

[Bruce S. <bas...@nc...>]

It's certainly not obvious what the problem is, but here are two small ideas:

1) Sometimes problems of this kind stem from a need to update the
graphics drivers. I'm inclined to think that's not your problem,
because you crash just trying to import visual, before creating any
objects.

2) I see that the visual machinery is in
/usr/lib/python2.6/site-packages. I don't know anything about
Arch-Linux, but on Ubuntu there was a recent change involving
dist-packages as opposed to site-packages. Is it conceivable that your
builds have put the newly compiled files somewhere not in
/usr/lib/python2.6/site-packages, but there is something in
/usr/lib/python2.6/site-packages that is corrupted?

Curious that VPython works on one of you machines but not on two others....

Bruce Sherwood

On Tue, Aug 17, 2010 at 12:04 PM, Nicolas Bigaouette
<nbi...@gm...> wrote:
> Just a followup of my own message... I've recompiled python and visual with
> debug symbols. Here is the new stack trace:
> http://pastebin.com/zihsxktb
>
> Thanx
>
> 2010/8/17 Nicolas Bigaouette <nbi...@gm...>
>>
>> Hi all,
>>
>> I've ported one of my script from PyMol to vpython and it's pretty neat.
>> But, I cannot use it on 2 of my machine as vpython is segfaulting. Just
>> importing visual is enough to segfault. Here is a gdb stacktrace:
>> http://pastebin.com/6vLzqjfb I couldn't see anything from it...
>>
>> I tried recompiling visual-5.32_release.tar.bz2 without success.
>>
>> I'm using ArchLinux x86_64
>> python 2.6.5-3
>> python-numpy 1.4.1-1
>> pkg-config 0.25-2
>> boost 1.43.0-1
>> gtkglextmm 1.2.0-4
>> libglademm 2.6.7-1
>> boost-libs 1.43.0-1
>> tk 8.5.8-2
>>
>> Thanx for any help!
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>

Re: [Visualpython-users] Segmentation Fault

[Nicolas B. <nbi...@gm...>]

Thanx for your answer.

It's certainly not obvious what the problem is, but here are two small
> ideas:
>
Indeed...


> 1) Sometimes problems of this kind stem from a need to update the
> graphics drivers. I'm inclined to think that's not your problem,
> because you crash just trying to import visual, before creating any
> objects.
>
My driver is up to date: xf86-video-intel 2.12.0-1


> 2) I see that the visual machinery is in
> /usr/lib/python2.6/site-packages. I don't know anything about
> Arch-Linux, but on Ubuntu there was a recent change involving
> dist-packages as opposed to site-packages. Is it conceivable that your
> builds have put the newly compiled files somewhere not in
> /usr/lib/python2.6/site-packages, but there is something in
> /usr/lib/python2.6/site-packages that is corrupted?
>
VPython is installed in /usr/lib/python2.6/site-packages/visual/. The
package is built by Arch devs with:

> pythondir=`python -c "from distutils import sysconfig; print
> sysconfig.get_python_lib()"`
> unset PYTHONPATH
>   ./configure --prefix=/usr \
>               --with-example-dir=$pythondir/visual/demos
>


Curious that VPython works on one of you machines but not on two others....
>
I just though about it too and the only difference between these machines is
that the failing one have integrated intel card, while the one which is
working has an nvidia. 3D is working fine on the integrated card: I run
KDE/kwin with composition on without any problem. I can even play some games
on these machines...

Re: [Visualpython-users] Segmentation Fault

[Bruce S. <bas...@nc...>]

This certainly sounds like a graphics driver problem. Just because
some 3D aspects work doesn't mean that all is well.

Bruce Sherwood

On Tue, Aug 17, 2010 at 5:09 PM, Nicolas Bigaouette
<nbi...@gm...> wrote:
>> Curious that VPython works on one of you machines but not on two
>> others....
>
> I just though about it too and the only difference between these machines is
> that the failing one have integrated intel card, while the one which is
> working has an nvidia. 3D is working fine on the integrated card: I run
> KDE/kwin with composition on without any problem. I can even play some games
> on these machines...
>

Re: [Visualpython-users] Segmentation Fault

[Nicolas B. <nbi...@gm...>]

2010/8/17 Guenter Schneider <Gue...@ph...>

> I am not surprised that visual runs on nvidia. I never had problems before.
> For ATI cards the following worked for me (note that my problem was never a
> complete crash on import but a seg fault for particular commands, but it may
> work for you).
>
> import wx
> import visual
>
> The rational is that wx has proper initialization for opengl while visual
> does not.
>
> Guenter
>

Wow, ok, so that made it:

> $ python
> Python 2.6.5 (r265:79063, Aug 17 2010, 14:03:21)
> [GCC 4.5.1] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> import wx
> >>> import visual
> >>> visual.sphere()
> i915_program_error: Exceeded max instructions
> <visual.primitives.sphere object at 0x22d47e0>
>
and I see a sphere.

I just tried my script and it seems to work. But it is really slow compared
to pymol with the same data, even on the intel card.. On the nvidia card,
it's blazing fast...

Re: [Visualpython-users] Segmentation Fault

[Samuel P. <sam...@gm...>]

Hi Nicolas,

I do experience the same "too slow" behavior on VPython with my
integrated Intel graphics card. Other OpenGL applications run smoothly. I
also get the weird message:
i915_program_error: Exceeded max instructions
<visual.primitives.sphere object at 0x2314c58>

Have you (or anybody in the list) found a fix for this annoyance? It makes
my VPython applications unusable!

Regards,
Samuel

On Wed, Aug 18, 2010 at 2:34 AM, Nicolas Bigaouette
<nbi...@gm...>wrote:

> 2010/8/17 Guenter Schneider <Gue...@ph...>
>
>> I am not surprised that visual runs on nvidia. I never had problems
>> before. For ATI cards the following worked for me (note that my problem was
>> never a complete crash on import but a seg fault for particular commands,
>> but it may work for you).
>>
>> import wx
>> import visual
>>
>> The rational is that wx has proper initialization for opengl while visual
>> does not.
>>
>> Guenter
>>
>
> Wow, ok, so that made it:
>
>> $ python
>> Python 2.6.5 (r265:79063, Aug 17 2010, 14:03:21)
>> [GCC 4.5.1] on linux2
>> Type "help", "copyright", "credits" or "license" for more information.
>> >>> import wx
>> >>> import visual
>> >>> visual.sphere()
>> i915_program_error: Exceeded max instructions
>> <visual.primitives.sphere object at 0x22d47e0>
>>
> and I see a sphere.
>
> I just tried my script and it seems to work. But it is really slow compared
> to pymol with the same data, even on the intel card.. On the nvidia card,
> it's blazing fast...
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>