[vtkusers] VTK 7.0.0.RC1 (and releases before): X-OpenGL depth test(?) problems when using QVTKRenderWindow

David Gobbi david.gobbi at gmail.com
Fri Jan 15 17:01:04 EST 2016


Hi Sophonet,

You can try using the vtkRenderWindow to get info from the context, to see
if it's different with QVTKRenderWindowInteractor than without.

There are two boolean methods for vtkRenderWindow:
   window.SupportsOpenGL()
   window.IsDirect()

And also this method that prints a bunch of info:
   print window.ReportCapabilities()

You might have to all Render() once before calling these methods.

 - David


On Fri, Jan 15, 2016 at 2:55 AM, Sophonet <vtk12af6bc42 at kant.sophonet.de>
wrote:

> Hi David,
>
> The problem occurs as well when running QVTKRenderWindowInteractor.py
> directly.
>
> I have tried picking the Mesa software renderer to exclude a graphics
> driver problem by
>
> LIBGL_ALWAYS_SOFTWARE=1 GALLIUM_DRIVER=softpipe python
> PATHTO_/QVTKRenderWindowInteractor.py, but same problem.
>
> The software renderer is chosen by setting the two variables as I could
> verify with
>
> LIBGL_ALWAYS_SOFTWARE=1 GALLIUM_DRIVER=softpipe glxinfo:
>
> ...
> OpenGL renderer string: Gallium 0.4 on softpipe
> OpenGL version string: 2.1 Mesa 9.2.3
> OpenGL shading language version string: 1.30
> ...
>
> I have found a related thread here (8 years old):
>
> http://osdir.com/ml/python-enthought-devel/2008-04/msg00491.html
>
> ...but as opposed to the problem described there, I do not see warnings on
> a missing depth buffer when turning on MESA_DEBUG=1 - only the errors with
> respect to glDrawBuffer / glReadBuffer (see my previous e-mail)... and in
> order to rule out device driver problems, I have forced software rendering
> as mentioned above.
>
> I would be happy to debug further if I would know how.
>
> Thanks,
>
>        sophonet
>
>
> Am 2016-01-14 21:39, schrieb David Gobbi:
>
>> Hi Sophonet,
>>
>> QVTKRenderWindowInteractor.py has a "main" and can be run as an
>> example.  If you run it, does it show the same problem?
>>
>> I recently tried it with VTK 7.0.0.rc1 and didn't see any depth buffer
>> issues (using Ubuntu 12.04).
>>
>>  - David
>>
>> On Thu, Jan 14, 2016 at 1:17 PM, Sophonet
>> <vtk12af6bc42 at kant.sophonet.de> wrote:
>>
>> Dear list,
>>>
>>> when trying to run a Qt Python application using the
>>> qt.QTVTKRenderWindow, there are depth test problems.
>>>
>>> Examples/Tutorial/Step1/Python/Cone.py works perfect, but with a
>>> simple QVTKRenderWindow example (based on
>>> https://gist.github.com/samueljohn/3090097 [1], needs two more
>>> imports though), there are depth test problems, i.e. it appears that
>>> for some reason no depth test is applied (so the last drawn
>>> primitive is always visible on top).
>>>
>>> When doing
>>>
>>> export MESA_DEBUG=1
>>>
>>> before running the two scripts, there is one warning in the
>>> Examples/Tutorial/Step1/Python/Cone.py,
>>>
>>> Mesa warning: couldn't open libtxc_dxtn.so, software DXTn
>>> compression/decompression unavailable
>>>
>>> [but everything looks fine],
>>>
>>> but many more with the QVTKRenderWindow
>>>
>>> Mesa: User error: GL_INVALID_OPERATION in
>>> glDrawBuffer(buffer=0x405)
>>> Mesa: User error: GL_INVALID_OPERATION in
>>> glReadBuffer(buffer=0x405)
>>> ...
>>>
>>> I am running Mesa 9.2.3 on my system (OpenSUSE 13.1), so I cannot
>>> switch to the OpenGL2 rendering backend (which might resolve the
>>> problem? - don't know...).
>>>
>>> Any hints?
>>>
>>> Thanks,
>>>
>>>      sophonet
>>>
>>
>>
>> Links:
>> ------
>> [1] https://gist.github.com/samueljohn/3090097
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtkusers/attachments/20160115/fa558b77/attachment.html>


More information about the vtkusers mailing list