[vtkusers] reproducible bug: NVidia 1.0-2802 + XFree86 4.1 + VTK 3.2

Charl P. Botha c.p.botha at its.tudelft.nl
Tue Mar 19 05:40:50 EST 2002


Dear Sir/Madam,

With the NVidia 1.0-2802 drivers and XFree86 4.1 there's a reproducible
problem with VTK (http://www.kitware.com/vtk/index.html), cvs checkout from
3.2 tree.  No new developments are performed on the 3.2 tree and this
software is considered to be near bug-free.

To reproduce, build VTK and then build Cone3.cxx in vtk/graphics/examplesCxx
by doing "make Cone3".  Running Cone3 results in a window with the cone
rendered, but as soon as one interacts with the rendering the cone
disappears.  By moving the window, then switching to wireframe and back to
surface mode ('w' and then 's') the cone returns and interaction continues
normally.  We could also tar up the whole build tree as well as included
binaries and make this available for you to test on.  This should work on a
GLIBC 2.2 system.

With the Mace.cxx example, it gets more interesting.  This makes use of two
actors, a sphere and a collection of cones on the sphere.  It seems that the
first actor (the sphere) is just not rendered (but the second one is) until
one does the move-wireframe-surface trick.

All of this works fine with the 1.0.2314 and previous drivers.

As you know, your boards and linux drivers are quite popular in the graphics
and visualisation communities and are often used in conjunction with VTK.
Your assistance in investigating this would be greatly appreciated.

If you need more information in tracking down this problem, please don't
hesitate to contact either Robert Belleman (robbel at science.uva.nl) or Charl
Botha (c.p.botha at its.tudelft.nl).  We could also tar up the whole VTK build 
tree as well as included binaries and make this available for you to test
on. This should work on a GLIBC 2.2 system.

Regards,
Charl P. Botha
Data Visualisation Group, Delft Technical University, The Netherlands
http://visualisation.tudelft.nl/

Robert Belleman
Section Computational Science, University of Amsterdam, The Netherlands
http://www.science.uva.nl/~robbel/

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/



More information about the vtkusers mailing list