This was posted to the OpenSceneGraph mailing list, but I'm not sure if the problem isn't with (my) SuSE implementation. My system is usually tweaked beyond all recognition, so determining if I broke it can be tricky. Could someone else verify that this works, or fails on their system? Here's where to get the goods: Prerequisites: http://openthreads.sourceforge.net/ http://www.andesengineering.com/Producer/download.html OSG: http://openscenegraph.sourceforge.net/ http://openscenegraph.sourceforge.net/downloads/index.html Crash course in configuration, building and installing: I went into the respective Make/makedef's and hardwired the INST_LOCATION = $(WHEREIPUTIT). I believe that was unnecessary, but I was having problems with the OSG install, so I tried to force the issue. I believe I can simply `make INST_LOCATION = $(OSG) install' assuming the environment is configures as shown below. Is that a correct assumption? I also put this in my ~/.bashrc export OPENTHREADS=$ORG/openthreads export OPENTHREADS_INC_DIR=$OPENTHREADS/include export OPENTHREADS_LIB_DIR=$OPENTHREADS/lib export OPENPRODUCER=$COM/andesengeneering/openproducer export PRODUCER_INC_DIR=$OPENPRODUCER/include export PRODUCER_LIB_DIR=$OPENPRODUCER/lib export OSG=$ORG/osg # added to LIBRARY_PATH, LD_LIBRARY_PATH and CPLUS_INCLUDE_PATH /************* Du Problem ******************/ When I run the teapot example from the OSG CVS, the surface has rectangular or parallellogram holes appearing throughout the image. In the xterm I see fallback Q_BIT fallback Q_BIT fallback Q_BIT fallback Q_BIT fallback Q_BIT fallback Q_BIT ... I'm running on SuSE 9.1 latest updates; a P4 with unsupported hyperthreading. This is what 3Ddiag has to say: 3Ddiag version 0.716 Verifying 3D configuration: Using 3dinfo ************************************************************ Verifying 3D configuration for 3D board "ATI Technologies Inc Radeon VE (1002@5159)": Tests for XFree86 configuration: Config File /etc/X11/XF86Config ... done. Driver ... done. Extensions ... done. Options ... done. Checking GL/GLU/glut runtime configuration: GL/GLU ... done (package XFree86-Mesa) glut ... done (package freeglut) Is anybody else experiencing this? -- Regards, Steven