Hi J.P, I've tried out your example and found that with paged databases and your supplied model everything works fine for the first viewer then on the subsequent viewer I get:
Warning: detected OpenGL error 'invalid operation' at after RenderBin::draw(..) Warning: detected OpenGL error 'out of memory' at after RenderBin::draw(..) Warning: detected OpenGL error 'out of memory' at after RenderBin::draw(..) For none paged databases, even ones with texturing I'm not getting any errors though, for any of the viewer runs. I really don't know what this means if anything, perhaps just a coincidence. Curiously your database uses straight osg::Geometry meshes for geometry, while the ones of mine were built with osgTerrain::TerrainTile. Your database will be only loading the topmost tile so the DatabasePager won't actually be even playing a part, so I doubt this is part of the equation. While I can't see a particular pattern emerging, I can at least reproduce the issue. I'll try and remove the multi-view element to the equation as I wouldn't have thought use of CompositeViewer will be a particular problem for VBO/Texture object usage. Robert. On Wed, Nov 25, 2009 at 8:28 AM, J.P. Delport <jpdelp...@csir.co.za> wrote: > Hi Robert, J-S, > > the results of my most recent tests: > > With old texture pool (#define USE_NEW_TEXTURE_POOL 1 commented out): > > ./test cow.osg -> All OK > > ./test earth_sphere_dt1.ive -> All OK (see model attached) > > ./test earth.ive -> sometimes corrupted model and textures, sometimes > segfault, looks like reading bogus data from the GPU (I used the top level > file from www.openscenegraph.org/data/earth_bayarea/earth.ive) > > With new texture pool: > > ./test cow.osg -> All OK > > ./test earth_sphere_dt1.ive -> textures missing on second viewer start, get > warning: > Warning: detected OpenGL error 'invalid operation' at After > Renderer::compile > > ./test earth.ive -> segfault or corrupted display. > > Latest test app also attached. I've tried to switch off unref of images > after apply, hope it worked. > > regards > jp > > J.P. Delport wrote: >> >> Hi J-S, >> >> Jean-Sébastien Guay wrote: >>> >>> Hi again, >>> >>>> I haven't had a chance to check your code yet, give me a few minutes and >>>> I'll check if there's anything amiss that I can see. >>> >>> I'm sorry to say I can't repro here on Windows Vista, VC++ 2005 on SVN >>> head. But it seems that you've pinned it down to the texture pool as I >>> thought, though it would seem to be OS specific? A bit weird. >>> >>> Sorry I couldn't help more. >> >> can you try with terrain as the parameter to the test app. For me this >> still fails on the second creation of the viewer. When zooming into the >> terrain the newly loaded textures are OK though. >> >> jp >> >>> >>> J-S >> > > -- > This message is subject to the CSIR's copyright terms and conditions, e-mail > legal notice, and implemented Open Document Format (ODF) standard. The full > disclaimer details can be found at http://www.csir.co.za/disclaimer.html. > > This message has been scanned for viruses and dangerous content by > MailScanner, and is believed to be clean. MailScanner thanks Transtec > Computers for their support. > > > _______________________________________________ > osg-users mailing list > osg-users@lists.openscenegraph.org > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > _______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org