HI Glenn, We did some work for 3.6.3 to smooth the transition between glTexImage and glTexStorage remapping to sized types where possible, perhaps this particular image format it's not remapped properly.
The best thing to do now is create a small example/provide data to illustrates the problem then we can work out what the short term application fix would be and what the longer term OSG fix might be if appropriate. Now that 3.6.3 is tagged I'll be taking a step back from testing and debugging on the OSG side. If there are bugs that appear in 3.6.3 that need to be addressed I'll return to another cycle of testing, debugging and working towards a 3.6.4 release. I won't push for 3.6.4 for another couple of months though - I have plenty of other work to get on with :-) Cheers, Robert. On Fri, 14 Sep 2018 at 14:50, Glenn Waldron <gwald...@gmail.com> wrote: > > Robert, > Transparency wasn't an issue (that was a missing GL_BLEND) but there are a > couple other interesting things I discovered with texture arrays: > > > 1) After loading a GIF, I had to manually call > image->setInternalTextureFormat(GL_RGBA8), otherwise I get this message and a > blank texture: > > Warning: Texture2DArray::applyTexImage2DArray_subload(..) given image do have > wrong dimension or internal format. > > The original internal texture format was "4" which doesn't appear to be legal > even for glTexImage; perhaps glTexStorage is being more strict. > > > 2) When using the LINEAR_MIPMAP_LINEAR minfilter, along with calling > tex->setResizeNonPowerOfTwoHint(false), the image appears black and I see > this message: > > Warning: detected OpenGL error 'invalid operation' at > StateSet::compileGLObejcts() compiling texture attribute Texture2DArray. > > > Again, these may just be glTexStorage strictly enforcing things that > glTexImage was lenient about. > > Glenn > > > On Fri, Sep 14, 2018 at 8:56 AM Robert Osfield <robert.osfi...@gmail.com> > wrote: >> >> On Fri, 14 Sep 2018 at 13:08, Glenn Waldron <gwald...@pelicanmapping.com> >> wrote: >> > False alarm on the transparency report -- sorry! >> >> That's a relief, didn't want to break the record of hearing of the >> first bug right after tagging a first release. >> >> Thanks for the testing, >> Robert. >> _______________________________________________ >> 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 _______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
