Hello, I'm testing fairly recent (16th Oct) autobuilds jogl jars on OSX 10.6, using a 2007 mac Mini with the Intel GMA 950 chipset, and found that if I set the stencil bits of the requested capabilities to 8, then I get the following error: Exception in thread "AWT-EventQueue-0" javax.media.opengl.GLException: stencil buffer n/a at com.jogamp.opengl.FBObject.attachRenderbuffer(FBObject.java:1389) at jogamp.opengl.GLFBODrawableImpl.initialize(GLFBODrawableImpl.java:130) at jogamp.opengl.GLFBODrawableImpl.contextRealized(GLFBODrawableImpl.java:321) ... If I set the requested stencil bits to 0, then there is no error. I was surprised about this at first, because if I query the chosen caps I get: GLCaps[rgba 0x8/8/8/8, opaque, accum-rgba 0/0/0/0, dp/st/ms: 24/8/2, sample-ext default, dbl, mono , hw, GLProfile[GL2/GL2.hw], offscr[fbo]] which indicates that the 24/8 bit configurations for the depth and stencil buffers, respectively is supported. On the other hand, it is very likely that only packed depth+stencil buffers are supported (see this this thread for more info http://lists.apple.com/archives/mac-opengl/2008/Aug/msg00089.html). I checked the source code of FBObject.java and found that the condition to use packed depth+stencil buffer is evaluated in the following line (804) packedDepthStencilAvail = fullFBOSupport || gl.isExtensionAvailable(GLExtensions.OES_packed_depth_stencil); fullFBOSupport is false on the GMA 950 for some reason (pgl.gl.hasFullFBOSupport() returns false), but I know for sure that the chipset supports packed depth and stencil because of tests I run earlier. Indeed, manually inspecting the extensions string returned by gl, I found GL_EXT_packed_depth_stencil. However jogl is checking for GL_OES_packed_depth_stencil instead, so I think this is the source of the problem.
(In reply to comment #0) > Hello, I'm testing fairly recent (16th Oct) autobuilds jogl jars on OSX 10.6, > using a 2007 mac Mini with the Intel GMA 950 chipset, Thank you for testing on this config, which I don't have available! > If I set the requested stencil bits to 0, then there is no error. I was > surprised about this at first, because if I query the chosen caps I get: > > GLCaps[rgba 0x8/8/8/8, opaque, accum-rgba 0/0/0/0, dp/st/ms: 24/8/2, sample-ext > default, dbl, mono , hw, GLProfile[GL2/GL2.hw], offscr[fbo]] Yes, thats a 'fake', i.e. onscreen dp/st/ms config .. sorry. > > which indicates that the 24/8 bit configurations for the depth and stencil > buffers, respectively is supported. On the other hand, it is very likely that > only packed depth+stencil buffers are supported (see this this thread for more > info http://lists.apple.com/archives/mac-opengl/2008/Aug/msg00089.html). > > I checked the source code of FBObject.java and found that the condition to use > packed depth+stencil buffer is evaluated in the following line (804) > > packedDepthStencilAvail = fullFBOSupport || > gl.isExtensionAvailable(GLExtensions.OES_packed_depth_stencil); > > fullFBOSupport is false on the GMA 950 for some reason full FBO support is only enabled if (GLContext): ( isGL3() || // GL >= 3.0 isExtensionAvailable(GLExtensions.ARB_framebuffer_object) || // ARB_framebuffer_object ( isExtensionAvailable(GLExtensions.EXT_framebuffer_object) && // All EXT_framebuffer_object* isExtensionAvailable(GLExtensions.EXT_framebuffer_multisample) && isExtensionAvailable(GLExtensions.EXT_framebuffer_blit) && isExtensionAvailable(GLExtensions.EXT_packed_depth_stencil) > (pgl.gl.hasFullFBOSupport() returns false), but I know for sure that the > chipset supports packed depth and stencil because of tests I run earlier. > Indeed, manually inspecting the extensions string returned by gl, I found > GL_EXT_packed_depth_stencil. However jogl is checking for > GL_OES_packed_depth_stencil instead, so I think this is the source of the > problem. Yes, can add GLExtensions.EXT_packed_depth_stencil to the above query for packedDepthStencilAvail. Thank you!
Fixed w/ commit dedf4ea7ddab1279602208fa194231ea65409251
Great, thanks a lot for the quick fix. Do the most recent autobuilds (jogl-b837-2012-10-23_19-29-45) already incorporate the fix? (In reply to comment #2) > Fixed w/ commit dedf4ea7ddab1279602208fa194231ea65409251