When trying to obtain the value of a GLXFBConfig's GLX_BIND_TO_MIPMAP_TEXTURE_EXT attribute (using glXGetFBConfigAttrib()), the driver returns -1 (GLX_DONT_CARE), while only GL_FALSE and GL_TRUE are allowed. See: http://lists.freedesktop.org/archives/compiz/2006-December/000957.html
The bug priority was upgraded (P2->high) with the bugzilla configuration change. I'm Changing the priority back to the normal one. Sorry for the spam.
Created attachment 9556 [details] [review] Set default values for bindTo fields. This patch sets some reasonable defaults for the various texture_from_pixmap related "bindTo" fields. After grepping around in the server, my gut feeling is that if GLX_EXT_texture_from_pixemap is supported, the server ought to send values for these fields to the client. It currently does not. In either case, I'm not terribly opposed to this work-around.
Created attachment 10938 [details] [review] Only return GL_FALSE or GL_TRUE for this attribute The previous patch had undesirable side effects ('libGL warning: 3D driver claims to not support visual 0x*' for most if not all visuals, and compiz didn't start either). This is a simpler patch which just makes sure that only GL_TRUE (if set to that by the server) or GL_TRUE (otherwise) is ever returned for this attribute. Ian, is this an acceptable approach? FWIW, it allows me to start stock compiz with the i915tex driver.
(In reply to comment #3) > The previous patch had undesirable side effects ('libGL warning: 3D driver > claims to not support visual 0x*' for most if not all visuals, and compiz > didn't start either). This is a simpler patch which just makes sure that only Any idea why? > GL_TRUE (if set to that by the server) or GL_TRUE (otherwise) is ever returned > for this attribute. Ian, is this an acceptable approach? FWIW, it allows me to > start stock compiz with the i915tex driver. This patch will return GL_FALSE for GLX_BIND_TO_MIPMAP_TEXTURE_EXT when bindToMipmapTexture is still set to GLX_DONT_CARE. Correct? I guess I don't see how that's different from pre-initializing bindToMipmapTexture to GL_FALSE.
(In reply to comment #4) > This patch will return GL_FALSE for GLX_BIND_TO_MIPMAP_TEXTURE_EXT when > bindToMipmapTexture is still set to GLX_DONT_CARE. Correct? Right. > I guess I don't see how that's different from pre-initializing > bindToMipmapTexture to GL_FALSE. I'm afraid I can't explain why, but it definitely does make a difference. Apparently pre-initializing these fields causes the configs to be treated differently in places where they shouldn't be.
What is the state of this bug ? Is the patch intended to make his way to mesa/git - works fine for me by the way - or is another fix expected ?
Patch pushed to master branch.
Mass version move, cvs -> git
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.