Stephen J Baker wrote: > > On Sat, 11 Sep 1999, Brian Paul wrote: > > > [cross-posting] > > > > I've implemented this extension in Mesa 3.1 and checked it into > > the CVS repository. I also implemented GLU_EXT_get_proc_address > > and GLX_EXT_get_proc_address. Only the core extension has been > > documented in the Mesa/docs/ directory at this time. > > I have a problem. > > We established that it should be illegal (or at least meaningless) > to call glGetFuncAddressEXT for an extension that's not in the > glGetString(GL_EXTENSIONS) list, or for an OpenGL vN.M function > address unless glGetString(GL_VERSION) is at least N.M. > > How does this work with glX and glu varients of gl*GetFuncAddress? The same policy would apply. If glXGetProcAddressEXT or gluGetProcAddressEXT returns a non-null pointer, you still have to query the extension string to determine if the function will work. In the case of GLU, since there's no dynamically loaded drivers, it would be trivial to implement the policy that non-null pointers guarantee extension availability. But to be consistant with the GL and GLX versions we shouldn't make that claim. > There is no glu or glX version of glGetString - AFAIK. GLX 1.1 and later have glXQueryExtensionsString, glXQueryServerString, and glXGetClientString. GLU 1.1 and later have gluGetString. -Brian
