Re: [oglbase-discuss] New GLX_ARB_get_proc_address specification

[Michael I. Gold]

At 09:46 AM 11/5/99 -0700, Brett Johnson wrote:
>
>I think we're looking at this from different perspectives, and that's the root
>of the miscommunication.  As I see it, the function pointer returned by a
>context independent GetProcAddress is explicitly *not* a pointer to a GL, GLX,
>or GLU function.  Instead, it is a pointer to a context independent piece of
>code that will do The Right Thing when called in the appropriate context.  If
>this code stub is called in an inappropriate context, it is an error, and it
>doesn't matter whether the code that would be called if the context was
>appropriate belongs to GL or GLX.

But it _does_ matter when were are defining error semantics.  Since this is
a GLX extension, perhaps we could simply define it as generating an XError
always, and never a GL_ERROR.  And, unless an error handler is installed,
this has the additional benefit of terminating the app when it makes a
mistake. =)

(Alternatively we could make the returned pointer context-dependent, and
this whole problem goes away, hehehe.)