[gameprogrammer] Re: GL 2.0 on Xorg w/ ATI's proprietary drivers

On Fri, May 12, 2006 at 01:05:06PM -0500, Bob Pendleton wrote:

> On Fri, 2006-05-12 at 11:26 +0200, Latimerius wrote:
> > On Thu, May 11, 2006 at 02:03:02PM -0500, Bob Pendleton wrote:
> > 
> > > On Wed, 2006-05-10 at 23:29 +0200, Latimerius wrote:
> > > > 
> > > > test.cpp: In function `int main()':
> > > > test.cpp:7: error: `glCreateProgram' undeclared (first use this 
> > > > function)
> > > > test.cpp:7: error: (Each undeclared identifier is reported only once 
> > > > for 
> > > >    each function it appears in.)
> > > 
> > > Aren't those compiler errors, not link errors? It looks like the symbol
> > > is not defined in gl.h or glext.h. I would go to /usr/include/GL and
> > > grep to see if glCreateProgram is defined anywhere. If it isn't then I
> > > would assume that I had an old version of the .h files and try to find
> > > new ones. If glxinfo says you have 2.0 then you have 2.0. 
> > 
> > I apologize, these *are* compiler, not linker errors.  I pasted into my
> > message wrong output (from a previous compilation when test.cpp didn't
> > have all the required headers included) and failed to notice before I
> > sent it.
> > 
> > Anyway, my mistake notwithstanding, there finally *are* link errors -
> > I'm sorry for the confusion.
> 
> No problem :-) If the APIs aren't in the .h files then it does not
> surprise me that they also cause link errors. But, I believe you said
> you checked the .dlls and the entry points are there?

Actually, I said I believed they weren't.  Headers are OK, however when
I extracted symbols from ATI's libGL.so there weren't any shader API
symbols that are part of GL 2.0.

That's basically what my question is about - if someone here manages to
compile a GL program that uses glCreateProgram() & comp., I'd be
interested where the definitions for the entry points come from.

        latimerius



---------------------
To unsubscribe go to http://gameprogrammer.com/mailinglist.html

Other related posts: