Upgrading OpenGL version

[Tristam MacDonald]

Is anyone familiar with the process of up[grading pyglet's supported OpenGL version? I assume that it is a script which operates on glext.h to generate the supported extensions?

I need to add support for geometry shaders, transform feedback, and a couple of other new extensions.

--
Tristam MacDonald
http://swiftcoder.wordpress.com/

[Andreas Schiefer]

On 6 Sep., 15:47, Tristam MacDonald <swiftco...@gmail.com> wrote:
> Is anyone familiar with the process of up[grading pyglet's supported OpenGL
> version? I assume that it is a script which operates on glext.h to generate
> the supported extensions?

I guess it's something like

$ python tools/gengl.py gl

called from within pyglet's root directory (run without the "gl"
attribute to see all possible modules).
But unfortunately it doesn't work on my computer (crashes while
accessing a method of a None object).

> I need to add support for geometry shaders, transform feedback, and a couple
> of other new extensions.

That would indeed be nice to have.


Andy

[Bruce Smith]

On Mon, Sep 6, 2010 at 8:15 AM, Andreas Schiefer <andreas....@gmail.com> wrote:

On 6 Sep., 15:47, Tristam MacDonald <swiftco...@gmail.com> wrote:
> Is anyone familiar with the process of up[grading pyglet's supported OpenGL
> version? I assume that it is a script which operates on glext.h to generate
> the supported extensions?

I guess it's something like

 $ python tools/gengl.py gl

called from within pyglet's root directory (run without the "gl"
attribute to see all possible modules).
But unfortunately it doesn't work on my computer (crashes while
accessing a method of a None object).

FYI, A few months ago Joe Wreschnig filed an issue http://code.google.com/p/pyglet/issues/detail?id=485

about a crash in that gengl.py script, and included a patch (but not a traceback). There was a bit more info in the pyglet-users thread at the time, so maybe you can figure out whether it's the same issue.

- Bruce Smith

[Joe Wreschnig]

The traceback is fairly useless and probably even misleading, because
the problem is not that some function got a None, but that there was a
misparse some steps earlier and exactly what happens when you get a
misparse is dependent on what exactly was in the header file after the
misparse (for example the misparse was harmless in many other header
files). The bug report goes into more detail.

[Andreas Schiefer]

On 6 Sep., 20:13, Bruce Smith <ores...@gmail.com> wrote:
> FYI, A few months ago Joe Wreschnig filed an issue
> http://code.google.com/p/pyglet/issues/detail?id=485
> about a crash in that gengl.py script, and included a patch (but not a
> traceback). There was a bit more info in the pyglet-users thread at the
> time, so maybe you can figure out whether it's the same issue.

Thanks for the hint (and Joe for the patch), I've applied the patch
and it worked just fine afterwards.

The patch is now in pyglet-trunk, and I've also regenerated most of
the OpenGL wrappers (except for the agl* and wgl* things).
All features needed by Tristam should be available now (but I haven't
tried them, just checked if they are defined).

Andy

[Tristam MacDonald]

Many thanks for taking the time to do this! I will checkout the trunk version and make a few tests in the next couple of days.