[Xcb] XCB and OpenGL, help and/or project ideas wanted

[RALOVICH, Kristóf]

On Thu, Jun 18, 2009 at 15:37, Vincent Torri <vtorri at univ-evry.fr> wrote:

>
>
> On Thu, 18 Jun 2009, Riku Salminen wrote:
>
> > On Wed, 2009-06-17 at 12:02 -0700, Barton C Massey wrote:
> >> Great to hear from you!  Wish it had been sooner, though.
> >> The person you need to talk to is Kristof Ralovich
> >> <kristof.ralovich at gmail.com>, who has already done quite a
> >> bit of the work you propose to do with OpenGL, and is
> >> currently working on more.
> >
> > I have already been in e-mail correspondance with Kristof and he has
> > been very helpful. I found his Google Summer of Code work on XCB and
> > OpenGL and contacted him.
>
> Try to talk with Jeremy Kolb too
>
> Vincent
> _______________________________________________
> Xcb mailing list
> Xcb at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/xcb
>

Hi all,

sorry for taking so long to respond, but I am currently engaged with my
exams at university.

First I would like to state a few FACTs about xcb-glx:
- provides C binding ONLY for X11 protocol NOT the glX API in current form,
- has anything to do with OpenGL is when INDIRECT rendering,
- is mostly integrated into Mesa already,
- doesn't know support the HUGE number of payload requests embedded into
glXRender, glXRenderLarge, glXVendorPrivate, and glXVendorPrivateWithReply
messages.

I think today the Xlib interface is still the de-facto standard for
interfacing X (probably this is the reason the low level networking in
XOrg's libX11 was rewritten using xcb instead of replacing it). But to make
the situation more fuzzy, the xcb library also provides an API that is a
replacement for some of Xlib's functionality.

Some sort of new API extending xcb-glx could also be possible, but I don't
see any advantages of it, since in my opinion the most important use cases
of OpenGL exclusively drive DIRECT rendering and mostly "skip" the glx
protocol. I am advising against introducing a new ad-hoc API (addressing X
only), and waiting for the whole industry to follow it. The point of OpenGL
is standardization and source compatibility (that is platform independence).

On the other hand I have high hopes for the EGL API (also standardized and
platform INDEPENDENT) with support for both OpenGL and OpenGL ES and running
with or without X or embedded! I strongly believe EGL has the potential to
free us from working with wgl, glX, and alike in the long run. Altough today
Mesa provides an initial implementation of EGL in git, it is still
experimental and not shipping. Having stable EGL ready to be included in
distributions would be a great progress for everyone!

Bye,
Kristóf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freedesktop.org/archives/xcb/attachments/20090623/9077760d/attachment.html