Felix Kühling wrote:
On Fri, 15 Aug 2003 13:13:13 -0700
Ian Romanick <[EMAIL PROTECTED]> wrote:
Felix Kühling wrote:

this is hopefully the final version of the per-screen dynamic GLX
extensions patch. Major changes from the last version:

- per-screen data in __GLXscreenConfigs
- __DRIscreen has an opaque pointer back to the screenConfigs
- Bumped the internal libGL API version to 20030813
- __glXExtensionBitIsEnabled must be per-screen too
- requires two helper functions to find the __GLXscreenConfigs from
 either a GLXcontext or a display, screen pair
- Updated some comments in glxextensions.c

To test that the screen finding functions actually work as expected we
should test at least one of the SwapInverval-related extensions and one
of the FBConfig-related ones. Can someone point me to some simple test
cases. I remember some posts about glxgears with SwapInterval support.
Is that in Mesa CVS? It is a bit inconvenient for me to download a whole
mesa CVS tree through a modem connection. Ian, IIRC you made those
changes in glxgears. Could you send me only the updated glxgears source?

I never noticed this before, but the files you have in lib/GL/glx are waaaaaaay out of date. There are a bunch of the functions you changed in glxcmds.c to get the __GLXscreenConfigs from the gc that don't have a gc anymore. The MESA_swap_frame_usage and OML_sync_control functions have been fixed to not use a gc (because there may not be one currently bound) since July 3rd. I did manage to work around all of the rejects, and I'm going to test it out.

Oops, that must be because I did that work on the config branch. If there were substantial changes in glx code after I created the branch then this work should probably go into the trunk. Feel free to commit your own version of the patch if you think it's ready. Then the vblank interrupt cleanup will have to wait until after the config branch merge or we decouple it from the configuration completely.

I've started working on doing this a piece at a time. I've got the first part done and in CVS already. I have a question about __glXFindGLXScreenConfigs. I see that it's basically a copy of glXFindDRIScreen (which I modified to use __glXFindGLXScreenConfigs). In glxcmds.c I see a bunch of routines (such as CreateContext) that do basically the same thing, but do it a totally different way.


        __GLXdisplayPrivate * priv = __glXInitialize(dpy);
        __GLXscreenConfigs * psc = &priv->screenConfigs[screen];

Would that be a better way to implement __glXFindGLXScreenConfigs?

I'll experiment around with it a bit, but I'm curious to hear other opinions.



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to