> ------ Forwarded message ------
> From: "Jeffrey H. Ingber" <jhingber@...>
> Subject: [E-devel] E/EVAS+DRI
> Date: Thu, 07 Dec 2000 19:00:34 -0500
> To: enlightenment-devel@...
> I have a particular (and reproducable) problem using e 16.5 with XF86
> 4.0.1 and DRI. I understand that e 16 is not being actively developed,
> but please allow me the time to explain a rather serious problem. I'll
> keep this short and to the point:
> The situation is as follows: A system with XFree86 4.0.x, DRI, a
> Matrox G400, and E is unable to to properly render OpenGL (especially
> GLU apps.) in a desktop window using direct rendering. Indirect
> rendering works, and direct rendering works properly in other WM's I
> have tested. When a program linked with libGLU (possibly libGL) creates
> OGL content in a window, X and the keyboard will become unresponsive,
> the mouse is able to move, one can SSH into the system and kill the X
> server, but the console never recovers and required a cold reboot. Even
> when X is killed, the desktop is still visible on the screen and the
> mouse continues to move. This is easily visible with the OGL plug for
> XMMS and Evas. Running these programs under E with the above hardware
> and X server config **WILL** lock the system.
> This problem makes using items such as Evas **impossible** when
> running E with the above configuration. This problem only seems to
> manifest itself under E, and **not** under other WM's. I haved posted
> this to the DRI mailing list and the common thread seem to be that all
> users experiencing this problem run E. I don't know which is the
> offending app, but I've tried other avenues, and this is the last one.
> Please see the following:
> The common thread is that all the users who reported this problem use
> E. This has been going on for over a year, and the DRI developers seem
> ignorant of the problem, as they are probably not using E. Frankly, I
> don't know if this problem is genuinely an e problem, or an E/DRI
> problem, but e is playing **some** part in this. Unfortunately, fixing
> this on my own is beyond my skill level.
> I am willing to have the appropriate hardware shipped N/C to E
> developers if you are willing to at least give this a look. Until there
> is some resolution to this, windowed OGL is **impossible** with the
> above hardware configuration. I'll buy an NVIDIA card before I ever get
> rid of E, but I'd like to hear some imput and thoughts on this subject.
> There is no specific kernel or XF86 config that causes this. Utah-GLX
> seems to work properly, though.
> Thanks for your time,
> Jeffrey H. Ingber (jhingber@...
Please file this in the DRI bug database.