On Thu, 4 Mar 2004 17:38:17 -0300 "Paulo R. Dallan" <[EMAIL PROTECTED]> wrote:
[snip] > > Hi Feliz, > > Made the tests - in fact, I had to patch the file manually - I noticed the > real change was only one line long, and I didn't know how to use the patch > program properly - what a shame!; ok, it is a newbie thing; give some credit > to me, I spent too much time in the Windows world! :) See man patch. One thing that takes getting used to is that patch reads the patch file from standard input. The other thing is, you need to get the -p option right. See man patch for more details. > > Basically, upgrading the mesa tree made a difference; Quake2, for example, > does not seem to freeze anymore; wolf3d-enemy territory also seems to accept > the OpenGL extentions without freezing anymore. But the difference was the > mesa update, not the patch. Good to hear. I'm not exactly sure why this fixed the lockups, but apparently the new way state is emitted is more sane than the old way. :) > > With the patch, I only had the impression (when trying quake2) that it was > slightly slower. But it was an impression, couldn't measure it for real. Ok. A small performance hit would be expected. But wasn't able to measure anything significant here. Probably it also depends on the application. Anyway, if it's not needed, the better. > > Another thing I notice was that after the mesa update I'm getting more > segmentation faults. Before, I had segfaults in some modes of quake2 - x11 > and SDL, OpenGL GLX being unplayable because of the mouse - so I play in Any segfaults that happen without OpenGL are definitely an application bug. > SDL_OpenGL - and in tux_aqfh; now, after the mesa update, in addition to > those, that continue, i got some in wolf3d-enemy territory too!!!!!: > > "LOADING... sounds > et.x86: savagetex.c:1847: savageBindTexture: Assertion `( � � � � � � � � � � � > Obj->DriverData != ((void *)0))' failed. > Received signal 6, exiting..." > > And: > > "200.221.1.6:27961 resolved to 200.221.1.6:27961 > et.x86: savagetex.c:1847: savageBindTexture: Assertion `(target != 0x0DE1) || > (tObj->DriverData != ((void *)0))' failed. > -------- UNRECOVERABLE ERROR --------" Ok. These are not segmentation faults but assertion failures. It happens when the programmer expressed an assumption in the code and it turns out that he was wrong. I just committed a fix for this (hopefully) to Mesa CVS. At least it solved a similar problem I was seeing with flight gear. Could you try it please? > Cube continues not to work, and neither > > I have more details if you need, just let me know, and I forward it to you or > to whom might be interested. > > Btw, I'm only e-mailing to you. If you find it interesting to other, please > feel free to forward this message! > > Best regards, > > Paulo Regards, Felix ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click -- _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
