On Tue, 2004-02-24 at 18:05, Emmanuel Fleury wrote: > > If I stop the X server and try to restart, I can't. I get the exact same > error than for the gnome-panel but for the X server (I guess, I have to > investigate this a little bit more now that I better know what is going > on).
The X server itself can't encounter this error, probably the display manager does. > My point is that it seems to be extremely consistant when it occurs. So, > I think that there is some determinism inside this bug since it started. It doesn't seem even as deterministic as that for others... doesn't the problem go away for you if you 'do something else' for a while? > The theory of the "wrong code" being remanent in the processor does not > really satisfy me because we have no real evidence of it (how could we > dump the registers of the processor or the cache memory ?). It's the only theory I've seen so far which could explain the semi-deterministic nature of the problem. > I'll try to do some investigations and I'll probably come back with a > summary of my though. I guess that I have to add this to the follow-up > to the bug #216933 as it was the first. Yes, unfortunately the BTS doesn't seem to forward comments to both bugs even though I merged them. -- Earthling Michel Dänzer | Debian (powerpc), X and DRI developer Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer
