On Sat, 2009-01-17 at 11:28 +0100, Florian Echtler wrote: > >> The bug causes a segfault in libGL.so, though even with debug info > >> enabled, I can't see the exact location. However, I can work around the > >> bug when, in freeglut, I replace > > Why would you do that? There is no improvement to freeglut and no > > benefit to its users. If the read drawable and the drawing drawable are > > always the same, just use glXMakeCurrent. You already have a tool that > > does *exactly* the job you need done. Why change to a tool that is > > intended to suit a different need? > Uh, I do that because I need to use freeglut with the opensource drivers > and DRI? It won't work otherwise. Note: freeglut currently contains > _glxMakeContextCurrent_, which I replaced by _glXMakeCurrent_, not the > other way round.
Apologies. I misread the original message. For whatever reason (blame it on being Friday), I thought you were replacing glXMakeCurrent with glXMakeContextCurrent, not the other way around. > If I remember correctly, I already posted this patch to the freeglut > mailing list last fall, where it was rejected, citing possible side > effects on other platforms. As I am not a regular freeglut developer, I > accepted that and (wrongly) concluded that it has to be fixed in Mesa. > > So let me summarize: > - there is a bug in Mesa which causes glxMakeContextCurrent to segfault, > - but it only occurs because freeglut doesn't call glXMakeCurrent, as it > really should, and passes invalid parameters, too. There are two bugs in freeglut: - It calls a GLX 1.3 function on a system that doesn't support GLX 1.3. - It passes the wrong kind of drawable ID to glXMakeContextCurrent. > If that is correct, could somebody with more authority than me > please convince the freeglut guys to fix it? As freeglut is the default GLUT > implementation on most distros, I consider this quite important for > acceptance of the opensource DRI drivers.. a _lot_ of software uses GLUT, > and consequently crashes. Do they have a bugzilla? Is there an open bug for this issue there?
signature.asc
Description: This is a digitally signed message part
_______________________________________________ xorg mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/xorg
