On Thu, Apr 17, 2008 at 3:35 PM, Andreas Färber <andreas.faerber@xxxxxx> wrote: > > Am 17.04.2008 um 22:56 schrieb Andreas Färber: > > > > > > > Am 17.04.2008 um 22:49 schrieb Ryan Leavengood: > > > > > On Thu, Apr 17, 2008 at 4:36 PM, Michael Lotz <mmlr@xxxxxxxx> wrote: > > > > > > > I'd be interested in comments on that topic before applying anything > > > > though. Is my understanding of the problem correct at all, do we care > > > > about those optimizations and how would others suggest we move forward > > > > with that? > > > > > > > > > > I am certainly not a compiler expert and don't have the time now to > > > research this aliasing issue, but I do recall hearing how Rene Gollent > > > cannot get a stable Haiku on his hardware when compiling with GCC4. So > > > it might be related to this issue or other alignment issues messing up > > > driver code and therefore causing a crash. So until we can really dig > > > into these issues it might be smart to go easy on the optimizations. > > > > > > > No KDL so far! Instead I get an apparently reproducible Tracker segfault > crash on restart, and one time the system started to get unresponsive and > hung apart from F12. So definitely a welcome improvement but not yet the > full cure. > > Andreas > Andreas, Just wondering if you are also going to be checking the BePorts ports on Haiku in both GCC2.95 and GCC4? I see the plan is to shift focus to Haiku, so might be good to test out on both, so if/when the move is made to GCC4 we'll have a better idea of where the various ports stand. -scottmc