On Sun, 11 Nov 2007 23:17:59 +0100 Grégory SCHMITT <gy.schmitt@xxxxxxxxx> wrote: > > Sofar, no definitive conclusion. There's nothing obvious inside e2, > > nor could I see anything obvious inside gtk (both 2.12.0 and 2.12.1 > > have been used here, with these same symptoms). My current focus is > > on gcc. I've seen a few recent posts about its buggy optimisation - > > not only in the current 4.2.2. > > > > Any of you that can provide further clues, please do. One thing would > > be to try compiling without optimisation (i.e. -O0 in Makefile.config > > or in your CFLAGS). It would be good even to find a test-case that > > will always fail, so we can do some exploration. > > I'll do that. Correct me if I'm wrong, but I feel enabling DEBUG in > Makefile.config won't help much, because the DEBUG flag essentially > outputs e2 behavior - which is imho OK, see below - and not GTK issues. The debug messages might indicate whether a problem 'elsewhere' is in fact triggered by something inside e2, but in this particular case, there's nothing obvious in the current messages, and no obvious way to decide where more messages might help. I've experimented with a bunch of small changes, but sofar nothing has reliably improved things, except - possibly - compiling without optimisation. > I tried e2 (svn build 685) with my previous system (which was > Fedora 7 not upgraded), and I'm positive these glitches didn't exist > before. So I'll try to see options in GTK or gcc. I looked at those - nothing obvious to change. BTW, another problem with the current code is a sporadic (and for me, _very_ rare) crash after changing out of the temp dir created by the unpack plugin. This one is similarly elusive. Again, the evidence suggests something going on outside of e2. Regards Tom -- Users can unsubscribe from the list by sending email to emelfm2-request@xxxxxxxxxxxxx with 'unsubscribe' in the subject field or by logging into the web interface.