On Sun, 2 Jan 2011 16:03:59 +0100 Peter van de Werken <pwerke...@a-eskwadraat.nl> said:
could you do: p *gcc (dump the whole gcc struct)? as such in valgrind i dont see any of your problems on restart/exit. it's a happy valgrind. i dont have extra modules loaded other than those that come with e17 itself (no e-modules-extra ones). > On Sun, Jan 02, 2011 at 12:50:05PM +0000, Mick wrote: > > On Sunday 02 January 2011 11:22:23 P Purkayastha wrote: > > > I am experiencing repeated crashes on exiting or restarting e17. It > > > is 100% reproducible here. I think it started about a week to two > > > weeks ago. One major update that happened just before the crashes > > > started was a system update in gentoo with "-hal" USE flag > > > (enlightenment was also updated). However, I did compile > > > enlightenment and e_dbus with hal USE flag and hald is running on my > > > system. > > I seem to have the same problem, running r55795 of libs & e. > Besides on e restart/exit, it also happens when rotating the screen with > xrandr. Attached are two gdb backtraces. > > > > I tracked down the crashes to the places module. Other gadgets do > > > not seem to trigger the crash. places module can't even be unloaded > > > once it is loaded (it gives no crashes on loading). The bt on trying > > > to unload places module is attached in the file gdb.places.txt > > I don't think it's caused by the places module, as I don't have it > loaded. The problem seems to be in e_gadcon.c, there the id_del func is > called with an invalid parameter (gcc->cf->id is invalid, because > gcc->cf is NULL, see the bottom of my backtraces). > > > I don't have a fix I'm afraid, but can confirm that I have been > > getting similar crashes when I try to remove some gadgets. The > > gadgets (e.g. weather, everything, etc) cannot be removed without > > having first removed the respective module. However, I have no > > crashes on starting or exiting E17. > > > > I am running revision 55768. > > Than it's very recent... looks like this bug is caused(/exposed?) > by r55771[1] where a 'gcc->cf = NULL' is added. > > regards, > Peter > > [1]: http://trac.enlightenment.org/e/changeset/55771 -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ras...@rasterman.com ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
