On Fri, 2009-05-08 at 06:44 +0200, Yves-Alexis Perez wrote: > On jeu, 2009-05-07 at 22:57 +0000, Debian Bug Tracking System wrote: > > Processing commands for cont...@bugs.debian.org: > > > > > reassign 522729 xfce4-notifyd > > Bug#522729: [banshee] crashes irregular > > Bug reassigned from package `banshee' to `xfce4-notifyd'. > > Hmhm, why reassigning? Bansheed crashed, not xfce4-notifyd. Maybe > there's a problem in xfce4-notifyd but I'm not sure banshee is supposed > to crash anyway :) > > But trying first with notification-daemon (or notification-daemon-xfce > if you don't want the gconf bloatness (which I guess you don't really > care about though)) might definitely be a good idea. Matthias, could you > do that? Sorry for my lack of explanation. I have tested, and can confirm that this bug _only_ happens with xfce4-notifyd. Both notification-daemon and notify-osd work fine with it. Tracing with dbus-monitor shows that the NotificationClosed signal which is sent out by xfce4-notifyd only has one integer, whereas in notify-osd and notification-daemon, there are two. The spec[1] also confirms that there should be two. Perhaps you could look into this?
As for why Banshee crashed, it's not Banshee's fault, but notify-sharp's fault. The exception occurred within notify-sharp code, while searching for that extra integer I was talking about (UINT32 reason, perhaps). [1] http://www.galago-project.org/specs/notification/0.9/x408.html#signal-notification-closed -- Regards, Chow Loong Jin
signature.asc
Description: This is a digitally signed message part