Hi Russel, On Thu, Sep 03, 2009 at 06:33:24AM +0100, Russel Winder wrote: > > > How come a problem in Seahorse causes stupid behaviour of viz with dbus? > > > This seems obscure to say the least!
> > viz contacts seahorse over dbus to verify GPG signatures on revisions. > Sorry to hammer on this one, but I guess there is some issue I am > missing. > Knowing nothing of the details, my reaction would be: > OK so Bazaar tries to contact Seahorse over DBus, there is an exception. > This exception is internal and should not be exposed to the user. The > exception should be caught by Bazaar. Bazaar knows this is a sequencing > issue and that a second try should work so try it. It works. User > probably doesn't notice the slight delay. > The fact that the DBus problem appears to be cured by a later Seahorse > is not a reason for Bazaar to abdicate its responsibility to the user, > i.e. not exposing them to unnecessary bizarrenesses. My feeling is that > Bazaar should be handling this exception not passing the buck. We could indeed add a workaround for this bug in seahorse and have a mechanism like that. I'm not sure it's worth the extra code complexity though, as the versions of the distributions that this version of bzr-gtk will end up in will also have the fixed version of seahorse. Cheers, Jelmer -- bzr-gtk mailing list [email protected] Modify settings or unsubscribe at: https://lists.canonical.com/mailman/listinfo/bzr-gtk
