Re: gconf_database_[dbus/corba]_notify_listeners



On Tue, 2005-03-08 at 10:25 +0100, Philip Van Hoof wrote:
> The function gconf_database_dbus_notify_listeners in the gconf-dbus
> branch isn't
> using the event-capabilities of DBUS. It's using a normal remote
> method call
> 
> It's currently using dbus_message_new_method_call. Wouldn't it be
> better if it
> used the dbus_message_new_signal stuff?
> 
> I'm asking this because I don't yet know much about DBUS :-). I do
> know it has, unlike
> some other IPC's, "events". So I'm trying to get an overall picture of
> how we can improve
> the DBUS layer.
> 

The way the dbus interface should work is that there are no "listener"
objects to invoke methods on. Use the signals mechanism.

The key reason to do this is to eliminate the concept of clients
registering with the daemon, which then kills the saved_state file and
associated nonsense.

Havoc





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]