[patch] direct service connections
hp at redhat.com
Wed Nov 3 21:22:30 PST 2004
On Sun, 2004-10-31 at 17:45 -0500, Colin Walters wrote:
> On Sun, 2004-10-31 at 17:50 -0500, Havoc Pennington wrote:
> Forgot to reply to this:
> > A harder one is that we probably have to split the object tree from the
> > connection objects. So e.g. signals would go out over all connections
> > using that object tree.
> Hm, complicated. So the libdbus level would filter duplicates out?
> I was thinking that with direct connections, you open up the possibility
> for two different processes to "own" the same service, at least in
> theory (in practice I don't think that will happen). But if that
> does happen, then why are we treating signals specially? What
> if a listening service wants to take some response to that signal -
> perhaps doing some method call on the originating object? In that
> case you also get duplicate method calls going back...unless there is
> filtering at the libdbus level, and that sounds hard.
I think this mostly just works - if you have a Proxy for a direct
connection to an object, and also a Proxy for a via-bus connection to
the same object, then sent the signal over both connections; you'd just
get the signal emitted on both proxies one time each. Same thing happens
if you have two Proxy both going via the bus.
More information about the dbus