[PATCH] Hal privilige seperation

[John (J5) Palmieri]

On Fri, 2006-01-20 at 17:36 -0500, John (J5) Palmieri wrote:
> On Fri, 2006-01-20 at 20:49 +0100, Sjoerd Simons wrote:
> > On Fri, Jan 20, 2006 at 10:30:39AM -0800, Artem Kachitchkine wrote:
> > > Artem Kachitchkine wrote:
> > > >
> > > >If hald process suddenly terminates without calling KillAll(), will 
> > > >hald-runner terminate too?
> > > 
> > > OK, I see now that KillAll() only kills children, not hald-runner 
> > > itself. So the question becomes: how and when is hald-runner terminated?
> > 
> > the runner does the following to setup it's dbus connection:
> >   dbus_connection_setup_with_g_main(c, NULL);
> >   dbus_connection_set_exit_on_disconnect(c, TRUE);
> 
> So the runner is directly connecting to HAL and not going through the
> bus?
> 
> If it is going through the bus this will not work and you should listen
> for the NameOwnerChanged signal on the org.freedesktop.Hal name.

Ah 
#define DBUS_SERVER_ADDRESS "unix:tmpdir=/tmp/hald-runner"

Looks good.

-- 
John (J5) Palmieri <johnp at redhat.com>