GObject signals

[Sadrul Habib Chowdhury]

* Sadrul Habib Chowdhury had this to say on [26 Jun 2008, 00:05:54 -0400]:
> * Gary Kramlich had this to say on [25 Jun 2008, 19:33:12 -0500]:
> > Sadrul Habib Chowdhury wrote:
> > 
> > <snip>
> > 
> > > To keep things plain and simple, we can have g_signal_connect_type,
> > > which will have a very similar signature as g_signal_connect, the only
> > > difference being the first parameter, which will be a GType, instead of
> > > a GObject *. To elaborate:
> > 
> > <snip>
> > 
> > I wouldn't advise using the g_* namespace, unless you're planning on
> > pushing that as a patch up to glib.  and if so, add a glib version check
> > around it when it gets accepted.
> > 
> 
> I renamespaced the function to purple_type_connect, and submitted a
> patch in GNOME bugzilla (#540256). (we can rename to g_* if/once the
> patch is accepted)

The patch is a WONTFIX with a very well reasoned explanation [1]. We
need to be careful of the states of the objects when the signals are
emitted. (we probably already do it in most places, since the purple
signals are emitted effectively at the same object states as for the
gobject signals, but we need to keep an eye out anyway)

We can have an additional purple_type_disconnect. I will probably add it
later tonight, if someone else doesn't do it by then.

[1] http://bugzilla.gnome.org/show_bug.cgi?id=540256#c4

> Sadrul
>