[Bug 27948] TpBaseConnection: support AddClientInterest, RemoveClientInterest

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=27948

--- Comment #7 from Simon McVittie <simon.mcvittie at collabora.co.uk> 2010-05-18 03:28:18 PDT ---
(In reply to comment #6)
> + * @self: a connection
> Aren't we using "a #TpConnection" as convention now?

IMO: you can if you want, but it's pointless to have as a convention, unless
the argument is of a different C type. When the C type is correct, we don't
need to hyperlink it again (gtkdoc will already get it right), but when the
type is not what you'd expect in order to reduce casting, it's necessary to
hyperlink it explicitly. So I think these usages are correct:

    /**
     * tp_proxy_get_stoat:
     * @self: a proxy
     * ...
     */
    Stoat *tp_proxy_get_stoat (TpProxy *self);

    /**
     * tp_proxy_get_badger:
     * @self: (type TelepathyGLib.Proxy): a #TpProxy
     * ...
     */
    Badger *tp_proxy_get_badger (gpointer self);

(In each case, gtk-doc will produce one hyperlink to TpProxy. In the latter
case, g-i would ideally pick up Proxy.get_badger as a method, but it doesn't -
I think danni filed a bug?)

> +  tp_connection_add_client_interest_by_id (test->conn,
> +      TP_IFACE_QUARK_CONNECTION_INTERFACE_AVATARS);
> This is just for the test right? Clients are not supposed to declare they are
> interested in avatars now, aren't they?

Correct. I'd have used Location and MailNotification (which both actually
benefit from this feature), but MailNotification is still in draft.

As currently implemented (and, IMO, as should be spec'd when I revise the
spec), it's always OK (albeit pointless) for a client to declare an interest in
an arbitrary interface.

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.