[Bug 27602] tp_connection_ensure_channel_async

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=27602

--- Comment #7 from Danielle Madeley <danielle.madeley at collabora.co.uk> 2010-04-12 21:42:10 PDT ---
(In reply to comment #5)
> > Move struct definitions from channel.h to channel-object.h
> 
> Hmm, I hadn't realised this would cause a circular reference. I'd much prefer
> it if you only moved the typedef, either into connection.h or some new central
> header (types.h or typedefs.h or some such), with the rest staying where it
> already is.
> 
> In other words, in connection.h or types.h or something:
> 
>     typedef struct _TpChannel TpChannel;
> 
> and in channel.h:
> 
>     /* TpChannel typedef is in connection.h */
>     typedef struct _TpChannelPrivate TpChannelPrivate;
>     typedef struct _TpChannelClass TpChannelClass;
> 
> TpProxy and TpDBusDaemon have a similarly circular relationship, with
> TpDBusDaemon typedef'd in proxy.h to resolve it.

Fixed.

> The new API needs adding to docs/reference/*sections.txt, which you'd have
> spotted if you'd run `make check` with gtk-doc enabled (please do this when
> writing patches for telepathy-glib). You'd also have spotted that t_c_c_c_f's
> doc-comment is labelled as if for t_c_e_c_f.

Fixed.

> > +  g_return_val_if_fail (g_simple_async_result_is_valid (result,
> > +        G_OBJECT (self), tp_connection_create_channel_finish), NULL);
> 
> The tag is conventionally the _async function, not the _finish function (I know
> TpAccount[Manager] uses the finish function in some places).

Fixed.

> > ... with the
> > + * %TP_CHANNEL_FEATURE_CORE feature ready on it.
> 
> Shouldn't these functions take an array of features as an optional argument? I
> expect it to be extremely common to want GROUP, for a start.

Fixed.

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