[Telepathy] Proposed standard MC API

[George Wright]

On Tuesday 14 August 2007 14:12:25 Sjoerd Simons wrote:
> I might be misunderstanding the proposed spec. But it doesn't indicate a
> default dir for chandlers and has a function SetSearchPath:
>   Sets the search path to find the .chandler files which define the Channel
>   Handlers.

Choosing a default search path is on the TODO. Any suggestions?

> Which makes you fall into the trap of deciding which chandler is best. If i
> got say 3 telepathy clients installed (GNOME, KDE, something else?).. Each
> will install their chandlers, so now a text channel comes in.... Which one
> should be started by the MC?

Interesting problem. There's a chandler priority setting in the API which can 
be set by the Telepathy client to put their chandlers at the top of the 
priority queue - perhaps this could be used?

> By explicitely telling the MC over dbus what the chandlers are (or in which
> dir they are) for you session your basically working around this problem.

Indeed, but the same effect can be achieved by the Telepathy client setting 
the search path correctly.

> Say i'm a happy empathy user and this being the future, so assume empathy
> does supports voip. And there is thus a chandler registered for VOIP by
> empathy.
>
> Now i start up some magicall sip specific (branded?) client.. Which wants
> to take over (all?) sip calls from empathy. And thus wants to install it's
> own chander that _overrides_ the current ones for SIP streamed media
> channels.
>
> Will this work with the current proposed API (and if so how).

This could use the priorities setting for the chandlers, but this has made me 
aware of a new problem. Say we have two Telepathy clients running, each 
supplying their own channel handlers for the same type of channel, and each 
client connected to the same Mission Control instance. How would Mission 
Control handle a new channel coming in?

> And a final remark:
>
> org.freedesktop.Telepathy.MissionControl.ChannelHandler has a signal to
> notify the MC that it can pass the channel to the next handler. Shouldn't
> there also be a signal to notify that it shouldn't _ever_ be passed to the
> next handler ?

Surely that's implicit if the signal isn't ever emitted?

Regards,

George

-- 
George Wright, http://www.gwright.org.uk
Collabora Ltd - http://www.collabora.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/telepathy/attachments/20070820/79896670/attachment.pgp