On Tue, Aug 16, 2011 at 8:34 PM, Peter Saint-Andre <[email protected]> wrote:
> I'm working my way through a backlog of open issues related to MUC, so
> you might see a few messages on the list here soon.
>
> Someone pointed out to me that the direct invitation method (XEP-0249)
> does not have feature-parity with the mediated invitation method because
> in XEP-0045 the inviter can indicate that the room is a continuation of
> an existing one-to-one thread:
>
> <message
>    from='[email protected]/desktop'
>    id='gl3s85n7'
>    to='[email protected]'>
>  <x xmlns='http://jabber.org/protocol/muc#user'>
>    <invite to='[email protected]'>
>      <reason>
>        Hey Hecate, this is the place for all good witches!
>      </reason>
>      <continue thread='e0ffe42b28561960c6b12b944a092794b9683a38'/>
>    </invite>
>  </x>
> </message>
>
> Contrast with XEP-0249:
>
> <message
>    from='[email protected]/desktop'
>    to='[email protected]'>
>  <x xmlns='jabber:x:conference'
>     jid='[email protected]'
>     password='cauldronburn'
>     reason='Hey Hecate, this is the place for all good witches!'/>
> </message>
>
> In version 1.1 of XEP-0249 we added the optional 'reason' attribute. I
> think we could similarly add an optional 'continue' attribute (with a
> datatype of boolean):
>
> <message
>    from='[email protected]/desktop'
>    to='[email protected]'>
>  <x xmlns='jabber:x:conference'
>     continue='true'
>     jid='[email protected]'
>     password='cauldronburn'
>     reason='Hey Hecate, this is the place for all good witches!'/>
> </message>
>
> We could also add an optional thread attribute:
>
> <message
>    from='[email protected]/desktop'
>    to='[email protected]'>
>  <x xmlns='jabber:x:conference'
>     continue='true'
>     jid='[email protected]'
>     password='cauldronburn'
>     reason='Hey Hecate, this is the place for all good witches!'
>     thread='e0ffe42b28561960c6b12b944a092794b9683a38'/>
> </message>
>
> Yes, it's unfortunate to be adding attributes to "jabber:x:conference"
> at this point, but the new attributes are backward-compatible.

This seems reasonable.

/K

Reply via email to