[Standards] LAST CALL: XEP-0166 (Jingle)

juha.p.hartikainen at nokia.com juha.p.hartikainen at nokia.com
Wed Nov 28 11:48:03 UTC 2007


Is this moreless the clients internal implementation issue, than 
related protocol. 


>-----Original Message-----
>From: standards-bounces at xmpp.org 
>[mailto:standards-bounces at xmpp.org] On Behalf Of ext Lauri Kaila
>Sent: 28 November, 2007 13:13
>To: XMPP Extension Discussion List
>Subject: Re: [Standards] LAST CALL: XEP-0166 (Jingle)
>Hi all
>There's one point in the protocol I'm not very fond of: 
>returning application-level errors (especially 
>unsupported-content) with iq response to session-initiate. 
>Because if one implements XEP-0166 and
>XEP-0167 (and video, file transfer, whiteboard, etc) as a 
>separate logical modules, it means that the generic Jingle 
>code must consult the upper-level module before it can send IQ 
>response. The generic Jingle code must have a separate state 
>for a not-acked session, instad of making the transaction 
>"atomic" by sending ack immediately.
>The alternative would be to ack session-initiate always, and 
>have a unsupported-content info carried with 
>session-terminate. I believe that would be easier to 
>implement, although the protocol flow may not seem optimal.
>Now I had my say. If you don't find this issue serious enough 
>(or even an issue) - fine, I'm quite happy with the current spec.

More information about the Standards mailing list