[Standards] XEP-0198 minor enhancement

John Williams (johnwi3) johnwi3 at cisco.com
Thu Jun 4 05:47:11 UTC 2015


Thanks Florian & Dave,

I made a major misunderstanding, and this now makes good sense. 

I was thinking the 'h' value in the failed stanza was the servers count of stanzas received by the session for delivery **to** the client. That is to say a <resume> using that value would require re-transmission of zero stanzas. I thought you were trying to make resume work with dropped packets - which could open up a can of worms. 

What you have in mind is that’s oldh is the servers count of packets received by the session **from** the client. As such it's just a way of allowing the client to know which stanzas made it to the server before the connection dropped. I can see that could be a useful enhancement.

== John (Jock) Williams ==

-----Original Message-----
From: Standards [mailto:standards-bounces at xmpp.org] On Behalf Of Florian Schmaus
Sent: Wednesday, June 03, 2015 12:35 AM
To: XMPP Standards
Subject: Re: [Standards] XEP-0198 minor enhancement

On 03.06.2015 01:48, John Williams (johnwi3) wrote:
> Thanks for the clarification.
> Hum,.. not sure how useful this is, since a lot of stanzas are of little long term interest (eg: chatstates), but as you describe it seems pretty harmless.

Consider the "XMPP user walks into the subway without GSM reception"
scenario which leads to a <failed/> once the user's device is back online. Without 'h' in <failed/> the client is unable to tell which messages have been received by the server right before the signal was lost. With 'h', the client is able to display some more visual feedback about the state of outgoing messages.

tl;dr: It's useful for mobile clients if you slightly missed the max resumption time.

- Florian



More information about the Standards mailing list