[Standards] s2s and gracelessly broken streams

Justin Karneges justin-keyword-jabber.093179 at affinix.com
Wed Mar 28 18:06:30 UTC 2007


On Tuesday 27 March 2007 9:44 pm, Mridul wrote:
> Justin Karneges wrote:
> > I'd also go as far as saying that we should decouple probing from being
> > explicitly related to the login process.  For example, if a client were
> > to log off/on within a very short timeframe, I'd say that the server
> > shouldn't probe the whole roster again.
>
> Currently, not doing so would be breaking a bunch of (implied) MUST
> requirements from xmpp right now.
> Or you mean we can adapt this into bis spec ?

I think it is okay to violate a MUST if you know what you are doing.

If there is anything in the text that says a server "MUST probe" when a client 
logs in, I think it is safe to return cached answers and probe some minutes 
later, without any harmful effect.  Because ultimately, it will look to the 
client as if you probed (and you are probing, but just not at the exact 
moment -- the cached presence is fresh enough).

Changing a MUST to a SHOULD here might actually be worse, because that might 
give license to not probe at all.

It comes down to good implementation judgment, but this is my own opinion.  I 
don't know if RFC-writers share it. :)

-Justin



More information about the Standards mailing list