[Standards-JIG] Re: JEP-0065 and MUC rooms

Peter Saint-Andre stpeter at jabber.org
Fri Jul 30 15:18:16 UTC 2004


In article <cba01a41040628143715904541 at mail.gmail.com>,
 Dave Smith <dizzyd at gmail.com> wrote:

> On Mon, 28 Jun 2004 23:22:46 +0300, Alexey Shchepin <alexey at sevcom.net> 
> wrote:
> > 
> > Hi!
> > 
> > When I want to send file to MUC participant, my client don't know his real 
> > JID,
> > and not able to calculate SHA1(SID + Initiator JID + Target JID).  Maybe it 
> > is
> > possible to avoid usage of real JIDs here?  And BTW, this JEP doesn't 
> > mention
> > that these JIDs should be normalized using stringprep profiles, otherwise 
> > you
> > can get different hashes.
> 
> Thanks for the comments, Alexey. It was my intention that the MUC
> participant would just use their MUC JID here -- i.e. room at host/user.
> 
> We'll get the JEP updated to require the use of stringprep'd JIDs when 
> hashing.
> 
> D.

The problem right now is that JEP-0045 discourages rooms from allowing 
IQs to/from room JIDs, even though it encourages messages to/from room 
JIDs. There may be good reasons for disallowing "IQ forwarding" through 
the room, but we need to think about this some more. One example: two 
room participants negotiate a file transfer via their room JIDs and the 
only workable option is In-Band Bytestreams. Now all that binary traffic 
is being routed through the room!

/psa




More information about the Standards mailing list