On Tue Sep 30 17:42:58 2008, Jonathan Schleifer wrote:
Am 30.09.2008 um 17:33 schrieb Dave Cridland:
And to cover our coversation elsewhere (through that funny
"Instant Messaging" thing), a downside of including <body> is
that a client might assume it's a reasonable alternative, whereas
otherwise it could bounce the message (type="error") which would
cause the sender to re-initiate the session.
Messages with unknown stuff are simply ignored, the RFC says so
IIRC.
Hmmm. I wonder if we could move toward changing that. Messages
without any known stuff feel like they should result in an error.
So yes, JS's problem is real, but the proposed cure of adding
<body> to IBB is worse than the disease, and I'll cheerfully
admit I hadn't thought this one through - sorry for jumping in
like that.
It solves the problem with clients that don't know IBB, thus will
just throw away the stanza and not warn - the user never knows he
lost a message.
It's certainly no worse than without encrypted streams.
Incidentally, both ends can check the session by using XEP-0199
inside the P2P XML stream. And XEP-0198 is also applicable here,
and much more useful than XEP-0184 on the IBB packets.
Think about small clients that only support the really necessary
stuff to use XMPP.
But then they wouldn't have E2E encryption anyway.
I suppose the point is that if you want to have a highly reliable
stream, you use XEP-0198 on it, and this applies whether you're
referring to a C2S or P2P stream.
Dave.
--
Dave Cridland - mailto:[EMAIL PROTECTED] - xmpp:[EMAIL PROTECTED]
- acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
- http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
