On 4 June 2018 at 15:43, Jonas Wielicki <jo...@wielicki.name> wrote: > I think this is a good point, and I’ve run into this quite a few times > when > implementing MUC. Branching on the message type where 'groupchat' is > somehow > special, but then again only if it does *not* come from the bare JID, and > if > it *does* come from the bare JID, you may be confused (but that happens, > in > fact). And non-groupchat messages from the bare JID have a different > semantics > than non-groupchat messages from the occupant JID (I’m only talking about > MUC > here). This is weird, and I think if MIX does away with that, that would > not > be too bad. > > I'm not sure that type=groupchat messages from the '45 chatroom bare jid are different from type=groupchat messages from the '45 occupant bare jid except in who they're from, are they? I mean, they're different because they're from the chatroom itself, of course, but beyond that?
Well, errors, I suppose, come from the bare jid, but they're not groupchat of course. > > A random side note: I found another argument against the variant where the > client resource is encoded together with the participant id in the > resource: > The client resource is -- in contrast to the channel name and participant > id > -- not under control of the MIX service. A client resource can be very > long > (although it probably isn’t right now), and that would break things. > Combining > the participant ID and channel name in the nodepart allows the MIX channel > to > control both parts to always fit in a nodepart (restricting the length of > the > channel name most likely). > > > kind regards, > Jonas > _______________________________________________ > Standards mailing list > Info: https://mail.jabber.org/mailman/listinfo/standards > Unsubscribe: standards-unsubscr...@xmpp.org > _______________________________________________ > >
_______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: standards-unsubscr...@xmpp.org _______________________________________________