[JDEV] jabberd patch

Richard Dobson richard at dobson-i.net
Tue Feb 25 09:30:23 CST 2003


----- Original Message -----
From: "Tijl Houtbeckers" <thoutbeckers at splendo.com>
To: <jdev at jabber.org>
Sent: Tuesday, February 25, 2003 2:52 PM
Subject: Re:Re: [JDEV] jabberd patch


> "Richard Dobson" <richard at dobson-i.net> wrote on 25-2-2003 15:30:10:
> >> Then again, not using this patch has it's own problems when two
> >> clients with auto-reconnect start to "fight" over the same
> >> connection.
> >>  So this patch is definatly usefull for those who want to avoid this
> >>  type of
> >> behaviour.
> >
> >That issue can be solved pretty easily, the client that is getting
> >logged off because another one is connecting just needs to be notified
> >of the reason of disconnection and if it is because another client has
> >logged on you pop up a box or something and dont try to auto-connect
> >again (just like the MSN protocol does). There can be issues with
> >logging off the old connection and letting the new connection in as
> >jabber does now, but I think it is much better than the reverse as
> >more serious problems happen doing it the other way. If the only
> >problems are things such as this that can be pretty easily solved then
> >we should leave it as is, this is all about a weighing of benefits
> >against problems and I think the current method jabber uses has
> >definate benefits that outweigh the problems pointed out so far.
>
> Until the problem is solved on the protocol level and enough clients
> support it this patch could still be usefull for some.

Thats fine but it must not be standard or default.

> Besides that it
> should be possible to prefer this method is some specific cases, so if
> we fix this we should also include the *possibility* for the server to
> deny a client during login because another session with the same
> resource excists. (This should be no more than adding an error code I
> think).

I dont think this is needed because if this fighting over session issue is
the only real problem then it is unnecessary once it has been fixed, having
this as standard or default is bad since it introduces several arguably
serious problems (sessions being left hanging around, clients being stuck
left logged in, clients not being able to log in and causing lots of
confusion for novice users. Also to make this patch work and be user
friendly clients will have to be modified anyway to support the new error
code, I think that instead they should just be fixed so they dont fight over
sessions anymore.

Richard





More information about the JDev mailing list