I think this is a bug...

[Brad Fitzpatrick]

I still maintain your fix isn't correct.  Did I miss your reply with a
reproducible test case, or better analysis?


On Fri, 16 Mar 2007, Pedro Melo wrote:

> Brad, any news on this one?
>
> Thanks,
>
> On Mar 9, 2007, at 9:44 AM, Pedro Melo wrote:
>
> > Hi,
> >
> > On Mar 9, 2007, at 1:18 AM, Brad Fitzpatrick wrote:
> >
> >> I'm not sure about this patch.
> >>
> >> When I look at that sub,
> >>
> >> IQ.pm....
> >> sub on_recv_from_client {
> >>     my ($self, $conn) = @_;
> >>
> >>     my $to = $self->to_jid;
> >>     if (! $to || $conn->vhost->uses_jid($to)) {
> >>         $self->process($conn);
> >>         return;
> >>     }
> >>
> >>     $self->deliver;
> >> }
> >>
> >>
> >> That seems to say, if there's no 'to' attribute, it's obviously
> >> intended
> >> for the server, not another user.  But if the 'to' is there, and
> >> the vhost
> >> uses that JID to represent itself (as opposed to handles_jid,
> >> which means
> >> the vhost's domain is the same as the 'to' domain... like the
> >> vhost for
> >> livejournal.com 'handles' brad at livejournal.com), then that means
> >> the same
> >> thing as no 'to' attribute, and the server should process it.
> >
> > You are forgetting IQ's addressed to local at domain (without
> > resource) that must also be processed by the server. uses_jid
> > compares the entire JID to the hostname. handles jid specifically
> > compares only the domain.
> >
> > Now that I look at my patch, it might need an extra check for bare
> > jids.
> >
> >> What IQ packet XML are you seeing that's spinning the server?
> >
> > I've finish the development where I was seeing that... :( I'll try
> > and rollback this patch and see if I can trigger it again....
> >
> > Best regards,
> > --
> > Pedro Melo
> > Blog: http://www.simplicidade.org/notes/
> > Jabber ID: melo at simplicidade.org
> > Use Jabber!
> >
> >
>
> --
> Pedro Melo
> Blog: http://www.simplicidade.org/notes/
> Jabber ID: melo at simplicidade.org
> Use Jabber!
>
>
>