CyrusSASL broken in jabberd-2.3.[01]
Hi, when trying to build 2.3.1 for openSUSE, I found that this change broke CyrusSASL support: https://github.com/jabberd2/jabberd2/commit/e134d30f4bb7d01e9674ef0338e9bc46e8cd2349 It adds three new local includes, which do not exist in the source tree. Tomasz, could it be that you forgot to git-add them, before committing that change? cu Reinhard
Re: XMPP connections
Hi, On Tue, 26 Nov 2013 at 11:41, Haider Ali wrote: Can anyone let me know that how XMPP handle so many connections. Since we know that we can only open 2 ^ 16 = 65536 ports ( connections ) with a single machine. But i came to know that single xmpp server can handle more connections than 65536. your assumption that there can only be a single connection per port number at a time is wrong. A combination of local IP address and port can be used for many connections at once as long as the remote IP address *or* the remote port number are different between each of them. Server sockets do this automatically with every call to accept(), but for client sockets you have to set the SO_REUSEADDR option to allow a second bind() while socket using the same port number is still open. cu Reinhard
Re: XMPP connections
On Tue, 26 Nov 2013 at 12:09, Eric Koldeweij wrote: The 65536 port limit is a client limit, not a server limit. A client can not have more than 65536 connections open to any other host at the same time (65536 is theoretical, in reality this number will be much lower). A client can even have more connections to a single server as long as all combinations of local and remote port numbers are distinct. So, the actual (theoretical) limit is 64k^2 connections per unique combination of local and remote IP addresses. Or in other words, it is the combination of local IP address, local port number, remote IP address and remote port number that uniquely identifies a TCP connection and if only one of these four values is different we are looking at two separate connections that can exist in parallel. cu Reinhard
Re: jabberd-2.2.16 release
On Fri, 4 May 2012 at 21:08, Tomasz Sterna wrote: Next jabberd2 release is available. Get 2.2.16 release at GitHub: https://github.com/downloads/Jabberd2/jabberd2/jabberd-2.2.16.tar.gz when I build this on the openSUSE Build Service, I get these reports from the automagic that checks for serious compiler warnings at the end of the build: I: A function uses a 'return;' statement, but has actually a value to return, like an integer ('return 42;') or similar. W: jabberd voidreturn out.c:635 I: Statement might be overflowing a buffer in strncat. Common mistake: BAD: strncat(buffer,charptr,sizeof(buffer)) is wrong, it takes the left over size as 3rd argument GOOD: strncat(buffer,charptr,sizeof(buffer)-strlen(buffer)-1) E: jabberd bufferoverflowstrncat main.c:776:24 I: Program returns random data in a function E: jabberd no-return-in-nonvoid-function log.c:258 I: Program returns random data in a function E: jabberd no-return-in-nonvoid-function log.c:258 cu Reinhard
Re: mu-conference: is the project alive?
Hi, On Sun, 8 May 2011 at 10:57, Tomasz Sterna wrote: Maybe it's time to adopt mu-conference to jabberd2 codebase? I think that would be a very good idea. cu Reinhard -- To unsubscribe send a mail to jabberd2+unsubscr...@lists.xiaoka.com
Re: jabberd with mu-conference
On Tue, 23 Feb 2010 at 18:28, Eric Schnoebelen wrote: How long is the secret? There is/was at least one bug on the secret processing code of some component that would cause such an error if the secret was *exactly* some even number of characters long. yep, that was it. Two minutes before your post arrived we found that it works for StuFF mc when he uses a much shorter secret than the 41 characters he had. :) I did some more tests and found that it fails when the password is 16 characters or longer, but couldn't find what's causing that limit when glancing over the respective areas in the router and mu-conference code. cu Reinhard -- To unsubscribe send a mail to jabberd2+unsubscr...@lists.xiaoka.com
Re: additional listen ports for c2s
Hi, On Wed, 10 Feb 2010 at 07:18, Tomasz Sterna wrote: iptables -t nat -A PREROUTING \ -p tcp --dport 80 -j REDIRECT --to-ports 5222 this works if jabberd listens on the wildcard address or on the primary IP address of the incoming interface. If it explicitly binds to a secondary address ($XMPPADDR), the DNAT target is needed, because it allows to specify the destination IP address for the redirection: iptables -t nat -A PREROUTING \ -d $XMPPADDR -p tcp -m tcp --dport 80 \ -j DNAT --to-destination $XMPPADDR:5222 cu Reinhard -- To unsubscribe send a mail to jabberd2+unsubscr...@lists.xiaoka.com
Re: jabberd-2.2.9 release
Hi, On Mon, 6 Jul 2009 at 10:12, Jorge Salamero Sanz wrote: is there any migration script to move from bd to sqlite ? debian packages have been using db for ages and we have to provide a migration path to move the users to sqlite. is there really a need to forcibly move existing users to sqlite just because the upstream default has changed? Installations are free to pick the storage backend they want and I think there is nothing wrong with keeping exising installations as db and only default new installations to sqlite. Or are there any plans to remove the db backend? cu Reinhard -- To unsubscribe send a mail to jabberd2+unsubscr...@lists.xiaoka.com
Re: Virtual domains for jabberd2
Hi, On Wed, 10 Jun 2009 at 16:03, Tomasz Sterna wrote: I would appreciate if you tried these and tell me what you think. I just tried it with jabberd 2.2.8.91 and found that on a freshly started jabberd, the first connection to each of the virtual domains fails and c2s logs SASL callback for non-existing host: ... with ... being the domain name I just tried to connect to. The second and any further connections work fine. c2s is configured with an empty id/ tag, in case that matters. cu Reinhard -- To unsubscribe send a mail to jabberd2+unsubscr...@lists.xiaoka.com