Logs for jdev

Show join/part/nick changes:

[00:38:14] * johnny left the chat.
[01:02:18] * johnny joined the chat.
[02:00:21] * MattJ left the chat.
[02:29:53] * johnny left the chat.
[02:30:05] * johnny joined the chat.
[06:23:10] * scippio left the chat.
[06:42:58] * rizinbox\40jabber.org joined the chat.
[06:58:49] * mlundblad_laptop joined the chat.
[07:04:55] * Kev joined the chat.
[07:09:20] * jadahl joined the chat.
[07:35:25] * jadahl left the chat.
[07:47:33] * Guus joined the chat.
[07:57:20] * Guus left the chat.
[07:57:48] * Guus joined the chat.
[07:57:53] * teo1 left the chat.
[07:57:55] * teo1 joined the chat.
[08:00:53] * ermine joined the chat.
[08:04:02] * loxs joined the chat.
[08:04:03] * loxs left the chat.
[08:05:39] * loxs joined the chat.
[08:05:39] * loxs left the chat.
[08:08:59] * jonas joined the chat.
[08:09:19] * loxs joined the chat.
[08:09:55] <loxs> does sleekxmpp do any async stuff? will I have troubles if I use it with non-async python libraries?
[08:36:47] * Tobias joined the chat.
[08:37:42] * Tobias left the chat.
[08:47:49] * Asterix joined the chat.
[09:21:22] * scippio joined the chat.
[09:23:30] * petermount joined the chat.
[09:33:23] * Tobias joined the chat.
[09:34:00] * Tobias left the chat.
[09:34:19] * Tobias joined the chat.
[09:36:23] * Treebilou joined the chat.
[09:41:00] * nabatt joined the chat.
[09:56:50] * Kev left the chat.
[09:57:03] * Kev joined the chat.
[10:01:12] * Guus left the chat.
[10:08:38] * MattJ joined the chat.
[10:11:59] * Guus joined the chat.
[10:16:45] * Treebilou left the chat.
[10:36:52] * Tobias left the chat.
[10:37:05] * Tobias joined the chat.
[10:37:20] * Tobias left the chat.
[10:37:50] * Tobias joined the chat.
[10:49:50] * rizinbox\40jabber.org left the chat.
[10:52:38] * Tobias left the chat.
[11:06:18] * rizinbox\40jabber.org joined the chat.
[11:20:56] * abhinavsingh joined the chat.
[11:22:21] * petermount left the chat.
[11:22:37] * petermount joined the chat.
[11:50:51] * Zash joined the chat.
[12:02:36] * abhinavsingh left the chat.
[12:06:35] * Florob joined the chat.
[12:11:09] * rizinbox\40jabber.org left the chat.
[12:11:33] * rizinbox\40jabber.org joined the chat.
[12:22:20] * abhinavsingh joined the chat.
[12:30:52] * Hugin joined the chat.
[12:38:00] * test joined the chat.
[12:38:25] * test left the chat.
[13:01:05] * Guus joined the chat.
[13:01:34] * Neustradamus joined the chat.
[13:06:01] * petermount joined the chat.
[13:06:58] * abhinavsingh joined the chat.
[13:07:03] * mlundblad_laptop joined the chat.
[13:07:15] * elmex joined the chat.
[13:11:08] * alfeberlin joined the chat.
[13:11:16] * alfeberlin left the chat.
[13:17:53] * tofu joined the chat.
[13:18:00] * loxs joined the chat.
[13:24:44] * Florob joined the chat.
[13:32:51] * abhinavsingh left the chat.
[13:40:05] * zanchin joined the chat.
[13:41:13] * abhinavsingh joined the chat.
[13:45:51] * scippio joined the chat.
[13:52:52] * abhinavsingh left the chat.
[13:53:03] * abhinavsingh joined the chat.
[13:56:02] * scippio left the chat.
[14:01:39] * tofu left the chat.
[14:04:52] * abhinavsingh left the chat.
[14:05:50] * abhinavsingh joined the chat.
[14:07:00] * scippio joined the chat.
[14:12:11] * Asterix joined the chat.
[14:15:22] * abhinavsingh left the chat.
[14:19:15] * Asim joined the chat.
[14:21:26] <Asim> hi all
[14:22:33] * Asim left the chat.
[14:28:34] * loxs left the chat.
[14:51:26] * scippio left the chat.
[14:51:44] * scippio joined the chat.
[15:03:07] * petermount left the chat.
[15:10:52] * Guus left the chat.
[15:12:53] * hawke joined the chat.
[15:14:53] * Link Mauve joined the chat.
[15:26:57] * deryni joined the chat.
[15:31:56] * Zash joined the chat.
[15:50:28] * scippio left the chat.
[15:52:16] * alfeberlin joined the chat.
[15:52:48] * MattJ joined the chat.
[15:53:04] * alfeberlin left the chat.
[15:53:46] * louiz’ joined the chat.
[15:54:20] <louiz’> hum. What was the "you must use a reserved nickname" error, on all @conference.jabber.org rooms? :(
[16:00:52] * mlundblad_laptop left the chat.
[16:08:12] * Tobias joined the chat.
[16:08:26] * Tobias left the chat.
[16:13:28] * teo1 joined the chat.
[16:14:21] * alfeberlin joined the chat.
[16:19:32] * Tobias joined the chat.
[16:28:54] * scippio joined the chat.
[17:29:45] * Florob left the chat.
[17:30:05] * Florob joined the chat.
[17:48:53] * Neustradamus left the chat.
[17:57:32] * smoku joined the chat.
[17:57:40] * smoku left the chat.
[18:06:59] * smoku joined the chat.
[18:10:23] * jadahl joined the chat.
[18:10:47] * justin joined the chat.
[18:10:52] * jadahl left the chat.
[18:11:09] * jadahl joined the chat.
[18:20:24] * Florob left the chat.
[18:22:05] * Tobias left the chat.
[18:35:50] * johnny joined the chat.
[18:45:24] * Xificurk joined the chat.
[19:06:09] * scippio left the chat.
[19:10:26] * scippio joined the chat.
[19:11:29] * Kev joined the chat.
[19:21:20] * Tobias joined the chat.
[19:21:51] * mlundblad joined the chat.
[19:24:55] * Treebilou joined the chat.
[19:27:09] * scippio left the chat.
[19:36:32] * abhinavsingh joined the chat.
[19:36:45] * teo1 left the chat.
[19:36:46] * teo1 joined the chat.
[19:40:21] <abhinavsingh> hello... does anyone know of a good resource about replication/distribution of jabber services (MUC/PubSub)...?
[19:40:37] <abhinavsingh> anything specific to ejabberd will be a great help
[19:41:07] <MattJ> I guess the ejabberd docs are the best place to start, then
[19:42:11] <abhinavsingh> then i guess i missed the part where they talked abt replication.... i thought they have no doc in thr on the site....
[19:42:17] <justin> abhinavsingh: clustering of components?
[19:42:19] <abhinavsingh> sorry will hunt again...
[19:42:25] <abhinavsingh> yup even components
[19:42:40] <abhinavsingh> infact all kind of jabber services
[19:43:09] <MattJ> Well clustering is implementation dependent, so it's up to each service
[19:43:35] * Xificurk left the chat.
[19:43:38] <abhinavsingh> so in case of component if i configure 2 ejabberd instances for
[19:43:39] <MattJ> I guessed you meant ejabberd's MUC/pubsub modules, which I (perhaps wrongly) assumed they had docs for
[19:43:49] <abhinavsingh> component.domain
[19:44:18] <abhinavsingh> will my component written in php,python (single instance)
[19:44:31] <abhinavsingh> get the traffic from both ejabberd instance?
[19:44:42] <MattJ> I doubt it
[19:44:43] <abhinavsingh> or do i have to run multiple instances of my component code
[19:45:12] <MattJ> and that wouldn't be clustering then, since you'd have a single component being the point of failure
[19:45:28] * Xificurk joined the chat.
[19:45:35] <MattJ> but I have no experience of ejabberd's clustering, other than what I've been told
[19:46:15] <Tobias> and you've been only told nice things about it, right?
[19:46:22] <abhinavsingh> right.... in that sense it doesn't bring plus of clustering....
[19:46:25] * Xificurk left the chat.
[19:46:39] * Xificurk joined the chat.
[19:47:20] <MattJ> Tobias, I wasn't going to mention anything :)
[19:47:25] <abhinavsingh> so before i go ahead an look for ejabberd config docs
[19:47:37] <abhinavsingh> can i clear my clustering doubts as it works with xmpp servers
[19:47:42] <abhinavsingh> here guys?
[19:47:56] <Tobias> MattJ, you don't have to, we can read between the lines
[19:48:01] <MattJ> abhinavsingh, feel free to try and get doubts cleared here :)
[19:48:10] <abhinavsingh> cool man....
[19:48:25] <MattJ> though as I said, clustering is rather implementation-specific, so it depends what your questions are
[19:48:34] <abhinavsingh> so now suppose i have 2 instances of ejabberd running
[19:48:49] <abhinavsingh> both accepting muc.domain.com
[19:49:24] <abhinavsingh> so if a new room gets created does it gets replicated to both
[19:49:29] <justin> i am just now investigating tigase clustering. and i believe each node only routes to external components on that node. so if you want three tigase nodes to route to one master muc node, then you have to add extra routing components. tigase is not ejabberd, but maybe this is interesting information and ejabberd is similarly designed
[19:49:48] <abhinavsingh> and both room occupants see messages irrespected from where the user is connected....
[19:50:09] <MattJ> abhinavsingh, that's all up to ejabberd
[19:50:12] <deryni> I believe the room actually gets created on only one node, but that's invisible to service users under normal circumstances.
[19:50:22] <deryni> But you'd have to ask in the ejabberd room for confirmation.
[19:51:00] <MattJ> justin, that model makes some sense for external components, since more often than not they won't have a clue about clustering
[19:52:05] <abhinavsingh> MattJ, i have a component which sits in the background
[19:52:18] <abhinavsingh> transparent to the connected users on the browsers/mobile etc
[19:52:49] <abhinavsingh> this component basically is the entry point into my application
[19:52:55] <abhinavsingh> for the outside user
[19:53:07] <abhinavsingh> and hence as a result all muc/pubsub traffic flows
[19:53:17] <abhinavsingh> through this component directly
[19:53:58] <abhinavsingh> so i will have to have at some time clustering of components too....
[19:54:35] <abhinavsingh> but i think jabber servers might not be able to help directly
[19:54:40] <abhinavsingh> with component clustering....
[19:55:35] <abhinavsingh> say my component is bind to
[19:55:41] <abhinavsingh> component.domain.com
[19:56:22] <abhinavsingh> so is there any general strategy/workflow
[19:56:41] <abhinavsingh> which can help out how indivisual instances of component script can handle
[19:56:51] <abhinavsingh> traffic to component.domain.com
[19:57:12] <abhinavsingh> from clustered ejabberd servers...
[19:57:45] <MattJ> There are many ways of tackling the problem
[19:57:57] <MattJ> which, if any, ejabberd supports, I don't know
[19:59:31] <abhinavsingh> can you elaborate on an example for tackling such case....
[19:59:36] <abhinavsingh> i think
[19:59:42] <abhinavsingh> of only one solution
[19:59:49] <abhinavsingh> where a central component handling
[20:00:02] <MattJ> Well which options are valid depends on the component too
[20:00:05] <abhinavsingh> can route traffic to indivisual script
[20:00:18] <MattJ> if the component has no state, you can easily run a component on each node
[20:00:22] * jkhii joined the chat.
[20:00:47] <MattJ> if it has state, stored in a database, and no local state, you can still do the same if all components use the same database
[20:01:38] <MattJ> if it has local state you can do the same with help from your server, to ensure that each JID is statically paired with a given component for its lifetime
[20:02:47] <MattJ> In all of these cases you can have a single component on a single node and have the server route all traffic there
[20:03:03] <MattJ> If that node goes down though, you no longer have a path to the component
[20:03:22] <abhinavsingh> right.... as such my component don't have a state attached but i want to make sure each JID is pariied to same component for it's lifetime
[20:03:39] <MattJ> As an alternative you could have the same component process connect to all nodes
[20:03:54] <justin> MattJ: the strategy i'm about to implement is all based on external components but with the assumption that the xmpp server doesn't attempt to route across nodes to find external components on its own
[20:04:02] <justin> which is at least the case for tigase
[20:04:09] <MattJ> and finally, you could have multiple component processes connect to multiple nodes and implement your own clustering in the component
[20:04:18] <MattJ> and that concludes just about every tactic I can think of :)
[20:04:51] <abhinavsingh> this ofcourse opens up a lot of options in my mind.... :)
[20:05:02] <justin> have a routing component, one instance per xmpp server node, that shares the common component domain name
[20:05:21] <justin> and then N instances of real component logic instances, with distinct domains
[20:06:18] <justin> stanza path is: client -> server node -> routing component -> server node(s) -> real component
[20:06:31] <justin> and then the same path on the way back, i guess
[20:07:12] <abhinavsingh> i think having own clustering in the components seems ideal for my use case.... though it brings in some additional maintenance/monitoring pain with itself.....
[20:08:11] <MattJ> justin, that just serves to highlight that no single size fits all in this case :)
[20:08:22] <abhinavsingh> @justin more or less thats the path i have as of now.... except that i have no routing thing done as of now.....
[20:08:52] <abhinavsingh> and i think i will experiement with 2 local ejabberd instances on my local box
[20:09:32] <abhinavsingh> and see which strategy can be worked out best....
[20:09:41] <abhinavsingh> thanks a lot guys :D
[20:14:41] * jadahl left the chat.
[20:14:49] * deryni left the chat.
[20:18:59] * jonas joined the chat.
[20:24:44] * jkhii left the chat.
[20:32:53] * Zash left the chat.
[20:41:45] * scippio joined the chat.
[21:00:06] * dwd joined the chat.
[21:03:12] * Zash joined the chat.
[21:04:33] * scippio left the chat.
[21:06:16] * Florob joined the chat.
[21:06:36] * scippio joined the chat.
[21:07:01] * deryni joined the chat.
[21:12:16] <dwd> Florob, Hey.
[21:12:29] <Florob> dwd, Hy
[21:12:34] <dwd> Florob, Turns out I was wrong - Fippo pointed out transition-needed is §6.2.14
[21:12:44] <dwd> Florob, In the current bis draft.
[21:12:48] <Zash> dwd!
[21:12:53] <dwd> Zash!
[21:13:58] <Florob> dwd, current as in? The one I had deemed current doesn't even have a 6.2.14
[21:14:04] <Zash> dwd: Haven't seen you here in a while. wb :)
[21:14:23] <dwd> Zash, Ah, last week was half-term, I was looking after the kids. Before that had my head down a bit.
[21:15:20] <dwd> Florob, Meh. §6.5.12.
[21:15:47] <Florob> dwd, yes, found it. Just wondering if I'm still looking at something I shouldn't be looking at
[21:16:22] <dwd> Florob, http://tools.ietf.org/html/draft-ietf-xmpp-3920bis-18#section-6.5.12
[21:22:58] * Neustradamus joined the chat.
[21:23:14] <dwd> Florob, Although your last does make you sound like you're not looking at XMPP specs at *all*...
[21:24:02] <Florob> dwd, ?
[21:24:56] <dwd> Florob, "[...] looking at something I shouldn't be looking at" - like, I dunno, young women covered in chocolate sauce.
[21:25:31] <Florob> dwd, that sounds like something one should be looking at :P
[21:25:44] <Zash> :D
[21:26:15] <Kev> /me wanders in.
[21:26:30] <dwd> /me denies lowering the tone again.
[21:27:32] <Zash> dwd: If your head is more up, I wanted to bug you about ssl stuff ;)
[21:27:52] <dwd> Zash, You as well?
[21:28:17] <dwd> Zash, Had an evening of it. And, for that matter, two weeks of it before that.
[21:28:46] <Zash> Heh :)
[21:29:07] <dwd> Zash, Anyway, I don't know anything about SSL at all.
[21:29:38] * scippio left the chat.
[21:29:52] * scippio joined the chat.
[21:29:57] <dwd> Zash, I know a bit about TLS and X.509, but nothing much about SSL.
[21:31:30] <Zash> /me is unsure about the diff between SSL and TLS :/
[21:32:12] <jonas> SSL is the netscape technology, TLS is the IEFT (i think) successor
[21:32:22] <dwd> jonas, Right.
[21:32:41] <dwd> jonas, Well, IETF. But yeah.
[21:32:50] <Zash> Ah, so something like Jabber vs XMPP then?
[21:32:58] <dwd> Zash, Very similar, actually.
[21:33:20] <justin> yeah Indigenous Education Foundation of Tanzania is something else
[21:33:22] <dwd> Zash, Except that TLS isn't actually compatible with SSL, instead there's a fallback between protocols.
[21:33:41] <Kev> So the same as Jabber/XMPP, then.
[21:33:45] <Zash> :D
[21:34:52] <dwd> /me waits for Zash's question. And tries to get the rabbit in while Zash types.
[21:35:42] <Zash> Anyways, I digged up what I think is how you get openssl to include SRVName stuff, and generated this: http://q.zash.se/173f178b.txt
[21:38:41] <Zash> dwd: the question beeing, is it correct?
[21:44:23] <dwd> Zash, http://dave.cridland.net/xeps/zash-cert-2.png
[21:44:26] * Asterix left the chat.
[21:45:07] <Tobias> dwd, does isode's cert tool base on openssl or does it use own libs?
[21:45:15] <dwd> Tobias, Both.
[21:45:21] <Tobias> ah, ok
[21:46:07] <dwd> Tobias, TLS is OpenSSL, but we do additional stuff for certificate verification and CRL checking, plus we *also* have a whole bunch of Java code which does the pretty stuff, and the CA itself.
[21:46:29] <Tobias> the GUI tool is java based?
[21:47:12] <dwd> Tobias, Yes. ALthough not "proper" Java; it uses various JNI stuff to native code libraries.
[21:47:23] <dwd> Tobias, Thus upsetting both the Java *and* C++ camps. :-)
[21:47:26] <Zash> !rfc 4985
[21:47:29] <Zash> ...
[21:47:53] <Zash> !xmpp-core
[21:48:16] <Tobias> dwd, yeah, so coding on it is joy for any developer :D
[21:48:18] <dwd> Zash - [21:48:01] [Isode Unclassified Internal] XMPPBot: Dave Cridland: rfc4985: internet x.509 public key infrastructure subject alternative name for expression of service name (august 2007) See: http://tools.ietf.org/html/4985
[21:49:26] <dwd> Tobias, IIRC, the C++ stuff is mostly the X.500/X.400 code. I *think* the CA is pure Java. Not entirely sure, though, it might pull in the CRL code, which is an aberration of '90's C++ as I recall.
[21:49:35] <jonas> oh, JNI, how wonderful
[21:49:50] <dwd> jonas, Yes, so much so it brings a tear to the GUI team's eye.
[21:49:58] <dwd> jonas, Shortly before the howling starts.
[21:50:15] <jonas> at work i'm fighting with multiple threads, some in java, some in C, synchronizing and sharing data
[21:50:26] <jonas> it's quite a pain actually
[21:50:30] <Zash> http://tools.ietf.org/html/draft-ietf-xmpp-3920bis-18#section-13.7.1.2
[21:51:48] <Zash> Do I win? :D
[21:52:13] <dwd> Hmmm... XmppAddr is deprecated? So what SAN is a server supposed to use for an (end-user) client certificate?
[21:53:04] <Zash> Waitwat
[21:53:27] * smoku left the chat.
[21:53:28] <dwd> Oh, I see. THis is only for server certs.
[21:53:35] * smoku joined the chat.
[21:54:22] <dwd> Zash, But anyway, yes, you do win.
[21:54:28] <Zash> \o/
[21:58:57] * smoku left the chat.
[22:02:41] <Zash> the openssl config file: http://q.zash.se/92788861.txt
[22:10:34] * Zash left the chat.
[22:11:22] * Zash joined the chat.
[22:12:40] * jonas left the chat.
[22:12:47] * Tobias left the chat.
[22:14:10] * Tobias joined the chat.
[22:20:25] * Treebilou left the chat.
[22:20:55] * nabatt joined the chat.
[22:21:11] * nabatt left the chat.
[22:21:24] * nabatt joined the chat.
[22:26:21] * Hugin joined the chat.
[22:32:20] * Zash left the chat.
[22:34:06] * Zash joined the chat.
[22:37:23] * Zash left the chat.
[22:37:27] * Zash joined the chat.
[22:46:26] * Tobias left the chat.
[22:58:25] * mlundblad left the chat.
[23:03:27] * nabatt left the chat.
[23:21:25] * abhinavsingh left the chat.
[23:21:37] * hawke left the chat.
[23:41:02] * justin left the chat.
[23:41:03] * Hugin left the chat.