Logs for jdev@conference.jabber.org
[00:56:11] * Florob left the chat.
[02:57:02] * Tobias joined the chat.
[02:57:47] * Tobias left the chat.
[02:58:30] * jcea left the chat.
[03:02:32] <> xnyhps: About what? (I'm usually wrong)
[03:09:55] * Tobias joined the chat.
[03:11:15] * Tobias left the chat.
[03:21:55] * darkrain left the chat.
[04:04:31] * darkrain joined the chat.
[06:17:55] * waqas left the chat.
[06:18:16] * waqas joined the chat.
[06:19:25] * Asterix joined the chat.
[06:24:56] * ermine joined the chat.
[06:25:48] * Alex joined the chat.
[06:26:20] * waqas left the chat.
[06:58:10] * scippio left the chat.
[07:08:07] * Asterix left the chat.
[07:31:10] <> I'm aware of the log issue, and a fix is on the way. Might take a short while, mind.
[07:34:43] <> Found the other one in our tracker ;)
[08:09:34] <> The session initiation is defined in RFC 3921, and as far as I can tell, we respond exactly as defined in that RFC. It's optional
in 6121, and only as a backward compatibility option.
[08:16:21] <> Ah, but the actual complaint is that you sent it to the bare jid instead of the server, yet received an answer from the server.
[08:22:42] * bear joined the chat.
[08:26:45] * Kev joined the chat.
[08:30:53] * Flow joined the chat.
[08:47:30] <> darkrain: I believe you mentioned there were some weird situations where the from didn't follow the RFC during roster retrieval
etc. (about #15879)
[08:50:33] * Lance joined the chat.
[08:52:51] <> edwinm: well yes, stanzas without a 'to' (from the client) are sent to the user's own bare JID
[08:53:15] <> and thus, the result should come from that as well. It seems that both m-link and facebook reply as the server instead
[08:55:00] * ralphm joined the chat.
[08:55:56] * bear left the chat.
[08:56:17] * bear joined the chat.
[08:56:31] * 0xAFFE joined the chat.
[08:57:40] <> Looks to me like a bug in the client, as the client should be sending this to the server, not its own JID.
[08:57:47] <> At least, in 3920. I've not found a reference to it in 6120 yet.
[08:58:11] <> Kev: you don't have to provide addressing
[08:58:32] <> This implies addressing it to your own JID
[08:59:59] <> No, the RFC says "to the server"
[09:00:17] <> Which also matches the examples.
[09:01:27] <> 6121 references it, but not explicitly includes it.
[09:01:52] <> 6121 just says "Do like 3921 to be kind to old clients".
[09:02:01] <> Yup, in 1.4
[09:02:06] <> Kev: 3920, section 8.1.2.1 says "When the server generates a stanza on its own behalf for delivery to the client from the
server itself, the stanza MUST include a 'from' attribute whose value is the bare JID (i.e., <domainpart>) of the server as
agreed upon during stream negotiation (e.g., based on the 'to' attribute of the initial stream header). "
[09:02:16] <> sorry, 6120
[09:02:22] <> So it seems like M-Link has a minor bug that it replies from the correct JID when a client that sends obsolete protocol sends
to the wrong JID.
[09:02:29] <> ralphm: Yes, I'm aware of teh addressing rules.
[09:02:46] <> ralphm: But 3921 says that the session start is sent to the server, not to the user's own JID or unaddressed.
[09:02:49] <> but I don't think this is applicable
[09:02:56] <> because it is an iq reply
[09:03:43] <> so I think point 3. applies there instead
[09:03:57] <> "sent to the server" is quite ambiguous
[09:04:14] <> cause hey, it is not going to send it to something else
[09:06:21] <> And regarding the addressing being vague in 3920, this is why it was made more explicit in 6120.
[09:07:22] <> Kev: as for the 'obsolete protocol' part of your statement, I think it also goes for resource binding.
[09:08:12] <> But just for facebook, not m-link, apparently.
[09:11:18] * scippio joined the chat.
[09:12:42] <> ralphm: The text says to send to the server, the examples show it addressed to the server - I think the intention here is
clear.
[09:13:58] <> but there, the client specifies the to attribute to be the server
[09:14:17] * bear left the chat.
[09:14:20] <> So, you are arguing that the client is misbehaving?
[09:14:37] <> (I hadn't notice this before)
[09:15:29] * bear joined the chat.
[09:17:31] <> ralphm: that's how I interpreted the section from 3921.
[09:18:14] <> Kev: cause I'm not sure what the best approach is here. I mean, if clients (rightfully) expect unaddressed stanzas to be replied
to from their own bare JID, the reply from m-link would seem unsollicited
[09:19:52] <> However, replying with not-acceptable or feature-not-implemented when no 'to' was given, seems undesirable
[09:24:11] <> I think the desired response here is for clients to stop doing session start :)
[09:25:45] <> Are there any servers out there that still require the session start?
[09:28:33] <> too bad that's not as easy to test with xnyhps's stuff
[09:28:53] <> (as you need to actually be able to log in)
[09:31:59] * ralphm left the chat.
[09:32:10] * ralphm joined the chat.
[09:37:40] <> edwinm: Surely not?
[09:39:20] <> Kev: I'm sure the percentage is not 0
[09:39:30] <> in deployed servers, that is
[09:39:38] <> I wouldn't expect many out there that are only 3921 or even pre-RFC. Current servers shouldn't require it at all. But older
implementations just might
[09:41:33] <> I don't believe Swift/Swiften/Sluift/Stroke do it, although I haven't checked.
[09:41:58] <> The Twisted support will try session establishment if advertised
[09:42:39] <> Even if it's marked as optional?
[09:42:55] <> yes
[09:46:11] <> edwinm: because 3921 doesn't specify a required attribute for the feature, there's no way for a client to determine that
[09:46:22] <> I think
[09:49:28] <> Well, if the server sends back it supports it, and it is optional, you can probably get away with not sending it.
[09:50:05] <> In stream:features, I mean.
[09:50:46] <> "probably" is not a great way to design client
[09:50:46] <> s
[09:50:46] <> I lie, Swiften et al. do do it.
[09:52:11] <> Although, interestingly, we have the same bug where we send it to bare-JID, not to the server.
[09:52:45] <> Apparently, when we did 6120, it was thought not a really big problem:
"Informational Note: Because there is no generic format for indicating that a feature is mandatory-to-negotiate, it is possible
that a feature that is not understood by the initiating entity might be considered mandatory-to-negotiate by the receiving
entity, resulting in failure of the stream negotiation process. Although such an outcome would be undesirable, the working
group deemed it rare enough that a generic format was not needed."
[09:53:21] <> Fortunately, all clients know about this feature, so will happily negotiate it.
[09:53:34] * Zash joined the chat.
[09:53:55] <> Kev: Twisted also sends no 'to'
[09:54:16] * Philonous left the chat.
[09:54:42] * Tobias left the chat.
[09:55:54] <> How interesting.
[09:56:35] * Philonous joined the chat.
[09:59:10] * aman joined the chat.
[09:59:26] <> Makes me wonder how many clients actually check if the sender of replies actually matches for iqs.
[09:59:43] <> (Or have ids which are highly unpredictable.)
[10:00:33] <> I'd guess that clients only look for the id attribute
[10:00:38] <> this is potentially very bad
[10:01:03] <> I never considered that before
[10:01:11] <> depends on which clients.
[10:01:16] <> Oh dear
[10:01:39] <> where I said clients, I mean client libraries
[10:02:14] <> Kaminsky bug all over again?
[10:02:22] <> xnyhps: Oh, I think they largely check the from (at least, we do) properly, for stream setup. But this one gets special-cased
that stanzas before resource binding are all accepted.
[10:02:45] <> *For streams that are set up.
[10:02:55] <> Kev: I actually mean *all* iqs
[10:03:08] <> (Because every iq received before resource binding have to be from the server)
[10:03:20] <> (from in the 'originating at' not 'addressed from' sense)
[10:04:02] <> I'm aware that there are client libraries with significant bugs in this area with security vulnerabilities around iqs from
unexpected sources - this was one of the significant reasons I had to write Stroke instead of using Smack.
[10:04:17] <> But I assumed it was a special case for that library, rather than a general problem.
[10:04:52] <> I looked at Verse, it has that issue :(
[10:05:12] <> Mine too
[10:06:29] <> iq responses are tracked only by id
[10:06:41] <> well, I guess we have something to talk about tomorrow
[10:06:44] <> for stanzaio i'm checking that froms are some variation of full/bare/domain of the sender or receiver's jids
[10:07:04] <> but i dont think sleek does yet
[10:07:04] <> ralphm: Would you mind wikiing it up, please?
[10:08:53] <> Lance: I just took a brief look at sleek, and it seems it only tracks by id, too
[10:08:58] <> Zash: And with numerically increasing ids, I see?
[10:09:12] <> xnyhps: Yes
[10:10:05] <> Kev: added to the list
[10:10:24] <> xnyhps: yes, Twisted/Wokkel, too
[10:11:04] <> ralphm: Thank you.
[10:13:17] <> http://swift.im/git/swift/tree/Swiften/Queries/Request.cpp#n78 is the Swiften code in question.
[10:14:18] * Tobias joined the chat.
[10:18:44] * ralphm left the chat.
[10:23:56] * MattJ joined the chat.
[10:29:47] * Kev left the chat.
[10:40:57] * 0xAFFE left the chat.
[10:41:01] * 0xAFFE joined the chat.
[10:47:54] <> Was there a change to the belief as to correct behaviour in the bind/session result from address? (I did try looking, in 6120/6121
for indications as to whether bind/session was being sent to the server-for-the-server versus being send to the server-for-the-account
but didn't find anything one way or the other. I didn't think to look in 3920/3921 for that.)
[10:48:37] <> deryni: yes...
[10:49:16] <> deryni: for session est., in 3921 the examples show explicitly addressing to/from the server's JID
[10:49:38] <> deryni: so m-link is doing it is specified there,
[10:50:00] <> deryni: however, many clients send it without a 'to' attribute
[10:50:16] <> The bind in 3920 section 7 does not have a to.
[10:50:33] <> deryni: bind is different indeed
[10:50:48] <> Ah, ok.
[10:52:13] <> So facebook is wrong and M-Link is right? And sessions should just stop being done? =)
[10:52:34] <> deryni: so a couple conclusions from this: a) many clients send session est. to the wrong address (i.e. none, instead of the
server), b) servers respond to that anyway, but some (like m-link) with a from that is the server's JID, c) many client libraries
don't actually check sender JIDs on IQ responses
[10:52:36] <> Are sessions actually doing anything in modern servers?
[10:52:45] <> No
[10:52:55] <> it is a no-op
[10:53:11] <> Right
[10:53:21] <> Why was it put in the standard in the first place?
[10:53:28] <> Legacy
[10:53:29] <> but especially conclusing c) above here is really bad
[10:53:36] <> conclusion
[10:54:02] <> Zash: no, we thought it was needed, and later we backtracked on that
[10:54:42] <> ralphm: Isn't it legacy as of 6120 ?
[10:54:51] <> that's why session est. is no longer defined in 612*1*
[10:55:17] <> gotta go pack now
[10:55:18] <> s/6120/current RFCs/
[10:55:28] <> I mean, what was the purpose when it was first introduced (in 3920, presumably)
[10:55:52] <> Philonous: the idea was some kind of 'ending' of negotiation
[10:56:17] <> I see
[11:02:22] * aman left the chat.
[11:09:34] <> I don't know that I accept the argument that the examples in rfc392X are normative about the destination of the session iq
(but I'm certain that's not a useful discussion to have at this point).
[11:10:23] <> deryni: there will be no other authority
[11:10:38] <> Assuming that stance though the M-Link behaviour is a minor bug (in that it responds) but it has the right address which makes
pidgin wrong in this case. And yeah, (c) is what this release of pidgin was fixing that broke this.
[11:11:00] <> deryni: so assume this is the entire truth
[11:11:13] <> awesome
[11:11:22] <> (that you were fixing it)
[11:11:59] <> Yeah. I'm fine with that assertion about session negotiation (I just don't know that I agree with the argument behind it but
I know it isn't worth arguing about).
[11:12:00] <> deryni: please try what happens (for many server implementations) when you don't send session establishment at all
[11:12:49] <> A plugin was written to allow pidgin to connect to facebook and M-Link despite these issues. I'm going to see if it can be
made to simply elide sending the session stanza at all and see if that's enough to get pidgin connecting to M-Link.
[11:12:58] <> I suppose a workable fallback would be this: try to send presence directly after binding a resource. If that fails (error
back) you have to do session establishment
[11:13:10] <> I am curious about which servers the fallback would be needed
[11:13:23] <> I don't know that pidgin actually cares whether session works or not (or just sends it because it always has). I haven't looked.
[11:14:14] * deryni left the chat.
[11:14:54] * deryni joined the chat.
[11:18:00] * bear left the chat.
[11:18:48] * Lance left the chat.
[11:23:40] * Flow joined the chat.
[11:31:19] * Lance joined the chat.
[11:39:04] * Flow left the chat.
[11:39:43] * Flow joined the chat.
[11:45:11] * Flow left the chat.
[11:46:32] * Flow joined the chat.
[11:49:49] * MattJ left the chat.
[11:50:23] * Lance left the chat.
[12:04:22] * Lance joined the chat.
[12:04:42] * Lance left the chat.
[12:04:57] * Ahmad joined the chat.
[12:05:59] <> اذهب مسلم@conference.mo.pp.ru
[12:07:01] <> Join مسلم@conference.mo.pp.ru
[12:07:16] * Lance joined the chat.
[12:08:37] <> _join مسلم@conference.mo.pp.ru
[12:08:48] * bear joined the chat.
[12:09:09] <> ioin مسلم@conference.mo.pp.ru
[12:10:04] <> Hi
[12:10:28] <> commads all
[12:11:00] <> Help
[12:11:10] <> help
[12:11:33] <> Commands all
[12:11:56] <> Commands all?
[12:12:06] <> Ahmad: quit that.
[12:12:31] * Isodetosi joined the chat.
[12:13:11] <> This commands boots
[12:13:55] * Ahmad joined the chat.
[12:15:00] <> I want to free you from the bot help?
[12:15:31] <> I'm Ahmed from Syria
[12:15:52] <> This room is for discussion of xmpp-related development.
[12:15:59] <> ahmed - this is a room for talking about xmpp/jabber dev
[12:16:14] <> not a room for bots or talking about bots
[12:18:40] <> Do you have another room for a chat?
[12:20:29] <> not on this server
[12:20:38] <> I'm studying English and I want to so I learn to look for a room
[12:20:50] <> ah ok
[12:21:10] <> this room and this server are really about xmpp development and not for general chat
[12:22:24] <> Give me the address please
[12:22:59] <> I personally don't have any general chat server addresses to give
[12:24:06] <> Thank you @}->--
[12:24:23] <> your welcome
[12:25:25] <> :-( I am very sad, because I did not find what I'm looking for
[12:26:22] <> I searched a long time I did not find
[12:27:24] * Kev joined the chat.
[12:27:53] * Florob joined the chat.
[12:30:37] <> I want to help find a chat room to help me in learning English, please
[12:32:07] * Irdis left the chat.
[12:32:23] <> Does http://search.wensley.org.uk/ still work?
[12:33:56] <> Thank you
[12:47:51] * Ahmad left the chat.
[13:13:08] * ermine left the chat.
[13:15:05] * Kev left the chat.
[13:17:08] * Irdis left the chat.
[13:22:59] * Flow left the chat.
[13:24:35] * ermine joined the chat.
[13:24:58] * edwinm left the chat.
[13:24:58] * Isodetosi left the chat.
[13:24:59] * edwinm joined the chat.
[13:25:00] * Isodetosi joined the chat.
[13:26:58] * Flow left the chat.
[13:30:07] * ThurahT left the chat.
[13:30:07] * ThurahT joined the chat.
[13:30:21] * Isodetosi left the chat.
[13:36:40] * ralphm joined the chat.
[13:42:30] * edwinm left the chat.
[13:53:41] * stpeter joined the chat.
[14:01:36] * jabberjocke joined the chat.
[14:10:37] * Florob left the chat.
[14:27:11] * jcea joined the chat.
[14:51:58] * Flow joined the chat.
[15:06:54] * Flow left the chat.
[15:13:42] * Flow joined the chat.
[15:17:15] * Asterix joined the chat.
[15:26:24] * ralphm left the chat.
[15:27:41] * Flow left the chat.
[15:42:02] * KevWalke joined the chat.
[15:50:13] * deryni left the chat.
[15:58:42] * scippio left the chat.
[16:03:44] * waqas joined the chat.
[16:11:18] * darkrain_ joined the chat.
[16:14:39] * aman joined the chat.
[16:22:38] * stpeter left the chat.
[16:23:13] * Zash left the chat.
[16:23:55] * Kev joined the chat.
[16:31:42] * deryni joined the chat.
[16:41:15] * jabberjocke left the chat.
[16:46:03] * stpeter joined the chat.
[16:51:52] * bear left the chat.
[17:02:57] * Zash joined the chat.
[17:14:06] * Kev left the chat.
[17:27:58] * Tobias left the chat.
[17:34:48] * Lance left the chat.
[17:39:33] * jabberjocke joined the chat.
[17:47:57] * jabberjocke left the chat.
[18:05:53] * datallah joined the chat.
[18:17:28] * Lance joined the chat.
[18:30:44] * 0xAFFE left the chat.
[18:30:57] * 0xAFFE joined the chat.
[18:34:39] * Lance left the chat.
[18:55:39] * Flow joined the chat.
[19:00:39] * ermine left the chat.
[19:09:45] * Zash joined the chat.
[19:11:00] * Zash left the chat.
[19:15:07] * 0xAFFE left the chat.
[19:15:10] * 0xAFFE joined the chat.
[19:17:56] * Lance joined the chat.
[19:21:00] * jabberjocke joined the chat.
[19:22:10] <> iChat:
<iq type='get' id='purple24c740bb'><vCard xmlns='vcard-temp'/></iq>
<iq type='result' id='purple24c740bb' from='user@chat.company.com/Office'><vCard xmlns='vcard-temp'>...</vCard></iq>
[19:30:24] <> Err, what?
[19:31:43] * aman left the chat.
[19:34:09] <> Just another version of the from address mismatch problem.
[19:35:54] * Lance left the chat.
[20:17:49] * Lance joined the chat.
[20:34:49] * Lance left the chat.
[20:41:16] * Maranda joined the chat.
[20:51:43] <> Zash, responded to your comment btw (the MUC prevents to send PMs).
[20:55:25] * Maranda left the chat.
[21:06:32] * 0xAFFE left the chat.
[21:18:07] * Lance joined the chat.
[21:23:19] * Alex left the chat.
[21:36:47] * Asterix left the chat.
[21:37:22] * Lance left the chat.
[21:39:55] * Lance joined the chat.
[21:45:05] * deryni left the chat.
[21:47:49] * Kev joined the chat.
[21:48:56] <> deryni: On jabber.org?
[21:49:07] * MattJ joined the chat.
[21:53:24] * Florob joined the chat.
[21:54:37] * Kev left the chat.
[21:56:17] * scippio joined the chat.
[22:16:13] * Florob left the chat.
[22:17:08] * Lance left the chat.
[22:23:13] * datallah left the chat.
[22:42:33] * deryni joined the chat.
[22:43:56] <> That was iChat the server, not iChat the client.
[22:48:33] * stpeter left the chat.
[22:50:22] * stpeter joined the chat.
[22:58:59] * Flow left the chat.
[22:59:13] * Flow joined the chat.
[23:01:48] * datallah joined the chat.
[23:12:05] * Zash left the chat.
[23:13:22] * Tobias joined the chat.
[23:15:27] * Flow left the chat.
[23:27:07] <> deryni: That's a new level of wrong.
[23:30:36] <> I must admit to being a bit surprised at how wrong that was.
[23:31:30] <> Makes me wonder how that's going to affect MUCs.
[23:31:48] <> For clients that request avatars for participants.
[23:48:56] <> I've never paid attention to how that works exactly, why would this affect that?
[23:57:04] * MattJ left the chat.
[23:58:22] <> Because those use JIDs where every participant is a different resource.
[23:58:59] <> Probably not relevant, but something to keep in mind if you allow stuff like that ^.