Logs for jdev

Show join/part/nick changes:

[00:06:02] * Philonous left the chat.
[00:21:30] * Lance left the chat.
[00:24:33] * Lance joined the chat.
[01:14:45] * singpolyma joined the chat.
[01:15:02] * darkrain_ left the chat.
[01:17:06] <singpolyma> I assume this has been discussed before, so feel free to point me at resources, but I've been thinking it would probably be useful on mobile devices to have a way to ask the XMPP server hold certain kinds of stanzas (presence notifications, most PEP) for delivery until requested. Basically just "pause these" and then "unpause these". That way the radio isn't waking up for people going online/offline when the user isn't even looking at the app.
[01:27:58] * Lance left the chat.
[01:30:41] * Lance joined the chat.
[01:33:23] * psa joined the chat.
[01:58:18] <Lance> singpolyma: yes, there was http://xmpp.org/extensions/xep-0273.html for dealing with that. i'm not sure about the state of server support for it though
[01:59:20] <Lance> you can also use privacy lists, i think, to block out presence updates
[02:00:22] <Lance> but for pep, you should just turn remove the +notify for features when you don't want the updates
[02:12:38] * Tobias_ joined the chat.
[02:15:01] <singpolyma> Lance: thanks :) if I turn off +notify, wouldn't the server have to re-query to know that I had done so?
[02:15:44] <Lance> yes, but only once
[02:16:01] <Lance> it should cache it afterward
[02:16:42] * Tobias left the chat.
[02:18:53] <singpolyma> but how would it know to re-query? Unless I also use the capability string thing, I guess
[02:19:23] <Lance> right, you'd have to recalculate your caps hash and send a presence update with it
[02:19:33] <Lance> it actually wouldn't be a bad idea for a server to just go ahead and compute the caps value for a given disco set with all of the +notify suffixes removed
[02:20:20] <Lance> if we make/update an informational xep about dealing with mobile clients, i think that should be added
[04:06:22] * psa left the chat.
[04:20:23] * Lance left the chat.
[04:27:19] * Lance joined the chat.
[04:47:31] * Lance left the chat.
[05:21:42] * Lance joined the chat.
[05:26:12] * vorner joined the chat.
[05:30:08] * Lance left the chat.
[05:47:32] * Alex joined the chat.
[05:47:45] * Alex left the chat.
[05:59:02] * vorner left the chat.
[06:11:40] * darkrain left the chat.
[06:11:40] * darkrain joined the chat.
[06:22:39] * Tobias_ left the chat.
[06:35:42] * ermine joined the chat.
[06:37:12] * Zash joined the chat.
[06:44:04] * Tobias joined the chat.
[06:47:39] * Tobias_ joined the chat.
[06:53:01] * Tobias left the chat.
[06:56:15] * Guus joined the chat.
[06:57:17] * vorner joined the chat.
[07:05:34] * Neustradamus left the chat.
[07:08:18] * Neustradamus joined the chat.
[07:14:19] * 0xAFFE joined the chat.
[07:19:57] * Alex joined the chat.
[07:20:56] * Neustradamus left the chat.
[07:24:42] * Neustradamus joined the chat.
[07:24:56] * Asterix joined the chat.
[07:26:48] * Schnouki joined the chat.
[07:32:28] * Lloyd joined the chat.
[07:39:21] * Bunneh left the chat.
[07:43:23] * jabberjocke left the chat.
[07:44:08] <dwd> singpolyma, Also google:queue. Let me sort that now.
[07:47:54] * Zash left the chat.
[07:54:10] <Kev> Maybe we should rename google:queue?
[08:04:22] * Tobias joined the chat.
[08:07:16] <dwd> Kev, Existing deployed client implementations, though.
[08:07:18] * Tobias_ left the chat.
[08:07:30] <Kev> Pfft.
[08:07:51] <dwd> Kev, It'd be nice to do so, I agree, and also change it from an <iq/> switch to a top-level element (since it is, after all, a stream mechanism).
[08:08:29] * jmedev joined the chat.
[08:10:19] <Kev> I don't see much value in not using iq, here.
[08:10:21] <Kev> Stream mechanism or not.
[08:10:39] <dwd> When would you ever want to route the exchange? WHat would that even mean?
[08:10:58] * Zash joined the chat.
[08:11:06] <Kev> Why would not wanting to route it mean it shouldn't be an iq?
[08:11:58] <dwd> I'd rather keep things which are purely stream mechanics out of routable elements.
[08:12:21] <dwd> http://www.packetizer.com/xml/papers/xml_compression_results.pdf seems very interesting, BTW.
[08:15:33] <Kev> Isn't the pretty graph suggesting that there's not much in it between EXI and zlib for compressing all but the non-trivial examples?
[08:16:26] <Alex> Kev: I think so
[08:16:55] <Alex> whats missing here is memory and cpu consumption, but I think zlib does a very good job
[08:18:13] <Alex> and I still think if you combine zlib with some predefined dictionary for xmpp it would be awesome
[08:18:54] <Alex> xzlib ;-)
[08:20:16] <Kev> all but the trivial examples, I meant.
[08:20:42] <Kev> Alex: The question there is whether zlib /needs/ a predefined dictionary for xmpp.
[08:30:32] * spencer.macdonald joined the chat.
[08:30:34] * spencer.macdonald left the chat.
[08:31:52] * MattJ joined the chat.
[08:32:11] * spencer.macdonald joined the chat.
[08:41:36] * jmedev left the chat.
[08:50:26] <dwd> The graph is suggesting that as the size of the XML to be compressed grows, the difference between pure zlib and EXI (which, remember, has a zlib stage in it) erodes pretty rapidly.
[08:51:25] * ermine left the chat.
[08:51:42] * ermine joined the chat.
[08:52:44] <Kev> And that it barely needs to grow to any size at all.
[08:52:52] <dwd> Right, about 1k.
[08:54:11] <dwd> I've a feeling that we could gain better compression if we used the recently received stanzas to inform the compression algorithm, actually, but I think it'd be painful to implement.
[08:54:26] <Kev> Doesn't zlib do that naturally?
[08:54:43] <dwd> No, it's a purely unidirectional algorithm.
[08:55:31] <dwd> I'd talking about using what amounts to a combination of '198 and '138 to use both the recently sent *and* recently acknowledged stanzas as "dictionary" fodder.
[08:55:39] <Kev> Oh, you mean using the B->A stanzas to poison the A->B dictionary?
[08:55:42] <Kev> Right.
[08:56:01] <Kev> I wonder how necessary that is, though.
[08:56:17] <dwd> I don't know. But I can't be bothered even trying it.
[08:56:54] <dwd> FWIW, you *can* get really quite impressive compression doing something similar with email - using the message you're replying to as a dictionary for the message you're sending.
[08:57:30] <dwd> (Given that the MTA can access the "dictionary message" via URLAUTH, this is actually relatively simple to implement, even)
[09:05:24] * Beanow_ joined the chat.
[09:25:23] * Neustradamus left the chat.
[09:26:36] * Neustradamus joined the chat.
[09:32:37] * spencer.macdonald left the chat.
[09:42:07] * spencer.macdonald joined the chat.
[09:42:09] * spencer.macdonald left the chat.
[09:42:32] * Philonous joined the chat.
[09:43:42] * spencer.macdonald joined the chat.
[09:56:47] * spencer.macdonald left the chat.
[09:57:09] * spencer.macdonald joined the chat.
[09:57:11] * spencer.macdonald left the chat.
[09:57:21] * spencer.macdonald joined the chat.
[09:57:23] * spencer.macdonald left the chat.
[10:02:23] * spencer.macdonald joined the chat.
[10:03:06] * Beanow_ left the chat.
[10:16:06] * spencer.macdonald left the chat.
[10:28:41] * Tobias left the chat.
[10:31:37] * Beanow_ joined the chat.
[10:32:29] * Beanow_ left the chat.
[10:33:23] * Asterix left the chat.
[10:40:53] * spencer.macdonald joined the chat.
[10:40:55] * spencer.macdonald left the chat.
[10:45:32] * spencer.macdonald joined the chat.
[10:46:53] * spencer.macdonald left the chat.
[10:49:02] * spencer.macdonald joined the chat.
[11:04:02] * Zash left the chat.
[11:04:24] * Zash joined the chat.
[11:12:28] * Tobias joined the chat.
[11:37:25] * spencer.macdonald left the chat.
[12:02:05] * jcea joined the chat.
[12:02:42] * Tobias left the chat.
[12:07:46] * spencer.macdonald joined the chat.
[12:07:56] * spencer.macdonald left the chat.
[12:10:19] * spencer.macdonald joined the chat.
[12:14:13] * spencer.macdonald left the chat.
[12:22:52] * Tobias joined the chat.
[12:35:21] * deryni joined the chat.
[12:36:12] * bear left the chat.
[12:37:25] * spencer.macdonald joined the chat.
[12:37:26] * spencer.macdonald left the chat.
[12:37:26] * spencer.macdonald joined the chat.
[12:37:26] * spencer.macdonald left the chat.
[12:37:28] * spencer.macdonald joined the chat.
[12:37:30] * spencer.macdonald left the chat.
[12:37:34] * spencer.macdonald joined the chat.
[12:37:35] * spencer.macdonald left the chat.
[12:37:37] * spencer.macdonald joined the chat.
[12:37:38] * spencer.macdonald left the chat.
[12:41:38] * spencer.macdonald joined the chat.
[12:46:42] * naw joined the chat.
[12:55:01] * deryni left the chat.
[13:09:37] * vorner left the chat.
[13:10:09] * vorner joined the chat.
[13:12:05] * deryni joined the chat.
[13:18:37] * spencer.macdonald left the chat.
[13:19:27] * spencer.macdonald joined the chat.
[13:19:28] * spencer.macdonald left the chat.
[13:24:19] * spencer.macdonald joined the chat.
[13:33:58] * naw left the chat.
[13:48:50] * xnyhps’ joined the chat.
[13:53:34] * xnyhps’ left the chat.
[13:55:50] * scippio left the chat.
[13:59:50] * 0xAFFE left the chat.
[14:00:27] * 0xAFFE joined the chat.
[14:05:48] * xnyhps’ joined the chat.
[14:09:41] * xnyhps’ left the chat.
[14:10:46] * xnyhps’ joined the chat.
[14:14:18] * mathieui joined the chat.
[14:22:42] * xnyhps’ left the chat.
[14:23:04] * xnyhps’ joined the chat.
[14:28:09] * spencer.macdonald left the chat.
[14:29:23] * spencer.macdonald joined the chat.
[14:29:24] * spencer.macdonald left the chat.
[14:30:36] * Neustradamus left the chat.
[14:32:06] * spencer.macdonald joined the chat.
[14:33:18] * xnyhps’ left the chat.
[14:34:10] * spencer.macdonald left the chat.
[14:34:32] * Zash left the chat.
[14:39:40] * Neustradamus joined the chat.
[14:44:54] * Alex left the chat.
[14:45:06] * darkrain_ joined the chat.
[14:57:59] * Neustradamus left the chat.
[14:59:38] * Neustradamus joined the chat.
[15:05:35] * Zash joined the chat.
[15:12:00] * psa joined the chat.
[15:16:52] * vorner left the chat.
[15:18:37] * spencer.macdonald joined the chat.
[15:18:37] * spencer.macdonald left the chat.
[15:20:59] <dwd> Oh, hey. I just realised something. At the IETF in Berlin, all those IETF folks wanting to join the chatrooms might not be able to if Google Talk's S2S drops by then.
[15:22:00] * spencer.macdonald joined the chat.
[15:22:21] <MattJ> Why do IETF folks use gmail for XMPP?
[15:23:05] <MattJ> and I have a feeling s2s isn't going away so soon, but then my worry is that it's going to stay half-alive for the foreseeable future
[15:23:16] <dwd> Well, yes - those IETF folk currently using gmail - that's certainly a reasonable percentage.
[15:23:26] <dwd> (Though insignificant by Google's reckoning, of course)
[15:23:42] <MattJ> Everything is insignificant by Google's reckoning ;)
[15:23:44] <dwd> It may well languish around for ages, yes.
[15:24:17] <Link Mauve> MattJ, for users of the official android client, s2s support has already been dropped.
[15:24:29] <Link Mauve> They don’t see me anymore in their roster.
[15:24:48] <MattJ> Right, but that's a client thing
[15:24:58] <MattJ> Do you see them btw?
[15:25:00] <Link Mauve> I guess those that use a non-Google XMPP client have the longest support possible.
[15:25:11] <psa> dwd: yes, I had that thought a few days ago
[15:25:20] <Link Mauve> I think I see them, but that may be a gmail bug.
[15:25:27] <Link Mauve> They never receive my messages though.
[15:26:11] <MattJ> Since it's 1999 again, who is working on the Hangouts gateway? :)
[15:26:15] <Zash> Let's just enforce TLS and be done wit hit
[15:26:18] <dwd> OK, so MattJ wins. S2S hasn't gone away, it's just broken.
[15:27:45] <Zash> Fun
[15:30:34] * fippo joined the chat.
[15:31:08] <psa> MattJ: I'd prefer not to go down the gateways route again
[15:31:29] <dwd> psa, Party like it's 1999?
[15:31:42] <psa> dwd: you betcha!
[15:31:50] <Kev> Yay, I'm back at Uni. Neato.
[15:32:01] <Kev> Back in my undergrad, no less.
[15:32:11] * spencer.macdonald left the chat.
[15:32:15] <psa> heh
[15:32:47] <MattJ> psa, gateways are a great idea, but in practice I think they fail in usability
[15:32:51] <MattJ> for some reason
[15:33:17] <Tobias> reason is mostly roster sync i think
[15:35:34] <singpolyma> biggest issue I used to have with gateways was removing a gateway and not having the roster items deleted
[15:36:02] * fippo left the chat.
[15:37:06] * fippo joined the chat.
[15:37:12] * spencer.macdonald joined the chat.
[15:37:36] <dwd> fippo, Kicking you was strangely satisfying.
[15:37:45] <Kev> I don't find that at all strange :D
[15:38:38] * spencer.macdonald left the chat.
[15:39:05] * fippo left the chat.
[15:39:07] * fippo joined the chat.
[15:40:30] * fippo left the chat.
[15:41:49] * fippo joined the chat.
[15:42:24] <fippo> i find it strange that m-link no longer seems to mirror the id in the presence returned
[15:42:47] <fippo> kev: next time i will poke you instead
[15:42:54] <Kev> Thanks!
[15:43:07] <Kev> fippo: Which id wherewhatwhywho?
[15:43:08] * Tobias left the chat.
[15:43:10] <dwd> fippo, That's a vague poitn in XEP-0045.
[15:43:25] <dwd> fippo, But in any case, you can tell which one is yours by the status code.
[15:43:59] <fippo> dwd: there is no security in that again forced joins. hence psyc has always required a kind of token that i really wanted to join
[15:44:32] <dwd> fippo, I'd be interested in hearing more about the threat model, there.
[15:44:32] <fippo> and putting that into the id of the presence when joining used to work.
[15:45:07] <fippo> old irc fears about server ops sending you a JOIN and making your client join a room
[15:46:09] <fippo> so back to the reason i originally joined...
[15:46:51] <fippo> it seems that messages are still received processed at google talk. when users switch back to the gmail chat the messages sent to them are part of the history
[15:47:23] <dwd> Oh, nice.
[15:49:04] * spencer.macdonald joined the chat.
[15:49:05] * spencer.macdonald left the chat.
[15:50:15] * spencer.macdonald joined the chat.
[15:50:55] <MattJ> fippo, I disagree with dwd about the id thing I believe, join Prosody MUCs instead ;)
[15:51:06] * vorner joined the chat.
[15:51:17] <MattJ> I think id preservation is important
[15:51:40] <MattJ> though I hadn't considered your issue before
[15:51:46] <fippo> mattj: now that works as expected (-:
[15:52:18] <MattJ> We had a discussion about it, perhaps on the list, and perhaps in the context of pubsub
[15:52:25] <MattJ> but my memory is vague
[15:54:44] <fippo> kev: that idea might be neat for FMUC btw
[15:54:58] <Kev> Preserving IDs?
[15:55:05] <fippo> yeah
[15:55:59] <Kev> Why for FMUC in particular?
[15:56:12] <fippo> so you can join x@y, get a "you're in" from x@z and know this is related to your initial join
[15:56:45] <Kev> But that never happens in the FMUC model - you only get stuff from the room that you joined.
[15:57:15] <fippo> ah... you designed around the issue (-:
[16:02:16] * StackOverflow joined the chat.
[16:06:27] * fippo left the chat.
[16:18:18] * Lance joined the chat.
[16:23:50] <dwd> MattJ, FWIW, you don't disagree with me, as such. I thought the correct method leant toward new ids, given the duplication - but since then I've seen various use-cases that simply won't work if the fan-out stanzas get new ids.
[16:24:08] <MattJ> Yay
[16:24:25] <MattJ> I guess I failed in thinking enough up to convince you back then :)
[16:30:58] * spencer.macdonald left the chat.
[16:31:25] <dwd> mathieui, I still think it's a hack. But I now think it's needed...
[16:31:40] <dwd> Ooops. I should check that tab expansion before hitting return.
[16:32:00] * Lloyd left the chat.
[16:32:32] <mathieui> yeah, it's a hack :p
[16:36:08] * mojomx joined the chat.
[16:39:18] * mojomx left the chat.
[16:41:00] * fippo joined the chat.
[16:44:12] <ermine> is there any good algorythm for client library to process stream features with given user's list of features?
[16:45:37] <ermine> or it is enough to try complete any extra features between tls and sasl negotations
[16:47:00] * Zash left the chat.
[16:47:05] * Zash joined the chat.
[16:47:12] <dwd> In general, encrypt, compress, authenticate - anything else is generally after that.
[16:48:27] <ermine> between sasl and bind?
[16:49:25] <fippo> has compress still the funny rule that you must be authenticated?
[16:49:57] <fippo> (even though that probably only matters when you're a server)
[16:50:00] <dwd> ermine, Depends on the extension.
[16:50:14] <dwd> fippo, Does it have that rule? If so, I've forgotten about it...
[16:52:07] <ermine> dwd: what xep feature has such rule?
[16:52:27] <ermine> dwd: to be placed in specific order
[16:52:50] <dwd> ermine, XEP-0198 defines a stream feature, and tells you when and how to negotiate it.
[16:52:56] <dwd> ermine, For example.
[16:53:29] * mojomx joined the chat.
[16:54:15] <fippo> dwd: it seems the problem with 0138 is in 0170
[16:56:41] * mojomx left the chat.
[16:57:26] <ermine> dwd: rfc6120 describes that you can use extra feature in any place - before mandatory ones or after them if they remains in stream:features
[17:02:00] <fippo> ermine: iirc on the server side the stragety to only offer what is negotiatable works quite well... mattj/dwd/kev: do prosody/m-link use something like that, too?
[17:02:23] <dwd> Yes, that being the only sane thing to do.
[17:04:18] <ermine> heh
[17:08:03] <ermine> it is easy for server
[17:10:25] <fippo> well, iff you assume the server does that, you can sort the list according to your own preferences and pick the first item
[17:12:03] <ermine> what about extra feature that could be mandatory?
[17:12:14] * xnyhps left the chat.
[17:13:08] <fippo> if it is mandatory to negotiate at that point, the server should not offer any other features
[17:15:41] * fippo left the chat.
[17:16:04] <ermine> ok
[17:18:07] * jcea left the chat.
[17:18:11] * fippo joined the chat.
[17:18:26] * jcea joined the chat.
[17:19:04] * xnyhps’ joined the chat.
[17:19:26] * 0xAFFE left the chat.
[17:23:39] * xnyhps’ left the chat.
[17:29:10] * fippo left the chat.
[17:29:12] * fippo joined the chat.
[17:40:27] * Schnouki left the chat.
[17:43:44] * psa left the chat.
[17:49:52] * Asterix joined the chat.
[17:52:51] * whatever joined the chat.
[18:00:23] * vorner left the chat.
[18:10:10] * bear joined the chat.
[18:18:46] * Schnouki joined the chat.
[18:19:18] * Schnouki left the chat.
[18:19:28] * Schnouki joined the chat.
[18:19:28] * Schnouki left the chat.
[18:25:05] * naw joined the chat.
[18:44:45] * fippo left the chat.
[18:47:00] * vorner joined the chat.
[18:49:11] * chris joined the chat.
[18:58:44] * chris left the chat.
[19:05:19] * xnyhps’ joined the chat.
[19:11:42] * xnyhps joined the chat.
[19:30:01] * xnyhps’ left the chat.
[19:32:29] * spencer.macdonald joined the chat.
[19:32:33] * spencer.macdonald left the chat.
[19:33:36] * Lance left the chat.
[19:33:39] * spencer.macdonald joined the chat.
[19:43:46] <singpolyma> Why do MUC messages have type=groupchat? Is it just weird historical leftovers?
[19:45:29] <dwd> It's so message stanza handling on the server can deal with them differently.
[19:45:33] <dwd> singpolyma, FWIW, I rather wish that pubsub messages had a different type, too.
[19:46:17] <singpolyma> But the server knows that the message is to=a muc compontent
[19:47:34] <Zash> Not if it's from a muc
[19:47:39] <Zash> and back to you
[19:48:24] <singpolyma> sure, but then the client knows it's a message from a chatroom it has joined
[19:48:26] <dwd> singpolyma, No, the server could find out, but it doesn't know, unless it happens to be a local built-in MUC component.
[19:48:44] <singpolyma> if it's not local, why does the server care?
[19:48:47] <dwd> singpolyma, And the client knows, but only if it's that session - and the server won't know.
[19:49:03] <dwd> singpolyma, For when the client goes offline suddenly, for example.
[19:51:28] * Lance joined the chat.
[20:08:29] * Lance left the chat.
[20:14:07] * Lance joined the chat.
[20:14:09] * spencer.macdonald left the chat.
[20:19:03] * spencer.macdonald joined the chat.
[20:24:19] * pidgin joined the chat.
[20:24:22] * pidgin left the chat.
[20:27:18] * pidgin joined the chat.
[20:31:31] * pidgin left the chat.
[20:31:40] * pidgin joined the chat.
[20:31:46] * pidgin left the chat.
[20:44:18] * Tobias joined the chat.
[20:44:26] * spencer.macdonald left the chat.
[20:45:20] * deryni left the chat.
[20:49:18] * spencer.macdonald joined the chat.
[21:24:31] * Florob joined the chat.
[21:26:11] * Florob left the chat.
[21:28:47] * Guus left the chat.
[21:29:18] * vorner left the chat.
[21:31:49] * jcea left the chat.
[21:32:02] * spencer.macdonald left the chat.
[21:55:05] * deryni joined the chat.
[21:59:47] * StackOverflow left the chat.
[22:00:22] * xnyhps left the chat.
[22:00:49] * Asterix left the chat.
[22:01:30] * xnyhps joined the chat.
[22:03:36] * Neustradamus left the chat.
[22:05:04] * Neustradamus joined the chat.
[22:19:22] * naw left the chat.
[22:34:55] * Neustradamus left the chat.
[22:36:10] * Neustradamus joined the chat.
[22:41:01] * bear left the chat.
[22:50:19] * boothj5 joined the chat.
[22:55:13] * Lance left the chat.
[22:55:38] * Lance joined the chat.
[22:57:32] * Lance left the chat.
[22:57:45] * Lance joined the chat.
[22:59:29] * bear joined the chat.
[23:06:14] * boothj5 left the chat.
[23:21:26] * jcea joined the chat.
[23:26:36] * Lance left the chat.
[23:26:48] * Lance joined the chat.
[23:37:12] * MattJ left the chat.
[23:38:07] * Zash left the chat.
[23:45:26] * pidgin joined the chat.
[23:45:29] * pidgin left the chat.
[23:56:43] * pidgin joined the chat.
[23:56:46] * pidgin left the chat.