Monday, February 12, 2018
xsf@muc.xmpp.org
February
Mon Tue Wed Thu Fri Sat Sun
      1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
       
             
XSF Discussion | Logs: http://logs.xmpp.org/xsf/ | Agenda https://trello.com/b/Dn6IQOu0/board-meetings

[00:12:19] *** ralphm shows as "online"
[00:19:08] *** jubalh has left the room
[00:19:09] *** jubalh has joined the room
[00:21:44] *** Steve Kille shows as "online" and his status message is "At Home"
[00:26:25] *** remko has joined the room
[00:26:26] *** remko shows as "away"
[00:37:14] *** Dave Cridland has left the room
[00:37:14] *** Dave Cridland has joined the room
[00:37:33] *** Steve Kille shows as "away" and his status message is "At Home"
[00:37:50] *** Dave Cridland has left the room
[00:37:59] *** Dave Cridland has joined the room
[00:53:07] *** lskdjf has joined the room
[00:57:45] *** la|r|ma has joined the room
[01:00:04] *** jubalh has joined the room
[01:09:16] *** la|r|ma has left the room
[01:09:23] *** la|r|ma has joined the room
[01:20:40] *** Guus has joined the room
[01:20:41] *** remko has joined the room
[01:20:41] *** remko shows as "away"
[01:24:34] *** jubalh has left the room
[01:29:45] *** jubalh has joined the room
[01:33:22] *** nyco has left the room
[01:33:27] *** nyco shows as "online"
[01:39:53] *** Guus has left the room
[01:57:11] *** moparisthebest has left the room
[01:58:12] *** Dave Cridland has left the room
[01:58:13] *** Dave Cridland has joined the room
[01:58:47] *** Dave Cridland has left the room
[01:58:54] *** Dave Cridland has joined the room
[01:59:10] *** moparisthebest has joined the room
[02:04:54] *** Zash has left the room
[02:12:19] *** lskdjf shows as "online"
[02:14:57] *** remko has joined the room
[02:14:58] *** remko shows as "away"
[02:22:08] *** moparisthebest has left the room
[02:24:14] *** moparisthebest has joined the room
[02:32:50] *** efrit has joined the room
[02:32:56] *** @Alacer has left the room
[02:32:58] *** @Alacer has joined the room
[02:39:33] *** Kev shows as "away" and his status message is "Weekend"
[02:39:33] *** Kev shows as "away" and his status message is "Weekend"
[02:39:36] *** lskdjf shows as "online"
[02:39:42] *** lskdjf shows as "online"
[02:52:10] *** Kev has left the room
[03:06:42] *** lskdjf shows as "online"
[03:08:22] *** lskdjf has left the room
[03:09:11] *** remko has joined the room
[03:09:11] *** remko shows as "away"
[03:10:44] *** uc has joined the room
[03:12:16] *** andy has joined the room
[03:18:16] *** efrit has left the room
[03:20:46] *** Guus has joined the room
[03:38:30] *** SamWhited has left the room
[03:39:53] *** Guus has left the room
[03:47:57] *** Ge0rG has left the room
[03:48:08] *** Ge0rG shows as "online"
[03:58:11] *** andy has left the room
[04:03:27] *** remko has joined the room
[04:03:27] *** remko shows as "away"
[04:11:52] *** andy has joined the room
[04:15:45] *** rion has joined the room
[04:28:03] *** jubalh has joined the room
[04:35:40] *** Syndace has left the room
[04:35:43] *** Syndace has joined the room
[04:46:00] *** jubalh has left the room
[04:46:30] *** jubalh has joined the room
[04:59:02] *** jubalh has left the room
[05:14:42] *** andy has left the room
[05:34:11] *** jubalh has joined the room
[05:34:12] *** suzyo has joined the room
[05:36:11] *** zinid has left the room
[05:37:09] *** zinid has joined the room
[05:47:47] *** suzyo has left the room
[05:51:34] *** jjrh has left the room
[05:51:34] *** jjrh has left the room
[05:51:34] *** jjrh shows as "online"
[05:53:07] *** jjrh shows as "online"
[06:03:53] *** Guus has joined the room
[06:12:35] *** Guus has left the room
[06:12:35] *** Guus has joined the room
[06:16:12] *** zinid shows as "online"
[06:21:53] *** Guus has left the room
[06:22:02] *** Guus has joined the room
[06:23:24] <Seve> daniel, Matrix already does that. http://mirror.onet.pl/pub/mirrors/video.fosdem.org/2018/H.1309 (Van%20Rijn)/matrix_webvr.webm
[06:29:14] *** zinid has left the room
[06:38:06] *** daniel has left the room
[06:43:52] *** Guus has left the room
[06:43:52] *** Guus has joined the room
[06:48:46] *** Guus has left the room
[06:48:47] *** Guus has joined the room
[06:52:00] <rion> looks cool
[06:53:21] *** Guus has left the room
[06:53:21] *** Guus has joined the room
[06:54:15] *** suzyo has joined the room
[06:54:53] *** Guus has left the room
[06:58:09] *** Dave Cridland shows as "online"
[07:01:22] *** tux has joined the room
[07:01:56] <zinid> daniel: Psi gets crazy when it receives muc self-presence, it renders "empty" participant in muc roster
[07:02:35] *** goffi has joined the room
[07:06:23] <rion> O_o
[07:06:54] *** lovetox has joined the room
[07:09:25] <zinid> yeah
[07:11:09] <rion> What server should I install to have such a self-presence? ejabberd?
[07:12:55] <zinid> rion, no, it's on local machine
[07:13:00] <rion> in any case last two day I rewrote muc roster. so it won't happen anymore. not commited yet
[07:13:08] <zinid> good
[07:13:26] <zinid> the problem is that now I'm not sure if other clients don't break though
[07:14:06] <zinid> rion, Psi requests MUC vCard on every join?
[07:14:14] <rion> yep
[07:14:46] <rion> I think with self-presence and caps I'll change this too
[07:18:15] <rion> and I have to improve avatars caching for mucs too as well as their requesting algo. otherwise it's a lot of traffic on join.
[07:20:12] *** Steve Kille shows as "online" and his status message is "At Home"
[07:21:25] *** Tobias has joined the room
[07:21:27] <zinid> yeah...
[07:21:34] *** lovetox has left the room
[07:21:38] <zinid> poezio has the same problem wrt avatars 😉
[07:23:35] *** mimi89999 shows as "online"
[07:25:58] *** remko has joined the room
[07:25:59] *** remko shows as "online"
[07:28:06] *** tim@boese-ban.de has left the room
[07:29:46] *** nyco has left the room
[07:29:51] *** nyco shows as "online"
[07:33:56] *** intosi shows as "away" and his status message is "Away"
[07:34:01] *** intosi shows as "online"
[07:35:49] *** remko shows as "away"
[07:36:18] *** Dave Cridland has left the room
[07:36:35] *** Dave Cridland shows as "online"
[07:38:10] *** Dave Cridland has left the room
[07:38:17] *** Dave Cridland shows as "online"
[07:40:23] *** tim@boese-ban.de has joined the room
[07:45:10] *** Steve Kille shows as "away" and his status message is "At Home"
[07:46:27] <zinid> gajim seems doing fine, doesn't render anything strange
[07:47:16] *** Dave Cridland has left the room
[07:48:52] *** Dave Cridland shows as "online"
[07:51:00] *** jonasw shows as "online"
[07:51:11] *** Dave Cridland has left the room
[07:51:18] *** Dave Cridland shows as "online"
[07:52:13] *** Dave Cridland has left the room
[07:52:21] *** Dave Cridland shows as "online"
[07:55:20] *** Dave Cridland has left the room
[07:56:30] *** Dave Cridland shows as "online"
[07:58:02] *** daniel has joined the room
[07:58:38] *** Dave Cridland has left the room
[07:58:39] *** Dave Cridland shows as "online"
[07:59:06] *** Dave Cridland has left the room
[08:01:45] *** Steve Kille shows as "online" and his status message is "At Home"
[08:02:08] *** Kev shows as "online"
[08:02:14] *** Dave Cridland shows as "online"
[08:07:20] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[08:08:07] *** SaltyBones has left the room
[08:08:10] *** remko shows as "online"
[08:08:45] *** uc has joined the room
[08:09:14] *** remko has left the room
[08:09:16] *** remko has joined the room
[08:09:16] *** remko shows as "online"
[08:11:50] *** tim@boese-ban.de shows as "online"
[08:12:02] *** bra has joined the room
[08:13:44] *** daniel shows as "online"
[08:15:17] *** zinid has left the room
[08:15:21] *** zinid has joined the room
[08:17:54] *** SaltyBones has joined the room
[08:19:03] *** Steve Kille has left the room
[08:22:35] *** bra shows as "away" and his status message is "Автостатус (неактивен)"
[08:23:36] *** Steve Kille has joined the room
[08:23:37] *** Steve Kille shows as "away" and his status message is "At Home"
[08:23:40] *** Steve Kille shows as "online" and his status message is "At Home"
[08:25:15] *** Steve Kille has left the room
[08:26:25] *** Dave Cridland has left the room
[08:26:25] *** Dave Cridland shows as "online"
[08:29:24] *** Steve Kille has joined the room
[08:29:32] *** Steve Kille shows as "online" and his status message is "At Home"
[08:30:01] *** Steve Kille shows as "online" and his status message is "Hampton"
[08:31:35] *** Kev shows as "away"
[08:33:05] *** Guus has joined the room
[08:34:07] *** Kev shows as "online"
[08:35:24] *** remko shows as "away"
[08:38:30] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[08:40:00] *** remko shows as "online"
[08:42:36] *** bra shows as "xa" and his status message is "Автостатус (неактивен)"
[08:43:20] *** tim@boese-ban.de shows as "online"
[08:48:42] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[08:49:08] *** Dave Cridland shows as "online"
[08:50:09] *** remko shows as "away"
[08:53:32] *** daniel has left the room
[08:53:36] *** daniel shows as "online"
[08:55:23] *** daniel has left the room
[08:55:26] *** daniel shows as "online"
[08:58:28] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[08:58:42] *** tim@boese-ban.de shows as "xa" and his status message is " (Nicht verfügbar wegen Untätigkeit seit mehr als 15 Minuten)"
[09:01:55] *** daniel has left the room
[09:01:59] *** daniel shows as "online"
[09:05:06] *** Guus has left the room
[09:05:06] *** Guus has joined the room
[09:06:21] *** Martin has joined the room
[09:08:02] *** daniel has left the room
[09:08:28] *** Dave Cridland shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[09:09:18] *** Dave Cridland shows as "online"
[09:09:51] *** daniel shows as "online"
[09:11:24] *** marc shows as "online"
[09:12:21] *** Guus has left the room
[09:12:21] *** Guus has joined the room
[09:15:53] *** Guus has left the room
[09:16:16] *** andy has joined the room
[09:19:32] *** tim@boese-ban.de shows as "online"
[09:21:46] *** tux shows as "dnd" and his status message is "Work work …"
[09:24:34] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[09:24:56] *** tim@boese-ban.de shows as "online"
[09:28:25] *** nyco has left the room
[09:29:15] *** nyco shows as "online"
[09:29:32] *** bra has left the room
[09:29:32] *** bra has joined the room
[09:31:03] *** andy has left the room
[09:31:32] *** Guus has joined the room
[09:32:33] *** winfried shows as "online"
[09:32:50] *** rion shows as "online"
[09:33:08] *** ralphm shows as "online"
[09:34:51] *** rion has left the room
[09:35:09] *** rion shows as "online"
[09:35:13] *** rion has left the room
[09:36:29] *** Steve Kille shows as "away" and his status message is "Hampton"
[09:37:15] *** rion shows as "online"
[09:37:58] *** hannes has joined the room
[09:38:04] *** rion has left the room
[09:38:07] *** rion shows as "online"
[09:39:06] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[09:44:01] *** jubalh has joined the room
[09:44:09] *** rion has left the room
[09:44:16] *** rion shows as "online"
[09:44:20] *** rion has left the room
[09:44:21] *** Zash has joined the room
[09:44:36] *** rion shows as "online"
[09:45:19] *** Steve Kille shows as "online" and his status message is "Hampton"
[09:48:15] *** @Alacer has left the room
[09:48:18] *** @Alacer has joined the room
[09:49:06] *** tim@boese-ban.de shows as "xa" and his status message is " (Nicht verfügbar wegen Untätigkeit seit mehr als 15 Minuten)"
[09:49:30] *** Guus has left the room
[09:49:30] *** Guus has joined the room
[09:49:53] *** jubalh has left the room
[09:50:32] *** rion has left the room
[09:50:34] *** rion shows as "online"
[09:50:52] *** rion has left the room
[09:51:12] *** ralphm shows as "online"
[09:53:21] *** tux shows as "away"
[09:54:11] *** jubalh has joined the room
[09:54:49] *** Alex has joined the room
[09:54:59] *** Alex shows as "online"
[09:55:11] *** jubalh has left the room
[09:57:20] *** ralphm has left the room
[09:57:52] *** tim@boese-ban.de shows as "online"
[09:59:18] *** rion shows as "online"
[10:00:53] *** Guus has left the room
[10:06:55] *** rion has left the room
[10:07:22] *** rion shows as "online"
[10:10:03] *** Steve Kille shows as "away" and his status message is "Hampton"
[10:11:54] *** jubalh has joined the room
[10:12:22] *** Steve Kille shows as "online" and his status message is "Hampton"
[10:12:29] *** jubalh has left the room
[10:12:44] *** ralphm shows as "online"
[10:13:31] *** jjrh has left the room
[10:15:37] *** jjrh shows as "online"
[10:18:42] *** winfried shows as "away" and his status message is "sssssttttt! my computer fell asleep"
[10:18:51] *** winfried shows as "online"
[10:19:49] *** rion has left the room
[10:19:53] *** rion shows as "online"
[10:20:57] *** daniel has left the room
[10:21:01] *** daniel shows as "online"
[10:21:29] <Seve> By the way, regarding the email I sent to JUser, is there a better mailing list for that, please?
[10:21:48] <jonasw> I still haven’t subscribed to juser :/
[10:22:20] <Seve> That's why I would like to send it somewhere else too, seems not many people are subscribed there
[10:22:40] <jonasw> Seve, I think operators would be more fitting, maybe
[10:22:43] *** rion has left the room
[10:22:45] <jonasw> possibly with a cross-post to members@
[10:23:05] *** rion shows as "online"
[10:23:19] <jonasw> I’m also not sure if a separate MUC for this is a good idea
[10:23:35] <Zash> Posted something to some KDE list?
[10:24:00] *** rion has left the room
[10:24:17] *** rion shows as "online"
[10:24:56] *** rion has left the room
[10:25:52] <jonasw> frankly, I’m not sure XMPP can fulfil the requirements at this instant in time
[10:25:55] *** rion shows as "online"
[10:27:07] <jonasw> > Stickers are available
[10:27:09] <jonasw> we doomed
[10:27:29] <Zash> Find out which are requirements and which are wishlist
[10:27:45] <jonasw> Zash, they don’t even know that yet
[10:27:50] *** rion has left the room
[10:27:50] <jonasw> > Requirements are not prioritized (incl. must-have vs. nice-to-have), that comes at a later step.
[10:28:01] *** rion shows as "online"
[10:29:22] *** rion has left the room
[10:29:29] *** rion shows as "online"
[10:29:43] *** Martin shows as "online"
[10:29:49] *** Martin shows as "away" and his status message is "Away"
[10:30:18] *** Martin has left the room
[10:31:46] *** jubalh has joined the room
[10:31:52] *** rion has left the room
[10:31:55] *** rion shows as "online"
[10:32:32] *** daniel has left the room
[10:32:34] *** daniel shows as "online"
[10:32:58] <Zash> jonasw: All are wishlist. There's online one requirement, that everyone you care about is already using it. :)
[10:34:19] *** andy has joined the room
[10:36:13] *** tim@boese-ban.de has left the room
[10:38:17] *** rion has left the room
[10:38:26] <Seve> jonasw, the point of the MUC is to discuss this topic with people interested in helping out. Almost everybody here would find this discussion annoying here.
[10:38:32] *** ralphm shows as "online"
[10:40:05] <jonasw> Seve, I’m running out of screen space for more MUCs.
[10:43:12] *** rion has left the room
[10:43:21] *** rion has joined the room
[10:43:41] *** tim@boese-ban.de has joined the room
[10:44:17] *** rion has left the room
[10:46:42] *** andy has left the room
[10:47:04] <marc> Ge0rG, jonasw can you review https://git.zapb.de/xeps.git/commit/?h=fix_xep_0401&id=52725e993987f205dc253cc5a4e6937fe3955d81 and merge please?
[10:47:27] *** uc has joined the room
[10:48:57] <jonasw> marc, EBUSY right now
[10:49:28] <jonasw> marc, if you refuse to make a github PR (which would really be the most useful thing for me), please send an email with the first line of the commit message as subject to editor@xmpp.org
[10:49:49] <marc> jonasw, I can do a GitHub PR if you like
[10:49:57] <jonasw> marc, that would be much appreciated, thanks
[10:50:02] *** jubalh has left the room
[10:50:03] <marc> jonasw, you're welcome
[10:50:12] <jonasw> gotta run now, see you
[10:50:13] *** jonasw shows as "away"
[10:50:20] *** SaltyBones shows as "online"
[10:50:20] <marc> bye
[10:50:34] *** Martin has joined the room
[10:50:39] *** Martin shows as "away" and his status message is "Away"
[10:50:43] *** Martin shows as "online"
[10:51:49] *** moparisthebest has joined the room
[10:52:17] *** rion has joined the room
[10:54:20] *** winfried shows as "away" and his status message is "sssssttttt! my computer fell asleep"
[10:55:52] <marc> done
[10:56:57] *** Alex has left the room
[10:57:09] *** winfried shows as "online"
[10:59:10] *** jjrh has left the room
[10:59:14] *** jjrh shows as "online"
[10:59:59] *** Zash has left the room
[11:00:48] *** rion has left the room
[11:00:50] *** rion has joined the room
[11:01:46] *** rion has left the room
[11:03:09] *** winfried shows as "away" and his status message is "sssssttttt! my computer fell asleep"
[11:03:54] *** jubalh has joined the room
[11:05:06] *** valo has left the room
[11:05:11] *** valo has joined the room
[11:06:55] *** jubalh has left the room
[11:08:27] *** rion has joined the room
[11:10:36] *** intosi shows as "online"
[11:13:09] *** winfried shows as "xa" and his status message is "sssssttttt! my computer fell asleep"
[11:14:49] *** andy has joined the room
[11:17:40] *** Zash has joined the room
[11:17:58] *** hannes has joined the room
[11:18:43] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[11:19:26] *** daniel has left the room
[11:19:30] *** daniel shows as "online"
[11:21:10] *** Holger shows as "online" and his status message is "I'm available"
[11:23:34] *** remko shows as "online"
[11:25:01] *** remko has left the room
[11:25:02] *** Zash shows as "online"
[11:25:03] *** remko has joined the room
[11:25:04] *** remko shows as "online"
[11:25:06] *** Zash shows as "online"
[11:26:59] *** jubalh has joined the room
[11:27:04] *** andy has left the room
[11:27:57] *** jubalh has left the room
[11:28:05] *** tux shows as "xa" and his status message is "Gone for Food"
[11:28:43] *** tim@boese-ban.de shows as "xa" and his status message is " (Nicht verfügbar wegen Untätigkeit seit mehr als 15 Minuten)"
[11:34:27] *** intosi shows as "online"
[11:35:04] *** remko shows as "away"
[11:40:46] *** ralphm shows as "online"
[11:42:42] *** Holger shows as "away" and his status message is "Auto-away (idle)"
[11:44:53] *** jonasw shows as "online"
[11:45:47] *** andy has joined the room
[11:47:10] <jonasw> zinid, regarding the server capabilities invalidation, how about a message?
[11:47:57] <jonasw> server changes caps -> server sends <message/> to local clients. for server-to-server links, it could simply close the link so that the peer re-opens it, thus getting new stream features
[11:51:07] *** zinid shows as "online"
[11:52:29] *** Alex has joined the room
[11:55:18] *** remko has joined the room
[11:55:18] *** remko shows as "online"
[11:56:16] <zinid> jonasw, closing and opening thousands of s2s will create avalanche effect
[11:56:25] <jonasw> right
[11:56:27] <zinid> we see this on jabber.ru for examlpe
[11:57:00] <zinid> so I just think about nonza
[11:57:06] <zinid> stream-level element
[11:57:20] *** MattJ shows as "online"
[11:57:30] <jonasw> zinid, thanks for your input, I’ll post a few suggestions to the mailing list and I hope you’ll comment on them :)
[11:57:37] *** tim@boese-ban.de shows as "online"
[11:57:41] <zinid> sure 😉
[11:59:06] *** suzyo has joined the room
[12:01:03] *** intosi has left the room
[12:01:47] *** andy has left the room
[12:04:07] *** ralphm shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[12:05:29] *** remko shows as "away"
[12:06:45] *** Dave Cridland has left the room
[12:08:10] *** moparisthebest has joined the room
[12:09:20] *** Dave Cridland shows as "online"
[12:10:58] *** tux shows as "dnd" and his status message is "Work work …"
[12:11:00] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[12:12:50] *** Dave Cridland has left the room
[12:12:59] *** Dave Cridland shows as "online"
[12:13:00] *** Dave Cridland shows as "online"
[12:14:07] *** ralphm shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[12:15:24] *** Dave Cridland has left the room
[12:15:50] *** ralphm shows as "online"
[12:18:17] <SaltyBones> Is there a good resource on the problems with message ids?
[12:18:29] *** Steve Kille shows as "away" and his status message is "Hampton"
[12:18:42] <jonasw> zinid, replied, happy to hear from you d)
[12:18:45] <jonasw> zinid, replied, happy to hear from you :)
[12:18:48] <jonasw> SaltyBones, I’m not sure
[12:18:53] <SaltyBones> I'm looking at XEP-0359 but the problems mentioned at the summit are not in there.
[12:19:05] <jonasw> maybe the summit notes as first starting point
[12:19:10] *** hannes has joined the room
[12:20:13] *** jubalh has joined the room
[12:20:17] *** andy has joined the room
[12:20:36] *** Dave Cridland shows as "online"
[12:20:43] *** jubalh has left the room
[12:20:53] *** Syndace has left the room
[12:20:57] *** Syndace has joined the room
[12:22:36] *** suzyo has joined the room
[12:23:21] <SaltyBones> 1.2 Stable IDs
Discussion on entropy ensues.
dwd: <stanza-ids> don't work for unique addressing because we don't trust clients to do them properly.
Kev: lots of possible attacks with spoofable stanza IDs.
[12:23:29] <SaltyBones> Not verbose enough.
[12:24:26] <jonasw> SaltyBones, I think the gist of that is: we need to generate stanza IDs on the server because weak/constructed stanza IDs are a problem. but the client needs to know the stanza ID for archive operations and possibly other things
[12:26:02] *** Martin shows as "online"
[12:26:09] *** Martin shows as "away" and his status message is "Away"
[12:26:38] <zinid> jonasw, well, dunno, for me message looks too complex
[12:26:39] *** Martin has left the room
[12:27:13] <zinid> can't we just send <c/> as a nonza?
[12:27:56] <jonasw> zinid, I think RFC 6120 would slap you in the face for that ;-)
[12:28:02] <jonasw> really, it needs negotiation
[12:28:14] *** matlag shows as "online"
[12:28:20] <jonasw> it’s a shame that we have no way for a client to send stream features :(
[12:28:39] <zinid> yeah, because some implementation may close a stream if they receive unrecognized nonzas
[12:28:45] <jonasw> yupp
[12:28:47] <jonasw> I know at least one.
[12:29:00] <zinid> however, I tried to implement such behaviour (closing a stream) and got lots of problems and left the idea
[12:29:26] <jonasw> never had issues so far; did you encounter things which sent unsolicited nonzas?
[12:29:54] <zinid> alas, I've already forgotten what that was exactly 😕
[12:31:02] <jonasw> zinid, the best we could do is probably say "okay, if the client sent a presence update with caps, we can send a nonza with a caps update"
[12:31:13] <jonasw> and similarly for servers ("offered caps in stream feature -> send nonza")
[12:31:34] <jonasw> but I don’t know enough about s2s to be sure that this works
[12:32:16] *** SaltyBones shows as "online"
[12:32:23] <jonasw> this would definitely need a namespace bump though
[12:33:17] <zinid> well, we're talking about new caps xep?
[12:33:21] <zinid> we can do it there
[12:33:29] <jonasw> yeah, it requires a namespace bump there too ;-)
[12:33:50] <zinid> can't we just negotiate this feature?
[12:33:51] *** Dave Cridland has left the room
[12:33:51] *** Dave Cridland shows as "online"
[12:33:56] <zinid> not sure how
[12:34:03] <jonasw> negotiation will add a round-trip
[12:34:21] <jonasw> which people will react very adversely to
[12:34:22] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[12:34:28] <zinid> ah
[12:34:35] <jonasw> but please comment on-list, hoping to get more input on that
[12:34:43] <zinid> I don't know what to say
[12:34:52] *** Dave Cridland has left the room
[12:35:05] <jonasw> what you said here, essentially?
[12:35:55] *** Dave Cridland shows as "online"
[12:36:13] *** la|r|ma has joined the room
[12:36:30] *** Dave Cridland shows as "online"
[12:36:42] *** lskdjf has joined the room
[12:37:38] *** andy has left the room
[12:39:44] *** Dave Cridland has left the room
[12:41:28] *** Dave Cridland shows as "online"
[12:41:30] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[12:42:36] *** remko has joined the room
[12:42:36] *** remko shows as "away"
[12:44:04] *** Dave Cridland has left the room
[12:44:12] *** Dave Cridland shows as "online"
[12:45:06] *** Dave Cridland has left the room
[12:45:06] *** Dave Cridland has left the room
[12:45:07] *** Dave Cridland shows as "online"
[12:45:27] *** Dave Cridland shows as "online"
[12:45:32] *** Dave Cridland has left the room
[12:45:41] *** Dave Cridland shows as "online"
[12:47:28] *** Dave Cridland has left the room
[12:47:29] *** Dave Cridland shows as "online"
[12:47:31] *** Dave Cridland has left the room
[12:47:41] *** Dave Cridland shows as "online"
[12:48:15] *** Dave Cridland has left the room
[12:48:24] *** Dave Cridland shows as "online"
[12:49:32] <SaltyBones> There are stanza-id and origin-id in XEP-0359. Somebody at the summit also mentioned "message-id" is that defined somewhere?
[12:50:05] <MattJ> They were probably referring to the id attribute, I guess?
[12:50:55] *** Dave Cridland has left the room
[12:51:02] *** Dave Cridland shows as "online"
[12:51:05] <SaltyBones> MattJ, this https://tools.ietf.org/html/rfc6120#section-8.1.3 ?
[12:51:26] <MattJ> Yes
[12:51:30] *** Dave Cridland shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[12:51:46] <SaltyBones> thx
[12:52:21] <MattJ> The reason XEP-0359 exists is because the id attribute is controlled (and can only be trusted by) the original sender of the stanza
[12:53:18] <MattJ> It's really just designed for tracking errors, though a couple of XEPs have re-used it for other purposes
[12:53:29] *** Dave Cridland has left the room
[12:53:48] *** Dave Cridland shows as "online"
[12:53:51] *** tux shows as "dnd" and his status message is "Work work …"
[12:53:58] *** tux shows as "away"
[12:56:20] *** Martin has joined the room
[12:56:25] *** Martin shows as "away" and his status message is "Away"
[12:56:29] *** Martin shows as "online"
[12:57:18] *** Dave Cridland has left the room
[12:57:36] *** Dave Cridland shows as "online"
[12:58:35] *** suzyo has joined the room
[12:58:36] *** Dave Cridland shows as "online"
[12:58:50] *** ralphm shows as "online"
[12:59:02] *** Dave Cridland has left the room
[12:59:15] *** ralphm shows as "online"
[12:59:16] *** Dave Cridland shows as "online"
[13:00:09] <SaltyBones> The id-attribute you mean?
[13:00:17] <jonasw> yes
[13:02:00] *** vanitasvitae shows as "online"
[13:04:30] <SaltyBones> So, I guess, stanza-id is used by MAM and origin-id is used to detect MUC reflections (that's what 0359 says anyway). And id-attribute is used for errors but the error will have the same id-attribute iiuc, right?
[13:04:45] <jonasw> yes, pretty much
[13:04:53] <jonasw> except that origin-id won’t work with all MUCs
[13:05:00] <SaltyBones> Why?
[13:05:10] *** andy has joined the room
[13:05:14] <jonasw> because not all MUCs may be able to reflect that ID for whatever implementation specific reason
[13:05:25] <jonasw> (just like not all mucs can reflect the message @id)
[13:06:25] *** Zash has left the room
[13:07:09] <MattJ> There is only one implementation I'm aware of that doesn't (didn't?) reflect @id, and I'm not even sure of the current status of that
[13:07:15] <SaltyBones> Okay, so why does the standard not just mandate that it be reflected?
[13:07:41] <MattJ> Simple oversight I suspect
[13:07:45] <Dave Cridland> SaltyBones, Non-MUC things pretending to be MUC, in part.
[13:08:04] <Dave Cridland> SaltyBones, Also by the time people had noticed this was a problem, enough implementations didn't.
[13:08:51] <SaltyBones> So, could this be fixed by changing the standard and requiring that origin-ids be reflected?
[13:09:10] *** Dave Cridland has left the room
[13:09:18] <SaltyBones> Dave Cridland, what do you mean by "non-muc things pretending to be muc"?
[13:09:24] <jonasw> SaltyBones, changing the standard doesn’t fix the implementations magically
[13:09:25] *** Dave Cridland shows as "online"
[13:09:26] <MattJ> Dave Cridland, enough = 1? Also in the case of bridging to non-MUC, isn't it as simple as 1) if the non-MUC supports acks, wait for the ack and reflect the id or 2) reflect the id immediately?
[13:09:33] <jonasw> SaltyBones, IRC transports are "non-muc things pretending to be MUC" for example
[13:09:34] <SaltyBones> jonasw, not magically but ...
[13:09:51] <Dave Cridland> MattJ, More than one, I think. One classic MUC implementation, but quite a few transports.
[13:10:02] *** uc has joined the room
[13:10:46] <Dave Cridland> FWIW, I've gone back and forth over whether MUC ought to reflect ids. It's a special case of maintaining the id on bradcast, which is itself both useful in some cases and bad in others.
[13:11:01] <SaltyBones> Why is this reflection necessary?
[13:11:33] <Dave Cridland> SaltyBones, Reflection isn't - we could add in a subelement which indicated the original id.
[13:12:04] <Dave Cridland> SaltyBones, Broadcast ids are harder to work around - there's a few protocols which use id as reference.
[13:12:21] *** winfried shows as "online"
[13:12:37] <Dave Cridland> This all said I *really* dislike having a zillion ids as subelements of stanzas when we already have an id attribute.
[13:13:43] *** Kev has left the room
[13:13:56] <SaltyBones> Well, what is the subelement necessary for? Is this just as an ACK to the client?
[13:14:03] <SaltyBones> What do you mean by broadcast-id?
[13:14:33] *** Dave Cridland has left the room
[13:14:35] <MattJ> SaltyBones, if you're asking why MUC reflects messages to the sender in the first place (some systems, e.g. IRC don't) - it has a few benefits
[13:14:47] <MattJ> Such as ensuring everyone in the room sees the same messages in the same order
[13:15:32] <jonasw> it also is natural as soon as multiple clients are in the play, somewhat like message carbons
[13:15:32] <Dave Cridland> SaltyBones, So the id attribute in the message stanza for a groupchat message sent to the MUC can be reused in the reflected message, and/or the broadcast messages to other participants in the group.
[13:15:33] <MattJ> In IRC if two people send a message at the "same time", the messages will be shown in a different order on both their clients
[13:15:46] <MattJ> and yes, multiple clients from one user in the room
[13:16:11] <jonasw> it allows MUCs to do fancy things to messages and still have everyone see the same thing
[13:16:19] <Dave Cridland> (As a side-note, Matrix achieves a similar goal by a complex hash chain)
[13:16:35] *** andy has left the room
[13:16:39] <MattJ> SaltyBones, and also servers modifying messages (e.g. in the Prosody MUC we automatically convert long multi-line messages to a pastebin link). The reflection allows the sending client to see what everyone else sees.
[13:16:46] *** remko has joined the room
[13:16:47] *** remko shows as "online"
[13:16:48] *** andy has joined the room
[13:16:56] *** Dave Cridland shows as "online"
[13:16:59] <jonasw> (in contrast to IRC, where you don’t see that your message was truncated)
[13:17:03] <SaltyBones> Okay, pretty convincing...
[13:17:03] <MattJ> Dave Cridland, and as many people at FOSDEM noted, much RAM
[13:17:32] <Dave Cridland> MattJ, Yes but BLOCKCHAIN.
[13:17:41] <jonasw> #bingo #triggered
[13:17:49] <SaltyBones> 🔲‍⛓️
[13:18:15] <Dave Cridland> I hope there's a ZWJ there.
[13:20:05] *** ralphm shows as "online"
[13:20:34] <SaltyBones> *Yes, there is.*
[13:21:15] *** Dave Cridland has left the room
[13:21:20] <SaltyBones> Okay, so we want reflected messages. Sounds fair. Let's suppose for a second that we can just make them mandatory in the standard and people would add them.
[13:21:37] <jonasw> they *are* mandatory
[13:21:44] <jonasw> the issue is with message IDs
[13:21:54] <jonasw> clients need to be able to recognize the reflection
[13:22:04] <SaltyBones> You mean id-attribute?
[13:22:12] <jonasw> or any ID really
[13:22:16] <jonasw> id attribute would be easiest
[13:22:28] <Dave Cridland> We could indicate in disco if ids were stable.
[13:22:29] *** andy has left the room
[13:22:42] *** Dave Cridland shows as "online"
[13:22:52] <jonasw> wasn’t that proposal rejected-ish?
[13:22:57] <SaltyBones> Okay, 1. is the origin-id used for anything else and 2. is the whole point of origin-id not to detect reflection?
[13:23:25] *** moparisthebest has joined the room
[13:23:33] <jonasw> speaking of MUC, I think the vote on https://github.com/xsf/xeps/pull/559 expired, didn’t it?
[13:24:16] <Ge0rG> I've attemtped to fix MUC reflected IDs some years back, not even in the normative language but merely by fixing the examples
[13:24:23] <Ge0rG> And got significant flack for it.
[13:25:26] <Ge0rG> I also attempted to make origin-id == message-id, but it was refused by the XEP author.
[13:25:27] *** Steve Kille shows as "online" and his status message is "Hampton"
[13:25:54] *SaltyBones groans.
[13:26:16] <Ge0rG> Now I'll just sit and wait, and ocassionaly proclaim: *I told you so!*
[13:26:19] <SaltyBones> So, is the point of origin-id to be used for reflection or is there something else?
[13:26:30] <Flow> SaltyBones, that is one use case
[13:26:31] *** lumi has joined the room
[13:26:35] <SaltyBones> Ge0rG, great, if you could also occasionally chime in with reasons for things and explanations that would be awesome.
[13:26:43] <SaltyBones> Flow, like?
[13:26:46] <Ge0rG> SaltyBones: it's mainly for reflection, except that MUCs are not guaranteed to keep non-body elements in the reflection
[13:26:55] *** remko shows as "away"
[13:27:07] <Flow> other use cases include finding your own messages in a archive
[13:27:18] <Ge0rG> SaltyBones: https://wiki.xmpp.org/web/XEP-Remarks/XEP-0045:_Multi-User_Chat#Matching_Your_Reflected_Message
[13:27:19] <SaltyBones> Flow, MAM uses stanza-id, right?
[13:27:34] <Flow> SaltyBones, it does
[13:27:36] <Ge0rG> SaltyBones: yes, but you don't know the stanza-id for the messages you sent
[13:27:45] <SaltyBones> Ge0rG, would you say this is a requirement by nature or just a problem with implementations?
[13:28:03] <daniel> i agree that the xep should specify that the sender should set origin-id=message-id
[13:28:05] <Ge0rG> SaltyBones: what?
[13:28:14] <Ge0rG> daniel: MUST set.
[13:28:21] <SaltyBones> Ge0rG, the fact that MUCs don't reflect non-body attributes.
[13:28:24] <daniel> yes
[13:28:25] <Ge0rG> because anything different is just a path into insanity
[13:28:28] *** winfried shows as "away" and his status message is "sssssttttt! my computer fell asleep"
[13:28:35] <daniel> (i didn't mean SHOULD)
[13:28:54] <SaltyBones> Ge0rG, also, why do I need the stanza-id for a message I sent?
[13:29:02] <Ge0rG> SaltyBones: have a look at biboumi, a modern IRC transport. It doesn't reflect your message ID, it doesn't reflect any non-body elements and it mangles multi-line messages
[13:29:06] <Ge0rG> Welcome to message tracking hell.
[13:29:29] <daniel> i think thinks will already break if some client doesn't do that and then expects a delivery reciept or something
[13:29:35] <Ge0rG> SaltyBones: because when you ask for a MAM sync, you'll get yout sent messages copied to you as well
[13:29:37] <daniel> when sending messages to Conversations i mean
[13:30:09] <Ge0rG> Yes, message correction, delivery receipts and anything else that references IDs is a mess already.
[13:30:15] <SaltyBones> Ge0rG, and why is that a problem? They should have their origin-id and be recognizable, right?
[13:30:44] <Ge0rG> SaltyBones: you can't rely on origin-id being there, and you can't rely on it matching the message-id.
[13:30:49] <SaltyBones> daniel, what breaks when a client does what?
[13:31:15] <Ge0rG> SaltyBones: but yes, you can match MAM archives based on origin-id
[13:31:21] <daniel> if an origin id is set Conversations will use that as a reference in the receipt
[13:31:41] <daniel> so if your client doesn't do id=origin-id and expects the receipt for the id it won't work
[13:32:11] <daniel> same with everything else that references ids
[13:32:29] <SaltyBones> So, there seems to be consensus at least in this MUC right now that attribute-id should be the same as origin-id?
[13:33:00] <daniel> i'd argue there is no good reason not do make this a MUST
[13:33:10] <daniel> i mean most client would naturally do this anyway
[13:33:16] <SaltyBones> Ge0rG, who was the xep-author who was against this and do you remember why?
[13:33:21] <jonasw> SaltyBones, Flow
[13:33:24] <daniel> because why would you generate a second id if you can reuse the existing one
[13:33:31] <zinid> Who is generating origin-id? A client?
[13:33:33] <Ge0rG> SaltyBones: https://mail.jabber.org/pipermail/standards/2017-September/033415.html
[13:33:36] <jonasw> zinid, the original sender
[13:33:38] <daniel> zinid, yes
[13:33:40] <jonasw> i.e. most of the times a client
[13:33:43] *** tux shows as "online"
[13:33:53] *** jonasw shows as "away"
[13:34:23] *** jubalh has joined the room
[13:34:34] <daniel> note that muc rewrites a irrelevant to the scenerio described above
[13:34:46] <Dave Cridland> jonasw, #559 expired, yes.
[13:34:51] <zinid> This is to track its own messages?
[13:34:59] <daniel> zinid, yes
[13:35:05] <daniel> or references to that
[13:35:25] <Dave Cridland> jonasw, But with no veto and the rest are +1, so apply it.
[13:35:30] *** Dave Cridland has left the room
[13:35:30] <SaltyBones> jonasw, what is this ID-rewriting MUC shit? :)
[13:36:01] <daniel> some mucs are known to change the 'attribute' id
[13:36:15] <Ge0rG> SaltyBones: have a look at examples 44 and 45 in https://xmpp.org/extensions/xep-0045.html#message
[13:36:21] <daniel> so tracking your own message are parsing references don't work
[13:36:42] *** Dave Cridland shows as "online"
[13:37:03] <zinid> daniel: why, you can rely on origin-id in this case
[13:37:18] <Ge0rG> If we have a MUC that rewrites message IDs, can we mandate that it also MUST rewrite references in all XEP payloads that reference message IDs, please?
[13:37:47] <Ge0rG> zinid: https://wiki.xmpp.org/web/XEP-Remarks/XEP-0045:_Multi-User_Chat#Matching_Your_Reflected_Message
[13:38:23] <daniel> zinid, in the case of muc? because most mucs won't remove child elements from the stanza
[13:38:29] *** winfried shows as "xa" and his status message is "sssssttttt! my computer fell asleep"
[13:38:47] <zinid> I'm lost
[13:39:05] <zinid> daniel: but that's exactly what's needed
[13:39:19] <daniel> now i'm lost :-)
[13:39:36] <zinid> So you can fetch origin-id and check if this is your id
[13:39:46] *** Kev has joined the room
[13:39:48] *** Kev shows as "online"
[13:40:06] *** jjrh has left the room
[13:40:10] *** jjrh shows as "online"
[13:40:37] <zinid> Why do you need to rely on message-id if you inject origin-id
[13:41:11] <daniel> i just said the sender should set id=origin-id
[13:41:12] <SaltyBones> zinid, I think there was a sub-discussion about forcing message-id = origin-id
[13:41:45] <daniel> and/or deal with delivery receipts that reference the origin-id instead of the id
[13:41:59] <daniel> which is made easier if you id=origin-id
[13:41:59] <zinid> Ok, I don't get it 🤔
[13:42:22] <daniel> doesn't matter. that client to client stuff anyway :-)
[13:44:31] *** la|r|ma has left the room
[13:44:46] *** la|r|ma has joined the room
[13:44:53] *** remko has joined the room
[13:44:54] *** remko shows as "online"
[13:45:09] <zinid> From what I understand we just need to ditch IRC transports
[13:45:17] <zinid> 😃
[13:45:21] *** jjrh has left the room
[13:45:24] *** jjrh shows as "online"
[13:45:36] <zinid> That's really their problem
[13:46:02] *** bra shows as "online"
[13:48:06] <daniel> oh yeah. or make irc transports track the messages themselves and re-add origin-id and stuff on reflection
[13:48:28] *** ralphm shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[13:48:38] <daniel> it's probably better if the irc transport does the right thing(tm) than letting each and every client figure it out
[13:49:08] <zinid> daniel: agreed, nobody said writing transport is simple
[13:49:15] <daniel> irc transport should also reassemble the message if they previously split it and stuff like that
[13:49:22] <daniel> because they know best if they did
[13:49:39] <SaltyBones> Hm.
[13:50:22] <SaltyBones> Okay, are there any actual problems with the current state of three IDs except that we don't like having so many?
[13:50:31] <daniel> yes
[13:50:36] <daniel> read what i said before
[13:51:02] <SaltyBones> Uh, assuming a well-behaved server?
[13:51:11] *** jonasw shows as "online"
[13:51:33] <daniel> yes
[13:51:42] <daniel> my argumentation has nothing to do with servers
[13:52:34] <jonasw> daniel, IRC doesn’t even reflect
[13:52:40] <jonasw> so the transport generates the reflections on its own
[13:52:43] <SaltyBones> Okay, can you please point out what you are referring to then?
[13:52:55] <jonasw> biboumi however decided to reflect the split version of the message, and there’s some argument to doing that
[13:53:07] <daniel> assume i sent: <message id="1"><body>hi</body><request/><origin-id="2"></message>
and then i will receive from conversations: <message><receipt id="2"></message>
[13:53:44] <zinid> who cares about IRC, really, are we designing a protocol for old nerds?
[13:54:22] *** Alex shows as "away" and his status message is "Auto-Status (untätig)"
[13:54:26] *** Alex shows as "online"
[13:54:26] <daniel> but we are talking about two different things. the how should irc transports behave has noting to do with the problem i'm describing
[13:54:36] <daniel> this applies to 1:1 chats as well
[13:54:41] <jonasw> daniel, yes, I don’t argue that
[13:54:54] <jonasw> I’m just replying to your message from "13:49:15Z" because I was away from keyboard.
[13:55:12] *** remko has left the room
[13:55:52] *** ralphm shows as "online"
[13:57:01] <daniel> the irc transport with the name i can not pronounce or remember does a few things that i don't like :-)
[13:57:27] <zinid> daniel: if we don't care about IRC then we probably don't need origin-id and thus there is no the problem you just described
[13:58:06] <daniel> there are non transport mucs which also rewrite the id
[13:58:09] <Ge0rG> daniel: https://lab.louiz.org/louiz/biboumi/issues/3283
[13:58:27] *** ralphm shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[13:58:31] <daniel> zinid, if it weren't only for transports i wouldn't care
[13:58:44] *** jonasw shows as "away"
[13:58:45] <daniel> Ge0rG, is that an issue to change the name to something i can remember?
[13:58:50] *** la|r|ma shows as "online"
[13:58:51] <daniel> or pronounce?
[13:58:52] <zinid> daniel: if they are abandoned, then I personally give zero fucks
[13:59:01] <Ge0rG> daniel: no, it's about maintaining IDs on reflection
[13:59:50] <Ge0rG> zinid: everything in XMPP is abandoned. Now stop giving fucks.
[14:00:29] <daniel> i should probably write my own irc transport
[14:00:40] <edhelas> biboumi is not good ?
[14:00:49] <Ge0rG> daniel: you should just patch biboumis two or three warts.
[14:00:50] <daniel> i haven't used irc though ever since counterstrike 1.6 came out
[14:00:52] <zinid> Ge0rG: not everything, so I still have a few fucks to give
[14:02:17] <daniel> there are a number of things i don't like that seem to be design decisions rather than bugs
[14:02:31] *** andy has joined the room
[14:02:34] <daniel> so i'm not sure if they even want me to change them
[14:04:13] <SaltyBones> daniel, so regarding your example, what you are saying is that the read receipt might reference the message-id or the origin-id and it is not properly specified which?
[14:04:14] <zinid> Ok, whatever, so what do you guys think about muc subscriptions?
[14:04:26] *** Alex shows as "away" and his status message is "Auto-Status (untätig)"
[14:04:30] <SaltyBones> Or are you getting at the fact that origin-id requires "by=" and is therefore sometimes not applicable?
[14:04:42] *** ralphm shows as "online"
[14:04:48] <daniel> > daniel, so regarding your example, what you are saying is that the read receipt might reference the message-id or the origin-id and it is not properly specified which?
This
[14:05:23] <daniel> not just read recepits but everything that references something
[14:05:27] <daniel> but yes
[14:05:31] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[14:05:36] *** jubalh has joined the room
[14:05:49] <zinid> daniel: isn't there a place in the xep which says you should copy the message-id?
[14:05:59] <daniel> zinid, no.
[14:06:06] <zinid> Really?
[14:06:09] <daniel> that's what Ge0rG and I want to add to the XEP
[14:06:19] <daniel> that's pretty much what the entire discussion is about :-)
[14:06:23] *** tim@boese-ban.de shows as "online"
[14:06:26] <zinid> So add it 😀
[14:06:40] <daniel> i can't without permission of the author
[14:06:59] <zinid> No shit?
[14:07:01] <SaltyBones> So, there is some problem that I am still not aware of....
[14:07:10] <SaltyBones> Flow, why is the by attribute in the origin-id?
[14:07:11] <zinid> Who's the author?
[14:07:15] <SaltyBones> The one that causes the privacy problems...
[14:07:31] <Kev> Not entirely true, BTW, that it's impossible to do things without the author. But it's the path of least resistance.
[14:07:49] <daniel> i don't think there is a by attribute in the origin id
[14:08:08] <SaltyBones> Ah, sorry that is only for stanza IDs. origin-id does not have by.
[14:08:20] <zinid> Peter is the author
[14:08:54] <SaltyBones> That might almost always be correct buy 0359 is Florian Schmaus
[14:09:10] <SaltyBones> That might almost always be correct but 0359 is Florian Schmaus
[14:09:56] *** Steve Kille shows as "away" and his status message is "Hampton"
[14:10:15] <zinid> > In addition, it SHOULD include an 'id' attribute that echoes the 'id' attribute of the content message.
[14:11:25] <zinid> And you want this SHOULD to be a MUST?
[14:12:06] *** lskdjf shows as "online"
[14:12:13] <SaltyBones> zinid, what some people here want is that origin-id = message-id
[14:12:13] *** lskdjf shows as "online"
[14:12:36] <SaltyBones> Ge0rG, you had an objection to that in https://mail.jabber.org/pipermail/standards/2017-September/033415.html but I don't quite understand it. Why do you want to know if somebody creates strong message-id?
[14:13:10] <zinid> SaltyBones: but the sender controls this itself, so what is a problem to set them equal?
[14:13:39] <daniel> zinid, we want wording that tells the client developer to do this
[14:13:46] <daniel> there is no 'problem'
[14:13:55] *** efrit has joined the room
[14:14:10] <Ge0rG> SaltyBones: I thought it would matter, but it doesn't, because there always can be malicious entities
[14:14:53] <zinid> daniel: if they don't then they only harm themselves, I guess
[14:15:19] <Ge0rG> zinid: no, they harm the other participants
[14:15:29] <zinid> Ge0rG: ah
[14:15:36] <SaltyBones> How?
[14:15:46] <zinid> Yeah, how?
[14:15:57] <Ge0rG> zinid: also, it's well possible that stanza ids are generated by the xmpp library, and origin ids by the client, causing a mismatch
[14:16:25] *** Martin shows as "online"
[14:16:28] <Ge0rG> By making all message references break
[14:16:39] *** Martin shows as "away" and his status message is "Away"
[14:16:43] <Ge0rG> We had that above already.
[14:16:45] <SaltyBones> Before we move to stanza-ids....
[14:16:59] <SaltyBones> There is nothing wrong with requiring origin-id = message-id?
[14:17:02] *** nyco shows as "online"
[14:17:09] *** Martin has left the room
[14:18:05] <zinid> If we require this then why origin-id is needed, wtf?
[14:18:48] *** nyco shows as "online"
[14:19:14] <SaltyBones> zinid, it's possible that it was a mistake
[14:19:40] *** Steve Kille shows as "online" and his status message is "Hampton"
[14:19:44] <SaltyBones> daniel, does requiring origin-id = message-id actually solve your problem or are there cases when stanza-id might be used to refer to a message instead??
[14:20:05] <daniel> It solves my problem
[14:20:09] *** jubalh shows as "online"
[14:21:20] *** andy has left the room
[14:21:41] *** jonasw shows as "online"
[14:21:44] *** andy has joined the room
[14:21:57] *** Kev shows as "away"
[14:22:44] <SaltyBones> daniel, wouldn't it be better to just mandate that clients always use the origin-id to avoid the problem of dealing with id-rewriting MUCs?
[14:24:26] *** Alex shows as "xa" and his status message is "Auto-Status (untätig)"
[14:26:21] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[14:26:27] <SaltyBones> I mean, certainly removing an unnecessary id would also be desirable but this might be a good intermediate step.
[14:26:37] *** bra shows as "away" and his status message is "Автостатус (неактивен)"
[14:26:37] *** ralphm shows as "online"
[14:26:52] *** Martin has joined the room
[14:26:59] *** Martin shows as "away" and his status message is "Away"
[14:27:03] *** Martin shows as "online"
[14:28:35] *** tim@boese-ban.de shows as "online"
[14:29:44] *** Kev shows as "online"
[14:33:35] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[14:36:19] *** ralphm shows as "online"
[14:39:29] *** winfried shows as "online"
[14:40:35] *** tim@boese-ban.de shows as "online"
[14:41:49] <SaltyBones> Would everything be better if clients would generate proper stanza-ids?
[14:42:15] <SaltyBones> For ...uh...some definition of proper that makes them UUIDish.
[14:43:05] <zinid> they should be strictly monotonically increasing, so we don't need XEP-0198
[14:43:11] <SaltyBones> And by "would everything be better" I specifically also mean, wouldn't that allow us to ditch the other IDs?
[14:43:12] <zinid> and not UUIDs
[14:43:45] <jonasw> zinid, "strictly monotonically increasing" is not gonna happen
[14:43:57] *** ralphm shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[14:44:50] <zinid> because?
[14:44:53] <SaltyBones> zinid, would increasing IDs really solve all the same problems as stream management? Seems unlikely?
[14:45:06] <jonasw> zinid, also, increasing IDs have security implications
[14:45:13] <jonasw> or rather: predictable IDs
[14:45:15] <zinid> SaltyBones, it solves in data replication, but of course XMPP is too unique
[14:45:30] <zinid> version vectors are based on such ids
[14:45:44] *** lumi shows as "away" and his status message is "(Idle 10 min)"
[14:46:03] <zinid> jonasw, what security implications?
[14:46:08] *** jubalh has left the room
[14:46:18] <jonasw> zinid, I think there were some IQ response injection attacks based on predictable IDs
[14:46:32] <jonasw> even though in that case you probably already made the mistake of not verifying the sender of the response
[14:46:37] *** bra shows as "xa" and his status message is "Автостатус (неактивен)"
[14:46:48] <jonasw> error injections would most likely work though because errors can come from entities different than the original recipient
[14:47:23] <zinid> jonasw, this can be fixed by adding routing information
[14:47:27] <SaltyBones> jonasw, maybe this is too much of an assumption for all of XMPP but don't we have transport encryption?
[14:47:45] <zinid> we anyway need this functionality already
[14:48:13] *** bra shows as "online"
[14:48:15] <jonasw> SaltyBones, that doesn’t stop me (jonas@zombofant.net) from sending an iq type="error" id="whateverIguessed" to="you" to break whatever you were doing
[14:48:20] <SaltyBones> zinid, I agree, if this is a problem I don't see why it cannot be exploited right now by somebody who randomly sees the message first
[14:48:33] *** SamWhited shows as "online"
[14:48:39] <jonasw> SaltyBones, if I can guess the ID, I can attack you from off-path
[14:48:47] <jonasw> I can’t do that when I can’t guess the ID
[14:48:54] <jonasw> if I’m on path, you’re right, everything is lost already in XMPP.
[14:49:04] <zinid> but if you cannot match incoming errors against requests you sent then you should really consider to change the job
[14:49:25] <jonasw> zinid, how’d you match incoming errors against requests other than the ID?
[14:49:36] <jonasw> you can’t really use from, because as I said, it might come from an entity you didn’t know about yet
[14:49:55] *** ralphm shows as "online"
[14:50:11] <MattJ> jonasw, no?
[14:50:18] <zinid> jonasw, you can use (from, ID) I think
[14:50:27] <MattJ> errors come from the original recipient that you addressed
[14:50:35] <jonasw> MattJ, even if an s2s error causes an error?
[14:50:37] <MattJ> Yes
[14:50:43] <jonasw> oh okay
[14:50:48] <MattJ> Otherwise it would be a nightmare
[14:53:07] <Dave Cridland> jonasw, There's a "by" attribute, if I remember right, that tells you the generator/reporter of the error.
[14:54:05] *** Dave Cridland has left the room
[14:54:14] <SaltyBones> So, I assume that if I run a h4xx0r server and try to send replies to messages that were not directed to me to some other server they will be discarded?
[14:54:30] <SaltyBones> And the same happens between client and server...?
[14:54:41] <MattJ> SaltyBones, by "replies", you mean error replies?
[14:55:11] <SaltyBones> Whatever magical interferring replies that jonasw was referring to :)
[14:55:15] <SaltyBones> yes, errors ;)
[14:55:39] <MattJ> For that to work you would have to know the original sent message id, and the original sender's full JID
[14:55:50] *** jubalh shows as "online"
[14:55:55] <SaltyBones> But if I did, I could?
[14:56:01] *** ralphm shows as "online"
[14:56:05] <MattJ> and then yes, you could send them whatever you wanted - it would be up to their client to match it up (or not) to one of its outgoing messages
[14:56:34] <SaltyBones> So, the statement that predictable IDs would allow me to spoof responses remains correct?
[14:56:35] <MattJ> so as said above, the client should use (from, id) to identify error responses, not just the id
[14:56:58] <SaltyBones> Because I cannot spoof "from"?
[14:57:02] <MattJ> Correct
[14:57:08] <SaltyBones> Yes, okay...
[14:57:12] <SaltyBones> That's what I was looking for. :)
[14:57:29] *** Dave Cridland shows as "online"
[14:57:34] *** remko has joined the room
[14:57:34] *** remko shows as "online"
[14:57:36] <MattJ> id "abc123" from userA@domainA is different from "abc123" from userB@domainB
[14:58:19] <SaltyBones> And servers only accept s2s with from=...@domainA if they know that the other server is in charge of domainA?
[14:59:34] <Dave Cridland> SaltyBones, That's what they're supposed to do, yes.
[14:59:37] *** remko has left the room
[14:59:43] <SaltyBones> ^^
[14:59:43] *** ralphm shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[14:59:56] <MattJ> SaltyBones, yes, they do
[15:00:07] <Dave Cridland> MattJ, No, they don't, but I love your optimism.
[15:00:10] <MattJ> If they don't, it's a bad bug or not an XMPP server
[15:01:24] <Dave Cridland> SaltyBones, So Metre (my S2S proxy mad thing) does this very strictly, and as a result I can't join this chatroom from my personal server.
[15:01:48] *** lumi shows as "online"
[15:01:55] <Dave Cridland> SaltyBones, Not sure if MattJ is saying that's a bad bug in this server, or if he's saying it's not an XMPP server. :-)
[15:02:04] <MattJ> Dave Cridland, for what reason?
[15:02:05] *** jubalh has left the room
[15:03:07] *** andy has left the room
[15:03:26] *** Alex shows as "xa" and his status message is "Auto-Status (untätig)"
[15:03:27] *** Alex shows as "online"
[15:03:28] <SaltyBones> Dave Cridland, so which messages do you get that don't pass?
[15:03:48] *** Alex has left the room
[15:04:15] <SaltyBones> Bunneh, version xmpp.org
[15:04:15] <Bunneh> SaltyBones: xmpp.org is running Prosody version 0.9.12 on Linux
[15:04:17] <Dave Cridland> MattJ, It's sending a response down the stream obtained by reversing the stream to/from, and not by reversing the stanza to/from. I forget the details, Zash knows.
[15:05:08] <Dave Cridland> MattJ, I don't know if Prosody would accept that or not. Metre doesn't, Openfire does (but because it knows it's multiplexed on the outbound, so weirdness.)
[15:06:02] *** ralphm shows as "online"
[15:06:09] <Dave Cridland> MattJ, For maximum fun, Openfire used to do similar "implicit" auth on outbound, but I stopped that as from 4.2.0.
[15:07:21] <MattJ> I don't really understand, but maybe Zash will enlighten me
[15:09:43] *** ralphm shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[15:09:49] <Dave Cridland> MattJ, Metre sends traffic for d.c.n -> muc.xmpp.org over a stream for d.c.n -> xmpp.org (after adding that domain via dialback).
[15:09:55] <Zash> How far back should I read to have any idea what you are talking about?
[15:10:18] <SaltyBones> 15:58 should suffice
[15:10:30] <MattJ> SaltyBones, and a time machine
[15:10:39] <Dave Cridland> MattJ, Prosody (0.9) responds to the traffic on the reverse stream, xmpp.org -> d.c.n - which it hasn't requested authorization for yet.
[15:11:03] <SaltyBones> Do you mean a timezone machine or a MAM machine?
[15:11:15] *** Alex has left the room
[15:11:35] <Zash> No MAM here
[15:11:36] <Dave Cridland> SaltyBones, For your amusement and mild confusion, XMPP authorizes streams by a 3-tuple of (from-domain, to-domain, direction).
[15:11:40] <Zash> Page up button tho
[15:11:51] *** ralphm shows as "online"
[15:12:03] <Dave Cridland> SaltyBones, Well. One or more of those 3-tuples anyway.
[15:13:11] <MattJ> Dave Cridland, who hasn't requested authorization for what? You mean Prosody connects to d.c.n as 'xmpp.org' and sends stanzas from 'muc.xmpp.org'?
[15:13:23] <Dave Cridland> MattJ, Yup, that.
[15:14:18] *** dwd has joined the room
[15:14:34] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[15:16:17] <Zash> The thing
[15:16:43] <Dave Cridland> MattJ, Like this: DEBUG 2018-02-12T15:14:35 /home/dwd/src/Metre/src/xmlstream.cc:95 : NS296914 - G
ot [399] : <?xml version='1.0'?><stream:stream xmlns:db='jabber:server:dialback'
xmlns:stream='http://etherx.jabber.org/streams' version='1.0' from='xmpp.org' t
o='dave.cridland.net' xml:lang='en' xmlns='jabber:server'><iq id='472-452048' ty
pe='error' to='dave.cridland.net' from='xsf@muc.xmpp.org/tim@boese-ban.de'><erro
r type='cancel'><not-acceptable xmlns='urn:ietf:params:xml:ns:xmpp-stanzas'/></e
rror></iq>

[15:16:57] *** ralphm shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[15:17:15] <dwd> This time I'm joined. Timing, I guess.
[15:17:17] *** ralphm shows as "online"
[15:19:58] <Zash> Uh, what was the thing with that again?
[15:20:35] <Dave Cridland> Zash, The bug itself? Responding to inbound traffic on a multiplexed stream by reversing the wrong domain pair.
[15:20:57] <MattJ> dwd, I don't see how timing would play a part
[15:21:15] <Zash> Dave Cridland: Do you remember where we discussed this?
[15:21:43] <Dave Cridland> MattJ, I think whether I can join is dependent on what streams are open, which is dependent on the session lifetime and activity.
[15:21:55] <Dave Cridland> MattJ, Not timing as in a race condition, mind.
[15:22:17] <Dave Cridland> Zash, Erm. 1:1 messages, I think, until we figured out it definitely wasn't a security issue.
[15:22:30] <Dave Cridland> Zash, Possibly some discussion in jdev.
[15:23:28] <dwd> FWIW, I have to put in something similar to an implicit auth for X2X, anyway. So this may provide a workaround.
[15:23:34] *** Martin shows as "online"
[15:23:48] *** Martin shows as "away" and his status message is "Away"
[15:24:13] *** Alex has joined the room
[15:24:18] *** Martin has left the room
[15:24:34] *** tim@boese-ban.de shows as "xa" and his status message is " (Nicht verfügbar wegen Untätigkeit seit mehr als 15 Minuten)"
[15:24:43] <Zash> MattJ: https://hg.prosody.im/trunk/file/0de0018bdf91/plugins/mod_s2s/mod_s2s.lua#l203
[15:25:04] <MattJ> Zash, exactly where I ended up
[15:25:31] <Zash> stanza from/to may differ from the stream to/from in case of dialback multiplexing
[15:27:26] *** tux shows as "dnd" and his status message is "Work work …"
[15:27:35] *** Steve Kille shows as "away" and his status message is "Hampton"
[15:33:41] *** Alex has left the room
[15:35:40] *** Alex has joined the room
[15:35:52] *** Holger shows as "online" and his status message is "I'm available"
[15:37:12] *** pep. has left the room
[15:37:16] *** pep. shows as "online"
[15:37:40] *** ralphm shows as "online"
[15:37:43] *** bra shows as "away" and his status message is "Автостатус (неактивен)"
[15:38:18] *** Alex has left the room
[15:38:52] *** Martin has joined the room
[15:38:55] *** Martin shows as "away" and his status message is "Away"
[15:39:02] *** Martin shows as "online"
[15:43:07] *** jubalh has joined the room
[15:43:44] *** jubalh has left the room
[15:43:46] *** jubalh has joined the room
[15:44:28] *** Steve Kille shows as "online" and his status message is "Hampton"
[15:44:32] *** tim@boese-ban.de shows as "online"
[15:44:38] <Dave Cridland> MattJ, You switching to FreeBSD?
[15:44:47] <Dave Cridland> MattJ, https://svnweb.freebsd.org/base?view=revision&revision=329166
[15:49:09] *** Holger has left the room
[15:49:21] *** Holger shows as "online" and his status message is "I'm available"
[15:50:08] <zinid> svn o_O
[15:52:40] <edhelas> gitis too mainstream
[15:53:07] <zinid> capitalist pigs use git!
[15:53:30] <Zash> gititis
[15:53:51] <Holger> NetBSD has Lua in the kernel for years!
[15:53:52] <Zash> Dave Cridland: Heh, nice
[15:54:08] <Holger> https://www.phoronix.com/scan.php?page=news_item&px=MTMwMTU
[15:54:11] <Holger> And it has CVS :-)
[15:54:24] <zinid> damn, I only get used to svn...
[15:56:01] <Dave Cridland> Holger, Can you download NetBSD yet, or is it still only available on tape?
[15:56:44] <zinid> Holger, cool stuff btw
[15:57:09] <SamWhited> More importantly, once I download it can I run it on any machines made after 1995? (this is always the problem I had trying to run NetBSD; a few of the ops people at work run it, but they all use *very* old machines)
[15:57:15] *Dave Cridland gets prepared to explain "tape" to the younger readers.
[15:57:41] *** bra shows as "xa" and his status message is "Автостатус (неактивен)"
[15:57:45] <Dave Cridland> SamWhited, Sure! It runs pretty well even modern machines like the Amiga 4000.
[15:58:03] *** nyco has left the room
[15:58:11] <Dave Cridland> (Sorry, that was 1993).
[15:58:29] <SamWhited> Exactly.
[15:58:37] <Holger> Well yes it's about as dead as XMPP :-P
[15:58:52] <SamWhited> I would like to use at *least* a P3.
[15:58:53] <Holger> But it always worked okay-ish for me on non-recent ThinkPads.
[15:59:15] <Holger> I no longer do that though.
[16:00:02] <Holger> cpu0 at mainbus0 apid 0: Intel(R) Core(TM) i3-2120T CPU
[16:00:15] <Holger> ... is the one box I still run NetBSD on.
[16:00:46] <Zash> Uh, was NetBSD the one being difficult, or OpenBSD?
[16:00:57] <Holger> Difficult?
[16:00:57] <SamWhited> Actually, jokes aside, that might be a nice thing to do with my broken old thinkpad; doesn't work very well as a laptop anymore, but it could be a {Free,Net,Open}BSD or Illumos "desktop".
[16:01:15] <Holger> Zash: Isn't all computers difficult except for Apple?
[16:01:39] <SamWhited> This is true… ⤴
[16:02:27] *** Dave Cridland has left the room
[16:02:34] <Zash> Et Apple, Holger
[16:02:57] <Dave Cridland> I find Apple incomprehensibly difficult to use, I must admit.
[16:03:12] <Dave Cridland> I've yet to figure out how to go up a directory consistently in file dialogs.
[16:06:11] <Kev> cd ..
[16:06:19] *** Dave Cridland shows as "online"
[16:06:19] <Kev> Happy to help.
[16:07:05] <Holger> Managing FreeBSD feels more or less like Debian to me, the two others feel a bit more basic but not really harder to use. All dead simple compared to the dances you had to go through when partitioning hard disks or getting X11 to run 20 years ago with any Linux or BSD.
[16:08:11] <Dave Cridland> Holger, Ah, X11. Kids these days don't know they're born.
[16:08:39] <Dave Cridland> Kev, Don't get me started on the antiquated command line tools Apple foists upon you.
[16:08:41] *** remko has joined the room
[16:08:42] *** remko shows as "online"
[16:08:43] <Holger> Yes, we're horribly old.
[16:08:49] *** nyco shows as "online"
[16:09:02] <Dave Cridland> Holger, We're very experienced. I'm sure that's what you meant.
[16:09:04] <Zash> I'm sure wayland will bring back the good old times for those who miss configuring Xorg
[16:09:20] <Holger> Precisely.
[16:09:25] <SamWhited> The only thing I want out of a machine is to sync an old ipod without Rhythmbox crashing…
[16:09:52] <Dave Cridland> SamWhited, Rhythmbox never crashes.
[16:09:59] <Dave Cridland> SamWhited, At that speed it's called "parking".
[16:11:12] <SamWhited> Rhythmbox is one of those people who parallel parks by backing up until they hit the car behind them, then driving forward until they hit the car in front.
[16:12:04] *** efrit has left the room
[16:12:05] *** jubalh has joined the room
[16:13:23] *** Alex has joined the room
[16:13:30] *** Alex shows as "online"
[16:14:23] <SaltyBones> So, why is it that a client cannot create the stanza-id?
[16:15:07] <SaltyBones> Is this a thing about malicious clients or what's going on there?
[16:15:15] *** Dave Cridland has left the room
[16:15:35] <jonasw> SaltyBones, no, unaware clients are sufficient; colliding IDs would lead to interesting issues with MAM
[16:15:38] *** Dave Cridland shows as "online"
[16:16:19] <SaltyBones> jonasw, but if the client is only unaware the server could just return an error informing the client that this ID has been used.
[16:16:54] <Dave Cridland> SaltyBones, But we could use, say, stream-id + attribute-id to form a stable, non-colliding reference identifier.
[16:17:14] <jonasw> SaltyBones, that’d be annoying to handle in the client, and the server would need an O(1) (or similar) way to determine that the ID has been used already...
[16:17:17] <SaltyBones> Yeah, or we could hash things and god knows what else...it seems very solvable...
[16:17:37] <Dave Cridland> SaltyBones, I mean, unless the client isn';t generating unique ids within its stream, in which case it's presumably not going to be fixed to use some other identifier.
[16:17:39] <jonasw> Dave Cridland, stream id + sm-counter?
[16:17:46] <jonasw> that’s verifiable by the server
[16:17:48] <jonasw> and predictable for both
[16:18:11] <Dave Cridland> jonasw, True. But I think we had some ideas on predictability outside the stream?
[16:18:16] <SaltyBones> throw a hash on top to avoid privacy issues, done
[16:18:33] <SaltyBones> Dave Cridland, I thought we dicussed those into oblivion earlier? :)
[16:18:47] <Dave Cridland> SaltyBones, Really I've lost track.
[16:19:00] <jonasw> Dave Cridland, hmmm ... hmac(stream-id, sm-counter)?
[16:19:03] <Dave Cridland> SaltyBones, I have the feeling that nobody quite knows the entire picture here.
[16:19:11] <SaltyBones> I almost lost track and I basically didn't work at all today and just discuss here. :p
[16:19:25] <Dave Cridland> jonasw, HMAC() requires a secret to be of any use.
[16:19:47] <jonasw> Dave Cridland, stream-id.
[16:20:01] <SaltyBones> Seems reasonable on first sight...
[16:20:02] <jonasw> (post-TLS obviously...)
[16:20:05] <Dave Cridland> jonasw, Is that secret? And why HMAC over a hash?
[16:20:21] <SaltyBones> Oh, that would leave out the hash...
[16:20:24] <jonasw> Dave Cridland, I’m fine with hash(stream-id || sm-counter) too, but that’s nearly an hmac ;-)
[16:20:28] <SaltyBones> hmac basically IS a hash
[16:20:32] <SaltyBones> yeah :)
[16:20:49] <Dave Cridland> jonasw, I'm pretty sure it's not. :-)
[16:20:56] <SaltyBones> I am very sure it is. :)
[16:20:58] <jonasw> that’s why I said "nearly"
[16:21:03] <jonasw> I think the concat is slightly different
[16:21:06] <SaltyBones> (yes, nearly)
[16:21:21] <SaltyBones> https://wikimedia.org/api/rest_v1/media/math/render/svg/4edcf0bd8b403c93564b8d7ea91338b3208dea03
[16:21:28] <jonasw> doesn’t matter anyways
[16:21:38] <Dave Cridland> jonasw, An HMAC is two nested hashed with concats and masks, yes.
[16:21:51] <SaltyBones> Okay, so we only disagree on our definition of nearly. ;)
[16:21:59] <Dave Cridland> jonasw, But the security properties are distinct.
[16:22:04] <jonasw> Dave Cridland, I don’t argue that
[16:22:16] <jonasw> (and I’m aware)
[16:22:56] <SaltyBones> Anyway, the suggestion of HMAC(key=stream-id, msg=sm-counter) seems good, doesn't it?
[16:23:04] <Dave Cridland> In any case, given a stanza with a known id on a given stream, we clearly want to be able to predict the MAM id.
[16:23:27] <SaltyBones> I mean, I am also fine with SHA-*(stream-id || counter-sm)
[16:24:04] <SaltyBones> Dave Cridland, but mam-id = stanza-id which is what I am proposing to set to this...
[16:24:13] <Dave Cridland> Question is, do we think the MAM id is likely to become public, and if so, can someone relatively easily figure out the next (or previous) value, and then do Something Bad?
[16:24:56] <SaltyBones> Dave Cridland, that might be the question bit if it only costs us a hash per message to not answer it might also not be...
[16:25:06] <SaltyBones> Dave Cridland, that might be the question but if it only costs us a hash per message to not answer it might also not be...
[16:25:15] <SaltyBones> Dave Cridland, that might be the question but if it only costs us a hash per message to not answer it, it might also not be...
[16:25:43] <Dave Cridland> What I'm wondering, you see, is whether we give clients an algorithm to generate attribute-id and then let them signal to the server that they're going to use globally-unique attribute-ids and the server is allowed to call them out if they don't.
[16:25:48] <jonasw> Dave Cridland, okay, right, this is about the MAM ID. this *probably* doesn’t matter, but if the goal is to consolidate all IDs at some point, choosing a way where the ID can become public is safer
[16:25:49] *** zinid has left the room
[16:26:07] <Dave Cridland> jonasw, That's roughly my thought, yes.
[16:26:48] <jonasw> and at that point, HMAC(…) seems like a sane choice.
[16:26:58] <Dave Cridland> jonasw, Well, that and "But we *HAVE* stanza ids - right there in the stanza!"
[16:27:14] <Kev> How would the server know that they didn't use a globally unique id?
[16:27:33] <SaltyBones> Kev, if the server can simply re-run the generation algorithm and compare results..?
[16:27:39] <jonasw> Kev, if the ID doesn’t follow the algorithm for a globally (predictable, for the server and client only) unique ID
[16:27:41] <Dave Cridland> Kev, Well, it'd know if they used no id at all, and if it saw a collision within a stream.
[16:28:28] <Kev> If we only cared about collisions within a stream, we could trivially solve all issues already.
[16:28:28] <SaltyBones> And that is how you can tell that Kev is a VIP
[16:28:36] <SaltyBones> when he says something three people reply!
[16:29:03] <Dave Cridland> Kev, Well, we care about collisions for a user's account for MAM, and no further.
[16:29:08] <Kev> But having a useful id to refer to a message would be jolly useful, and I don't see how we can get there yet.
[16:29:52] <SaltyBones> Okay, so you're saying when two servers federate and you assume one of them is malicious how can the other make sure he doesn't get duplicate IDs?
[16:30:07] <Kev> Or non-malicious server with a malicious client.
[16:30:10] <Dave Cridland> SaltyBones, Do they care? ANd if not, should they?
[16:30:12] <jonasw> SaltyBones, you don’t use other servers stanza-ids in your own MAM
[16:30:22] <SaltyBones> Well, the malicious client would be detected by the server.
[16:30:23] <Kev> jonasw: Bloody useful if you could, though, no?
[16:30:37] <Kev> SaltyBones: How would the server know it's malicious, if it doesn't know all globally generated ids?
[16:30:44] *** Alex has left the room
[16:30:52] <SaltyBones> Kev, it can simply check if the client generates their IDs correctly...
[16:30:54] <Dave Cridland> Kev, Again, why would a server care?
[16:30:55] <Kev> Without exploding the size of IDs, at least.
[16:30:56] <jonasw> (we should mandate the use of shakespeare-inspired adjectives in each bloody sentence in this room, by the way)
[16:31:12] <jonasw> Kev, how would it be useful?
[16:31:36] <Kev> I'd like to send a reference to a stanza, where the reference makes sense in my archive, the MUC/MIX archive, and the user's archive.
[16:31:37] *** tim@boese-ban.de shows as "away" and his status message is " (Abwesend wegen Untätigkeit für mehr als 5 Minuten)"
[16:31:42] <Dave Cridland> jonasw, I don't think "bloody" is Shakespeare. "Submarine" is, though, as far as I remember.
[16:31:45] <SaltyBones> Kev, you simply make it a hash or hmac of something that both the client and the server know and the server can check the calculation.
[16:32:15] *** Holger shows as "away" and his status message is "Auto-away (idle)"
[16:32:16] <Kev> "simply" :)
[16:32:28] <SaltyBones> Kev, but doesn't that only involve you, the muc/mix server and $other_client? i.e. only one server...?
[16:32:31] <Dave Cridland> Kev, Is the (from, id) attributes of the stanza sufficient? If not, why?
[16:32:43] <Zash> If this is for MAM purposes, then I will cry if I can't stick some time based bits into the ID
[16:32:48] <Kev> Dave Cridland: No, because the from gets rewritten.
[16:32:56] <Dave Cridland> Kev, When?
[16:33:03] *** tim@boese-ban.de shows as "online"
[16:33:09] <SaltyBones> Zash, what?
[16:33:12] <Kev> In a MUC or a MIX.
[16:33:33] <Zash> My MAM ids are basically yyyy-mm-dd-random
[16:33:41] <jonasw> Kev, those are different use-cases I think.
[16:34:01] <Kev> jonasw: If we cherry-pick the trivial use cases, it's going to be trivial to solve them :)
[16:34:06] <jonasw> Kev, this is where origin-id makes more sense. It is generated by the client. If the client doesn’t ensure that there’s enough entropy in there, references to their message won’t work.
[16:34:08] <SaltyBones> Zash, and this is important because?
[16:34:09] <Dave Cridland> Kev, Ah, so given a message from a MUC fanout, do you want to be able to identify the originating id? I'm not actually sure you always want to allow that.
[16:34:32] *** Alex has joined the room
[16:35:01] <SaltyBones> I still don't get it. Why does from get rewritten?
[16:35:09] <SaltyBones> And what does that have to do with anything? :D
[16:35:32] <jonasw> SaltyBones, when you send a message from saltybones@yourdomain.example/client-resource to this MUC, we see the message from xsf@muc.xmpp.org/SaltyBones
[16:35:37] <jonasw> so the from is being rewritten.
[16:35:52] <jonasw> and if references are based on (from, id) pairs, you can’t use the same reference we all can use for your message
[16:36:20] <jonasw> (simplified, of course we could still refer to the reflection of the message, but that would break down with MIX, too, I think)
[16:36:26] *** Dave Cridland has left the room
[16:36:34] *** Dave Cridland shows as "online"
[16:36:42] <Kev> jonasw: Well, no, because if you can fake someone else's id, suddenly you can maliciously change the target of a reference.
[16:37:09] <jonasw> Kev, good point. so we indeed need something like (from, id) :(
[16:37:17] <SaltyBones> Wait what?
[16:37:24] <zinid> lol
[16:37:25] <SaltyBones> How can you now fake stuff again?
[16:37:26] <jonasw> but given that all our fanouts and from-rewriting things do actually do reflections, I’m not sure if that’s so bad, Kev?
[16:37:57] <jonasw> your local archive woudl still be able to resolve everything
[16:38:00] <Kev> jonasw: Well, it enforces a round-trip before you can refer to or correct anything, which isn't ideal.
[16:38:19] <jonasw> Kev, you normally know the "from" you’ll be having.
[16:38:26] <jonasw> and since you set the origin-id yourself, you already know everything you need
[16:38:44] <jonasw> (meh, races with server-enforced nickname changes in MUC)
[16:38:44] <Kev> Point.
[16:38:56] *** winfried shows as "away" and his status message is "sssssttttt! my computer fell asleep"
[16:39:26] <jonasw> (but I guess those will be rare enough for us not to care)
[16:39:27] <SaltyBones> halp! :(
[16:39:30] *** ralphm shows as "online"
[16:39:37] *** Dave Cridland has left the room
[16:39:39] <Kev> SaltyBones: Because there are potentially malicious entities.
[16:39:51] *** tim@boese-ban.de has left the room
[16:40:04] *** daniel has left the room
[16:40:05] *** Dave Cridland shows as "online"
[16:40:07] *** daniel shows as "online"
[16:40:14] <SaltyBones> Are these only clients or servers as well?
[16:40:14] <Kev> jonasw: Still doesn't help where a user generates the same id twice, I think, and you're left with ambiguous references/corrections/whatever.
[16:40:17] <Kev> SaltyBones: Both.
[16:40:34] <SaltyBones> So why do you connect to a malicious server and expect things to work?
[16:40:40] <SaltyBones> Maybe that's not a good approach.
[16:40:54] <jonasw> Kev, that’s not an issue. put 256 bits of entropy in there and you are officially allowed to not care
[16:41:02] <Kev> Ah, if you have a mechanism for identifying malicious intent on the Internet, you could be a very famous person.
[16:41:05] <jonasw> Kev, and if we want to have defined behaviour in that case, always assume the most-recent message
[16:41:23] *** ralphm has left the room
[16:41:32] <Kev> jonasw: That only works for accidents, rather than manipulation, no?
[16:41:37] *** marc has left the room
[16:41:50] <SaltyBones> Kev, but you cannot manipulate as long as the server is checking.
[16:41:50] <jonasw> Kev, indeed
[16:41:58] <jonasw> but if you manipulate your own origin-ids, it’s your own fault?
[16:42:01] <SaltyBones> Of course if we have malicious servers now a bit more work is required. :)
[16:42:13] <Kev> And I'm concerned that 'most recent' falls apart if you can manage different people receiving different subsets.
[16:42:18] <jonasw> can one inflict harm by producing duplicate origin-ids?
[16:42:42] <jonasw> I’m not convinced that this is actually an issue.
[16:42:44] <Kev> Most likely.
[16:42:54] <Kev> If I can cause the 'wrong' message to get corrected.
[16:42:59] <jonasw> maybe with something like Reactions based on (from, origin-id) references.
[16:43:02] <Kev> Or a like to apply to the wrong message.
[16:43:42] <Kev> If I can manipulate you into liking a Godwin instead of something sensible, that's not ideal.
[16:43:45] *** andy has joined the room
[16:43:55] *** daniel shows as "online"
[16:44:18] <jonasw> meh
[16:44:18] *** bra shows as "online"
[16:44:32] <SaltyBones> But messages aren't authenticated anyway, if you are a malicious server you can claim to have received whatever likes from me...
[16:44:39] <jonasw> Kev, then the only way is a MUC/MIX stanza ID generated by the MUC/MIX and waiting for a round-trip before references can be done.
[16:44:55] <Kev> SaltyBones: Other way around.
[16:44:59] <jonasw> SaltyBones, all it needs for now is a malicious client though
[16:45:00] *** daniel shows as "online"
[16:45:14] <jonasw> SaltyBones, or a malicious client+server pair if we do the hmac-stanza-id thing for origin-id
[16:45:22] <jonasw> I can still run my own server and make it fake origin-ids
[16:46:21] *** daniel has left the room
[16:46:24] *** daniel shows as "online"
[16:46:41] *** andy has left the room
[16:46:42] *** andy has joined the room
[16:46:49] <zinid> your own malicious server?
[16:46:49] <jonasw> Kev, I’d however argue that actively, from a client/own server side, make participants in a MUC/MIX receive only parts of the conversation would be rather trikcy
[16:47:00] <jonasw> and targeted parts even
[16:47:22] <jonasw> and that without them suspecting that things are broken in funny ways and thus not trusting the system anymore
[16:47:25] <Kev> jonasw: Ah, that's ok then, no exploits have ever involved doing anything tricky :)
[16:47:39] <jonasw> Kev, I see your point, but I’m not sure this is actually something which is reasonably exploitable.
[16:47:43] <jonasw> but yes
[16:47:55] <jonasw> unless we let the ID be generated by the fanout service, there’s no way to be sure I’m afraid
[16:48:04] <jonasw> I mean, I don’t have an issue with that, I’m fine with that even.
[16:48:11] <Kev> Yes. I don't see a way of solving this, which was why I brushed it under the carpet at the Summit.
[16:48:14] <jonasw> complicates things though
[16:48:56] *** winfried shows as "xa" and his status message is "sssssttttt! my computer fell asleep"
[16:49:46] *** Guus has joined the room
[16:50:13] *** jjrh has left the room
[16:50:17] *** jjrh shows as "online"
[16:51:54] *** marc has left the room
[16:52:02] <jonasw> Kev, I’m having a really hard time constructing a successful attack which wouldn’t be seen by the victim in the MUC case
[16:52:10] <jonasw> even when I omit arbitrary messages
[16:52:13] <jonasw> to only some participants
[16:52:21] <jonasw> (which would be really really tricky to achieve targetedly I think)
[16:53:10] <jonasw> ah, now I have it
[16:53:56] <jonasw> this is the scenario:
1. user A, "bad statement", origin-id=1
2. user A, "good statement", origin-id=1
3. user B, like (from=user A, origin-id=1)

if (2) isn’t seen by all users, they see user B liking "bad statement" instead of "good statement"
[16:53:59] *** tux has joined the room
[16:55:12] <jonasw> with an arbitrary amount of messages between (2) and (3), it is also not too difficult to make people not see (2) in a MAM-less MUC.
[16:55:42] <Kev> Indeed.
[16:55:57] *** Dave Cridland has left the room
[16:56:32] <jonasw> I don’t think there’s a way unless the MUC does things there
[16:56:55] <jonasw> (i.e. if the MUC generates the ID?)
[16:56:55] <SaltyBones> But, all it requires to prevent that is for the MUC to check that the ID is unique just like the assumed-non-malicious server did in our previous discussionn...
[16:56:56] <jonasw> (i.e. if the MUC generates the ID)
[16:57:09] <jonasw> SaltyBones, but that is a rather expensive check to do
[16:57:22] <jonasw> you need to record IDs for all eternity, or have a defined way to generate the ID which can be executed by clients
[16:57:33] <jonasw> the latter would be tremendously tricky to get to work synchronously
[16:57:41] <jonasw> but possible...
[16:57:59] <SaltyBones> Which is a little more tricky because there is no obvious common "secret" like the stream-id but a simple Hash("counter") would do
[16:58:14] <jonasw> something like hash(message-counter || presence-id), where presence-id is an ID assigned to the client on join
[16:58:16] <SaltyBones> Kev, not that in this case I used the word simple again but I was totally lying...
[16:58:22] <SaltyBones> Kev, note that in this case I used the word simple again but I was totally lying...
[16:58:32] <jonasw> hash counter doesn’t really work with multi-session nicks I think, it would lead to collisions or races
[16:58:58] <SaltyBones> jonasw, yeah it was just to get the idea. Indeed you would at least need some salt.
[16:59:27] <SaltyBones> But if the server gives you the salt on join and you use that to generate the IDs it should be good again.
[16:59:30] <jonasw> that would be verifable by the MUC and the MUC would drop stanzas which do not adhere to this schema
[16:59:38] <jonasw> (or rather, reject)
[16:59:49] <SaltyBones> and since the counter can be per-muc there is no issue with having it
[16:59:50] *** Dave Cridland shows as "online"
[16:59:51] *** mimi89999 shows as "online"
[16:59:55] <jonasw> per client even
[16:59:57] <SaltyBones> I think :p
[17:00:02] <jonasw> Kev, ^
[17:00:03] <SaltyBones> well, yes per client and per muc
[17:00:11] <SaltyBones> just saying it doesn't seem to be a privacy issue
[17:00:14] <jonasw> SaltyBones, yeah
[17:00:18] *** Holger shows as "online" and his status message is "I'm available"
[17:00:30] *** remko has joined the room
[17:00:31] *** remko shows as "online"
[17:00:32] <jonasw> so origin-id = one-way-function(presence-id, message-counter), where presence-id is assigned on MUC join
[17:00:45] *** andy has left the room
[17:00:52] <jonasw> only need to define what happens when message-counter wraps around or becomes large or something
[17:01:05] <jonasw> (same thing for SM by the way, SM has wraparound semantics)
[17:01:23] *** efrit has joined the room
[17:01:39] <SaltyBones> O_o
[17:01:52] <SaltyBones> Y U SEND SO MANY MESSAGES!?!!
[17:01:58] *** ralphm has joined the room
[17:02:01] <jonasw> SaltyBones, itym "y u have so long sessions"
[17:02:27] *** andy has joined the room
[17:02:35] <SaltyBones> still, if this is defined somewhere we can probably renegotiate the salt at that point...
[17:02:36] <jonasw> while wrapping around a 64bit counter in a single session is sure challenging, we need to be prepared if this is to be solid :)
[17:02:40] <jonasw> yeah
[17:02:50] *** Maranda has joined the room
[17:03:04] <zinid> I hear counter?
[17:03:10] <zinid> sorry, I don't track the discussion
[17:03:40] <jonasw> zinid, I’m not going to repeat everything you can read in the backlog :)
[17:03:54] <zinid> as you wish
[17:04:00] <zinid> not that I care
[17:04:08] <SaltyBones> counter works in this case because you can restart counting when rejoining the muc
[17:04:14] <zinid> whatever you will end up will be shit anyways, so
[17:05:39] *** Maranda has left the room
[17:05:52] <zinid> you already created 4 fucking ids: stanza-id, origin-id, attribute-id and SM id
[17:06:01] <zinid> maybe it's time to stop and think?
[17:06:18] <jonasw> what is an sm ID?
[17:06:23] <zinid> h
[17:06:25] <jonasw> right
[17:06:44] <jonasw> zinid, all of this is about reducing the number of IDs, so I thnik this is kinda what we’re doing?
[17:07:06] <zinid> jonasw, from what I see you want to add yet another counter
[17:07:16] <jonasw> zinid, to generate origin-ids, yes
[17:07:20] <jonasw> for MUCs and MIXes
[17:07:25] <jonasw> to make them verifiable by the service
[17:07:45] <jonasw> zinid, or rather, we’re replacing origin-id by some one-way-function(counter)
[17:08:23] <zinid> how this will cover sm id?
[17:08:41] *** waqas has joined the room
[17:09:11] <SaltyBones> We haven't discussed sm-id, yet. Only the other three.
[17:09:18] <SaltyBones> I actually have no clue what sm-id is. :)
[17:09:31] <zinid> SaltyBones, then I need to wait while you recognize the problem 😉
[17:09:57] <SaltyBones> zinid, so reducing 4 -> 2 is not enough, eh? :p
[17:10:18] <jonasw> zinid, sm-id stays, but stanza-id (and attribute-id) could potentially become one-way-function(stream-id, sm-id)
[17:10:20] <zinid> 2 is enough, however I think first should be counter and second should be routing information, and not an ID
[17:10:52] <jonasw> stanza-id being used for the archive only, origin-id being used throughout the network together with the sender address for refernecing a specific message
[17:10:52] *** lovetox has joined the room
[17:11:40] <zinid> what ID will be used to sync?
[17:11:58] *** ralphm has joined the room
[17:12:53] <zinid> with SM IDs you just create a pointless queue in c2s state
[17:14:36] *** Ge0rG has left the room
[17:14:59] *** Ge0rG shows as "online"
[17:15:24] <SaltyBones> Okay, I think we are trying to solve a problem that is very orthogonal to stream management. But we have only discussed this a bit and it might very well not work even for what we want it to do.
[17:15:49] <zinid> SaltyBones, I don't think SM should be separated from archive
[17:15:52] <zinid> it's pointless
[17:16:37] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[17:16:37] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[17:16:38] <SaltyBones> Why?
[17:16:53] <jonasw> zinid, stanza-id is used for ysnc
[17:16:56] <jonasw> (for MAM sync that is)
[17:17:03] *** Dave Cridland shows as "online"
[17:17:10] <zinid> SaltyBones, because why you need this separation?
[17:17:17] *** dwd shows as "online"
[17:17:29] <zinid> SaltyBones, you keep messages in MAM and then put them in c2s SM queue
[17:17:48] <zinid> why can't you just inc(counter), store in MAM and send via c2s?
[17:18:24] <SaltyBones> Why are you asking me that? I am 100 % sure that you know more about SM than I do!
[17:18:56] <zinid> what I know about SM is that I need to maintain stupid queues inside c2s processes, which sucks as hell
[17:19:10] <zinid> even though a client can request those messages via MAM
[17:20:39] <zinid> if you receive an ID out of order, you just reconnect and ask everything started from the ID you received
[17:20:52] <zinid> and server will send you this from archive
[17:20:55] <zinid> and vice versa
[17:22:01] <SaltyBones> But SM-IDs are per-stream not per-message, right?
[17:22:02] *** andy has left the room
[17:22:13] *** tux shows as "online"
[17:22:22] <SaltyBones> At least that's what it looks like in https://xmpp.org/extensions/xep-0198.html
[17:22:22] <zinid> when you connect a server, you provide last seen ID and it will resend you everyting what's greater than this ID
[17:22:39] <zinid> SaltyBones, yes, they are totally separated instances
[17:22:57] <SamWhited> What if the thing you missed was an IQ or a presence that isn't stored in MAM?
[17:23:18] <SamWhited> If you temporarily disconnect and miss something, SM acks allow you to find out. Doesn't help as much if it only covers messages.
[17:23:26] <zinid> SamWhited, I think we can drop them
[17:23:31] *** ralphm has joined the room
[17:23:46] <zinid> SamWhited, we already don't care about IQs with Push, so why would we start care?
[17:23:57] <zinid> try to make jingle call when I'm in "push" mode
[17:24:15] <SamWhited> Does that not work? That does seem like something we need to care about to me
[17:24:33] <zinid> SamWhited, but we don't and we cant with push stuff
[17:24:45] <jonasw> shouldn’t an IQ trigger a push? :-O
[17:24:52] <zinid> SamWhited, what if I want to receive your software version and you're in push mode?
[17:24:56] <SamWhited> That seems like a problem that needs fixing then, not an example of something being done right that we should copy.
[17:25:04] *** Holger shows as "away" and his status message is "Auto-away (idle)"
[17:25:05] *** Steve Kille has left the room
[17:25:29] <zinid> anyway, you can keep IQs in MAM if you prefer
[17:25:52] <zinid> you need to keep subscription requests for sure there
[17:26:11] <zinid> we keep them already, but in a separate database due to historic reasons only
[17:27:22] *** Martin has left the room
[17:27:32] *** andy has joined the room
[17:28:28] *** Steve Kille has joined the room
[17:28:29] *** Steve Kille shows as "online" and his status message is "Hampton"
[17:28:43] <jonasw> IQs (which are inherent request-response, with exactly one response) in MAM sounds like a terrible idea.
[17:29:47] <SamWhited> yah, now you have to try and figure out which IQs are time sensitive, which are important to store, etc. it seems like a comlicated way to work around having a stanza counter.
[17:30:43] *** bra shows as "away" and his status message is "Автостатус (неактивен)"
[17:31:10] *** bra shows as "online"
[17:31:26] <zinid> SamWhited, but you have to decide now too: for example, after 5 minutes of inactivity (by default) a server just bounce all IQs from SM queue
[17:31:37] <zinid> *bounces
[17:31:42] *** Steve Kille shows as "online" and his status message is "At Home"
[17:31:47] <SamWhited> I'm not saying it's perfect, just that this doesn't seem like a good fix.
[17:32:02] <zinid> a fix? the behaviour is the same
[17:32:07] *** SaltyBones has left the room
[17:32:14] <zinid> but we let a client to decide which IQ to reply
[17:33:04] <SamWhited> Except now you have to store all stanzas in MAM, or not store some stanzas and risk those being the ones that are dropped and you don't know you missed something. The point of SM is to make sure you know if you lose something, which is often important.
[17:33:32] *** andy has left the room
[17:33:36] <zinid> do you really want to loose incoming call?
[17:33:56] <zinid> that's how it works now: if somebody calls you and you're not connected you loose any track
[17:34:32] <SamWhited> As I said, I'm not claiming that the current solution is good; just that this makes it worse.
[17:34:42] <zinid> no, it makes it better
[17:35:06] <zinid> you just need to define what to store and what not, well, it's too much to redesign, yes
[17:35:31] <SamWhited> If you don't store everything you now lose the ability to detect connection drops though
[17:35:58] <zinid> sigh
[17:36:22] *** Dave Cridland has left the room
[17:36:48] *** Guus has left the room
[17:36:49] *** Guus has joined the room
[17:38:55] <zinid> not sure how will you address missing call though
[17:39:01] <zinid> with your approach
[17:39:22] *** Dave Cridland shows as "online"
[17:40:12] <SamWhited> I do not have an approach; I agree that is a problem that needs solving though.
[17:40:20] *** Dave Cridland has left the room
[17:40:31] *** Dave Cridland shows as "online"
[17:40:32] *** Dave Cridland has left the room
[17:41:16] *** Dave Cridland shows as "online"
[17:41:25] *** Dave Cridland has left the room
[17:41:35] *** lumi shows as "away" and his status message is "(Idle 10 min)"
[17:42:02] *** Dave Cridland shows as "online"
[17:44:29] *** Dave Cridland has left the room
[17:47:26] *** Dave Cridland shows as "online"
[17:47:38] *** Steve Kille shows as "away" and his status message is "At Home"
[17:48:45] <SaltyBones> It seems like we need storage for management messages and storage for actual messages. I don't see why these should be mixed up.
[17:48:53] *** Guus has left the room
[17:49:43] *** rion has left the room
[17:49:44] *** Dave Cridland has left the room
[17:50:28] *** Dave Cridland shows as "online"
[17:52:23] *** SaltyBones shows as "online"
[17:53:49] *** Guus has joined the room
[17:54:13] *** Holger shows as "online" and his status message is "I'm available"
[17:54:44] *** remko has joined the room
[17:54:44] *** remko shows as "online"
[17:55:19] <zinid> we need to decide what to store and what not
[17:55:22] *** Dave Cridland has left the room
[17:55:35] *** Dave Cridland shows as "online"
[17:55:42] <SaltyBones> hm..that distinction was wrong, yeah
[17:55:46] *** Steve Kille shows as "online" and his status message is "At Home"
[17:55:56] <SaltyBones> there should be a storage until delivered and an optional long term storage
[17:56:02] *** Dave Cridland has left the room
[17:56:04] <SaltyBones> users might not even want mam :p
[17:56:17] *** Dave Cridland shows as "online"
[17:56:19] <zinid> if you don't want mam then just lose messages
[17:57:44] <SaltyBones> Okay, let me rephrase, some users might not want long-term storage of messages...
[17:58:12] <zinid> not sure why this should be specific to any approach
[17:59:01] <SaltyBones> Because I think that's what MAM is intended for and that's why using it for all messages as you suggest is strange and might have unexpected results if people built it under the assumption that it is for long term storage...
[17:59:13] *** bra has left the room
[17:59:14] *** bra has joined the room
[17:59:49] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[17:59:49] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[17:59:53] <zinid> servers can drop MAM archives on reconnect for example, what's the problem?
[18:00:01] <SaltyBones> Okay...
[18:00:04] <zinid> or later, when it's delivered
[18:00:09] <SaltyBones> this is just me confusing MAM and the other thing again...
[18:00:16] <SaltyBones> for the one-millionth time...
[18:00:49] <Ge0rG> What if I want my messages both on my pc and my mobile? I can't just drop MAM when my pc is online
[18:01:23] <zinid> you can if everything is delivered
[18:02:43] <zinid> anyway, what you are trying to say is a partial replication, and this problem is very hard to resolve
[18:02:56] <SaltyBones> what is the other thing again, server side archiving?
[18:03:05] *** Dave Cridland shows as "online"
[18:03:05] *** dwd shows as "online"
[18:04:33] *** ralphm has joined the room
[18:05:34] <SaltyBones> zinid, does MAM know if everything was delivered?
[18:05:54] *** Seve/SouL shows as "online"
[18:06:51] *** Seve/SouL has left the room
[18:07:00] *Ge0rG had a very concerning realization about guessable IDs and packet filters in smack earlier today.
[18:07:41] <zinid> SaltyBones, well, no, I think you cannot know that in general case
[18:07:54] <SaltyBones> zinid, but that is what SM does, right?
[18:08:16] <zinid> SaltyBones, well, if you introduce acks, then yes
[18:08:48] <SaltyBones> Hm.... :)
[18:09:21] <Holger> All this is orthogonal to the question whether having two separate stanza/message queues is sane. I agree with zinid that it isn't.
[18:09:59] <Zash> Two whatnow
[18:10:31] <Ge0rG> I'm saying for many years now that we need to replace 0184, 0198, 0280 and 0313 with one single proper message syncing thing.
[18:10:41] <Holger> SM is already mostly just an optimization, and I think we should fix the remaining issues to make stream management superfluous.
[18:10:49] <SaltyBones> I just want to point out that this is orthogonal to our earlier discussion about IDs :)
[18:11:04] <zinid> SaltyBones, I said "no" in general case because we have Bysantine General Problem, it's unresolvable, what we can gurarantee is sequential consistency
[18:11:05] *** Dave Cridland has left the room
[18:11:07] <Ge0rG> SaltyBones: please show your ID before being allowed into this room :P
[18:11:13] *** Dave Cridland shows as "online"
[18:11:34] *** Tobias has joined the room
[18:11:34] <SaltyBones> Holger, does anybody already know what those issues are?
[18:12:02] *** lumi shows as "online"
[18:12:30] *** daniel has left the room
[18:12:44] *** Guus has left the room
[18:12:55] *** Guus has joined the room
[18:12:59] <Zash> Ge0rG: Kidnap some server devs and lock yourself in a room until that one single proper message syncing thing to rule them all is properly implemented and XEP'd
[18:13:03] *** daniel shows as "online"
[18:13:18] <Holger> SaltyBones: Syncing of outgoing messages (in a sane way) and maybe avoiding some round trips during session startup.
[18:13:36] <zinid> Zash, and watch them die 😉
[18:14:14] <Ge0rG> Zash: I'll lock you and zinid up in that room, and watch the live stream
[18:15:45] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[18:15:45] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[18:16:10] <SaltyBones> Holger, so, can we just turn off stream management and leave MAM and see what happens to find out?
[18:16:47] <Holger> Sure you can :-)
[18:16:53] <SaltyBones> Or do you think fixing up MAM is a bad idea and something new is required?
[18:17:17] *** jonasw shows as "away"
[18:18:04] <Zash> SaltyBones: That's what I did, actually. Haven't died from SM-lessness yet.
[18:18:29] *** Steve Kille shows as "away" and his status message is "At Home"
[18:18:51] <Holger> SaltyBones: I think you can already implement proper sync with MAM as-is. But something new is required to let clients implement this in a sane way, without having to de-duplicate and whatnot.
[18:19:09] <zinid> that's my point: what we need is to store messages and some other restricted stuff and call it a day
[18:21:40] <SaltyBones> Holger, so would unique message IDs solve this?
[18:23:59] <SaltyBones> I mean, at least de-duplication would be reasonably easy then, I guess.
[18:23:59] *** jjrh has left the room
[18:24:06] *** jjrh shows as "online"
[18:24:17] *** Dave Cridland shows as "online"
[18:24:17] *** dwd shows as "online"
[18:24:27] <SaltyBones> Of course some sort of counter makes more sense in this case...
[18:25:04] <Zash> SM has counters. MAM has server-issued, guarante to be unique message ids.
[18:25:16] *** ralphm has joined the room
[18:25:25] <SaltyBones> Why are IDs not a counter?
[18:25:31] <Holger> Not sure what sort of uniqueness we need to solve what problem. Didn't read the backlog, sorry :-) The thing missing for proper sync of outgoing messages is an algorithm to compute the MAM IDs of outgoing messages.
[18:25:37] <SaltyBones> I mean the MAM-IDs....or is that the same as the stanza-ids?
[18:25:47] <Holger> (Sorry, typed this before reading the last few messages.)
[18:25:55] <zinid> Zash, now answer the question "get me last messages I didn't receive" with current SM and MAM approach 😉
[18:25:59] <SaltyBones> Holger, what do you mean by outgoing? Messages that we sent?
[18:26:07] <Holger> Yes.
[18:26:27] *** lovetox shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[18:26:31] <zinid> if you say "time", then no
[18:26:34] <Zash> zinid: query after = id of last message I saw
[18:26:35] <SaltyBones> Okay, that should be solved by the hash idea....if it works :p
[18:26:52] <Zash> cry over outgoing messages sent after that
[18:27:11] <Holger> SamWhited: I need to know their IDs so I can tell the server "give me all messages since $id".
[18:27:12] <SaltyBones> But indeed if we want to query MAM by a "point in time" or "counter" unique IDs are not really the best thing :)
[18:27:38] *** remko has joined the room
[18:27:39] *** remko shows as "online"
[18:27:54] <SaltyBones> Holger, but then the server still has to search the MAM archive for that ID and give you everything after it....
[18:27:55] <zinid> Zash, define after in distributed system 😉
[18:27:59] <SaltyBones> So a counter would be much better.
[18:28:13] <Holger> SaltyBones: Sure?
[18:28:19] <Zash> zinid: MAM archive ids on incoming messages
[18:28:21] <SaltyBones> Right?
[18:28:26] <Zash> zinid: after is a MAM term
[18:28:34] <zinid> Zash, but you need some ordering
[18:28:34] *** pep. has left the room
[18:29:38] <Zash> zinid: huh
[18:29:48] *** lovetox shows as "online"
[18:29:54] <Zash> zinid: XMPP streams are ordered
[18:30:53] <zinid> yes, but you need to maintain a timestamp index in the database
[18:31:05] <Zash> You've lost me
[18:31:08] *** Dave Cridland has left the room
[18:31:09] *** Guus has left the room
[18:31:09] *** Guus has joined the room
[18:31:45] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[18:31:45] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[18:31:48] *** Dave Cridland shows as "online"
[18:32:26] <zinid> well, you're probably right that if we assume timestamp ordering then we don't need counters and SM at all
[18:33:09] <Zash> Huh?
[18:33:29] <Zash> In a MAM query, 'after' is a field that carries a MAM archive ID
[18:33:50] <zinid> so, what's the ordering? 🙂
[18:33:54] *** pep. shows as "online"
[18:34:01] *** andy has joined the room
[18:34:02] <zinid> ID+1?
[18:34:23] <zinid> from what I understand you use timestamp+id, which means time ordering
[18:34:25] <Zash> Ordering?
[18:34:47] <Zash> As I said, you lost me. I have no idea what any of us are talking about anymore.
[18:34:57] <zinid> ok
[18:35:30] *** vanitasvitae has left the room
[18:41:41] *** lskdjf has left the room
[18:41:45] *** Dave Cridland shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[18:42:22] *** dwd shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[18:42:53] *** Guus has left the room
[18:43:15] *** Tobias has joined the room
[18:43:37] *** Dave Cridland has left the room
[18:44:05] *** Dave Cridland shows as "online"
[18:46:34] *** remko has joined the room
[18:46:34] *** remko shows as "online"
[18:46:53] *** daniel has left the room
[18:46:59] *** daniel shows as "online"
[18:47:39] *** rion has joined the room
[18:47:51] *** Dave Cridland has left the room
[18:49:00] *** lumi shows as "away" and his status message is "(Idle 10 min)"
[18:49:02] *** Dave Cridland shows as "online"
[18:52:17] *** suzyo has joined the room
[18:52:22] *** Dave Cridland has left the room
[18:52:32] *** Dave Cridland shows as "online"
[18:53:45] *** andy has left the room
[18:53:48] *** bra shows as "away" and his status message is "Автостатус (неактивен)"
[18:56:31] *** ralphm has joined the room
[18:57:08] *** Dave Cridland has left the room
[18:57:15] *** Dave Cridland shows as "online"
[18:57:32] *** rion has left the room
[18:58:42] *** Kev shows as "away"
[19:00:20] *** Syndace has left the room
[19:00:23] *** Syndace has joined the room
[19:03:29] *** Dave Cridland has left the room
[19:03:40] *** Dave Cridland shows as "online"
[19:03:44] *** blabla has left the room
[19:03:48] *** blabla has joined the room
[19:04:41] *** Dave Cridland has left the room
[19:05:20] *** lumi shows as "online"
[19:05:20] *** Dave Cridland shows as "online"
[19:06:36] *** ralphm has joined the room
[19:09:18] *** Dave Cridland has left the room
[19:09:26] *** Dave Cridland shows as "online"
[19:10:06] <Ge0rG> https://summerofcode.withgoogle.com/organizations/?sp-page=5 no xsf?
[19:10:46] <zinid> BEAM community is there 😛
[19:12:45] *** Kev shows as "online"
[19:13:10] *** andy has joined the room
[19:13:49] *** bra shows as "xa" and his status message is "Автостатус (неактивен)"
[19:14:37] *** jjrh has left the room
[19:14:56] *** jjrh shows as "online"
[19:16:42] *** remko has joined the room
[19:16:42] *** remko shows as "online"
[19:16:51] *** ralphm has joined the room
[19:18:25] *** andy has left the room
[19:21:44] <moparisthebest> SaltyBones: like 98% of what you were talking about is here https://github.com/moparisthebest/xmpp-ircd
[19:22:18] *** Dave Cridland has left the room
[19:22:19] *** Dave Cridland shows as "online"
[19:22:21] *** Dave Cridland has left the room
[19:22:22] *** Dave Cridland shows as "online"
[19:22:26] *** Dave Cridland has left the room
[19:22:55] <moparisthebest> Basically it works if IRC users don't want nickserv or chanserv but they do and I never got back to it :)
[19:23:00] *** Dave Cridland shows as "online"
[19:23:37] <SaltyBones> moparisthebest, you mean of a way to connect to a muc using the irc protocol?
[19:23:43] <moparisthebest> Yes
[19:24:16] <moparisthebest> Running that makes a muc look like an IRC server to an IRC client
[19:24:28] *** Dave Cridland has left the room
[19:24:42] <SaltyBones> I'm pretty sure nobody wants that. At least I don't! :D
[19:24:55] <SaltyBones> But I think it might provide an excuse for people who have to convince irc users ;)
[19:25:44] *** Dave Cridland shows as "online"
[19:26:16] *** dwd shows as "online"
[19:26:31] *** Dave Cridland has left the room
[19:26:31] *** Dave Cridland shows as "online"
[19:26:32] *** Dave Cridland shows as "online"
[19:28:11] <moparisthebest> SaltyBones: yesterday you said
[19:28:14] <moparisthebest> Hm...maybe the transport should be the other way round.

Offer an IRC server that connects to MUCs.
[19:28:25] <moparisthebest> That's what I was referring to
[19:28:25] <vanitasvitae> Ignite didnt make it into GSoC
[19:28:28] <vanitasvitae> :(
[19:29:02] *** jubalh has joined the room
[19:29:15] *** Dave Cridland has left the room
[19:30:03] *** Dave Cridland shows as "online"
[19:30:10] *** daniel has left the room
[19:30:13] *** daniel shows as "online"
[19:31:17] *** jubalh has left the room
[19:31:18] *** jubalh has joined the room
[19:32:30] <SaltyBones> moparisthebest, I know, I said that regarding the discussion of the KDE folks looking for an IM solution
[19:32:31] *** andy has joined the room
[19:32:53] <Flow> `> * Ge0rG had a very concerning realization about guessable IDs and packet filters in smack earlier today
Care to share?
[19:33:01] *** marc has joined the room
[19:33:09] *** jubalh has left the room
[19:33:34] <Flow> Or is it just something in the ancient smack library yaxim uses?
[19:34:30] *** Dave Cridland has left the room
[19:34:38] *** Dave Cridland shows as "online"
[19:34:41] *** moparisthebest shows as "online"
[19:35:08] <moparisthebest> SaltyBones, right so they could use a MUC, but also have an IRC server that IRC users could use
[19:35:15] <Dave Cridland> moparisthebest, I like this, BTW.
[19:35:18] <moparisthebest> and everyone would end up in the same place, but it'd be a muc
[19:35:37] <Dave Cridland> moparisthebest, Is the GPLv3 your addition or from telepaatti?
[19:36:25] *** ralphm has joined the room
[19:36:35] *** la|r|ma shows as "online"
[19:36:41] *** lskdjf shows as "online"
[19:36:48] *** lskdjf shows as "online"
[19:37:10] <Dave Cridland> Anyone know what servers support XEP-0288 these days?
[19:37:26] <Dave Cridland> Does this Prosody instance, for example?
[19:37:34] <moparisthebest> Dave Cridland, looks like the original telepaatti is gone, but GPLv3 goes back to at least the next fork looks like
[19:38:13] *** andy has left the room
[19:38:14] *** andy has joined the room
[19:38:20] <moparisthebest> honestly I kind of abandoned it because I like rust so much more than python nowadays but can't be asked to rewrite everything yet haha :)
[19:38:22] <Flow> -xep 288
[19:38:23] <Bunneh> Flow: Bidirectional Server-to-Server Connections (Standards Track, Draft, 2016-10-17)
See: https://xmpp.org/extensions/xep-0288.html
[19:38:29] <moparisthebest> I still run an IRC server I would love to rm -rf
[19:38:56] <SamWhited> > I need to know their IDs so I can tell the server "give me all messages since $id".

I know, apparently I wasn't clear about something, sorry. I'm not suggesting we get rid of MAM or leave everything exactly as it is today, just that MAM doesn't cover some important parts of SM and probably can't be made to cover it without significant downsides.
[19:39:06] <Ge0rG> Flow: it's affecting the old smack for sure, I'll have a look into smack 4 and let you know
[19:39:16] <SamWhited> (sorry for the long delay, got pulled into a meeting and then was AFK for a bit)
[19:39:49] *** winfried shows as "online"
[19:40:22] *Dave Cridland greps his logs
[19:40:26] *** Dave Cridland has left the room
[19:40:26] *** Dave Cridland shows as "online"
[19:40:37] <Dave Cridland> So there's actually only one server I talk to that does bidi. That's scary.
[19:41:44] *** Dave Cridland has left the room
[19:41:52] *** Kev shows as "away"
[19:41:54] *** Kev shows as "online"
[19:42:24] *** Holger shows as "away" and his status message is "Auto-away (idle)"
[19:43:24] *** Dave Cridland shows as "online"
[19:44:07] *** jonasw shows as "online"
[19:44:48] *** moparisthebest has left the room
[19:44:48] *** moparisthebest shows as "online"
[19:45:00] *** Dave Cridland has left the room
[19:45:02] *** winfried has left the room
[19:45:32] <Flow> Ge0rG, let me know if you didn't just re-discover CVE-2014-0364
[19:45:40] *** Dave Cridland shows as "online"
[19:46:34] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[19:46:40] *** Dave Cridland shows as "online"
[19:46:47] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[19:46:47] *** dwd shows as "online"
[19:47:43] *** Holger shows as "online" and his status message is "I'm available"
[19:48:44] *** Dave Cridland has left the room
[19:49:37] *** Syndace has left the room
[19:49:42] *** Dave Cridland shows as "online"
[19:50:20] *** Syndace has joined the room
[19:51:43] *** jubalh has joined the room
[19:51:54] *** Dave Cridland has left the room
[19:52:19] *** Dave Cridland shows as "online"
[19:52:35] *** Dave Cridland has left the room
[19:52:46] *** Dave Cridland shows as "online"
[19:53:45] <Ge0rG> Flow: it's related but different
[19:54:27] <Zash> Dave Cridland: Is it mine? :)
[19:55:03] *** marc has left the room
[19:55:37] <Dave Cridland> Zash, No, Lance's. Although actually grep might have gone into Annoying Binary Mode.
[19:55:56] <SaltyBones> -I ?
[19:55:57] *** Dave Cridland has left the room
[19:55:58] *** Dave Cridland shows as "online"
[19:56:18] <Dave Cridland> -a, but yeah.
[19:56:19] *** Dave Cridland has left the room
[19:56:20] *** Dave Cridland shows as "online"
[19:56:22] *** Syndace has joined the room
[19:56:28] *** Dave Cridland has left the room
[19:56:34] *** Dave Cridland shows as "online"
[19:56:44] *** Dave Cridland has left the room
[19:56:55] <Dave Cridland> OK, so that's actually lots of servers doing bidi. Happy bunny, now. I've been putting the support into Metre.
[19:57:08] *** Dave Cridland shows as "online"
[19:57:15] <Zash> Oh? Hm, feeling up for doing a survey?
[19:57:39] <Dave Cridland> Zash, What sort of survey?
[19:57:54] <Zash> "How many servers do 288?"
[19:58:06] <Zash> Hooooold on now
[19:58:13] <Zash> Is that the same number as ...
[19:58:44] <Zash> https://www.youtube.com/watch?v=azEvfD4C6ow !!!
[19:58:47] *** blabla shows as "online"
[19:59:09] *** Dave Cridland has left the room
[19:59:10] *** Dave Cridland shows as "online"
[19:59:41] *** lumi shows as "away" and his status message is "(Idle 10 min)"
[20:00:31] *** tux shows as "online"
[20:00:39] *** tux has left the room
[20:00:53] <Dave Cridland> Zash, Looks like it's PSYC and Prosody.
[20:01:28] <Zash> Prosody doesn't do it out of the box, you need to go a bit out of your way to install a community module.
[20:01:37] *** Dave Cridland has left the room
[20:01:59] <Dave Cridland> Really? Quite a few people have, then.
[20:03:04] *** Dave Cridland shows as "online"
[20:03:57] *** lskdjf shows as "online"
[20:05:12] *** Dave Cridland has left the room
[20:05:31] *** Dave Cridland shows as "online"
[20:05:35] <Zash> Dave Cridland: Question is, how much self-selection bias is there among people that have you in their roster? :)
[20:05:43] <Dave Cridland> Zash, Lots.
[20:07:48] *** ralphm has joined the room
[20:09:11] *** Kev shows as "away"
[20:09:13] *** Kev shows as "online"
[20:11:01] *** remko has joined the room
[20:11:02] *** remko shows as "online"
[20:13:42] *** jubalh shows as "online"
[20:16:54] *** Dave Cridland has left the room
[20:16:54] *** Dave Cridland shows as "online"
[20:16:56] *** Dave Cridland has left the room
[20:17:02] *** Dave Cridland shows as "online"
[20:18:53] *** winfried has joined the room
[20:20:17] *** jubalh has left the room
[20:20:18] *** jubalh shows as "online"
[20:20:55] *** andy has left the room
[20:27:51] *** Holger shows as "away" and his status message is "Auto-away (idle)"
[20:27:58] *** blabla has left the room
[20:28:34] *** vanitasvitae shows as "online"
[20:30:06] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:30:22] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:31:09] *** lumi shows as "online"
[20:35:26] *** daniel has left the room
[20:35:31] *** daniel shows as "online"
[20:36:16] *** MattJ shows as "away"
[20:37:42] *** blabla shows as "online"
[20:37:45] *** daniel has left the room
[20:37:50] *** blabla shows as "online"
[20:38:19] *** moparisthebest shows as "online"
[20:40:06] *** Dave Cridland shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[20:40:06] *** dwd shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[20:41:21] *** Dave Cridland has left the room
[20:41:48] *** Dave Cridland shows as "online"
[20:43:05] *** jubalh has left the room
[20:46:38] *** Dave Cridland shows as "online"
[20:46:38] *** dwd shows as "online"
[20:46:46] *** remko has joined the room
[20:46:46] *** remko shows as "online"
[20:51:35] *** Dave Cridland has left the room
[20:51:38] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:51:38] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:52:05] *** zinid shows as "online"
[20:52:07] *** Dave Cridland shows as "online"
[20:53:08] *** zinid has left the room
[20:53:32] *** Dave Cridland shows as "online"
[20:53:32] *** dwd shows as "online"
[20:54:06] *** Dave Cridland has left the room
[20:54:31] *** Dave Cridland shows as "online"
[20:56:55] *** Dave Cridland has left the room
[20:57:03] *** Dave Cridland shows as "online"
[20:58:10] *** daniel has left the room
[20:58:20] *** Alex has left the room
[20:59:32] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:59:32] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[20:59:41] *** Dave Cridland has left the room
[20:59:50] *** Dave Cridland shows as "online"
[21:00:12] *** dwd shows as "online"
[21:00:55] *** Dave Cridland shows as "online"
[21:02:34] *** Dave Cridland has left the room
[21:03:28] *** bra has left the room
[21:03:28] *** bra has joined the room
[21:03:52] *** Dave Cridland shows as "online"
[21:04:41] *** daniel has joined the room
[21:04:56] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[21:05:04] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[21:05:48] *** Steve Kille shows as "online" and his status message is "At Home"
[21:07:46] *** SamWhited has left the room
[21:08:46] *** Dave Cridland shows as "online"
[21:09:22] *** dwd shows as "online"
[21:09:59] *** Alex has joined the room
[21:12:59] *** lovetox shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[21:14:29] *** marc has joined the room
[21:14:46] *** lovetox shows as "online"
[21:15:27] *** Tobias has joined the room
[21:16:28] *** marc shows as "online"
[21:16:59] *** Dave Cridland has left the room
[21:17:45] *** Dave Cridland shows as "online"
[21:22:41] *** dwd shows as "online"
[21:24:04] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[21:24:57] *** Steve Kille shows as "away" and his status message is "At Home"
[21:26:45] *** remko has joined the room
[21:26:45] *** remko shows as "online"
[21:31:13] *** bra has left the room
[21:31:13] *** bra has joined the room
[21:31:26] *** Dave Cridland shows as "online"
[21:32:10] *** dwd shows as "online"
[21:32:56] *** daniel has left the room
[21:34:19] *** bra has left the room
[21:34:20] *** bra has joined the room
[21:34:53] *** daniel has joined the room
[21:36:02] <Dave Cridland> .
[21:36:11] <Zash> ,
[21:40:36] *** vanitasvitae shows as "online"
[21:41:44] *** Dave Cridland has left the room
[21:41:47] *** moparisthebest has left the room
[21:41:47] *** moparisthebest shows as "online"
[21:41:56] *** Dave Cridland shows as "online"
[21:43:33] *** lovetox shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[21:50:12] *** lovetox shows as "online"
[21:51:29] *** jubalh shows as "online"
[21:52:11] *** jonasw shows as "away"
[21:52:55] *** andy has joined the room
[21:52:59] <Kev> M-Link does bidi, but we disabled it because of bug reports from Prosody about us not accepting stanzas down the right streams. Some of which I'm starting to question :)
[21:53:28] <Dave Cridland> :-)
[21:53:59] *** mimi89999 shows as "online"
[21:54:11] *** Dave Cridland has left the room
[21:54:22] *** Dave Cridland shows as "online"
[21:55:45] *** Dave Cridland has left the room
[21:55:49] *** bra has left the room
[21:55:49] *** bra has joined the room
[21:56:02] *** Dave Cridland shows as "online"
[21:57:24] *** Kev has left the room
[21:59:38] *** bra has left the room
[21:59:38] *** bra has joined the room
[22:00:35] *** tux has left the room
[22:05:57] *** tux shows as "online"
[22:07:33] *** jubalh has left the room
[22:12:18] *** pep. has left the room
[22:12:22] *** pep. shows as "online"
[22:12:57] *** moparisthebest has joined the room
[22:15:45] *** vanitasvitae has left the room
[22:17:59] *** Dave Cridland has left the room
[22:18:07] *** dwd has joined the room
[22:20:22] *** Dave Cridland shows as "online"
[22:23:24] *** blabla has left the room
[22:23:30] *** blabla shows as "online"
[22:25:03] *** tux has left the room
[22:26:13] *** dwd shows as "online"
[22:28:32] *** marc has left the room
[22:28:45] *** SaltyBones has left the room
[22:28:55] *** andy has left the room
[22:30:57] *** remko has joined the room
[22:30:57] *** remko shows as "online"
[22:32:54] *** SamWhited shows as "online"
[22:40:06] *** vanitasvitae has left the room
[22:40:16] *** vanitasvitae shows as "online"
[22:41:31] *** daniel has left the room
[22:41:46] <dwd> So, Metre now does Bidi.
[22:47:13] *** daniel has joined the room
[22:47:54] *** marc has left the room
[22:52:18] *** lumi has left the room
[22:54:25] <dwd> OK, this is weird. Metre is successfully negotiating Bidi with various Prosody servers. OK, great. But absolutely nothing ever tries to negotiate bidi with it, despite it offering the feature.
[22:58:47] *** jjrh has left the room
[22:58:52] *** jjrh shows as "online"
[23:00:13] *** Holger shows as "online" and his status message is "I'm available"
[23:00:31] *** Dave Cridland has left the room
[23:02:09] *** Dave Cridland shows as "online"
[23:02:17] *** Holger has left the room
[23:02:19] *** Holger shows as "online"
[23:02:30] *** dwd has left the room
[23:02:50] *** Neustradamus has left the room
[23:05:11] *** Dave Cridland has left the room
[23:05:38] *** daniel has left the room
[23:05:58] *** nyco has left the room
[23:06:08] *** nyco shows as "online"
[23:07:09] *** Dave Cridland shows as "online"
[23:08:55] *** Neustradamus has joined the room
[23:15:14] *** lskdjf shows as "online"
[23:15:19] *** lskdjf shows as "online"
[23:17:06] *** mimi89999 shows as "online"
[23:17:19] *** daniel has joined the room
[23:23:36] *** Steve Kille shows as "online" and his status message is "At Home"
[23:25:13] *** remko has joined the room
[23:25:14] *** remko shows as "online"
[23:26:28] *** Holger shows as "away" and his status message is "I'm away"
[23:26:29] *** Holger shows as "online" and his status message is "I'm available"
[23:31:32] *** Dave Cridland shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[23:31:32] *** dwd shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[23:31:42] *** daniel has left the room
[23:32:26] *** daniel has joined the room
[23:35:13] *** blabla has left the room
[23:40:05] *** lovetox shows as "away" and his status message is " (Away as a result of being idle more than 5 min)"
[23:40:05] *** jjrh has left the room
[23:40:33] *** jjrh shows as "online"
[23:41:00] *** jjrh has left the room
[23:41:03] *** jjrh shows as "online"
[23:41:32] *** Dave Cridland shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[23:41:32] *** dwd shows as "xa" and his status message is " (Not available as a result of being idle more than 15 min)"
[23:41:41] *** jjrh has left the room
[23:41:45] *** jjrh shows as "online"
[23:42:05] *** jjrh has left the room
[23:42:11] *** Dave Cridland has left the room
[23:42:31] *** jjrh shows as "online"
[23:42:36] *** Dave Cridland shows as "online"
[23:45:54] *** lovetox shows as "online"
[23:46:50] *** Steve Kille shows as "away" and his status message is "At Home"
[23:51:33] *** lovetox has left the room
[23:54:51] *** jjrh has left the room
[23:54:54] *** Dave Cridland has left the room
[23:54:57] *** jjrh shows as "online"
[23:56:07] *** Dave Cridland shows as "online"
[23:57:09] *** SamWhited has left the room
[23:57:32] *** Dave Cridland has left the room
[23:58:59] *** Dave Cridland shows as "online"