Home
ejabberd
ejabberd@conference.jabber.ru
Thursday, 16 August 2007< ^ >
badlop has set the subject to: Здесь говорят только по-английски! | Only english here! | website: http://ejabberd.jabber.ru/ | Erlang/OTP R11B-5 released; any interesting feature?
Room Configuration

GMT+4
[00:00:03] sss leaves the room
[00:00:06] <bomer> i'm converting from a previous apt-get install ejabberd installation :)
[00:00:28] sss joins the room
[00:03:38] <ermine> proper ejabberd is installed into erlang/lib/
[00:07:48] <badlop> <legoscia> hm… could i use the bugzilla for my PEP hack? <--- this is important, and i forgot to ask mremond: where to track the bugs that are found in ejabberd-modules?
[00:08:54] <badlop> if he does not give any solution, jabber.ru/bugzilla can be a nice solution for your pep :)
[00:09:56] <ermine> badlop: why do you want to ask mremond? Who is mremond?
[00:12:18] <bomer> if you use mod_ctlextra to add user1@server2 <mailto:user1@server2> to user1@server1 <mailto:user1@server1>'s roster, will the presence of user1@server2 <mailto:user1@server2> be sent to user1@server1 <mailto:user1@server1> if user1@server1 <mailto:user1@server1> is not on user1@server2 <mailto:user1@server2>'s roster?
[00:12:47] <bomer> both users using different domains
[00:18:18] <badlop> mremond is the person that hosts and administers ejabberd-modules svn repo, and ejabberd Jira
[00:18:53] <badlop> legoscia: is your pep work commited in ejabberd-modules?
[00:19:40] <Kev> yes, I'm running it :)
[00:19:49] <legoscia> yes, it is
[00:19:55] <legoscia> there are a few things left to do…
[00:20:31] <ermine> ah, jira is not integrated with ejabberd-modules svn?
[00:21:29] <legoscia> it's more of a social than a technical distinction, i think
[00:22:05] <legoscia> i see ejabberd-modules as the place for strange experiments that p1 people may or may not be interested in having in their bug trackers ☺
[00:23:22] <ermine> ☺ strange symbol
[00:23:28] <legoscia> my favorite
[00:23:41] <legoscia> everyone should have a smiley key on their keyboard
[00:24:09] <ermine> avatar?
[00:24:23] <ermine> next generation of smiles
[00:24:40] <ermine> small video icons
[00:24:44] <legoscia> what's wrong with unicode smilies? ☺
[00:24:51] <legoscia> there are _three_ different ones!
[00:25:09] <legoscia> ☹ ☺ ☻
[00:25:22] <ermine> eh
[00:26:20] <badlop> legoscia: so, would it help you a 'PEP' component here?
http://www.jabber.ru/bugzilla/describecomponents.cgi?product=ejabberd
[00:27:19] <legoscia> exactly
[00:29:46] <badlop> aleksey: can you create a PEP component in ejabberd@bugzill, and assign it to legoscia?
[00:30:30] ermine pokes aleksey
[00:31:25] <badlop> aleksey, ermine, teo: do you *disagree* if I mention that Jira is official ejabberd bug tracker since today (15 august 2007) in http://ejabberd.jabber.ru/bugzilla ?
[00:31:43] <badlop> there's no need to close bugzilla, or disable account registration
[00:32:47] <ermine> badlop: you are not p1's manager and cannot say nothing officialy
[00:33:12] <ermine> badlop: you can only cite
[00:33:28] <aleksey> badlop: no
[00:33:54] ermine simply scoffs
[00:34:31] <badlop> aleksey: :P
a) you do not disagree ---> you agree with me
b) yoy not like the idea ---> you disagree with me
[00:34:43] <aleksey> a
[00:36:41] <ermine> reminds three parts
[00:36:48] <aleksey> legoscia: pep component is added
[00:37:46] <badlop> legoscia: now you can link to your personal bug tracker in docs and in http://ejabberd.jabber.ru/mod_pep :)
[00:37:58] <legoscia> thanks!
[00:38:08] <badlop> <ermine> badlop: you can only cite <--- ah, right, i'll just cite and link to sources
[00:38:40] Christian leaves the room
[00:43:20] <ermine> badlop: right
[00:44:03] <legoscia> you could put it this way: there is an entity called process-one, that does most ejabberd bug fixing these days. if you want their attention, post your bugs ---> there
[00:44:05] <badlop> ermine: if you look closely, the bug tracker page mentions bugzilla first, and jira later
[00:44:15] <badlop> even if mremond considers jira the main bug tracker
[00:44:30] <legoscia> if you want other people to fix your bug, bugzilla is <---- here ☺
[00:44:52] <badlop> i edited that text this way, because p1 is manager of ejabberd, but not the author
[00:45:16] <badlop> now that both the author and the managers agree in this topic, i can finally give more importance to Jira
[00:45:49] <badlop> in that page
[00:50:35] <ermine> manager or hijacker... :) well, bad and lazy hijacker, like fat cat
[00:50:53] <legoscia> oh, hijack it yourself ☺
[00:52:40] teo leaves the room
[00:53:38] <ermine> badlop: it seems p1 is sure that everything will be solved for their advantage.
[00:54:56] <legoscia> but for whose disadvantage?
[00:56:17] <ermine> for users (not customers)
[00:57:24] <legoscia> i'm an ejabberd user (jabber.se). what should p1 do to serve me better?
[00:58:42] <ermine> you are advanced user and are able to read erlang code and even write it yourself
[00:59:24] niekie leaves the room
[00:59:38] bomer leaves the room: Disconnected
[00:59:43] <ermine> i.e you are able to install from svn and check its security holes
[01:00:11] <ermine> no?
[01:00:29] <legoscia> yes, that's true
[01:01:45] <ermine> do you know about holes in installers produced by p1?
[01:01:55] <legoscia> no, i don't
[01:03:12] <ermine> do you should know?
[01:03:24] <ermine> as user or as developer -- at your choice
[01:03:56] <ermine> or as comunity member (if community exists still)
[01:04:03] kostix joins the room
[01:04:03] <legoscia> personally i wouldn't even touch such an installer, but you're right, the installer should be open
[01:04:43] <Kev> I used the installer, and it was a completely different layout from svn
[01:04:45] <Kev> very confusing
[01:05:03] <ermine> legoscia: well, partially they are were in svn and there were fixed (not by p1)
[01:07:35] <ermine> legoscia: it is interesting why we (russians) knows about most of holes existed in svn or installers
[01:08:26] <ermine> something is wrong
[01:27:22] legoscia leaves the room
[01:33:01] <badlop> many mistakes and holes that were not conveniently reported, right?
[01:35:19] <badlop> last week I found a problem with Psi's xep33 implementation
[01:35:48] <badlop> so, i joined their chatroom, chatted with Psi developers to confirm if it's a bug
[01:36:05] <badlop> and i'll submit to their bug tracker soon
[01:37:21] <badlop> I also found problems with Openfire's and Tkabber's xep33 implementations: in all cases i contacted the authors of the code to verify, and to report
[01:37:59] <badlop> if anybody finds bugs in my code, i'm very happy to receive the reports, and i'll try to fix them
[01:38:11] <ermine> badlop: i said -- most of holes were fixed in svn (not by p1), so they were reported.
[01:39:03] <badlop> and there were fixed <--- ah, i didn't understand this part
[01:40:03] <ermine> badlop: what do we should report to get new releases?
[01:40:33] <ermine> and to whom?
[01:43:01] <badlop> would you like frequent ejabberd releases (new version every 4 or 6 months)?
[01:44:07] <ermine> why not every month or after every found hole?
[01:44:41] <badlop> do you like openfire's monthly release cycle?
[01:44:46] <ermine> there can be different releases -- main, bugfix and security fix
[01:46:38] <ermine> badlop: 1.1.3 is crap security fix, there were some more important fixs insvn at that time
[01:46:56] <TobiasFar> badlop: i guess each quarter year would be enouth
[01:47:03] <TobiasFar> *enough
[01:47:21] <badlop> there can be different releases -- main, bugfix and security fix <--- ah, i like this idea :)
[01:48:08] <ermine> it does not matter how often - every 4 month, every month, every day
[01:48:41] <TobiasFar> ermine: it does matter since you seem to complain about not having enough releases ;)
[01:48:50] <kostix> I thibk a good rule of thumb might be: one release per one-two major features (like new XEP implemented)
[01:49:30] <badlop> each quarter year <--- i agree, and i already chatted with teo and mremond, and all we agree that 1-year releases are bad
[01:50:44] <ermine> TobiasFar: let me check which version of your server and try to crash it :)
[01:51:11] <badlop> ermine: bad luck, he seems to use svn :S
[01:51:13] <ermine> if there were known bugs since you had time upgrade it
[01:51:17] <TobiasFar> i think quarter year is enough time to develop new features
[01:51:30] <TobiasFar> ayena.de uses maybe a 4 days old svn
[01:51:36] sss leaves the room
[01:51:49] <TobiasFar> but i'm in here through jabber which use a 1 year old svn hahaha
[01:52:09] sss joins the room
[01:52:14] <TobiasFar> that reminds me of checking whether i still can hang up openfire :)
[01:53:01] <ermine> 1 year? If that server user sql, let'me try sql injector
[01:53:23] <TobiasFar> was just guessing...i guess the latest release which is pretty old
[01:53:28] ermine tries remember when this hole was entered to svn
[01:54:57] <elmex> badlop: does xep33 really have implementations? wow
[01:59:48] sss leaves the room
[01:59:49] <elmex> the replyto/room/noreply stuff is crazy
[02:00:34] <badlop> currently, psi has support for sending TO
[02:00:39] sss joins the room
[02:00:44] thomas leaves the room
[02:01:04] <badlop> tkabber shows info if it receives a message with xep33 data
[02:01:28] <elmex> looks like i have to implement it in Net::XMPP2 soon if there are already so many users :)
[02:01:59] <badlop> openfire implementation seems slightly experimental: it does all the MUST, but few of the SHOULD and MAY
[02:02:06] <elmex> badlop: xep33 is going to be fun with 100 replyrooms :)
[02:02:37] <elmex> imagine i send you a message with 100 replyrooms and you want to answer
[02:02:39] <badlop> i hope the revised xep will be more sane
[02:03:00] <elmex> the whole reply-cruft is crazy
[02:04:52] <badlop> elmex: reading the current version, i understand that 1, and only 1 reply is sent
[02:05:14] <elmex> here MAY be more than one replyto or replyroom on a stanza, in which case the reply stanza MUST be routed to all of the addresses.
[02:05:18] <TobiasFar> badlop: yeah...wen i've tested wildfire's xep33 imlementation over one year ago it was acting really slow
[02:05:42] <badlop> elmex: ouch
[02:05:52] <elmex> badlop: it's even a -MUST- :)
[02:06:08] <badlop> i only read the first line, not the second :S
[02:06:20] <badlop> i bookmark this topic to discuss with stpeter
[02:06:34] <elmex> replyto is overdesig IMO, from attribute of the mssage should be enough
[02:06:51] <elmex> this is not smtp after all
[02:08:28] <elmex> if at all replyto/room/whatever more feels like a shim-feature
[02:09:01] sss leaves the room
[02:11:42] <elmex> badlop: i wonder how the interface for replyroom might look like
[02:13:31] badlop is fortunate to not need to worry about xep33 human interface, however, he must care about performance, security and spam
[02:14:27] <elmex> i mean, imagine this: client opens message window with a message wil lots of replytos. now user types in reply and hits enter
[02:14:38] <elmex> client will sent out to all the replytos
[02:16:06] <elmex> "A recipient SHOULD trust a stanza's extended addressing headers only as much as it trusts the sender of the stanza." - does that mean my users need to read the xep? :)
[02:17:14] <badlop> if the stanza is received from a contact not in roster, the client could show a big warning to the user
[02:17:28] <elmex> yea
[02:23:59] <elmex> gn
[02:24:00] <elmex> 8
[02:24:02] elmex leaves the room
[02:28:26] ermine leaves the room
[02:29:05] niekie joins the room
[02:34:07] sss joins the room
[02:37:57] <badlop> done: http://ejabberd.jabber.ru/bugtracker
and http://ejabberd.jabber.ru/bugzilla redirects to there
[02:43:02] sss leaves the room
[02:57:35] badlop leaves the room
[03:16:03] TobiasFar leaves the room
[05:01:26] niekie leaves the room
[10:45:29] badlop joins the room
[10:49:27] ermine joins the room
[11:11:53] Kev leaves the room
[11:28:45] ermine leaves the room
[11:28:47] ermine joins the room
[11:29:06] ermine leaves the room
[11:29:20] ermine joins the room
[11:30:38] sss joins the room
[11:41:10] sss leaves the room
[11:41:11] sss joins the room
[12:04:56] notKev joins the room
[12:32:26] kostix leaves the room
[12:36:27] sss leaves the room
[12:36:52] sss joins the room
[12:41:56] legoscia joins the room
[14:20:40] elmex joins the room
[14:23:12] elmex leaves the room
[14:43:06] TobiasFar joins the room
[14:50:10] zapw joins the room
[14:50:11] <zapw> hi
[14:53:17] <zapw> https://support.process-one.net/browse/EJAB-307 , is it just me or epam , reads /etc/passwd file directly
[14:53:31] zinid joins the room
[14:56:10] <badlop> i didn't find 'passwd' in the patch, did you?
[14:57:31] <zapw> no
[14:57:41] <zapw> i wonder if pam_unix talks directly to /etc/passwd
[14:57:46] <zinid> epam doesn't read passwd file, it uses pam api for that. however some pam modules read it
[14:58:03] <zapw> or calls getpwnam
[14:58:08] <zapw> or getpwent
[14:58:29] <zapw> zinid, do you know about pam_unix ?
[14:58:40] <zinid> zapw: yes
[14:59:02] <zapw> it reads /etc/passwd directly? or uses nss library
[14:59:12] <zinid> zapw: pam_unix needs read access on passwd and shadow
[14:59:55] <zapw> i have passwd: mysql and shadow: mysql , in nsswitch.conf
[15:01:17] <zinid> zapw: i don't know about pam mysql
[15:01:24] <zapw> it isnt pam mysql
[15:01:31] <zapw> it's libnss-mysql
[15:01:34] <zinid> ah
[15:07:16] <zapw> sorry for repeating , so pam_unix does read /etc/passwd and /etc/shadow directly? it dosent use calls that end reading from nsswitch.conf ?
[15:10:16] <zinid> zapw: http://www.kernel.org/pub/linux/libs/pam/Linux-PAM-html/sag-pam_unix.html
[15:11:23] <zinid> zapw: A helper binary, unix_chkpwd(8), is provided to check the user's password when it is stored in a read protected database. This binary is very simple and will only check the password of the user invoking it.
[15:11:58] <zapw> evil
[15:12:03] <zinid> zapw: if you want to check the passwords of any users you must be root
[15:12:15] <zinid> zapw: yep :)
[15:16:02] kostix joins the room
[15:16:04] Turyzaki joins the room
[15:18:16] <Turyzaki> Никто в делфи не шарит?
[15:18:16] <Turyzaki> Got it...
[15:19:08] <kostix> sharit, tolko ne tut
[15:19:19] <Turyzaki> Where?
[15:19:34] <zinid> nada perepisat' yozhika na delphi
[15:20:04] <Turyzaki> I need some jabber client libraries for delphi, but cant find them :-(
[15:20:10] <kostix> zinid: йа -- za!
[15:20:55] <Turyzaki> What is hedgehog written on?
[15:20:57] <kostix> Turyzaki: Exodus contains jabberlib written in Delphi; also there's a set of Delphi components called IP@Work -- you can get it for as low as $400
[15:21:20] <Turyzaki> I checked everything
[15:21:25] <Turyzaki> 1:
[15:22:37] <Turyzaki> Ip works doesn't contain such volume as joining groupchats, it doesn't support conferences,
[15:22:50] <badlop> Turyzaki: ask in the jdev mailing list
[15:23:10] <Turyzaki> Where is it?
[15:23:13] sss leaves the room
[15:23:31] <badlop> www.jabber.org
[15:23:33] <zinid> Turyzaki: http://www.jabber.org/about/lists.shtml
[15:23:47] <Turyzaki> 2: jopl on exodus doesn't compile well :-( sl it doesn't work :-(
[15:25:27] <Turyzaki> I used to download and install jabbercom, but it also doesn't work
[15:26:05] <kostix> Turyzaki: it's just hard to decouple from the client + you have to throw in some external libraries
[15:26:11] <Turyzaki> By the way... I got ip works for free ... Hh
[15:27:20] <Turyzaki> I've made my oxn library, but it's not complite
[15:27:31] <Turyzaki> Own
[15:28:21] <Turyzaki> Ет всегда вы тока англииским?
[15:28:51] <zinid> Turyzaki: yes
[15:29:09] <Turyzaki> I like it :-)
[15:30:10] <Turyzaki> Damn, i'm so sick and tired of unsuccessfull compilining jopl... :-(
[15:32:42] <Turyzaki> Ok, thanks for help ;-)
[15:32:56] Turyzaki is now known as 2pac
[15:37:43] 2pac leaves the room
[16:15:40] zinid leaves the room: Replaced by new connection
[16:15:41] zinid joins the room
[16:37:45] sss joins the room
[16:39:00] <zapw> zind: it just accoured to me unix_chkpwdl
[16:39:05] <zapw> is used only when authenticating
[16:39:19] <zapw> auth , not in account modules
[16:44:40] chirayu joins the room
[16:49:05] bomer joins the room
[16:49:11] <chirayu> Hi , everybody .. i want to give permission on some host like example.com ..using mod_filter.. how it can be perform
[16:49:31] <chirayu> i have more then one host on the server
[16:50:02] <bomer> with mod_ctlextra with add-rosteritem, what are "subs"?
[16:50:46] <chirayu> mod_filter
[16:51:05] <bomer> sorry, i wasn't replying to you chirayu :)
[16:51:59] <chirayu> i want to restrict some host's user denying some services
[16:52:13] <chirayu> like transport
[16:52:18] <bomer> i personal have no idea
[16:52:30] <chirayu> ok thanks
[16:52:44] <bomer> well hang around, someone might be back soon that can help you
[16:53:14] <chirayu> ok ...thanks for the complement..
[16:53:32] <bomer> what? that wasn't a complement ....
[16:54:29] <chirayu> ok sorry...thnks for u r help..
[16:54:48] <bomer> what? i'm not helping you at all ... actually, i'm totally wasting your time
[16:58:03] <chirayu> is anybody has idea . plz reply me
[17:06:59] <legoscia> chirayu: you can use the example at http://ejabberd.jabber.ru/mod_filter as a starting point
[17:08:24] <chirayu> thanks for ur reply .. i have all ready tried that all things.
[17:09:17] <chirayu> see . what my problem is ... i have two host like example.com, and example .net.... ok
[17:09:36] <chirayu> i am using transport on both host
[17:10:19] <chirayu> i just want to allow some transport user on host example.net
[17:10:33] <chirayu> using mod_filter
[17:10:56] sss leaves the room
[17:11:51] <chirayu> if i assing some access rule on example.net ... its working for me but .. only for outgoing msg
[17:12:04] <chirayu> not incoming msg
[17:12:34] <chirayu> u have seen that last example on above link
[17:12:57] sss joins the room
[17:13:44] <chirayu> in that only for one host its give but i want to allow that thing on example.com but not on example.net
[17:14:36] <chirayu> sorry my mistake ... allow all user of msn transport on on example.com
[17:15:00] <chirayu> but allow some user on example.net
[17:15:26] <chirayu> u get my problem...
[17:17:25] sulci leaves the room
[17:20:04] sulci joins the room
[17:22:46] chirayu leaves the room: Replaced by new connection
[17:22:52] chirayu joins the room
[17:28:11] zapw leaves the room
[17:29:02] sss leaves the room
[17:29:02] sss joins the room
[17:33:54] <bomer> is there anyway at all to add someone on an other vhost to your roster without having to be added to their roster, but still see presence?
[17:40:58] <notKev> yes
[17:41:08] <notKev> you don't need someone to be on your roster to grant them a subscription to your presence
[17:41:08] zinid leaves the room: Logged out
[17:47:16] <bomer> if i add someone from another domain to my roster, i do not see their presence until i am added to theirs. if they are on the same domian, then it works fine
[17:49:01] <notKev> most clients join together the 'auth' and 'add to roster' items.
[17:49:11] <notKev> but in the protocol they're distinct and you can do one without the other
[17:49:59] <bomer> i can't seem to make it work between vhosts
[17:50:24] <notKev> odd
[17:50:56] <bomer> i just installed 1.1.3, going to see if that helps anywa
[17:51:00] <bomer> any*
[17:55:48] <bomer> yep, same thing. example: user1@server1 is on the roster for user user2@server2 Both users are logged in, but the presence still shows offline
[17:56:18] <notKev> and has user2@server2 given auth?
[17:57:36] bomer leaves the room: Disconnected
[18:31:39] bomer joins the room
[18:33:08] <bomer> I get this error when I try to use ejabberdctl pushroster: E(<4010.1278.0>:ejabberd_hooks:187): {{badmatch,{error,enoent}}, any ideas?
[18:35:38] <legoscia> enoent means that some file doesn't exist
[18:37:14] <bomer> hmm... i put the xml file in /var/lib/ejabberd/ebin, where the other beam files are
[18:39:05] <bomer> awwww, specify the full path, even if its in the pwd, and even if its looking in the ebin directly
[18:40:30] <bomer> do you know if its possible to create sub-groups using a template like this?
[18:43:22] chirayu leaves the room
[18:44:08] sss leaves the room
[18:46:02] ermine leaves the room
[18:46:07] ermine joins the room
[18:46:26] zapw joins the room
[18:46:29] <zapw> hi
[18:46:51] <zapw> how to use aclpopulate
[18:47:44] <badlop> zapw: what is aclpopulate?
[18:48:02] <zapw> http://ejabberd.jabber.ru/aclpopulate , Author by you
[18:48:16] <zapw> Author: Badlop Allows automated addition of users to a Shared Roster Group
[18:48:45] <zapw> question is how do i add multiplie shared_roster_Groups
[18:48:50] <zapw> in that syntax
[18:49:32] <badlop> that stuff is two years old, and i never used it again, so i completely forgot how to use it
[18:49:47] <badlop> i hope that the documentation i wrote in 2005 is useful still today
[18:50:10] <zapw> badlop, it didnt catch on?
[18:50:25] <badlop> and maybe it does not even work in newer ejabberd releases...
[18:50:46] <zapw> i think there really should be some way to add shared rosters in config file insted of going to the web interface
[18:51:25] <badlop> me too, so i published this patch as open source
[18:51:35] <badlop> you can modify it as much as you can
[18:51:44] <badlop> *want
[19:03:33] Christian joins the room
[19:06:25] sss joins the room
[19:11:42] <bomer> hey badlop, can i make sub-groups in the rosters using a template?
[19:14:25] <badlop> do you mean nested groups of contacts in the roster?
[19:14:30] <bomer> yes
[19:15:08] <badlop> can you do this with a Jabber client?
[19:15:14] <bomer> yes
[19:15:36] <deryni> Which clients support that?
[19:15:56] <badlop> how? in tkabber i set group names like this:
Workers::Spain::Madrid
[19:16:10] <bomer> i'm using a xmmp client developed locally for the company i'm at. this is the first time i've seen this work my self
[19:16:45] <notKev> badlop: that's the way to do subgroups. It's not all that clean, but it does work.
[19:16:57] <notKev> I don't know which clients support it though; Psi doesn't
[19:17:07] <bomer> do you name the groups like in the templates and the client will draw the groups nested correctly?
[19:17:23] <badlop> yes, i think so
[19:17:32] <bomer> i'm going to see if this client will do the same
[19:17:49] <deryni> I really dislike that XEP.
[19:17:56] <notKev> Informational XEP-0083, Nested Roster Groups, is Active (last updated 2004-10-11): http://www.xmpp.org/extensions/xep-0083.html
[19:18:16] <badlop> deryni: i don't have opinion in the XEP, but i like the feature :)
[19:18:32] <notKev> I dislike having to query for the delimiter
[19:18:32] <kostix> works just great in Tkabber
[19:18:42] <notKev> but apart from that it does the job
[19:18:52] <deryni> I agree that the ability can be occasionally useful (though I've never needed it myself), but I very much dislike overloading group names like that.
[19:20:01] <notKev> well, there's something to be said for the backwards compatability
[19:20:12] <notKev> a client which doesn't support it will still show something which makes sense
[19:21:40] <bomer> that did not seem to work for me. oh well, to me its not that important. if the boss does, then i may revisit it
[19:33:27] bomer leaves the room: Disconnected
[20:01:15] notKev leaves the room
[20:23:01] bomer joins the room
[20:25:41] aleksey leaves the room
[20:27:03] aleksey joins the room
[20:28:31] <bomer> is it not possible to add users to people's roster using ejabberdctl that are across different vhosts (but on the same server) without both users needing to be on each others roster before presence is seen?
[20:32:01] <deryni> What *exactly* are you doing and what *exactly* is happening?
[20:32:31] Kev joins the room
[20:32:34] <deryni> Also, do you understand how XMPP rosters and presence authorizations work?
[20:33:16] sss leaves the room
[20:33:32] sss joins the room
[20:33:55] <bomer> i have several vhosts. some of these vhosts would like to have all of the members of certain vhosts added to their rosters. however, they don't want to be on those other people's rosters, unless they wanted to add the other vhost them selfs
[20:33:55] <bomer> not fully no
[20:36:04] bomer leaves the room: Replaced by new connection
[20:40:23] sss leaves the room
[20:43:18] bomer joins the room
[20:43:25] <bomer> sorry about that ...
[20:44:28] <bomer> i can do this: ejabberdctl add-rosteritem bill vhost1 bill vhost2 Bill vhost1 from If I do that, it will add the user bill@vhost1 to bill@vhost2's roster, without prompting bill@vhost1. presence is shown and bill@vhost2 is not on bill@vhost1's. The problem is, it adds the user without a group.
[20:46:05] <deryni> That is failing to set the group correctly?
[20:47:46] <bomer> well, if i replace "from" with "to" It does add the group correctly, and in the order its suppose to. Then, it would add bill@vhost2 to bill@vhost1's roster and the group is created correctly. Using "from" it seems to only add the requester's username to the targets roster without prompting and presence is seen without the target being on the requesters roster.
[20:49:00] <badlop> and you put both?
[20:49:44] <bomer> if i do both, the group is created on bill@vhost1 and the users are added on both rosters. However, bill@vhost1 is not put in a group on bill@vhost2's roster at all.
[20:49:56] <bomer> it's outside of any group
[20:50:13] <badlop> maybe a bug
[20:50:24] <bomer> i've tried with 1.1.2 and 1.1.3
[20:50:26] <badlop> i'll take a look in some minutes
[20:51:00] <bomer> that would be much appreciated. i have been banging my head on my desk for a few hours now
[21:12:00] <badlop> bomer: ok, now i remember
[21:12:30] <badlop> add-rosteritem was initially supposed to add only 1 item in 1 roster
[21:13:03] <badlop> usually, presence is added in both sides, which means 1 item in 2 different rosters, right?
[21:14:15] <bomer> well, i would like to be able to have bob@vhost2 on my list. I'm bill@vhost1, but I dont want to be on his list. But i want to see his presence
[21:15:19] <badlop> ok, that means that you will have an item 'to', and he will have an item 'from'
[21:15:39] <badlop> note that, in jabber, he can remove that item, and then you'll not see his presence
[21:15:54] <bomer> but that will add me to his list as well, right?
[21:16:06] <badlop> yes, that's how jabber works
[21:16:21] <badlop> each individual has the right to decide who can view his presence
[21:17:08] <bomer> i'm seeing this functionality with jabberd. i'm to convert us to ejabberd, and this is one of the things my boss wants to see. this actually works just fine with shared rosters, but shared rosters dont work across virtual hosts, which is what i'm aiming for.
[21:19:47] <badlop> i'm looking at http://ejabberd.jabber.ru/shared-roster-all
[21:20:20] <badlop> and the examples do not show how to setup 'spies' who can see but are not seen
[21:20:37] <Kev> bomer: I think I had a brain fart earlier, so please ignore what I said
[21:20:45] <Kev> I shouldn't chat while I'm working
[21:20:58] <bomer> http://www.process-one.net/docs/ejabberd/guide_en.html#htoc64
[21:21:35] <bomer> you can be a spy. make a group called everybody, members: @all@ no display groups. Make a new group called viewall. Add your self to that group, display group: everybody
[21:21:43] <bomer> and now you see everybody.
[21:24:36] <badlop> ah, i'll try it :)
[21:26:12] <bomer> so it has to be possible to do what i want. i can make this work between vhosts, but its basically a bug in add-rosteritem in ejabberdctl. It will not create the group like I would like, as i mentioned earlier. If I could even get that funcationality, that would be enough for me to move forward
[21:26:56] <deryni> Do shared rosters actually add people to your roster or do they just pretend they are there?
[21:27:09] <bomer> but i can in fact, add a person from one vhost, to another persons roster on another vhost, with authorization, and with presence to be seen. it just doesn't add them to a group
[21:27:28] <bomer> looks like they are on my roster to me. if i delete somebody though, they will be added again when you refresh your roster (login out/in)
[21:27:49] <bomer> with authorization should have been *without*
[21:28:09] <badlop> deryni: mod_shared_roster does not modify roster table in DB, they are added dinamically to the xmpp stanzas i think
[21:28:50] <deryni> If that's the case then that is likely how 'spying' works, since nothing in the rosters themselves needs to change.
[21:29:43] <bomer> if they would work between vhosts, my job would be easy
[21:30:13] <deryni> They can't, by definition.
[21:30:38] <bomer> i dont understand why members couldn't contain @all@vhost
[21:32:09] <badlop> bomer: because it's a feature that was not implemented :)
[21:32:17] <deryni> Note: I don't know the vhost code at all.
But given that, I would imagine that each vhost is essentially its own server, and as such doesn't know anything about the users/status/presence/etc. of users on another vhost.
[21:32:57] <bomer> basically, if I do "ejabberdctl add-rosteritem bill tomsyard.com bill bobsyard.com Bill BobsYard from", then bill@tomsyard.com is added to bill@bobsyard.com but without a group, and without an authorization request.
[21:32:59] <badlop> bomer: i've modified mod_ctlextra: now add-rosteritem adds only 1 rosteritem: if you want to add the other, just swap the arguments in the shell script :)
[21:33:29] <deryni> Authorization 'requests' are unrelated to anything we have been discussing here, really.
[21:34:20] <bomer> well with presence being sent it may
[21:34:57] <bomer> badlop: so i should be able to add a roster listing to a person's roster without anything being sent to both sides? will presence be sent as well?
[21:35:00] <deryni> Authorization requests are just the method through which authorization is granted.
[21:35:15] <deryni> The authorization involved in seeing presence has nothing to do with the requests themselves.
[21:35:29] <deryni> Given that you have access to the db involved.
[21:36:26] <bomer> aww ok
[21:36:58] <badlop> bomer: jabberd allows spies by an implementation-specific trick which violates xmpp RFCs; this is ok for a private server like yours. and it seems nobody cared to implement that in ejabberd yet
[21:37:36] <badlop> since this feature may be nice for business servers, you can submit it to bug tracker :)
[21:37:53] <bomer> interesting. then i guess i dont understand what your update is to mod_ctlextra
[21:38:30] <badlop> my add-rosteritem command also violates xmpp rfcs: it allows the admin to modify rosters without permission of the user
[21:38:46] <bomer> and do you have any links with information on the rfc violation in jabberd by chance?
[21:39:08] <deryni> I don't believe the RFCs say anything about the server not being able to do that.
[21:39:11] <badlop> note that this violation provides a good feature, so it's not a bug :)
[21:39:22] <bomer> haha
[21:39:53] <badlop> however, for public servers like gmail and jabber, it's important that each user can grant and revoke access to his presence, right?
[21:39:59] <badlop> *jabber.org
[21:40:41] <bomer> yes i agree. the feature should be available however for private servers to enable if needed
[21:40:43] <badlop> http://www.xmpp.org/rfcs/rfc3921.html#roster
[21:40:56] <deryni> Shared rosters are hard to make work sanely, as evidenced by the issues with every implementation of them I've yet seen.
[21:49:32] <bomer> i still see this as a bug. when i run: "ejabberdctl add-rosteritem bill tomsyard.com bill bobsyard.com bill tomsyard from" then bill@tomsyard.com is added to bill@bobsyard.com's roster. i see presence. bill@bobsyard.com is not in bill@tomsyard.com
[21:51:53] <badlop> you see presence?
[21:51:59] <bomer> yes
[21:52:38] <bomer> only when i do a from
[21:53:01] <bomer> if i do a to and add the contact the other way, i do not see presence. only when i do a from, like the command above
[21:54:12] zapw leaves the room: offline
[21:54:28] <badlop> the person that is spied has the spy with type FROM or BOTH in the roster, right?
[21:55:33] <bomer> i've only done: add-rosteritem bill bobsyard.com bill tomsyard.com bill bobsyard from
[21:55:53] <bomer> then bill@bobsyard.com showed up on the roster for bill@tomsyard.com, but not under a group, and with presence.
[21:56:03] <bomer> bill@tomsyard.com is not on bill@bobsyard.com's roster.
[21:56:07] <badlop> check the rosters
[21:56:12] <badlop> you can use web interface
[21:56:24] <bomer> i have. i've cleared the rosters as well
[21:57:22] <badlop> oh, strange
[21:57:57] <bomer> indeed. if this "feature", hehe, would add the contact in a group, that would be just awesome
[21:58:53] <badlop> did you update my module to svn, compile, install and restart?
[21:59:11] <bomer> not yet, should it be available now?
[21:59:45] <badlop> yes
[21:59:50] <bomer> yup i see it now, one sec
[22:01:24] Christian leaves the room
[22:01:26] <bomer> loaded, 2 accounts logged in, cleared both rosters.
[22:02:07] <badlop> now you must add items one at a time
[22:02:21] <badlop> so, a FROM to the first, and TO to the other
[22:02:36] <badlop> so, you can now set proper nicknames and groups to each one
[22:02:40] <bomer> {"init terminating in do_boot",{undef,[{ejabberd_ctl,start,[]},{init,start_it,1},{init,start_em,1}]}}
[22:03:20] <bomer> i get that when i run the above ejabberdctl command
[22:03:46] <bomer> doh!
[22:04:02] <bomer> my pwd was not correct... sorry, wasn't with the beam files
[22:07:11] <bomer> SWEET!
[22:07:51] <bomer> i think this is going to work exactly how i need it to, just gotta get my syntax right. but i was able to add a user in a group, with presence, without that user being on the other's list
[22:07:55] <bomer> thanks badlop!
[22:08:23] <badlop> i still think it's strange
[22:08:33] kostix leaves the room
[22:09:24] <bomer> ejabberdctl add-rosteritem bill tomsyard.com bill bobsyard.com bill bobsyard to
[22:09:30] <bomer> ejabberdctl add-rosteritem bill bobsyard.com bill tomsyard.com bill tomsyard from
[22:09:47] <bomer> if i do both of those, then bill@bobsyard.com is added to the bobsyard group on the roster for bill@tomsyard.com
[22:15:52] <bomer> this is absolutly perfect! thanks again man
[22:26:56] bomer leaves the room
[22:29:47] Christian joins the room
[22:43:27] sss joins the room
[23:41:54] <TobiasFar> i recently updated my ejabberd to newer trunk and have a bug now. I use a 'ö' in the name of a contact in my roster but now it's a 'ö'.
[23:42:04] <TobiasFar> it's worked before with the same client
[23:59:31] <ermine> which db storage?
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!