CHAT LOG 1

[16:14:37] --- LOGGING STARTED
[16:14:52] --- pgmillard has joined
[16:15:44] <pgmillard> foo
[16:15:45] <pgmillard> bar
[16:15:46] --- rob has joined
[16:15:46] <pgmillard> test
[16:15:52] <pgmillard> nice.
[16:15:53] <pgmillard> working.
[16:15:56] <pgmillard> + logs
[16:15:56] <rob> pong
[16:15:57] <rob> nice ;)
[16:16:00] <rob> muc?
[16:16:04] <pgmillard> yes.
[16:16:08] <rob> sweeeet
[16:16:20] <rob> hmm
[16:16:22] <rob> one
[16:16:22] <rob> word
[16:16:22] --- ChatBot has joined
[16:16:23] <ChatBot> [pgmillard]: My client could whoop your client any day
[16:16:23] <rob> lines
[16:16:24] <rob> :)
[16:18:52] --- kibozer has joined
[16:28:01] --- hawke has joined
[16:42:59] * pgmillard drops a pin
[16:43:01] <pgmillard> :)
[16:43:36] <rob> hehe
[16:51:23] --- brian has joined
[16:53:04] <rob> pgmillard: opinion please ..
[16:53:19] --- pcurtis has joined
[16:53:20] <ChatBot> [pcurtis]: "Erk"
[16:53:31] <rob> if a client sends an invisible broadcast presence, should that be forwarded to the users other sessions? if so, as invisible or unavailable?
[16:53:46] <pgmillard> rob: LOL
[16:54:01] <pcurtis> rob: in mu-c, invisible == unavailable
[16:54:09] * pgmillard wonders what it means to be invisible to yourself.
[16:54:25] <pgmillard> rob: gimme a min.
[16:54:49] <rob> yeah I know .. invisible goes out as unavailable to everyone on your roster .. I'm just wondering if the presence that goes to your other sessions is to inform other sessions of your actual state .. or something else
[16:54:52] <rob> np, ta
[16:55:21] <pgmillard> unavailable IMO.
[16:55:50] <rob> cool .. thats easier to implement, so I was leaning that way
[16:56:42] <rob> invisible is such an abomination anyway
[16:57:51] * pcurtis screams .....
[16:58:06] <pcurtis> invisible? an abomination?
[16:58:08] <pcurtis> hehehehe
[16:58:20] * rob grins
[16:58:25] <pcurtis> now it I could just figure out Objective C ....
[17:04:55] * pcurtis recommends http://www.pimphats.com
[17:05:23] <hawke> for Objective C?
[17:05:34] <pcurtis> no, for a general alugh
[17:05:37] <pcurtis> laugh, even
[17:06:25] * hawke recommends http://www.livejournal.com/community/pro_scurvy/
[17:08:50] <pcurtis> hawke: that's sick.
[17:09:12] <hawke> It's a parody of pro_anorexia
[17:09:16] <hawke> which is truly sick
[17:12:42] <hawke> So, how does invisibility work in Jabber?
[17:13:03] * hawke looks for a jep.
[17:13:28] <hawke> hey hey...there it is!
[17:18:33] <rob> hawke: its wrong
[17:18:34] <rob> the jep I mean
[17:19:15] <hawke> as in, does not describe how invis actually works?
[17:19:40] <rob> well .. it describes a way that invisibility could work .. but not how it actually works in implementations
[17:19:42] <rob> :P
[17:19:45] <rob> its pretty simple
[17:19:52] <hawke> k...
[17:19:54] <rob> you send <presence type='invisible'/>
[17:19:58] * pgmillard prefers the jinc impl.
[17:20:12] <pgmillard> rob: it uses <presence type="visible"/>
[17:20:13] <rob> the server broadcasts unavailable to your roster
[17:20:27] <hawke> ok..
[17:20:33] <rob> but the server keeps you in the running for the top sessions, so you can still get messages
[17:20:44] --- mogul has joined
[17:20:45] <rob> pgmillard: so its implemented in the jep 18 way?
[17:20:50] <rob> interesting .. I wonder where that came from
[17:20:55] <rob> wondered even
[17:21:26] <hawke> So, it has the auto-status failings described in the jep?
[17:21:26] <pgmillard> rob: yeah, it looks like JEP-18 is what we implement.
[17:21:29] <mogul> Can anyone tell me where there's a MUC server publicly accessible?
[17:21:37] <hawke> muc.jabber.org
[17:21:39] <pgmillard> mogul: your using one.
[17:21:44] <mogul> Oh ok
[17:21:45] <hawke> is this muc now?
[17:21:45] <pgmillard> mogul: right now.
[17:21:48] * pgmillard nods
[17:21:50] <hawke> sweet
[17:21:53] <mogul> I thought jabber.org was still doing gc1.0
[17:22:02] <pgmillard> mogul: I switched it today.
[17:22:20] --- mogul has left
[17:23:03] <pgmillard> rob: the _really_ crappy thing about the non-JEP-18 way (the 1.4.2 way) is that the client must maintain it's OWN avails list.
[17:23:14] <rob> yeah
[17:23:14] <pgmillard> ie, who I'm not invisible to.
[17:23:18] <pgmillard> which totally blows.
[17:23:26] <pgmillard> JEP-18 allows the server to manage that.
[17:23:33] * rob nods
[17:23:42] <hawke> so, j2 does, or does not do jep-18?
[17:23:49] <rob> j2 does it the 1.4 way
[17:23:52] <pgmillard> rob: I'd lobby for -18 support in j2 :)
[17:23:57] <pgmillard> 1.4 is just plain wrong.
[17:24:01] <pgmillard> IMO
[17:24:02] <rob> I'm not aware of any (non-jinc) clients that do it
[17:24:05] <rob> so there wasn't much point
[17:24:07] * hawke agrees with pgm
[17:24:11] <pgmillard> Exodus does invisible.
[17:24:21] <hawke> psi does invis
[17:24:28] <rob> nah .. I'm in favour of ditching it entirely in favour of privacy lists
[17:24:34] <rob> nono, I mean do invis in the -18 way
[17:24:35] <pgmillard> well, thats true.
[17:24:48] <rob> adding types to toplevel packets is evil :P
[17:24:48] * pgmillard shrugs
[17:25:24] <hawke> privacy lists?
[17:25:40] <pgmillard> hawke: yeah. check the IETF IM doc.
[17:25:50] <hawke> k
[17:26:02] <pgmillard> http://www.jabber.org/ietf/draft-ietf-xmpp-im-19.html#privacy
[17:28:33] --- mjw99 has joined
[17:28:39] <hawke> Nice. Does anybody do privacy lists though?
[17:29:39] <rob> j2
[17:29:54] <rob> tkabber does as well, though I think it may be the older jep 16 lists
[17:33:00] * pgmillard nods.
[17:33:09] --- mjw99 has left
[17:33:13] <pgmillard> the jinc server does an earlier version of them also.
[17:33:21] <pgmillard> like draft 12 or something.
[17:46:48] --- davel has joined
[17:51:35] --- torp has joined
[17:51:56] <torp> hello
[17:52:05] --- Caveman has joined
[17:52:42] --- Caveman has left
[17:54:35] <torp> weird problem here: transports running in a different process have suddenly stopped connecting to the main jabber server process, something about connection rate limiter
[17:58:02] --- torp has left
[18:01:24] --- pcurtis has left
[18:03:24] --- pcurtis has joined
[18:03:24] <ChatBot> [pcurtis]: "Erk"
[18:04:08] <pcurtis> ugh
[18:09:01] <pgmillard> pcurtis: ?
[18:09:13] <pgmillard> pcurtis: BTW, conference.j.org is mu-c now.
[18:14:22] <pcurtis> I see that
[18:14:34] <pcurtis> but no one can change the subject/
[18:14:37] <davel> whats with the users _and_ the participants
[18:14:49] <pcurtis> huh?
[18:15:16] <davel> I have both users and participants
[18:15:22] <davel> and moderators
[18:15:39] <pcurtis> you should update tkabber
[18:15:51] <pcurtis> moderators are the room admins
[18:16:04] <pcurtis> and Participants are the great unwashed.
[18:17:22] <davel> yeah i know that -- just u have users as well :/ -- Latest version is 0.9.5beta.
[18:17:39] <pcurtis> hmmm
[18:17:40] <davel> u=i
[18:17:52] <davel> maybe its just leftover presence
[18:18:02] --- hawke has left
[18:18:03] <pcurtis> I ahve 0.9.6 and I only see the moderators and participants
[18:18:54] <davel> cvs?
[18:19:46] <davel> yeah im sure its cause ive still got presence stuff for the conference (non-muc room) least closing it fixes :)
[18:20:39] --- hawke has joined
[18:21:32] <davel> pcurtis: so did u cvs that presumably?
[18:21:38] <pcurtis> yes
[18:21:48] <pcurtis> I try to keep up
[18:22:09] <pcurtis> I want a client with privacy lists ..... ;)
[18:22:31] <pcurtis> pgm: why didn't you disable private chats?
[18:23:58] <hawke> Is there a jabber python lib?
[18:24:13] <pcurtis> yeah .... there is
[18:24:21] <pcurtis> pyjab, IIRC
[18:24:54] <hawke> I suppose I should've just googled. ;-)
[18:25:11] <pcurtis> hehe
[18:30:23] <pgmillard> pcurtis: why would I?
[18:30:37] * pcurtis sighs
[18:30:40] <pcurtis> nevermind
[18:30:47] --- pcurtis has left
[18:30:58] <hawke> pgm: I was kinda wondering that myself.
[18:31:03] <hawke> annoying people, maybe?
[18:34:38] <pgmillard> maybe.
[18:34:42] <pgmillard> Blocking in exodus works.
[18:35:23] <davel> b4n
[18:35:36] <hawke> after n
[18:35:36] --- davel has left: Lost connection
[18:36:23] --- pgmillard has left
[18:37:08] --- pcurtis has joined
[18:37:08] <ChatBot> [pcurtis]: "Erk"
[18:37:29] <pcurtis> grrrr
[18:37:36] <pcurtis> not everyone uses Exodus ....
[18:39:59] <hawke> brb
[18:40:00] --- hawke has left
[18:40:08] --- Hawke has joined
[18:40:29] <pcurtis> oh, no.
[18:40:33] <Hawke> What?
[18:40:34] <pcurtis> time to update tkabber again.
[18:40:39] <pcurtis> :)
[18:40:40] <Hawke> Why?
[18:40:48] <pcurtis> I'm a week behind
[18:40:57] <Hawke> hehe
[18:41:20] <Hawke> This version is working better than the 0.9.5 beta in Debian...
[18:41:30] <pcurtis> good
[18:41:31] <pcurtis> brb
[18:41:35] --- pcurtis has left
[18:43:08] --- pcurtis has joined
[18:43:09] <ChatBot> [pcurtis]: "Erk"
[18:43:18] <pcurtis> hmmm
[18:43:28] <pcurtis> still has that annoying SSL pop up.
[18:43:35] <Hawke> ??
[18:43:48] <Hawke> Didn't see anything here...
[18:43:56] <pcurtis> when tkabber connects with SSL, it pops up a warning about self-sgned certs.
[18:44:07] <Hawke> Ahhhh
[18:44:19] <Hawke> Heh, it doesn't like me to shift-backspace. <grin>
[18:45:12] <Hawke> Does Customize -> Warnings let you disable that?
[18:45:19] <pcurtis> hmmm
[18:45:31] <Hawke> Strange, even though I have that on, and my server is self-signed ssl, it doesn't pop up for me.
[18:46:16] <pcurtis> let me try that ....
[18:46:20] --- pcurtis has left
[18:46:30] --- pcurtis has joined
[18:46:30] <ChatBot> [pcurtis]: "Erk"
[18:46:41] <pcurtis> yup ... that's it.
[18:46:49] <pcurtis> that's new in this week's release
[18:47:55] <Hawke> brb
[18:47:58] --- Hawke has left
[18:50:52] --- xiabre has joined
[18:51:01] --- xiabre has left
[19:04:04] --- shane has joined
[19:04:25] <shane> cool...when did this happen ?? [20:07:14] *** This room supports the MUC protocol.
[19:06:04] --- JohnnyRocket0 has joined
[19:09:04] --- boo has joined
[19:17:39] <rob> shane: about three hours ago
[19:18:32] <shane> nice
[19:28:03] <JohnnyRocket0> so... once jabberd 1.4 is totally obsoleted.. how will muc be used?
[19:28:13] <JohnnyRocket0> that's one of the few things i haven't figured out yet..
[19:28:26] <pcurtis> JCR
[19:28:37] <shane> there's a standalone app.... preach on pcurtis.. lol
[19:28:51] <JohnnyRocket0> jabber common runtime? or what?
[19:28:56] <JohnnyRocket0> whatcha mean?
[19:28:58] <pcurtis> JohnnyRocket0: I have built a standalone wrapper for mu-c
[19:29:10] <pcurtis> http://jabber.terrapin.com/
[19:29:20] <JohnnyRocket0> does it work yet?
[19:29:27] <pcurtis> It's running on a few servers now.
[19:29:44] <JohnnyRocket0> cool
[19:29:47] <JohnnyRocket0> thanks for the info
[19:29:50] <JohnnyRocket0> will be very useful
[19:29:54] <pcurtis> yes
[19:30:02] <pcurtis> I working on Yahoo-t next.
[19:30:32] <pcurtis> I hope to have JCR ready enough to take any jabberd 1.4 transport and make it standalone.
[19:30:37] <rob> pcurtis: btw, I'll have it up on jabber.org.au in the next few days
[19:30:47] <pcurtis> :)
[19:30:59] <pcurtis> nice. good thing I'm sleeping when you're awake ....
[19:31:08] <shane> heh
[19:31:13] * rob grins
[19:33:25] <pcurtis> rob: I have ficve mu-c + JCRs at work ... no leaks, no crashes. More than two weeks now.
[19:33:56] --- hawke has joined
[19:33:57] <pcurtis> and our j2 servers are holding up pretty well. Except for the annoying little leak in the 'sm'.
[19:34:14] <rob> yeah .. did you read what I sent you yesterday?
[19:34:28] <pcurtis> yes.
[19:34:45] <rob> what say you?
[19:34:46] <pcurtis> The counts of nodes, domains, and resources isn't climbing very much.
[19:35:06] <pcurtis> logins with the same info don't seem to cause any growth.
[19:35:23] <pcurtis> it's just the logged out people's stuff is still in the hash.
[19:35:45] <pcurtis> no biggie .... the size is pretty stable. Higher than I'd like, but stable.
[19:36:15] <pcurtis> All of the NAD stuff frees perfectly, so I'm focusing only on the logins/logouts
[19:38:06] <rob> ok .. I'll probably leave it for now. I'm planning a new memory allocation scheme with proper expiry and stuff anyway for the next round of dev, so I'll use that when it materialises
[19:38:12] * rob prepares an rc2 release
[19:38:18] <rob> fixed so many bugs this week - yay!
[19:39:47] <pcurtis> :)
[19:39:50] <pcurtis> :(
[19:39:58] <pcurtis> now I have to rebuild ....
[19:40:12] <pcurtis> and figure out the Solaris LDAP stuff again.
[19:40:43] <rob> it should patch the same ..
[19:40:52] <rob> if you send it to me, I can include it ;)
[19:41:22] <pcurtis> I think it will need to be a config option.
[19:41:46] <rob> possibly
[19:41:47] <pcurtis> basically, delete all references to lber.h and -llber
[19:41:56] <pcurtis> and it work perfectly.
[19:42:35] <rob> which version of solaris is this?
[19:42:39] <rob> coz I have access to 8
[19:42:46] <pcurtis> 2.7+
[19:43:04] <rob> hrm .. my sol8 box has ldap.h and lber.h
[19:43:23] <pcurtis> erk
[19:43:29] <pcurtis> right ... -llber
[19:43:36] <rob> no liblber tho
[19:43:37] * pcurtis is in HP mode
[19:43:40] * rob smiles
[19:43:47] <rob> lemme see what I can do ..
[19:44:06] <pcurtis> no biggie.
[19:44:22] <pcurtis> I just chomped up the ac.helper, IIRC
[19:45:16] <rob> yeah, thats what I'll do .. just seeing if I can make it autodetect
[19:45:21] --- boo has left
[19:45:47] <rob> hmm .. openldap libldap declares a dep on liblber .. I wonder if I really need to explicitly link it ..
[19:45:48] * pcurtis is afraid of autotools
[19:46:19] <rob> yeah, me too
[19:46:35] <rob> I never know what I'm doing
[19:46:48] <rob> actually, automake/conf is easy .. libtool is scary
[19:46:53] <pcurtis> hehehe
[19:46:58] <pcurtis> true enough
[19:47:10] <pcurtis> and there isn't any good docs
[19:47:16] <pcurtis> at least not that I can find
[19:47:41] <rob> yeah .. the autobook is decent, but way out of date .. and the infopages are comprehensive, but unreadable .. typical gnu :(
[19:48:16] * pcurtis nods
[19:48:31] <pcurtis> I tried to wade through the infodocs ... I really did
[19:48:39] <pcurtis> really!
[19:49:00] <rob> I'm still trying :(
[19:52:08] --- fizk has joined
[19:52:55] --- fizk has left
[20:06:12] <shane> anyone here a net::jabber guru? :)
[20:07:20] <shane> i'm trying to figure out the difference between PresenceSend and MUCJoin when wanting to join a muc room.
[20:07:47] <pcurtis> good question
[20:07:53] <pcurtis> lemme me try it.
[20:07:54] <pcurtis> brb
[20:07:59] --- pcurtis has left
[20:08:02] <hawke> perhaps "your mom" is the difference.
[20:08:40] --- pcurtis has joined
[20:08:40] <ChatBot> [pcurtis]: "Erk"
[20:09:36] <pcurtis> <presence to='jdev@conference.jabber.org/pcurtis'><x xmlns='http://jabber.org/protocol/muc'/></presence>
[20:09:44] <pcurtis> that's the difference
[20:09:55] <shane> ahhh
[20:10:09] <pcurtis> you're indicating to the mu-c room you can handle all the muc namespaces and such
[20:10:22] * shane nods..
[20:10:37] <pcurtis> otherwise, mu-c won't return all the roles, etc.
[20:10:55] <shane> makes sense.
[20:11:36] <pcurtis> <presence xmlns='jabber:client' to='pcurtis@www.terrapin.com/home' from='jdev@conference.jabber.org/rob'><x xmlns='jabber:x:delay' stamp='20031204T22:15:46' from='jdev@conference.jabber.org/rob'/><x xmlns='http://jabber.org/protocol/muc#user'><item role='moderator' affiliation='admin'/></x></presence>
[20:11:49] <pcurtis> quite a bit of stuff comes back ... ;)
[20:12:01] <pcurtis> !time
[20:12:01] <ChatBot> [time]: Fri Dec 05, 2003 02:15:36 UTC
[20:13:11] <pcurtis> it's nice because you can find out how long rob has been loitering in the room, etc. Who are the admins ...
[20:13:56] <shane> lol
[20:15:06] * rob chuckles
[20:15:08] <rob> I'm an admin?
[20:15:10] <rob> neat :)
[20:15:17] <pcurtis> yup
[20:15:18] <shane> so, if i *don't* MUCJoin, is it safe to assume muc will not listen to room management commands from the bot?
[20:15:20] <rob> I should get a client that actually allows me to nuke ppl ;)
[20:15:31] <pcurtis> shane: not sure about that.
[20:16:00] <pcurtis> I know that if you don't indicate the mu-c namespace, you won't be able to admin the room.
[20:16:07] <pcurtis> kick and ban might work, however.
[20:16:12] <shane> ok..
[20:16:20] <pcurtis> better to send it ....
[20:16:42] <pcurtis> rob: yeah
[20:16:51] <shane> well, i'll change the channel code to use MUCJoin anyway...may as well since i'll be dealing in muc rooms anyway. wish i had found that little function last night.. lol
[20:16:52] <pcurtis> you're the school marm tonight.
[20:17:57] --- pcarrillo has joined
[20:18:47] --- pcarrillo has left
[20:21:24] <rob> pcurtis: did you have to make any code changes at all to authreg_ldap?
[20:21:35] <pcurtis> none
[20:21:44] <pcurtis> except for our special aliases
[20:22:09] <rob> good. just trying to find a common define in the two headers to test on .. its no longer any good to test for LDAP_VENDOR_VERSION
[20:22:17] <rob> I think LDAP_API_VERSION will do ..
[20:22:44] <pcurtis> VENDOR_VERSION is not consistent
[20:23:23] <rob> yeah, I know that (now)
[20:23:30] <rob> it was when I was assuming openldap
[20:23:46] <pcurtis> yes
[20:23:51] <pcurtis> I had to nuke that test
[20:24:12] --- zion has joined
[20:24:18] <pcurtis> it *may* be enough to compile and link test lldap and the lldap+llber
[20:24:56] <rob> no its cool - both debian and redhat have built properly against just ldap
[20:25:12] <rob> which is the right way to do it anyway .. use lib deps, thats what they're there for
[20:25:35] <pcurtis> :)
[20:25:42] <pcurtis> one question about privacy ....
[20:25:49] <pcurtis> (in jabber2)
[20:26:07] <pcurtis> does the client have to activate a privacy list after each login?
[20:26:45] <rob> it has to activate one for the session, but only if it only wants to override the default
[20:26:56] <rob> you can set a default list, and a session list
[20:27:01] <rob> default hangs around after you log out
[20:27:03] <rob> session doesn't
[20:27:06] <rob> session overrides default
[20:27:10] <pcurtis> ok ... so j2 always will activate the default list on login.
[20:27:17] <rob> yup, if you have one
[20:27:20] <pcurtis> :)
[20:27:22] <pcurtis> coolio
[20:27:38] * pcurtis finishes up the jhatbot privacy plugin
[20:29:08] <rob> bah ..
[20:30:29] <rob> solaris can't handle a NULL passed as a %s arg to *printf .. it cores
[20:30:38] <rob> every other platform prints "(null)" instead
[20:30:49] <shane> ever look at code you wrote and say to yourself, "wtf were you thinking?" ?
[20:30:51] <rob> but I can't wrap it, coz they're varargs functions
[20:30:57] <rob> shane: all the time .. c2s
[20:31:16] <rob> so I'm going to have to put a check on every call to log_debug
[20:31:21] <rob> sx_debug, log_write, etc
[20:31:34] <rob> sigh .. I don't need this shit right now !@$#
[20:31:37] <shane> or wrap your own printf :)
[20:31:57] <zion> that seems like a better solution
[20:32:19] <zion> a simple sed script could then replace all printf calls
[20:32:20] <rob> shane: I can't .. coz its stdargs, there's no way to get to the arguments and then turn them back into a ...
[20:32:41] * shane hmms
[20:32:48] <rob> unless there's some way to populate a va_list directly .. but I don't think there is
[20:36:36] --- fizk has joined
[20:37:40] --- fizk has left: Disconnected
[20:38:40] <shane> of course, i can't find what i'm looking for when i'm looking for it... somewhere i have my own *printf commands...
[20:38:46] <shane> er, functions.. heh
[20:41:35] <rob> yeah, I have those too .. but they aren't wrappers
[20:42:20] --- pcurtis has left
[20:47:18] --- port8001 has joined
[20:53:07] --- port8001 has left
[20:53:12] <zion> rob: ever think about using apr?
[20:53:36] <zion> it is designed to solve problems like that
[20:57:53] <JohnnyRocket0> the apache portable runtime?
[21:00:42] <rob> well .. thinking about apr for 2.1
[21:01:05] <rob> though I already have ripped certain bits of code (snprintf/vsnprintf) from apr anyway
[21:01:36] <rob> though the guy asking thinks that it may be suboptimal because its not optimised for the platform like libc might be
[21:01:43] <rob> I'm not sure thats a legitimate concern though
[21:02:46] <rob> there's the verdict on wrapping varargs - you can't
[21:02:47] <rob> http://mail.python.org/pipermail/python-list/2001-December/076970.html
[21:02:55] <rob> time for another apache rip ;)
[21:03:35] <zion> I think ppl have a tendency to focus on pointless performance concerns
[21:03:44] <zion> i.e. apr printf is too slow
[21:04:11] <zion> if you are that good of a dev that that is a legitimate concern, then I envy you sir
[21:04:21] <rob> I don't think about optimisations too much .. I get it working first, then find the bottlenecks and fix it
[21:04:39] <rob> I could worry about performance issues all day .. but it wouldn't help a single bit
[21:04:49] <rob> why bother optimising something thats not on the criticial path
[21:05:08] <rob> and log_debug certainly isn't - if you're running debug logging in production, you're an idiot
[21:05:17] * rob is running debug logging in production, fwiw
[21:05:24] <rob> and pointer tracking, and everything else
[21:05:26] <zion> ;hehe
[21:05:49] <shane> well, on core functions it *can* add to performance loss across the board, but i can't imagine in this case it'll matter. that's what faster cpu's are for (hey, it works for Microsoft)
[21:06:21] * shane goes back to perl...
[21:06:25] <rob> yes .. I agree core/critical path stuff should be the first candidates for optimisation
[21:06:36] <rob> but I'm not going to do it until someone is actually having a problem
[21:06:43] <shane> absolutely.
[21:06:45] <rob> why write code that I don't need and can't test?
[21:23:19] --- infiniti has joined
[21:24:11] <infiniti> http://www.kdedevelopers.org/node/view/254
[21:27:32] --- Jasonatoffice has joined
[21:28:51] <Jasonatoffice> Howdy!!!!
[21:28:53] <rob> the answer, as always, is awareness of the issues .. not using C is not going to make your code invulnerable, and you're a fool if you go in expecting that
[21:28:58] <rob> hey jason
[21:29:05] <Jasonatoffice> :))
[21:29:30] <Jasonatoffice> Man, the last two hours Ive been tutor'ng uers on using transports :)
[21:29:47] <infiniti> rob: well, that is true also. but he does raise a valid point, that most of the listed software was written in C, and most had buffer overflow exploits.
[21:30:11] <infiniti> there are certainly other things to worry about, but these C programs keep getting nabbed by buffer overflows
[21:30:52] <rob> infiniti: yes .. and the reason for that is that C give you plenty of rope, and most people aren't aware of just how easy to get it wrapped around your neck
[21:31:08] <rob> that doesn't mean that C is inherently insecure
[21:31:13] <rob> any bug could be a security hole
[21:31:21] <infiniti> of course not
[21:31:29] <infiniti> but are the openssl developers not aware of C's power?
[21:31:51] <rob> and using high-level languages doesn't guarantee you anything .. you don't know if the interpreter (usually written C) has holes or not
[21:32:22] <infiniti> well, tim's suggestion was to use C++. unlike saying all servers should be coded in python or java, using C++ is pretty easy to swallow
[21:32:29] <shane> I don't think you can blame the language... blame the programmers.
[21:32:38] <rob> a) writing crypto is hard. b) people make mistakes. c) why assume the openssl programmers are any better at hacking than you or me?
[21:32:57] <infiniti> shane: you are right
[21:33:11] <rob> I think its very very dangerous to say "write software in this language, because it'll make you less prone to security problems"
[21:33:17] <rob> you're just asking for trouble
[21:33:19] <rob> shane: agreed
[21:34:07] <shane> there's a ton of diagnostics tools out there to help avoid overruns.. if you don't use them, you're asking for trouble, imo.
[21:34:54] <shane> and i'll admit it, i'm guilty too.
[21:35:07] <rob> so am I - I know j2 is full of string problems
[21:35:13] <rob> dunno where, but there's no way its secure
[21:35:26] <infiniti> i think it comes down to this... the more you have to write the same thing over and over, the more likely there will be a mistake.
[21:35:32] <shane> hence why i run it as a nonpriv user in a chroot() environment ;)
[21:35:44] <rob> shane: smart :)
[21:35:44] <infiniti> no one is perfect, and so you just gotta take that into consideration when you are programming and/or choosing a language for your project
[21:36:26] <rob> infiniti: agreed. paranoia is good - if you go in thinking that any tool you use is perfect and won't bite you in the arse, then you're a fool
[21:36:36] * rob is narky today, sorry for the tone
[21:36:45] <shane> there's a lot of things that can be done to prevent serious issues with overruns as well.. take linux's grsecurity for example..
[21:37:03] <infiniti> well, consider this. why not use C++ just for string and array classes, but stop there?
[21:37:19] <shane> for me it's cos i hate C++ lol
[21:37:48] <infiniti> well, you have to ask yourself... why am i manually moving memory around each time i deal with this string?
[21:37:52] <infiniti> is it for performance or what?
[21:38:26] <rob> infiniti: are you asking me specifically? or just rhetoricals? :)
[21:38:32] <shane> for me, mainly performance.. my background is in assembly, so i still have the mindset of only using what i need to use.
[21:38:47] <infiniti> rob: err, rhetorical i guess. not you-you :)
[21:39:16] <rob> shane: I'm exactly the same .. started with assembly
[21:39:23] <rob> 6800 .. not this dogs breakfast intel crud ;)
[21:39:25] <infiniti> shane: i can appreciate that, sometimes you have to write low level code given your situation
[21:39:46] <infiniti> but are servers in need of that?
[21:40:03] <shane> heh...i'm not sure if i want to admit where i stated in assembly...wasn't 6800 ;)
[21:40:18] <infiniti> no one writes code in assembly for servers really. i think we've all decided that it was better to have the portability of C than to get the extra speed bonus of asm.
[21:40:26] <rob> (actually, I did 6502, but its the same thing)
[21:40:50] <rob> its also a learning experience ..
[21:41:02] <shane> but if i run into a need to code for Z80's, i'm all set...
[21:41:09] <infiniti> haha
[21:41:17] <rob> I don't write software for the outcome .. I write it to learn
[21:41:28] <rob> I don't learn anything if I use someone elses lib
[21:41:57] <rob> yes, I know it means I make mistakes that others have avoided, that I could have avoided
[21:42:02] <infiniti> hmm, don't put that quote on the jabberd website :P
[21:42:21] <rob> heh
[21:42:21] <shane> lol
[21:42:31] <rob> I have never ever claimed that I write jabberd for the users
[21:42:35] <rob> and I never will ;)
[21:42:43] <rob> I write it for me :)
[21:42:44] <infiniti> maybe so, but everyone depends on you :)
[21:43:00] <rob> I make no claims that it works, or is bug free, or won't blow things up
[21:43:09] <rob> perhaps .. thats their choice, not mine :)
[21:43:18] <shane> coding for me was always a hobby.. a hobby that turned into a career..
[21:43:24] <rob> I started some real project planning gear this morning ..
[21:43:40] <rob> http://jabberd.jabberstudio.org/vision/goals.shtml
[21:43:44] <rob> still wip
[21:45:13] <infiniti> looks good
[21:46:19] <rob> yeah .. we need it
[21:46:29] <rob> noone knows where the project is going
[21:46:40] <shane> yeah, does look good.
[21:46:48] <rob> so now I can outline tasks and stuff, and tie them back into the goals
[21:46:58] <rob> if it doesn't fit, then its out .. that, or goals need changing
[21:47:00] <rob> its useful :)
[21:47:19] <rob> there's more to do .. tasks, milestones, architecture, etc ..
[21:47:52] * JohnnyRocket0 is going to make the best jabberd2 admininistration
[22:01:31] --- Jasonatoffice has left: Replaced by new connection
[22:04:17] --- jaramir has joined
[22:04:47] <jaramir> hi there :o)
[22:05:07] <infiniti> yo
[22:05:07] <jaramir> i got a a problem with jabberd2: it installs fine but it dosn't let me register. how can i solve?
[22:05:22] <rob> need more info that that
[22:05:43] <jaramir> C2S : sx (io.c:228) encoding 138 bytes for writing: <stream:error xmlns:stream='http://etherx.jabber.org/streams'><not-authorized xmlns='urn:ietf:params:xml:ns:xmpp-streams'/></stream:error>
[22:05:52] <jaramir> the error seams to be that..
[22:06:05] <jaramir> server is @ slack.it and is open..
[22:08:03] <rob> again, need more than that .. is anything around that line that says why you're not authorised?
[22:08:14] <jaramir> tryng again..
[22:09:17] <jaramir> C2S : sx (io.c:270) 18 ready for writing C2S : sx (io.c:228) encoding 138 bytes for writing: <stream:error xmlns:stream='http://etherx.jabber.org/streams'><not-authorized xmlns='urn:ietf:params:xml:ns:xmpp-streams'/></stream:error> C2S : sx (chain.c:79) calling io write chain C2S : sx (io.c:289) handing app 138 bytes to write
[22:10:03] <jaramir> what kind of info whould i look for?
[22:10:05] <infiniti> i think we want the part where the client sends data
[22:10:25] <infiniti> like, stuff before that
[22:10:26] --- jaramir has left: Lost connection
[22:10:27] <rob> dude, read the logs .. is there anything that says why?
[22:10:27] <rob> if you can't decipher it, can you post it somewhere?
[22:10:36] --- pgmillard has joined
[22:10:36] <ChatBot> [pgmillard]: My client could whoop your client any day
[22:10:47] <rob> /msg
[22:10:50] <rob> meh
[22:11:00] <rob> shane: can't priv msg from gaim .. but thanks ;)
[22:11:08] <shane> lol
[22:11:23] <rob> shane: actually, what's your jid?
[22:11:33] <shane> shane@jabber.silicondairy.net
[22:12:16] <shane> so need SRV records to work in all clients...
[22:12:17] <rob> cute
[22:12:54] <rob> no, gaim can't send priv msgs through groupchat .. it gets confused when constructing a jid from your nick
[22:13:06] <rob> just tries to send to='nick'
[22:13:07] <rob> very silly :)
[22:13:13] <pgmillard> that sucks
[22:13:25] <shane> i was more referring to my need to have the 'jabber' part in my JID :)
[22:13:43] <rob> well there's that too, indeed
[22:14:27] <infiniti> shane: as a pre-SRV solution, you might be able to set the jid and hostname separately in your client.
[22:14:44] <shane> yeah, psi supports that
[22:16:03] <pgmillard> shane: exodus daily builds support SRV lookups.
[22:16:04] <pgmillard> FWIW.
[22:16:10] <shane> unfortunately, being an ISP and offering Jabber to our customers, we need to keep things as simple as possible for the "not so technically inclined" user. so, i opted for the jabber.* address
[22:16:52] <rob> pgmillard: where's that exodus release?
[22:16:56] <rob> the new one ..
[22:17:02] <shane> pgmillard: very cool :)
[22:17:02] --- jaramir has joined
[22:17:13] <pgmillard> rob: http://exodus.jabberstudio.org/daily/
[22:17:14] <pgmillard> for now.
[22:17:23] <pgmillard> 0.9 will be out Real Soon Now (tm).
[22:17:33] <pgmillard> I think the .pl translation finally got updated today.
[22:17:43] <jaramir> here again, gaim crashed..
[22:17:50] <rob> excellent :)
[22:17:51] --- sean has joined
[22:17:58] <rob> waiting for 0.9
[22:18:09] <pgmillard> rob: maybe next week sometime.
[22:18:17] <shane> i considered doing some router port forwarding tricks, but figured it wasn't worth it.
[22:18:20] <pgmillard> I'm still trying to pin down some rogue cores as well.
[22:18:48] <jaramir> C2S : sx (io.c:183) decoded read data (134 bytes): <iq id="msg_2" type="set"> <query xmlns="jabber:iq:auth"> <username>jaramir</username> C2S : <resource>Gossip</resource> C2S : </query> C2S : </iq>
[22:19:01] <jaramir> think the request is this..
[22:19:21] <infiniti> hmm, there is no password or digest :)
[22:19:24] <rob> there's no password in that
[22:19:32] <rob> pgmillard: sweet :)
[22:19:36] <rob> finally, a real client ;)
[22:19:41] <pgmillard> heh
[22:19:47] <shane> heh
[22:19:53] <infiniti> :O
[22:19:56] <sean> I have a .Net client that I would like you guys to check out. Please let me know what you think. http://www.transactim.com/products/messenger/download.htm
[22:19:57] <jaramir> C2S : sx (io.c:183) decoded read data (163 bytes): <iq id="msg_3" type="set" to="jaramir@slack.it"> <query xmlns="jabber:iq:register"> <username>jaramir</username> C2S : <password>uyiuygiuy</password> C2S : </query> C2S : </iq>
[22:20:20] <infiniti> rob: hey we're getting there. xmpp's on the back burner until i have everything doing the error stuff right.
[22:20:42] <rob> infiniti: ;)
[22:20:47] <rob> just stirring
[22:21:05] <pgmillard> does anyone know the ACL's around browsing w/ mu-c ??
[22:21:23] <pgmillard> can some rooms not get disco'd if I'm not a member/admin/owner ??
[22:21:28] <shane> i'm usually in Linux... so, unfortunately, no Exodus for me.. tried it in wine for the hell of it.. wasn't too successful;)
[22:22:07] --- sean has left: Disconnected
[22:22:15] <infiniti> transactIM, eh?
[22:23:18] <rob> shane: yeah, I want to get it going under wine
[22:23:23] <jaramir> ok, maybe i got it..
[22:23:41] <jaramir> where is the "timeout on registration" configured?
[22:23:52] <shane> strangely, winex got further than wine, but still didn't work.
[22:24:35] --- sean has joined
[22:24:35] --- sean has left
[22:24:35] --- sean has joined
[22:25:33] --- jaramir has left
[22:25:36] --- jaramir has joined
[22:26:57] <jaramir> it's working a bit weird, but working.. got to hit the "register" button for variuos time really fast..
[22:27:35] --- jaramir has left
[22:27:38] <infiniti> well, gossip is a bit new to the table, probably just client bugs
[22:28:20] --- sathe has joined
[22:28:39] --- sathe has left
[22:28:44] --- jaramir has joined
[22:29:59] <jaramir> infiniti: about the registration timeout?
[22:30:47] --- cmullins has joined
[22:32:56] <infiniti> or the login stuff, i dunno
[22:33:02] <cmullins> Did you base TransctIM on any sdk, or write it from scratch?
[22:33:19] <sean> TransactIM uses jabber.net
[22:33:44] <cmullins> fun, fun. :)
[22:34:09] --- bill.mallin has joined
[22:34:09] --- jaramir has left
[22:34:10] <cmullins> How easy / tough was developing a client against jabber.net?
[22:34:28] <sean> Early version were buggy.
[22:34:36] <cmullins> (I'm curious.. as I've never tried. I've only really used my/our SDK)
[22:34:46] <sean> No its no problem
[22:34:53] <sean> I had to integrate SSL myself
[22:35:01] <sean> Now its no problem
[22:35:03] <cmullins> Which SSL package did you use?
[22:35:05] <cmullins> Mentalis?
[22:35:48] <sean> yep
[22:35:57] <sean> There are a few bugs there to
[22:36:04] <cmullins> We're using them as well, but having some big problem with them under heavy load.
[22:36:27] <sean> Using them for a server or client?
[22:36:44] <cmullins> just SSL2/3/TLS encrypted sockets in general.
[22:37:01] <sean> I have not seen any performance issues.
[22:37:23] <sean> I have seen some ocassional exceptions that cause it to hang on initialization....
[22:37:25] <cmullins> For a single client on a machine, they're fine - no performance issues at all.
[22:37:52] <sean> I am a firm believer in c/c++ for servers
[22:38:05] <cmullins> I used to be. I spent enough years doing it.... :)
[22:38:23] <cmullins> Do you have SASL yet in TransactIM ?
[22:38:39] <sean> No
[22:38:45] <sean> Thats the authentication stuff?
[22:38:49] <cmullins> Any plans for it?
[22:38:58] <sean> Wouldnt that sit in the server?
[22:39:07] <cmullins> It's needed on both sides of the equation.
[22:39:22] <cmullins> The client needs to support it, as does the server.
[22:39:35] <sean> No plans
[22:39:36] <cmullins> Rob's baby, Jabberd2 supports it beautifully.
[22:39:57] <sean> I need to check if they are adding it to jabber.net
[22:40:30] <cmullins> I wonder. I know Joe added StringPrep - I dunno what his SASL plans are.
[22:40:34] <sean> What is it that you are building?
[22:40:37] --- JD has joined
[22:40:53] <cmullins> I work for Winfessor - we do .NET based XMPP stuff....
[22:41:03] <sean> thats what I thought
[22:41:05] --- Jasonatoffice has joined
[22:41:22] <sean> good night
[22:41:25] --- sean has left
[22:41:26] <cmullins> 'night.
[22:41:32] --- bill.mallin has left
[22:50:33] --- Jasonatoffice has left
[22:54:48] --- infiniti has left
[22:55:19] --- cmullins has left
[22:57:23] --- jmrbcu has joined
[22:57:33] <jmrbcu> hi everyone
[22:58:07] --- jmrbcu has left
[22:58:14] <JD> hi
[23:04:10] --- kirjava has joined
[23:04:50] <kirjava> what is file tranfer proxy?
[23:05:17] <kirjava> (its listed in the new jabber.org/user/publicservers.php page)
[23:05:35] <JD> it's part of a JEP for doing file transfers through a server in order to transverse firewalls
[23:05:49] <JD> it borrows a bit from SOCKS5 with some jabber stuff on top of it
[23:07:10] --- Kelan has joined
[23:07:24] <Kelan> how does the email gateway work?
[23:08:33] <Kelan> ok later all!
[23:08:35] --- Kelan has left
[23:08:42] <JD> well, the email gateway is just an smtp server that's also a jabber gateway..
[23:08:42] <kirjava> thanks
[23:08:56] <kirjava> JD: too late ;)...
[23:09:01] <JD> heh, yeah
[23:10:17] * kirjava buggers off to try fenchruch (php jabber client) under windows for the first time
[23:14:22] --- highway has joined
[23:14:43] <highway> Hi
[23:15:14] <JD> hi
[23:15:28] --- kirjava has left
[23:18:05] --- JD has left
[23:27:39] --- highway has left
[23:30:18] --- arciphera has joined
[23:30:50] <arciphera> whenever I log onto the MSN gateway, it automatically changes my MSN display name to (null).
[23:31:08] <arciphera> anyone know what exactly causes this and/or how to stop it from happening?
[23:31:23] --- arciphera is now known as ARC
[23:31:42] --- rob has left
[23:38:10] --- JohnnyRocket0 has left: Disconnected
[23:44:00] --- ARC has left
[23:44:55] --- cmullins has joined
[23:45:35] --- cmullins has left