[00:04] cloudhead: kriskowal: did you find the answer?
[00:04] kriskowal: 42?
[00:04] cloudhead: Is there even a high level way to do that?
[00:04] cloudhead: : )
[00:04] cloudhead: "does anyone know what the lowest level API available in javascript for enqueueing a callback to the event loop?"
[00:04] kriskowal: oh, no.  i still haven't heard of anything lower-level than setTimeout
[00:04] cloudhead: ok
[00:05] cloudhead: cause I was looking for a way to do this too
[00:05] cloudhead: but it doesn't seem to be possible yet, unless you use exec
[00:10] isaacs: cloudhead: in javascript proper, setTimeout is the only way
[00:10] isaacs: in node, you can use the process.Timer class
[00:11] isaacs: look for GLOBAL.setTimeout in noe.js
[00:11] isaacs: *node.js
[00:11] the_undefined has joined the channel
[00:12] cloudhead: isaacs: but then you would just use 0 as the time value?
[00:12] isaacs: cloudhead: yep
[00:12] isaacs: i've done that in browser js plenty
[00:12] cloudhead: so potentially, I could return a promise before the timeout goes through
[00:12] cloudhead: even if the timeout is 0
[00:13] isaacs: cloudhead: yes.
[00:13] cloudhead: nice
[00:13] isaacs: setTimeout(fn, 0) is not strictly "0"
[00:13] isaacs: it's 0 + (process queue overhead)
[00:13] orlandov: isn't that undefined?
[00:13] isaacs: which is quite small
[00:13] cloudhead: so it'll fork?
[00:13] isaacs: cloudhead: no, it'll just happen later.
[00:13] cloudhead: ok
[00:13] cloudhead: cool.
[00:14] isaacs: like, after the stuff that's already on the process queue
[00:15] jed has joined the channel
[00:17] cloudhead: isaacs: so is it better to use process.Timer, or is it equivalent to setTimeout
[00:18] isaacs: cloudhead: i'd use setTimeout, personally, just becasue then the code is more clear.
[00:18] isaacs: setTimeout is like a 4 line function wrapping process.Timer
[00:18] cloudhead: ah ok, good
[00:21] cadorn has joined the channel
[00:26] hassox: isaacs: thanx heaps for you help yesterday mate
[00:26] hassox: really worked out a lot from that session :)
[00:26] isaacs: no problem, hassox 
[00:26] hassox: where is everyone located?
[00:26] hassox: 
ACTION is in Melbourne Australia

[00:27] isaacs: 
ACTION is in Silicon Valley, CA, where the internet lives.

[00:27] hassox: lol
[00:31] deanlandolt: 
ACTION is in Washington, DC, where jsconf.us lives.

[00:32] hassox: we have kangaroos
[00:32] hassox: does that count for something?
[00:32] isaacs: hassox: and australians
[00:32] hassox: ;)
[00:32] binary42 has joined the channel
[00:42] alexiskander has joined the channel
[00:44] deanlandolt has joined the channel
[00:46] inimino: setTimeout isn't even JavaScript proper
[00:46] isaacs: inimino: well, it's in all the browsers, and in node.
[00:46] isaacs: it's common, and old.
[00:47] inimino: yes, it's something like DOM0
[00:47] isaacs: inimino: except, not terrible and evil
[00:48] inimino: the Web is built on DOM0 :-)
[00:48] inimino: well, the parts of the Web that are mostly JavaScript ;)
[00:49] isaacs: oh, yeah, true that
[00:49] isaacs: dom0 isn't strictly evil
[00:49] orlandov: chaotic neutral?
[00:49] isaacs: hehe
[00:50] isaacs: quick, someone write a blog post exploring the DnD alignments of various programming languages and environments.
[00:50] inimino: heh
[01:06] deanlandolt has joined the channel
[01:08] deanlandolt has joined the channel
[01:09] Connorhd__ has joined the channel
[01:09] martyn__ has joined the channel
[01:10] kriskowal: isaacs NERD
[01:10] kriskowal: i love doing that.
[01:12] deanlandolt has joined the channel
[01:14] deanlandolt has joined the channel
[01:21] alexiskander has joined the channel
[01:25] rictic has joined the channel
[01:30] alex-desktop has joined the channel
[01:35] bryanl has joined the channel
[01:44] jed: what's the most reliable way of getting the protocol/scheme from a request?
[01:45] iamstef has joined the channel
[01:59] mikeal has joined the channel
[02:22] blakemizerany has joined the channel
[02:32] malkomalko has joined the channel
[02:39] dnolen has joined the channel
[03:02] pdelgallego has joined the channel
[03:08] deanlandolt has joined the channel
[03:10] deanlandolt has joined the channel
[03:12] deanlandolt has joined the channel
[03:13] erikcorry|away has joined the channel
[03:16] deanlandolt has joined the channel
[03:19] shfx has joined the channel
[03:24] erichocean: w00t! only 37 functions away from a Solaris v8 build: http://pastie.org/746652
[03:28] micheil has joined the channel
[03:31] rictic: Oh hey, v8 supports __defineGetter__ and __defineSetter__ … sweet
[03:42] micheil: hmm.. what are some usages for them?
[03:42] deanlandolt: micheil: they do what they sound like...and it's pretty broadly useful
[03:43] micheil: an example maybe?
[03:47] micheil has joined the channel
[03:52] hassox: is there any documentation on V8 or do you have to read the sauce to see what's what?
[04:00] micheil: yeah, you actually drink the sauce
[04:00] micheil: no, there's docs on google's website for it
[04:00] hassox: thanx
[04:05] micheil: by the way, it's source code, not sauce code
[04:05] binary42 has joined the channel
[04:06] hassox: I know
[04:06] hassox: I always call it the sauce
[04:06] dnolen has joined the channel
[04:07] hassox: dunno why..
[04:07] Connorhd__ has joined the channel
[04:07] RJ2 has joined the channel
[04:11] inimino: secret's in the sauce
[04:12] micheil: 
ACTION is trying to think of how to structure the api to the IMAP module he's working on

[04:13] deanlandolt: hassox: i think v8's more a juice than a sauce ;)
[04:14] inimino: ...IMAP?
[04:25] hassox: that's tru
[04:44] jazzychad: when node accepts a TCP connection, does it spin the connection off to a different local port or pool the connections?
[04:49] eddanger has joined the channel
[04:51] micheil has joined the channel
[04:51] jazzychad: seems to pool them on the same port from what i can tell
[04:56] jcrosby has joined the channel
[05:04] mikeal has joined the channel
[05:12] kriszyp: is it best to submit bugs to http://github.com/ry/node/issues or to the mailing list?
[05:21] inimino: kriszyp: I'd say the mailing list
[05:22] rakeshpai has joined the channel
[05:25] kriskowal has joined the channel
[05:26] kriskowal has joined the channel
[05:32] scudco has joined the channel
[05:32] scudco has joined the channel
[05:37] mikeal has joined the channel
[05:38] dnolen has joined the channel
[05:47] JimBastard has joined the channel
[05:54] erichocean: w00t! v8 now compiles on Solaris 10!
[05:56] JimBastard: nerdgasm
[06:19] eddanger has joined the channel
[06:23] kriskowal has joined the channel
[06:26] aurynn: now port it to sparc!
[06:28] jcrosby has joined the channel
[06:29] logicuce has joined the channel
[06:30] onar_ has joined the channel
[06:37] eddanger: Hi All, I am having a problem where after the 3rd request to my hello world app, it has: Segmentation fault: 11 (core dumped)
[06:37] eddanger: server is FreeBSD 7.2...
[06:48] blakemizerany has joined the channel
[06:55] hassox has joined the channel
[07:00] erichocean: aurynn: now I'm getting the rest of node.js working ;-)
[07:05] JimBastard: peeeace
[07:20] rakeshpai has joined the channel
[07:40] scudco has joined the channel
[07:49] JimBastard has joined the channel
[07:51] JimBastard: hello party people
[07:51] Connorhd: morning
[07:53] JimBastard: do you know if there are any xmpp modules for node? like using websockets i guess?
[07:54] Connorhd: don't think I've seen any
[07:54] JimBastard: word
[07:59] kriskowal has joined the channel
[08:03] rakeshpai_ has joined the channel
[08:23] michaelk^ has joined the channel
[08:39] the_undefined has joined the channel
[08:40] ryah: hello
[08:41] kriskowal: hey
[08:41] kriskowal: if i undestand clocks, it is absurdly early for you, ryah, and absurdly early for me.
[08:41] kriskowal: i mean, late.  late.
[08:43] JimBastard: clocks tend to be linear
[08:43] kriskowal: radial, really.  helical in some sense, maybe more than one.
[08:43] JimBastard: unless traveling at high speeds
[08:43] kriskowal: yeah, and stretchy at high speed
[08:44] kriskowal: not a good way to loose weight though, but maybe live longer
[08:44] JimBastard: heh
[08:45] kriskowal: mr. bastard, what is your github handle?
[08:45] JimBastard: yo ryah ive got a team competing in a programming contest this weekend, node is gonna be part of my stack ill be showing it off
[08:46] JimBastard: http://github.com/Marak/
[08:46] erikcorry|away has joined the channel
[08:46] kriskowal: i've only heard the term "stack" in competition in the context of cal tech's senior ditch day
[08:47] kriskowal: wherein the seniors, in an attempt to keep the lower classmen from pranking their rooms, leave complex puzzles for them.
[08:48] JimBastard: k
[08:49] kriskowal: best wishes and g'night.
[08:55] micheil has joined the channel
[08:55] ryah: JimBastard: cool
[08:58] JimBastard: mostly automated deployment, http testing, and doing compacts and builds of client side JS
[08:58] JimBastard: you da the man
[09:04] mikeal has joined the channel
[09:14] JoePeck_ has joined the channel
[09:15] erikcorr1|away has joined the channel
[09:15] logicuce has joined the channel
[09:17] iamstef has joined the channel
[09:18] frodenius has joined the channel
[09:20] elliottcable has joined the channel
[09:22] Yuffster has joined the channel
[09:23] shfx has joined the channel
[09:23] jchris has joined the channel
[09:27] teemow has joined the channel
[09:30] halorgium has joined the channel
[09:37] skampler has joined the channel
[09:38] johan-s has joined the channel
[10:15] logicuce has joined the channel
[10:16] micheil has joined the channel
[10:32] pdelgallego has joined the channel
[10:44] malkomalko has joined the channel
[10:57] bry has joined the channel
[11:22] micheil: ryah: are you still about?
[11:31] chrisfarms has joined the channel
[11:51] JoePeck_ has joined the channel
[11:52] micheil: does anyone know if it's possible to export an object from a module as the default module..?
[11:53] micheil: I mean, I have the object SMTP, and I want to be able to setup in a piece of code to have var SMTP = require("libs/smtp");
[11:54] micheil: without needing to tell it to point exports.SMTP to SMTP
[11:54] micheil: by doing var SMTP = require("libs/smtp").SMTP;
[11:58] jspiros has joined the channel
[12:13] bry has joined the channel
[12:16] JimBastard has joined the channel
[12:29] micheil: morning JimBastard 
[12:30] JimBastard: morning
[12:32] jasondavies has joined the channel
[12:35] jasondavies: hmm I'm getting a version mismatch when running node-waf, is it safe to just change the version in bin/node-waf to 1.5.10?
[12:37] jasondavies: ls
[12:37] jasondavies: oops wrong window
[12:38] JimBastard: i can haz root
[12:39] micheil: JimBastard: make me a sandwhich =P
[12:39] JimBastard: sudo
[12:48] micheil: hmm..
[12:56] jasondavies has joined the channel
[12:58] micheil: elliottcable: Mu looks like a nice name, but is there any code to go with the name?
[13:04] alex-desktop has joined the channel
[13:09] micheil: the_undefined: hmm.. that was an interesting article you did on rightjs and the cargo cult, I'd actually never heard that term before..
[13:10] the_undefined: micheil: richard feynman came up with the concept of cargo cult science
[13:10] micheil: hmm..
[13:10] micheil: well, either way, I can see your point
[13:11] micheil: in many respects, it could be similar to compare Node.js to php+apache
[13:11] the_undefined: I was a little harsh, but mostly because I like writing articles that piss people off, they have a much higher comment conversion rate ;)
[13:11] micheil: ones apples, the other is oranges
[13:12] the_undefined: yeah, it's not fair to do that comparison
[13:12] the_undefined: unless you put it in clear context
[13:12] micheil: true
[13:12] micheil: hmm.. I wish I could figure out the best way to write this api..
[13:12] the_undefined: for example lets say you want to write a chat application, apache+php might still be viable due to the fact that your team members know 
[13:13] the_undefined: micheil: which API?
[13:13] micheil: one for an SMTP lib
[13:13] micheil: the_undefined: the is true, I know in my case, I know js better then any of the other languages I've studied, which makes me love node just the little bit more.
[13:14] micheil: I'm not sure whether this api should be very low-level or high-level, or middle level like ruby's implementation is
[13:14] micheil: python's is very low level I think, I can quite comprehend what they
[13:14] micheil: 're doing though.
[13:15] JimBastard: mmmmmmmm
[13:15] JimBastard: TCPIP protocols
[13:15] JimBastard: whats up with the bittorrent port
[13:16] micheil: hmm?
[13:16] the_undefined: micheil: why are you writing the lib? What is the goal?
[13:16] micheil: the_undefined: to see if I can, and because I think it'd be useful
[13:17] the_undefined: micheil: hm, bad inspiration for an API :)
[13:17] micheil: yeah, true
[13:17] michaelk^ has joined the channel
[13:18] the_undefined: I don't know enough about smtp to be really useful
[13:18] micheil: well, I don't know if I want to have something that has a few public methods and a heap of private, or a fully public api
[13:18] the_undefined: Usually I do it by writing some code of how I want to use the library
[13:18] the_undefined: just a few sample use cases
[13:19] the_undefined: where I make up how the library would talk to me
[13:19] micheil: hmm./.
[13:19] micheil: okay
[13:19] the_undefined: there are some flaws because you may forget about some semantics, but its usually a good start
[13:20] micheil: I guess the best way to go would be something like:
[13:20] micheil: var smtpConn = SMTP.connect(host, port, options...);
[13:21] micheil: erm
[13:21] micheil: no
[13:22] micheil: the_undefined: and you have one creepy penguin
[13:24] the_undefined: micheil: creepy penguin?
[13:24] micheil: http://debuggable.com/posts/burning-apple:4ad4a03b-6998-4bb8-b8a6-6c43cbdd56cb
[13:25] JimBastard: tldw
[13:32] the_undefined: haha
[13:47] kriszyp has joined the channel
[13:55] micheil: ryah: I really have no idea how you read those histograms you like reading...
[14:33] n8o has joined the channel
[14:44] jspiros has joined the channel
[14:44] deanlandolt has joined the channel
[14:48] alexiskander has joined the channel
[14:50] n8o_ has joined the channel
[15:02] micheil has joined the channel
[15:08] binary42 has joined the channel
[15:12] rolfb has joined the channel
[15:20] ryah: micheil: histograms?
[15:21] micheil: yeah
[15:21] jed has joined the channel
[15:21] micheil: the ones that benchmark between node and other servers, eg, the one between node and the facebook one
[15:21] ryah: oh - its reponse time
[15:22] ryah: left is is faster
[15:22] jed_ has joined the channel
[15:23] n8o has joined the channel
[15:23] micheil: okay...
[15:23] micheil: hmm.. what code editors do people use on os x?
[15:25] dnolen has joined the channel
[15:26] jasondavies: 
ACTION uses vim

[15:27] micheil: I'm looking for something a little more.. gui based.
[15:27] micheil: On my ubuntu box I use Gedit, with a fair few plugins
[15:27] alexanderte has joined the channel
[15:28] jasondavies: a lot of OS X people seem to swear by textmate
[15:29] jasondavies: there is gvim of course but it's probably not what you're looking for when you say gui-based
[15:35] micheil: I've been using textmate, but I can't figure out how to get tab completion in it
[15:35] inimino: 
ACTION recommends vim

[15:39] rolfb has joined the channel
[15:40] aurynn: I swear by textmate.
[15:41] inimino: aurynn: you mean vim
[15:41] alexanderte has joined the channel
[15:41] alexanderte: then*
[15:41] aurynn: no
[15:41] aurynn: I outgrew vim
[15:41] inimino: see, vim is for the young and hip, textmate for the old and infirm!
[15:42] frodenius: hrhr
[15:43] aurynn: it was more that gvim was irritating, and ran poorly on osx10.3, at the time.
[15:43] inimino: ah
[15:44] inimino: 
ACTION recommends real vim with a nice terminal emulator

[15:45] aurynn: 
ACTION refuses to edit text in the terminal

[15:45] micheil: inimino: you're proly using uxrvt?
[15:45] micheil: ;p
[15:45] inimino: I am
[15:45] aurynn: I have ridiculously high-resolution bitmapped displays.. why would I use a legacy console system?
[15:45] inimino: urxvtcd in fact :)
[15:46] ashb: aurynn: there are gui versions of vim. http://code.google.com/p/macvim/
[15:46] ashb: or gvim
[15:46] aurynn: ashb, but not very good.
[15:46] ashb: macvim is very nice
[15:46] inimino: aurynn: because nothing so far comes close the power of unix :)
[15:47] inimino: aurynn: and a grid of character cells turns out to be about perfect for a text editor UI anyway
[15:48] inimino: all those pixels help draw nice glyphs!
[15:48] aurynn: 
ACTION checks out MacVim

[15:48] ashb: the google code project is a nice cocoa version. not the old crappy carbon version
[15:49] micheil: hmm.. night folks
[15:49] logicuce has joined the channel
[15:50] aurynn: hm. I remember this.
[15:50] aurynn: 
ACTION wonders where her vimrc went

[15:50] bryanl has joined the channel
[15:51] ashb: i've often thought of switching to another editor, but i do too much remote editing and working on more than one machine
[15:51] ashb: and i want something i can use (mostly) the same in all environments
[15:51] aurynn: I use vim when I need to do spot edits on remote systems
[15:51] aurynn: but not for primary development
[15:52] ashb: function print () { print.apply(undefined, arguments); }
[15:52] ashb: oops
[15:52] ashb: :/
[15:53] aurynn: hm. So far this is sensible, a vim that uses the OS key chords
[15:53] aurynn: some of them, anyway
[15:53] quirkey has joined the channel
[15:53] ashb: aurynn: let me know how you get on. I've actually been thinking of switching to textmate. so aview from 'that side of the fence' would be good
[15:54] ashb: (just cos i fancy a change)
[15:55] micheil: ashb: textmate is... different
[15:55] micheil: but I come from using Gedit, so, eh.
[15:55] ashb: heh
[15:58] aurynn: I really, really like TextMate
[15:58] aurynn: and that it lacks a comparable alternative really locks me to OSX
[15:58] nrstott: E text editor
[15:59] nrstott: works well on windows, uses textmate bundles.  linux version under development
[15:59] aurynn: I don't use windows
[15:59] nrstott: source code is released if you want to hack on it to make it work better on linux
[15:59] micheil: 
ACTION feels like just doing ssh -X to grab at Gedit.

[15:59] nrstott: it uses cygwin on windows, so it basically works on linux
[15:59] aurynn: I've been paying attention to the linux build, but it looks like it died
[15:59] nrstott: but it's not stable
[15:59] nrstott: i dont get how people use gedit ;0
[16:00] nrstott: id rather use gvim
[16:02] ashb: nrstott: ugh cygwin
[16:05] nrstott: ashb, ugg, i agree
[16:05] ashb: 2 things i dont like about it: 1) you can't use the utils outside of a cygwin/bash shell
[16:05] ashb: 2) it still uses the failsome cmd.exe
[16:05] ashb: which is the worst terminal ever invented
[16:11] aurynn: yeah, E looks pretty dead at this point.
[16:37] nefD: vim for the win!
[16:37] nefD: vim + NERDTree + ctag & TagList = the best evar
[16:38] nefD: and to make things even better, the terminator terminal app, which provides for tabbed sessions in the same window
[16:39] logicuce has joined the channel
[16:48] frodenius: yeah, terminal tabs are so awesome!
[16:52] johan-s has joined the channel
[16:54] nefD: i get kind of OCD with my desktop space, so having multiple terminal windows bugs the shart out of me
[16:55] inimino: 
ACTION recommends xmonad

[16:55] RayMorgan has joined the channel
[16:56] nefD: hmm, thanks, I'll check that out
[17:06] dnolen has joined the channel
[17:14] sudocarl has joined the channel
[17:24] ryanmcgrath has joined the channel
[17:33] teemow has joined the channel
[17:36] RayMorgan_ has joined the channel
[17:41] mikeal has joined the channel
[17:53] scudco has joined the channel
[18:06] sudoer has joined the channel
[18:17] aguynamedben has joined the channel
[18:19] eddanger has joined the channel
[18:20] isaacs has joined the channel
[18:20] jcrosby has joined the channel
[18:26] dnolen has joined the channel
[18:28] ericflo has joined the channel
[18:28] logicuce has joined the channel
[18:34] erichocean: so now we need a node.js-based buildbot
[18:35] erichocean: I recommend porting this one: http://integrityapp.com/
[18:37] jtoy has joined the channel
[18:39] mikeal has joined the channel
[18:48] isaacs: erichocean: we could probably port Jake pretty easily
[18:48] isaacs: oh, you said buildbot, not build tool
[18:49] erichocean: yeah; I've read the buildbot source cover to cover; it wasn't pretty
[18:49] erichocean: node.js would be a much better fit IMO
[18:53] qFxo has joined the channel
[19:02] soitgoes has joined the channel
[19:03] alexiskander has joined the channel
[19:19] bryanl has joined the channel
[19:38] Yuffster_ has joined the channel
[19:44] inimino has joined the channel
[19:48] inimino has joined the channel
[20:03] malkomalko has joined the channel
[20:06] RayMorgan has joined the channel
[20:14] jazzychad: erichocean: would runcoderun.com work for your auto-build needs?
[20:17] jed has joined the channel
[20:30] hornbeck has joined the channel
[20:35] isaacs: kriszyp: you around?  got an idea about http://www.pastie.org/746255
[20:35] isaacs: it looks like the JSON.stringify() happening on the module exports obj is strictly for debugging purposes, and serves absolutely no purpose most of the time.
[20:43] brandon_beacher has joined the channel
[20:49] erichocean: jazzychad: hmm, I can't tell; the website doesn't say how it works
[20:49] erichocean: but given the pricing, I'd say "no"
[21:08] kriszyp: isaacs: that's what I was thinking too
[21:08] kriszyp: haven't looked at the code
[21:21] jazzychad: erichocean: i think it pulls repos and runs a build script and reports sucess/fail
[21:21] jazzychad: i'm not sure how it all works, but one of my friends works on it, so it rang a bell
[21:23] blakemizerany has joined the channel
[21:28] jamiew has joined the channel
[21:38] rtomayko has joined the channel
[21:38] logicuce has joined the channel
[21:41] the_undefined has joined the channel
[21:51] mikeal has joined the channel
[22:00] RayMorgan has joined the channel
[22:01] hassox has joined the channel
[22:01] hassox: good morning 
[22:01] hassox: quick q for anyone out there
[22:02] hassox: does a http.Response object emit an event when response.finish() is called?
[22:05] isaacs: kriszyp: i'm thinking about a way to remove those debug() statements in node.js without losing the debugging ability, but it's a bit tricky.
[22:05] isaacs: it crossed my mind to dig into the build tools and strip them out unless you do a ./configure --debug or something
[22:06] isaacs: but that scares me.
[22:06] isaacs: ryah_away: you around?  any comments re ^
[22:07] isaacs: my other idea was to create an internal module called "debug", and only switch it on if the env var is there.
[22:15] the_undefined: isaacs: I don't think response.finish() gives you any event other than the callback you can pass into
[22:16] isaacs: the_undefined: hrm.... context?
[22:16] isaacs: 
ACTION lost the conversation where we were talking about that, apparently.

[22:16] isaacs: 
ACTION 's stack overflows sometimes

[22:16] the_undefined: isaacs: nvm. I think I was trying to talk to hassox 
[22:16] the_undefined: :)
[22:16] hassox: oh hey
[22:16] isaacs: ok, great.  still not senile.
[22:16] the_undefined: too much beer
[22:16] hassox: oh you can pass a callback into it?
[22:16] the_undefined: isaacs: well, at least we can't proof it yet ;)
[22:17] the_undefined: hassox: yeah, I think so
[22:17] hassox: that would be sufficient
[22:17] hassox: any ideas on how I would do taht?
[22:17] hassox: that
[22:17] the_undefined: let me check first
[22:18] hassox: it doesn't look like it's in the dox
[22:18] the_undefined: hassox: nvm, you cannot pass a function to it
[22:18] the_undefined: you only can for HttpClient
[22:19] the_undefined: hassox: but, I think you can add a "close" listener to the underlaying tcp connectino
[22:19] the_undefined: something like this:
[22:19] hassox: hrm
[22:19] the_undefined: req.connection.addListener("close", function(){ ...});
[22:19] the_undefined: that should work
[22:19] hassox: that would do the job too
[22:19] hassox: awesome :D
[22:19] hassox: t
[22:19] hassox: hanx
[22:20] inimino: hassox: note that the connection isn't guaranteed to be closed when the request finishes
[22:20] hassox: hrm
[22:21] the_undefined: inimino: a freaking keep-alive
[22:21] hassox: ah yeah
[22:21] hassox: good point
[22:21] hassox: thanx :)
[22:21] hassox: well that still leaves me with the issue of finish
[22:21] hassox: I guess I could hijack the function and wrap it
[22:22] hassox: but I really really don't wanna do taht
[22:22] the_undefined: message.emit("sent");
[22:22] the_undefined: let me see where this hooks into
[22:24] aurynn has joined the channel
[22:24] gjritter has joined the channel
[22:25] the_undefined: hassox: try res.addListener('sent', function() {})
[22:25] hassox: kk 1 sec
[22:27] hassox: the_undefined: u are the man/woman!
[22:27] erichocean: cool
[22:27] hassox: thanx the_undefined :D
[22:27] the_undefined: hassox: np ;)
[22:28] the_undefined: hassox: FYI, I just looked through lib/http.js 
[22:28] hassox: kk
[22:28] hassox: 
ACTION must lose the fear of the js sauce

[22:29] hassox: it's a lot more readable than I thought it was a week ago ;)
[22:35] rolfb has joined the channel
[22:35] the_undefined has joined the channel
[22:38] hassox: anyone know of a good style guide for js that doesn't suck?
[22:42] jazzychad: hassox: i know several people like dojo's style guide http://dojotoolkit.org/developer/StyleGuide
[22:42] hassox: thanx jazzychad 
[22:42] hassox: I'll give it a read
[22:43] jazzychad: hassox: also some suggestions on the node wiki http://wiki.github.com/ry/node/contributing
[22:44] hassox: awesome :)
[22:44] hassox: just a quick q
[22:44] hassox: do you use ; all the time?
[22:44] hassox: or only when you're on multi statements on one line?
[22:45] jazzychad: personally, i do it all the time
[22:45] jazzychad: 100%
[22:46] jazzychad: not that this reason applies to node, but the main reasons i do it is out of habit (C coder) and ...
[22:46] jazzychad: it's the only way to minify the code later into one line
[22:46] hassox: oh that's a good point
[22:46] hassox: but is that really a concern in node?
[22:47] hassox: I can't see myself minifying in node
[22:47] jazzychad: well, i don't see many people needing to minify node code
[22:47] jazzychad: haha "node code"
[22:47] isaacs: hassox: i use ; most of the time.
[22:47] isaacs: for compression, it really doesn't matter.  \n is the same number of bytes as ;, and you don't compress ssjs, anyhow.
[22:47] isaacs: i use ; because then i *know* that i'm not chaining.
[22:48] hassox: chaining?
[22:48] isaacs: and i often do blah\n\t.method(...)\n\t.method(...) etcl
[22:48] jazzychad: right
[22:49] isaacs: http://gist.github.com/259097
[22:49] isaacs: methods that return "this"
[22:49] hassox: ah right
[22:49] hassox: oh cool :)
[22:49] isaacs: when i don't see a ; on a line, i'm mentally looking for the next method.
[22:49] hassox: I didn't realise you could split it onto multiple lines
[22:49] isaacs: yeah, js is extra dangerously flexible like that ;)
[22:50] jazzychad: i'm so used to using ;'s that code w/o them (python) just looks naked
[22:50] hassox: lol
[22:50] jazzychad: crockford has some screwy examples of what happens when js assumes where to put the ; for you
[22:50] jazzychad: i just like to avoid all that
[22:50] rictic has joined the channel
[22:51] gjritter: you have to be careful with splitting lines in javascript, particularly with returns
[22:51] gjritter: http://gist.github.com/259100 is an example that does something you probably wouldn't expect...
[22:51] jazzychad: oh hey mr. nerve
[22:52] gjritter: hey mr. jazzy
[22:52] isaacs: yes, return\n is dangerous.
[22:52] jazzychad: quite
[22:52] isaacs: that's why OTB is so very very bad in JS.
[22:52] jazzychad: otb?
[22:53] isaacs: otb = "one true brace", or "put braces on their own lines".  in JS, braces are butterflies.
[22:53] isaacs: } else { etc.
[22:53] jazzychad: yes yes, right
[22:53] hassox: } else { << - isn't good?
[22:53] isaacs: hassox: it IS good
[22:53] jazzychad: no, that is good
[22:53] hassox: oh ok
[22:53] hassox: phew ;)
[22:53] isaacs: that's what i'm saying.  make your elses into butterflies
[22:53] hassox: that's what I've been doing
[22:53] isaacs: } butterfly P
[22:53] hassox: ah gotcha
[22:53] isaacs: s/P/\{
[22:53] erichocean: is anyone aware of coming changes to the coupling lib in node.js/deps?
[22:54] jazzychad: i used to do OTB a long time ago... then the style guide at work forced butterfly (they called it cuddling)
[22:54] isaacs: a brace on its own line can be interpreted as a code block.
[22:54] jazzychad: so i've changed my ways thankfully
[22:54] isaacs: and js doesn't have block-scope, but it does have arbitrary blocks (for no raisin)
[22:54] hassox: a code block? 
[22:54] hassox: js has those?
[22:54] hassox: can you pass them around n stuff?
[22:54] isaacs: hassox: crazy, rite?
[22:54] isaacs: nope.
[22:54] isaacs: you can't do *anything* with them.
[22:54] hassox: ah right
[22:54] isaacs: they do nothin.g
[22:54] hassox: oh
[22:54] hassox: that kinda sucks
[22:54] isaacs: it also has line labels.
[22:54] hassox: 
ACTION was getting excited 

[22:54] jazzychad: isaacs: futurama reference?
[22:55] isaacs: jazzychad: yes, it is canon in my religion.
[22:55] jazzychad: that's the greetest
[22:55] isaacs: ^_^
[22:55] isaacs: line labels, but no gotos.  code blocks, but no block scope.
[22:56] isaacs: combine that with "," ending a statement, arbitrary whitespace, and expressions-as-valid-statements, and you've created a language in which an object literal can be very badly misinterpreted.
[22:56] hassox: sounds like quite a few gotchas
[22:57] gjritter: it's a powerful language with a lot of accidents
[22:57] jazzychad: nicely put
[22:57] isaacs: gjritter: i agree.
[22:57] isaacs: http://gist.github.com/259105
[22:59] jazzychad: i put "javascript: the good parts" on my wishlist
[22:59] hassox: what's with the trailing "," you mentioned?
[22:59] isaacs: hassox: you can combine statements with ,
[22:59] isaacs: so you can do things like this:
[22:59] isaacs: return (a=b, b = c, c = d)
[22:59] isaacs: the combined statement evaluates to the value of the last statement.
[23:00] isaacs: that lets you do things like: for (var i = 0, l = arr.length; i < l; i ++)
[23:00] hassox: so in this case the value of d ?
[23:00] isaacs: right
[23:00] hassox: ok
[23:00] hassox: so it's not like splatting in rub
[23:00] hassox: ruby
[23:00] hassox: where you can say
[23:00] hassox: elem1, elem2 = [1,2]
[23:00] hassox: can you use it like that?
[23:00] isaacs: hassox: actually, that's destructuring
[23:01] isaacs: in the plans for es5, implemented in spidermonkey today
[23:01] hassox: cool
[23:01] isaacs: var [foo, bar, baz]  = [1, 2, 3]
[23:01] isaacs: and even: var {foo} = {foo : "bar"}
[23:01] hassox: what's that one do?
[23:01] isaacs: in your example, elem1 would be left as-is, and elem2 would be set to the array.
[23:02] isaacs: that sets teh var "foo" to the "foo" member of the obj
[23:02] isaacs: in this case, "bar"
[23:02] hassox: yeah that would be cool
[23:03] jazzychad: oh yeah, useful
[23:03] isaacs: hassox: unfortunately, it's not widely available yet.
[23:05] jazzychad: oh right, since there are people here now... i had a question last night...
[23:05] jazzychad: when you are accepting tcp connections in node, does it spin off the connection on a new local socket, or pool all connections on the listening port?
[23:06] jazzychad: sorry... *new local port
[23:07] isaacs: ... crickets...
[23:07] jazzychad: hah
[23:08] jazzychad: my cursory investigation with netstat seems to indicate they all stay on the same port
[23:10] jazzychad: well, that was a fruitful conversation :)
[23:10] gjritter: funny, I saw a tweet about irc crickets earlier today :-)
[23:10] bryanl has joined the channel
[23:11] isaacs: jazzychad: send it to the mailing list.  i'm curious myself.
[23:11] isaacs: i'm sure *someone* knows.  at least ryah_away probably does, i mean, he wrote the thing, right?
[23:12] jazzychad: i could dig through the source i suppose
[23:12] jazzychad: i was just curious about scaling stuff, if one could have a huge number of open connections
[23:13] jazzychad: not like 2^16 ports isn't huge already, but still
[23:14] gjritter: There was a thread on the mailing list about how to scale to 100000 connections. I think most people ran into OS/benchmarking difficulties above several hundred simultaneous connections.
[23:14] gjritter: running out of sockets, sockets not being closed by ab fast enough, etc.
[23:15] RayMorgan_ has joined the channel
[23:15] jazzychad: hmm, that's interesting
[23:15] logicuce has left the channel
[23:16] gjritter: http://groups.google.com/group/nodejs/browse_thread/thread/d19df83c7fd1f3ca
[23:17] gjritter: this is not directly node related, but some of the linux kernel parameter settings would be helpful for node as well: http://www.metabrew.com/article/a-million-user-comet-application-with-mochiweb-part-1/
[23:22] jazzychad: woah what a thread.. thanks
[23:27] inimino: hassox: I use semicolons only where necessary
[23:27] RayMorgan_ has joined the channel
[23:29] the_undefined: anybody interested in discussing promises/deferreds might want to join #commonjs
[23:29] the_undefined: having a little discussion there
[23:30] logicuce has joined the channel
[23:35] logicuce has left the channel
[23:35] isaacs: how would you folks feel about something like require("debug").log(thing, "category")
[23:35] isaacs: to replace sys.debug()
[23:35] isaacs: and sys.error()
[23:36] isaacs: then it could have a mechanism to flip a switch to turn on/off certain types of messages.
[23:36] isaacs: require("debug").show("core", "warning", "info", "error")
[23:36] isaacs: require("debug").throw("error")
[23:37] isaacs: i'm finding that now that i'm used to categorized debugging with Y.log(), i'm really missing it in node.
[23:40] jazzychad: i've written a consolelog module.
[23:40] jazzychad: http://github.com/jazzychad/consolelog.node.js
[23:40] jazzychad: feel free to fork/tweak
[23:40] isaacs: jazzychad: i think that your thing is a bit higher-level that my debug module.
[23:41] isaacs: but maybe you could put my thing to use somehow.
[23:41] jazzychad: mm, oh i see what you're saying now
[23:41] isaacs: i'm looking to fix the way that debug() is used in node.js
[23:42] isaacs: it does some things that cause unexpected side-effects in some cases.
[23:42] RayMorgan_ has joined the channel
[23:44] mikeal has joined the channel
[23:46] cloudhead has joined the channel
[23:49] hassox: lads
[23:49] hassox: if I'm making a function to add a callback to a queue
[23:50] hassox: would callback or addCallback be more appropriate?
[23:50] hassox: callback(fn) or addCallback(fn)
[23:50] malkomalko_ has joined the channel
[23:52] jazzychad: i would say addCallback seems to be the convention
[23:52] mediacoder: hassox: hmm.. addCallback() and addErrback() are used on Promise success/error
[23:52] hassox: eyah thought so
[23:52] hassox: thanx guys
[23:52] mediacoder: i would say addCallback implies a promise-success ..not sure if your callback is that :-)
[23:53] hassox: not quite
[23:53] hassox: it _may_ run
[23:54] mediacoder: hm, dunno.. might confuse people maybe, if your callback is not related to a success :-)
[23:55] jazzychad: but just 'callback' would imply that it would be the only callback (to me)
[23:56] jazzychad: and may overwrite whatever callback(s) is already registered
[23:56] jazzychad: maybe pushCallback ?
[23:57] mikeal has joined the channel
[23:57] hassox: this is a callback that will execute unless something further down the stack decides it wanted to finish things
[23:58] hassox: stack == callback queue
[23:58] jazzychad: maybeCallback
[23:59] jazzychad: so like a last ditch/default callback ?
[23:59] hassox: ah good call
[23:59] hassox: thats nice
[23:59] hassox: kinda yeah
[23:59] jazzychad: defaultCallback(fn)
[23:59] jazzychad: tho i guess that could also be confusing :/
[23:59] hassox: yeah