[00:00] _ry: isaacs: we need to tell v8 about the rethrow thing
[00:00] isaacs: _ry: kk
[00:00] isaacs: _ry: i've been meaning to write up a simple reduction that just uses v8 and shows the problem, but haven't gotten to it
[00:01] _ry: isaacs: it would be really cool to get a little example to put into deps/v8/test/cctest/test-api.cc
[00:06] _ry: okay i'm going to trash the old code now
[00:06] isaacs: _ry: trash the old code?
[00:06] maushu: http://pastebin.com/PSY4iYZ0
[00:06] _ry: evcom http_old
[00:07] isaacs: ah, cool
[00:10] _ry:  15 files changed, 2 insertions(+), 5672 deletions(-)
[00:10] _ry: :P~
[00:21] jedschmidt has joined the channel
[00:21] CIA-77: node: 03Ryan Dahl 07master * rc72967d 10/ (19 files in 5 dirs): Remove http_old tcp_old - http://bit.ly/cbG7Rp
[00:23] maushu: 5672 deletions(-)
[00:23] maushu: AHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHhhh-
[00:23] maushu: *huff* *huff*
[00:23] maushu: -hhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh!
[00:24] CIA-77: node: 03Ryan Dahl 07master * r3d6ebbd 10/ wscript : Don't look for GnuTLS in wscript - http://bit.ly/bacsXj
[00:24] inimino: :D
[00:27] maushu: Bye GnuTLS, we will NOT miss you.
[00:28] binary42 has joined the channel
[00:30] CIA-77: node: 03Ryan Dahl 07master * r9926dac 10/ src/node_net2.cc : Remove getaddrinfo from net2.cc - http://bit.ly/dn4sNh
[00:30] _ry: love cutting out stale code
[00:31] maushu: Tell me about it.
[00:31] _ry: what else can i trash?
[00:31] maushu: inimino
[00:33] inimino: maushu: yeah?
[00:34] maushu: Don't mind me. Just answering ry's question.
[00:34] maushu: Humans are so complicated. I can't find any useful API for them! Drat.
[00:36] inimino: hehe
[00:44] binary42 has joined the channel
[00:48] kriskowal has joined the channel
[00:54] kriskowal: _ry v0.1.92 does not build, mac os 10.6; similar log to the gist i posted yesterday
[00:54] kriskowal: i suspect an ssl config problem
[00:55] kriskowal: oh, looks more like a 64/32 bit mismatch
[00:56] kriskowal: with /opt/local/lib/libssl.dylib.  i'm not even sure what package manager is responsible for my /opt tree
[00:58] kriskowal: builds fine if i disable ssl in wscript
[01:00] _ry: kriskowal: hm. fuck.
[01:00] _ry: i forgot about that problem
[01:00] kriskowal: could be specific to my configuration
[01:00] _ry: that's mac ports?
[01:00] _ry: ssl?
[01:01] kriskowal: eh, probably
[01:01] kriskowal: i've used macports in the past
[01:02] _ry: hm
[01:02] _ry: not sure how to solve that issue
[01:02] _ry: should i, like, detect what arch libssl is?
[01:02] _ry: man. why does apple suck so bad.
[01:03] kriskowal: reality distortion field leaves a vacuum
[01:05] ssteinerX: kriskowal: if you've got a mixed macports/not-macports env, it could be, um, challenging
[01:05] kriskowal: might pose a question to the mailing list; detecting the architecture would probably be the way to go, i just don't know how to do that
[01:05] _ry: kriskowal: what if you uninstall ssl and reinstall it
[01:05] _ry: ? :)
[01:05] ssteinerX: apple should bless one package manager 
[01:06] kriskowal: yeh, i talked about that during my interview.
[01:06] ssteinerX: kriskowal: did you install ssl through macports?  If not, port installing it should get it ahead of the other one for building
[01:06] kriskowal: my interviewer, now a friend, replied with a question.
[01:06] ssteinerX: kriskowal: "why are you so confrontational?"
[01:06] ssteinerX: ;-)
[01:06] kriskowal: "what do you imagine some of the complications might be?"
[01:07] ssteinerX: kriskowal: that it would make it apple's responsibility?
[01:07] kriskowal: yeh, it's unfortunate.
[01:07] ssteinerX: that would be way more "complicated" than letting package management suck the life out of developing on OS X
[01:07] _ry: yes
[01:08] _ry: anyway - it's a terrible operating system
[01:08] mape: hehe
[01:08] mjr_: Is there a recommended way to install openssl for use with node?
[01:08] _ry: if it weren't for the beautiful suspend and wifi auto connection, no one would use it
[01:08] mjr_: homebrew?
[01:08] ssteinerX: _ry: but it sure is pretty...the Megan Fox of operating systems...
[01:08] _ry: mjr_: are you also having problems?
[01:09] mjr_: no, but I've never tried ssl.
[01:09] _ry: mjr_: i use macports
[01:09] mjr_: I've been meaning to, so I figured I'd see if I had the same problem.
[01:10] ssteinerX: I was able to build node about two weeks ago, with a macports setup -- is some specific .configure option causing all this agita?
[01:10] kriskowal: _ry eesh. my port command is broken. i guess i better get my house in order.
[01:11] kriskowal: and, yeah. /opt is managed by macports for me
[01:13] mjr_: wait, my OSX seems to come with openssl.  Is there a need to install it from a package?
[01:13] _ry: mjr_: no
[01:16] mjr_: hmm, well the tls test passes on my machine.  I guess it works then?
[01:16] ssteinerX: i just pulled from git, how come .configure's still looking for gnutls >= 2.5.0, i thought that just got ripped out
[01:16] maushu: gnutls... fighting to its last breath.
[01:16] mjr_: Speaking of last breath, I need to go home.
[01:23] bpot has joined the channel
[01:29] JimBastard has joined the channel
[01:29] JimBastard: hey hey 
[01:29] JimBastard: javascript party
[01:30] JimBastard: im hanging out here with tobey from http://github.com/tobeytailor/gordon
[01:30] JimBastard: hes still stranded in the us
[01:33] JimBastard: sup binary42 
[01:34] binary42: JimBastard: Heya.
[01:34] binary42: JimBastard: You know... I'm craving one of those spicy burgers.
[01:35] JimBastard: ahaha its dhaba tonight
[01:35] JimBastard: im getting hungry though if you are around you should come by
[01:35] mape: JimBastard: planes wonky?
[01:35] JimBastard: yeah
[01:35] mape: still?
[01:35] JimBastard: volanownd
[01:35] JimBastard: volcanowned
[01:35] JimBastard: volcanownd
[01:35] JimBastard: ^_^
[01:35] mape: I have multiple friends that were abroad that are back now
[01:35] binary42: JimBastard: Yum... that is tempting.
[01:35] ssteinerX: volpwnd
[01:36] binary42: JimBastard: I was debating the make something vs. go out to eat thing.
[01:36] binary42: JimBastard: You guys ordering out?
[01:36] binary42: 
ACTION was going to be in Dublin till that volcano decided to grab attention.

[01:37] mape: funconf?
[01:37] bacon has joined the channel
[01:37] binary42: Yeah.
[01:37] polotek has joined the channel
[01:38] binary42: mape: Rescheduling to the end of summer it seems. So I have hopes of getting there eventually.
[01:38] mape: mm
[01:40] polotek has left the channel
[01:41] inimino: hm...
[01:41] inimino: 
ACTION looks at docs

[01:41] inimino: ooh, pretty
[01:41] JimBastard has joined the channel
[01:42] _ry: needs anchors 
[01:42] _ry: i hope mape will redo it with markdown.js
[01:42] mape: I'll look at it
[01:43] inimino: _ry: is calling response.write() repeatedly supposed to be safe?
[01:43] polotek has joined the channel
[01:43] _ry: inimino: yes
[01:43] inimino: because it looks like bytes are being dropped
[01:43] _ry: mape: if i linke to a certain function, it gets covered up: http://nodejs.org/api.html#fs-truncatesync-97
[01:43] inimino: alright, I'll make a test case
[01:44] mape: _ry: yeah..
[01:44] MattJ: _ry: I already told him!
[01:44] MattJ: and sent a patch :)
[01:44] _ry: mape: :)
[01:44] _ry: yey open source
[01:44] mape: Easiest is probly to just drop the top header moving around
[01:44] inimino: _ry: is there a drain event?
[01:45] _ry: inimino: on net.Stream, yes
[01:45] MattJ: mape: Just remove that header entirely, would suit me fine
[01:45] inimino: that would be probably a quick fix, but I should find the bug anyway
[01:45] markwubben has joined the channel
[01:46] MattJ: mape: The TOC frame could also scroll if you enter a section that is out of view
[01:46] _ry: mape: http://github.com/ry/node/blob/9926dacd14c39276299712ced4a83fb043f27162/AUTHORS#L68
[01:46] _ry: is that your correct name?
[01:46] JimBastard: mmm cowboy kisses are inc
[01:47] mape: _ry: Yeah
[01:47] MattJ: I need my name there... *goes to find some trailing whitespace he can remove*
[01:47] mape: Why? Are you making fun of it?
[01:47] mape: ;)
[01:47] JimBastard: thats shrimp and jalepeon wrapped in bacon deep friend served with spicy mayo
[01:47] _ry: mape: you should enter it into git so it goes on your commits
[01:47] mape: I have?
[01:48] mape: oh mape
[01:48] Aria has joined the channel
[01:49] mape: There we go
[01:52] mape: _ry: Btw the site, should that be fixed as well so it matches the api doc page?
[01:53] _ry: mape: maybe
[01:53] _ry: i actually think opposite? :)
[01:54] _ry: i don't know
[01:54] mape: Hehe k, so.. Revert doc changes?
[01:54] _ry: 
ACTION sucks at website design

[01:54] _ry: i mean what would you change?
[01:54] _ry: font?
[01:55] siculars has joined the channel
[01:55] mape: Yes
[01:56] mape: not sure it is a big deal
[01:56] mape: Guess it is good for the project to have a nice page to make it more "official" and "serious"
[01:57] _ry: yeah
[01:57] mape: But since you hate logos I assume you don't put as much time thinking about that
[01:58] RayMorgan has joined the channel
[02:02] brainproxy: okay so i finally figured out the stupid callback parse thing; it probably should have been obvious to me; but I needed to take the chunk coming back from the redis client and do
[02:02] brainproxy: foo = '' + chunk
[02:02] brainproxy: and then parse poo
[02:02] brainproxy: foo
[02:02] brainproxy: :)
[02:03] brainproxy: anyway, you guys might be rolling your eyes, but I thought i would share
[02:03] brainproxy: (node rocks)
[02:04] _ry: brainproxy: chunk.toString()
[02:05] brainproxy: gracias
[02:06] brainproxy: hmm, that doesn't seem to work when chunk is json
[02:14] baconbits has joined the channel
[02:29] quirkey has joined the channel
[02:32] polotek: hey, has anyone checked out using v8-juice when writing node libraries?
[02:33] MattJ: I don't think it's open sauce
[02:33] polotek: New BSD
[02:34] MattJ: Ah, not http://www.v8juice.co.uk/ then :)
[02:35] polotek: http://code.google.com/p/v8-juice/
[02:35] polotek: sorry
[02:35] MattJ: Wow that's a stretched acronym :)
[02:36] polotek: yeah he REALLY wanted to get that play on words going
[02:36] MattJ: I personally haven't looked at the C++ V8 API
[02:36] MattJ: and I guess most libs are written in Javascript
[02:37] polotek: yeah, I'm realizing that's not obvious.  I'm talking about addons with c/c++ bindings
[02:38] polotek: libxmljs in particular
[02:39] polotek: I've spent the last week really trying to get the lay of the land in terms of the v8 community
[02:39] davidwilhelm has joined the channel
[02:39] polotek: there's a general consensus that the documentation blows
[02:40] polotek: and not a lot of info about people writing libraries with embedded v8
[02:42] davidwilhelm: I found it easy to install and try the demo, but got an 'error connecting to server' using the chatroom demo
[02:42] dnolen has joined the channel
[02:45] Aria: Yagh, I hate WAF
[02:46] davidwilhelm: signoff
[02:50] binary42 has joined the channel
[02:52] quirkey has joined the channel
[02:52] isaacs has joined the channel
[02:55] inimino: _ry: failing test case: http://gist.github.com/377412
[02:56] inimino: 
ACTION tries 0.1.92

[02:57] inimino: _ry: hm... it seems this is fixed since 0.1.90
[02:58] _ry: inimino: ?
[02:58] _ry: is it broke
[02:59] inimino: _ry: the test case that I just gisted was broken in 0.1.90 but now it seems to work
[02:59] inimino: so I guess you already fixed it :)
[02:59] _ry: Array(1025).join('x')
[02:59] _ry: ^-- this is cool
[03:00] inimino: oh, yeah, it's a cute trick
[03:00] _ry: kind of unnecessary though
[03:00] _ry: :)
[03:00] _ry: i mean, the big array object
[03:01] inimino: yes
[03:01] inimino: before that I had an array full of actual data, this made the test file much smaller ;)
[03:02] inimino: alright, time to test my actual server again...
[03:02] inimino: _ry: btw, did you know of any recently-fixed bugs that would cause that?
[03:05] _ry: inimino: my guess would be 684740c23200be67368fb1b225140e06c1fec295
[03:06] inimino: ah
[03:06] creationix: _ry: so there weren't any api changes in this release right?
[03:07] creationix: well, I guess the error thing is considered an api change
[03:07] creationix: it will break some code
[03:07] _ry: creationix: the error thing, yeah
[03:07] _ry: other than that, no
[03:07] creationix: cool, I've been testing all my stuff and everything seems to run fine
[03:08] ncb000gt has joined the channel
[03:08] ditesh|cassini has joined the channel
[03:09] _ry: creationix: there won't be much more API changes
[03:09] inimino: :)
[03:09] creationix: nice, stability is on the way
[03:09] _ry: just trying ot fix some bugs before 0.2
[03:09] polotek: +1
[03:09] _ry: the 'error' thing is going to be break code
[03:09] _ry: but that will be good
[03:10] creationix: alright, howtonode.org is running on v0.1.92 :)
[03:11] creationix: I'm trying to remember what has error events to catch, I guess streams
[03:11] inimino: I think I am going to upgrade from 0.1.25 to 0.1.92 tonight
[03:12] creationix: inimino: good luck, that a big jump
[03:12] inimino: if anybody wants to talk to possibly one of the oldest living node processes, now's your chance ;)
[03:12] inimino: creationix: yeah, I already did most of the work and testing
[03:12] creationix: inimino: you should check with felixge, he's got some old ones on transload.it
[03:13] creationix: how do you check the age of a process in linux anyway?
[03:14] inimino: ps
[03:14] inimino: ps u | grep node
[03:14] inimino: or something like that
[03:14] inimino: ps auxww | grep node
[03:23] JimBastard has joined the channel
[03:25] rictic has joined the channel
[03:27] PyroPete1 has joined the channel
[03:38] tekky has joined the channel
[03:39] Aria: Woot! html5 parser now round-trips through the parser for my test code.
[03:39] _ry: Aria: nice
[03:39] _ry: Aria: code
[03:40] _ry: ?
[03:40] Aria: http://github.com/aredridel/html5
[03:40] Aria: Depends on http://github.com/aredridel/env-js (which is patched to work as commonJS modules, at least for dom)
[03:41] hobson has left the channel
[03:42] kriskowal has joined the channel
[03:42] hobson has joined the channel
[03:42] polotek has left the channel
[03:44] admc has joined the channel
[04:03] steadicat has joined the channel
[04:06] admc has joined the channel
[04:07] lifo_ has joined the channel
[04:10] icey has joined the channel
[04:11] crodas has joined the channel
[04:20] crodas has joined the channel
[04:20] micheil: hey, _ry we no longer need the | sed ... \ line in make doc, as the id's are being added dynamically via javascript.
[04:24] Aria: Oh and v0.1.92 fails to build using --system without adding CARES to the list of libraries to compile against, _ry
[04:26] crodas has joined the channel
[04:32] Ned__: when using the http server, how do I determine if a client has disconnected ?
[04:32] Ned__: I figured there would be an event on the response or something to say EOF
[04:32] Ned__: ?
[04:36] creationix: _ry: where should I put nvm on the node wiki, technically it's not a module and works with any node version, it's written in bash
[04:36] creationix: maybe make a separate section for tools like that?
[04:37] Yuffster has joined the channel
[04:38] RayMorgan has joined the channel
[04:38] noonat: Oh man, the new docs are pretty :D
[04:39] Aria: Aren't they!
[04:39] noonat: Ned__: You should be able to listen for 'end' on the underlying request.connection object, I think
[04:39] JimBastard has joined the channel
[04:40] Ned__: noonat: okay, thanks :-)
[04:40] JimBastard: whats the general approach to adding timeouts to http client requests? i want to stop the request if it doesnt respond after 20 seconds or so
[04:43] JimBastard: set a timer outside the request and remove the timer
[04:43] noonat: JimBastard: You could probably use the built-in setTimeout on the stream, too
[04:44] JimBastard: is there any sample code floating around?
[04:47] Aria has joined the channel
[04:48] kriskowal has joined the channel
[04:49] kriskowal: _ry reinstalled macports and reinstalled openssl; all is well now
[04:51] silentrob has joined the channel
[04:56] noonat: JimBastard: http://gist.github.com/377476
[04:57] JimBastard: heh that looks way cleaner
[04:57] JimBastard: thanks ill try it out noonat 
[05:01] towski has joined the channel
[05:02] JimBastard: sup towski 
[05:02] JimBastard: :-D
[05:02] towski: you JimBastard
[05:02] towski: how's it going?
[05:03] JimBastard: working on some node reverse proxy stuff
[05:03] JimBastard: having a javascript party in nyc right now lol
[05:03] Aria: Woot! Several big web pages went through my HTML5 parser without damage!
[05:03] JimBastard: me, binary, the gordon guy, and indexzero
[05:03] Aria: (Testing with a proxy!)
[05:03] JimBastard: flash gordon owns
[05:04] towski: github link? :)
[05:05] JimBastard: ill do ya one better
[05:05] JimBastard: http://paulirish.com/work/gordon/demos/
[05:06] JimBastard: thats SWF => JS in realtime
[05:06] JimBastard: on the browser
[05:11] micheil: JimBastard: can it override <object> / <embed> tags?
[05:12] JimBastard: i mean i think that would be a job for your selection engine?
[05:12] JimBastard: read the wiki on github
[05:12] JimBastard: swf goes in canvas comes out or something
[05:12] JimBastard: maybe svg
[05:12] JimBastard: its magic
[05:12] JimBastard: vml support wouldnt be impossible
[05:16] micheil: he should talk to the guy behind Raphaël js
[05:18] mikeal1 has joined the channel
[05:23] JimBastard: dude
[05:23] JimBastard: they are friends
[05:23] JimBastard: we all hang out too at jsconf
[05:23] JimBastard: :-D
[05:23] JimBastard: javascript party
[05:38] silentrob has joined the channel
[05:40] noonat has joined the channel
[05:53] mikeal1 has joined the channel
[05:56] bolson has joined the channel
[05:57] JimBastard has joined the channel
[05:57] JimBastard: mikeal1 you alive
[05:57] JimBastard: mikeal
[05:58] mikeal1: yeah
[05:58] mikeal1: here
[05:58] JimBastard: do you think you could take a look at some proxy code i got going?
[05:58] JimBastard: http://gist.github.com/377497
[05:59] JimBastard: i dont seem to be catching the timeout correctly
[05:59] JimBastard: maybe its closing the request instead of timing out?
[06:00] mikeal1: if request.connection the client object?
[06:00] mikeal1: is it basically c?
[06:00] mikeal1: because i usually only get connection errors on the Client
[06:01] JimBastard: not sure
[06:01] JimBastard: im adding debug out on every event now
[06:01] mikeal1: i never do timeout stuff
[06:01] mikeal1: i just add a listener for "error" on the client
[06:03] JimBastard: hrmm
[06:03] JimBastard: maybe im making this too hard 
[06:03] JimBastard: the error event might cover it?
[06:04] JimBastard: ohh hell yeah that worked
[06:04] JimBastard: mad easy
[06:05] JimBastard: thanks mikeal1
[06:12] bolson has left the channel
[06:12] bolson has joined the channel
[06:29] _ry: JimBastard: the timeout might be broken
[06:29] JimBastard: its fine
[06:29] JimBastard: i got around it
[06:29] JimBastard: im working on some of the child process stuff now
[06:30] JimBastard: i need to sys.exec('node /path/to/custom.js') or so
[06:30] JimBastard: not sys.exec now
[06:31] kriskowal has joined the channel
[06:32] JimBastard: yeah thats working nicely
[06:32] JimBastard: <3 node
[06:32] JimBastard:     exec  = require('child_process').exec,
[06:32] JimBastard: now i gotta figure out some linux sandboxing
[06:38] bpot has joined the channel
[06:44] JimBastard: mmm break time
[06:45] binary42 has joined the channel
[06:50] cruxst has joined the channel
[06:50] binary42 has joined the channel
[07:00] sudoer has joined the channel
[07:01] isaacs has joined the channel
[07:03] kriskowal has joined the channel
[07:12] ditesh|cassini has joined the channel
[07:12] Nohryb has joined the channel
[07:19] JimBastard has joined the channel
[07:20] JimBastard: does anyone know any other cool node apps to play with besides like expressjs and fab? i guess wheat the blogging engine thing, but anything else?
[07:20] JimBastard: like an application someone might actually use for real stuff? that actually works?
[07:31] towski has joined the channel
[07:31] teemow has joined the channel
[07:37] CodeOfficer has joined the channel
[07:39] pdelgallego has joined the channel
[07:44] derbumi has joined the channel
[07:46] Elfix has joined the channel
[07:51] nsm has joined the channel
[08:08] jspiros has joined the channel
[08:10] pkrumins: anyone experiencing weird behavior with the latest node? particularly sting objects somehow get turned into buffer objects, and I can't call string functions on them anymore, like str.toLowerCase, str.substring, etc.
[08:10] jspiros has joined the channel
[08:10] pkrumins: s/sting/string/
[08:11] pkrumins: oh!
[08:11] pkrumins: it seems that 'data' callback now stores everything in buffer
[08:11] pkrumins: so i have to call toString to get it!
[08:13] rtomayko has joined the channel
[08:13] pkrumins: maybe that is true
[08:14] CIA-77: node: 03Jérémy Lal 07master * r610faf7 10/ wscript : Drop UDNS from node.uselib. Add CARES for --system. - http://bit.ly/cPziEC
[08:17] pkrumins: yep taht is true.
[08:17] pkrumins: got to call stream.setEncoding().
[08:18] pkrumins: or use toString
[08:21] ewdafa has joined the channel
[08:22] zackattack has joined the channel
[08:27] pkrumins: is it now absolutely required to setEncoding for streams?
[08:28] pkrumins: I have node itself croak when two stream get proxied through it.
[08:32] pkrumins: this fails for me now res.write(data, {'Content-Type': 'text/plain' });
[08:32] pkrumins: with
[08:32] pkrumins: the following traceback http://paste.pocoo.org/show/205546/
[08:33] pkrumins: however if I change that line to res.write(data), it succeeds
[08:33] isaacs has joined the channel
[08:34] kixxauth has joined the channel
[08:35] jamesduncan has joined the channel
[08:35] pkrumins: Can't set ecoding on it either
[08:35] pkrumins: TypeError: Object #<a ServerResponse> has no method 'setEncoding'
[08:38] pkrumins: figured it out.
[08:38] pkrumins: i had a bug in res.write(data, {'Content-Type': 'text/plain' });
[08:38] pkrumins: you don't send headers in write()
[08:39] noonat: writeHead :)
[08:39] pkrumins: yes sir
[08:39] ahc has joined the channel
[08:43] noonat: as for setEncoding: you can pass the encoding in as the second param to write
[08:45] qFox has joined the channel
[08:45] pkrumins: noonat, roger.
[08:49] javajunky has joined the channel
[08:50] piranha has joined the channel
[09:00] mcarter has joined the channel
[09:14] ditesh|cassini has joined the channel
[09:18] xla has joined the channel
[09:34] towski has joined the channel
[09:36] jamesduncan has joined the channel
[10:09] zackattack has joined the channel
[10:12] xla has joined the channel
[10:24] pjb3_ has joined the channel
[10:25] ec: ’sup everyone
[10:26] ec: I’ve been … away … for a while.
[10:26] ec: Seems I missed a lotl 0.1.*92*!?
[10:27] qFox has joined the channel
[10:28] nefD has joined the channel
[10:33] xla has joined the channel
[10:39] jamesduncan_ has joined the channel
[10:40] ditesh|cassini has joined the channel
[10:45] jamesduncan_ has joined the channel
[10:51] maritz has joined the channel
[11:00] ssteinerX has joined the channel
[11:14] Nohryb has joined the channel
[11:16] martoche has joined the channel
[11:16] botanicus has joined the channel
[11:19] jedschmidt has joined the channel
[11:20] hellp has joined the channel
[11:30] towski has joined the channel
[11:44] jamesduncan has joined the channel
[11:44] ditesh|cassini has joined the channel
[11:54] xla: heya
[11:55] xla: which lib in userland has a fast and solid implementation of a mixin function?
[12:01] Nohryb has joined the channel
[12:09] sveimac has joined the channel
[12:10] piranha has joined the channel
[12:14] technoweenie has joined the channel
[12:16] technoweenie has joined the channel
[12:29] herbySk has joined the channel
[12:37] botanicus has joined the channel
[12:37] tav has joined the channel
[12:41] teemow has joined the channel
[13:16] pagameba has joined the channel
[13:17] pagameba: hi, new to node, trying to install mongodb-native via kiwi (kiwi -v install mongodb-native) and getting invalid seed on 0.7.0 (many tar errors) - anything I can do about this?
[13:19] pagameba: sorry 0.7.1 is the version it is trying to install
[13:22] creationix has joined the channel
[13:23] botanicus has joined the channel
[13:32] pagameba: :(
[13:39] hsuh has joined the channel
[13:49] Nohryb has joined the channel
[14:04] creationix: xla: what do you need mixin to do?
[14:04] qFox has joined the channel
[14:05] creationix: pagameba: have you tried installing the driver manually
[14:06] pagameba: creationix: newbie, how would I do that?  clone from git? and then? :D
[14:06] creationix: I guess so, the mongo driver should say in the readme
[14:07] creationix: none of the package managers are complete yet, kiwi is one of the more matured ones, but nothing is official yet
[14:07] pagameba: looking at the readme ...
[14:08] creationix: Are you using this one? http://github.com/orlandov/node-mongodb
[14:08] maushu has joined the channel
[14:09] polotek has joined the channel
[14:10] creationix: hmm, I'm not sure which is the most stable, there is also http://github.com/christkv/node-mongodb-native.git
[14:11] creationix: xla: well, if you just need a shallow copy, then my proto library has a good mixin it adds to Object.prototype http://github.com/creationix/proto
[14:11] pagameba: the native one
[14:13] creationix: pagameba: I'm afraid I haven't used either, so I'm not too much help, I think this site uses mongo http://nodeblogs.com/
[14:14] creationix: the site is by Christian, so I assume he's using his own driver
[14:14] pagameba: I don't really need to use it but I'm following a tutorial on setting up a simple blog using node.js, express and mongodb :)
[14:14] polotek: anybody else think it's a little confusing that the conventin is "native" means pure-js and non-native means C addon?
[14:15] pagameba: what I'm really looking to do is set up a twitter oauth thing ... this was just a way of getting started with node.js
[14:15] creationix: polotek: I agree, I try to use "pure-js" when I mean there are no C parts
[14:16] creationix: pagameba: ahh, makes sense, If you do get ahold of tjholowaychuck, ley him know the kiwi build isn't working, he'll probably be interested to know
[14:17] pagameba: I posted a comment on the github issue list with the symptoms
[14:17] creationix: pagameba: great, are you on the mailing list?  There has been discussion of oauth stuff lately
[14:18] pagameba: creationix: yes, following the oauth discussions with a bit of interest :)
[14:18] pagameba: creationix: there is no 'install' docs, I assume this is something that people just 'know' how to do with nodejs?  How do I make a package like mongodb-native accessible in my app.js?
[14:19] creationix: well, if there are no C parts, then you just need to get the lib folder into your search path
[14:19] pagameba: is it just require('/path/to/git/checkout/location/')
[14:19] pagameba: ah - search path ... which is set where?
[14:20] creationix: the "Modules" section of the docs explains it http://nodejs.org/api.html
[14:20] polotek: creationix:  yeah that makes sense re: naming
[14:21] creationix: I usually just symlink the lib/* to my ~/.node_libraries/ folder
[14:21] creationix: pagameba:  ^
[14:22] polotek: pagameba: you can also set NODE_PATH in your environment and add paths there
[14:22] polotek: export NODE_PATH=/path/to/your/lib
[14:25] Nohryb has joined the channel
[14:26] botanicus has joined the channel
[14:26] gwoo has joined the channel
[14:47] xla: creationix: thx I thought about proto, but I couldn't remember the name
[14:48] creationix: xla: was a shallow copy what you wanted?
[14:48] xla: creationix: yay
[14:49] zackattack has joined the channel
[14:49] creationix: I'll take that as a yes :)
[14:52] pagameba: ok, I used polotek's suggestion to create NODE_PATH, the example is running now - thanks all
[14:52] alex-desktop has joined the channel
[14:54] polotek: pagameba:  cheers
[15:02] Nohryb has joined the channel
[15:05] Aria has joined the channel
[15:11] pagameba: ok, next question - does anyone know how to get a sass file to be included in a haml template rendered by the express engine?
[15:11] pagameba: at least I think that is what I want to know :D
[15:12] gwoo has joined the channel
[15:15] pagameba: I'm doing this tutorial http://howtonode.org/express-mongodb and the page rendering doesn't seem to use the sass style sheet
[15:15] pagameba: so i think something is missing
[15:19] maushu has joined the channel
[15:20] polotek: haml and sass aren't necessarily connected
[15:21] martoche has joined the channel
[15:21] polotek: ^-- pagameba 
[15:21] polotek: Once you set up your route for the css file just add the link tag to your head of your haml file
[15:21] pagameba: polotek: ok ... I see a route in my app for pulling in a css file from the sass file
[15:22] pagameba: polotek: how to add a link tag, that is the question :)
[15:23] polotek: Check out the docs for haml.  It's pretty simple.  Start with %link
[15:24] creationix: pagameba: I run the howtonode site, that particular howto isn't mine, but I'm thinking of writing a new one
[15:25] creationix: what is it you're looking for in a howto (besides examples that actually work as described)
[15:26] pagameba: creationix: just trying to get this simple blog howto working so I can see how all the pieces fit together for running a simple site
[15:26] pagameba: creationix: then onto the horribly complicated stuff :)
[15:26] creationix: as far as haml, visionmedia's version is almost exactly as the ruby version, my haml-js is slightlyl modified
[15:27] creationix: pagameba: one of the difficult but terribly interesting things about node is that traditional mvc based apps don't translate well to node
[15:27] creationix: it's much better suited to other yet-to-be-solidified patterns
[15:27] pagameba: polotek: I can't find anything on %link
[15:27] maushu: Anti-patterns.
[15:28] pagameba: creationix: :)
[15:28] creationix: besides, I wouldn't say that using a huge framework like express is a good place to start learning node
[15:28] creationix: it abstracts away a lot and gives you a much more rubyesque style of coding
[15:28] pagameba: creationix: ok
[15:29] polotek: creationix:  agreed
[15:29] creationix: that's like using rails to learn ruby
[15:29] pagameba: 
ACTION wonders where he should start?

[15:29] creationix: what is the sample site you want to build?
[15:29] pagameba: I did get the whole blog example working, including the mongodb thing so that is good
[15:29] creationix: I'm assuming http server with templates and some sort of persistence?
[15:30] pagameba: the only thing that didn't work was including the style sheet :)
[15:30] pagameba: creationix: I'm building a twitter application
[15:30] pagameba: the concept is simple, and I have all the pieces working in PHP
[15:30] creationix: so a twitter clone or a site that wraps the twitter web service
[15:31] polotek: pagameba:  the "hello, world" of framework learning projects :p
[15:31] polotek: I've built at least 5 on different platforms
[15:31] pagameba: creationix - it polls the direct messages of a particular user and looks for key words
[15:31] pagameba: creationix: and uses that to modify the profile pic of users that have authenticated the service to make those changes using oauth
[15:32] creationix: interesting
[15:32] pagameba: creationix: in its simplest state you might sign up, oauth authenticate the app then send a direct message about how you are feeling and it will update your profile pic with an emoticon
[15:33] polotek: creationix:  most of the people I talk to about node are interested in http. a good tutorial going over that would go a long way as a foundation I think
[15:33] pagameba: creationix: so its mostly a server-side thing
[15:33] pagameba: perhaps with high load, you can never tell how twitter-related things are going to take off :)
[15:34] creationix: I think I'll do one on writing a simple http site with templates, post data, and persistence
[15:34] creationix: using no libraries, just pure node
[15:34] creationix: pagameba: sounds like fun
[15:34] polotek: how would you do persistence?
[15:34] pagameba: but it needs a bit of a front end to fire off the oauth request to twitter and capture the resulting tokens into a db
[15:35] creationix: If you data is just a cache of the web service you're scraping, then a simple in-memory hash will do fine
[15:35] creationix: node-dirty is pretty simple if you want to save the hash to disk and query it
[15:35] pagameba: creationix: perfect!
[15:35] creationix: I have yet to create a node app that needs a database
[15:36] creationix: my sousaball game has persistent data, but I just store the levels as json documents in real files
[15:36] pagameba: creationix: a howto setup a site to oauth authenticate with twitter would be awesome - render a simple page with the authenticate url in an <a> tag and a handler for the callback with the user's token :)
[15:36] polotek: creationix: yeah that's what I figured. works fine while the app is running.
[15:36] creationix: it would be awesome, but I'd have to learn that myself first
[15:37] polotek: might be a pain on restart if the json files get large
[15:37] creationix: well, none of my sites store per-user data
[15:37] creationix: so that helps with scaleability
[15:39] polotek: also http://www.countdowntooauth.com/
[15:40] polotek: :)
[15:40] ncb000gt: pagameba: I just worked on some things with Tornado and have been considering moving that to node due to libs
[15:40] ncb000gt: So, if I do move it over I'll try to write something up like that
[15:40] pagameba: ncb000gt: cool
[15:41] sh1m has joined the channel
[15:42] Aria: Hm. Any idea what an ... exception, maybe? ... saying just "node.js:164:9" means?
[15:42] Aria: No full backtrace.
[15:45] creationix: Aria: looks like an exception, sucks that you don't get the trace though
[15:45] polotek: Aria:  what's the message that came with that? usually if all you get is a reference to node.js it's a problem in the C bindings somewhere
[15:45] polotek: usually bad arguments
[15:48] derbumi has joined the channel
[15:48] herbySk: 164:9 is the "process.loop(), so if an exception is generated inside loop, you may as well have no stacktrace...
[15:48] Aria: No message, that's the thing.
[15:49] Aria: Hm, alright. A problem closing file descriptors or something, perhaps, or overflowing a buffer (I /am/ using Buffer)
[15:49] Aria: And I might indeed be passing bad data to the Buffer methods.
[15:54] binary42 has joined the channel
[15:54] polotek: Aria: I don't have a lot of experience with buffer stuff, if you gist an example somebody smarter will probably help out
[15:55] Aria: Hehe, aright. If only this didn't invoke 5000 lines of code.
[15:55] Aria: I'm still looking for a tighter test case.
[15:58] Aria: Yagh. The downcasing of header names in the HTTP stuff bugs the heck out of me.
[15:58] ncb000gt: polotek: your streaming twitter stuff work with oauth?
[15:58] polotek: ncb000gt:  nope
[15:58] ncb000gt: ok
[15:59] polotek: I'm planning to update and release the lib this weekend, but I don't have much experience coding for oauth
[15:59] ncb000gt: hmm
[15:59] polotek: if someone can point to a nice implementation I can take a crack at porting
[15:59] ncb000gt: there is a node-oatuh lib somewhere
[15:59] brainproxy: haha, nice node is pushing ~40mbps right now, and the vm it's running in is hardly breaking a sweat
[16:00] Aria: Hehe. Gotta love evented code.
[16:00] ncb000gt: braiproxy: sweet :)
[16:00] brainproxy: of course redis is preforming on par with node
[16:00] brainproxy: :)
[16:00] ncb000gt: polotek: http://github.com/ciaranj/node-oauth
[16:01] polotek: brainproxy: qps for redis?
[16:01] brainproxy: qps?
[16:01] ncb000gt: queries
[16:01] polotek: queries per second.  just curious
[16:01] _ry: like i said, you'll saturate your bandwith
[16:01] brainproxy: hmm, i'm not sure, let me ballpark it
[16:01] brainproxy: is there a way I can get that stat
[16:01] brainproxy: from noe
[16:01] brainproxy: from redis i mean
[16:01] _ry: people should stop worrying about multicore :)
[16:01] brainproxy: with redis-cli
[16:02] mape: has node even been deployed on a project and it turned out it couldn't handle the preasure?
[16:02] brainproxy: this is not an answer but I just fired up this test a couple of min ago
[16:02] brainproxy: and redis says
[16:02] brainproxy: total commands processed
[16:02] brainproxy: 2522828
[16:02] _ry: if you've only got a 100mb connection to the internet, i think a single core will be enough - unless you're calculating primes
[16:03] polotek: damn, _ry did say that I remember giving him a good "pshaw!" when I read it
[16:03] _ry: mostly web apps just shuffle data around
[16:04] _ry: take something from over there - maybe eval a template - send it over there
[16:04] brainproxy: my very conservative estimate atm is 15000+ qps
[16:04] _ry: it's really sad that in the current state of things you need like 100 rails processes running at 200mb each to serve a decent website
[16:05] mape: brainproxy: That sounds high..
[16:05] ncb000gt: _ry: agreed
[16:05] mape: Hello world pushes like 6k on a random machine?
[16:05] _ry: i agree - it does sound high
[16:05] brainproxy: well, each request results in 4 queries to/from redis
[16:05] brainproxy: i have three clients blasting it at 1000 rps
[16:06] brainproxy: so 3000 * 4
[16:06] brainproxy: plus I have a handful of clients balsting at 100 rps
[16:06] brainproxy: w/ respect to node I mean
[16:06] _ry: 
ACTION away for the weekend. ciao

[16:07] brainproxy: have a good one ry
[16:08] brainproxy: alright, i added two more 1000 rps clients to the system
[16:08] brainproxy: 5000 * 4
[16:08] brainproxy: and I have others too
[16:09] brainproxy: the vm running the server has 4 cores, none going over 40% util atm
[16:09] ncb000gt: a single query would include the to and from...so is it 2 queries or actually 4?
[16:09] brainproxy: talking redis queries
[16:09] brainproxy: each request results in two to node from the server
[16:09] brainproxy: i mean two to redis
[16:09] brainproxy: and then a worker thread works the queue
[16:09] brainproxy: ever ms
[16:09] brainproxy: so it does a read/write
[16:09] ncb000gt: ok
[16:10] brainproxy: so 4
[16:10] ncb000gt: gotcha
[16:10] brainproxy: and then I have two clients catching all the output as well
[16:10] brainproxy: they're getting absolutely hammered of course
[16:10] polotek: someone should post this thread on twitter so I can retweet.  it's like nerdporn
[16:10] brainproxy: so all in all, it's pretty amazing
[16:10] ncb000gt: polotek: hah
[16:11] maushu has joined the channel
[16:11] polotek: this is the main reason web developers end up being sys admins, managing clusters and whatnot. it sucks
[16:11] stepheneb has joined the channel
[16:12] polotek: One machine with plenty of ram and a couple load balanced node processes could probably put a lot of other setups to shame
[16:12] brainproxy: i am going to work out my own logic so I can have multiple node processes playing catch up front
[16:12] brainproxy: and multiple workers on the backside of redis
[16:12] brainproxy: but it will still be simpler
[16:12] brainproxy: than a lot of the stuff that's done these days
[16:13] brainproxy: btw, i'm using jcoglan's faye library
[16:13] brainproxy: to channelize
[16:13] brainproxy: the http requests
[16:13] brainproxy: it's friggin awesome
[16:13] polotek: I've been poking around faye too. what do you mean by "channelize"
[16:13] brainproxy: faye + redis pubsub = a web heaven
[16:13] brainproxy: polotek: you can have clients name channels
[16:13] maushu: node cluster for the win.
[16:13] brainproxy: and the server can too
[16:14] brainproxy: the named channels are multiplexed on top of the xhr
[16:14] brainproxy: that's behind the scenes
[16:14] brainproxy: faye = the bayeux protocol
[16:14] brainproxy: implemented in node.js
[16:14] brainproxy: my hope now is to couple named channels to class instances
[16:15] brainproxy: and push clients classes via express and nodejuice
[16:15] polotek: I understand the tech but not understanding why you're channelizing
[16:15] polotek: sorry for being dense
[16:15] brainproxy: becase each channel can represent something different
[16:16] brainproxy: and then the client side and server side programming becomes easier
[16:16] brainproxy: well to me a tleast
[16:16] brainproxy: I can invoke methods for incoming messages
[16:16] brainproxy: based on channels
[16:17] brainproxy: maybe it's dumb, i dunno, but it's been fun to work with
[16:17] brainproxy: there might be better ways
[16:18] polotek: sounds cool but I'm not getting a mental model. good luck though
[16:19] brainproxy: ah, it crapped ot :(
[16:19] brainproxy: *out
[16:20] Yuffster has joined the channel
[16:21] polotek: question for the room about module design
[16:21] polotek: I have an object that allows the user to read a stream
[16:21] sveimac has joined the channel
[16:21] polotek: internally it reads the data events and processes them, then emits it's own events based on the outcome
[16:22] polotek: right now when you invoke the "stream" method it returns a custom eventemitter that you can attach listeners to
[16:23] polotek: another module I ran across has a similar approach, accept the module object itself is used as the event emitter and you attach listeners directly to it
[16:23] polotek: wondering what the pros and cons of each approach are
[16:24] ncb000gt: polotek: it sounds like the second is one less step and little less mem.
[16:24] Aria: The first approach lets you delay emitting events until handlers are attached.
[16:24] polotek: Aria: correct
[16:24] Aria: Since you can wait until the function that gives the handler to return.
[16:24] polotek: but you can wrangle the api on the second approach to support that
[16:25] Aria: Right -- just give this as the argument.
[16:25] Aria: I myself would do that.
[16:25] polotek: so 2 votes for #2?
[16:25] polotek: here's why I like #1
[16:25] polotek: essentially the module obj is a factory
[16:26] polotek: you can call stream multiple times and get different emitters
[16:26] Aria: And would you want separate emitters?
[16:26] polotek: rather than creating a new object multiple times
[16:26] ncb000gt: what would be the benefit of multiple emitters 
[16:26] ncb000gt: as opposed to a single emitter?
[16:27] polotek: stream takes options, so successive calls could have different options
[16:27] Aria: 
ACTION nods.

[16:27] polotek: but internally the same module obj could reuse resources
[16:27] Aria: That makes sense, then. That's a lot like the createServer method in node's net and http modules.
[16:27] polotek: like it only needs one http client
[16:27] polotek: and can spawn multiple requests
[16:28] polotek: I would think client/clientrequest objects are more expensive than the eventemitter stream objects
[16:28] polotek: but I could be bass ackwards on that
[16:28] Aria: Probably. Though quite likely not importantly so.
[16:28] polotek: yep, you're right. I didn't even realize that.
[16:29] polotek: well the reason I say that is because requests open connections
[16:29] martoche has joined the channel
[16:29] polotek: more to manage
[16:29] Aria: Yeah.
[16:30] polotek: but honestly that's a moot point, in my case I do need a request for each instance of stream
[16:30] polotek: because the options change the request
[16:30] polotek: could reuse the client object that gets returned from createClient though
[16:30] crodas has joined the channel
[16:34] JimBastard_ has joined the channel
[16:34] JimBastard_: zzzz
[16:34] JimBastard_: has anyone come up with a good module hot loading strategy that doesnt require compiling node?
[16:34] ncb000gt: Scala?
[16:34] ncb000gt: :)
[16:35] ncb000gt: i haven't seen anything
[16:36] noonat: JimBastard_: Are there good ones that do require compiling node?
[16:37] JimBastard_: there is one, but its outdated
[16:37] noonat: Whereabouts?
[16:37] JimBastard_: i actually dont need to use require() in my case
[16:37] JimBastard_: i just need to fileWatch a json file
[16:37] JimBastard_: that when it gets updated the application will know
[16:37] JimBastard_: so i guess i can just fs.read the file instead of doing a require
[16:37] silentrob has joined the channel
[16:38] noonat: Ah, yeah, that makes it simpler. You could just deep merge it into the existing object, then.
[16:38] JimBastard_: but yeah umm hot require
[16:38] JimBastard_: http://romeda.org/blog/2010/01/hot-code-loading-in-nodejs.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A%20LiminalExistence%20%28Liminal%20Existence%29
[16:39] JimBastard_: but i think thats built against an older version of node
[16:39] JimBastard_: _ry i think this has been talked about a lot. is there any plan for hot module reloading in the core?
[16:39] maushu has joined the channel
[16:39] brainproxy: ry said he's out for the weekend
[16:40] JimBastard_: got ya
[16:40] JimBastard_: does anyone know the state of the union for hot loading of node js modules?
[16:40] brainproxy: not sure; has anyone tried jcoglan's helium in the context of node.js
[16:41] brainproxy: would that even work
[16:41] polotek: I seem to remember some thread saying it was hard to do really robust file watching
[16:41] polotek: don't quote me
[16:41] mape: doesn't work just requiring it again?
[16:41] dnolen has joined the channel
[16:43] polotek: mape: are you saying require checks modification status before just returning cached version?
[16:43] polotek: docs don't talk much about require. probably should, it's fairly important
[16:43] mape: I wasent aware it cached them so guess dont
[16:44] polotek: yeah module caching is definitely part of the spec
[16:44] polotek: there might also be a require.cached(...) that will make sure it doesn't cache or something like that
[16:45] ncb000gt: http://github.com/blaine/node/blob/hotload/src/node.js#L840
[16:45] ncb000gt: shows as 850 for me
[16:45] ncb000gt: but the highlighting is weird
[16:45] polotek: no require.cache or cached
[16:45] ncb000gt: that checks the cache for the module
[16:46] polotek: don't know where I got that from
[16:46] ncb000gt: but it's not doing file io unless it has to
[16:46] brainproxy: i killed terminal.app!
[16:46] brainproxy: never done that before
[16:46] ncb000gt: lol
[16:46] mape: so what if you just add false && ?
[16:46] polotek: brainproxy:  win
[16:47] ncb000gt: brainproxy: time to install linux
[16:47] mape: it doesn't cache the modules?
[16:47] ncb000gt: 
ACTION ducks

[16:47] brainproxy: ncb000gt: it's becaus I'm monitoring all my traffic realtime through it, which is kind of dumb
[16:48] ncb000gt: brainproxy: don't worry, i'm just being an ass
[16:48] polotek: mape: it does, and there's now way to by pass the cache
[16:48] brainproxy: np :)
[16:48] polotek: I bet _ry would accept a patch though
[16:48] brainproxy: Chesterton wrote a famous poem about asses .. they too have their day
[16:49] brainproxy: anyway, I think I'm done with node for the weekend, ya'll have fun
[16:50] ncb000gt: later
[16:52] ncb000gt: also, like an idiot i was in the wrong repo when looking, http://github.com/ry/node/blob/master/lib/module.js#L212
[16:52] ncb000gt: this is the check
[16:52] ncb000gt: but, again, it's not doing any io to see if the file timestamp has changed
[16:53] ncb000gt: bbiab
[16:54] polotek: now I remember
[16:54] polotek: the problem is with the io call to check the file stats
[16:54] polotek: if the call is async then require becomes async which is no good
[16:55] polotek: if the call is sync then require becomes a bottleneck
[16:55] polotek: because the way node idioms are developing, people use require quite a lot
[16:55] noonat: imo hot-loading shouldn't be part of the main require() -- it should be opt in, something on top of require
[16:55] noonat: or a method of the module itself
[16:55] polotek: noonat:  agreed
[16:55] noonat: you don't want require() to start returning a different module halfway through your bootstrapping
[16:56] noonat: so, in that case, whatever code is doing hot-loading is fine to be async
[16:56] polotek: module.reload() ?
[16:56] noonat: you could emit a reloaded event as well, so other modules could watch for it and update their references if needed
[16:57] JimBastard_: back
[16:57] JimBastard_: so yeah
[16:57] JimBastard_: reloading modules in pure js is not possible?
[16:57] noonat: that's one of worst problems with hot-loading: if someone has var BarClass = require('bar').BarClass, even if you replace BarClass when you hot-load, their reference would still point at the old one
[16:57] JimBastard_: oofa
[16:58] javajunky has joined the channel
[16:58] JimBastard_: so you think my case filewatch some json and then just eval it?
[16:58] noonat: so you have to be intelligent about it on the consumer end
[16:58] JimBastard_: bypass the module system?
[16:58] noonat: JimBastard_: Yeah
[16:58] JimBastard_: so file watch fires when you do anything to the file right? even read it?
[16:58] JimBastard_: i think i noticed that last night at 4am or so while my shit looped
[16:58] noonat: Yeah, you have to check the timestamp if you want to watch for changes
[16:58] noonat: mtime or whatever
[16:59] mikeal1 has joined the channel
[16:59] JimBastard_: i think you get curr and prev
[16:59] twomashi has joined the channel
[16:59] noonat: Yep -- if (curr.mtime > prev.mtime)
[16:59] twomashi has left the channel
[17:00] deanlandolt: JimBastard_: yeah -- a watch fires on every read
[17:00] deanlandolt: you can compare last read to last write to detect if it's a write
[17:01] JimBastard_: of course
[17:01] JimBastard_: deanlandolt you dirty hippy
[17:01] deanlandolt: ha
[17:01] JimBastard_: i love how i thought everyone i'd meet from online would be prim and proper
[17:01] JimBastard_: little did i know they were mostly degenerates like me
[17:01] deanlandolt: certainly not in the javascript world...and i love it
[17:01] JimBastard_: ahaha
[17:02] towski has joined the channel
[17:02] maushu: JimBastard_, sorry, you're wrong.
[17:02] JimBastard_: i thought RayMorgan was some 50 year old man and deanlandolt was like like 28 year old suit and tie guy
[17:02] maushu: We are worse than you.
[17:02] JimBastard_: hee hee
[17:02] JimBastard_: if you insist
[17:02] JimBastard_: me, im a 400lb african american guy who is really into naruto
[17:03] RayMorgan has joined the channel
[17:03] Aria: Those are surprisingly common.
[17:03] maushu: ^
[17:03] Aria: Well, maybe not 400lb. But y'know.
[17:03] polotek: JimBastard:  ditto - about 200lb
[17:03] JimBastard_: sup RayMorgan i was just talking about you
[17:03] JimBastard_: polotek: :p
[17:04] derbumi has joined the channel
[17:04] RayMorgan: JimBastard_: were you now?
[17:04] JimBastard_: lol yeah
[17:04] JimBastard_: i thought RayMorgan was some 50 year old man and deanlandolt was like like 28 year old suit and tie guy
[17:04] JimBastard_: ^^^
[17:04] JimBastard_: ahaha
[17:04] RayMorgan: lol
[17:04] JimBastard_: hows Vegas?
[17:04] RayMorgan: IRC fail
[17:04] RayMorgan: I am actually in SF right now for jquery conf
[17:04] JimBastard_: ohh nice
[17:05] JimBastard_: i forgot that is this weekend
[17:05] RayMorgan: yeah
[17:05] JimBastard_: they talking about jquery.template at all?
[17:05] RayMorgan: a little larger than jsconf was ;)
[17:05] JimBastard_: ahaha
[17:05] RayMorgan: I think there is a talk on it later
[17:05] mape: they finalized anything?
[17:05] JimBastard_: word, im looking forward to see what JQ is going with templating and databinding
[17:05] mape: Didn't they go of the deep on on that?
[17:06] JimBastard_: its time for JS to kick it up with virtualized components
[17:06] JimBastard_: so we can have million record list boxes
[17:06] binary42 has joined the channel
[17:06] mape: {$name} :/
[17:07] noonat: mape: WTF? Seriously?
[17:07] mape: last I read
[17:08] mape: Didn't seem like a lot of in template logic either
[17:08] noonat: Guhh... what do they think this is, PHP? :(
[17:08] CodeOfficer has joined the channel
[17:08] mape: They seem to think they are smarty
[17:09] CodeOfficer has joined the channel
[17:09] webben: noonat: they like their dollar signs don't they? ;)
[17:09] noonat: Anything using Smarty as a design guideline is doomed to fail
[17:09] JimBastard_: mustache is for templates
[17:09] herbySk: hey you who are not on *bsd, could you please ru na little code for me?
[17:10] alxc1 has joined the channel
[17:10] JimBastard_: sup herbySk ?
[17:10] herbySk: I'm interested in running
[17:10] herbySk: ./node test/simple/test-http-304.js >out
[17:10] herbySk: and see the contents of the 'out' file
[17:10] polotek: am I the only one who doesn't like mustache?
[17:11] JimBastard_: polotek: do you not want a mustache ride?
[17:11] JimBastard_: :-D
[17:11] JimBastard_: i mean
[17:11] JimBastard_: :-{{)
[17:11] noonat: I think it's scientifically impossible for mustaches not to be cool.
[17:11] JimBastard_: >:-{{)
[17:11] JimBastard_: mustache crockford is not amused
[17:11] noonat: ^ case in point
[17:12] JimBastard_: mustache is simple enough
[17:12] JimBastard_: similiar to ctemplate
[17:12] polotek: http://pixelspread.com/images/beards_trust_2.jpg
[17:13] polotek: I know the name of my new module that will stand in stark contrast to mustache
[17:13] herbySk: JimBastard_: what's sup?
[17:13] JimBastard_: lol polotek 
[17:13] noonat: polotek: hahaha
[17:13] polotek: seriously though, I just don't like the take on control flow.
[17:13] polotek: I like my control structures to be visible
[17:14] polotek: not dependent on the data that gets passed in
[17:14] polotek: other than that I think it's fine
[17:14] polotek: and I can't escape the fact that Mu is screaming fast on node
[17:15] mape: Muuu?
[17:15] polotek: mustache for js
[17:16] RayMorgan has joined the channel
[17:16] piranha has joined the channel
[17:18] herbySk: pls, could someone ru nthat line in your node? I'd lie to see the output...
[17:18] herbySk: ./node test/simple/test-http-304.js >out
[17:18] admc has joined the channel
[17:18] herbySk: For me, the contents of 'out' file is not the same as when I run it without output redirection
[17:18] JimBastard_: okay i got you herbySk one sec
[17:20] JimBastard_: herbySk:  i dont have that test
[17:20] softdrink has joined the channel
[17:20] herbySk: it's in 0.1.92... hm. pity :-/
[17:21] noonat: herbySk: it's the same for me on osx
[17:21] noonat: http://gist.github.com/377790
[17:21] JimBastard_: yeah sorry
[17:21] JimBastard_: not really in the mood to update today
[17:21] JimBastard_: not that i dont like punishing myself
[17:21] herbySk: noonat: thanx
[17:21] inimino: updating is the node way
[17:21] JimBastard_: i just have 10 modules or so 
[17:21] JimBastard_: yeah
[17:21] zackattack has joined the channel
[17:21] JimBastard_: my submodules dont appreciate it
[17:21] inimino: if your code isn't broken, you're not keeping up
[17:21] JimBastard_: i usually wait a week or so
[17:22] JimBastard_: hows it going inimino long time to chat
[17:22] inimino: good, good
[17:22] polotek: JimBastard:  http://github.com/visionmedia/nvm
[17:22] JimBastard_: i feel like i havent had any problems that meritted your expert advice lately
[17:22] JimBastard_: what about NVM polotek ?
[17:22] polotek: easy to manage multiple versions of node
[17:23] JimBastard_: yeah i mean i guess so
[17:23] JimBastard_: im not trying to do that
[17:23] inimino: hehe
[17:23] JimBastard_: for me its , pull , make , test
[17:23] JimBastard_: if it falls im rolling back
[17:23] JimBastard_: fails*
[17:23] polotek: 
ACTION has to run

[17:23] JimBastard_: i could see that being useful though
[17:23] polotek: later guys
[17:23] JimBastard_: peace man
[17:23] JimBastard_: talk to you later
[17:23] inimino: I run tests against the latest, then install it when everything works
[17:23] polotek has left the channel
[17:24] JimBastard_: i meant test as in run my apps
[17:24] herbySk: for me in freebsd the stdout fd gets silently closed :-/ either throws EPIPE or SIGPIPE silently kills node, so I only have the first line
[17:24] JimBastard_: not run the node tests
[17:24] inimino: JimBastard_: yeah, me too
[17:24] piranha_ has joined the channel
[17:24] JimBastard_: herbySk: is it like that for all tests or just that one?
[17:24] mikeal1: JimBastard: you get that proxy thing worked out
[17:24] mikeal1: i fell asleep :)
[17:24] JimBastard_: yeah man
[17:24] JimBastard_: it was actually really really easy
[17:24] herbySk: for all test that use child processes... after child process ends, the main process' stdout is closed in the background
[17:24] mikeal1: awesome
[17:24] JimBastard_: once i dove into your code mikeal1 
[17:25] JimBastard_: proxying in node is like nothing
[17:25] mikeal1: totally
[17:25] JimBastard_: its built for shit like that
[17:25] herbySk: http-204 is just one of them
[17:25] mikeal1: i want to keep that around as a really simple example
[17:25] JimBastard_: ive got a cool reverse proxy working
[17:25] maushu: It also makes great coffee.
[17:25] herbySk: s/204/304/
[17:25] JimBastard_: im setting up a fileWatched routing table for the reverse proxy
[17:25] mikeal1: i was starting to add feature to it and then I decided to rip them out and keep it simple
[17:25] JimBastard_: so it doesnt use a DB
[17:25] JimBastard_: yeah mean
[17:25] JimBastard_: well the other thing
[17:25] mikeal1: that code is kind of like a template i use whenever i need new proxy code
[17:26] JimBastard_: is that yours is setup to be a true proxy server
[17:26] JimBastard_: with a PAC file
[17:26] JimBastard_: i guess
[17:26] JimBastard_: i needed more of a http proxy
[17:26] JimBastard_: but its great boiler plate code
[17:26] JimBastard_: err, http proxy as in using a browser and typing in a url and getting proxy passed
[17:26] mikeal1: yeah, that's what it's for
[17:26] JimBastard_: opposed to pointing the browser to a proxy server
[17:26] mikeal1: yeah, totally
[17:26] JimBastard_: :-)
[17:26] herbySk: JimBastard_: for all test that use child processes... after child process ends, the main process' stdout is closed in the background http-304 is just one of them
[17:26] mikeal1: hrm....
[17:26] mikeal1: i could make it better tho
[17:27] mikeal1: if the url isn't a full url I could use the host header
[17:27] mikeal1: that would make things simpler
[17:27] JimBastard_: im doing something like that
[17:27] mikeal1: i have code to do that in all the other proxies i have
[17:27] Aria: Hehe. And I'm writing a forward proxy.
[17:27] mikeal1: i have a caching reverse proxy and a load balancer rever proxy that both do that
[17:27] JimBastard_: so right now im doing proxy pass based on unique URL, it would be awesome to do it on subdomain
[17:28] JimBastard_: maybe i could point *.mydomain.com and then use node to forward to right site?
[17:28] JimBastard_: will that work?
[17:28] mfeiri has joined the channel
[17:28] mikeal1: totally
[17:28] mikeal1: that's what balance is designed to do
[17:28] JimBastard_: ?
[17:28] JimBastard_: node-balance?
[17:28] JimBastard_: i have node-balancer in my private git repo
[17:28] JimBastard_: (my own roll)
[17:28] mikeal1: http://github.com/mikeal/balance
[17:29] mikeal1: http://github.com/mikeal/balance/blob/master/test/run.js
[17:29] JimBastard_: yeah
[17:29] mikeal1: you create a proxy and then you get a "route" event which you use to tell it which port/host to route the request to
[17:29] JimBastard_: word
[17:30] mikeal1: there is a ton of code in there to work around a couple bugs in node
[17:30] JimBastard_: my approach is a bit diffirent
[17:30] JimBastard_: its a bit simplier imho, but im probaly misunderstanding something or i missed a major edge case
[17:30] JimBastard_: i just do a simple lookup when the request comes in based on a JSON table
[17:30] JimBastard_: there any route object or anything
[17:31] mikeal1: if you wanted a file to be the routing table you would still need the file watcher and that would just update the routing table for your route event handler
[17:31] JimBastard_: yeah im doing that, just no routing system
[17:31] JimBastard_: its literally one line
[17:31] mikeal1: i needed mine to use routes defined in couchdb documents
[17:31] mikeal1: and update whenever the db updated
[17:31] JimBastard_: of course
[17:31] mikeal1: so i have a changes feed listener to couchdb that updates the table
[17:31] JimBastard_: ill probaly do the same
[17:31] JimBastard_: you are couch watching instead of file watching
[17:32] maushu has joined the channel
[17:32] mikeal1: you'll want to handle this
[17:32] mikeal1: http://github.com/mikeal/balance/blob/master/lib/balance.js#L84
[17:32] JimBastard_: what is method HEAD ?
[17:32] mikeal1: there is a bug in node's handling of HTTP HEAD requests
[17:33] mikeal1: a HEAD request returns only the headers for what would be a GET request
[17:33] mikeal1: it's logged, it'll be fixed eventually i'm sure
[17:33] mikeal1: there is also a bug, which i don't have a workaround for
[17:33] mikeal1: really large chunked responses won't emit an end event
[17:33] JimBastard_: damn
[17:34] JimBastard_: well i plan on running a shit ton of requests through this proxy
[17:34] mikeal1: http://github.com/ry/node/issues#issue/89
[17:34] mikeal1: http://github.com/ry/node/issues#issue/77
[17:34] mikeal1: there are tests checked in for both of them
[17:35] alxc1: hi people, just came across nodejs, installing it on freebsd now & look forward to playing with it! was just wondering how long it's been about?
[17:36] derbumi has joined the channel
[17:37] mikeal1: when was JSConfEU last year?
[17:37] ashb: 18months or so?
[17:37] mikeal1: September?
[17:37] ashb: november i thought?
[17:37] mikeal1: that's when it was kind of like "announced"
[17:37] mikeal1: but people knew about it before that
[17:38] JimBastard_: sup alxc1 welcome
[17:38] JimBastard_: node is amazing
[17:38] ashb: so long as you don't mind things changing underneath you ever once in a while :)
[17:38] JimBastard_: the learning curve to get going is a slight incline, but its not a very big hill
[17:39] JimBastard_: aye
[17:39] JimBastard_: its still early adopter status, but it gets better everyday. its dangerously close to being production ready
[17:39] JimBastard_: i mean, a bunch of people use it in prod already
[17:39] alxc1: ok well i was looking to experiment plugging v8 into some things... but your existing classes look like they contain a lot of things i'd like that i'd waste time building myself :)
[17:40] JimBastard_: make sure to read the nodejs.org docs
[17:40] JimBastard_: then howtonode.org
[17:41] alxc1: ok cool :) have been flicking through the docs while i wait for a load of my ports to upgrade...
[17:41] JimBastard_: personally, im only doing pure js modifications to node, i.e. CommonJS modules
[17:41] JimBastard_: so i dont mess around with the C code
[17:42] JimBastard_: i havent even tried to custom compile node outside of how it ships
[17:42] JimBastard_: but we are working on a new project that will require some modifications to the httpServer
[17:43] alxc1: ok well i'm not going to pretend I'm going to able to help you guys develop your code right now but it's done building so i'm gonna go have a play around ;) 
[17:43] ncb000gt: I haven't modified the node c/c++ libs but there are a handful of c/c++ modules floating around
[17:44] ncb000gt: it's pretty straight forward
[17:46] JimBastard_: heh i wasnt asking for help lol
[17:55] mape: JimBastard_: Wait a sec, I'll help you
[17:56] JimBastard_: lol wut
[17:56] mape: Just stay steady and I'll fix it
[17:56] JimBastard_: ?
[17:56] mape: There, no worrys :)
[17:57] JimBastard_: so i have a question actually
[17:58] JimBastard_: if inside my module i have a global var, and i have a fileWatcher in play
[17:58] JimBastard_: if i modify the global var on fileWatcher update
[17:58] JimBastard_: will further incoming requests have access to the new value in the global?
[17:58] ashb: i think so, yes, unless you (manage to) reload the module
[17:59] JimBastard_: time to test it out
[18:07] xla: heya, is there a way to close an open http.ClientResponse?
[18:07] mikeal1: JimBastard: that will all work as long as the consumer of the table doesn't accidentally cache the value
[18:07] mikeal1: a better way to do it would be to only expose a lookup function and keep the table private to the module
[18:07] aho has joined the channel
[18:07] JimBastard_: i mean, its almost private to the module
[18:07] JimBastard_: always*
[18:08] JimBastard_: if you do var foo; at the top of the module, its not being exported
[18:08] JimBastard_: sorry im bad with terminology
[18:08] mikeal1: right, you're setting the routing table to an attribute on exports right?
[18:08] JimBastard_: naaah
[18:08] JimBastard_: its all inside the same module
[18:08] mikeal1: oh ok
[18:08] mikeal1: nevermind then
[18:08] mikeal1: that'll all be fine
[18:09] JimBastard_: what time zone are you in mikeal1
[18:09] mikeal1: PDT
[18:09] mikeal1: i'm in Oakland, CA
[18:09] JimBastard_: got ya
[18:09] JimBastard_: cool stuff
[18:09] JimBastard_: you do JS work out there?
[18:10] mikeal1: i work for couch.io
[18:10] mikeal1: and i worked at Mozilla before that
[18:10] mikeal1: everything i do now is js, when i was at Mozilla it was js and Python
[18:11] JimBastard_: ohh shit
[18:11] JimBastard_: were you at jsconf?
[18:11] mikeal1: nope
[18:11] mikeal1: missed the tickets :(
[18:11] JimBastard_: i defintely was drinking with jan____ a little
[18:11] mikeal1: but jan is helping put together JSConfEU so I'm having him keep a ticket for me
[18:11] JimBastard_: he said i wasnt a pirate since i left on sunday afternoon instead of raging more
[18:11] JimBastard_: got ya
[18:11] mikeal1: hahahaha
[18:11] JimBastard_: thats good stuff
[18:12] noonat: Arrggg, why are you swallowing my exceptions, child_process.exec? :(
[18:13] noonat: When I add a sys.puts() before the function that calls it, the exception gets properly thrown too @_@
[18:15] jedschmidt has joined the channel
[18:16] ncb000gt: anyone having issues with the latest node wherein you get "Error: ECONNREFUSED, Connection refused" when trying to start an http server with createServer?
[18:16] ncb000gt: latest being HEAD
[18:17] martoche_ has joined the channel
[18:18] martoche_ has joined the channel
[18:19] towski has joined the channel
[18:23] martoche has joined the channel
[18:25] nsm has joined the channel
[18:26] noonat: yearg... race condition i think
[18:26] martoche has joined the channel
[18:28] martoche_ has joined the channel
[18:29] kriskowal has joined the channel
[18:31] jamesduncan has joined the channel
[18:31] jbrantly has left the channel
[18:32] martoche has joined the channel
[18:34] martoche has joined the channel
[18:36] martoche has joined the channel
[18:38] jedschmidt: wow, the new style for the node docs is great.
[18:38] mape: fabolous
[18:40] martoche has joined the channel
[18:41] martoche has joined the channel
[18:45] martoche has joined the channel
[18:47] MattJ: /almost/ great :)
[18:51] hsuh has joined the channel
[18:52] Nohryb has joined the channel
[18:52] ncb000gt has joined the channel
[18:53] RayMorgan has joined the channel
[18:54] mape: Anyone here tried node juice?
[18:55] ncb000gt: mape: not i
[18:55] mape: Just got it working, kinda neat
[18:56] ncb000gt: cool
[18:56] ncb000gt: what is it? haven't heard of it
[18:56] mape: Allows you to develop frontend stuff and you pipe it through node juice so it auto updates the page
[18:56] mape: CSS/js and whatnot when files change
[18:59] noonat: wow, that is awesome.
[19:00] tlrobinson_ has joined the channel
[19:00] mape: If you look at http://mape.me:8080/index.html
[19:03] hsuh: mape: why it keeps reloading everything?
[19:03] mape: hsuh: Because I'm saving the html file
[19:03] noonat: hsuh: that's the whole point :)
[19:03] maushu has joined the channel
[19:03] ncb000gt: interesting
[19:05] mape: Give that to clients
[19:05] rictic has joined the channel
[19:05] mape: let them keep a page up while the site is being developed ;D
[19:06] noonat: eff clients, i have wasted way too much of my life hitting cmd-r :P
[19:06] mape: Hehe
[19:06] mape: Yeah and you are suppose to be able to hook it into apache as well if you are doing non node stuff
[19:07] piranha has joined the channel
[19:07] mape: looks better in ff
[19:07] mape: chrome blinks
[19:07] mape: ff just updates
[19:07] quirkey has joined the channel
[19:08] Nohryb has joined the channel
[19:09] hsuh: never seen a project gain so much momentum so fast
[19:10] ncb000gt: hsuh: node has def. caught on fast
[19:10] ncb000gt: people like doing work in JS, there just need to be more libraries
[19:11] ncb000gt: but, that'll come with time
[19:11] hsuh: so.. as na outsider i think "v8 was written for a browser, how can it also be appropriate for a server?"
[19:11] ncb000gt: hsuh: v8 is purely a JS parser
[19:11] malkomalko has joined the channel
[19:12] ncb000gt: it isn't dependent on the browser, the Chrome source adds browser specifics to the mix which is why you can write JS that work with chrome
[19:12] mape: The one thing I see missing is the rails for node
[19:12] ncb000gt: rails/django/something
[19:12] ncb000gt: there isn't a standard web dev framework yet
[19:13] hsuh: yes, but i mean... it was probably designed to work well with the browser, i suppose
[19:13] jedschmidt: mape: express seems pretty close to fulfilling the rails role.
[19:13] ncb000gt: hsuh: no, it was designed to work well with JS. :)
[19:13] mape: The only issue I've heard is the gc
[19:13] ncb000gt: ragardless of it's context
[19:14] mape: jedschmidt: Still havent seen someone build a site with it yet?
[19:14] mape: Seems most things are "this is a web framework, here is a function that takes a string and outputs it at a path"
[19:14] mape: Not really doing the generic stuff most web projects do
[19:14] ncb000gt: lol
[19:14] admc has joined the channel
[19:15] ncb000gt: I had been working on something for that, but there are a lot of moving parts to a web framework
[19:15] mape: And I'm sure express does more then that :P Just yeah.. Noone seems to use them at least then
[19:15] ncb000gt: express looked close to filling that sort of role, but i haven't played with it enough
[19:15] jedschmidt: well, now that the API is less of a moving target, that should change.
[19:15] mape: Or a good example to show people how to structure a large project
[19:15] mape: Most people I talked to just say "meh it will end up a large pile of messy code"
[19:16] jedschmidt: mape: i think the (fab) approach should help there.
[19:16] mape: Would be nice to have something to show 
[19:16] hsuh: mape: there's a djangode on your site :)
[19:16] jedschmidt: mape: you can define sites as modularly as you like, and plug them into each other.
[19:16] bpot has joined the channel
[19:16] mape: jedschmidt: Yeah I tried looking at that but I'm to thick to figure it out and the examples were to basic to really add any meat
[19:17] mape: At least from what I understood
[19:17] mape: Waiting for that video pres on jsconf to be released for fab :)
[19:17] mape: jedschmidt: Heard from anyone doing anything fun with it?
[19:17] jedschmidt: mape: did you check out the annotated preso slides online? that seems to help a lot of folks.
[19:17] jedschmidt: mape: naw, it's a bit early for that.
[19:18] jedschmidt: i'm really dreading the video. i think i talked wayyyy to fast.
[19:18] mape: Yeah took at look at them, can't say it helped me though. I'm not very bright though so might just be me
[19:19] ncb000gt: jedschmidt: from what i saw (feedback wise) it went well
[19:19] mape: I'm just wondering how the modules thingy will work out
[19:19] mape: Seems like people will reinvent the wheel not wanting to depend on others
[19:19] jedschmidt: mape: ha ha. well, examples help a lot. i'm probably going to end up writing the (fab) site itself in (fab), so that should serve as a good example.
[19:19] mape: jedschmidt: Add some dogfood yeah :)
[19:22] mape: But so yeah, how will people manage to depend on eachothers modules? And if they don't won't that just lead to a lot of redundant modules?
[19:23] mape: Unless it goes into core, which it probly shouldn't it seems like there needs to be a larger "player" that gathers modules into somekinda soup to nuts module for easy web develoment
[19:23] jedschmidt: mape: that's a tough nut to crack.
[19:23] mape: That is kept up to date and whatnot
[19:23] ncb000gt: mape: if you look at the python community, there are tons of modules/libraries, the good ones win out
[19:24] ncb000gt: but they have tools to help manage those libs, easy_setup, pip, virtualenv
[19:24] ncb000gt: using those tools is essential to making sure you're working with the libs you want to
[19:24] polo7 has joined the channel
[19:24] ncb000gt: but also providing a way to try things out
[19:24] polo7: Hi!
[19:24] ncb000gt: hey polo7
[19:24] polo7: New here :)
[19:25] ncb000gt: welcome :)
[19:25] towski has joined the channel
[19:25] polo7: Thank you
[19:25] mape: Owell, I guess we'll have to wait for a too cool for school module to show up and sweep everyones feet
[19:25] ncb000gt: mape: the python guys use pypi for the "base" but tools like pip allow you to specify tarballs, git repos, etc
[19:25] ncb000gt: haha
[19:26] mape: Or you could copypaste php.js and call it a day ;)
[19:26] ncb000gt: mape: i do think it's essential for the community to come up with something around it, and it looks like people are trying
[19:26] ncb000gt: lol
[19:26] ncb000gt: npm, nodules, kiwi are all examples of attempts at this
[19:27] mape: Yeah, I guess the slowing down of API changes will help
[19:27] ncb000gt: maybe i'll work on a virtualenv for node...
[19:27] ncb000gt: as a way to break things out
[19:27] ncb000gt: it will a lot
[19:29] mape: jedschmidt: You have any idea on the ETA of those jsconf talks?
[19:29] ncb000gt: mape: http://bitbucket.org/ianb/virtualenv is the source repo for the python version
[19:30] ncb000gt: that'd be a better question for voodootikigod_
[19:30] jedschmidt: mape: i think voodootikigod_ said a month or more.
[19:30] mape: Figured he was out rapeing and pilaging
[19:30] mape: Ok
[19:30] ncb000gt: haha
[19:30] ncb000gt: he typically is
[19:31] jedschmidt: mape: he's got a move to finish and a kid on the way, so he's plenty busy.
[19:31] ncb000gt: Last year he released a video per month
[19:31] ncb000gt: for a while
[19:31] mape: Ok, just eager to listen to the talks :)
[19:31] mape: Weren't there loads of videos?
[19:32] mape: Seemed to be quite a bit of people talking this year at least
[19:35] derbumi has joined the channel
[19:35] ncb000gt: yea, there were track A and B talks this year, last year track B seemed somewhat like an afterthought
[19:36] ncb000gt: most likely because people still wanted to talk, but had like 10-15 minutes of a presentation
[19:36] ncb000gt: or just some idea and wanted to get some chatter going
[19:36] mape: Yeah
[19:36] ncb000gt: track A had the main presentations
[19:36] mape: those are usually the neat ones
[19:36] mape: small, quick and interesting
[19:36] ncb000gt: heh, that's the thing about jsconf, they are all interesting
[19:37] ncb000gt: I spoke at the first one, all of track A was quite awesome, and track B had some interesting stuff going 
[19:37] ncb000gt: Plus, this year, MW2?
[19:38] mape: modern warfare? :P
[19:38] ncb000gt: yep, they had it on site
[19:38] ncb000gt: heh
[19:39] mape: oh, sit down, have a beer, play kinda thing?
[19:39] ncb000gt: yea
[19:39] ncb000gt: take a break from info. overload
[19:39] mape: The interactions are king :)
[19:39] ncb000gt: :)
[19:39] hsuh: is nodewiki running anywhere?
[19:39] ncb000gt: they've done a great job with it for sure
[19:40] ncb000gt: hsuh: not sure about a module, but there is the wiki on github if that's what you're looking for.
[19:40] mape: http://github.com/gjritter/nodewiki ?
[19:40] hsuh: yep
[19:45] alxc1: hmm ok, can anyone tell me, in reference to the v8 docs, does one running nodejs script operate in a v8 'context', with all the scripts started using '#node script.js' running in different contexts in one instance of the v8 virtual machine?
[19:45] mikeal1 has joined the channel
[19:45] magcius has joined the channel
[19:45] magcius: _ry: you around?
[19:45] mape: Hehe find it fun how nodejs is defacto js on the servern nowadays
[19:46] ncb000gt: _ry: is out
[19:46] magcius: ah
[19:46] mape: All articles mention it
[19:46] mape: for the weekend :)
[19:46] mape: /re ry
[19:46] magcius: for some reason scons from waf is failing
[19:46] mape: http://ajaxian.com/archives/oni-rocket
[19:46] magcius: python "/home/node/tools/scons/scons.py" -C "/home/node/build/default/" -Y "/home/node/deps/v8"
[19:46] magcius: fails
[19:47] ncb000gt: alxc1: i believe so
[19:47] pjb3 has joined the channel
[19:47] magcius: but scons -C "/home/node/build/default" -Y "/home/node/deps/v8" works
[19:47] magcius: and they are both the same version
[19:48] sprsquish has joined the channel
[19:48] teadict has joined the channel
[19:48] ncb000gt: mape: what's interesting is that most people/companies seemed to be focusing on JS for web frameworks, while node was written with the mindset of true serverside/desktopside
[19:49] mape: I'm just glad it is so handy
[19:49] alxc1: ncb000gt: ok cheers
[19:49] mape: not having to run java or other clumpsy stuff
[19:50] ncb000gt: meh, c/c++/java/etc... it's still running in some level of interpreted env.
[19:50] joshbuddy has joined the channel
[19:51] ncb000gt: java is just more verbose in terms of the code...that said most of the V* stuff I've seen so far isn't far from Java
[19:51] ncb000gt: s/v*/v8/
[19:53] mape: is node doing summer or code? 
[19:53] ncb000gt: tho, I am really happy that node is around. It's easy and fast.
[19:53] ncb000gt: not sure
[19:53] mape: Or is that a pita to get into?
[19:53] ncb000gt: i'm not so sure it's a PITA to get into
[19:54] ncb000gt: but you have to have some really defined goals iirc
[19:54] mape: Makes sence
[19:55] mape: Seems like a neat way to get some semi funding/dev time
[19:55] mape: Seems like that is an issue now
[19:57] ncb000gt: welcome to open source. ;D
[19:57] mape: Hehe yeah
[19:58] alxc1: ncb000gt: ok with that in mind... do you know if script.runInNewContext be used to create new contexts that will continue to exist even after the context that created it has expired? thanks
[19:58] ncb000gt: brb, phone
[19:58] magcius: okay, what happened to "multipart"?
[19:59] ashb: _ry decided it sucked
[19:59] mape: soc?
[19:59] ashb: soc = Google Summer Of Code
[20:00] mape: Yeah I know
[20:01] mape: Does ry work full time on node? Or sideproject?
[20:02] ashb: he works for Joyent i think
[20:02] mape: Just find it strange he manages to do so much for node on a side basis
[20:03] sprsquish: Joyent is doing stuff with node so if he's not on it full time, he's on it a lot over there
[20:03] ashb: last i heard joyent are using node sa part of a message queue consumer
[20:03] mape: k
[20:03] binary42 has joined the channel
[20:03] qFox: http://developers.slashdot.org/story/10/04/24/1734241/Proof-of-Concept-For-Ajax-Without-JavaScript :(
[20:04] cadorn has joined the channel
[20:04] Nohryb has joined the channel
[20:05] qFox: actually. nevermind (the linked page took ages to load). i can do that without xml. a form, a target and an iframe will make that work :/
[20:05] ashb: man someone doesn't get the point of client side JS
[20:05] mape: <type 'exceptions.OSError'>
[20:05] mape: heh
[20:06] mape: love the header
[20:07] qFox: the 404 is annoying
[20:07] qFox: takes you a minute or so to figure out that its actually teh 404 :/ especially if that's the first attempt
[20:09] ncb000gt: alxc1: not sure, sorry for the delay
[20:10] alxc1: ncb000gt: thanks, no worries!
[20:11] polo7: Hey, I have some difficulties with the module system! someone can help me please?
[20:11] mape: ask :)
[20:11] polo7: cool
[20:12] polo7: I'm using the express 'Class' module
[20:12] polo7: I think I shouldn't be calling it for each files but just once in the main file right ?
[20:12] polo7: Ok but if I write in the main file :
[20:12] polo7: var Class = require('support/class/lib/class').Class;
[20:13] polo7: it won't be visible in the other files as var is local to this file right ?
[20:13] polo7: so I should write global.Class = ... ?
[20:14] ashb: globals are evil.
[20:14] polo7: but I thought express already did it
[20:14] ashb: require it every place you need it
[20:14] polo7: yeah I suppose
[20:14] ashb: its not very expensive
[20:14] ashb: infact its dead cheap
[20:15] polo7: ok !
[20:15] polo7: and what if I write :
[20:15] polo7: without the var
[20:15] polo7: I saw that in express source
[20:15] ashb: then the gremlins will climb out of your monitor and bite your face off
[20:15] ashb: i.e. don't do it, its evil
[20:15] polo7: lol
[20:16] polo7: ok thx
[20:16] ncb000gt: ashb: it's awful when it happens! :(
[20:16] ashb: i noe rite!?
[20:16] ncb000gt: I'm lucky my C-k is fast!
[20:16] ncb000gt: <.<
[20:17] polo7: also? :)
[20:18] polo7: I want to declare a new function for Object
[20:18] polo7: like :
[20:18] polo7: Object.prototype.cast
[20:18] polo7: where should I declare it ?
[20:18] ashb: don't put things on Object.prototype
[20:18] polo7: :s
[20:18] ashb: certainlly not enumerable things
[20:19] ncb000gt: lol
[20:19] ashb: you can if oyu wan't, but be warnings some things might sudenly break
[20:19] ashb: &
[20:19] polo7: mmm
[20:19] ncb000gt: polo7, if you need casting functions and such, just throw them in a module
[20:19] polo7: and if I want to add it to every instance of Class
[20:20] ncb000gt: instead of doing obj.cast(class)
[20:20] ncb000gt: you'd do something like mod.cast(obj, clazz)
[20:20] polo7: yeah ok not bad :)
[20:22] Aria has joined the channel
[20:22] magcius: ashb: how am I supposed to read multipart form data onw?
[20:23] zackattack has joined the channel
[20:23] ncb000gt: magcius: first you sacrifice a chicken
[20:23] magcius: I don't care that ry thought it sucked
[20:24] mape: application/blackmagic ?
[20:24] ncb000gt: lol
[20:24] magcius: You shouldn't be breaking backwards compatibility this quickly. My application worked a month ago
[20:24] ncb000gt: magcius: how were you doing it?
[20:24] magcius: Screw node.js if all it's going to do is break shit
[20:24] magcius: ncb000gt: using the multipart module?
[20:24] mape: is the module broken? Wasen't it just moved out of core?
[20:24] ncb000gt: ahh, I suspect it's still possible through the res events
[20:25] martoche has joined the channel
[20:26] ncb000gt: also, _ry said that the api's would stablize around .2
[20:26] ashb: magcius: i agree with you - i think _ry's stance on backcompat is a fucking joke
[20:30] magcius: now
[20:30] magcius: I don't use node.js as a web framework
[20:30] magcius: I use it as a lint.js server for code coverage
[20:30] ncb000gt: magcius: there looks to be something to do with multpart here http://github.com/isaacs/multipart-js/
[20:30] magcius: So I need multipart data
[20:31] magcius: ncb000gt: I don't care. Tell that ry that if he wants some actual real-world uses he shouldn't treat Node as a toy
[20:31] magcius: His toy.
[20:32] qFox: or wait for a stable version and stick with it.
[20:32] mikeal1 has joined the channel
[20:32] ashb: magcius: ^^ that is my exact view
[20:32] ashb: which is why i don't use node
[20:32] qFox: also, wasnt he going to freeze the api?
[20:32] ashb: he doesn't have to freeze the API
[20:32] noonat: would you rather he doesn't release it until he is ready for the api to be stable?
[20:32] ashb: just manage backcompat nicer
[20:32] maushu: He removed multipart because it was too much code for something simplr.
[20:32] magcius: noonat:  yes
[20:32] noonat: magcius: you're insane :P
[20:33] mape: magcius: Then just wait and don't use it? 
[20:33] qFox: right
[20:33] noonat: i think node has been perfectly clear about it's stability
[20:33] magcius: noonat: if he did that, things wouldn't break every 5 months when I bork my node.js install and forget which version I'm using
[20:33] noonat: if you don't want to use something with api instability, then don't use it until it's stable
[20:33] magcius: The problem is that these kinds of libraries aren't developed out of a use case to support.
[20:33] qFox: if you really dont want him to release it beforehand. just dont use it till he does. let us play around with the toy
[20:33] noonat: no one is forcing you to, and they're not lying and saying it is stable
[20:33] ncb000gt: magcius: it's new.
[20:33] ncb000gt: give it some time.
[20:33] magcius: ncb000gt: if 2 years is "new"
[20:33] maushu: At least he shares his toys, unlike you. :p
[20:34] qFox: hehe
[20:34] maushu: Sharing is caring.
[20:34] magcius: maushu: I develop my toys out of refactoring a bigger, messier toy, so I can't go ahead to try and break stuff just because I feel like it
[20:35] magcius: If he actually wrote an app in Node.js using any of the features, he'd actually stop doing stupid shit that makes no fucking sense
[20:35] maushu: ...what stupid shit?
[20:35] magcius: Because I don't want to wait for another 5 months where whatever stdlib I had then is gone
[20:36] magcius: maushu: gradually replacing the stdlib until
[20:36] magcius: until the latest hotness 5 months ago no longer works
[20:36] magcius: and you're stuck not getting stupid bug fixes
[20:36] maushu: replacing stdlib?
[20:37] magcius: maushu: I've written three versions of this jslint standalone server
[20:37] magcius: maushu: because each time he decided to phase out something useful because he didn't feel it was cool anymore
[20:37] magcius: It's hell
[20:37] RayMorgan has joined the channel
[20:37] mape: There is no way you can stay on older versions?
[20:37] polo7: I don't get it, why can't you stay with an older version of node and wait for multipart to come back ?
[20:38] polo7: lol
[20:38] magcius: Because the older version didn't support IPv6
[20:38] RayMorgan has joined the channel
[20:38] maushu: ...congratulations, you just found out how nice it is to develop projects for frameworks that are also under heavy development.
[20:38] ncb000gt: polo7: it's not likely it's coming back if it has been removed
[20:39] polo7: ry will write his own ^^
[20:39] ncb000gt: magcius, i take it you are uploading files? is there instead a way you can pull code from the source repository you're using instead?
[20:39] ncb000gt: no mimeparts needed.
[20:39] magcius: ncb000gt: this is local code
[20:40] magcius: ncb000gt: I'm using it to do live error highlighting in emacs
[20:40] magcius: and _ry forbid that I add an IPv6 IP address to do my actual work at Verizon which is support network testing tool
[20:40] magcius: tools
[20:40] mape: fork? :P
[20:40] magcius: I look on the Google Group, oh IPv6 is fixed in v0.9.8, so I upgrade to that tag, and multipart is removed
[20:41] magcius: great
[20:41] magcius: now I have to rewrite it to the latest greatness hoping not to hit another bug and have to rewrite again
[20:41] qFox: 
ACTION excepts that you'll have to rewrite it a few more times still

[20:42] qFox: excepts? really? ... expects.
[20:45] MattJ: Nah, ry said the API is more or less stable now afaik
[20:45] MattJ: I've also had to change code because of the API changes
[20:45] MattJ: But I'm not complaining, if those API changes improve Node
[20:45] ncb000gt: agreed.
[20:49] zimbatm has joined the channel
[20:50] zimbatm has joined the channel
[20:50] zimbatm: hello
[20:51] joshbuddy has joined the channel
[20:52] ncb000gt: zimbatm: hi
[20:52] zimbatm: I would like to discuss something... I'd like to propose to move src/node.js to the lib/ directory
[20:53] zimbatm: Having that .js file in src/ has caused some confusion to newcomers
[20:53] ncb000gt: heh, _ry isn't here for the weekend
[20:53] zimbatm: and since lib/ is now also statically compiled..
[20:53] zimbatm: doh
[20:53] ncb000gt: heh
[20:53] ncb000gt: probably something for the google group?
[20:54] zimbatm: yeah, I guess so
[20:55] admc has joined the channel
[20:56] saikat has joined the channel
[20:58] mfeiri_ has joined the channel
[21:17] zimbatm: posted: http://groups.google.com/group/nodejs/browse_thread/thread/f518a4a0c37068bd
[21:22] derbumi has joined the channel
[21:22] creationix has joined the channel
[21:24] inimino: alxc1: there is one v8 process per node invocation
[21:25] polo7: where can I find a list of all functions implemented for objets and arrays ?
[21:25] polo7: I came across each merge map ...
[21:25] ashb: https://developer.mozilla.org/en/Core_JavaScript_1.5_Reference/Objects/Array
[21:25] polo7: but never seen a doc
[21:25] ashb: v8 will have most of those
[21:25] ashb: everything except Array.map (the static version)
[21:26] ashb: and others like that
[21:26] polo7: ok thx
[21:26] jspiros has joined the channel
[21:26] polo7: this is for arrays 
[21:26] polo7: and objects ?
[21:26] creationix has left the channel
[21:26] ashb: take a guess :)
[21:26] jspiros has joined the channel
[21:26] creationix has joined the channel
[21:26] polo7: yeah shouldn't be far
[21:27] creationix: polo7: http://wiki.github.com/ry/node/ecma-5mozilla-features-implemented-in-v8
[21:27] ashb: the stuff that is Requires JS 1.9.3 might or might not be in v8, but they should be soon - they are added in ECMAScript 5
[21:28] creationix: polo7: Object.getOwnPropertyNames(Array.prototype)
[21:28] ashb: polo7: avoid anything that is __wunderBared__
[21:29] polo7: thanks percfect
[21:29] jspiros has joined the channel
[21:29] creationix: since we now have Object.getOwnPropertyNames, we can see the prototype of anything
[21:29] ashb: use defineProperty or getOwnPropertyDescriptor instead
[21:29] polo7: oki
[21:30] polo7: creationix, you're the one who wrote Do right ?
[21:30] creationix: yep, why
[21:31] polo7: I thought I could use it but it was too much work to modify my code
[21:31] polo7: so I use your chain function
[21:31] polo7: and modified it to use callbacks
[21:32] creationix: polo7: have you seen my Step library
[21:32] creationix: it works with node style callbacks
[21:32] polo7: it was very useful
[21:32] creationix: Do works with continuable style
[21:32] polo7: nop
[21:33] polo7: it is not on github ?
[21:34] alxc1: inimino: oh i see... so really in most cases, nodejs is only running one context in each process unless the script demands otherwise?
[21:34] ashb: its alwasy only running one context
[21:34] ashb: i don't thin v8 will work any other way
[21:34] polo7: first I thought I could modify your convert function to convert my Class object
[21:34] derbumi has joined the channel
[21:35] polo7: but I didn't really figure it out how
[21:35] alxc1: ashb: on the contrary v8 is designed for many simultaneous contexts from what i've been reading, and that was a feature i wanted to take advantage of.
[21:35] ashb: alxc1: link?
[21:35] ashb: cos in chrome its process per tab
[21:36] alxc1: oh really
[21:36] ashb: yes. its the central idea of chrome
[21:36] ashb: process per tab, process per plugin per tab
[21:36] alxc1: ok i'm mixing up terminology
[21:37] alxc1: there's many contexts in one *instance* of v8, right?
[21:37] ashb: what do you mean by instance and context?
[21:37] alxc1: All i'm trying to do is read their docs :P http://code.google.com/apis/v8/embed.html#contexts
[21:38] alxc1: it says many, unrelated javascript applications (i.e. a context) in one instanc of v8 (what i meant by 'process')
[21:38] ashb: i think its one context
[21:39] ncb000gt: I believe there is one for each script. but there is a main context
[21:39] polo7: creationix, ok I found your step lib. Yeah, I kind of rewrote that :s
[21:40] creationix_ has joined the channel
[21:40] creationix: polo7: cool
[21:40] polo7: too bad!
[21:41] alxc1: ok so what i'd like to understand is, and sorry if this is getting a bit convoluted, if you invoke another script by calling 'node script.js' is it going to merely create a new context in an existing v8 instance if there is one already running on the system, or is it going to create a whole new v8 instance? thanks for thinking it over.
[21:41] ashb: its going to create a whole new process
[21:42] ncb000gt: what ashb said
[21:42] ashb: whole new *everything*
[21:42] ashb: the only way it could reuse stuff created was if you did a fork.
[21:42] ashb: but then you would'nt be calling `node script.js`
[21:42] alxc1: hmm ok thanks, that's great, i just needed to understand that.
[21:43] herbySk: alcx1: but you can create a new context inside node if you want, see Script in docs.
[21:44] mape: has there been any more pokes at the webworker multi core node thingymado?
[21:45] alxc1: herbySk: yah that is what i will do, so i'll invoke a 'master' script and go from there, with just one instance of v8 :) cool. i'm spawing the nodejs as an erlang port node, so i didn't want to go about creating loads of them if it was going to lead to lots of redundancy.
[21:55] isaacs has joined the channel
[21:57] isaacs: _ry: you around?
[21:58] mape: he is away
[21:58] mape: for the holiday
[21:59] herbySk: ah, I wanted him as well....
[22:01] creationix: mape: As long as you don't mind communicating only over serialzable channels, then I find the childProcess model works great.
[22:01] isaacs: alxc1: what's this about contexts?
[22:01] creationix: evalctx right?
[22:01] isaacs: i can tell you, at least in the current HEAD, node script.js will run script.js in exactly the same context as node.js and everything else.
[22:02] creationix: process.evalctx(fs.readFileSync('script.js'))
[22:02] isaacs: there is no evalcx going on right now, if that was a question (just skimmed the tail of the logs)
[22:02] isaacs: creationix: no t
[22:02] isaacs: just cx, not ctx
[22:02] creationix: ahh, right
[22:02] isaacs: also, you should provide a context object.
[22:02] creationix: I should know better, I use it for the sandbox in wheat
[22:02] isaacs: process.evalcx(code, { require: require }, "filename")
[22:03] isaacs: of course, giving it "require" pretty much makes the "sandbox" useless ;)
[22:03] creationix: well, this keeps it from messing with existing globals
[22:03] creationix: or is evalcx not a new global context
[22:04] creationix: though it works great for modules that don't use exports, you can see what got added to the context object you passed in
[22:05] sudoer has joined the channel
[22:05] maushu: Private stuff is a must in sandboxes.
[22:07] alxc1: isaacs: i wanted to clarify the best way to create multiple contexts in the same v8 VM using nodejs, rather than lauching multiple instances of the VM with one context each, but i found out what i needed :)
[22:08] isaacs: alxc1: yah, process.evalcx is the way to go then.
[22:09] isaacs: alxc1: or you can check out require("script") which has more granular utilities.  herbySk wrote that, i think.
[22:10] alxc1: isaacs: wait... no i was going to use Script, .evalcx doesn't appear to be in the docs...?
[22:10] brainproxy has joined the channel
[22:10] darkf has joined the channel
[22:10] alxc1: isaacs: yah ok script it is, all sorted cheers.
[22:11] noonat: evalcx is gone, script replaced it
[22:11] isaacs: ah, i see
[22:12] qFox: does make more sense though :)
[22:17] herbySk: well, gone only from the docs, in code there is process.evalcx = process.binding('evals').Script.runInNewContext
[22:17] herbySk: It depends on ry how long will the evalcx have a grace period before gone actually
[22:18] alxc1: Well if you want an outsider's perspecitve, Script sounds intuitive enough to me and i've only know about nodejs for a couple of hours :P
[22:18] mcarter has joined the channel
[22:19] mape: $.getScript
[22:24] creationix: hmm, the new documented script object looks handy
[22:27] inimino: isaacs: that Safari issue should be fixed
[22:27] isaacs: oh, cool
[22:27] isaacs: cuz i've got a lot of old code to reformat properly.  who wrote this shit anyhow?
[22:27] isaacs: 
ACTION is working on npm...

[22:27] inimino: isaacs: if you want to send me the list of Firefox extensions you have installed I might see if I can reproduce that too
[22:28] inimino: I just updated my server from node 0.1.25 to 0.1.92
[22:28] rictic has joined the channel
[22:29] zimbatm has joined the channel
[22:30] isaacs: inimino: so, i'm not sure what this should be doing?
[22:31] inimino: isaacs: did the stuff on the right show up this time?
[22:31] isaacs: nope
[22:31] isaacs: i'm in safari, nothing there.
[22:31] isaacs: just blank
[22:32] inimino: argh
[22:32] isaacs: inimino: it'd be easier if there was an example that just took a string and returned a string, and i'd plug it into a node program
[22:32] inimino: alright, I'll test it later, don't have Safari in front of me at the moment
[22:32] isaacs: inimino: i'm just gonna run it on the command line anyhow.
[22:33] inimino: sure, but it's easier to write it this way, ...when it works
[22:33] inimino: oh, hm, hang on
[22:33] isaacs: speak for yourself :)
[22:34] inimino: it really is fun
[22:34] inimino: hm, yeah I see the blank output too
[22:34] inimino: 
ACTION investigates

[22:35] creationix: speaking of npm...
[22:35] inimino: erm, no, spoke too soon
[22:35] creationix: I should get into this package management contest before it's too late ;)
[22:35] inimino: heh
[22:36] ncb000gt: lol
[22:36] creationix: the more the merrier right?
[22:37] ncb000gt: that's one view
[22:37] ncb000gt: :)
[22:37] inimino: isaacs: hm, I'll make a modularized version and you can play with that in V8
[22:37] inimino: isaacs: but send me that list of extensions you have in Firefox
[22:37] noonat: what we really need is a package manager manager
[22:37] inimino: this really should work
[22:37] creationix: noonat: now you're talking
[22:38] RayMorgan has joined the channel
[22:38] creationix: so a package middle manager
[22:38] creationix: is that a form of middleware?
[22:38] isaacs: haha
[22:39] creationix: If only I wasn't swamped trying to finish school and move to California, I'd have time to whip something up
[22:40] maushu: We need a package manager manager.
[22:40] maushu: To manage package managers.
[22:40] ashb: i really wish i had more time to work on my doc render site thing
[22:41] Aria: 
ACTION laughs.

[22:42] Aria: Hey, as long as I'm not forced to use something like Rubygems, I'm happy.
[22:42] inimino: isaacs: hm, it works for me on Safari on OS X too
[22:42] creationix: If I ever do make it to California, I need to meed up with all the noders there
[22:42] ashb: i just want something that lets you search packages and view its docs like search.cpan.org
[22:42] creationix: ashb, github?
[22:42] ncb000gt: creationix: where are you now?
[22:42] creationix: ncb000gt: North Dallas
[22:42] isaacs: ashb: you seen the js-registry thing?
[22:42] ncb000gt: ahh
[22:43] ashb: creationix: github search is too general
[22:43] ashb: isaacs: no?
[22:43] isaacs: ashb: check out http://registry.npmjs.org/
[22:43] ashb: creationix: also github doesn't (yet) render JSDoc
[22:43] ashb: isaacs: that downloaded a file...?
[22:44] isaacs: ashb: it's json
[22:44] isaacs: ashb: i highly recommend the json view extension for firefox
[22:44] ashb: isaacs: what i wan't is just an online rendering of the docs for any and every package i can find
[22:44] ashb: i find firefox so jarring on OSX i don't use it
[22:44] creationix: isaacs: is there such a thing for chrome?
[22:44] isaacs: creationix: not sure
[22:45] isaacs: ashb: we could make the registry support something like /packagename/version/README or something
[22:45] ncb000gt: creationix: it's just a string in chrome
[22:45] ncb000gt: doesn't download
[22:45] ncb000gt: well...
[22:45] ncb000gt: doesn't download as a file.
[22:45] ncb000gt: :)
[22:45] isaacs: ashb: the vast majority of packages have a README.md in the root
[22:45] ashb: isaacs: README is okay, but for a large package you want class docs too
[22:45] creationix: well, most js files aren't going to have jsdocs
[22:45] creationix: the language doesn't lend itself to automatic doc generation
[22:45] isaacs: ashb: good point.  well, good luck with that. ;)
[22:46] ashb: isaacs: yeah. i know :)
[22:46] ashb: creationix: my long term goal (ha!) is to support every doc format used in any commonJS package
[22:46] creationix: sounds like fun
[22:46] ashb: but well i've just started a new 3 month contract so its been put on hold
[22:46] ashb: plus i have an awesome domain for it - http://packag.es >_>
[22:47] ncb000gt: lol
[22:47] creationix: isaacs: how do I get my packages in your npm registry?
[22:48] isaacs: creationix: you do npm publish <tarball url>
[22:48] ashb: (oh and once i've got the docs up i'll try to start mirroring all the dists as well)
[22:48] isaacs: creationix: i think you might have to do npm register email password first
[22:48] ashb: since to render the docs i'll need to get the content anywya
[22:49] isaacs: and i need to integrate with the new crypto stuff
[22:50] ashb: new crypto stuff?
[22:52] rictic has joined the channel
[22:52] bmizerany has joined the channel
[22:53] isaacs: ashb: require("crypto")
[22:53] ashb: ah
[22:53] isaacs: so your email/pw isn't exposed in plain text in your home dir
[22:53] isaacs: base64, but still, not crypted
[22:53] ashb: use ssh private keys ^__^
[22:55] inimino: isaacs: er, whoops, nevermind, I see the problem in Safari now
[22:55] isaacs: ashb: that's the plan.
[22:55] ashb: cool
[22:55] isaacs: use ~/.ssh/id_rsa as the crypt key
[22:55] inimino: 
ACTION accidentally tested Firefox on OS X instead before

[22:56] isaacs: ahh
[22:56] ashb: problem with?
[22:56] inimino: I really should set up some kind of VNC or something so I don't have to run between rooms
[22:56] inimino: ashb: this webpage IDE-type thing I have
[22:56] ashb: ah
[22:57] ashb: inimino: btw OS 10.5+ has VNC server built in
[22:57] inimino: is there something like VNC for OS X?
[22:57] inimino: ah...
[22:57] ashb: 10.4 has it too actually i think
[22:57] ashb: System Prefernce -> Sharing
[22:57] ashb: 'Screen Sharing'
[22:57] inimino: ah
[22:58] inimino: I think that has 10.4 on it, so it's probably there
[22:58] Nohryb has joined the channel
[22:59] ssteinerX has left the channel
[22:59] maushu: ubuntu 10.4 > osx 10.4
[22:59] ssteinerX has joined the channel
[23:07] isaacs: inimino: i've not had good experiences in the past with IDE-type things, especially webpage ones.
[23:07] inimino: hehe
[23:08] inimino: isaacs: do you like REPLs?
[23:08] isaacs: inimino: yeah.
[23:08] inimino: then you'll probably like this
[23:08] isaacs: inimino: but for stuff like this, my favorite repl is bash
[23:08] inimino: bash is nice :)
[23:09] maushu: bash makes the world go around
[23:09] maushu: and around
[23:10] maushu: Everyone knows that the universe is coded in bash.
[23:10] inimino: I'm going to have to work out this Safari thing later
[23:10] inimino: some Web Workers API difference
[23:20] softdrink has joined the channel
[23:33] pjb3 has joined the channel
[23:38] mikeal1 has joined the channel
[23:45] sudoer has joined the channel
[23:46] jtoy has joined the channel
[23:49] cloudhead has joined the channel
[23:56] kriskowal has joined the channel
[23:58] Nohryb has joined the channel