[00:08] emyller has joined the channel
[00:44] Connorhd has joined the channel
[00:45] hassox_ has joined the channel
[00:50] alan1 has joined the channel
[01:02] Biscuits: Hmmm, yay, playing with Parser combinators is nice :D
[01:06] inimino: :)
[01:10] isaacs_ has joined the channel
[01:26] sudoer has joined the channel
[01:34] Biscuits: Yay, simple parser combinators that are interriptible on any character :p
[01:34] Biscuits: You can even feed it data in pieces
[01:40] jed has joined the channel
[01:45] LordMetroid: Anyone who had experience using Node.js with CouchDB?
[01:53] aguynamedben has joined the channel
[01:54] inimino: Biscuits: got some code to show? :-)
[01:55] Biscuits: What's a nice pastebin ?
[01:56] Biscuits: nvm
[01:56] Biscuits: http://pastebin.com/m686ad498
[01:56] Biscuits: It's work in progress though
[01:56] Biscuits: but seems to do a pretty nice job
[01:57] Biscuits: I'm concerned about the memory usage though
[01:57] inimino: in a meeting right now but will read it after
[01:57] Biscuits: Let me know what you think :)
[02:03] elliottcable: ARGH
[02:03] elliottcable: ARGHGHGHGHG
[02:03] elliottcable: http://gist.github.com/245217
[02:04] elliottcable: I wonder if I could convince ryah_away to move away from asciidoc
[02:04] Biscuits: asciidoc ?
[02:04] ashb: thats more a problem with docbook than asciidoc isn't it?
[02:04] inimino: what's asciidoc?
[02:05] inimino: I guess I should have looked at the paste before asking
[02:05] elliottcable: the node documentation is generated with it
[02:05] inimino: looks like a docbook problem to me
[02:06] elliottcable: 
ACTION shrugs

[02:06] elliottcable: I have no idea what either is, beyond that ‘they make the Node.js documentatiuon’
[02:06] Biscuits: Don't suppose there's a quick way to array comparison in JS ?
[02:06] Biscuits: Trying to keep my parser library thing compatible with both arrays and strings
[02:06] elliottcable: so I suppose I meant, drop ‘wahtever bit isn’t working consistently’
[02:06] Biscuits: but == only works on string
[02:06] elliottcable: shouldn’t need internet access to build Node /-:
[02:07] ashb: ["a"] == "a"
[02:07] ashb: true
[02:07] ashb: == will stringify (or call valueOf) in some cases
[02:07] elliottcable: hence why I studiously avoid it
[02:07] elliottcable: lol
[02:08] elliottcable: why the poop is this requiring an external network object
[02:08] elliottcable: v.v
[02:08] ashb: tweak the build scripts to not run the doc commands
[02:09] elliottcable: uh
[02:09] elliottcable: I could build node without the documentation
[02:09] elliottcable: that’s my point, I want the documentation
[02:09] ashb: tough :)
[02:09] ashb: read the plain text version
[02:09] elliottcable: I *have* node, I *want* the documentation.
[02:09] elliottcable: your response is unsatisfactory.
[02:10] ashb: the fix it. thems the choices
[02:10] elliottcable: would if I could
[02:11] inimino: Biscuits: function arrEq(a,b){if(a.length!=b.length) return false; etc}
[02:13] Biscuits: Hmm, manual way it is then
[02:22] kriskowal has joined the channel
[02:37] inimino: 
ACTION looks

[02:37] elliottcable: wat
[02:38] inimino: at Biscuits' code
[02:39] Biscuits: Probably not very efficient
[02:39] Biscuits: but it seems to get the job done on small things
[02:39] sudocarl: Biscuits, what exactly does it do?
[02:40] Biscuits: sudocarl: Parser combinators library :)
[02:40] Biscuits: Parses data
[02:40] sudocarl: ah
[02:40] Biscuits: e.g. to parse a list of words into an array
[02:40] Biscuits: var parseDigit=new SatisfyParser(function(chr) { return (chr>='0' && chr<='9') }).f(function(c) { return c-'0'; });
[02:40] Biscuits: var parseAlpha=new SatisfyParser(function(chr) { return ((chr>='a' && chr<='z')||(chr>='A' && chr<='Z')); });
[02:40] Biscuits: var parseAlNum=parseDigit.or(parseAlpha);
[02:40] Biscuits: var parseWord=parseAlNum.some().f(function(l) { return l.reduce(function(a,b) { return a+b; },""); });
[02:40] Biscuits: var parseWhitespace=new SatisfyParser(function(chr) { return (chr==' ' || chr=='\n' || chr=='\r' || chr=='\t'); }).some().f(function(x) { return x.length; });
[02:40] Biscuits: var parseListOfWords=parseWord.list(parseWhitespace).eof();
[02:40] sudocarl: ohhh sweet
[02:40] Biscuits: err
[02:41] inimino: Biscuits: your arrayFlatten can be made more efficient
[02:41] Biscuits: That code was a bit more than I expected
[02:41] Biscuits: inimino: Exactly why I made it a function, I had a suspicion it could be ;)
[02:41] inimino: return Array.prototype.concat.apply([],a)
[02:41] inimino: or something to that effect
[02:42] elliottcable: hahah inimino 
[02:42] inimino: elliottcable: yes?
[02:43] Biscuits: inimino: So any other suggestions ?
[02:44] inimino: haven't read further yet :)
[02:44] inimino: what's line 8 about?
[02:45] elliottcable: var probe = require('probe').probe; probe(acquire, ‘acquire’); probe(acquire.absolute, 'acquire.absolute');
[02:45] elliottcable: woo
[02:45] Biscuits: Just my general Class definition
[02:45] elliottcable: works great
[02:45] elliottcable: github.com/elliottcable/probe
[02:45] elliottcable: a bit more verbose alternative to sys.puts(sys.inspect())
[02:45] elliottcable: for anybody who wants it
[02:45] malkomalko has joined the channel
[02:45] inimino: Biscuits: if it's intended to be called with 'new', the return value is not used
[02:45] inimino: (and *ahem* JavaScript has no classes)
[02:45] Biscuits: I always return this; from any class constructor, so I know it's a class constructor :)
[02:45] inimino: (which it's not ;-)
[02:46] inimino: elliottcable: probe?
[02:46] sudocarl: Biscuits, do you have a bit of a more outlined use of your library?
[02:46] elliottcable: inimino: yep
[02:46] sudocarl: like a demo file?
[02:46] inimino: oh, I see
[02:46] inimino: 
ACTION will look at that too

[02:47] Biscuits: inimino: function f() { return {toString:function() { return "should never see this";}}; }; sys.debug(new f());
[02:47] Biscuits: inimino: try that, in my v8 console it actually returns that object returned by return :p
[02:48] Biscuits: sudocarl: It's for creating quick parsers for strings
[02:49] inimino: Biscuits: actually, if you return something, it seems to be used
[02:50] inimino: (but that doesn't mean you should do that)
[02:51] Biscuits: Hmmm, well, I don't see a downside to it, and it safely implies how the function should be used :/
[02:52] inimino: well, I don't think it's intuitive to JavaScript programmers
[02:52] inimino: I don't recall ever seeing it anyway
[02:53] Biscuits: hmmm fss, just trying it on some larger strings
[02:53] Biscuits: takes ages to complete :(
[02:53] sudocarl: Biscuits, did you consider doing a recursive-descent parser instead?
[02:53] sudocarl: or maybe that wouldn't work...
[02:53] sudocarl: hmm
[02:53] inimino: 'new' does its own thing, it's not really the same as a normal function call
[02:53] inimino: but the spec does say that if the function returns a value of type object, it should be used
[02:53] sudocarl: wonder what the python parse library does
[02:54] Biscuits: sudocarl: Might do once we hit that subject in the course I'm currently following on languages & compilers ;)
[02:54] sudocarl: huh interesting I am in a similar course right now :)
[02:54] inimino: this looks pretty recursive-descent-ish to me so far
[02:55] inimino: is it not?
[02:55] sudocarl: hmm I dunno
[02:55] sudocarl: I think of recursive descent as a more structured system formed of mutally-recursive procedures
[02:55] sudocarl: these each work alone
[02:56] inimino: true
[02:58] Biscuits: bah, on 2000+ characters it gets really slow :(
[02:58] sudocarl: :(
[02:58] inimino: Biscuits: anyway, looks interesting
[02:58] inimino: yes, it's never going to be fast
[02:59] Biscuits: Wish JavaScript was lazily evaluated like Haskell, bet that would speed it up some
[02:59] sudocarl: Biscuits, you might want to look how "http://pyparsing.wikispaces.com/" is implemented...  its a very very fast parser implemented in python which uses BNF as its declaration syntax
[02:59] sudocarl: something like that would be amazing in JS
[02:59] Biscuits: sudocarl: I believe someone already showed me a BNF->JS converter somewhere
[02:59] sudocarl: really? I would like to see that
[03:00] Biscuits: Look around in the logfiles
[03:01] sudocarl: kk
[03:01] inimino: probably me
[03:01] inimino: though it was PEG not BNF
[03:01] inimino: http://boshi.inimino.org/3box/PEG/build/demo.html
[03:01] Biscuits: Ah, yeah, that one :)
[03:02] Biscuits: Looks a whole lot like BNF though :/
[03:02] sudocarl: inimino, is that dynamic?
[03:02] inimino: Biscuits: what would really speed it up is if JavaScript was compiled, and had a type system like Haskell, and a very smart compiler
[03:02] inimino: the semantics are different from BNF but the syntax is similar
[03:03] inimino: PEG is a more flexible and convenient formalism IMO
[03:03] inimino: sudocarl: in what sense?
[03:04] sudocarl: like can I change the grammer on the page and get it to recompile?
[03:04] sudocarl: and reparse?
[03:04] Biscuits: inimino: Is that code yours? if so, what kind of license is it ?
[03:04] inimino: Biscuits: yes it's mine, as for the license I don't know, maybe MIT?
[03:05] inimino: sudocarl: oh, no, sorry haven't added that yet
[03:05] Biscuits: And did you run build_output to some kind of JS minimizer? or do you just code really dense ?
[03:06] inimino: Biscuits: most of that code is the way I write it, and some of it is generated code
[03:06] Biscuits: Ah, k
[03:07] inimino: (you can tell the generated code because it has more whitespace ;-)
[03:07] Biscuits: heh :p
[03:07] sudoer has joined the channel
[03:07] sudocarl: inimino, no worries  :)
[03:07] inimino: (actually the generated code is mostly the wall-of-text and the wall-of-numbers about halfway down)
[03:08] inimino: also it has full Unicode support which that demo doesn't really demonstrate
[03:08] inimino: (including characters outside the BMP)
[03:09] Biscuits: Hmmm, downside to this is that it's not streaming like Node.js
[03:10] Biscuits: hmm btw
[03:10] Biscuits: LALR ECMAscript on google gives me:
[03:10] Biscuits: http://jscc.jmksf.com/
[03:10] Biscuits: Seems rather promising
[03:11] inimino: js/cc is more mature
[03:12] inimino: and it's LALR(1) which is of course an entirely different kettle of fish
[03:44] inimino: sudocarl: I made the input editable (but not the grammar) http://boshi.inimino.org/3box/PEG/build/demo.html
[04:00] cloudhead has joined the channel
[04:05] jed has joined the channel
[04:13] aguynamedben has joined the channel
[04:22] kriskowal_ has joined the channel
[04:39] alan1 has joined the channel
[04:55] sudoer has joined the channel
[05:13] jtoy has joined the channel
[05:18] hassox has joined the channel
[05:19] atmos has joined the channel
[05:26] isaacs has joined the channel
[05:43] jtoy has joined the channel
[05:46] voodootikigod has joined the channel
[05:46] voodootikigod: !log
[05:49] softdrink1 has joined the channel
[05:57] jtoy has joined the channel
[06:16] sprsquish has joined the channel
[06:57] jed has joined the channel
[07:25] isaacs has joined the channel
[07:58] the_undefined has joined the channel
[08:16] johan-s has joined the channel
[08:39] sixtus42 has joined the channel
[09:17] the_undefined has joined the channel
[09:52] cmlenz has joined the channel
[10:08] saikko has joined the channel
[10:32] hassox has joined the channel
[10:40] sixtus42_ has joined the channel
[11:13] Biscuits has joined the channel
[11:18] webben has joined the channel
[11:29] rolfb has joined the channel
[11:36] webben has joined the channel
[11:51] sixtus42_ has joined the channel
[11:55] sixtus42 has joined the channel
[12:01] malkomalko has joined the channel
[12:07] rakeshpai has joined the channel
[12:32] olivvv has joined the channel
[12:44] soveran has joined the channel
[12:46] kendes has joined the channel
[13:46] webben has joined the channel
[13:50] mdg has joined the channel
[14:07] voodootikigod_ has joined the channel
[14:08] voodootikigod_: morning
[14:20] jan____: mr tikigod
[14:21] voodootikigod_: anyone know why the window.btoa function is not available in v8?
[14:21] voodootikigod_: jan____: how goes it
[14:23] jan____: voodootikigod_: trying to find a good coffee maker.
[14:23] jan____: think I settled on the gaggia calssic
[14:23] voodootikigod_: i see
[14:23] jan____: classic
[14:23] jan____: coworking needs coffee
[14:23] voodootikigod_: tough decisions in life eh?
[14:25] jan____: Don't get me started :)
[14:26] voodootikigod_: hehe
[14:26] voodootikigod_: i got basic auth support for http client working
[14:26] voodootikigod_: and added a bunch of crypto to node
[14:26] voodootikigod_: going to extrapolate out to a node_ext package
[14:30] sudoer has joined the channel
[14:33] ashb: voodootikigod_: the whole window object isn't available is why :)
[14:33] voodootikigod_: but normally things like atob would be available in v8
[14:33] voodootikigod_: IE not normally a dom engine thing, more of a js thing
[14:34] ashb: sure - its on the window object which means it comes from the browser, not the JS engine tho
[14:45] jan____: voodootikigod_: cools
[14:45] jan____: ashb: window is the global object in the browser
[14:45] voodootikigod_: ashb: cool will just continue to use the javascript version
[14:45] jan____: so window.foo could be a JS global or a browser global
[14:46] ashb: jan____: yes. but odds are anything on it apart from Date, Array etc are from the browser not the JS engine
[14:47] frodenius: every js coder should know the difference between js primordials and host objects
[14:47] voodootikigod_: its all good
[14:47] voodootikigod_: frodenius: but which lies where is a a case by case basis
[14:47] voodootikigod_: frodenius: spidermonkey has atob
[14:48] frodenius: well, not really, the things you can 100% rely on are the objects outlined in the language spec
[14:49] dims has joined the channel
[14:49] voodootikigod_: true, but then again the things outlined in the spec 1) arent fully implemented in v8 and 2) are in v8 but not in the spec ;)
[14:49] frodenius: heh
[14:50] voodootikigod_: anyways
[14:50] voodootikigod_: for now i will just use my javascript crypto lib
[14:50] voodootikigod_: and eventually move it down into the c level 
[14:50] voodootikigod_: where it properly belongs
[14:51] sixtus42: voodootikigod: are you aware of http://github.com/waveto/node-crypto ?
[14:51] voodootikigod_: was not
[14:52] sixtus42: voodootikigod: same guy bringing tls to node
[14:52] mediacoder: voodootikigod_: check out github...there is already a c node crypto lib
[14:52] mediacoder: exactly
[14:52] voodootikigod_: yea
[14:52] voodootikigod_: excellent
[14:52] mediacoder: he will add some more ciphers soon
[14:53] voodootikigod_: any reason why base 64 was left out
[14:53] voodootikigod_: not crypto
[14:53] voodootikigod_: (at least not worht anything)
[14:53] voodootikigod_: but necessary for basic auth
[14:53] mediacoder: hmmm
[14:53] mediacoder: should be there
[14:54] mediacoder: atleast you can b64 encode encryptions
[14:54] mediacoder: so guess "plain basse64" should be also there
[14:57] voodootikigod_: hmmm
[14:58] adam___ has joined the channel
[14:58] voodootikigod_: can't seem to get the module to load
[14:58] voodootikigod_: will play with it later
[14:58] voodootikigod_: thanks for the pointer though
[14:59] cloudhead has joined the channel
[15:04] kendes has joined the channel
[15:05] sixtus42_ has joined the channel
[15:06] emyller has joined the channel
[15:13] ryah: morning
[15:15] Nailor: Evening :p
[15:18] JoePeck has joined the channel
[15:24] ryah: JoePeck: you're from rochester - cool
[15:25] ryah: my favorite american city
[15:28] Biscuits: ryah: Quick question, if your aim was to support multiple OSs, why did you make the IO stuff such a direct POSIX map ? I'm pretty sure Win32 has poor POSIX support :/
[15:29] ashb: it has equivalent functionality
[15:30] ashb: in most cases
[15:30] ashb: infact windows has working natvie overlapped (i.e. async) IO by all accounts
[15:30] Biscuits: Hmm, CreateFile has that as far as I know
[15:30] Biscuits: But I've also read that there's no real stat equivalent in win32 :/
[15:31] ryah: all of the posix stuff works on windows afaik
[15:31] ashb: do do what? get file size or the other stuff?
[15:31] ashb: file size there most certainly is
[15:31] ryah: "posix" is not the best name though
[15:31] ryah: maybe "fs" or "sys" or something would be better
[15:32] ryah: i hate having all these modules.
[15:32] ashb: fs is what commonJS went with
[15:32] ryah: yeah? maybe i should hcange to that
[15:33] ashb: well right now there is a fs-base which is low level non pretty api cos we couldn't decide on a nicer DWIMy form
[15:33] Biscuits: Anyway, let's see if anything from Node.js wants to compile on win32 :p
[15:33] ashb: Biscuits: good luck :)
[15:33] ashb: i know for a fact it wont without some changes
[15:33] ashb: dlopen etc at the very least
[15:34] ryah: yeah, its going to take some work
[15:34] ashb: also msvc or gcc (mingw)?
[15:34] ryah: would like to do msvc
[15:34] Biscuits: I'd prefer MSVC :/
[15:34] ashb: tho i guess node itself doesn't use any/much template stuff
[15:35] ryah: since i guess thats what windows people use
[15:35] ashb: depends who you are aiming at
[15:35] ashb: if its just 'getting it running on Win for JS devs' then most wouldn't even want to compile it themeselves
[15:35] ryah: Biscuits: getting it to compile in cygwin (is that what it's called?) probably will be easy
[15:35] Biscuits: I'm pretty sure V8 on Windows compiles with MSVC
[15:36] ryah: it does
[15:36] ashb: cywgin/msys always feels like a bit of a cop-out to me.
[15:37] ryah: yeah, need to compile for real at some point
[15:37] ryah: but as a stop-gap it will work
[15:43] Biscuits: hmm, udns seems to have a win32 port too
[15:43] ryah: yeah - going to remove udns real soon
[15:43] Biscuits: ?
[15:44] ryah: well in the next two weeks
[15:45] Biscuits: Why? and what's going to be the replacement ?
[15:45] ashb: replacing it with?
[15:46] ryah: replacing it with c-ares
[15:46] ryah: because it doesn't do /etc/hosts resolutions
[15:46] Biscuits: udns doesn't do hosts ?
[15:47] ashb: does c-ares support windows hosts file (yes there is one somewhere)?
[15:47] ryah: probably - it's a pretty heavy library
[15:48] ryah: udns is a lot more elegant - but i really need /etc/hosts resolution
[15:48] Biscuits: well, c-ares has proper win32 support, so that seems to be fine :)
[15:50] CIA-28: node: 03jed 07master * ra9ea21f 10/ lib/http.js : fixed check on header type, to check member instead of object. - http://bit.ly/4JW04D
[15:50] CIA-28: node: 03Rhys Jones 07master * r5b1a535 10/ (4 files in 3 dirs): Add HTTP client TLS support - http://bit.ly/5tcxBl
[15:51] ryah: i can't wait to get back to germany and have some time to hack
[15:52] JoePeck: ryah: you know Rochester? Did you go to school here?
[15:53] JoePeck: ryah: also, nice presentation as jsconf. I'm a fan of the evented IO approach, and the presentation is a great resource I can use to show other people its advantages
[15:53] ryah: JoePeck: i lived there for a long time. its great. where do you live?
[15:54] JoePeck: ryah: Rochester New York. I actually live in East Rochester
[15:54] ryah: JoePeck: do you know rocwiki? or ant hill cooperative?
[15:54] JoePeck: ryah: I know rocwiki =)
[15:54] ryah: i started that with some friends :)
[15:55] JoePeck: haha, very cool =)
[15:55] ryah: http://rocwiki.org/Wiki_Community/History :)
[15:56] Biscuits: ryah: Just a quick question on the porting, why did you opt to put the dns module directly in Node.js ? why not just make it a seperate module ?
[15:57] rakeshpai: How easy is to talk to say something written in C from JS in node?
[15:57] voodootikigod_: !log
[15:57] JoePeck: ryah: neat!
[15:57] Biscuits: rakeshpai: If it's a console application: really easy
[15:57] Wes-: holy smokes
[15:58] ryah: Biscuits: i'm going to use it for, like, resolving hosts for tcp clients
[15:58] rakeshpai: Biscuits: by using the available apis for stdin and stdout, you mean?
[15:58] Wes-: Rochester is just across the lake and down the road 60 miles from here
[15:58] Biscuits: rakeshpai: process.createChildProcess should do the trick. Just launches another application and pipes output and error to your own functions, and allows you to write stuff to it.
[15:59] rakeshpai: Biscuits: cool... will investigate
[15:59] Wes-: ryah: does c-ares do the whole nsswitch.conf thing?
[15:59] ryah: Wes-: i don't know
[16:00] Wes-: ryah: FWIW, that's pretty key around here, 'cause we also use ldap, nis+ and nis besides dns and files. But we're unusual.
[16:00] Wes-: does it do resolution from primitives, or does it somehow wrap libresolv?
[16:01] ryah: it reimplements the resolver
[16:05] sztanpet has joined the channel
[16:06] kendes has joined the channel
[16:08] Biscuits: http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#WIN32_PLATFORM_LIMITATIONS_AND_WORKA
[16:08] Biscuits: Hmmm
[16:10] ryah: i think libev support on windows is pretty good. 
[16:10] Biscuits: I'm beginning to wonder if it wouldn't be easier just to re-implement parts of Node.js for Win32 instead of porting it :/
[16:10] Biscuits: That's the libev doc that states otherwise :/
[16:10] Biscuits: directly from their CVS repos
[16:10] ryah: no, it states that windows sucks
[16:11] Biscuits: ah, ok
[16:11] Biscuits: Yeah, sorry, think I slightly misread that :/
[16:12] ryah: marc lehmann hates windows but tries very hard to support it as much as possible
[16:12] ryah: i think libeio is similar
[16:12] Biscuits: I still wonder if a simple standard Win32 event/message loop, together with a simple message window and the Winsock built-in asynchronous, and the event-message-to-window stuff would be better on windows :/
[16:13] ryah: node is heavily tied to libev
[16:13] Biscuits: I mean the whole map file descriptors to sockets so select can take em seems like a case of trying to get a certain model working on a totally different model...
[16:14] ashb: libeio is files or sockets too?
[16:14] Biscuits: I think libeio is for the files
[16:14] ryah: yeah
[16:14] ashb: windows has its own builtin async file IO calls
[16:15] mikekelly: ryah: was that twitter message earlier a deliberate bright eyes reference?
[16:15] ryah: mikekelly: embarrassingly, yes
[16:15] mikekelly: haha, cool album :)
[16:15] ryah: ashb: yeah - it would be cool if libeio compiled to aio on linux and windows aio on windows..
[16:16] ashb: yeah
[16:16] ryah: and GCD (or whatever) on mac
[16:16] ashb: GCD isn't for aio
[16:16] ashb: its just a thread pool i thought?
[16:17] ryah: well, maybe it's a better thread pool than libeio's?
[16:18] ashb: probably yeah
[16:31] nefD has joined the channel
[16:32] emyller has joined the channel
[16:42] Wes-: I don't know squat about libeio, but if you're going to do libaio on linux, you should do libaio everywhere that is POSIX / SUSv3
[16:42] Wes-: (which pretty much means everywhere but windows)
[16:43] ryah: does mac support libaio?
[16:43] ryah: afaik it doesn't
[16:44] Wes-: ryah: Really? I didn't think you could get a SUSv3 stamp without aio.h
[16:44] Wes-: Lemme dig
[16:47] Wes-: Hmm, my Leopard box as an aio man page, says it scooped aio from FreeBSD3.   Has aio_cancel, _error, _read, _return, _suspend, _write and lio_listio
[16:47] Wes-: I'm not sure if that's enough for the POSIX AIO tag or not
[16:50] Wes-: Yeah, aio_fsync man page is missing, but other than that, seems to be okay against my POSIX realtime extensions list for AIO.   There *is* an aio_fsync in /usr/include/sys/aio.h, though.
[16:52] Wes-: So, ryah, I'm pretty sure you can count on aio.h virtually everywhere meaningful but Windows.  You might need to throw appropriate #defines to request appropriate levels of POSIX, though.
[16:53] bentomas has joined the channel
[16:54] Wes-: *educated guessing* - _POSIX_SOURCE and _POSIX_C_SOURCE=200112L
[16:54] Wes-: May also want _XOPEN_SOURCE definitions for SVR4-derived unices, e.g. Solaris
[16:55] Wes-: Latest _XOPEN_SOURCE=600, but 600 implies C99, which will generate compiler warnings or errors (depending on compiler and pedantry) with C++ source code
[16:55] Wes-: 500 is probably enough for AIO
[16:55] Wes-: alternatively you can sequester AIO code in C modules and build then as C99 with extern C linkage from your V8-using code
[17:05] RayMorgan has joined the channel
[17:05] johan-s has joined the channel
[17:10] ryah: Wes-: ah, good to know
[17:10] ryah: i wonder if aio just uses pthreads internally or if it has lower level hooks
[17:11] Wes-: ryah: That's probably implementation-specific, however, I am fairly confident that at least on SVR4-descendents (e.g. Solaris) it is much lower level
[17:11] Wes-: IIRC Oracle as of version 8 uses aio for low-level access when the backing store is a file system
[17:12] Wes-: historically they have not been shy of poking deep into kernel, esp on primary platforms like sun
[17:18] ryah: Wes-: http://lists.schmorp.de/pipermail/libev/2008q2/000335.html
[17:20] Wes-: ryah: Wow, thanks for that! -- FWIW threads on Solaris would be solaris-threads, not posix-threads, so slightly cheaper, but they are still LWPs
[17:21] emyller has joined the channel
[17:29] aguynamedben has joined the channel
[17:34] richtaur has joined the channel
[18:11] isaacs has joined the channel
[18:40] erikg has joined the channel
[18:50] cruxst has joined the channel
[18:50] isaacs: ryah: i'm curious, do you have a job?
[19:03] orlandov: what's the difference between libeio and libev?
[19:04] kriskowal has joined the channel
[19:05] ashb: libev is generaly event framework (sockets, timeouts etc,) libeio is evented local file IO
[19:05] ashb: at least thats my understanding
[19:08] orlandov: ashb: that's what i thought, but aren't they both just file descriptors?
[19:08] orlandov: or is there some more subtlety there
[19:09] ashb: orlandov: you can't select on files typically
[19:09] orlandov: o rly
[19:09] orlandov: fair enough :)
[19:09] ashb: or select doesn't return sensible values (yes you can always read or some sucgh)
[19:09] orlandov: right, i always assumed it was the same but that figures
[19:23] the_undefined_ has joined the channel
[19:24] the_undefined_: wow 64 people
[19:24] the_undefined_: is there a party tonight?
[19:24] the_undefined_: I should get a beer, just in case :)
[19:25] jan____: or a case of beer.
[19:27] inimino: ryah: about setting HTTP headers, how about just having a response.headers object that you can modify at will until the first sendBody() call?
[19:27] ashb: what does node do if you get an exception thrown after you've sent status/headers?
[19:29] inimino: it just dies
[19:29] ashb: any indication in the output stream? stderr?
[19:29] ashb: (we'just just wondering what to spec for async for JSGI in this case)
[19:29] inimino: ashb: if you don't catch it
[19:29] inimino: yes, you should get a stack trace on stderr from v8
[19:30] ashb: anything sent to client?
[19:30] inimino: nothing special, the server just gets killed
[19:31] inimino: (so the client would see that the connection was closed but that's it)
[19:32] ashb: k
[19:32] nefD: hmm.. from everything i've seen so far, node would likely be pretty well suited for browser based games.. ..!
[19:32] inimino: of course if somebody was writing JSGI support for node they could always wrap the request handler in try/catch and send a nice 500 page
[19:32] inimino: nefD: sure!
[19:33] ashb: inimino: not for the async promise version :)
[19:33] nefD: I especially like that the client side code isn't really having to do anything special to interface with it.. standard jQ/proto/dojo ajax calls appear to be sufficient
[19:33] inimino: ashb: oh
[19:34] inimino: 
ACTION shifts perspective

[19:34] deric_skibotn has joined the channel
[19:35] inimino: ashb: well, s/wrap in try\/catch/install some global error handler of last resort, which probably isn't implemented yet/
[19:36] n8o has joined the channel
[19:36] inimino: ashb: is the async JSGI stuff written up anywhere yet?
[19:36] ashb: commonJS mailing list
[19:37] ashb: there's an example, but it doesn't show a streaming case
[19:37] inimino: ok, yeah I skimmed those emails
[19:37] ashb: (yet, kriszyp is going to post a new example that does)
[19:38] orlandov: ashb: are you talking about the jfgi draft proposal thread?
[19:39] orlandov: oops.. jsgi
[19:39] ashb: orlandov: yes and a chat we're having in another irc channel
[19:39] orlandov: for this scgi server i hacked together i'm targetting a jsgi interface, and it'd be cool to see how that stuff should be handled (async input etc)
[19:39] inimino: "JSGI Level 0/A Draft 2 Proposal on the wiki..." thread
[19:40] orlandov: ya
[19:49] Connorhd has joined the channel
[19:53] rakeshpai has joined the channel
[19:58] deric_skibotn has left the channel
[20:00] Connorhd_ has joined the channel
[20:02] sixtus42 has joined the channel
[20:06] rakeshpai_ has joined the channel
[20:06] path[l] has joined the channel
[20:12] jed has joined the channel
[20:15] sudocarl has joined the channel
[20:15] sudocarl: morning everyone :)
[20:16] path[l]: hello :), just watching the js conf video
[20:17] sudocarl: cool :)
[20:28] orlandov: is there a magic incantation to building the node_postgres c++ project? node-waf configure build
[20:28] orlandov: throws up a bunch of errors
[20:28] sudocarl: can you put your errors up?
[20:28] sudocarl: make a gist or pastie?
[20:28] orlandov: sudocarl: it's not finding the includes
[20:28] orlandov: sec
[20:28] sudocarl: did node install fine?
[20:29] sudocarl: and are you using the latest version?
[20:30] orlandov: newest as of friday i guess
[20:30] sudocarl: kk
[20:30] orlandov: from github
[20:30] orlandov: http://pastebin.com/d6b370453
[20:31] sudocarl: huh weird
[20:31] sudocarl: and your node installed fine?
[20:31] sudocarl: can you run node?
[20:31] orlandov: yep
[20:31] orlandov: i am running waf-build like: ../node/tools/waf-light configure build
[20:31] orlandov: not sure if that's correct
[20:32] sudocarl: did you try to just run "node-waf configure build"?
[20:32] orlandov: and i see a node dir in /usr/local/include
[20:32] sudocarl: node-waf should be installed in your bin
[20:32] orlandov: oh shit, it is too
[20:32] orlandov: i'm retarded :P
[20:32] sudocarl: :)
[20:33] sudocarl: no worries hope it fixes the problem
[20:33] orlandov: thanks man...
[20:33] orlandov: yeah it compiles now
[20:33] sudocarl: :D
[20:33] voodootikigod_ has joined the channel
[20:33] orlandov: yay id10t error :)
[20:33] path[l]: lol
[20:35] the_undefined_: got video file uploading from iPhone to node.js working, good stuff :)
[20:39] quirkey has joined the channel
[20:40] sudocarl: orlandov, just to let you know I had to make some minor changes to that module to get it to work: 
[20:40] sudocarl: http://github.com/carlsverre/node_postgres
[20:40] sudocarl: incase your's doesn't work feel free to use mine
[20:40] sudocarl: yours*
[20:40] orlandov: sudocarl: cool, i was actually just using it to figure out how C++ extensions are structured
[20:40] sudocarl: I will be putting a bunch of work into it hopefully later today
[20:40] sudocarl: ohh :D sounds good
[20:41] orlandov: rad! there could definately be better rdbms support in node :)
[20:44] sudocarl: indeed
[20:44] sudocarl: :)
[20:44] okito has joined the channel
[20:44] sudocarl: I am an idiot because I decided that since I love node so much I want to do a project which requires an rdbms on it...
[20:44] sudocarl: oh well :)
[20:44] sudocarl: gotta flex my c++ skillz sometime
[20:44] sudocarl: (they are lacking)
[20:44] orlandov: sudocarl: hah i totally hear ya
[20:45] orlandov: why must i love things i can't have
[20:45] sudocarl: haha
[20:45] sudocarl: so far I have spent more time working on other node modules which my project needs than actually working on the project.... :(
[20:45] orlandov: i'm trying to learn how reuse the node_net stuff do implement wire protocols in C++
[20:45] sudocarl: oh well :)
[20:45] sudocarl: oh nifty
[20:45] orlandov: but yeah, C++ is kind of a headfuck after 10 years of disuse
[20:46] sudocarl: no kidding
[20:46] sudocarl: I can't stand it
[20:46] sudocarl: I script too much :)
[20:46] ashb: could be worse - it could make heavy use of templates :)
[20:46] sudocarl: haha
[20:46] orlandov: yeah, same
[20:46] orlandov: oh god
[20:46] sudocarl: templates make me cry
[20:46] sudocarl: I prefer brainfuck to c++
[20:46] sudocarl: http://en.wikipedia.org/wiki/Brainfuck
[20:46] orlandov: see, i like my sanity
[20:46] sudocarl: haha
[20:48] the_undefined_: sudocarl: My approach with node has been to run a persistent pool of php/ruby/whatever processes and communicate with them over stdio for stuff node does not support yet
[20:48] sudocarl: http://99-bottles-of-beer.net/language-c++-108.html versus: http://99-bottles-of-beer.net/language-brainfuck-101.html
[20:48] sudocarl: its almost as bad :D
[20:48] sudocarl: the_undefined_, thats not a bad idea :)
[20:49] the_undefined_: it's not pretty, but it lets me write the APIs I want, make them look async and replace them with better stuff as it is being developed
[20:49] the_undefined_: I think I should make that a general module
[20:49] sudocarl: the_undefined_, but considering that node needs developers contributing to the core modules I don't mind spending my time working on the
[20:49] sudocarl: them
[20:49] sudocarl: the_undefined_, a module like that would be very nice though
[20:50] the_undefined_: sudocarl: oh, I contribute lots of stuff too. But some things like writing a mysql connector is outside my abilities and time available so I work on stuff I can finish in reasonable amounts of time instead
[20:50] the_undefined_: I really admire people like rhys who contributed the SSL stuff
[20:50] sudocarl: the_undefined_, fair enough :
[20:50] the_undefined_: tons of work
[20:50] sudocarl: :)
[20:50] sudocarl: no kidding
[20:50] sudocarl: luckily ry has already put together a pretty good postgre connector
[20:50] sudocarl: otherwise I would probably do what your doing
[20:51] the_undefined_: hehe
[20:51] the_undefined_: well I use MySql, so no love for me : )
[20:51] the_undefined_: MySql sucks, but postgres sucks harder
[20:51] the_undefined_: ;)
[20:51] orlandov: oh no you didn't!
[20:52] the_undefined_: orlandov: AFAIK it has no support for full text search
[20:52] sudocarl: really?
[20:52] the_undefined_: which, yes, your RDBM shouldn't do - but man is it useful to not add something new to your stack when you need it
[20:52] sudocarl: haha fair enough
[20:52] the_undefined_: I could be wrong
[20:52] ashb: the_undefined_: mysql isn't a database :P
[20:52] path[l]: I have a question. Im watching the node.js video and Im confused about something. If I've built up a heavy session worth of data and then Im going to get stuff from a database and I want a callback, I have to send all my data to the callback to it can resume functioning with all that data. How does node perform when stuff like this happens
[20:52] sudocarl: haha ashb
[20:52] sudocarl: lulz
[20:52] the_undefined_: I just talked to some people who use postgres in big enterprise stuff
[20:53] ashb: sudocarl: its a nice fast key-value store, but it aint a database
[20:53] the_undefined_: and they told me they were longing for mysql after a while even so its worse on paper :)
[20:53] ashb: the_undefined_: did you tell them they are doing it wrong?
[20:53] ashb: cos they really must be
[20:53] sudocarl: path[l], so basically when you make a callback it can access any variables in the defining scope...
[20:53] the_undefined_: well, doing it wrong can be tons of fun and exactly what you want in the beginning of building a system ...
[20:53] cmlenz has joined the channel
[20:53] sudocarl: so there is no copying it just continues execution as if it was executing directly in the scope
[20:53] sudocarl: atleast I think thats what happens
[20:54] path[l]: so where does all this get persisted? 
[20:54] path[l]: when the process exits
[20:54] sudocarl: the process doesn't exit
[20:54] sudocarl: node.js runs constantly
[20:54] sudocarl: so data is persisted across requests
[20:55] the_undefined_: ashb: anyway, they might have been idiots for all I know - so I'm not going to argue against postgres if smarter people are having a good time with it :)
[20:55] ashb: the_undefined_: it certinaly works when it comes to things like VIEW etc.
[20:55] sudocarl: the_undefined_, the thing I love about postgres is stored procedures...  my company essentially has an entire game running in stored procedures which makes it insanely fast
[20:55] ashb: and it has a non-brain dead query optimizier
[20:55] sudocarl: we only ever transfer about 1 line of code to the DB
[20:55] the_undefined_: sudocarl: I guess I hate stored procedures
[20:56] sudocarl: and our php just acts like a middleman
[20:56] sudocarl: the_undefined_, now thats a fair thing...  half the people I talk to love them and the other half hate them
[20:56] path[l]: hmm ok Im looking at the webserver example. I can service multiple requests, each of them might hit the db and it has a callback for each. Until the callback is called, all the data for each request gets persisted in memory then?
[20:56] the_undefined_: I love the idea, but hate working with it
[20:56] sudocarl: but if you want some serious DB speed I havn't seen anything like it
[20:56] sudocarl: path[l], yes
[20:56] sudocarl: the_undefined_, :D
[20:57] the_undefined_: sudocarl: fair enough, but I see it as a way to optimize - not to design your apps upon
[20:57] path[l]: so then I assume if you want to build a webserver with a loadbalancer you need some sticky sessions like thing? 
[20:57] the_undefined_: I will go crazy if I need to join a team working with a db full of stored procedures
[20:57] the_undefined_: :)
[20:57] path[l]: or no I guess not
[20:57] path[l]: only the callback has to be on the same machine
[20:57] sudocarl: path[l], yea...
[20:58] sudocarl: path[l], I assume that it gets more complex as the app gets bigger but since node can act like a local cache and the callbacks will maintain the data i think it should be fine under a load balancer
[20:58] orlandov: anyone know of other pedagogical explamples of writing Node extensions in C++? the postgres example ain't doing it for me
[20:58] path[l]: hmm ok
[20:58] sudocarl: the_undefined_, I think I would too...  I am a flash guy not a sql guy :D
[20:59] the_undefined_: sudocarl: flash, huh : ). What type of stuff do you work on / use node for?
[20:59] sudocarl: games :)
[20:59] the_undefined_: sudocarl: LINK :)
[20:59] sudocarl: I work for a video game company called relevant games
[20:59] sudocarl: http://www.saveourfrogs.com
[20:59] the_undefined_: sudocarl: "We donate 100% of the money earned from advertising to charity" -> how do they pay you?
[21:00] sudocarl: its not that great
[21:00] sudocarl: :D
[21:00] sudocarl: haha 1 sec
[21:00] path[l]: hmm there's a node.js irc server too?
[21:00] sudocarl: so basically we donate 100% of ADVERTISING profits
[21:00] sudocarl: which comes out to be about...  10% of our profits
[21:00] sudocarl: we do micropayments
[21:01] sudocarl: I wasn't supposed to say that
[21:01] sudocarl: but your not my target demographic
[21:01] sudocarl: :D
[21:01] mediacoder: say what? ;-)
[21:01] sudocarl: nuttin :)
[21:01] sudocarl: haha
[21:01] sudocarl: eh its still better than what most companies do
[21:01] path[l]: lol
[21:02] sudocarl: funny enough our partner for our game savethefrogs (the savethefrogs foundation) has found that his donations have gone up by over 100% since we partnered with him
[21:02] sudocarl: so it is working out :)
[21:02] mediacoder: path[l]: yea, there is a proof-of-concept irc server in node, which ry used to demo node
[21:02] the_undefined_: sudocarl: nice
[21:03] sudocarl: but savethefrogs was sorta a test project for our company
[21:03] sudocarl: our next game I am hoping will be alot more fun
[21:03] sudocarl: :)
[21:03] sudocarl: but its underwraps for now :)
[21:03] the_undefined_: sudocarl: any plans on using node?
[21:04] mediacoder: sudocarl: and already running productive on node?
[21:04] sudocarl: I would love to :)
[21:04] sudocarl: not yet
[21:04] mediacoder: ah :-)
[21:04] sudocarl: I am getting into node as of last week
[21:04] sudocarl: its alot of fun but I need to ensure that its super stable
[21:04] sudocarl: right now were running on nginx + postgres
[21:04] sudocarl: php middleman + stored procedures backend
[21:04] sudocarl: also a apache server to serve alot of our misc stuff
[21:05] ashb: why bother with the apache?
[21:05] sudocarl: 99% of our page load time is querying FB
[21:05] sudocarl: I think its cause we started with apache and our svn server is linked to it
[21:05] sudocarl: I dunno I didn't handle that side of things
[21:05] pdelgallego has joined the channel
[21:05] sudocarl: its not serving very much so we could probably switch it
[21:05] sudocarl: we just don't care lol
[21:05] mediacoder: sudocarl: thats a socket-server based game on php-sockets?
[21:05] sudocarl: no
[21:06] rtomayko has joined the channel
[21:06] sudocarl: its currently post request based
[21:06] sudocarl: our next game/games will all be socket based
[21:06] mediacoder: just started a simple socket-server and policy server for flash :-D
[21:06] sudocarl: :) fun eh
[21:06] mediacoder: yea..especially with node
[21:07] sudocarl: I am writing a pretty fully-featured game framework for flash which has a powerful socket client in it
[21:07] sudocarl: it will power all of relevant game's games
[21:07] sudocarl: and one of the other developers is writing a low level socket server in c
[21:07] mediacoder: you goind to release it at some point? or just internal project?
[21:07] sudocarl: for now its internal
[21:07] sudocarl: but I want to push for release
[21:07] path[l]: hmm the video was good, now I wanna try to build something
[21:08] sudocarl: path[l], hello world!!
[21:08] sudocarl: :)
[21:08] path[l]: lol maybe
[21:08] mediacoder: SPDY implementation would be great
[21:08] path[l]: gtg to bed now, gonna install tomorrow
[21:08] mediacoder: :-)
[21:08] sudocarl: path[l], night :)
[21:08] sudocarl: brb
[21:08] path[l]: night
[21:09] sudocarl: fuck!!  I just downloaded a iso for windows 7 and its in hebrew
[21:09] sudocarl: V_VV
[21:09] ashb: serves you right ;)
[21:09] sudocarl: its a non-cracked iso
[21:09] sudocarl: I have a key for it
[21:09] sudocarl: legal
[21:09] sudocarl: I just need a iso
[21:09] sudocarl: for vmware
[21:10] ashb: dont think that makes it any more legal mind
[21:10] sudocarl: haha I think it does?
[21:10] sudocarl: I bought the key
[21:10] sudocarl: I just don't have the cd
[21:10] sudocarl: whats not legal about that?
[21:10] malkomalko has joined the channel
[21:10] the_undefined_: sudocarl: paying for windows is like buying stocks in oil companies ... except YOU will not become rich :)
[21:11] ashb: if you care read the EULA i guess
[21:11] sudocarl: hahaha sounds good
[21:11] sudocarl: the_undefined_, I totally agree :)
[21:11] sudocarl: I run 99% linux
[21:11] sudocarl: just need windows for a couple things
[21:11] sudocarl: like adobe
[21:11] sudocarl: <3 adobe
[21:15] Atmoz: 
ACTION runs photoshop in virtualbox, but must reboot to play games :/ Windows dies the day gaming companies make games work on linux.

[21:15] sudocarl: :) so true!
[21:17] sztanpet has joined the channel
[21:22] rakeshpai has joined the channel
[21:27] rakeshpai has joined the channel
[21:29] LordMetroid has joined the channel
[21:29] LordMetroid: Wow, node.js is really easy
[21:30] mediacoder: and awesome
[21:31] inimino: and fun!
[21:31] LordMetroid: I only compiled and copy pasted the hello world example but I don't think I have been up and running a webserver in such a short amount of time before
[21:32] LordMetroid: Now I just need to start writing something useful
[21:32] coolguy_: ...
[21:32] coolguy_: PROFIT!!
[21:33] LordMetroid: :)
[21:33] LordMetroid: Is it feasible to write a whole e-commerce site using node.js?
[21:33] sudocarl: LordMetroid, yes!
[21:34] sudocarl: LordMetroid, do it!
[21:34] LordMetroid: I will
[21:34] sudocarl: LordMetroid, good.
[21:34] sudocarl: [+] Made someone do something.  done.
[21:34] LordMetroid: lol
[21:34] sudocarl: time for a nap
[21:34] sudocarl: lol
[21:35] orlandov: i see there's another outliner user :)
[21:36] sudocarl: actually I don't but I am intrigued :)
[21:36] sudocarl: I use that format for my personal time tracking application
[21:36] inimino: 
ACTION uses ✓

[21:36] sudocarl: inimino, yay old school!
[21:37] sudocarl: er I am assuming you mean pen and paper
[21:37] sudocarl: but I could be wrong
[21:37] sudocarl: lots of systems use unicode names nowadays
[21:37] inimino: no, I mean Unicode text files
[21:37] inimino: in a browser window, as it so happens
[21:37] sudocarl: nice :)
[21:38] sudocarl: using your nifty revision system?
[21:38] inimino: yeah
[21:38] sudocarl: wow its multipurpose
[21:38] sudocarl: is it running on node?
[21:38] sudocarl: I just assumed it did I never asked lol
[21:38] inimino: it's like my filesystem :)
[21:38] inimino: it is now!
[21:38] sudocarl: sweet :)
[21:38] sudocarl: have you noticed a speedup/slowdown?
[21:38] inimino: the previous version was prototyped in Apache with bash CGI :-)
[21:39] sudocarl: ouch
[21:39] skampler: hohoho
[21:39] skampler: now that's an improvement
[21:39] inimino: it only had one user, so I never noticed the speed one way or the other
[21:39] inimino: everything is just instant when you have one user per VPS, no matter how inefficient the code
[21:40] sudocarl: haha lol
[21:40] sudocarl: so true
[21:40] sudocarl: where are you running it?
[21:40] inimino: old one is running on slicehost, new one on linode
[21:41] okito has left the channel
[21:41] sudocarl: oo how are you liking linode?
[21:41] sudocarl: thats in cali right?
[21:41] inimino: it's very nice so far
[21:41] mediacoder: 
ACTION wishes eiterone accepted paypal :-/

[21:41] inimino: they have a few datacenters, I think mine might be in Dallas
[21:41] sudocarl: oh cool
[21:45] sudocarl: inimino, do all linode instances have the same amount of ram/cpu?
[21:45] inimino: I think they have different plans
[21:45] inimino: I think mine is whatever they offer for $20/mo
[21:46] sudocarl: ah I see
[21:46] sudocarl: so you have 360mb of rab
[21:46] sudocarl: ram
[21:46] inimino: I'm using 32bit Arch
[21:46] sudocarl: thats pretty impressive actually
[21:46] sudocarl: for only 20$
[21:47] coolguy_: linode is fantastic
[21:47] coolguy_: great customer service and cheap
[21:47] sudocarl: has anyone hosted with webfaction?
[21:51] nefD: i used to use linode.. i highly recommend them
[21:51] nefD: slicehost is another, similar provider, but I haven't tried them
[21:53] inimino: both are good in my experience
[21:53] inimino: linode has a nicer control UI IMO
[21:56] binary42 has joined the channel
[21:58] sudocarl: and a ping of 58 ms isn't too bad...
[21:59] sudocarl: in comparison I have a shared server also hosted in dallas and I am gettign 103 ms on it
[21:59] sudocarl: V_V
[21:59] sudocarl: silly shared
[21:59] sudocarl: need to upgrade sometime
[21:59] nefD: when I went from shared hosting to vps hosting, I got supremely spoiled by having root level ssh access to it.. I can't imagine going back to shared, now
[22:00] sudocarl: sounds about right
[22:01] sudocarl: well I have a homeserver which I use for 99% of my stuff
[22:01] sudocarl: only my website is hosted in dallas
[22:01] sudocarl: so its not too bad
[22:01] nefD: *nod*
[22:01] sudocarl: but I totally understand cause whenever I have to login to it and mess with something its a headache
[22:01] nefD: i'd do the same, but the lack of a dedicated ip and my wife wanting to be in windows all the time makes a hosted solution more suited to my needs :P
[22:01] inimino: I went from dedicated to VPS for development
[22:02] sudocarl: inimino, interesting
[22:02] sudocarl: do you develop on your vps?
[22:02] sudocarl: or locally?
[22:02] inimino: VPS is quite a bit better fit for me
[22:02] sudocarl: I would kill if there was any lag in keystrokes
[22:02] sudocarl: even 50ms response is too much
[22:02] sudocarl: my homeserver is much nicer with a .037ms response time :)
[22:02] inimino: sudocarl: I always do everything remotely
[22:03] sudocarl: preferred editor?
[22:03] nefD: personally, I prefer either vim'ing from an ssh (its really not too bad) or mounting my server locally over ssh and editing on my own box
[22:03] inimino: vim
[22:03] sudocarl: nice me too
[22:03] sudocarl: hail to zee vim
[22:03] nefD: oh yea, vim pwns
[22:03] sudocarl: I am running windows right now with 3 xming windows running terminal sshed into my homeserver
[22:03] sudocarl: :) its my way of getting linux on windows
[22:04] nefD: get taglist+ctags and nerdtree, plus some tinkering with the mouse and tabbed interface, you get a suweet lightweight editor
[22:04] sudocarl: yuppers
[22:04] sudocarl: my vimrc file has grown disgustingly large though
[22:04] nefD: i've been on this neverending quest to find *something* i can tolerate editing with in windows.. suffice to say, I haven't found it yet
[22:04] sudocarl: I like putting in f5/f6 commands for every language haha
[22:04] sudocarl: notepad++ is all I can find
[22:04] nefD: haha yea
[22:04] sudocarl: and even then it has its quirks
[22:05] sudocarl: although the one area where windows shines is in as3 development
[22:05] sudocarl: I really love flashdevelop
[22:05] nefD: i can remember when ultraedit wasn't so ginormous and served the purpose pretty well
[22:05] nefD: ahh yea
[22:05] nefD: ive done a bit of as3, but i spent most of my time messing with haxe
[22:05] sudocarl: haha haxe is pretty fun
[22:05] nefD: indeed!
[22:06] sudocarl: but once you get to know (and love) as3 there is really no comparison
[22:06] sudocarl: my three fav languages right now are...  python, as3 and probably js
[22:06] sudocarl: either js or sml
[22:06] sudocarl: sml because I am sadistic 
[22:06] nefD: honestly though, it was my only real option for developing flash in linux
[22:06] sudocarl: thats true
[22:06] mediacoder: that has atleast changed much
[22:06] sudocarl: flash linux support sorta sucks
[22:06] sudocarl: which is sad
[22:07] nefD: the client isn't too bad, just a little laggy, especially with sound
[22:07] sudocarl: mediacoder, is there a new alternative?
[22:07] mediacoder: well, flexbuilder for linux works
[22:07] sudocarl: yea
[22:07] skampler: who needs flash when you've got javascript?!
[22:07] sudocarl: skampler, amen!
[22:07] mediacoder: sudocarl: well, not for fancy WYSIWYG
[22:07] sudocarl: haha
[22:07] nefD: actually, i'd save JS is better for fancy ui and wysiwyg, especially if you go with something like dojo
[22:07] nefD: but for games (my main area of interest), flash is hard to beat when it comes to graphics and responsiveness
[22:08] mediacoder: yea.. i think flash will have its hard times now with html5 and highperformance js engines
[22:08] sudocarl: who needs objective c when you have cs5?  :D  Sorry I dunno if you guys have heard but cs5 will be able to build flash apps directly for the iphone
[22:08] nefD: indeed
[22:08] sudocarl: thats what I am looking forward to
[22:08] sudocarl: cause obj c is retarded
[22:08] inimino: I look forward to the death of Flash with great anticipation
[22:08] nefD: iphone development seems like a precarious thing to hang your hat on, what with the seemingly opaque app review process I keep hearing about
[22:08] mediacoder: hehe
[22:09] sudocarl: nefD, no kidding...  but my company wants iphone apps
[22:09] nefD: >__<
[22:09] sudocarl: so I would prefer to develop in as3
[22:09] sudocarl: cause it makes sense
[22:09] sudocarl: :D
[22:09] nefD: indeed
[22:09] sudocarl: I think that silverlight has a small chance of doing well..
[22:09] sudocarl: but I think it wasn't launched correctly
[22:09] sudocarl: I like the fact that it supports threading
[22:10] sudocarl: which flash sadly does not
[22:10] mediacoder: sudocarl: yea..i think that is one future of the flash IDE.. building native mobile apps for platform X
[22:10] sudocarl: mediacoder, bingo
[22:10] nefD: as3 is based on ecma3, yea?
[22:10] sudocarl: web systems will turn to other things
[22:10] sudocarl: yea
[22:10] sudocarl: thats another reason why its great :D
[22:10] sudocarl: its just like js
[22:10] nefD: explains why the syntax is like js
[22:10] nefD: yea
[22:10] sudocarl: inimino, whats your beef against flash?
[22:10] inimino: sudocarl: it's proprietary
[22:10] cmlenz: as3 is based on the dead ecma4
[22:11] sudocarl: lol :)
[22:11] sudocarl: cmlenz, thanks for clarifying
[22:11] nefD: hrm, i had read somewhere it was based on ecma3 + parts of the ecma4 specs
[22:11] sudocarl: cmlenz, I wasn't sure which ecma* it was based on
[22:11] inimino: (with all that entails)
[22:11] sudocarl: inimino, well thats fair
[22:12] cmlenz: adobe is probably pretty unhappy about how the ecma4 process turned around
[22:12] sudocarl: adobe is unhappy about alot of stuff
[22:12] sudocarl: like applle
[22:12] sudocarl: apple*
[22:12] inimino: ES4 was far too big a change, I'm rather glad it died
[22:13] sudocarl: and as for js taking over the flash world I am totally behind it, but browsers need to get better before that will happen
[22:13] inimino: but I'm glad the process got moving again and I hope Harmony will be ... better
[22:13] inimino: sudocarl: they are getting much better, and standards are moving again
[22:14] inimino: Flash's days are numbered :-)
[22:14] sudocarl: harmony?
[22:15] sudocarl: ohh the new version of ecmascript
[22:15] sudocarl: cool
[22:15] inimino: Harmony is a set of proposals and ideas that represent the next, or maybe next again, ES standard
[22:15] inimino: (after ed. 5)
[22:15] inimino: I hope they get some of their process issues straightened out too
[22:19] sudocarl: anyone want to share screenshots of their monitors?
[22:19] sudocarl: I am interested to see how people work
[22:20] orlandov: sudocarl: dual lcd's mounted to the wall... FTW!
[22:21] sudocarl: luls
[22:21] sudocarl: s/s/z
[22:21] sudocarl: http://imgur.com/hZ2nR.jpg
[22:22] orlandov: sudocarl: typically i have a maximized terminal on my first monitor and firefox maximized on my second
[22:22] sudocarl: haha I split up alot more :)
[22:22] sudocarl: maximized terminal == alot of wasted space
[22:22] orlandov: nah
[22:22] sudocarl: most of my lines are less than 80 chars
[22:22] orlandov: i use screen
[22:22] inimino: 
ACTION uses xmonad

[22:22] sudocarl: ohhh
[22:22] sudocarl: well that works  :)
[22:23] orlandov: i <3 that screen has vertical split now
[22:23] orlandov: on ubuntu at least
[22:23] sudocarl: yea screen is pretty nifty
[22:23] sudocarl: but its broken on my home-server
[22:23] sudocarl: ever since I updated V_V
[22:23] sudocarl: I need to reformat+reinstall
[22:23] sudocarl: cause something is fucked on it
[22:24] orlandov: i_i
[22:24] sudocarl: what?
[22:24] orlandov: cryingface
[22:29] the_undefined_: ryah: CommonJS testing is ready :)
[22:29] the_undefined_: ryah: see mailing list
[22:30] sudocarl: the_undefined_, whats that?
[22:31] LordMetroid: Hmm, so node.js is just a server, if I want to add functionality of a web server I will have to code it myself?
[22:31] the_undefined_: sudocarl: http://wiki.commonjs.org/wiki/Unit_Testing/1.0
[22:31] the_undefined_: sudocarl: http://github.com/felixge/node/commit/a7f5f51c45e7358c545afec052da845f9ba52ed0
[22:31] the_undefined_: and http://github.com/felixge/node/commit/de78e7a917c891f3fc0404ecd380dcd1b55e7ec2
[22:31] sudocarl: LordMetroid, http://nodejs.org/api.html#_http
[22:31] sudocarl: the_undefined_, lol :D thanks
[22:32] the_undefined_: basically its a cleaner & nicer assertation suite than mjsunit
[22:32] the_undefined_: which was ratified as a standard
[22:32] the_undefined_: by commonjs
[22:32] sudocarl: uh nifty
[22:33] the_undefined_: lets see how well this went, its kinda a massive patch :D
[22:33] sudocarl: no kidding lol
[22:34] sudocarl: atleast alot of the basic changes are just find+replace
[22:34] sudocarl: :)
[22:34] the_undefined_: sudocarl: pretty much everything I did is, yes
[22:34] sudocarl: yay its the ping pong test
[22:34] sudocarl: my fav
[22:34] the_undefined_: karl rewrote assert.js which is hard
[22:35] the_undefined_: the CommonJS reference implementation was highly dependent on their util.js class
[22:35] sudocarl: well nicely done job :)
[22:35] sudocarl: LordMetroid, did you get my link?
[22:35] LordMetroid: sudocarl, yes, I've read that
[22:35] sudocarl: LordMetroid, there are also a couple frameworks to help you out.  my personal fav right now is node-router
[22:35] sudocarl: http://wiki.github.com/ry/node
[22:35] LordMetroid: If I want to create easily configurable virtual hosts... I would need something more than this, right?
[22:35] sudocarl: there is a list of frameworks
[22:35] sudocarl: uhhhh
[22:35] the_undefined_: sudocarl: node router does look nice
[22:35] sudocarl: i am going with a yes
[22:36] orlandov: LordMetroid: what are you looking to do?
[22:36] sudocarl: the_undefined_, http://github.com/carlsverre/node-router
[22:36] sudocarl: the_undefined_, my changes :)
[22:36] sudocarl: the_undefined_, more coming down the pipeline
[22:36] the_undefined_: sudocarl: cool
[22:36] the_undefined_: love to see this stuff :)
[22:36] the_undefined_: I'm contemplating to rewrite my companies blog in node.js over christmas
[22:36] the_undefined_: http://debuggable.com/
[22:36] sudocarl: DO IT
[22:36] sudocarl: :)
[22:37] LordMetroid: orlandov, I want to create an e-commerce toolset and reach different tools through different virtual hosts
[22:37] the_undefined_: its pretty simple, and it would probably be insanely fast afterwards :)
[22:37] orlandov: LordMetroid: did you want node.js to do all of that? or were you just going to like reverse proxy to node.js for some stuff?
[22:37] the_undefined_: ok going to bed now
[22:37] LordMetroid: I don't know
[22:37] the_undefined_: cya guys
[22:37] orlandov: l8
[22:38] LordMetroid: orlandov, can I do it with node.js?
[22:39] LordMetroid: I suppose it should be a matter of simply parse some config files and setup the docroots for different incomming requests
[22:40] orlandov: i would personally use something like nginx to serve most stuff, and then delegate some requests to node.js
[22:40] mediacoder: yea, thats also how im using node
[22:41] orlandov: yeah, like static files i'd let a frontend web server dole out
[22:41] mediacoder: behind a proxy and let the dumb webserver do the dirty serving work
[22:41] orlandov: ya
[22:41] LordMetroid: So how would I do that?
[22:42] LordMetroid: 
ACTION goes and downloads nginx

[22:42] mediacoder: install nginx
[22:42] binary42_ has joined the channel
[22:42] mediacoder: and proxy the requests to node running on localhost
[22:42] orlandov: yeah, you'll have to do a "man nginx" or some googling
[22:42] orlandov: i personally use Cherokee for httpd
[22:43] orlandov: but any front end server would do
[22:43] orlandov: woot, successfully cargo culted a c++ extension
[22:44] mediacoder: orlandov: what is it about? i was also going to dive into the c extension world at some point :-)
[22:45] orlandov: mediacoder: one sec
[22:45] mediacoder: (would love to see some wrapper around some image manipulation lib :-)
[22:46] inimino: ooh
[22:47] inimino: the GIMP library would be nice
[22:48] Biscuits: Hmmm
[22:48] inimino: then maybe the Web-based photoshop killer will be backed by node.js
[22:48] Biscuits: You wouldn't want to run that in your main thread though :/
[22:48] sudocarl: inimino, that would be sweet
[22:48] inimino: sure, definitely needs to be async
[22:48] sudocarl: er mediacoder  :)
[22:49] sudocarl: you could write a tcp wrapper to gimp
[22:49] sudocarl: heck there probably already is one
[22:49] sudocarl: then just dispatch jobs to it
[22:49] wil_ has joined the channel
[22:49] inimino: probably
[22:49] inimino: yes
[22:50] mediacoder: well, that wouldnt need a c-wrapper, wouldnt it? (i wish i wasnt that noobish here :-D)
[22:50] sudocarl: nope that could be done in js
[22:50] inimino: then build a Web UI, <canvas> for client-side preview, GIMP library on the backend for heavy-duty image processing :)
[22:50] mediacoder: hehe
[22:50] sudocarl: see: http://github.com/creationix/postgres-js
[22:50] sudocarl: for a similar example
[22:51] mediacoder: well, i was more thinking about captchas..but that also sounds great :-)
[22:51] sudocarl: mediacoder, hahaha
[22:51] sudocarl: you don't need gimp for that
[22:51] sudocarl: http://www.google.ca/search?q=imagemagick+captcha&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a
[22:52] sudocarl: node's child-process support would be fine for that
[22:53] mediacoder: yea.. i was also looking into this..but thaught a lower-level interface like for example gd for php (or also imagemagick) for nodejs would be also kool
[22:53] sudocarl: :)
[22:53] ashb: mediacoder: gd sucks
[22:53] sudocarl: but just remember that things like this exist and work: http://caca.zoy.org/wiki/PWNtcha
[22:53] mediacoder: yea..that was just an example :-P
[22:54] ashb: cairo or imlib2 are better
[22:54] ashb: ah k
[22:55] inimino: CAPTCHA sucks
[22:55] inimino: broken idea
[22:55] sudocarl: yea
[22:56] sudocarl: there are too many easy ways around it
[22:57] orlandov: mediacoder: i would use a job queue, rather than try to do it fro mnode
[22:57] orlandov: use something like gearman or TheSchwartz
[22:58] ashb: dont use theschwartz if order of jobs is important
[22:58] pjb3 has joined the channel
[22:58] orlandov: ashb: ah yeah... i actually hacked it to do that :)
[22:59] ashb: ditto :)
[23:00] ashb: there are better options now than TheSchwartz
[23:01] sudocarl: here is my favorite way of breaking captchas: http://www.bluehatseo.com/creating-an-army-of-free-captcha-typers/
[23:01] sudocarl: :)
[23:01] ashb: mech turk?
[23:02] sudocarl: nope this method is free
[23:02] sudocarl: you make a web-proxy
[23:02] ashb: ah
[23:02] sudocarl: and then write a script which grabs a captcha off a reg form and gets the next person who wants to use the proxy to typoe it in
[23:02] sudocarl: type*
[23:02] orlandov: ah the old switcheroo
[23:02] sudocarl: :-0
[23:02] sudocarl: indeed
[23:13] sixtus42_ has joined the channel
[23:13] isaacs_ has joined the channel
[23:16] fictorial has joined the channel
[23:16] fictorial: hey guys
[23:17] fictorial: so I had an instance wherein I needed a sequence of functions that return Promise objects called one after another... so I created PromiseChain
[23:17] fictorial: but it got me wondering if anyone has any sort of "Promise utils"
[23:18] isaacs_: fictorial: I did something like that here: http://github.com/isaacs/npm/blob/master/src/queue.js
[23:19] fictorial: isaacs_: ok I'll look thanks
[23:19] isaacs_: although, in that case, you provide it with a list of things, and a function
[23:19] fictorial: btw, were you from Yahoo?  
[23:19] isaacs_: the function returns a promise, and the invocation returns a promise that emits success when all the items are successfl
[23:19] isaacs_: 
ACTION is indeed from Yahoo

[23:19] fictorial: I quit Yahoo 2 weeks ago to focus on Fictorial full time
[23:20] isaacs_: oh, neat
[23:20] fictorial: scary but exciting
[23:20] fictorial: 
ACTION checks out Queue

[23:20] orlandov: fictorial: good for you :)
[23:20] isaacs_: i'm saving a nest egg to do the same thing next year some time
[23:20] isaacs_: s/Fictorial/sisterTrain/
[23:20] fictorial: yeah, that's what I did - ads in iPhone games pay handsomely enough for me to do it
[23:21] fictorial: well, that and consulting
[23:21] fictorial: so, we'll see
[23:21] isaacs_: nice
[23:21] isaacs_: i tend to tread carefully where my rent is concerned.
[23:21] isaacs_: i like living indoors and alone
[23:21] fictorial: hehe
[23:21] fictorial: Yeah, I "ran the numbers" and it worked out for about 6-8 months so that was enough for me to dive in and try it
[23:22] isaacs_: so, when i've got a year or so in the bank, i'll be more comfortable ditching the regular direct deposit gig.
[23:22] isaacs_: nice
[23:22] orlandov: fictorial: what's your venture (if you can say)?
[23:22] fictorial: I'm in "stealth mode" :)
[23:22] isaacs_: that seems like the way to do it.  if you're self-funded you can play a bit more, try things, fail a little, etc.
[23:22] pjb3 has joined the channel
[23:22] fictorial: it has to do with iPhone gaming
[23:22] fictorial: I'll either rule the world or fail spectacularly
[23:22] orlandov: fictorial: nice! well,best of luck to ya
[23:23] isaacs_: fictorial: failing educationally is better.
[23:23] fictorial: thanks!
[23:23] fictorial: indeed :)
[23:23] isaacs_: though less entertaining ;)
[23:23] orlandov: it's only a failure if you dont' learn
[23:23] sudocarl: fictorial, when your ready to release more information my company relevantgames.com would love to hear about what your doing :)
[23:23] isaacs_: orlandov: no, sometimes it's still a failure.
[23:23] fictorial: sudocarl: awesome
[23:23] orlandov: isaacs_: sshh i'm being supportive!
[23:23] sudocarl: lulz isaacs 
[23:23] isaacs_: 
ACTION has been a part of some very educational failures that hurt just the same.

[23:24] isaacs_: no pain, no gain!
[23:26] sudocarl: the best part about crud is being able to say that you write code about crud
[23:26] sudocarl: tehee
[23:32] isaacs: fictorial: getting back to what you were saying... a "promise utils" lib would be killer, imo.
[23:33] fictorial: yeah
[23:33] fictorial: you're Queue idea is very similar to my PromiseChain
[23:33] fictorial: I also have a half-ass PromiseGroup
[23:33] isaacs: especially for cases when you have like 15 things to do, but they can be in any order.
[23:33] isaacs: and you want to fire a listener for when they're all done
[23:34] fictorial: PromiseGroup takes a bunch of Promises and waits until they're all done (pass or fail) then emits results (n failed, m passed, etc.)
[23:34] isaacs: right
[23:34] ashb: sounds perfect for a test runner
[23:34] pjb3: I have a question based on something I saw in Ryan's JSConf.eu presentation
[23:34] fictorial: ashb: heh, yeah
[23:34] isaacs: whereas Queue waits for item n before running item n+1
[23:34] pjb3: When he demoed stat
[23:34] pjb3: first it does
[23:35] pjb3: var promise = stat("/etc/passwd")
[23:35] isaacs: fictorial: so... got a github link you feel like sharing with the class?
[23:35] pjb3: then it adds the success callback to promise
[23:35] pjb3: so, when does it actually make the call out to the OS?
[23:35] ashb: in a thread in the pool
[23:36] pjb3: Will the success handler be on promise before the stat finishes?
[23:36] isaacs: pjb3: yes
[23:36] isaacs: pjb3: if you're familiar with browser js, it's very similar to setTimeout(fn, 0)
[23:36] fictorial: isaacs: let me test these a bit then I'll post em yeah
[23:36] pjb3: how do you know that, just because it's the next thing you do and you know the stat OS call will take longer?
[23:37] fictorial: nothing earth-shattering
[23:37] isaacs: pjb3: no, it's not a race condition.  node actually enforces this.
[23:37] pjb3: isaacs: how?
[23:37] isaacs: pjb3: rtfs ;P
[23:37] ashb: isaacs++
[23:37] pjb3: what would happen if I did this?
[23:37] pjb3: var promise = stat("/etc/passwd")
[23:38] ashb: rtfs like isaacs said and find out
[23:38] pjb3: for(var i=0; i < 1000000; i++) { whatever() }
[23:38] pjb3: promise.addCallback(function(s) {} ... )
[23:38] isaacs: pjb3: like i said, it's not a race condition or a timing issue.  the current executing process will block the execution and the stat() will wait patiently until it's done.
[23:39] orlandov: to be fair i don't think the way promises are intuitively obvious, even looking at the code, specially for soemone just taking a look now
[23:39] isaacs: orlandov: yes, there's a learning curve.
[23:39] isaacs: pjb3: this, however, would be dangerous: p = stat("/etc/passwd"); setTimeout(function () { p.addCallback(foo) })
[23:39] sudocarl: pjb3, the thing is that promise can wait until it gets a "newListener" event before firing off the success callback
[23:40] sudocarl: issacs, wouldn't that still work?
[23:40] sudocarl: because promise should wait until there is a listener?
[23:40] sudocarl: before firing the callback?
[23:40] pjb3: sudocarl: yeah, that's what I was wondering
[23:40] isaacs: sudocarl: afaik, it'll be a problem.
[23:40] ashb: TIAS
[23:40] isaacs: you can do stuff without assigning a callback ever, and just an errBack
[23:40] isaacs: because in that case, the setTimeout is *also* deferring execution until later, so you end up with a race condition.
[23:41] ashb: depends. it sohuldn't ever crash
[23:41] ashb: it might just mean you dont get your callback fired
[23:41] isaacs: however, last tiem i played with that type of pattern was a while back, and it's entirely possible that ryah or someone else added some magic in there.
[23:41] isaacs: ashb: right, what i'm saying is, if you add your callback in a setTimeout fn, then it might not be added before the emitSuccess() happens
[23:41] ashb: yeah
[23:41] isaacs: but you kind of have to go out of your way to make that mistake.
[23:42] ashb: depends if anyone has made it fire if its compelted when you add the first onSuccess
[23:42] isaacs: ashb: as i understand it, it just puts a todo item on the process queue
[23:42] ashb: no idea
[23:42] ashb: i've used node for all of 5 mins :D
[23:43] isaacs: 
ACTION has not rtfs enough :)

[23:43] inimino: it's the same as the browser's single-JavaScript-thread event loop
[23:43] ashb: i'm just here to steal ideas, steal info and spread FUD >_>
[23:43] kriskowal: you guys need to read ref-send and mark miller's dissertation
[23:43] isaacs: inimino: that's how i understand it, yes
[23:43] rakeshpai has joined the channel
[23:43] inimino: which as far as I know isn't really well-documented anywhere
[23:43] pjb3: I just tested it, adding the callback in a setTimeout doesn't work
[23:44] inimino: well, actually, it's documented now in HTML5
[23:44] kriskowal: promises aren't just an API; there's a lot that needs to happen under the hood for them to provide the reliability and composability features they're supposed to guarantee
[23:44] inimino: I should say there isn't any good introductory material that I'm aware of
[23:44] kriskowal: if you want to learn about promises, check in at #erights
[23:46] isaacs: pjb3: yep.  it should never work, actually.
[23:47] isaacs: as in, not even a race condition, just a flat out failure.
[23:47] pjb3: yeah, but if you do for(var i=10000000; i > 0; i--) { puts(i) }
[23:47] pjb3: between stat("whatever")
[23:47] pjb3: and adding the callback, the callback does fire
[23:48] isaacs: pjb3: that should work, yes
[23:48] isaacs: because it's actually all happening in a single thread.
[23:48] pjb3: so as you said, it's not a timing issue
[23:48] isaacs: the stat itself occurs in a process pool.
[23:48] kriskowal: some stuff on promises http://waterken.sourceforge.net/web_send/
[23:48] inimino: it's an event loop
[23:48] kriskowal: the lib Q API is the important bit.
[23:48] kriskowal: it's not just an event loop.
[23:49] inimino: reading about the browser JavaScript execution model would probably be highly enlightening
[23:49] kriskowal: at least, it's not supposed to be.
[23:49] kriskowal: promises are supposed to guarantee stable propagation of causes and effects
[23:49] kriskowal: through return values and arguments, with very reliable order of resolution
[23:49] isaacs: kriskowal: i'm pretty sure they do in node.  (if not, posting a bug would probably be a great help)
[23:50] pjb3: yeah, I can't get my head around how this works: http://gist.github.com/245894
[23:50] pjb3: somehow, both callback get fired
[23:51] isaacs: pjb3: becasue, until it gets to the END of your current block of executing code, you can add as many callbacks as you want.
[23:51] inimino: pjb3: both event handlers are added, so both are fired
[23:51] inimino: pjb3: there is only one event
[23:51] isaacs: you could loop i for hours.
[23:51] isaacs: it's not a timing issue.
[23:51] kriskowal: ah, pjb3, a promise is a "state" event.  once a promise has been "resolved", all further callbacks are guaranteed to execute in successive events
[23:52] kriskowal: a "promise" is just that; a promise that all registered callbacks will execute exactly once when the promise has been resolved
[23:52] isaacs: pjb3: the setTimeout case fails because it's handing the program some code to run "later", 
[23:52] pjb3: isaacs: so that stat executes once the interpreter has eval'd everything in the file?
[23:52] ashb: no the stat will fire off in a thread almost straight away
[23:52] isaacs: pjb3: the blocking c-level stat() call happens in the process pool, pretty son.
[23:52] isaacs: *soon
[23:53] isaacs: but the *promise* wrapper provided by libeio won't do its thing until node is done with the current code.
[23:53] pjb3: but the callbacks aren't fired until it's parsed everything?
[23:53] isaacs: it'll sit there with teh results of the stat() call, waiting patiently until its turn comes up
[23:53] kriskowal: it's not a matter of parsing.  but it is a matter of the current thread of execution completing.
[23:53] inimino: parsing is something else altogether
[23:53] pjb3: isaacs: yeah, that's what I'm trying to grasp, how it knows when it is it's turn
[23:54] isaacs: so, guys, but s/parse/execute/ and pjb3's comment is right.
[23:54] ashb: so the timeout can't fire until you get back into the event loop
[23:54] pjb3: right, I'm sure I'm getting the parsing/executing/evaling language mixed up
[23:54] ashb: which could happen via a .wait() or an explicit loop() call
[23:54] isaacs: pjb3: imagine that all your code is wrapped in a loop.
[23:54] isaacs: when it gets to the end of the loop, it checks to see if there's anything on its todo list, and then it does it.
[23:54] isaacs: adding a callback adds something to the todo list.
[23:55] pjb3: isaacs: awesome, that clears it up for me
[23:55] isaacs: setTimeout() also adds something to the todo list.
[23:55] pjb3: so the call to stat("/etc/passwd") is a "todo-list" item as well, at least conceptually, right?
[23:55] ashb: yes
[23:55] isaacs: pjb3: yes, exactly.
[23:55] kriskowal: um.
[23:56] inimino: isaacs: now go write the blog post that explains the browser/node event loop in an introductory fashion and then we can just point people to it :)
[23:56] isaacs: kriskowal: broad strokes.
[23:56] isaacs: :)
[23:56] isaacs: inimino: not a bad idea, actually
[23:56] kriskowal: no, it's a promise.  any callback you attach to the promise gets added to the todo list when the promise is resolved.
[23:56] isaacs: oohhhh... couldn't restrain yourself, could you kriskowal!
[23:57] kriskowal: if it helps, imagine i said that in the voice of the simpson's comic book guy
[23:57] pjb3: so at least logically, I can think about the order of operations that is happening in the gist I posted as
[23:57] pjb3: 1.  when I tell you to, call stat
[23:57] pjb3: 2.  when you call stat, do this
[23:57] pjb3: 3.  right now, do a for loop
[23:57] pjb3: 4.  when you call stat, also do this
[23:57] pjb3: 5.  ok, call stat now
[23:58] kriskowal: 6. that first thing happens
[23:58] kriskowal: 7. that second thing happens
[23:58] isaacs: pjb3: as long as when you say "call stat", you mean "get the results of the stat call and inform the listeners"
[23:58] crux has joined the channel
[23:58] pjb3: isaacs: yeah, that's what I mean
[23:58] isaacs: actually checking in with the OS about the nature of the file is done in the process queue almost immediately
[23:58] isaacs: and asynchronously
[23:58] isaacs: oops, not process queue, process *pool*
[23:58] isaacs: pjb3: incidentally, the reason why (imo) this is such a powerful pattern is that it encourages small bits of code that don't block one another and only care about order when it really matters.
[23:59] isaacs: code written in that fashion tends to be easier to reason about.
[23:59] kriskowal: on a side note, the node api is a bit out of date if it's tracking the commonjs proposal.
[23:59] pjb3: I'm sure I could look this up somewhere, but does stat take a second argument as a callback function?
[23:59] inimino: kriskowal: which proposal?
[23:59] pjb3: because stat("/etc/passwd", function(s) { ... } )