[00:06] martyn_ has joined the channel
[00:13] rob42 has joined the channel
[00:13] rob42: how have you guys found the performance of node.js vs, say, ruby+eventmachine?
[00:13] rob42: I'm finding the performance to be incredible
[00:14] Tim_Smart: rob42: Cool :p
[00:15] rob42: :)
[00:15] tmpvar has joined the channel
[00:16] rektide: re: avro & node-avro, tmpvar did post this: http://groups.google.com/group/nodejs/msg/a2020366da9bf67d
[00:16] rektide: oh hello back
[00:16] tmpvar_ has joined the channel
[00:27] tmpvar_ has joined the channel
[00:33] tmpvar has joined the channel
[00:35] cadorn has joined the channel
[00:37] isaacs: Tim_Smart _ry felixge: Hey, when/if you guys get a second, I'd love to hear any thoughts on this proposal: http://groups.google.com/group/nodejs/browse_thread/thread/7e4cb5dd8ef87d76
[00:39] felixge: isaacs: I ran into this before, but I couldn't describe it as well as you have
[00:39] Tim_Smart: isaacs: Will it solve the issue requiring a dir with a index.js, and then requiring the index.js directly?
[00:39] isaacs: Tim_Smart: i didn't touch that, but it could be done at the same time
[00:39] isaacs: Tim_Smart: i tried to just focus on teh problem at hand.
[00:39] Tim_Smart: I guess __filename would do it, yeah
[00:40] isaacs: Tim_Smart: well, if you did require("./foo") and then later did require("./foo/index"), then they'd have separate entries at the top level of the cache
[00:40] Tim_Smart: I'm currently trying to run modules in a sandbox using evalcx
[00:40] felixge: isaacs: your solution looks too complicated
[00:41] isaacs: felixge: i agree.  but i'm not sure how to make it any simpler.
[00:41] felixge: isaacs: in fact, ryan is probably going to bring up removing the module system all together again :D
[00:41] isaacs: hahaha
[00:41] felixge: I'm wondering why we need module caching
[00:42] isaacs: felixge: it cuts down load time
[00:42] felixge: how much?
[00:42] isaacs: felixge: also, 2 other things
[00:42] felixge: you should never use require() during runtime unless you want to hot reload something
[00:42] isaacs: felixge: 1. if you require a module you might not want to do the setup again, just get the old exports.
[00:42] felixge: and I don't care if my script starts up in 20ms or 200ms
[00:42] isaacs: felixge: 2. circular requires necessitate SOME kind of cache
[00:42] felixge: ah, right
[00:42] felixge: the circular stuff sucks
[00:43] isaacs: a.js does require("b").  b.js does require("a")
[00:43] isaacs: if it's cahced, you jsut return the "unfinished" object
[00:43] isaacs: incidentally, for the same reason, i need a package registry to do circular dependencies in npm
[00:43] isaacs: but that's another godawful long thread
[00:43] felixge: :D
[00:44] felixge: anyway, looking at the problem I have no better solution to offer
[00:44] felixge: but I'm on my 2nd club mate for tonight, miracles may occur
[00:44] felixge: :)
[00:44] isaacs: i think it's a good idea to have a bundled module loading system.  and the code is absolute spaghetti atm
[00:44] felixge: yes, if ryan kills the module system I will kill myself
[00:44] felixge: :)
[00:44] Tim_Smart: isaacs: Why are we skipping functions with evalcx?
[00:44] Tim_Smart: for the sandbox
[00:45] isaacs: Tim_Smart: because function scopes destroy the sandboxing, and v8 yells at you when you try to run a function from context A in context B
[00:45] isaacs: (at least, it did last year, don't know if that's supported somehow)
[00:45] Tim_Smart: Time to find out :p
[00:45] isaacs: hehe
[00:45] isaacs: Tim_Smart: you can't do the ->ToObject()->Clone() on the function, though
[00:45] isaacs: Tim_Smart: because it wont' be a function after you do that
[00:46] Tim_Smart: ToFunction()?
[00:46] isaacs: Tim_Smart: i don't think that functions have a Clone(), do they?
[00:46] Tim_Smart: no, why don't we just pass by reference?
[00:46] isaacs: Tim_Smart: because they're different contexts
[00:48] isaacs: Tim_Smart: let's say you're running some malicious code.  and that malicious code does someObj.__defineGetter__("foo", function () { require("sys").eval("rm -rf /") })
[00:48] Tim_Smart: right, so v8 sandboxes aren't as cool as spidermonkey ones
[00:48] isaacs: Tim_Smart: if by "cool" you mean "hideously insecure", then, yes
[00:48] carsonm: Does anyone know if the net2 branch is the future? Or maybe parts of it?
[00:48] Tim_Smart: isaacs: I want to get require() into the sandbox :p
[00:48] Tim_Smart: etc
[00:48] isaacs: Tim_Smart: i see
[00:49] isaacs: well, try it
[00:49] isaacs: ie, remove the line that does the continue() if it's a function, and see if you can pass functions back and forth
[00:49] Tim_Smart: I have, but you got a continue for functions
[00:49] Tim_Smart: ok I will :p
[00:49] isaacs: i was going off of what i'd remembered from times long gone about functions being tied to their context
[00:50] Tim_Smart: good times, it gets passed in as a function
[00:50] isaacs: weird
[00:50] isaacs: Clone()?
[00:50] isaacs: that just seems so... unsafe.
[00:51] isaacs: i mean, your evalcx code could get at callee.caller and whatnot
[00:52] Tim_Smart: v8 might null those values or something. Will find out I guess
[00:54] jed_ has joined the channel
[00:56] isaacs: Tim_Smart: oh, i wonder if you'll just get a permission error, since the context's SecurityToken isn't the same?
[00:56] Tim_Smart: testing now
[00:56] isaacs: that's how iframes can talk to one another (or can't, if they're on different domains)
[00:56] isaacs: nice
[00:56] mjr_ has left the channel
[00:57] dekz has joined the channel
[01:05] ditesh has joined the channel
[01:07] joshbuddy has joined the channel
[01:09] isaacs: Tim_Smart: so, what's the verdict?
[01:09] Tim_Smart: that its a big hack :p
[01:09] isaacs: hahaha
[01:10] isaacs: i mean, can you pass functions in and actually get a working require() into evalcx?
[01:10] isaacs: cuz that'd be kinda neat
[01:12] isaacs: Tim_Smart: at the very least, it'd make andy and edwin happy
[01:12] isaacs: Tim_Smart: and i suppose, even if the functions were a leak, it would be like a box-model security type thing.  ie, you only have access to what i show you, so it's up to me to be explicit and only share what's safe to share
[01:16] Tim_Smart: Yeah well, atm I got it going almost, except instanceof in the node context doesn't point to the right constructors
[01:19] isaacs: Tim_Smart: oh, right, huh?  since it created a new copy of those things
[01:19] isaacs: but i think that's a pretty reasonable caveat
[01:19] Tim_Smart: Yeah, new context = new Function, Array etc etc
[01:20] voodootikigod_ has joined the channel
[01:20] isaacs: the alternative is to go through all the things in the sandbox, test its constructor, etc.
[01:20] Tim_Smart: well I got node-repl going :p
[01:20] isaacs: OH! hey, you should probably do clone() on arrays, too
[01:20] isaacs: yeah, the repl's one place where cx really can be awesome
[01:20] isaacs: you can make the .clear a LOT less hacky
[01:20] creationix has joined the channel
[01:20] isaacs: like, it can *actually* give you a brand new context
[01:21] isaacs: anyway, because arrays are byref, I could get your array in my cx code, and do "ar.__proto__.push = myEvilFunction"
[01:22] Tim_Smart: you should just clone everything like you do with objects, as everything in Javascript is an object pretty much
[01:22] isaacs: the alternative would be to change the __proto__ in the EvalCX() c++ function to point everything at the sandboxed version, which is wonky
[01:22] isaacs: Tim_Smart: but then if you pass 3 in, it's not 3, it's Number(3)
[01:22] isaacs: which isn't the same thing, because JavaScript is a dick.
[01:23] Tim_Smart: OK, we can eliminate the exceptions I guess
[01:23] isaacs: that's what IsObject() did, i thought.
[01:23] isaacs: ie, any native (non-object) would return false from that
[01:23] Tim_Smart: Yeah arrays are objects
[01:23] isaacs: so arrays return true from IsObject()?
[01:23] Tim_Smart: not sure, will have to test that one
[01:24] isaacs: i'd think it would, since v8::Array inherits from v8::Object
[01:24] mikeal: arrays are objects with a length property
[01:25] mikeal: :)
[01:25] Tim_Smart: Yeah it should then
[01:25] mikeal: oh, this is in C land
[01:25] isaacs: oh, weird, v8::Date *doesn't* inherit from v8::Object
[01:25] mikeal: i needed to scroll up
[01:25] isaacs: mikeal: hehe
[01:26] mikeal: the Date object is all kinds of fucked up
[01:26] isaacs: Tim_Smart: yeah, assuming that v8 is sane here, you'd have to test for IsObject(), IsDate() and IsExternal()
[01:26] isaacs: if any of those are true, then Clone()
[01:26] Tim_Smart: isaacs: You can fix your patch if you want :p
[01:26] isaacs: hahah
[01:27] isaacs: Tim_Smart: well, when i say "you" i really mean "the human team".  who actually types the code is irrelevant
[01:27] isaacs: Tim_Smart: it's us vs the machines
[01:27] Tim_Smart: hopefully we are winning, but no-one knows for sure
[01:27] isaacs: so, otoh, maybe it makes sense to not clone anything, and just say "if you hand evalcx an object, then evalcx will have your object"
[01:27] joshbuddy has joined the channel
[01:27] mikeal: that's how SM works
[01:28] isaacs: right, so then it's up to the caller to be safe
[01:28] isaacs: or not
[01:28] isaacs: Tim_Smart: how's that sound to you?  would that make it easier for you to do what you're trying?
[01:28] Tim_Smart: Sounds good
[01:29] felixge: do you guys need some decent date libraries?
[01:29] felixge: * know
[01:29] felixge: something to make things easy like adding '1 hour' to a date object?
[01:29] isaacs: felixge: philip tellis ported php's date() function to js, i believe
[01:29] isaacs: date() and strtotime
[01:30] felixge: isaacs: got a link?
[01:30] felixge: does it have tests?
[01:30] felixge: :)
[01:30] Tim_Smart: nodejs_v8: Object.getOwnPropertyNames(new Date())
[01:30] nodejs_v8: Tim_Smart: []
[01:30] Tim_Smart: nodejs_v8: Object.getOwnPropertyNames(Date)
[01:30] nodejs_v8: Tim_Smart: ["now", "arguments", "UTC", "length", "name", "prototype", "caller", "parse"]
[01:31] Tim_Smart: Date.parse('25/4/2010')
[01:31] isaacs: felixge: ok, so, it's strftime, not strtotime: http://hacks.bluesmoon.info/strftime/
[01:31] Tim_Smart: nodejs_v8: Date.parse('25/3/2010')
[01:31] nodejs_v8: Tim_Smart: NaN
[01:32] Tim_Smart: nodejs_v8: You fail
[01:32] nodejs_v8: Tim_Smart: Exception: SyntaxError: Unexpected identifier
[01:32] felixge: Tim_Smart: Date.parse is just picky
[01:32] isaacs: Tim_Smart: i was hoping to see "You is undefined"
[01:32] inimino: carsonm: it's the future
[01:33] Tim_Smart: isaacs: OK, I'm going to try pass everything by reference
[01:33] felixge: damn, I really don't want to write date libraries at 2 am :(
[01:33] felixge: *sigh*
[01:34] inimino: if you're writing the date library you can make it be whatever time you like
[01:34] Tim_Smart: felixge: Go to bed then :p
[01:34] felixge: Tim_Smart: I got my agenda, and sleep ain't on it until thursday :(
[01:36] inimino: js> var hr=3600*1000; new Date(+new Date + 1*hr)
[01:36] gbot2: inimino: Tue Mar 16 2010 22:36:22 GMT-0400 (EDT)
[01:37] inimino: JavaScript date handling is pretty painful.
[01:37] inimino: felixge: what kind of API do you want?
[01:37] inimino: there's a Ruby-like one if you like that sort of thing
[01:38] felixge: I guess this will do:
[01:38] felixge: nodejs_v8: new Date(+new Date - 60 * 60 * 1000)
[01:38] nodejs_v8: felixge: {}
[01:38] felixge: nodejs_v8: new Date(+new Date - 60 * 60 * 1000).toString()
[01:38] nodejs_v8: felixge: "Wed Mar 17 2010 00:37:45 GMT+0000 (UTC)"
[01:38] felixge: nodejs_v8: new Date(+new Date - 2 * 60 * 60 * 1000).toString()
[01:38] nodejs_v8: felixge: "Tue Mar 16 2010 23:37:53 GMT+0000 (UTC)"
[01:38] felixge: yeah, works
[01:38] felixge: inimino: I actually want as little 3rd party code as possible. Anything without unit tests is not acceptable
[01:39] mjr_ has joined the channel
[01:39] isaacs: _ry: you seeing the evalcx stuff above?  http://github.com/isaacs/node/commit/1cfb0d41ae29dc654437ee48e5187169d471f14b
[01:39] isaacs: Tim_Smart: fixed like we talked about here: http://github.com/isaacs/node/commit/1cfb0d41ae29dc654437ee48e5187169d471f14b
[01:40] Tim_Smart: isaacs: You wouldn't need to copy it back wouldn't you? everything is referenced
[01:40] isaacs: Tim_Smart: well, but, when you turn it into a new context global, you're actually not getting the same object.
[01:41] isaacs: you have to copy in order to pick up new globals that may have been created
[01:41] isaacs: or modified
[01:41] Tim_Smart: Ah ok
[01:42] Tim_Smart: yeah, other than that, those were the changes I made on my local branch
[01:42] carsonm: inimino: thanks, I'm glad I noticed someone talking about it.  I was looking for unix domain support.
[01:44] Tim_Smart: yay, fully securable modules
[01:45] stepheneb has joined the channel
[01:48] temoto has joined the channel
[01:49] temoto: Hello. Does Node.js use aio_* family for file IO?
[01:52] rektide: holy shit i didnt know evalcx was eval(code,obj)
[01:52] rektide: word
[01:53] Tim_Smart: isaacs: I tapped a filename argument to evalcx
[01:53] isaacs: Tim_Smart: word
[01:53] Tim_Smart: anyway, time to commit this shi
[01:53] isaacs: Tim_Smart: you should reply to my patch thread with your filename update
[01:54] Tim_Smart: sure
[01:54] isaacs: that's a good idea.  didn't think of that
[01:54] isaacs: rektide: yeah, it defaults to {} if you don't pass in anything
[01:57] okito has joined the channel
[01:58] Tim_Smart: isaacs: Going to have to change 'instanceof Function' to 'typeof === 'function''
[01:58] Tim_Smart: due to a different sandbox for every module
[01:59] isaacs: Tim_Smart: wow, that's a bit lame
[01:59] isaacs: what's this for?
[01:59] Tim_Smart: only 2 lines need changing though
[01:59] Tim_Smart: addListener
[01:59] Tim_Smart: and removeListener
[01:59] isaacs: Tim_Smart: wait, what are you doing with evalcx?
[01:59] Tim_Smart: isaacs: Fully securable modules
[01:59] isaacs: ok... to replace node's existing require(), or be an add-on to it?
[01:59] temoto has left the channel
[02:00] Tim_Smart: isaacs: Nothing will change, except a modules global context won't leak into the next one
[02:01] Tim_Smart: And you can require, without worrying about your globals getting messed up
[02:01] isaacs: oh, so you're swapping out the process.compile() with a process.evalcx?
[02:01] Tim_Smart: yeah
[02:01] isaacs: i think that might destroy express.
[02:01] rektide: process.evalcx is enough to get env-js going
[02:01] rektide: Tim_Smart: 
[02:02] rektide: i havent looked at, what is it, runtime.js or what not
[02:02] rektide: there's one file with all the rhino deps
[02:02] Tim_Smart: isaacs: Why, what does express do, naughty things?
[02:03] isaacs: Tim_Smart: it's only naughty if you say no to it.
[02:03] isaacs: ;P
[02:03] Tim_Smart: isaacs: No!
[02:03] isaacs: hehe
[02:03] isaacs: yeah, it puts stuff on the global object
[02:03] Tim_Smart: rawr >.>
[02:03] isaacs: but you could probably just expose "global" on the context obj, and if you really WANT to use globals, you can.
[02:03] isaacs: but at least then it won't be by accident.
[02:04] isaacs: or, you could write a completely separate module loader, and we could move that out of node.js (or greatly reduce it)
[02:04] Tim_Smart: global already has global attached to it
[02:04] isaacs: Tim_Smart: right, but i mean, as the context obj, you pass in {global:global}
[02:04] Tim_Smart: And I'm just doing a 1 level deep copy
[02:04] isaacs: so you can still do global.foo to pollute things
[02:04] Tim_Smart: to the sandbox object
[02:04] isaacs: i see, so express could still work
[02:04] Tim_Smart: yeah
[02:05] isaacs: it'd just have to do global.foo = bar rather than just foo = bar
[02:05] isaacs: for the record, i don't approve of global pollution.  but people like express.
[02:05] Tim_Smart: It breaks coffeescript though >.>
[02:06] Yuffster has joined the channel
[02:06] carsonm: Would it be a waste of time to make a simple fcgi interface for node.js that could be used with nginx, etc?
[02:07] rektide: no
[02:07] carsonm: It would just be one less thing to think about node having to handle.
[02:07] rektide: just.... not super easy
[02:07] carsonm: I think it will actually be pretty easy
[02:07] rektide: well, its being able to not think about anything your other http server already has setup
[02:08] rektide: there's no other real advantage
[02:08] dnolen has joined the channel
[02:08] brandon_beacher has joined the channel
[02:09] joshbuddy has joined the channel
[02:09] carsonm: yeah, it almost seems like the http lib shouldn't be part of node itself
[02:09] rektide: its not, its in a module
[02:09] carsonm: yeah but it comes with it is all
[02:10] rektide: yes
[02:10] carsonm: unlike things like mq or database support
[02:11] Tim_Smart: isaacs: http://github.com/Tim-Smart/node/compare/module-sandbox
[02:13] isaacs: Tim_Smart: instead of using the wrapper, why not just create the sandbox obj with those values in it?
[02:13] mjijackson has joined the channel
[02:13] Tim_Smart: I did that initially
[02:13] Tim_Smart: But I forget why I reverted
[02:14] mjijackson: ashb: Didn't you point to an example some time last week of how one could extend markdown-js with their own extension?
[02:14] isaacs: people bitch about require() putting their code in a function unexpectedly.  that'd kinda solve the problem.
[02:15] isaacs: ok, i gotta run.  Tim_Smart, i'm gonna be all over require() stuff tomorrow for the module.id wonkiness that is breaking npm a little bit these days.
[02:16] mjijackson: ashb: Sorry ... *dialect* is the word that is used in the code.
[02:16] isaacs: so i'll review your kit in more detail then.  feel free to email me about it if you do more tonight.
[02:16] dandean has joined the channel
[02:16] Tim_Smart: isaacs: OK cool
[02:16] erikvold has joined the channel
[02:18] aryounce has joined the channel
[02:21] ryandotsmith has joined the channel
[02:21] ryandots_ has joined the channel
[02:22] cdorn has joined the channel
[02:23] captain_morgan has joined the channel
[02:24] captain_morgan: felixge, Where does the "node-stress" module you're using in Dirty v2 come from?
[02:25] felixge: captain_morgan: /Users/felix/.node_libraries/node-stress :)
[02:26] captain_morgan: Uhh huh..
[02:26] captain_morgan: You have it up anywhere? 
[02:26] felixge: captain_morgan: let me push it
[02:26] captain_morgan: doesn't look to be on github
[02:26] captain_morgan: thanks
[02:27] felixge: no guarantees, no docs - but it's been helpful for me so far
[02:28] felixge: captain_morgan: http://github.com/felixge/node-stress
[02:28] captain_morgan: I can pretty well guess what it does.. just didn't want to rewrite it to run the benchmarks
[02:28] captain_morgan: rad, thanks
[02:28] felixge: captain_morgan: it should work fairly well for "sync" tests. Async support is also in there, but it will run all tests at the same time unless you put them in seperate files
[02:29] captain_morgan: ok
[02:29] captain_morgan: and is this something you are continuting to work with/develop?
[02:29] felixge: captain_morgan: I'd like to work on it more, yes
[02:29] felixge: I like the main idea of it a lot
[02:29] felixge: you tell it how many "runs" you want and how much time to spend on each run
[02:30] felixge: and it will double the "iterations" variable passed to your benchmark until the benchmark takes long enough to run
[02:30] felixge: this way the benchmark should have absolutely no effect on what is being benchmarked, and you don't have to figure out the right values
[02:31] felixge: one thing I'd like to do is support for slopes to give some ideas on the Big-O traits of a function
[02:31] felixge: but not sure when I'll find the time
[02:31] quirkey has joined the channel
[02:31] captain_morgan: sounds cool
[02:31] captain_morgan: I'll be knock this around a little over the next couple days
[02:32] captain_morgan: thinking I will be using Dirty for a project i'm working on
[02:33] felixge: captain_morgan: :)
[02:33] felixge: let me know if you run into any issues
[02:33] felixge: the branch version should work well right now
[02:33] felixge: but I'm going to polish it up further before JSConf
[02:34] captain_morgan: cool
[02:34] captain_morgan: I'll give you some feedback on how it goes
[02:34] felixge: captain_morgan: cool, thanks
[02:34] captain_morgan: so far SQLite, CouchDB, and Mongo have not worked out
[02:35] captain_morgan: or I'm being fickle
[02:41] Tim_Smart: hah, just released securable modules would break the entire test suite
[02:44] ditesh|cassini has joined the channel
[02:48] BryanWB has joined the channel
[02:50] JimBastard_ has joined the channel
[02:50] JimBastard_: JAVASCRIPT PARTY
[02:50] JimBastard_: sup Tim_Smart 
[02:51] jed_ has left the channel
[02:51] Tim_Smart: ello JimBastard_ 
[02:51] tmpvar_ has joined the channel
[02:57] tmpvar: yo JimBastard 
[02:57] JimBastard_: hey hey
[02:58] felixge: captain_morgan: what was the problem? Limitations of the db? Or client situation in node?
[03:00] felixge: anyway, 4am ... I think it's bed time
[03:01] RayMorgan has joined the channel
[03:02] jashkenas has joined the channel
[03:10] CIA-77: node: 03isaacs 07master * r31ed37f 10/ (src/node.cc test/simple/test-eval-cx.js): 
[03:10] CIA-77: node: evalcx shouldn't be too fancy
[03:10] CIA-77: node: After getting some feedback from Mikeal Rogers and Tim Smart, it was decided
[03:10] CIA-77: node: that evalcx should not try to do any fancy security stuff, and instead leave
[03:10] CIA-77: node: that in the hands of the user. To comply more with spidermonkey, everything
[03:10] CIA-77: node: is passed in, and objects are passed in by reference rather than being
[03:10] CIA-77: node: cloned. - http://bit.ly/adob7a
[03:12] binary42 has joined the channel
[03:24] blazzy_ has joined the channel
[03:29] BryanWB has joined the channel
[03:29] rtl has joined the channel
[03:32] r11t has joined the channel
[03:32] rednul has joined the channel
[03:33] JimBastard_ has joined the channel
[03:34] JimBastard_ has joined the channel
[03:39] ditesh|cassini has joined the channel
[03:54] joshbuddy has joined the channel
[03:54] joshbuddy has joined the channel
[03:54] BryanWB has joined the channel
[04:00] joshholt has joined the channel
[04:01] markwubben has joined the channel
[04:03] ssteinerX has left the channel
[04:06] creationix has joined the channel
[04:13] joshholt_ has joined the channel
[04:30] ryandotsmith has joined the channel
[04:30] PyroPeter has joined the channel
[04:31] quirkey has joined the channel
[04:32] jed has joined the channel
[04:44] isaacs has joined the channel
[04:45] JimBastard_ has joined the channel
[04:58] cloudhead has joined the channel
[05:10] mikeal has joined the channel
[05:12] okito has joined the channel
[05:12] chakrit1 has joined the channel
[05:13] chakrit2 has joined the channel
[05:17] mattly: is there a writeup somewhere on how to use narwhal stuff in node?
[05:21] bpot has joined the channel
[05:22] tlrobinson: mattly: it's still somewhat experimental, but kriskowal would be the person to ask
[05:22] mattly: heh
[05:24] tlrobinson: mattly: basically you need kris's branch of node until ryan merges some of the changes in. eventually i'd like narwhal to be an add on to a stock build of node http://github.com/kriskowal/node
[05:24] aguynamedben has joined the channel
[05:25] mattly: tlrobinson: gotcha
[05:25] mattly: i'm considering porting a fledgling rack (ruby) app to node, i think it would be a good fit, but i'm looking for an implementation of wsgi/rack and jack seems the most mature
[05:29] creationix has left the channel
[05:30] dnolen has joined the channel
[05:50] kriszyp has joined the channel
[05:58] mcarter has joined the channel
[06:01] jed has joined the channel
[06:05] gf3 has joined the channel
[06:06] Yuffster has joined the channel
[06:10] teemow has joined the channel
[06:11] isaacs has joined the channel
[06:35] jed has joined the channel
[06:39] kris928 has joined the channel
[06:39] micheil: why do I always not catch tim >_>
[06:40] kris928: Hey all, I'm trying to setup a new node.js machine except it is leopord and not snow leopord this time. Node.js installed ok but now kiwi cannot find express which I installed. I'm getting an error that process.fs is undefined in kiwi. Any ideas?
[06:41] r11t has joined the channel
[06:43] micheil: kris928: last I knew, kiwi wasn't part of node, it's more then likely an issue with kiwi
[06:43] kris928: I think it has something to do with node.js not installing completely because some system libraries were not found
[06:43] micheil: kris928: probably best to open up a bug / issue on the project's github issues page
[06:43] kris928: when I run configure, I get "not found" on ssy/eventfd.h and sys/epoll.h
[06:44] micheil: kris928: try doing starting node-repl
[06:44] micheil: no
[06:44] micheil: I don't think they are used on os x
[06:44] micheil: or create a empty file, do "var fs = require("fs") " in it
[06:44] micheil: and try running it
[06:44] kris928: I just got the same thing working on snow leopord machine earlier today so I'm not sure it's a problem with kiwi. I'll try that
[06:46] kris928: yeah process.fs.stat("foo") after require("fs") fails in node-repl too
[06:48] JimBastard__ has joined the channel
[06:48] kris928: micheil, any ideas why that might be?
[06:49] JimBastard__: wow it seems i have a few friends logged in
[06:50] micheil: kris928: sorry, no, I haven't been working with node for a little while
[06:50] kris928: micheil, nevermind I know why that is in repl - strange that this worked earlier today
[06:57] jed has joined the channel
[07:03] tmpvar: night
[07:08] kris928: micheil, FYI got it all working. Looks like the most recent version of node is not compatible with kiwi, I just reverted to an older revision for now.
[07:14] BryanWB has joined the channel
[07:15] JimBastard__: i love how my github languages tab usually comes up 100% javascript
[07:19] kjeldahl has joined the channel
[07:47] Gruni has joined the channel
[07:48] CIA-77: node: 03Krishna Rajendran 07master * r84277ea 10/ (src/node_dns.cc test/disabled/test-dns.js): Check for callback argument in PTR lookup - http://bit.ly/aeeu8U
[07:50] sveisvei has joined the channel
[08:25] markwubben_ has joined the channel
[08:26] markwubben_ has joined the channel
[08:27] ssteinerX_ has joined the channel
[08:43] voodootikigod_ has joined the channel
[08:45] Gruni has joined the channel
[08:46] QuietKnight has joined the channel
[08:54] sveimac has joined the channel
[08:56] hassox has joined the channel
[09:01] Tim_Smart has joined the channel
[09:15] Tim_Smart has joined the channel
[09:22] felixge has joined the channel
[09:22] felixge has joined the channel
[09:23] javajunky has joined the channel
[09:24] jed has joined the channel
[09:32] ashb: mjijackson: yeah towards the end of markdown.js
[09:32] ashb: search for Maruku
[09:33] lifo has joined the channel
[09:35] tisba has joined the channel
[09:39] sveimac has joined the channel
[09:39] lifo_ has joined the channel
[09:44] Tim_Smart: ashb: Where was that JS implementation of Markdown again? Might replace showdown with it
[09:48] ashb: http://github.com/evilstreak/markdown-js/
[09:48] ashb: i should merge it so its 1 file
[09:48] ashb: it doesn't need to be 2
[09:49] sztanpet: god i hate plain markdown with a passion, github flavored is ok
[09:49] ashb: 'falvoured' is part of the problem with markdown - there are like 7000 different flavour
[09:50] ashb: but one of the things we wanted to do with that markdown impl was make it easy to add dialects without having to re-write everything
[09:52] Tim_Smart: ashb: Is there any HTML -> Markdown for JS?
[09:53] bridge has joined the channel
[09:53] Tim_Smart: At present I am using a PHP solution called markdownify
[09:53] ashb: Tim_Smart: not yet - its not something we've needed to do
[09:54] Tim_Smart: OK. That's cool
[09:54] ashb: html -> md is a lot easy than plain text -> md tho
[09:54] ashb: oh and one fairly big caveat: we dont yet support inline HTML
[09:54] ashb: cos I personally think its an abomination ;)
[09:55] ashb: patches welcome tho
[09:56] Tim_Smart: Ah OK. Some guy got into a argument with me because he didn't like <pre><code> tags
[09:56] Tim_Smart: Which Markdown uses, and unfortunately he is very opinionated
[09:57] ashb: what's his argument?
[09:57] Tim_Smart: (btw I prefer <pre><code> tags)
[09:59] Tim_Smart: ashb: http://userscripts.org/topics/48140#posts-229090
[10:00] ashb: 'its visually unappelaing' has *nothing* to do with the markup
[10:02] ashb: yeah i dont buy his argument
[10:02] Tim_Smart: Yeah he tries to justify himself later on down :D
[10:03] ashb: and i'm not quite sure what classifies as 'ugly' to him
[10:04] ashb: but *anyway*
[10:04] ashb: markdown.js gives you an intermediate representation on JsonML when its still markdown
[10:04] ashb: so he could make it output WTF he likes for codeblocks
[10:05] Tim_Smart: Yeah, he is just hard to please
[10:06] Tim_Smart: I gave him a code suggestion later on, and the same guy pretty much pulled a tantrum, and muttered something bout security
[10:06] ithinkihaveacat has joined the channel
[10:07] Tim_Smart: then decided I was the worst thing that ever crossed his path
[10:07] Tim_Smart: so meh
[10:11] felixge has joined the channel
[10:14] ashb: :)
[10:33] BryanWB has joined the channel
[11:03] joshholt_ has joined the channel
[11:05] ryandotsmith has joined the channel
[11:06] ryandots_ has joined the channel
[11:13] mjijackson_ has joined the channel
[11:16] Joshua__Holt_ has joined the channel
[11:17] joshholt_ has joined the channel
[11:18] joshholt_ has joined the channel
[11:19] joshholt_ has joined the channel
[11:20] MattJ has joined the channel
[11:20] Joshua__Holt_ has joined the channel
[11:22] joshholt_ has joined the channel
[11:27] Joshua__Holt_ has joined the channel
[11:30] voodootikigod_ has joined the channel
[11:37] kjeldahl has joined the channel
[11:43] binary42 has joined the channel
[11:43] kjeldahl has joined the channel
[12:00] mooit has joined the channel
[12:11] drostie has joined the channel
[12:18] kjeldahl has joined the channel
[12:25] maritz has joined the channel
[12:31] alex-desktop has joined the channel
[12:34] drostie has joined the channel
[12:35] ryandotsmith has joined the channel
[12:39] ryandotsmith has joined the channel
[12:47] mjijackson_ has joined the channel
[12:49] mjijackson_ has joined the channel
[12:54] kjeldahl has joined the channel
[12:54] pmuellr has joined the channel
[12:57] davidsklar has joined the channel
[12:59] ryandotsmith has joined the channel
[13:07] kriszyp has joined the channel
[13:11] isaacs has joined the channel
[13:12] quirkey has joined the channel
[13:13] BryanWB has joined the channel
[13:21] kjeldahl has joined the channel
[13:22] kjeldahl has joined the channel
[13:30] khaase has joined the channel
[13:30] khaase: _ry: poke
[13:31] JAAulde has joined the channel
[13:31] JAAulde: morning all
[13:36] mjijackson has joined the channel
[13:37] joshholt_ has joined the channel
[13:37] alexmchale has joined the channel
[13:38] mjijackson_ has joined the channel
[13:49] paul_ has joined the channel
[13:50] confounds has joined the channel
[13:55] gf3 has joined the channel
[13:59] rtl has joined the channel
[14:01] ashb has joined the channel
[14:05] binary42 has joined the channel
[14:09] rolfb has joined the channel
[14:17] stepheneb has joined the channel
[14:18] alex-desktop has joined the channel
[14:31] aryounce has joined the channel
[14:33] n8o_ has joined the channel
[14:45] Yuffster has joined the channel
[14:51] jed_ has joined the channel
[14:53] brapse has joined the channel
[14:55] jed__ has joined the channel
[14:57] jed___ has joined the channel
[15:05] alexiskander has joined the channel
[15:11] jherdman has joined the channel
[15:13] gwoo has joined the channel
[15:23] dnolen has joined the channel
[15:25] kjeldahl has joined the channel
[15:26] creationix has joined the channel
[15:26] kjeldahl has joined the channel
[15:30] jherdman has joined the channel
[15:31] bentomas has joined the channel
[15:32] qFox has joined the channel
[15:32] bentomas: creationix:  I can't get node-persistence to work, what version of Node is it compatible with currently?
[15:32] bentomas: (specifically sqlite)
[15:32] creationix: good question
[15:32] creationix: 0.1.30 was the last I tested I think
[15:32] creationix: I'm going to change the API this week though, be warned
[15:33] stepheneb has joined the channel
[15:33] joshbuddy has joined the channel
[15:33] bentomas: yeah, you mentioned that, but I'm trying to get some sort of connection to sqlite up an running. I can change my code later this week if necessary
[15:34] creationix: well, good luck
[15:34] bentomas: thanks!
[15:35] creationix: I should have the new version tonight (about 13 hours)
[15:35] jed_ has joined the channel
[15:35] bentomas: oh wow, well maybe I can find something else to do in the mean time
[15:40] christkv has joined the channel
[15:41] MattJ has joined the channel
[15:49] jed_ has left the channel
[15:54] dnolen has joined the channel
[15:57] deanlandolt has joined the channel
[15:57] RayMorgan has joined the channel
[16:02] QuietKnight has joined the channel
[16:04] khaase has joined the channel
[16:07] okito has joined the channel
[16:08] joshbuddy has joined the channel
[16:08] joshbuddy has joined the channel
[16:10] tlynn has joined the channel
[16:14] blackdog has joined the channel
[16:19] javajunky has joined the channel
[16:26] mrjjwright has joined the channel
[16:29] kjeldahl has joined the channel
[16:29] tjholowaychuk has joined the channel
[16:30] aryounce has joined the channel
[16:35] tisba has joined the channel
[16:36] r11t has joined the channel
[16:38] brandon_beacher has joined the channel
[16:40] khaase has joined the channel
[16:40] r11t_ has joined the channel
[16:45] r11t__ has joined the channel
[16:51] bpot has joined the channel
[16:55] RayMorgan_ has joined the channel
[17:01] stephenlb has joined the channel
[17:05] RayMorgan has joined the channel
[17:08] stepheneb has joined the channel
[17:09] bpot_ has joined the channel
[17:09] okito has joined the channel
[17:13] aguynamedben has joined the channel
[17:16] pyrotechnick has joined the channel
[17:17] pyrotechnick: is /doc/api.txt the source file of http://nodejs.org/api.html?
[17:17] codeswing has joined the channel
[17:18] whoahbot has joined the channel
[17:20] ashb: pyrotechnick: for the moment, yeah think so
[17:21] pyrotechnick: ashb: aight
[17:21] pyrotechnick: cheers
[17:21] joshholt__ has joined the channel
[17:21] tjholowaychuk: its asciidot i think
[17:21] tjholowaychuk: something like that
[17:21] xueyu has joined the channel
[17:21] ashb: it's moving to markdown iirc
[17:21] ashb: creationix: speaking of which - markdown.js is now just the one file
[17:28] creationix: ashb: thanks
[17:28] creationix: I'll look into it, probably tomorrow
[17:30] JimBastard: sup creationix 
[17:31] pyrotechnick: creationix: good work on markdown ^w^ KISS FTW
[17:31] dandean has joined the channel
[17:32] mjijackson has joined the channel
[17:33] pyrotechnick: creationix: i just found a silly typo in your "what is this" how to node article if u wanna checkit
[17:33] creationix: sure, what is it?
[17:34] pyrotechnick: pull requested
[17:36] pyrotechnick: how many of you prodogs use coffeescript?
[17:36] jherdman: hey JimBastard. did you ever figure out your AOP problem from yesterday?
[17:37] mattly: pyrotechnick: +1
[17:37] JimBastard: not really, its a shitty problem. the real solution like voodootikigod pointed out is to step back and create proper tests
[17:37] JimBastard: im gonna use window.onerror as a stop-gap for now
[17:37] pyrotechnick: mattly: whatchu writing?
[17:37] JimBastard: so i have SOME idea if the client errors
[17:38] pyrotechnick: node stuff i presume...
[17:38] mikeal has joined the channel
[17:38] tjholowaychuk: where is this markdown.js ?
[17:38] mattly: pyrotechnick: porting over a rack-based cms type app thing
[17:38] JAAulde: Is the owner of other FreeBSD build slave here?
[17:39] mattly: sadly there don't seem to be any decent rack-like interfaces that work well without promises, yet
[17:39] pyrotechnick: mattly: you've seen jack?
[17:39] pyrotechnick: mattly: oh...
[17:39] pyrotechnick: mattly: old code?
[17:39] creationix: tjholowaychuch: I think he's talking about http://github.com/evilstreak/markdown-js
[17:39] mattly: pyrotechnick: yeah they're all pre 0.1.30 or something
[17:39] creationix: I'm currently using a modified version of showdown, but I plan on switching to the other once it's ready
[17:39] pyrotechnick: mattly: is there much code in jack?
[17:39] mattly: pyrotechnick: and jack is narwhal, which i understand is sort-of runable on node but that's still experimental
[17:39] pyrotechnick: ohh
[17:40] tjholowaychuk: creationix: ah cool, I have a way lighter implementation in the works
[17:40] tjholowaychuk: just no repo yet
[17:40] creationix: sounds cool
[17:40] creationix: I'm a fan of lightweight
[17:40] pyrotechnick: creationix: we've put haml-js to good use
[17:40] pyrotechnick: http://github.com/DaNCeR/requisition-test
[17:40] ashb: creationix: tjholowaychuk: yeah thats the one
[17:41] creationix: as long as I can extend it with custom tags
[17:41] ashb: tjholowaychuk: our needs are stated in the reeadme
[17:41] creationix: pyrotechnick: cool
[17:41] ashb: mainly its 1) not direct text->text, 2) well-formed html 3) not have to re-write everything to add a new dialect (of which there are about 7000)
[17:41] pyrotechnick: there was something my BP wanted to ask you but i forget now
[17:42] mattly: pyrotechnick: part of the problem is that the *sgi interface specifies that the app's function should _return_ something, which becomes very tricky or even impossible in callbacks
[17:42] mattly: without promises
[17:42] pyrotechnick: there are still promises i believe
[17:42] pyrotechnick: i've seen some >0.1.30 code with them
[17:43] ashb: I'm a big fan of processing steps exposing their intermeidate forms
[17:43] mattly: i modified jsgi-node to pass in a respond function in addition to the request and then i just call respond
[17:43] mattly: but
[17:43] pyrotechnick: mattly: let me have a look
[17:43] mattly: Error: Promise has been removed. See http://groups.google.com/group/nodejs/msg/0c483b891c56fea2 for more information.
[17:43] mattly:     at new <anonymous> (node.js:11:11)
[17:43] mattly: that's from both "require('promise')" and "require('events').Promise"
[17:43] pyrotechnick: oh reallly
[17:44] pyrotechnick: grr im trying to remember this code
[17:44] mattly: i may just write my own :D
[17:44] mattly: or :'(
[17:45] atmos has joined the channel
[17:45] pyrotechnick: mattly: we were heaps keen on jack, so i'm kinda sad now i hear the implementations have sunken down the sink
[17:45] pyrotechnick: i didnt much like the syntax though
[17:46] pyrotechnick: so maybe this was a good thing ultimately
[17:46] ashb: what syntax?
[17:46] pyrotechnick: like chaining them all together
[17:46] ashb: oh the promise sytnax?
[17:46] pyrotechnick: no jack
[17:46] pyrotechnick: i didnt mind promises TBH
[17:46] ashb: jack is a module - what syntax does it hack?
[17:46] ashb: *have
[17:46] pyrotechnick: i had kind of dreamt for similar facilities in PHP
[17:47] ashb: or am i thinking of a differetn jack?
[17:47] pyrotechnick: jack like rack for node
[17:47] ashb: (entierly possible - there are like 40 JS projects named jack i'm sure)
[17:47] ashb: yeah - its just a module isn't it - what syntax does it have?
[17:48] mattly: well it's not the syntax so much as it's not really evented
[17:48] ashb: jack != JSGI
[17:48] mattly: right
[17:48] mattly: ejsgi looks promising
[17:48] pyrotechnick: mattly: http://howtonode.org/do-it-fast
[17:49] ashb: can't remember exactly what ejsgi looks like anymore
[17:49] ashb: but lst time i looked at the async options for jsgi it was my preference.
[17:49] mattly: http://github.com/isaacs/ejsgi
[17:49] ashb: i'm not really a fan of the boilerplate you have to use to use promises
[17:50] mattly: honestly i'd rather do what i did to esjgi, just modified it to pass in a respond function to call
[17:50] pyrotechnick: i imagine you could substitute the promise stuff in ejsgi for that in do
[17:50] ashb: mattly: i propsed that a while ago i think
[17:50] khaase has joined the channel
[17:50] mattly: function app(request, respond) { respond({status: 200, headers:{}, body:["Hello World"]}) }
[17:51] ashb: mattly: that doesn't let you stream the body tho
[17:51] pyrotechnick: lol
[17:51] pyrotechnick: http://github.com/ithinkihaveacat
[17:51] pyrotechnick: "Like JSGI, but using streams for the request bodies."
[17:51] creationix: I actually don't use Do much, I have a minimilist version of flow-js I use for everything now
[17:51] pyrotechnick: is he lying then?
[17:52] tjholowaychuk: I find if you abstract your routines enough you dont really need some fancy thing like Do and friends
[17:52] tjholowaychuk: at least I have not come across a need for them yet
[17:53] ashb: http://gist.github.com/251804
[17:53] ashb: mattly: the first form there
[17:53] JimBastard: im using event emitters and named events
[17:53] JimBastard: my event chain isnt too complex though
[17:53] ashb: http://gist.github.com/251577
[17:53] mattly: ashb: interesting
[17:54] ithinkihaveacat: pyrotechnick: ? that description on my github page comes from isaacs's ejsgi (mine's a fork)
[17:54] ashb: mattly: stolen wholesale from Perl's plack
[17:54] mattly: heh
[17:54] pyrotechnick: ithinkihaveacat: i know, i was lolling at your name
[17:54] ashb: why bother infenting something when you can seal it :)
[17:54] pyrotechnick: ^w^ because it's awesome
[17:54] rauchg has joined the channel
[17:55] pyrotechnick: ithinkihaveacat: i'm curious though, have you verified whether or not you have a cat yet?
[17:55] ithinkihaveacat: pyrotechnick: heh, okay
[17:56] mattly: ashb: where does the responder come from in that gist?
[17:56] ashb: the [EA]JSGI server passes it in
[17:56] ashb: doesn't really need the extra function
[17:57] mattly: ah
[17:57] ithinkihaveacat: pyrotechnick: i still can't be certain
[17:57] pyrotechnick: ithinkihaveacat: man that cow video on your blog is uberdisturbing
[17:57] ashb: could just as easily be a property on the request, or a 2nd argument
[17:57] mattly: maybe "request.respond()"
[17:58] ashb: tbh its probably pretty damn similar to EJSGI
[17:58] ithinkihaveacat: pyrotechnick: sometimes i think i do, sometimes not
[17:58] ashb: mattly: http://github.com/isaacs/ejsgi/blob/master/examples/baseline.js
[17:58] ashb: compare to that
[17:58] mattly: i think the main problem is that the jsgi specification says that a function should return its response
[17:58] ashb: sure - EJSGI isn't strictly JSGI and doesn't have to be
[17:58] mattly: and without promise-like contraptions that's just hard to do
[17:59] mattly: right
[17:59] ithinkihaveacat: pyrotechnick: yeah, they're belgian cows, they're great.  they were curious about me, and i'm from the city so i was curious about them as well
[17:59] ashb: to my mind it just should have the same feel
[17:59] pyrotechnick: ithinkihaveacat: you should write that underneath it
[17:59] ashb: the one major different between my proposal and EJSGI as it stands is:
[17:59] mattly: I like request.respond({status:200, body:["Hello World"]}) as an api
[17:59] ashb: my proposal you can't get hold of a stream until you've sent the headers
[17:59] ashb: explicitly.
[18:00] mattly: nicely mirrors rack's app.call
[18:00] ashb: mattly: sure - but thats sync
[18:00] mattly: hm
[18:00] mattly: ok so i'm still new to node :D
[18:00] mattly: and thinking this way
[18:00] pyrotechnick: HES STILL DEMANDING INPUT IN HIS HELLO WORLD APP
[18:01] deanlandolt: ashb: have you given any thought to the async forEachable object as stream?
[18:01] ashb: speaking of hello worlds apps: http://blog.ianbicking.org/2010/03/16/web-server-benchmarking-we-need/
[18:01] ashb: deanlandolt: not givne much thought to ajsgi since december :)
[18:01] stepheneb_ has joined the channel
[18:02] ashb: deanlandolt: but an explicit 'write this chunk' rather than 'here let me return the next chunk' fits better with my mental models
[18:02] deanlandolt: heh...yeah, same here...i've been using the async forEachable thing (using kriszyp's LazyArray stuff) for quite a while now and i'm loving it though
[18:02] deanlandolt: ashb: that's what the asyn cforEachable does
[18:02] tmpvar has joined the channel
[18:03] mattly: bah
[18:03] tmpvar_ has joined the channel
[18:03] mattly: maybe i'll just do the simplest thing and say fuck all
[18:03] pyrotechnick: has anyone run into anyone else using node for gameservers?
[18:03] deanlandolt: just like with current jsgi where you can return a response -- the body must have a forEach functoin...so when you're setting it up it you have a "write this chunk" setup...
[18:03] mattly: and build some sort of middleware thing on top of the standard http stuff
[18:03] xueyu_ has joined the channel
[18:03] deanlandolt: the "promise" part is really only to communicate that you're done writing
[18:03] creationix: http://github.com/ry/node_docs/blob/master/index.js#L12
[18:04] khaase has left the channel
[18:04] creationix: pyrotechnick: I'm making one in the next couple of months
[18:04] pyrotechnick: creationix: no way rly?
[18:04] QuietKnight has joined the channel
[18:04] ashb: deanlandolt: not a fan of promises :/
[18:04] creationix: yep, though I don't think I'll release anything till after my talk at TXJS
[18:05] pyrotechnick: creationix: let it bake, let it bake
[18:05] pyrotechnick: creationix: what technologies for the client? you're not tackling webgl are you?
[18:05] creationix: no, either canvas or raphaeljs
[18:05] pyrotechnick: raphael is svg right?
[18:06] creationix: svg and vml
[18:06] pyrotechnick: works in ie too yeah?
[18:06] creationix: so it works in IE
[18:06] pyrotechnick: sick
[18:06] pyrotechnick: we're probs gunna draw our UI with that
[18:06] ashb: raphaeljs++
[18:06] mjr_ has joined the channel
[18:06] ashb: you can't use CSS with it, but apart from that its wonderful
[18:06] pyrotechnick: to hell with css
[18:06] creationix: I think I'll use raphael since it's got built-in tweening and stuff
[18:06] creationix: handy for games
[18:07] pyrotechnick: yeah and the performance looks pretty damn impressive too
[18:07] creationix: for simple stuff, I'll find how far I can push it
[18:07] tjholowaychuk: I want to port cocos2d to node *drool*
[18:07] pyrotechnick: i've seen some people rewriting the flash runtime in it
[18:07] creationix: their demo page is crap slow on my netbook though
[18:08] pyrotechnick: tjholowaychuk: how does that have anything to do with serverside js?
[18:08] tjholowaychuk: pyrotechnick: it does not really, but why limit node to servers? writing a game in JS vs objective-c would be much more fun
[18:08] mjr_: It'd be pretty handy to make graphics in node and then ship them out to browser clients.
[18:09] tmpvar_: creationix, nice work on the docs
[18:09] pyrotechnick: tjholowaychuk: i see i see, are there any bindings to like SDL for javascript or something
[18:09] creationix: thanks
[18:09] tmpvar has joined the channel
[18:09] tjholowaychuk: pyrotechnick: not that I know of, started some the other day but cant find time right now
[18:09] tmpvar: stupid xfce keeps launching 2 copies of xchat 
[18:09] creationix: tjholowaychuck: I'd love bindings to webview (the html renderer from webkit)
[18:10] creationix: for local games and apps
[18:10] pyrotechnick: mjr_: you can, i have seen bindings to imagemagik and stuff
[18:10] tmpvar: creationix, what about bindings to something like cairo?
[18:10] pyrotechnick: mjr_: on the node modules page
[18:10] creationix: works too, it's just that most javascript people already know html css and canvas
[18:10] QuietKnight: I've written a basic PNG encoder in pure JS, but havent released it (yet)
[18:11] pyrotechnick: QuietKnight: that's intense
[18:11] tmpvar: lol
[18:11] tjholowaychuk: to many choices lol
[18:11] tjholowaychuk: im having a hard time even choosing between box2d and chipmunk
[18:11] tjholowaychuk: ive only used chip
[18:11] QuietKnight: planning on using it in a game. Dynamic graphics on the client if possible (data urls), delegating to server is not possible (node)
[18:11] pyrotechnick: creationix: can you tell me anymore about your game?
[18:12] tjholowaychuk: cairo's api is nice tho, not as ugly as most c shit
[18:12] mjr_: creationix: just sent you a pull request.  Let me know what you think of my example code.  I love your doc framework.
[18:12] stepheneb has joined the channel
[18:12] tmpvar: tjholowaychuk, i think the cool thing about cairo is that it works everywhere
[18:12] tjholowaychuk: that too
[18:12] creationix: pyrotechnick: it will be real-time battles in space across the internet
[18:12] tjholowaychuk: I will look into it some more
[18:13] tmpvar: creationix, you are going to need udp support for that, right?
[18:13] pyrotechnick: sockets
[18:13] pyrotechnick: are probs fast enough in our experience
[18:13] tmpvar: tcp = slow, udp = lightning
[18:13] mjr_: You can do some pretty reasonable gaming over TCP most of the time.
[18:13] creationix: well, browsers don't do udp
[18:14] pyrotechnick: yeah i was having this debate with my BP yesterday
[18:14] ashb: tcp isn't strictly slow
[18:14] pyrotechnick: people are unaware that many triple a game titles have been using tcp for years
[18:14] mjr_: I'm doing live voice over HTTP, and it is often great.
[18:14] tjholowaychuk: and cairo has REAL docs OMG haha, thats a first
[18:14] creationix: besides I'll keep the traffic to a minimum and use websockets so it should be plenty fast.
[18:14] tmpvar: ^_^
[18:14] QuietKnight: creationix: have you done actual tests yet for the client-side? Its funny because I've also looked into a web-based space game (think Subspace or Cosmic Rift), but found that SVG is too slow and canvas kills the CPU
[18:14] tmpvar: tjholowaychuk, you going to take the dive?
[18:14] tjholowaychuk: sure am
[18:15] tjholowaychuk: after a bit of research of course, but I will check it out
[18:15] creationix: I did make a game a while back and found that svg was too slow then
[18:15] creationix: sousaball.creationix.com
[18:15] mjr_: Over many mobile networks, TCP ends up being as fast as UDP because of air interface retransmission.
[18:15] tmpvar: nice, hit me up when/if you get a project setup.. ive been meaning to dive in for a while now
[18:15] creationix: it's canvas now
[18:15] pyrotechnick: i dont want to make you guys cry or anything
[18:16] pyrotechnick: but
[18:16] tjholowaychuk: tmpvar: sounds good
[18:16] pyrotechnick: if you have the audacity to click this link
[18:16] pyrotechnick: http://ie.microsoft.com/testdrive/Default.html
[18:16] pyrotechnick: youll see some nice canvas demos (they all seem to work on chromium nightlies)
[18:16] JimBastard: my friend is at MIX and hes all like yo the new IE is awesome, im like. microsoft makes an internet browser? really?
[18:16] pyrotechnick: JimBastard: lol. at least they're trying with ie9
[18:16] tjholowaychuk: fuck lol i hate them
[18:16] tjholowaychuk: so bad
[18:17] pyrotechnick: JimBastard: it's pages like this that give me hope
[18:17] ashb: JimBastard: they still haven't got standard support right
[18:17] tmpvar: lol
[18:17] tjholowaychuk: if they deprecated ALL their older browsers then I will stop hating
[18:17] JimBastard: surely it wont have 0-day errors that will root my machine
[18:17] ashb: http://twitter.com/jaffathecake/status/10614034213
[18:17] QuietKnight: I wound up just going pure HTML/JS for my game. Doesn't scale quite as well as canvas, but also doesnt max out the CPU
[18:17] QuietKnight: there is a great canvas library out there for games.. let me see if I can find it..
[18:18] pyrotechnick: QuietKnight: sweet
[18:18] ashb: webGL is better suited, no?
[18:18] mjr_: Ha.  SVG-oids works really well in Safari.
[18:18] tmpvar: chrome only games
[18:18] pyrotechnick: mjr_: yeah thems microsoft bitches can actually chrome
[18:18] QuietKnight: http://code.google.com/p/cakejs/
[18:19] QuietKnight: webGL is great, but I want to support as many browsers as possible :)
[18:19] pyrotechnick: webgl will be in all next major revisions
[18:19] pyrotechnick: which i dare say is less than 12 months off
[18:20] pyrotechnick: we've targeted july for our first game / platform /engine / editor
[18:20] tjholowaychuk: shit apparently I have cairo installed already
[18:20] tjholowaychuk: weird
[18:20] QuietKnight: http://www.effectgames.com/effect/ <-- also cool
[18:20] tjholowaychuk: do any of you guys make money off these games?
[18:20] pyrotechnick: we're not on target so we're kinda hoping it's a bit later than we first anticipated, more like 12 like i say now instead of 6 months
[18:20] pyrotechnick: tjholowaychuk: that's the plan
[18:20] QuietKnight: mine is still in early development, so no. But hopefully eventually
[18:20] tjholowaychuk: whats the business model for that
[18:21] QuietKnight: pyro, I didn't catch the beginning of the convo. Are you working with creationix or somebody else?
[18:21] pyrotechnick: tjholowaychuk: for us it's taking on steam in the webgl universe
[18:21] softdrink: node = blindingly easy to create data fakers for web dev
[18:21] pyrotechnick: we see theyve changed to webkit so we're expecting webgl games from them
[18:21] pyrotechnick: so we're kinda gunna try take a stab at webgl distribution bundled with the tech to make the games
[18:21] tjholowaychuk: hmm
[18:21] tjholowaychuk: sounds cool
[18:21] Gruni has joined the channel
[18:21] pyrotechnick: and make a pilot game or two as tech demos
[18:21] isaacs has joined the channel
[18:21] tjholowaychuk: I think it would be sweet to have a MMO RTS type thing maybe
[18:22] creationix: I'm just making a game for my TXJS talk in June
[18:22] tjholowaychuk: I guess people could make facebook games or something for revenue 
[18:22] QuietKnight: oh gotcha
[18:22] creationix: hopefully it will live beyond my talk though
[18:22] pyrotechnick: tjholowaychuk: yeah that's slated as our first game but we're deciding atm wether we want to switch
[18:22] pyrotechnick: tjholowaychuk: basically it's MMO simcity
[18:22] JimBastard: hey creationix i was thinking about yout idea last night about extending object for getting process.mixin type behaviors
[18:22] tjholowaychuk: I wrote a starcraft clone a while back haha was pretty fun just css / html tho
[18:22] pyrotechnick: but with a heavy emphasis on transport
[18:22] JimBastard: that might be a fun thing to work on at the hackathon
[18:23] creationix: sure, it wouldn't take more than an hour
[18:23] JimBastard: yaaaa
[18:23] QuietKnight: tjholowaychuk- did you do pathfinding? I found that to be one of the more difficult things
[18:23] JimBastard: i should send those invites out
[18:23] pyrotechnick: webgl people, are any of you keen to setup a community?
[18:23] pyrotechnick: kinda like a meta-blog
[18:24] tmpvar: QuietKnight, A* is fairly simple
[18:24] QuietKnight: yea, that wasnt the hard part
[18:24] QuietKnight: the hard part was creating a navigation mesh for A* to run on
[18:25] pyrotechnick: a* fails in js
[18:25] pyrotechnick: greedy FTW
[18:25] tmpvar: yeah
[18:25] creationix: I'm pretty sure I won't need A for my game (space physics are easy)
[18:25] creationix: A*
[18:26] pyrotechnick: creationix: is your ssjs going to be opensource
[18:26] creationix: of course
[18:26] creationix: after the presentation
[18:26] pyrotechnick: we'll totally commit
[18:26] Aduros has joined the channel
[18:27] QuietKnight: My final solution was: determine vertices from a tile-based map, run a constrained delaunay triangulation (removing "non-walkable" edges), then run A* on that.
[18:27] RayMorgan has joined the channel
[18:27] QuietKnight: the constrained delaunay triangulation was the hard part, for a number of reasons. Anyways, fun, but difficult
[18:27] tmpvar: yeah QuietKnight its pretty intensive (possible too much so for js)
[18:28] QuietKnight: I got it :)
[18:28] tmpvar: oh? where? :P
[18:28] tmpvar: pics or it isnt real
[18:28] QuietKnight: haha
[18:28] orlandov: QuietKnight: sorry i haven't been following the conversation, but on the topic of A* have you seen http://greenmayo.com/dev/CarterHenry/
[18:29] tmpvar: grid based a* is really easy haha
[18:29] pyrotechnick: he just walked through a wall...
[18:30] pyrotechnick: and again
[18:30] pyrotechnick: through a diagonal
[18:30] QuietKnight: grid-based, yea
[18:30] QuietKnight: taking a grid and "vectorizing" it is hard
[18:30] QuietKnight: one sec for pic
[18:30] pyrotechnick: maybe because he is a ninja
[18:31] orlandov: ninjas have no need for vectors!
[18:32] dnolen has joined the channel
[18:32] QuietKnight: http://img202.imageshack.us/img202/1245/pathing.png
[18:33] QuietKnight: oh yea. And theres a path "smoothing" technique after A*
[18:33] tmpvar: very nice
[18:33] QuietKnight: orange edges are navigation mesh. blue is result of A*. green is after smoothing
[18:34] pyrotechnick: dude that is ultra
[18:34] tmpvar: now for 3d :P
[18:34] unomi has joined the channel
[18:34] QuietKnight: tmpvar- hell no
[18:35] QuietKnight: yea, I've got it to the point where I can have two dudes walking around on the map (pathing), and real-time communication via websockets and node
[18:35] pyrotechnick: how come ur using the vertex and not the edges for a* or is that just how it's drawn?
[18:35] QuietKnight: pretty cool stuff
[18:36] tmpvar: QuietKnight, agreed that is cool, nice job
[18:36] pyrotechnick: your path seems to use the vertex too
[18:36] QuietKnight: pryo, eh? Using vertexes cause I need to get from point A to point B
[18:36] pyrotechnick: QuietKnight: hmm
[18:36] QuietKnight: obviously those are connected by the edges, each with a weight (the distance)
[18:37] pyrotechnick: that's waypoint though not a navmesh :(
[18:37] QuietKnight: no
[18:37] QuietKnight: a waypoint walks from predefined points
[18:37] tmpvar: i think the waypoint is calculated off of the navmesh
[18:37] QuietKnight: my start and end points are not predefined
[18:37] pyrotechnick: lol
[18:38] QuietKnight: furthermore, path smoothing definitely leaves the "waypoints"
[18:38] pyrotechnick: your green line should run through the center of each edge along the path
[18:38] pyrotechnick: i mean your blue line
[18:39] pyrotechnick: and then your green line should slightly deviate from this and curve for natural movement
[18:39] QuietKnight: http://www.ai-blog.net/archives/000152.html
[18:39] pyrotechnick: i write pathfinding algorithms for THQ ^w^
[18:39] pyrotechnick: well i implement them
[18:40] QuietKnight: err
[18:40] QuietKnight: do you mean why don't I use the edge midpoints as vertices in the nav mesh as well?
[18:40] pyrotechnick: no i mean you dont use vertices with a navmesh
[18:40] pyrotechnick: you use polygons
[18:41] pyrotechnick: you should have an adjacency graph of edges not vertices
[18:41] QuietKnight: dude
[18:41] QuietKnight: the polygons are the triangles
[18:41] QuietKnight: based on delaunary triangulation
[18:41] QuietKnight: then the vertices of the triangles are the nodes in the nav mesh
[18:41] pyrotechnick: yes
[18:41] pyrotechnick: and then the cost of cross an edge should be the distance between the center of each triangle
[18:42] pyrotechnick: lol
[18:42] pyrotechnick: navmeshes done have nodes
[18:42] pyrotechnick: they have polygons, be it triangles, but that's a polygon all the same
[18:42] pyrotechnick: i believe you have waypoints there
[18:42] pyrotechnick: but that's better in a maze anyway
[18:43] pyrotechnick: please trust me
[18:43] pyrotechnick: i'll try find some pictures
[18:43] pyrotechnick: see this?
[18:43] pyrotechnick: http://theory.stanford.edu/~amitp/GameProgramming/polygon-navmesh-faces.png
[18:43] tmpvar: yes
[18:43] QuietKnight: I was just trying to find that
[18:44] QuietKnight: I used that site (even picture) extensively
[18:44] QuietKnight: read the rest of that page though
[18:44] joshbuddy has joined the channel
[18:44] QuietKnight: you can use different things for the nav mesh
[18:44] QuietKnight: the center point (as in this pic), or the vertices, of the midpoints of the edges, etc
[18:44] QuietKnight: look at the rest of the pix there
[18:44] pyrotechnick: "Polygonal map"
[18:45] pyrotechnick: okay
[18:45] QuietKnight: can you link the actual page?
[18:45] QuietKnight: I seem to be having issues
[18:45] pyrotechnick: http://theory.stanford.edu/~amitp/GameProgramming/MapRepresentations.html
[18:45] tmpvar: node related question: who here is running node in production?
[18:45] QuietKnight: thanks
[18:45] pyrotechnick: WHAT A NODE Q IN HERE WTF?
[18:46] pyrotechnick: we're running a node proxy in production
[18:46] QuietKnight: this is what I'm doing: http://theory.stanford.edu/~amitp/GameProgramming/polygon-navmesh-vertices.png
[18:46] pyrotechnick: yeah i can see how you were mislead
[18:46] tmpvar: pyrotechnick, what sort of load does it see and how stable is it?
[18:46] QuietKnight: and a quotation from the page: "Using vertices is better for obstacle avoidance, and if you’re using path smoothing it won’t negatively affect path quality."
[18:47] QuietKnight: pyro- mislead?
[18:47] pyrotechnick: neither have anything to do with obstacle avoidance really that's quite a blind statement to make
[18:47] gJ|Alex has joined the channel
[18:48] QuietKnight: pathing is obstacle avoidance, no?
[18:48] pyrotechnick: pathfinding is pathfinding, it's your mesh that dictates obstacle avoidance 
[18:48] QuietKnight: fair enough
[18:49] pyrotechnick: but in your game
[18:49] pyrotechnick: do your characters walk through the center of the rooms or around the corners?
[18:49] QuietKnight: both? they follow the green line
[18:49] pyrotechnick: around the walls*
[18:50] steadicat has joined the channel
[18:50] pyrotechnick: i think you will have problems with larger rooms
[18:50] pyrotechnick: the only time that green path isnt following the edges is at the start and endpoints
[18:50] QuietKnight: in the first room it walks through it goes through the center of the room
[18:51] QuietKnight: it goes from "door" to "door"
[18:51] QuietKnight: and there is not an edge there
[18:51] QuietKnight: in the mesh. It does that due to path smoothing
[18:51] pyrotechnick: did you do this hybrid approach thing?
[18:51] QuietKnight: the blue line does walk around walls
[18:51] QuietKnight: no
[18:51] pyrotechnick: id say it's just the drawing maybe
[18:51] QuietKnight: you see the blue and green lines, right?
[18:51] pyrotechnick: i dunno dude if it looks good then it doesnt matter it's just i've never seen this before
[18:52] QuietKnight: I feel like you're thinking I walk the blue line
[18:52] pyrotechnick: nah im not silly
[18:52] brandon_beacher has joined the channel
[18:52] pyrotechnick: just i can forsee problems
[18:52] pyrotechnick: like with a wide character
[18:52] pyrotechnick: you'll want it to walk through the middle of the doorway
[18:52] QuietKnight: I have to create a new mesh for different character radius
[18:53] pyrotechnick: do you'll need to consider the edges and have them influence the path
[18:53] QuietKnight: I also have a radius in here (half a square)
[18:53] QuietKnight: *already
[18:53] pyrotechnick: please show me when it's in action dude i'm keen
[18:54] pyrotechnick: is that drawing with canvas?
[18:54] QuietKnight: yea
[18:54] pyrotechnick: that's awesome
[18:54] QuietKnight: final product doesnt use canvas, just used it for drawing the paths and such
[18:54] pyrotechnick: what will it draw with
[18:55] QuietKnight: plain ol HTML/Javascript
[18:55] pyrotechnick: orly
[18:55] tmpvar: pyrotechnick, no love :P
[18:55] pyrotechnick: that was you saying that performance sucked aye
[18:55] QuietKnight: characters will be rendered on the fly using that png encoder I mentioned:) and data uris
[18:55] JimBastard: yeah javascript party invites have been sent
[18:55] pyrotechnick: tmpvar: so distracted, all you need to do is mention games or anything affiliated and i'm gone
[18:55] tmpvar: word
[18:55] JimBastard: caus there aint no party like a javascript party caus a javascript party got node
[18:55] tmpvar: yeah, i hear yah.. 
[18:55] pyrotechnick: umm
[18:56] tmpvar: JimBastard, we got some lan action wut wut
[18:56] JimBastard: tim and i started to rework the node-persistence api a bit last night, its looking better already
[18:56] pyrotechnick: it's much better than our ruby solution which was using eventmachine so it was all streamed and evented
[18:57] pyrotechnick: QuietKnight: do you have a blog?
[18:57] QuietKnight: no, although I've done enough cool things with this game that I should start one
[18:57] pyrotechnick: definately
[18:57] pyrotechnick: we're at that stage too
[18:57] pyrotechnick: my BP said to me today he's like we're so overdue for a blog i have about 1001 articles in my head, i fully agreed
[18:58] mjr_: tmpvar: I've got a couple of node servers in production, but they are in roles that can tolerate outages.
[18:58] pyrotechnick: tmpvar: did you ask about stability?
[18:58] QuietKnight: pyro- what you mentioned earlier is very interesting. You're talking about Valve's Steam, right?
[18:58] pyrotechnick: i'm yet to see one crash
[18:58] pyrotechnick: yes
[18:58] pyrotechnick: QuietKnight: yes
[18:59] tmpvar: pyrotechnick, how much load to they experience?
[18:59] QuietKnight: thats a great idea to be able to use webgl with a large audience
[18:59] pyrotechnick: we run 30 on a single VPS, 512 megs of ram although the traffic to each node is quite low because of that
[19:00] pyrotechnick: QuietKnight: i'm not 100% sure that that's their motivation for moving to webkit, it might just be that they're preparing for their entry into mac and linux but i have a feeling they're going to try webkit out
[19:02] QuietKnight: well its a good catch and if it pans out could be quite lucrative
[19:02] pmuellr_ has joined the channel
[19:03] QuietKnight: btw, I'm fairly new to Node, but it has made writing the backend for this stuff an absolute breeze. So thanks all you core devs :P
[19:03] pyrotechnick: tmpvar: i am told they have peaked at 30K req/hour but we've had some crashes i didnt know about
[19:03] jherdman has joined the channel
[19:03] pyrotechnick: we are running a patched node though
[19:04] pyrotechnick: which is untested
[19:04] pyrotechnick: for more ssl support
[19:04] tmpvar: ah
[19:04] tmpvar: alright, thanks
[19:05] pyrotechnick: is that alot, i really have no idea i'm a programmer we've outsourced the hosting for this project, been there done that, never going back ever again
[19:05] pyrotechnick: they're just (almost) blind proxies though
[19:06] QuietKnight: tmpvar- have you seen the post on google re: trying to max out Node?
[19:06] tmpvar: ill look
[19:06] QuietKnight: not production runs, but quite interesting
[19:06] whoahbot has joined the channel
[19:06] pyrotechnick: the only real logic they do is screwing with headers so they appear transparent
[19:06] QuietKnight: http://groups.google.com/group/nodejs/browse_thread/thread/d19df83c7fd1f3ca/c7c6883bf85ae3a1
[19:06] pyrotechnick: i wouldn't exactly call them app servers
[19:07] QuietKnight: and in particular: http://groups.google.com/group/nodejs/msg/afb6e7273042f90a
[19:08] QuietKnight: again, not production, but does kind of give an idea of what might be possible
[19:09] QuietKnight: pyro- one other small thing you might be interested in, and vaguely on the topic of SSL. I can't support SSL (due to legacy websockets stuff), so I'm implementing SRP for login authentication. Its pretty cool, and has a side-effect of providing private session keys for encrypting certain pieces of traffic
[19:11] pyrotechnick: legacy websockets stuff?
[19:11] pyrotechnick: that sounds like a PITA
[19:11] pyrotechnick: cool solution though
[19:13] javajunky has joined the channel
[19:15] pyrotechnick: QuietKnight: how often do you regen the keys?
[19:16] jed_ has joined the channel
[19:17] drostie has joined the channel
[19:19] QuietKnight: I don't really plan on using the keys (as in I don't plan on encrypting much traffic)
[19:19] QuietKnight: I could change my mind, but I'm not sure how necessary encrypting "Player A moves to square (1,1)" is
[19:21] QuietKnight: but its just nice to have them around, in case I do want to encrypt certain pieces of information (using a CC from in-game, for instance)
[19:21] okito has joined the channel
[19:23] pyrotechnick: mmm
[19:23] pyrotechnick: CC requires ssl though at least in non 3rd world countries
[19:23] pyrotechnick: but i get your point
[19:24] pyrotechnick: QuietKnight: have you found canvas to be really that slow?
[19:24] QuietKnight: it requires certification (not sure if SSL is part of that certification; that you may know). But yea, the point's the same.
[19:24] QuietKnight: not slow, just cpu consuming
[19:24] QuietKnight: similar to flash
[19:25] QuietKnight: I also didn't spend a whole awful lot of time with it. I considered writing my own scenegraph with certain optimizations I was thinking about, but decided against it
[19:25] pyrotechnick: when using a draw loop?
[19:25] QuietKnight: yea
[19:25] bentomas has joined the channel
[19:26] pyrotechnick: i imagine you can code quite well seeing that pathfinding stuff so this isnt an attack on you but i've seen some fucking dodgy draw code aye i hope canvas doesnt get a bad name from code like that
[19:26] QuietKnight: CAKEJS is quite amazing, but it re-renders every object on each loop
[19:26] QuietKnight: I was using cakejs
[19:26] QuietKnight: which yea, isnt exactly optimized
[19:26] pyrotechnick: so you were contemplating doing like vnc kinda shit?
[19:26] pyrotechnick: i think flash does that now too
[19:27] pyrotechnick: i forget what they're called
[19:27] pyrotechnick: redraw regions or something
[19:27] QuietKnight: yea, partial updating or rendering or whatever
[19:27] QuietKnight: yea
[19:27] pyrotechnick: that came with AS3.0 if i'm not mistaken
[19:27] QuietKnight: which means I'd need to calc bounds, and see what was stacked on top of each other, and redraw those objects + affected objects
[19:28] QuietKnight: definitely possible
[19:28] QuietKnight: just... yea.
[19:28] pyrotechnick: yeah not fun
[19:28] QuietKnight: other thing was IE6&7.
[19:28] pyrotechnick: so you're going for the masses?
[19:28] Aduros has joined the channel
[19:28] pyrotechnick: that's something we've completely written off with webgl which has kind of been nice really
[19:29] pyrotechnick: is basically not having to support anything but an A++++++++++ (we're talking almost a year from now) browser
[19:29] QuietKnight: I don't think I'm trying to do anything extra to support IE 6
[19:29] QuietKnight: if it works, great. If not, oh well. Gonna require at least IE7. But yea, I'd like that share.
[19:29] pyrotechnick: but i'm a web dev by trade so i feel your pain dude trust me
[19:29] pyrotechnick: it's a shit
[19:29] QuietKnight: haha, yea, webgl is a crazy goal :) which is why I liked your Steam idea so much
[19:30] pyrotechnick: that's where we want to be dude
[19:30] pyrotechnick: distribution
[19:30] pyrotechnick: but we think with that comes having a trademark engine too and some flagship games
[19:30] jan____ has joined the channel
[19:30] rnewson has joined the channel
[19:31] pyrotechnick: kinda hit the nail on the head though with the first game
[19:31] QuietKnight: I'm mostly doing this for the heck of it. Ever since I was 11 I wanted to write my own MMO. Now I finally can.
[19:31] pyrotechnick: ^w^
[19:31] cpleppert has joined the channel
[19:31] pyrotechnick: we decided on an MMO simcity
[19:31] QuietKnight: I think my chances of actually "making" it are slim-to-none
[19:31] pyrotechnick: dude
[19:32] pyrotechnick: no way i've seen your pathfinding, it's so good i couldnt even understand it, i doubted it
[19:32] pyrotechnick: anyway, at first i didn't really realise what that meant until i started coding it
[19:32] pyrotechnick: then i was like dudes...
[19:32] pyrotechnick: what is simcity
[19:32] pyrotechnick: and they're like it's a game
[19:32] pyrotechnick: and i'm like yeah...but what type of game
[19:32] pyrotechnick: and they're like a simulator
[19:32] pyrotechnick: and i'm like yeah but what does it simulate
[19:32] pyrotechnick: and they're like worlds
[19:33] pyrotechnick: and i'm like yeah but what do you do when you actually play simcity
[19:33] pyrotechnick: and they're like you build worlds
[19:33] pyrotechnick: and i'm like yeah and what is that
[19:33] pyrotechnick: and they're like AHHHHHH a game editor
[19:33] pyrotechnick: so
[19:34] pyrotechnick: between toolbars and camera movements and picking and spawning and rearranging and tweaking values and all that nonsense we've begun our game editor inadvertently 
[19:34] QuietKnight: haha
[19:34] QuietKnight: that, not too far of a leap there
[19:34] QuietKnight: *that=yea
[19:34] JimBastard: what is up with this recurring theme on the mailing list of russian programmers posting messed up benchmarks 
[19:34] pyrotechnick: i'm heaps keen to see what everyone is hacking together though dude
[19:34] QuietKnight: did you take a look or know about that "effect games" link I sent earlier
[19:35] pyrotechnick: i think i clicked it
[19:35] QuietKnight: might give you some ideas
[19:35] QuietKnight: its a games editor
[19:35] rnewson has left the channel
[19:35] QuietKnight: havent actually used it, but I've seen some of the finished games, and they're quite god
[19:36] pyrotechnick: fuck me dude is this webgl
[19:36] QuietKnight: nope
[19:36] QuietKnight: pure HTML/JS
[19:36] isaacs: JimBastard: inorite!?
[19:36] QuietKnight: I don't think they use canvas/svg, but they may
[19:36] JimBastard: lol
[19:37] JimBastard: i think its russian developers not wanting to believe anything is close to nginx, which is kinda funny caus node and nginx are very diffirent things
[19:37] pyrotechnick: lol yeah i was just typing that up JimBastard
[19:38] pyrotechnick: have you seen the nginx website
[19:38] JimBastard: yes?
[19:38] pyrotechnick: it's scarerely communist 
[19:39] pyrotechnick: it looks like it was designed by westwood
[19:39] QuietKnight: pyro- also, http://waroflegends.jagex.com/   recent game that is sorta MMO SimCity. I think its kinda popular, but dead simple graphics/gameplay. Again, might give you some ideas :)
[19:39] isaacs: mikeal: hey, is the js-registry sync working for you in the latest node?
[19:40] mikeal: i removed the sync script
[19:40] mikeal: you use couchapp now
[19:40] mikeal: npm install couchapp
[19:40] isaacs: yeah, that's what i mean
[19:40] isaacs: mikeal: i'm getting an odd crash in the event loop when http.js tries to emit("close")
[19:40] mikeal: let me make sure
[19:40] isaacs: pulling from your latest repo
[19:41] jan____: mikeal: isaacs: Hi!
[19:41] mikeal: works for me
[19:41] isaacs: weird!
[19:41] isaacs: hi, jan____ 
[19:42] isaacs: i'm getting this: http://gist.github.com/335638
[19:42] mikeal: so
[19:42] mikeal: either there is no couchdb there
[19:43] mikeal: or it's some kind of auth failure
[19:43] isaacs: hmm....
[19:43] isaacs: there is definitely couchdb there
[19:43] mikeal: http://couch.izs.me/jsregistry
[19:43] mikeal: 404
[19:43] isaacs: mikeal: it's a vm
[19:43] pyrotechnick: QuietKnight: what's your github so i can keep in contact
[19:43] isaacs: so, unless you can point your /etc/hosts at my laptop, you're not gonna see it
[19:43] cloudhead has joined the channel
[19:44] mikeal: haha
[19:44] mikeal: hrm......
[19:44] isaacs: but i've got a jsregistry db in it, shows up in futon just fine
[19:44] isaacs: and it's admin party
[19:44] mikeal: super weird
[19:44] mikeal: try removing the design document and then syncing
[19:45] isaacs: this doesn't spit out any errors: $ couchapp --sync --couch http://couch.izs.me/jsregistry
[19:45] mikeal: that's because there is no ddoc
[19:45] isaacs: k
[19:45] mikeal: it's not actually doing anything
[19:45] mikeal: i just added an error when you do that
[19:45] n8o has joined the channel
[19:46] isaacs: k
[19:46] mikeal: so
[19:46] isaacs: if you feel like it, wanna put some sys.debug()s in there just saying what's happening?
[19:46] mikeal: did you already setup the vhost?
[19:46] isaacs: yeah
[19:46] mikeal: ah ha!
[19:47] isaacs: the virtual host is not the same hostname, though
[19:47] isaacs: couch.izs.me is just plain couch
[19:47] mikeal: you're trying to sync to the vhost which is probably rewriting to /design/app/_rewrite
[19:47] isaacs: js-registry.izs.me is the vhost
[19:47] QuietKnight: pyro- check pm
[19:47] mikeal: oh, damn
[19:47] mikeal: there goes that theory
[19:47] mikeal: so when you still http://couch.izs.me/jsregistry in your browser you see the JSON for the database
[19:48] isaacs: yep
[19:48] isaacs: i see this: {"db_name":"jsregistry","doc_count":0,"doc_del_count":0,"update_seq":0,"purge_seq":0,"compact_running":false,"disk_size":79,"instance_start_time":"1268823985499100","disk_format_version":5}
[19:48] mikeal: and you're sure the sync fails?
[19:48] isaacs: yeah
[19:48] mikeal: because the error looks like it happens after
[19:48] mikeal: ok
[19:48] isaacs: well, couch doesn't have the data, it looks like
[19:48] isaacs: but yeah, the crash does seem to come after.
[19:48] isaacs: when it's closing the connection
[19:48] isaacs: (or something...)
[19:48] mikeal: what does couchdb say?
[19:48] mikeal: anything?
[19:49] jherdman has joined the channel
[19:49] mikeal: is there some kind of crash?
[19:50] isaacs: nope
[19:50] isaacs: couchdb is not reporting anything.
[19:51] isaacs: not so much as a GET
[19:51] isaacs: or even a HEAD
[19:51] isaacs: it's ok.
[19:52] isaacs: i'm gonna rewrite require() first, and then if it's still breaking this evening and you haven't seen it break yet, i'll take a look.
[19:52] isaacs: the WAY it's breaking is very interesting.
[19:52] jherdman has joined the channel
[19:52] mikeal: it looks like it's never connecting
[19:52] isaacs: yeah
[19:53] mies has joined the channel
[19:53] isaacs: oh.... wait...
[19:53] isaacs: so, http:393 is incoming.emit("end")
[19:53] isaacs: it's failing at the end of the uploaded data
[19:54] isaacs: ok, this is fascinating...
[19:54] mikeal: hrm......
[19:56] mikeal: i don't see how I'm suppose to have a handler for this error
[19:56] mikeal: when http can't connect
[19:56] isaacs: well, http is connecting
[19:56] isaacs: but it's failing at the end of the message for some reason
[19:56] mikeal: because from what I can tell, ClientRequest doesn't emit an error, and response is never emitted
[19:56] mikeal: right now i'm just trying to add proper handling when you do try to connect to a site that's not there
[19:57] dnolen has joined the channel
[19:57] isaacs: well, i'm pretty sure that's not it
[19:57] isaacs: because futon can do PUTs just fine, and curl PUTs work
[19:57] isaacs: there's something about couchapp's messages that isn't quite right.  and it's wokring for you, so it's some very interesting edge case.
[19:58] Tim_Smart has joined the channel
[19:58] brandon_beacher has joined the channel
[19:59] pyrotechnick: isaacs: you broke webchat with that massive json
[19:59] mikeal: a while back I had an issue with proper host headers, but that should be fixed
[19:59] isaacs: pyrotechnick: whoops, sorry!
[19:59] mikeal: oh no wait
[19:59] mikeal: that's not fixed in this code
[19:59] pyrotechnick: ^w^
[19:59] mikeal: dammit
[19:59] mikeal: oh no, i'm wrong, it's in here
[20:00] QuietKnight: pyro- hi. I had said check pm, but I think you crashed?
[20:00] pyrotechnick: isaacs: not your fault, they should be doing some overflow: hidden or something at least
[20:00] pyrotechnick: LEARN2CSS
[20:00] isaacs: pyrotechnick: true that.  looked fine in colloquy :)
[20:00] mikeal: yeah, i can't figure this out
[20:01] pyrotechnick: i kinda miss colloquy, my mate got me addicted to limechat
[20:01] mikeal: _ry: what do I listen to, and on which objects, to see if my HTTP client failed to connect?
[20:01] QuietKnight: mIRC ftw
[20:01] isaacs: mikeal: ping me if you figure it out.  i gotta run for a bit.
[20:01] mikeal: i will
[20:01] isaacs: thanks!
[20:02] mikeal: oh
[20:02] mikeal: wait, nevermind
[20:02] mikeal: that wouldn't be it
[20:05] mikeal: ah HA
[20:05] mikeal: "error" on client AND request
[20:11] nalply has joined the channel
[20:17] nalply: _ry: I did some work on net2 Buffer. I added a 'd' format to Unpack (for IEEE 754 double) and also wrote a pack() method. Interested?
[20:18] joshbuddy_ has joined the channel
[20:25] felixge has joined the channel
[20:25] felixge has joined the channel
[20:30] tlrobinson_ has joined the channel
[20:30] stephenlb has joined the channel
[20:33] okito has joined the channel
[20:46] tilgovi has joined the channel
[20:46] pmuellr has left the channel
[20:47] mattly has joined the channel
[20:55] nalply has left the channel
[21:02] CIA-77: node: 03Ryan Dahl 07net2 * r04c06b9 10/ (25 files in 6 dirs): child process now use net.Socket - http://bit.ly/cbMyyf
[21:03] _ry: http://github.com/ry/node/blob/d96c52694a56e10f2ba5db239680cb6a0af02120/test/fixtures/echo.js
[21:04] _ry: oops
[21:04] _ry: http://github.com/ry/node/blob/04c06b91493762dea34d91c34f2a33ebaa919eca/test/fixtures/echo.js
[21:04] JimBastard: echo echo
[21:04] _ry: ^-- tell me that isn't beautiful
[21:04] _ry: true encoding free pipes
[21:04] _ry: all on a single code base
[21:05] JimBastard: if you say so boss
[21:05] JimBastard: ima file this under, "thar be magic thar"
[21:06] _ry: http://github.com/ry/node/blob/04c06b91493762dea34d91c34f2a33ebaa919eca/test/simple/test-child-process-ipc.js
[21:07] _ry: in particular 
[21:07] _ry: http://github.com/ry/node/blob/04c06b91493762dea34d91c34f2a33ebaa919eca/test/simple/test-child-process-ipc.js#L20
[21:07] inimino: that is nice :)
[21:08] maritz: heelp, i can't motivate myself to start on the new node.js project i have planned. i just keep watching more and more the office and firefly... :(
[21:08] JimBastard: _ry:  in RE to your comment on the mailing list about node projects. tim_smart and i are close to releasing a pretty sweet open-source node app and web service
[21:08] inimino: _ry: I guess things are going well with net2 :)
[21:08] maritz: think that should be fixed before any other bug in node
[21:09] tmpvar: _ry, nice one
[21:09] JimBastard: one that ACTUALLY does stuff, not just theortical framework code
[21:10] Tim_Smart: _ry: Time for WebWorkers?
[21:10] Tim_Smart: :p
[21:10] _ry: very close now
[21:10] Tim_Smart: _ry: What is left to do?
[21:11] _ry: i need to get the http client running on net2 and file i/o using buffers
[21:11] Tim_Smart: Does Web Workers need that to work?
[21:12] _ry: well - net2 needs to be merged
[21:12] okito has joined the channel
[21:13] _ry: someone just posted http://img696.imageshack.us/img696/5114/resulti.png on twitter
[21:13] _ry: but i don't see how that's possible
[21:13] _ry: they clearly made some errors :)
[21:14] Tim_Smart: yeah, I would assume that too
[21:15] Tim_Smart: 2x faster than tornado? Must be dreaming...
[21:15] JimBastard: fffffuuuuu twisted
[21:15] _ry: well - especially twisted 5 times slower than torando
[21:16] _ry: there is no way
[21:16] _ry: maybe the horizonal axis is mislabled, maybe it's sec/requests 
[21:17] davidsklar: ha!
[21:17] atmos: haha
[21:17] atmos: i love how thin doesn't even have a bar
[21:20] atmos: irb
[21:20] atmos: lulz
[21:20] maritz: it has 1px! :D
[21:22] maritz: okay, but that 1px is only #7c7c7c while the normal border of the bars is #000... so it has 0 < N < 1 px.
[21:27] tmpvar: wow, thats really impressive
[21:28] tmpvar: wonder if there is an article associated with it heh
[21:28] mattly: nice
[21:29] softdrink: *perk*
[21:30] softdrink: found it: http://github.com/lucasts/Async_Test
[21:31] mattly: it'd be interesting to see thin on ruby 1.9 vs. 1.8, it looks like they ran it on 1.8
[21:32] ashb: softdrink: what kind of benchmakrs are they?
[21:32] ashb: simple hello world ones? BAH!
[21:32] ashb: hellow world is fucking useless as a benchmark
[21:32] softdrink: ashb: damn right.
[21:33] mattly: benchmarking: ur doin it wrong
[21:34] softdrink: we should write a small suite of benchmarks, and do something like the programming language shootout
[21:34] _ry: softdrink: yes
[21:34] ashb: softdrink: i linked to a post by the wsgi guy earlier
[21:34] ashb: he has good ideas
[21:34] softdrink: except for the whole no-free-time thing...
[21:34] softdrink: 
ACTION scrolls

[21:35] _ry: arlo is working on integrating benchmarks into the buildbot
[21:35] mattly: http://jan.prima.de/~jan/plok/archives/175-Benchmarks-You-are-Doing-it-Wrong.html#variables
[21:36] softdrink: i'm either blind or it scrolled off my buffer ashb
[21:36] softdrink: or both.
[21:36] ashb: softdrink: check my delicious: ashberlin
[21:36] softdrink: k
[21:37] ashb: on mobile connection right now so fiddly for me to do it
[21:37] softdrink: thankie. *reads*
[21:38] softdrink: that reminds me:  is anyone fronting node with nginx? why/why not?
[21:40] JimBastard: im using nginx as our web server
[21:40] mikeal: node is already great at handling concurrent connections, so i don't put anything in front of it
[21:40] JimBastard: nginx serves all static content, then we proxy_pass all dynamic requests
[21:40] JimBastard: i dont see any reason to serve static content from node
[21:40] mikeal: non-blocking disc IO vs blocking :)
[21:40] softdrink: *nod*
[21:41] mikeal: nginx uses blocking IO for serving files off the disc
[21:41] softdrink: 
ACTION moves everything into a ramdisk

[21:41] JimBastard: but if its all static files how many times does it have to actually load the file from HD ?
[21:41] blackdog has joined the channel
[21:42] JimBastard: with node we had no caching solution so we were re-reading our static files via the file module EVERYTIME
[21:42] mikeal: you could do the same in node without much work
[21:42] blackdog: is it just me or has github been sucking all day?
[21:42] mikeal: all week
[21:42] mikeal: fork queue is BROKEN
[21:42] JimBastard: feel free to go into #github and bitch at tekkub
[21:42] JimBastard: PLEASE say that i sent you
[21:42] blackdog: :)
[21:46] softdrink: JimBastard: what sort of site are you using node on?
[21:50] JimBastard: its a free web hook platform
[21:50] JimBastard: hook.io
[21:50] JimBastard: we havent launched yet
[21:50] JimBastard: v0.1 in a couple of weeks i hope
[21:50] softdrink: huh. interesting
[21:51] JimBastard: it will be a free service and an open-source software project
[21:51] JimBastard: with the same name for both
[21:52] kjeldahl has joined the channel
[21:52] JimBastard: im actually hosting a hackathon tommorow where creationix and myself (and a few others) will be working on node stuff
[21:52] JimBastard: and a lot of the node stuff revolves around getting that project to work
[21:52] JimBastard: node-persistence, node-oauth, node-twitter
[21:52] JimBastard: we've been making a bunch of contributions
[21:53] softdrink: i wish we were using node at work instead of a bunch of little ruby things
[21:55] JimBastard: we use ruby at work
[21:55] JimBastard: this is my own project
[21:56] hassox has joined the channel
[21:56] derferman has joined the channel
[21:57] softdrink: i've pretty much decided on node for a personal project since i can more easily express myself in js than anything else these days
[21:57] isaacs has joined the channel
[21:57] dandean has joined the channel
[22:07] alexiskander has joined the channel
[22:09] devinus has joined the channel
[22:09] devinus: will node WebWorker's actually be threads?
[22:10] tilgovi has joined the channel
[22:10] _ry: devinus: no
[22:11] _ry: JimBastard: have you seen http://github.com/orlandov/node-sqlite
[22:11] _ry: JimBastard: still under development - but it's the shit
[22:11] JimBastard: _ry:  i've seen most of the db wrappers for node
[22:12] _ry: in-process async 
[22:12] JimBastard: the goal is to have a common interface for dealing with multiple database vendors
[22:12] JimBastard: so ideally i would take that driver and port it into node-persist
[22:12] JimBastard: take his code, put the common wrapper around it
[22:12] devinus: _ry: aren't they threads in the browser?
[22:13] JimBastard: im working with creationix on it this week (hes visiting from texas) we having a node hackathon party on thursday night
[22:13] _ry: devinus: i don't know. depends on the implementation
[22:13] JimBastard: a few noders are coming
[22:13] mjr_: It'd be great if we had the same API for sqlite than HTML5 has, even though it is ugly and complicated.
[22:13] _ry: devinus: in chrome it's processes
[22:13] _ry: devinus: like god intended
[22:13] devinus: _ry: in node they will be proesses?
[22:14] _ry: yes
[22:14] devinus: _ry: oh thank god, thats all i wanted to know
[22:14] devinus: _ry: want to utilize some cores :)
[22:14] Tim_Smart has joined the channel
[22:16] _ry: fork and conquer
[22:16] QuietKnight has left the channel
[22:17] _ry: 
ACTION away

[22:19] mikeal: anyone know if do supports parallel chains?
[22:20] Tim_Smart: mikeal: E.g. ?
[22:21] mikeal: like
[22:21] Tim_Smart: I was going to make a prototype for doing parallel actions
[22:21] mikeal: do.parallel(Do.chain(one, two),Do.chain(three, four))(function (results) {})
[22:22] mikeal: run the two chain pairs in parallel
[22:28] tmpvar has joined the channel
[22:28] JimBastard: i think Do can do parrells
[22:28] JimBastard: also kry zips library?
[22:30] tmpvar has joined the channel
[22:36] JimBastard: http://github.com/yourmother
[22:36] JimBastard: damn you binary42 
[22:36] binary42: Heheh.
[22:36] JimBastard: i just forked yourmother
[22:36] JimBastard: i sent her a pull request.....and she accepted
[22:36] binary42: JimBastard: You asked for accounts. At least she follows you.
[22:36] KungFuHamster has joined the channel
[22:36] JimBastard: we'll be merging later tonight
[22:38] JimBastard: i think i might know a trick for getting thousands and thousands of watchers on a project
[22:38] JimBastard: i might try it when i get home for fun
[22:38] atmos: MAKE A GAME OMG
[22:38] binary42: Create thousands of accounts?
[22:38] JimBastard: naaah ill pm you
[22:38] binary42: Amazon's mechanical turk.
[22:39] pedrobelo has joined the channel
[22:45] okito has joined the channel
[22:49] gf3 has joined the channel
[22:53] mattly has joined the channel
[22:56] CIA-77: node: 03pyrotechnick 07master * r8aaffe7 10/ doc/api.txt : Fixing a typo in the fs.readFile example. - http://bit.ly/doCgJq
[22:56] CIA-77: node: 03Ryan Dahl 07master * r2d7e86e 10/ (82 files in 12 dirs): Upgrade V8 to 2.1.5 - http://bit.ly/9b0cB2
[22:58] Tim_Smart: Whats new in V8 2.1.5?
[22:58] JimBastard: they took out some of the tomatoe juice and replaced it with carrot juice
[22:58] Tim_Smart: Performance improvements ^^
[22:59] _ry:  82 files changed, 3732 insertions(+), 104751 deletions(-)
[22:59] _ry: yes
[22:59] _ry: :)
[22:59] Tim_Smart: That is a lot of deletion :
[22:59] Tim_Smart: :p
[23:00] _ry: (they just deleted a data file from the tests)
[23:00] Tim_Smart: oh
[23:00] JimBastard: thats like my coworker bragging about his influence on our project (i was like installing a 10k library and running the installer script doesnt count)
[23:01] JimBastard: impact rather, not influence
[23:02] mattly: so is there like, some "how to do async properly 101" writeup somewhere?
[23:02] mattly: i figure as long as i'm dinking around with this stuff i may as well do it right
[23:04] JimBastard: have you read www.howtonode.org ?
[23:04] JimBastard: there are a few articles that might help in regards to event flow
[23:04] mattly: been skimming it
[23:04] mattly: ah ok
[23:04] JimBastard: thats targetted for node.js obviously though
[23:04] mattly: right
[23:06] Tim_Smart: mattly: Have you done much browser Javascript?
[23:07] mattly: Tim_Smart: yes
[23:07] Tim_Smart: Well, just think event listeners
[23:07] mattly: mostly with prototype.js but getting more into event-driven with jquery recently
[23:07] mattly: ok
[23:08] orlandov has joined the channel
[23:08] Tim_Smart: The whole "when this finishes, call this function" deal
[23:08] orlandov: .
[23:18] mrjjwright has joined the channel
[23:20] voodootikigod has joined the channel
[23:24] brianm has joined the channel
[23:29] rnewson has joined the channel
[23:31] r11t has joined the channel
[23:31] orlandov: omg... node has isArray()
[23:31] orlandov: <3
[23:32] derferman has joined the channel
[23:33] rnewson: is it possible to start a new request from a response callback? it hangs for me, so I'm assuming I'm blocking the event loop by doing something stoopid.
[23:33] rnewson: use case is I need to do a HEAD call before a PUT.
[23:37] devinus has joined the channel
[23:39] creationix has joined the channel
[23:39] Tim_Smart: rnewson: Is this for HTTP Client?
[23:40] rnewson: yes.
[23:40] Tim_Smart: rnewson: Close the connection first?
[23:40] rnewson: could I have been that stupid? let's find out...
[23:41] Tim_Smart: I'm not sure >.< You have to listen for the 'data' and 'end' event, so maybe chuck the request in the end event
[23:41] mjr_: it is possible to start a new HTTP request from a response callback.  That's how you can use node as a proxy.
[23:41] rnewson: hm, it's not that.
[23:42] rnewson: I need to proxy the original call still, so I can't close that one. hmm.
[23:42] rnewson: it's for couchdb. I need to do a HEAD to get the ETag so I can pass it in an If-Match on the PUT.
[23:42] rnewson: but the PUT might have a request body.
[23:43] Tim_Smart: Have 2 http clients?
[23:45] rnewson: that sounds right to me, will try. thanks.
[23:45] mjr_: Here's an example of a proxy to couchdb:
[23:45] mjr_: http://gist.github.com/335873
[23:47] rnewson: mjr_: thanks, had something pretty much like that.
[23:48] nicketynick has joined the channel
[23:48] nicketynick: heh, there are a bunch of people idling in #nodejs, which I just accidentally joined
[23:50] Tim_Smart: lawl
[23:53] nicketynick: is nodejs threadsafe? i'm writing an extensions module. should I make an effort to make sure its reentrant?
[23:53] orlandov: are you planning on using threads for something?
[23:53] nicketynick: no i'm not, but I don't want to break other people's assumptions
[23:54] rnewson: gotta run, thanks for suggestion.s
[23:54] orlandov: in general, you shouldn't have to worry about thread safety
[23:55] felixge has joined the channel
[23:55] felixge has joined the channel
[23:55] _ry: nicketynick: thread safe? it's thread-free
[23:55] nicketynick: i'm looking through, the nodejs source and I think I'm seeing a bunch of global state floating around. so I guess I don't have to worry
[23:55] orlandov: threads are used very sparingly in node, only as a last resort
[23:55] orlandov: to wrap sync stuff
[23:56] _ry: nicketynick: you can make the assumption that all node functions are synchronous
[23:56] _ry: there are no threads
[23:56] orlandov: asynchronous? :)
[23:57] nicketynick: _ry: got it. thank you
[23:57] _ry: synchronous in state
[23:57] orlandov: ah yeah
[23:57] orlandov: iow: only one thing happening at a time?
[23:58] _ry: yes