--- Log opened Mon Nov 26 00:00:14 2012
00:02 < warz> is it okay to make use of callbacks in such a way to where you dont need to wait for the results? sort of like dispatching processing tasks but immedietly responding with a http resosponse, for example.
00:02 < warz> where the processing is known to take longer, but as long as it finishes i dont care if its done by the end of the request
00:03 < warz> or are they not guaranteed to complete if you do not wait on them?
00:03 < ljharb> warz: you shouldn't need a callback if you don't care when it finishes, right? just fire it off and forget about it
00:03 < Zarel> So, socket.remoteAddress appears to be copying the IP in the X-Forwarded-For header.
00:03 < Zarel> How do I get the real IP address of a socket?
00:03 < ljharb> Zarel: "real"? if it was forwarded by a proxy, than the proxy IP is the real one
00:03 < Zarel> ljharb, yes. I'm trying to get the proxy IP.
00:05 < ljharb> oh i see, it's forwarding the internal ip
00:05 < chilts> warz: whether you listen to the callback in that case or not is up to you (e.g. you might want to log OK/Failure cases just so you know what happened)
00:05 < ljharb> to the client
00:05 < chilts> but yes, if it's longer than ther equest that's ok :)
00:05 < warz> ljharb, thats basically what im doing, im just not handling the callback basicallyu.
00:05 < Zarel> ljharb, to be exact, I'm experiencing issues with users forging the X-Forwarded-For header.
00:05 < warz> didnt know if that was "bad"
00:05 < Zarel> And then the forged IP becomes socket.remoteAddress and I don't know how to get their actual IP.
00:05 < warz> chilts, ah ok
00:06 < ljharb> Zarel: forging? why would that matter - IP addresses are never something you should use for security.
00:06 < ljharb> Zarel: just by being on the same university campus, or in the same office, i can get the same IP as someone else.
00:06 < Zarel> ljharb, I understand, but it's really the best I can do.
00:06 < ljharb> Zarel: that's no security at all. and it's not the best - you can easily add a session key or CSRF token or something.
00:07 < Zarel> I'm trying to ban a set of known untrustworthy IPs. -_-
00:07 < ljharb> you should be able to just do that with iptables
00:07 < Zarel> Session keys and CSRF tokens don't solve that problem at all.
00:07 < chilts> warz: just out of interest, what are you doing which is fire-and-forget?
00:07 < chilts> sometimes analytics is like that and some other things
00:07  * chilts is curious
00:07 < Zarel> Okay, fine, I'd like to restrict a set of known untrustworthy IPs from doing certain things.
00:08 < _numbers> lol i love the vanilla js framework http://vanilla-js.com/
00:08 < ljharb> Zarel: but "socket.remoteAddress" is on the client. you'd need to restrict it on the server, right?
00:09 < Zarel> ljharb, I don't understand. socket is the socket of the client connecting to the server. I'm trying to get the client's IP, but socket.remoteAddress isn't the client's real IP.
00:09 < Zarel> Say I'm writing a forum. A user has been banned and is no longer allowed to post. How do you propose that be done, if not by the user's IP?
00:10 < chilts> _numbers: that's brilliant :)
00:11 < ljharb> you ban by account, since anonymous posting is asking for trouble.
00:11 < Zarel> Session keys and CSRF and all that don't work if it's an open-to-the-public forum. And yes, I know there are proxies and all that to get around IP bans and IP bans can have collateral damage, but it's really as close as it gets.
00:11 < warz> chilts, nothing that interesting. im just hacking together a web front-end for this document converter module i wrote.
00:11 < Zarel> The user can create a new account; that's what IP banning prevents.
00:11 < warz> and in an effort to just "get it working", im firing and forgetting the actual convert and upload to S3 part, for now.
00:11 < ljharb> lol it's really easy to get a new ip. all you need is somewhere to ssh. but ok
00:11 < chilts> warz: have you seen connect-stream-s3?
00:11 < warz> in which i intend to move out to something like a message queue and worker, later.
00:12 < chilts> I wrote it, it uses AwsSum :D
00:12 < Zarel> Also; allowing anonymous posting and banning spammers by IP has worked for me well enough until someone found a way to spoof an IP.
00:12 < warz> chilts, no is it l33t?
00:12 < chilts> heh
00:12 < warz> ill look at it. im just using knox, currently.
00:12 < chilts> it's just connect middleware which uploads files from your forms
00:12 < chilts> this might not be your use-case :)
00:12 < chilts> if so, ignore me :D
00:12 < Zarel> Anyway, either way, I presume there's a way to get a socket's true remote address in Node.js, and if not, I'd imagine that would be a bug.
00:12 < warz> hadn't heard of awsum, either
00:13 < ljharb> Zarel: yep, of that i'm sure as well - i just don't know what it is, and googling didn't turn up anything
00:13 < mmalecki> Zarel: what do you mean by 'true remote address'?
00:13 < warz> oh i see, its a route middleware. ok i might come back at look at that.
00:13 < Zarel> mmalecki, I mean, the remote IP address of the socket.
00:14 < mmalecki> this is exactly what you get from socket.remoteAddress
00:14 < chilts> Zarel: have you tried req.connection.remoteAddress?
00:14 < Zarel> It seems like in the current version of Node, it's actually
00:14 < Zarel> socket.remoteAddress = socket.headers['x-forwarded-for'] || socket.actualRemoteAddress;
00:14 < chilts> warz: http://awssum.io/ :)
00:14 < chilts> does all the AWS services
00:14 < Zarel> This seems to be a recent change, too, since googling says that socket.remoteAddress is the IP address of the proxy, when it clearly isn't
00:15 < Zarel> chilts, where do I get 'req'?
00:15 < ljharb> what version of node?
00:16 < Zarel> 0.8.3
00:16 < ljharb> hm, ok. i guess you could upgrade to 0.8.9 but that shouldn't matter.
00:22 < Zarel> chilts, req.connection.remoteAddress worked
00:22 < Zarel> Thank you so much!
00:25 < chilts> heh, no worries
00:25 < EvilJordan> does anyone have any similar experiences where io.of('/namespace').emit('blah') never reaches the client?
00:28 < Disil> How does require work for modules?
00:29 < Disil> like a module has require('steamkit'), do I have to do anything so it can find it?
00:32 < chilts> Disil: npm install, or npm install steamkit
00:35 < Disil> How does install work? Does it install it to the current folder?
00:37 < Disil> I guess so
00:39 < dominictarr> gozala, hey! whats up?
00:41 < Disil> yay it mostly works
00:41 < mscdex> Disil: it installs under node_modules
00:41 < mscdex> for direct dependencies
00:42 < Disil> well I had to do npm install steam/irc inside of the folder of the repo
00:43 < mscdex> you can also just do `npm install` to install the dependencies in ./package.json
00:46 < Disil> well this is pretty simple, but I suck with JS. How can I modify steam.logOn(details.username, details.password) to pass details.authCode if it exists?
00:47 < SomeoneWeird> middleware
00:48 < ljharb> Disil: just pass details.authCode - if it doesn't exist, it will be `undefined`
00:48 < ljharb> or middleware
00:48 < Disil> oh, okay
00:50 < Disil> I hate Steam Guard =I
00:50 < ljharb> what are you trying to do with steam anyways?
00:51 < Disil> it's a bot to relay group chat messages to an IRC channel
00:51 < Disil> https://github.com/seishun/irc-steam-relay
00:51 < ljharb> ah
00:51 < Disil> ugh, now my webhost is being slow as fuck
00:52 < Disil> I need my emails for my auth code ;-;
00:53 < MI6> joyent/node: Ben Noordhuis v0.8 * 121ed91 : tls: fix tls.connect() resource leak  The 'secureConnect' event listener  - http://git.io/Cq6E8w
00:56 < Disil> holy shit it works perfectly
01:11 < MI6> joyent/libuv: Ben Noordhuis master * 99d3102 : unix: drop pthread dependency in threadpool.c  Replace the pthread code w - http://git.io/l6iiFw
01:13 < sinclair|work> has anyone here used the http-proxy module?
01:18 < Disil> Hmm, what would be an easy way to do user permissions? I don't know Node that well :c
01:21 < mscdex> Disil: read the permissions from the database when you load the user's session?
01:21 < sinclair|work> am looking to proxy a request out to a external host using node-http-proxy
01:22 < Disil> Well I've never done anything like it in JS.
01:22 < sinclair|work> however, the host headers don't seem to be proxied through, for example, the host is domain.com running on port 80
01:22 < sinclair|work> the proxy is localhost on port 8000 (as per their examples)
01:23 < mscdex> Disil: if you're using express, there's session middleware for most popular databases
01:33 < TehShrike> isaacs: you had mentioned that you thought "new" was the most efficient way to create objects, but it looks like that might not be the case http://jsperf.com/cost-of-a-function-call/3
01:39 < sinclair|work> TehShrike: Firefox does better with object literals 
01:40 < TehShrike> It depends on the version, apparently
01:40 < TehShrike> Which was also really interesting to me
01:40 < TehShrike> Check out the difference between the different versions of Chrome, or Firefox
01:41 < TehShrike> That surprised me
01:43 < TehShrike> I'm not sure how I feel about the huge performance dropoff after Chrome 22
01:43 < TehShrike> But it's probably not good
01:45 < sinclair|work> TehShrike: ever dealth with the node-http-proxy?
01:45 < sinclair|work> *dealt
01:45 < TehShrike> No, I haven't used a proxy yet, though I have my-little-proxy on my list of proxies to check out once I need one
01:46 < TehShrike> https://github.com/jesusabdullah/my-little-proxy
01:47 < sinclair|work> TehShrike: im having an issue with the host headers
01:47 < sinclair|work> TehShrike: for example, i hit http://localhost:8000, i want to forward that to http://domain.com (port 80)
01:48 < TehShrike> Well, like I said, I haven't used it yet, but that seems like a thing that mlp does very easily
01:48 < TehShrike> FRIENDSHIP IS JAVASCRIPT!
01:49 < sinclair|work> TehShrike: i like the little graphics they have, but i think i'll stick with this node-http-proxy
01:49 < TehShrike> :-P
01:50 < sinclair|work> TehShrike: did you say the otherday you wanted to look at TS ?
01:50 < DukeofURL> hey, does anyone have any node book recommendations? i'm looking for more in depth information
01:50 < sinclair|work> TehShrike: if you did, im happy to spend a few minutes getting you started :)
01:51 < TehShrike> sinclair|work: Yeah, I'm keeping an eye out for a good project to try it on
01:51 < sinclair|work> TehShrike: want a quick overview
01:51 < TehShrike> I checked out the docs the day it came out, but just haven't used it yet
01:51 < TehShrike> sinclair|work: do you have one written up? :-o
01:51 < TehShrike> DukeofURL: info about what?
01:51 < DukeofURL> node
01:51 < sinclair|work> TehShrike: was going to tell you about it over IRC
01:51 < TehShrike> DukeofURL: it's changed a good bit over the last year, general books would be tough
01:51 < TehShrike> sinclair|work: I'll learn whatever you feel like saying :-)
01:51 < DukeofURL> hmm
01:52 < TehShrike> DukeofURL: there are lots of good JavaScript books to recommend
01:52 < DukeofURL> well its mostly cuz im still not that solid on node's architecture
01:52 < TehShrike> DukeofURL: node.js is JavaScript, plus the libraries that come with node
01:52 < DukeofURL> or hell even good js software design 
01:52 < TehShrike> aah, now that is a thing that has books written about it
01:53 < TehShrike> Opinionated jerk though Crockford may be, you're probably still best off checking out his JavaScript: The Good Parts
01:53 < TehShrike> Personally, I also learned good stuff from JavaScript Patterns
01:53 < DukeofURL> im relatively new to javascript and node, i never did web js, coming from c/c++ and java im still trying to wrap my head around javascripty stuff
01:54 < DukeofURL> javascript patterns sounds like something for me
01:54 < TehShrike> You'll definitely want JS: The Good Parts too
01:54 < DukeofURL> ok
01:55 < TehShrike> Apply grains of salt liberally while taking it
01:55 < DukeofURL> thanx
01:55 < TehShrike> But still, it's pretty essential reading
01:55 < TehShrike> Also very good: http://eloquentjavascript.net/
01:55 < DukeofURL> ok cool 
01:55 <@mbalho> www.jsforcats.com
01:57 < DukeofURL> aww thats cute 
01:58 < TehShrike> aah, good link
01:58 < TehShrike> That's a recent one. <3 maxogden
02:01 < DukeofURL> anyway, thanks for the links y'all
02:02 < rump> can someone help a brother out with a fundamental...  how can i get result out of zlib.gzip()?  i thought this was the whole closure thing and this should work?  http://paste.ubuntu.com/1386896/
02:02 < pomke> mbalho: dawwww jsforcats <3
02:03 < pomke> I'm so sending that to a friend, I was teaching her some JS a little while ago
02:06 < bnoordhuis> rump: what exactly is the problem? it's compressing right?
02:06 < rump> bnoordhuis: yes, i just cant return it
02:06 < rump> how do i get it out?
02:06 < rump> its a closure/js question not a zlib question
02:06 < bnoordhuis> rump: your compress function should take a callback
02:06 < rump> yes im using it there
02:07 < rump> but i want to wrap around it or whatever
02:07 < bnoordhuis> rump: read carefully. i mean your var compress = function(buf) { ... } should take a cb
02:07 < rump> oh
02:07 < bnoordhuis> rump: zlib.gzip() is asynchronous, hence your compress() function should be too
02:08 < rump> hm
02:08 < rump> where to, heh
02:09 < rump> mind boggled
02:09 < rump> cb to cb to cb but i never get a value back hehe
02:10 < rump> is this that trick ive seen where you just return a function
02:10 < rump> so fuzzy in here
02:11 < bnoordhuis> rump: no
02:12 < rump> more context: http://pastebin.ubuntu.com/1386911/
02:12 < jwulf> rump: https://gist.github.com/4146219
02:12 < jwulf> pass a callback to compress
02:12 < jwulf> inside the callback from gzip, pass the result to the callback
02:13 < rump> ok then where do i plug in cb for like the example i just pasted
02:13 < rump> how does it eventually ever get back
02:14 < jwulf> rump, you need to do one of two things here
02:14 < jwulf> make it more async
02:14 < jwulf> or use an async library to make things synchronous
02:15 < rump> im using async in this file already
02:15 < jwulf> async and step are two good libraries for that
02:15 < rump> but if id love to hear a bird eyes view of both and what it takes
02:15 < rump> to learn
02:15 < rump> if you can spare it
02:15 < jwulf> so you can use async.series to force it to be synchronous
02:16 < rump> ah so the only problem is console.log(after) is literally being executed before gzip returns?
02:16 < jwulf> or waterfall
02:16 < jwulf> rump, yes
02:16 < rump> ok thats so much clearer
02:17 < rump> i thought was scoping craziness
02:17 < jwulf> by putting the console.log inside the zlib callback you'll see the expected result
02:17 < jwulf> but your outer scope returns before the inner scope does
02:17 < rump> no easy way to return out of the parent function from in the callback?
02:17 < jwulf> so the only way to get the result back is to pass in a callback from the parent
02:17 < jwulf> then call the callback inside the zlib callback
02:18 < jwulf> that's the usual way
02:18 < rump> so where does that passed callback go
02:18 < jwulf> so compress takes a callback
02:18 < rump> to get it back in my object literal
02:18 < jwulf> it's available inside your zlib callback because of closure
02:18 < jwulf> that takes us back to what i said earler
02:19 < jwulf> your two options are to make your other code more async
02:19 < jwulf> or to use the async library to handle that by making it "synchronous"
02:19 < jwulf> actually it's still async, but the async library handles that for you, and you program like it was synchronous
02:19 < jwulf> so put the compress in series with the assignment to the object literal
02:20 < jwulf> or in a waterfall
02:20 < jwulf> async.series
02:20 < jwulf> async.waterfall
02:20 < bnoordhuis> i suggest you first get comfortable with synchronous vs. asynchronous code before you start using a lib like async
02:21 < rump> yeah im missing the bigger picture
02:21 < rump> tricky tricky
02:21 < jwulf> rump, this will become quite an async spaghetti fest without async
02:22 < jwulf> without async what you need to do here is break your different steps into separate functions
02:22 < rump> so maybe i compress it and then emit an event saying compressed and then i can set it to my object
02:22 < jwulf> and instead on relying on program flow, as you are doing
02:22 < jwulf> yeah, like that
02:22 < jwulf> but you don't emit an event
02:22 < jwulf> you call a callback
02:22 < jwulf> it's callbacks all the way down
02:22 < rump> ohh
02:22 < rump> ok the blur is fading
02:23 < jwulf> so you pass in step n+1 to step n as a callback
02:23 < jwulf> when step n completes, it calls the callback
02:23 < rump> how come events are so underutilized in examples n such.. they seem like they would simplify a lot
02:23 < jwulf> which passes execution to step n+1
02:23 < rump> i guess callbacks are the same just for some reason it seems more hectic
02:24 < ljharb> rump: as i told you earlier, events, promises, and async can each be used to help.
02:24 < jwulf> rump, it's functionally equivalent, but not as declarative
02:24 < jwulf> using callbacks is more imperative programming
02:24 < jwulf> events is more declarative
02:24 < jwulf> it makes it more clear what's happening
02:25 < jwulf> gzip.onComplete = nextStep;
02:25 < jwulf> or gzip (myData, nextStep)
02:25 < rump> so instead of here set this property to result of compress... i have to say, go compress and when you're done set this property
02:25 < jwulf> ^ exactly
02:25 < jwulf> so you break everything up into small functions that do one thing
02:25 < jwulf> and chain them together, either using a library like async
02:25 < jwulf> or by manually injecting callbacks
02:25 < jwulf> async makes it more declarative
02:26 < rump> js is pretty verbose eh
02:26 < jwulf> it can be
02:26 < jwulf> but with async it's not so much
02:26 < jwulf> async the library i mean
02:26 < ljharb> if you do it right, your overall architecture will be less verbose :-)
02:27 < timeturner> what's the easiest to use documentation generator
02:27 < timeturner> from source comments
02:27 < timeturner> in jsdoc format
02:27 < rump> timeturner: dox and markdox seems saweet
02:27 < timeturner> rump: thanks, will check those out
02:27 < ljharb> timeturner: self-documenting code
02:28 < rump> timeturner: or if you prefer annotated style check docco or variants
02:28 < rump> i gotta throw all this code away and paradigm shift from scratch haha
02:28 < rump> what a mess
02:30 < rump> i was using async.sequence without callbacks or parameters
02:30 < rump> they were throwing me
02:30 < niggler> in soviet russia, errors throw you?
02:30 < rump> err, results, callbacks, confusing how it applys and not a lot of documentation
02:33 < ljharb> substack: you here? i'm ready to be converted to "everything is streams" for this little script i'm writing and i think you're the acolyte to teach me
02:35 < sshaginyan> Can someone explain to me async loops. For example I have a loop for(var i = 0; i < 10; i++) if this loop were async does that mean that all iterations would happen in parallel? i at 1 does not depend on i at 0?
02:36 < sshaginyan> If this is so can someone tell me how to read from a file async line by line.
02:36 < ljharb> sshaginyan: no, it means the execution happens in an unreliable order. JS only has one thread.
02:36 < sshaginyan> If I don't care about the order does that mean it's faster?
02:36 < ljharb> sshaginyan: and you wouldn't have an async loop with a numbered index, because you could get any line back at any time
02:37 < sshaginyan> yes thats what I want.
02:37 < sshaginyan> But can this happen on file IO on lines?
02:37 < sshaginyan> I don't care which line 
02:38 < ljharb> i have a feeling that it will probably give you every line in order, but the point is you wouldn't have to care about that
02:39 < ljharb> probably you'd create a read stream, and add an event listener that fires when each chunk (line) is received?
02:40 < RLa> how big the file is
02:40 < sshaginyan> ljharb: Can I show you some code I wrote?
02:40 < sshaginyan> really big
02:41 < sshaginyan> please critique my code http://pastie.org/5434194
02:41 < RLa> oh, payroll processing, isn't cobol used for that?
02:42 < RLa> anyway, i think i have seen a module to read stream line-by-line
02:42 < ljharb> sshaginyan: first critique: always use curly braces around conditionals :rageface:
02:42 < RLa> that would free you from doing \n search and all that
02:43 < ljharb> http://www.quora.com/What-is-the-best-way-to-read-a-file-line-by-line-in-node-js
02:44 < RLa> sshaginyan, it was the carrier module: https://github.com/pgte/carrier
02:46 < sshaginyan> RLa: Do you think it would be more impressive for an interviewer if I didn't use any libraries?
02:47 < RLa> it's impressive to reinvent wheel?
02:47 < sshaginyan> no
02:47 < RLa> well, that what it feels like
02:47 < sshaginyan> k
02:47 < ljharb> RLa: reinventing wheels is not impressive
02:47 < RLa> for real world apps you will probably want to use some external modules anyway
02:47 < ljharb> sshaginyan: it's good to prove you could reinvent a wheel. but otherwise it's silly to do so.
02:47 < niggler> woah woah woah sshaginyan this is a pure recruiter-free job-free space
02:48 < RLa> is it?
02:48 < sshaginyan> sorry
02:49 < niggler> well one at a time
02:49 < sinclair|work> ljharb: isn't node a reinvention of the wheel?
02:49 < ljharb> serverside javascript that doesn't suck? nope, nobody's done that before.
02:49 < RLa> you could also write multiple versions, one with the library, and other without it
02:50 < sshaginyan> RLa: That's what I'm going to do
02:50 < niggler> sinclair|work most things are reinvention pf the wheel
02:50 < sinclair|work> ljharb: i do find myself coding a lot of things that i would have leveraged by more established frameworks 
02:51 < ljharb> sinclair|work: within the problem space, javascript on the server, those frameworks don't exist.
02:51 < RLa> i think that sometimes it's not easy to find library to do what you want
02:51 < sinclair|work> ljharb: all for the sake of JS on the server?
02:51 < RLa> but it might exists there, somewhere
02:52 < niggler> sinclair|work that's the wrong way of looking at it
02:52 < sinclair|work> ljharb: functionality nodejs isn't doing anything new tho
02:52 < ljharb> of course it's not, that's not what i meant, and you know it.
02:52 < sinclair|work> niggler: i like node btw 
02:52 < niggler> how do you get reliable development from unreliable developers
02:52 < ljharb> i think i should clarify: when one is trying to achieve goal Foo, the only code you should write is code that solves that problem. everything ancillary to that, if solved elsewhere, should reuse esxisting solutions.
02:53 < sinclair|work> ljharb: so, i should plug .net into nodejs?
02:53 < sinclair|work> that's a lot of bang for buck
02:53 < ljharb> sinclair|work: that's not a solution.
02:53 < niggler> javascript as a CLR target
02:53 < ljharb> solutions that use .net, require .net to be valid. we're using node, so nothing in .net is a valid solution.
02:53 < sinclair|work> niggler: nah, CLR invocation with native node modules (in windows of course)
02:54 < sinclair|work> var system-io = requireclr('System.IO.dll");
02:54 < niggler> ive actually been thinking about Basic or VB in javascript
02:55 < RLa> why is clr so cool?
02:55 < sinclair|work> system_io.FileStream('yay.txt');
02:55 < ljharb> niggler: i'm sure somebody
02:55 < ljharb> 's 
02:55 < ljharb> done that.
02:55 < RLa> c# in js would make some sense tho
02:55 < ljharb> if there's php.js, i'm sure there's <insert language>.js
02:55 < ljharb> RLa: how would it make any?
02:56 < sinclair|work> RLa: im used to binding JS to .net, not the other way around
02:56 < jwulf> RLa: it's called typescript
02:56 < jwulf> typescriptlang.org
02:56 < niggler> phpjs is a different effort ljharb
02:56 < RLa> hm, indeed
02:56 < _numbers> ah node-grunt is the equivalent of ruby guard gem
02:56 < ljharb> niggler: no i'm not talking about phpjs (which is horrible code), i'm talking about php dot js
02:56 < ljharb> php.js is a php runtime written in javascript.
02:56 < niggler> they are trying to write javascript equivalents of php functions
02:56 < ljharb> i know. and they do it very badly.
02:56 < ljharb> i recommend you don't use any of their code.
02:57 < jwulf> sinclair|work, I think javascript demonstrates that each generation needs to invent the wheel for itself
02:57 < niggler> ljharb item #3 in DDG is php vm, the first two (including one identifying itself as php.js) is the phpjs project
02:57 < sinclair|work> jwulf: that's a nice way to view thing :)
02:58 < ljharb> niggler: http://phpjs.hertzen.com versus http://phpjs.org - two different things. i was talking about the former.
02:58 < blackcoat> Speaking of reinventing the wheel, I'm surprised by the wide array of competing testing tools for node:
02:58 < blackcoat> https://github.com/joyent/node/wiki/modules#wiki-testing
02:58 < ljharb> and again, http://phpjs.org is mostly junk.
02:58 < niggler> DDG search shows a few pages pointing to the latter that uses the phraes "php.js"
02:58 < niggler> sadly, those are #1 and $2
02:58 < RLa> and sometimes you need to reinvent the wheel to get the cruft out of things that has accumulated there with times
02:59 < niggler> sorry upon reading its php.Jjs
02:59 < RLa> so more like to get a clean state to solve todays problems, not yesterdays
02:59 < sinclair|work> niggler: i don't think there is a project for node to have CLR bindings
02:59 < niggler> sinclair|work node is a bunch of sugar atop v8.  v8 itself could be brought into the CLR
02:59 < sinclair|work> niggler: tho, i think i recall a long time ago someone here writing up the entirety of win32.h in js
03:00 < sinclair|work> niggler: oh, ive done that
03:00 < ljharb> niggler: not sure why you keep emphasizing ddg, who cares about that. they're different projects and i was referring to the php vm to make a point about all sorts of languages being runnable in JS.
03:00 < sinclair|work> binding v8 to things is old news
03:00 < sinclair|work> niggler: binding nodejs to the clr however...
03:00 < sinclair|work> (sounds like a challenge)
03:01 < niggler> will you accept said challenge sinclair|work?
03:01 < sinclair|work> niggler: i would if i could get native modules working on azure :(
03:02 < sinclair|work> niggler: i expect that writing managed clr assemblies to do heavy lifting would be way more accessible than writing native node modules
03:02 < sinclair|work> (and would bring a lot of .net heads over to node)
03:03 < sinclair|work> not to mention you would have the entire .net framework at your disposal, which means fast tracking integration into the typical ms enterprise
03:03 < jtomasrl> if i have a .js with a list of anonymous functions how can i export them?
03:04 < mdedetrich> a general question when dealing with cookies/sessions
03:04 < mdedetrich> is the general expection to set a massive expiry time to a cookie (like 40 years)
03:05 < blackcoat> 1 year is typical
03:05 < sinclair|work> niggler: !
03:05 < mdedetrich> or to set a timeout to a more standard time (like 30 days) and then just create a new cookie when the original expires
03:05 < niggler> lol
03:05 < freewil> mdedetrich, it depends on your purpose, if you need a secure login to something important (banking) then you will want to set it lower (probably < 30 mins)
03:05 < blackcoat> Indeed.
03:06 < mscdex> jtomasrl: exports.foo = function() { };
03:07 < MI6> joyent/libuv: bnoordhuis created branch uv_work_cancel - http://git.io/NsXbrQ
03:07 < jtomasrl> mscdex: im getting throw new TypeError('options.handlers must be an Array of Functions');
03:08 < mscdex> jtomasrl: what are you actually trying to do?
03:09 < jtomasrl> mscdex: instead of having a my routes function in app.js im moving them into a separate file
03:10 < aj910martin> Have issues understanding routing. I am using express. Here is what is working for my index.jade page http://pastie.org/5434263. But I am not able to what routes.index and exports.index is doing here
03:10 < mscdex> and? where does options.handlers come into play? i'm still missing more information here :-)
03:10 < aj910martin> *not able to understand what
03:10 < jtomasrl> mscdex: dunno :/
03:11 < mscdex> aj910martin: exports.index is the function you're executing for the path /
03:11 < mscdex> jtomasrl: got some code you can gist?
03:11 < aj910martin> mscdex: in the term `routes.index`, `index` suggests the file, right? 
03:12 < mscdex> aj910martin: anything you attach to exports is returned by require()
03:12 < jtomasrl> mscdex: https://gist.github.com/4146427
03:13 < mscdex> aj910martin: so your 'index' function is being called when a request comes in for the root path
03:13 < aj910martin> mscdex: but how does it know the index function is in the routes/index.js file?
03:14 < mscdex> aj910martin: because you required it?
03:14 < mscdex> aj910martin: index.js is a special filename
03:15 < mscdex> aj910martin: if you require a directory and it has "index.js", it will load that
03:15 < aj910martin> mscdex: I just mentioned `routes = require('./routes')` Does that include all files under routes folder
03:15 < aj910martin> mscdex: Ok
03:16 < mscdex> jtomasrl: for one thing, you have a double new which is unnecessary
03:16 < mscdex> jtomasrl: createServer already returns a new AuthServer object
03:16 < aj910martin> mscdex: And what is the recommended way to add routes for more files. For example I add a bunch of new pages under views, can I use wild cards to perform routing?
03:17 < mscdex> aj910martin: it's in the express docs, but yes you can use wildcards and placeholders as far as strings go, or a regex iirc
03:18 < mscdex> aj910martin: usually what i do is group related REST-like actions together in one file
03:18 < jtomasrl> mscdex: right, but it didnt solve any problem :/
03:18 < aj910martin> mscdex: Ok. Makes sense
03:18 < mscdex> aj910martin: like for a User model, you'd have a module that exports index, create, remove, update, etc functions
03:18 < blackcoat> Does anyone have a preferred library for integration testing a node app? I'm looking at Zombie.js, and wanted to know if anyone had other recommendations.
03:18 < mscdex> aj910martin: that operate on User
03:18 < warz>  in the npm semver for dependencies, what does the "~" do? i tried looking for it in semver docs but didnt find it. does it do something like within X number of versions lower or higher, or does it basically do >= ?
03:19 < chovy> blackcoat: i use selenium myself.
03:19 < mscdex> jtomasrl: right, still looking at it
03:19 < aj910martin> mscdex: Ya. In my case there are no models. I am basically dealing with a personal website. So there'll be randomly-named pages
03:20 < chovy> warz: it says upgrade 0.3.2 to 0.3.3, but not to 0.4.0
03:20 < blackcoat> chovy: Good point. I had only been looking at headless solutions.
03:20 < chovy> blackcoat: i'm only do small app though, i only test my custom model functions with mocha and should.js
03:21 < blackcoat> chovy: I'm already using mocha + should for my unit tests.
03:21 < chovy> most of my testing is in selenium though. i setup a new db for testing and clear it out before and after. Selenium creates all the users, etc.
03:21 < mscdex> jtomasrl: i don't see anything else wrong. is it your express version?
03:21 < blackcoat> On a related note, I've been fine using should.js, but I've seen a lot of people favoring Chai lately. Does anyone have a preference between the two?
03:22 < jtomasrl> msxdex: im using Restify
03:22 < mscdex> ohhh
03:22 < mscdex> no idea then :/
03:23 < MI6> joyent/libuv: Ben Noordhuis uv_work_cancel * 8187985 : unix, windows: add uv_work_cancel() - http://git.io/cFRv7Q
03:25 < jtomasrl> mscdex: i change the server.use to server.get and worked :S
03:34 < niggler> before i embark on this quixotic quest, anyone implement basic in nodejs?
03:35 < jwulf> niggler, go all the way -implement 8088 assembler in node.js, so that we can get real fine-grained control of our apps....
03:35 < mscdex> jwulf: go all the way - implement punch cards in node.js
03:36 < jwulf> lol
03:36 < jwulf> asynchronous batch processing!
03:36 < jwulf> it's the node.js way
03:37 < ljharb> niggler: http://www.calormen.com/applesoft/
03:37 < ljharb> http://www.codeproject.com/Articles/25069/JSBasic-A-BASIC-to-JavaScript-Compiler
03:37 < ljharb> you're not the first person to have that terrible idea.
03:38 < niggler> ljharb the terrible idea is far larger than basic->JS :)
03:38 < jwulf> where terrible == awesum
03:40 < mscdex> wait, awesum is a module
03:40 < mscdex> :-)
03:55 < rump> i hear a lot of fuss about handling err in callbacks.. but wont they bubble and eventually throw/exit all on their own?
03:55 < ljharb> that's thrown exceptions. errors in callbacks are just values passed to functions, which will be ignored if not handled.
03:55 < ljharb> and you wouldn't want your whole server to crash from an unhandled exception anyways
04:02 < rump> fair enough :)
04:07 < rump> i guess that is bugging me about these callbacks... is it totally removes the modularity of my functions
04:07 < rump> i mean maybe not now but i'd like to readDir without it automatically reading each file and then automatically doing another 3 callsbacks on it
04:08 < rump> ah wait thats why add cb to the function parameters
04:08 < ljharb> rump: modularity? if you want them reusable, have them take the callback in the first place
04:09 < ljharb> yes
04:09 < rump> so wait, is pretty much every function i ever make gonna have the (err, my, many, params, cb) signature?
04:09 < ljharb> yup
04:10 < freewil> using callbacks isnt really that much different than using return values
04:10 < freewil> you call a function with some arguments and instead of returning a value, you pass it to a callback
04:10 < freewil> only callback functions
04:10 < ljharb> yup
04:11 < freewil> var add = function (a, b, cb) { cb(a + b) };
04:11 < freewil> ^ that's not a callback function so it doesnt have an err argument
04:11 < freewil> it takes a callback argument instead
04:11 < freewil> var add = function (a, b, cb) { cb(null, a + b) };
04:11 < freewil> correction: pass null for err
04:12 < sinclair|work> freewill: function add(a:number, b:number, callback: (result:number) => void) : void {  callback(a+b); }
04:12 < rump> ok tryin to wrap brain about when its appropriate to pass err
04:13 < freewil> sinclair|work, great give me an example in java now
04:13 < rump> ok only in call backs
04:13 < rump> so when you're starting a series of callbacks, how do you execute the first in the series?
04:13 < rump> var blah = fn() {}(); ?
04:14 < ljharb> sinclair|work: seriously. i thought we went through this. use typescript if you want, but nobody in here wants to hear about it.
04:14 < sinclair|work> freewil: add(a, b, cb) is a poor example of callbacks
04:15 < sinclair|work> ljharb: i thought i had you on ignore
04:15 < ljharb> that's pretty cowardly if you did
04:15 < sinclair|work> ljharb: how so?
04:15 < ljharb> this is node.js. not node.ts. don't put typescript in here please.
04:15 < ljharb> because it's running away from your problems.
04:15  * sinclair|work notes he has SomeoneWeird on ignore
04:15 < ljharb> you're doing something antisocial, and you don't like that i'm commenting on it.
04:15 < niggler> what stops microsoft from hooking TS into nodejs?
04:16 < ljharb> rump: by convention, everything should pass an err. if there's no error, pass null
04:16 < SomeoneWeird> lawl.
04:16 < ljharb> niggler: nothing. but it still wouldn't belong in this chatroom.
04:16 < niggler> SomeoneWeird you need to do a dance
04:16 < sinclair|work> ljharb: im not running away from problems, im ignoring them, thats a good thing
04:16  * SomeoneWeird dances
04:16 < ljharb> it's really not.
04:16 < sinclair|work> how so?
04:16 < ljharb> sinclair|work: i'm not trolling you, nor insulting you, nor pestering you all the time. i'm just asking you not to talk about irrelevant things in this room.
04:16 < aandy> you know, like how they say ignorance is bliss
04:17 < sinclair|work> aandy: +1
04:17 < ljharb> how so? coping mechanisms might be a good way to work around a problem, but they don't solve it.
04:17 < aandy> that was actually directed as a +1 for ljharb but yeah
04:17 < ljharb> lol
04:17 < sinclair|work> ljharb: sorry, i really don't want this conversation, 
04:18 < ljharb> sinclair|work: i'll also add, supplying a TS example nobody asked for in a JS room is you trolling us.
04:18 < ljharb> if you don't want the conversation, stop mentioning typescript in here. it's that simple.
04:18 < sinclair|work> ljharb: oh who cares
04:18 < ljharb> obviously i do, and obviously i'm not the only one.
04:18 < ljharb> and obviously you do too, or else you'd take the path of least resistance and stop talking about it.
04:18 < sinclair|work> see, its pointless even talking about it
04:18 < niggler> strictly speaking this is a node chatroom
04:18 < ljharb> niggler: node dot JS
04:18 < niggler> so if there was an issue with npm and typescript
04:19 < dnz-> bor
04:19 < dnz-> bro
04:19 < dnz-> wat
04:19 < dnz-> wat
04:19 < dnz-> wta
04:19 < dnz-> wat
04:19 < dnz-> wat
04:19 < ljharb> niggler: yeah i'm not saying the topic is verboten, i'm just saying that it needs to be relevant.
04:19 < dnz-> wat
04:19 < dnz-> wta
04:19 < dnz-> harrow
04:19 < dnz-> boten anna ?
04:19 < sinclair|work> ^ see, this is annoying, unacceptable behaviour
04:19 < ljharb> oh goody, a chat flooder.
04:19 < niggler> e.g. if 'npm install typescript' failed
04:19 < dnz-> that dnz guy
04:19 < dnz-> what a fucking ass
04:19 < niggler> dnz- having fun with gray codes?
04:19 < ljharb> sinclair|work: i'll agree that i put him far higher than you on the "unacceptable" list.
04:20 < dnz-> niggler: nah just trying to get thru the day
04:20 < dnz-> all sick n stuff
04:20 < sinclair|work> ljharb: that's ok, i put you low on the code experience list
04:20 < ljharb> sinclair|work: see, that's just insulting, childish, and unnecessary.
04:20 < aandy> i think the morale is quite clear: there's probably a chan like #typescript or a very active google groups
04:20 < ljharb> and inaccurate, of course, but i'm not going to bother defending myself there.
04:21 < sinclair|work> ljharb: you just ranked me lower than a channel spammer, that's kinda insulting
04:21 < ljharb> yep - i already said, i'd never be so fucked up as to go to ##typescript and talk about anything other than typescript.
04:21 < dnz-> IM SO MAD AN INTELLECTUAL ABOUT PROGRAMMING JAVASCRIPT.
04:21 < ljharb> sinclair|work: actually i said you were less "unacceptable"
04:21 < sinclair|work> oh, sorry then
04:21 < sinclair|work> i misread
04:21 < ljharb> sinclair|work: so please reread things if you'd otherwise throw out an insult.
04:21 < dnz-> people on the internet are all reasonable and unacceptable
04:21  * sinclair|work apoligiezes
04:22 < aandy> i don't go posting python questions in my knitting facebook group, we try to use the correct channel. if it's not really about node, but kind of is, i generally put a disclaimer at the beginning, like "dunno if off topic, but <pre question>?"
04:22  * sinclair|work apologizes
04:22 < mdedetrich> does anyone know hot to $set multiple fields in the mongo.db driver
04:23 < ljharb> mdedetrich: i normally use #mongoosejs, and .set takes a hash of attributes. not sure on the native driver.
04:23 < sinclair|work> ljharb: do you know how many conversations i see in here on people talking about writing modules, and style, and general basic things?
04:23 < niggler> dunno if off topic, but is the following question OT: I am fully aware that node isnt javascript but it seems to run a version of ECMAscript.  Microsoft claims that typescript is also an ECMAscript.  does node support typescript?
04:23 < sinclair|work> ljharb: "what do you think about this style", or "why doesn't this work right"
04:23 < ljharb> sinclair|work: of course. but for all those things, if someone gets riled up, people engage in the discussion, or learn not to bring it up.
04:23 < sinclair|work> ljharb: i really wish i could suggest TS in those situations, 
04:23 < ljharb> sinclair|work: you seem to no longer wish to engage in a "merits of TS" discussion, which is totally your right. however, you haven't yet learned not to discuss it.
04:24 < ljharb> sinclair|work: you are more than welcome to, if you are *also* willing to tolerate and engage in the inevitable reactions.
04:24 < sinclair|work> ljharb: see, the reactions are silly
04:24 < mdedetrich> ljharb: nvm its working
04:24 < ljharb> you don't have the right to talk. you do, however, have the right to converse.
04:24 < sinclair|work> ljharb: id rather not
04:25 < dnz-> politically correct statements
04:25 < ljharb> sinclair|work: again, that's fine. but you can't speak, and also ignore the reactions - it's childish, rude, antisocial, et cetera. either be silent, or be engaging.
04:25 < sinclair|work> ljharb: being trolled by 3 or 4 anti MS zealots is a waste of my time
04:25 < ljharb> then don't bring it up in the first place.
04:25 < sinclair|work> id rather not engage
04:25 < ljharb> it's not that complex
04:25 < dnz-> you can ignore me all day
04:25 < dnz-> you will still be mad inside
04:25 < dnz-> intil you call me an asshole
04:25 < dnz-> on mirc
04:25 < polyrhythmic> dnz is right
04:25 < polyrhythmic> peace comes from inside
04:25 < ljharb> dnz-: mirc? that'd mean i was using windows :-)
04:26 < dnz-> mirc is a colloquial term for irc in my circles
04:26 < sinclair|work> ljharb: its not childish to ignore, its better than me having to mention TS anymore than than i would if id been defending it 
04:26 < sinclair|work> ljharb: ignoring prevents language wars
04:26 < sinclair|work> and they suck
04:26 < ljharb> sinclair|work: so does not bringing it up at all.
04:26 < dnz-> pfft ruby is the best
04:26 < dnz-> js is like a second rate ruby
04:26 < sinclair|work> ljharb: no, not at all
04:26 < aandy> dnz-: :D
04:26 < dnz-> there should be a client side integrated ruby interpreter in all browsers
04:27 < niggler> dnz- ruby2js
04:27 < sinclair|work> ljharb: a developer comes in here and is familiar with python, and knows nothing about js, you can talk to them in terms of python to help them out
04:27 < ljharb> sinclair|work: not bringing it up prevents language wars just the same as ignoring people. better, even.
04:27 < dnz-> yeah but python is better than ruby
04:27 <@mbalho> node.js channel patrons, read the channel topic please. thanks, the mgmt
04:27 < aandy> what about dependencies? then gems would make sites load sooo slow. bad throughput! ;)
04:27 < ljharb> sinclair|work: the day a C# developer comes in here trying to learn JS, you're welcome to analogize with TS. that's not what you've been doing.
04:27 < rump> ljharb: so how do i go on to the 3rd jump without having to declare it in the 1st at line 8 as extra parameter?  i guess i just dont see an easy way to do this modularly http://pastebin.ubuntu.com/1387302/
04:27 < dnz-> there should be python2ruby2js2html7
04:27 < sinclair|work> ljharb: read mbalho's comment, i don't want to continue this conversation
04:27 < mdedetrich> back to typescript again
04:27 < mdedetrich> I see
04:28 < aandy> ironically throwing troll bait IS the same as engaging. it's like saying "javascript sucks. oh i'll just leave this right here, muhaha *off to lunch*"
04:28 < dnz-> there should be a C# interpreter in all browsers
04:28 < sinclair|work> ljharb: continue to bug me, and your on ignore
04:28 < niggler> dnz- i think vb and perl and php are missing
04:28 < dnz-> yeah
04:28 < ljharb> sinclair|work: i read his comment, and i'm being friendly. ignoring people isn't friendly. neither is inciting anger by mentioning things you don't wish to discuss.
04:28 < dnz-> well actually
04:28 < mdedetrich> scala > C#
04:28 < dnz-> there was vb in IE
04:28 < mdedetrich> by loads
04:28 < mdedetrich> as a language
04:28 < dnz-> for client side scripting
04:28 < dnz-> but i dont think its supported for some time now
04:28 < mdedetrich> if I was going to make any language work in browser
04:29 < mdedetrich> it would be scala
04:29 < mdedetrich> not C#
04:29 < sinclair|work> ljharb: you are preventing the channel from talking about node while trying to tell me what i can and can't talk about 
04:29 < dnz-> mdedetrich: what about scheme
04:29 < sinclair|work> ljharb: so stop it please
04:29 < ljharb> you are doing the same by bringing up typescript, which has nothing to do with node.
04:29 < dnz-> ljharb: what about coffeescript
04:29 < mdedetrich> dnz-: scheme is way too esotiric
04:29 < ljharb> i'd be happy to drop the topic, but you continue to bring it up.
04:29 < dnz-> mdedetrich: what about tcl
04:29 < dnz-> or orlang
04:29 < aandy> [05:26] < ljharb> sinclair|work: you are preventing the channel from talking about node while trying to tell me what i can and can't talk about
04:29 < dnz-> *erlang
04:29 < sinclair|work> ljharb: ts is just as applicable for node as js is
04:29 < aandy> fixed that for ya
04:29 < aandy> seriously, stop
04:29 < mdedetrich> erlang is a language designed for specific domain
04:30 < dnz-> no its not
04:30 < dnz-> what u know bout erlang son
04:30 < niggler> erlang is the lingua franca and should be supported in every possible context
04:30 < mdedetrich> well yeah, its designed for concurrent systems
04:30 < dnz-> ^ this guy
04:30 < mdedetrich> I wouldn't want to program in erland for the client side
04:30 < dnz-> niggler: erlang should be on bus stops
04:30 < _Tristan> Hello. This doesn't work, and I don't know why. It's probably something simple... can anybody help? http://pastebin.com/0cxWDisp
04:30 < mdedetrich> *erlang
04:30 < dnz-> erlang is the best language
04:31 < niggler> !pastebin @_Tristan
04:31 < dnz-> its like bacon of programming languages
04:31 < dnz-> internet is obsessed with it
04:31 < niggler> ok that needs to be added to purr
04:31 < dnz-> there needs to be erlang native cpus
04:31 < mdedetrich> scala is the best language imho
04:31 < _Tristan> is pastebin.com a deeply unacceptable tool?
04:31 < dnz-> yes
04:31 < dnz-> its for hacker to hack on
04:31 < dnz-> they put 0 days on it
04:31 < sinclair|work> there seems to be a fascination with erlang of late
04:31 < dnz-> yeah erlang is great
04:31 < dnz-> you should learn it
04:31 < dnz-> it wouldnt waste ANY of your time
04:32 < dnz-> not a single second
04:32 < dnz-> would be better spend sitton on broken glass
04:32 < dnz-> with no pants on
04:32 < sinclair|work> dnz-: please stop spamming
04:32 < exit2> any advice on something to build to strengthen my node knowledge?
04:32 < mdedetrich> erlang is very good at doing what its designed for
04:32 < mikeputnam> http://sprunge.us/   for great pastebin justice
04:32 < dnz-> this aint spam son, this is convesation
04:32 < ljharb> dnz-: i can't find the npm module that crashes your entire server if you have coffeescript anywhere in your app, but that's my reply to your question :-)
04:32 < dnz-> ljharb: what about coffeescript ?
04:33 < mdedetrich> it makes you stay up all night
04:33 < mdedetrich> thats what
04:33 <@mbalho>  /join #programminglanguageopinions
04:33 < dnz-> what about late nights ?
04:33 < dnz->  /join #latenightlanguageoinions
04:33 < mdedetrich> eww inions
04:33 < sinclair|work> exit2: yeah, something with streams
04:33 < rump> lost how to tie these two functions together and get contents of files to play with in serial function 3.  http://pastebin.ubuntu.com/1387317/
04:33 < niggler> coffeescript is one of the many reasons programmers drink coffee
04:33 < mdedetrich> *onions
04:34 < niggler> in an effort to be parsimonious, coffeescript introduced many programmers to 4 AM
04:34 < dnz-> niggler: it took them a while to catch on
04:34 < dnz-> i think some still havent
04:34 < aandy> i write in cs just for the novelty. "watcha doin?", "drinkin' and writin' some coffee HAHAHAHHAHHAHAHAH ha, oh wait come back"
04:35 < sinclair|work> niggler: lol
04:35 < dnz-> so if im doing concurrent stream handling, asynchronously, why would i want to do it in nodejs and NOT erlang ?
04:35 < sinclair|work> niggler: hows your ruby
04:35 < dnz-> the asnwer better be shiny and red
04:35 < dnz-> on a pimp ring
04:36 < niggler> the right answer is "I am competent enough to put it on a resume" although that bar is fairly low
04:36 < ljharb> rump: fyi generally i'd throw new Error(err) so you get a backtrace
04:36 < dnz-> wrong answer
04:36 < dnz-> its all about pimp rings
04:36 < niggler> i don't claim to understand the zen of ruby
04:36 < rump> ljharb: ok ty
04:36 < ljharb> rump: also /\.ejs$/.text() will always be false
04:36 < dnz-> GEN-YOU-EYEN DIA-MONDS BABY
04:36 < sinclair|work> niggler: its a bit mysterious 
04:36 < niggler> dude dnz- you need better jews on your jewry
04:36 < ljharb> rump: .test takes an argument and you're not passing it one.
04:36 < dnz-> damn son
04:36 < rump> ljharb: ah yeah it was working before i just didnt paste it over correctly
04:36 < dnz-> ok back to work, you guys argue more baout programming languages now
04:37 < dnz-> also
04:37 < sinclair|work> niggler: mind you, i completely ignored ruby and did other things
04:37 < dnz-> WRITE EVERYTHING IN C
04:37 < niggler> dnz- better check your gold -- you may have been tungsten'd
04:37 < sinclair|work> niggler: want to help me write a node clr binding?
04:38 < dnz-> omg i need help to write my javascripts
04:38 < dnz-> write it yourself sinclair|work 
04:38 < rump> ljharb: but im lost in how to link them and take the file contents to a 3rd function
04:38 < niggler> wanna help me write a vba to javascript translator?
04:38 < dnz-> haha
04:38 < dnz-> di mdim dim
04:38 < dnz-> dim dim dim
04:38 < niggler> the xls format makes me cry
04:38 < dnz->  ' commends and shit
04:38 < sinclair|work> niggler: with a clr binding, you got all the compiler as a service good ness you could ever want! (that includes VB.NET if you were do inclined)
04:38 < sinclair|work> not sure why....
04:39 < badem4o> Who said xls
04:39 < badem4o> are you kidding me
04:39 < dnz-> haha
04:39 < dnz-> so mad
04:39  * dnz- &
04:39 < aandy> someone just ruined his day
04:39 < aandy> lol
04:39 < niggler> haha
04:39 < niggler> i want to replace excel
04:39 < polyrhythmic> someone already replaced excel with JS
04:39 < niggler> no one made a serious effort to replace excel
04:39 < polyrhythmic> did a whole presentation at CascadiaJS. it's available to stream somewhere
04:39 < sinclair|work> niggler: vbs is quite possibly the worst thing ive ever had to read
04:40 < niggler> polyrhythmic with vba etc?
04:40 < sinclair|work> sure, why not
04:40 < polyrhythmic> nobody tell the finance industry how terrible VB and excel is, they might have to stop making so much money
04:41 < polyrhythmic> PLT Hulk must have come in through the window today and smashed this plae up
04:41 < polyrhythmic> *place
04:41 < aandy> _Tristan: sorry, you got kind of lost in the spam. did you solve it? you're not emitting the event you're listening for
04:41 <@mbalho> polyrhythmic: which talk http://www.youtube.com/channel/UCIP244iNzbn4iEkDOgczvcQ
04:41 < _Tristan> aandy: yeah, I solved it, I was using call instead of apply. Thanks, though
04:41 < aandy> ok :)
04:42 < niggler> polyrhythmic simplymeasured?
04:42 < rump> ljharb: hmm, maybe a decent way is to just add the results of the "modular functions" to a shared variable outside of their scopes?  thats how i could get return values i guess
04:43 < rump> use that cb as function(err, data) { toplevelvar = data; } ?
04:43 < polyrhythmic> mbalho: https://github.com/cascadiajs/cascadiajs.github.com/blob/master/proposal/excel-in-the-browser_halffullheart.md http://www.youtube.com/watch?v=LHlElf1aJJY
04:45 < rump> hmm if my async function returns a variable and i invoke like var myvar = asyncfunc();  can i use the return value in top to bottom execution or it will be undefined?
04:46 < rump> haha jeez, i read like 2 js books and it all seemed breezy and i was ecstatic but when feet hit the pavement i cant figure out basic control flow without turning into throw away spaghetti
04:46 < rump> its real convenient when book examples only show callback functions like console.log(data) and never have to piece it all together
04:49 < rump> fun though :)
04:51 < polyrhythmic> rump: it gets easier and easier
04:51 < rump> yeah i get the impression that its all fun and enjoyable and then there are like 2-3 really odd thinks to grok but once they click its no problem
04:52 < polyrhythmic> you can't use callbacks directly in top to bottom flow exactly, without some kind of flow control lib
04:53 < polyrhythmic> it's good practice to write code in a top to bottom style where possible, for readability
04:54 < polyrhythmic> niggler: yes simplymeasured, just saw your hilight
04:56 < polyrhythmic> rump: really in programming, nothing happens instantly... therefore thinking in callbacks imho better reflects how the real world works
04:57 < polyrhythmic> the cascadiajs talk on johnny-five was a nice example of that
04:59 < mscdex> darn, here i thought he would be supporting parsing the binary XLS format ;-)
05:00 < niggler> yeah im similarly disappointed mscdex
05:01 < mscdex> i started a module awhile back for doing that, but didn't get into parsing the actual word/excel data
05:01 < mscdex> just the meta data
05:02 < mscdex> that guy wasn't kidding about the documentation for the file formats
05:02 < niggler> the biggest hurdle is in figuring out the correct recalculation order
05:02 < niggler> *at least, for xls
05:02 < rump> polyrhythmic: http://pastebin.ubuntu.com/1388342/ i just cant figure out how to link all those functions together but keeping them modularly and without using async
05:02 < rump> whats the standard way of doing this?
05:02 < niggler> its not documented, which is why most implementations skip it
05:03 < rump> i just want to readDir->readFile->addSite->compress->serve
05:03 < rump> ive had each function working and doing as i want but as soon as i try and link it crashes
05:03 < rump> and everything was perfect until i had to compress with zlib which doesnt have a Sync version
05:03 < rump> because i was only using sync versions of these functions i was shielded for quite a while
05:06 < rump> gross
05:10 < polyrhythmic> rump: personally, I would probably use Promises, because they have built-in error bubbling to handle the nested erros
05:11 < polyrhythmic> *errors, although some people prefer to program as you are with the plain callbacks
05:11 < rump> ok i havent read up on promises yet.. thought they were depreciated
05:11 < polyrhythmic> rump: they aren't built into Node but the Q lib is good
05:12 < rump> cant i just listen for event 'error' and call it a day
05:13 < _numbers> why do ppl idolize TJ
05:13 < rump> _numbers: 200+ projects and you've probably used half
05:13 < rump> isnt that enough
05:14 < timeturner> TJ is a programming super-being
05:14 < _numbers> ah a big contributor in terms of porting libs to javascript
05:14 < timeturner> honestly, he has made a lot of good modules
05:14 < polyrhythmic> rump: it's a matter of personal preference (as we have seen in discussion today :))
05:14 < timeturner> that's the most likely reason
05:14 < polyrhythmic> rump: in regards to promises vs. events I mean
05:15 < rump> id like to learn all 3
05:15 < rump> i mean im tempted to just throw async in and enjoy the black magic
05:15 < rump> but thats not really helping me... i hate relying on magic without understanding
05:16 < warz> whys it gotta be black?
05:16 < _numbers> in terms of github repos he actually has closer to 119 public
05:17 < rump> warz: cause its the type that creates fear and mystery and not giggles
05:17 < rump> bobo doesn't bring this magic to birthday parties
05:19 < polyrhythmic> rump: async isn't magic... all programming functions are asynchronous under the hood
05:19 < rump> polyrhythmic: to me i mean
05:19 < rump> i want to understand it so its not
05:20 < niggler> warz microsoft brings the white magic
05:20 < polyrhythmic> rump: if you want to understand it, write it the 'ugly' way with callbacks everywhere
05:21 < rump> did that, had it working
05:21 < rump> mind was boggling
05:21 < polyrhythmic> exactly
05:21 < rump> considered php for a moment
05:21 < polyrhythmic> programming is about abstractions... I can't write ASM without a manual either
05:21 < polyrhythmic> using an async lib is no different than writing JS instead of ASM
05:22 < rump> ah, fair enough
05:22 < rump> but how do all these other guys seem to do it so effortlessly
05:22 < L10n> practice
05:22 < polyrhythmic> practice... and because they are stubborn tbh
05:22 < rump> speaking of tj, i read his code and theres never more than 3-5 lines together and it all meshes
05:22 < polyrhythmic> they refuse to do it another way until they are used to it
05:22 < rump> yea well i want that hehe
05:23 < rump> guess i need to spend the night and go read other peoples source
05:23 < mdedetrich> man the one thing that I hate about moving to a dynamic language
05:23 < mdedetrich> is the number of retarded mistakes you make
05:23 < mdedetrich> and then end up having to debug
05:23 < rump> prob is, it seems like all the docs i read avoid this and their callbacks are just console.log(result)
05:24 < rump> or i read module code and its easy cause they export single values
05:24 < polyrhythmic> rump: it's always good to further your understanding of what's going on under the hood, because then you know why you are using the abstraction. but then when it comes to being productive, it is often time to use the abstraction
05:24 < rump> never proper examples i can read
05:24 <@mbalho> go read github.com/substack for a while
05:24 < rump> polyrhythmic: oh for sure... soon as i understand it ill probably never consider it again :)
05:24 < polyrhythmic> rump: pick an open source project that interests you and dive in... that will be more real-world
05:24 < _numbers> anyone here tried Mincer and Grunt? did you end up using one over the other?
05:24 < polyrhythmic> I used to read the jQuery source... and I still do sometimes
05:25 < polyrhythmic> even they gave up and use promises now though :) if that tells you anything
05:26 < polyrhythmic> any of the node libs that do file IO (like you are doing) should be good examples
05:26 < mdedetrich> jquery source is kinda ugly
05:26 < polyrhythmic> since it's hard to even pretend that IO is synchronous
05:26 < mdedetrich> its not a really good example
05:27 < polyrhythmic> jquery source has gotten uglier over the years... hence why I 'used to'
05:27 < mdedetrich> if you think IO in js is synchronous, you are in the wrong world buggy =D
05:27 < mdedetrich> *buddy
05:30 < imm> i like buggy better *
05:37 < sshaginyan> Can someone tell me how this returns true? /[a-zA-Z].*/.test("J0hn")
05:38 < substack> sshaginyan: that matches a single a-zA-z character and zero or more characters
05:38 < substack> you probably want to anchor that with ^ and/or $
05:39 < substack> >> /^[a-zA-Z]+$/.test('J0hn')
05:39 < purr> substack: (boolean) false
05:39 < substack> >> /^[a-zA-Z]+$/.test('John')
05:39 < purr> substack: (boolean) true
05:39 < polyrhythmic> regexpal.com ftw
05:40 < CoverSlide> .* basically means match anything. the [a-zA-Z] matches the first character only and continues to the next expression
05:40 < CoverSlide> so it would pretty much be the same if you left out the .* entirely
05:40 < CoverSlide> actually learning regex,
05:41 < CoverSlide> priceless
05:42  * substack wrote a lot of perl in high school
05:42  * Oddman feels sorry for substack
05:42 < substack> Oddman: perl is fun
05:42 < Oddman> I hated it
05:42 < Oddman> functions... ugh
05:42 < substack> sub stack {}
05:43 < ljharb> substack: got a second to help me out with some stream stuff?
05:44 < substack> sub stack { $ perl -le'sub stack { $_=shift; s/(e+)/q{O}x(length$1)/e; uc } print stack "beep"'
05:44 < substack> wups
05:44 < substack> $ perl -le'sub stack { $_=shift; s/(e+)/q{O}x(length$1)/e; uc } print stack "beep"'
05:44 < substack> BOOP
05:45 < substack> ljharb: you should just ask your questions instead
05:45 < ljharb> ha, fair enough
05:45 < substack> asking to ask is too much overhead
05:45 < dnz-> wahat u know about overhead perl son
05:46 < substack> dnz-: http://www.perlmonks.org/?node_id=563923
05:46 < dnz-> substack: y u want ur links son
05:46 < dnz-> i aint clickin dat
05:46 < dnz-> might b a viruz
05:46 < dnz-> lul
05:46 < dnz-> XD
05:47 < ljharb> so i'm trying to write a node script that generates a sprite from a directory of images - and all the existing modules i've seen don't seem to work for me. i'm ok going with base64-encoded images in a single css file. basically, i want to read from the fs, and for each file, append a chunk to the output css file
05:47 < ljharb> but i don't know anything about streams so i don't know where to start
05:47 < ljharb> (i'll also happily adjust my methods if there's a better option)
05:47 < substack> ljharb: I wrote some stream docs https://github.com/substack/stream-handbook
05:48 < dnz-> hi, im tryin to build a spaceship but i dont know shit about welding so if you guys can help m out dat wud b gr8
05:48 < substack> ljharb: you could create a readable stream and emit 'data' chunks for each file
05:48 < substack> fs.readdir() to grab the file list
05:48 < substack> then emit 'end' when you're done
05:49 < dnz-> emit chunks like when u  drink a bottle of vodka right after spaghetti and meatballs
05:49 < dnz-> or like
05:49 < ljharb> hm, ok, thanks - i'll start with your docs
05:49 < dnz-> chikin curry
05:49 < dnz-> 14 meatpies and a case of rum
05:49 < dnz-> get it up ya
05:51 < imm> good looking docs on stream substack !
05:52 < ljharb> substack: can i use .pipe() similar to: stream.pipe(base64encode).pipe(function (encoded) { var output = encoded; return output; }).pipe(writeToCSSFile) ?
05:52  * myndzi dances
05:53 < myndzi> my output is 30% the size of deflate and getting smaller :D
05:53 < myndzi> yay exploitation of purpose-specific patterns
05:53 < substack> ljharb: use through(fn) for the middle part https://github.com/dominictarr/through
05:54 < ljharb> so stream.pipe(base64encode).through(modifyOutput, writeToCSSFile)?
06:00 < ljharb> substack: so in this case, i'm reading N images, and encoding their data to base64, and dumping something like ".foo { background-image: url(" + base64encodeddata + "); }" to the css file - so the stream is going to give me chunks of image data, right? so how do i ensure that all the encoded data arrives in the right order, and in between the prefix and suffix strings?
06:06 < _numbers> anyone using jade or haml + handlebars with connect/express? what libs did you go with? 
06:27 < dileep> anybody have jump start node.js pdfd
06:34 < ljharb> why would stream.pipe(otherstream) return undefined?
06:34 < sinclair|work> ljharb: is encoding images in your css a good idea?
06:35 < ljharb> sinclair|work: possibly not. i have 2-300 icon files, that have to be kept as separate files. however, when displaying them to the user, i want it done as a sprite. so basically anything that lets me do it in one http request, instead of N, is what i want.
06:36 < ljharb> sinclair|work: the first option i considered was using a node module to build an image sprite/css file combo, but those didn't work for me, and then i found a module that base64encodes them, and that seemed more straightforward
06:36 < sinclair|work> ljharb: why not store all your icon assets in 1 one file, and map them to css classes?
06:36 < ljharb> because they're iOS app tab bar icons. they have to be separate.
06:36 < ljharb> and i don't want to manually make a sprite :-)
06:36 < sinclair|work> ljharb: could automate it
06:36 < ljharb> so i figured i'd learn more about streams by writing the process myself, since i couldn't find a module that easily does it
06:37 < ljharb> sinclair|work: that's what i initially was looking for. a module that generates the sprite, and the css background-positions. but it didn't work as advertised.
06:37 < sinclair|work> ljharb: this is trivial stuff, ill code it for you 
06:37 < ljharb> thanks, i'd love to see even a very basic working example
06:38 < sinclair|work> 1) name the image assets as css class you want them as
06:38 < ljharb> yeah that part's fine, i can use a parsed filename as a css class
06:38 < sinclair|work> 2) create a image containing all icons, and store their positions in that image
06:38 < sinclair|work> 3) map offsets to css classes
06:38 < sinclair|work> done
06:39 < ljharb> right - so the "create an image" step is the one that's difficult
06:39 < ljharb> because i want to do it in a nice, nonblocking fashion
06:39 < ljharb> otherwise i'd just write a php script (since the rest of the app's in php)
06:39 < sinclair|work> why do you need this automated?
06:39 < ljharb> because i'm going to be given another 100 icons to add next week, and more in the future, and automation > manual labor :-)
06:40 < ljharb> it doesn't have to be on-the-fly, just a one-off
06:40 < sinclair|work> what about the html canvas node plug in?
06:40 < sinclair|work> whatever its called, node-canvas or something
06:40 < ljharb> hm, lookin
06:40 < ljharb> g
06:41 < sinclair|work> https://github.com/LearnBoost/node-canvas
06:41 < ljharb> hm, that requires a lib called cairo
06:41 < ljharb> i already have imagemagick, so i'd rather use that if i need image manipulation
06:42 < sinclair|work> Cairo... wtf
06:42 < ljharb> also, i don't know anything about canvas, so i'd have to learn how it works :-)
06:42 < ljharb> i don't see a real problem with a single css file that has all the icons in it
06:43 < ljharb> it seems straightforward enough, and i could write it easily with promises, but i haven't gotten my head around streams yet
06:43 < sinclair|work> ljharb: i do, changing icons means changing your css 
06:43 < dileep> how i connect ma html,css file connectiviry with node.js also with mysql
06:43 < tantom> wc
06:44 < ljharb> sinclair|work: i'll never change an icon, only add, and adding icons means changing my css too. as a sprite, i'd have to change the image resource anyways, so i'd lose the benefit of caching in either case
06:44 < ljharb> (this would be a standalone css file, i don't concat assets)
06:44 < sinclair|work> why did they use Cairo instead of SDL
06:45 < sinclair|work> maybe they weren't keen on exposing the other multimedia aspects
06:45 < sinclair|work> fair enough i guess
06:46 < sinclair|work> ljharb: i still think its not a good idea
06:47 < ljharb> i'm still open to doing the image and css classes separately, if there's a module that will work for me
06:47 < ljharb> i started with node-spritesheet but it has lots of bugs
06:48 < sinclair|work> ljharb: its a trivial script to do what your asking
06:51 < sinclair|work> ljharb: http://aheckmann.github.com/gm/
06:51 < sinclair|work> ljharb: nodejs's image support is a bit on the lackluster side, use PHP
06:52 < sinclair|work> ljharb: or find a tool to do it
06:53 < ljharb> :-) i can do it in php, i'd just rather not
06:53 < ljharb> and i know it's a trivial script, but i lack the understanding of streams to make it trivial.
06:53 < ljharb> i'll check out gm too
06:55 < NaN> hi there
06:56 < NaN> I had installed node on my fedora machine, but now my chromium isn't working anymore and I can't re-install it because a v8 dependency problem, what can you recommend me?
06:58 < the_regulator_gu> what's up
07:16 < DTrejo> mikeal tweeted some point about node engineers having a salary premium; anyone remember what it is?
07:25 < _numbers> ljharb: not sure if this helps you, but i am thinking of using mincer + stylus + stylus-image to handle my css sprite compilation
07:25 < ljharb> yeah, i saw lots of solutions for stylus but i'd rather not add stylus
07:26 < ljharb> i'm slowly figuring out how to do it myself with streams, but nothing's been very informative in helping me do that.
07:27 < ljharb> all the docs i've read just tell me how to use other people's stream modules, but not really how to make my own, or how to add my own into a chain. also chaining pipes seems broken
07:31 < timeturner> ljharb: substack wrote a huge article on streams. It's somewhere on github.
07:58 < ljharb> timeturner: yes, i'm afraid it's not been helpful for me either.
08:04 < sinclair|work> ljharb: done it
08:05 < sinclair|work> ljharb: http://pastebin.com/B0QvC7Zf
08:05 < sinclair|work> ljharb: i even write it for you in ts because i know how much you like TS
08:07 < ljharb> why is the mime type necessary? isn't it just writing files?
08:07 < ljharb> oh for the data uri, right
08:08 < ljharb> so that's not using streams at all. thank you for writing it, but that wasn't the goal, otherwise i'd just have done it in php :-)
08:08 < sinclair|work> ljharb: you can mess with the buffer writing on line 52
08:09 < ljharb> totally, except that i don't understand how streams work :-) so what i'm really looking for is a basic example of the actual stream code i need
08:10 < sinclair|work> ljharb: replace the StringBuffer with a file stream, and call the write methods on that
08:11 < GreenDog> wow, that's the first time i've actually seen someone use ts
08:11 < ljharb> GreenDog: probably the only ;-)
08:11 < sinclair|work> GreenDog: that's ok, professionals use TS
08:11 < ljharb> professionals use something that's been around a month and a half?
08:12 < ljharb> sinclair|work: so the fs.createReadStream - how can you know that the "data" chunk will represent the entirety of the file?
08:12 < sinclair|work> ljharb: i don't
08:12 < GreenDog> i'm a professional!
08:13 < sinclair|work> ljharb: you can build up the data however and wait for the end or close or whatever it is, but as your css images are 64kb limited (or near abouts, you wont have an issue)
08:13 < sinclair|work> ljharb: ive been using nodejs as long as ive been using ts
08:14 < ljharb> so less than 2 months?
08:14 < GreenDog> whoa
08:14 < sinclair|work> something like that
08:14 < ljharb> sinclair|work: ok so i know i can listen to 'data', and i can listen to 'end' - so you're saying i can build up the buffer in 'data', but only write it to the file in 'end'?
08:14 < GreenDog> but yer super smart!
08:14 < ljharb> that's basically what i've cobbled together so far
08:14 < sinclair|work> ljharb: thats all there is to it
08:14 < sinclair|work> nothing more
08:15 < sinclair|work> however, there was one glaring omission from node
08:15  * sinclair|work wants a memorystream
08:15 < sinclair|work> or a stream.ReadWriteStream
08:15 < mike029> hey everyone, im interested in using node.js for a project involving drupal and an ios/android app. could anyone point me in the right direction to start learning? i figured i should just ask the expert source first instead of wasting all my time googling with no result. im a complete newbie to node.js, but im decently familiar with jquery and javascript to the level that i can guess most things correctly
08:15 < ljharb> there definitely are duplex streams.
08:15 < sinclair|work> ljharb: i dont really want to write my own
08:16 < guor> mike029: node.js doesn't work on ios afaik
08:16 < sinclair|work> ljharb: but as one doesn't exist i might
08:16 < sinclair|work> ljharb: for example...
08:16 < mike029> guor, sorry, i should have been more clear, im using phonegap
08:16 < sinclair|work> var memorystream = new MemoryStream();
08:16 < sinclair|work> readstream.pipe(memorystream);
08:16 < ljharb> sinclair|work: wait - why'd you bother making a StringBuffer class rather than just using an array, and pushing onto it?
08:16 < GreenDog> are you using node.js as a backend server?
08:16 < guor> mike029, is the node part supposed to run on the phone?
08:17 < guor> in that case, I don't see how phonegap helps
08:17 < sinclair|work> ljharb: because its an object, and i wanted to annoy you with its candy syntax
08:17 < mike029> guor, that's why im asking haha, i have no idea where to begin
08:17 < mike029> phonegap is a website wrapped in a native container
08:17 < ljharb> k :-) definitely overuse of nobjects. "new" is a code smell.
08:17 < guor> apple is not going to allow any javascript runtime with JIT to run on the ios
08:17 < mike029> it's just a webpage, but you get access to native stuff like gps, camera, etc
08:17 < sinclair|work> ljharb: i was looking for a memorystream
08:17 < ljharb> ah
08:17 < GreenDog> i think node.js is not what you are looking for
08:17 < sinclair|work> ljharb: and opted for a StringBuffer insstead
08:18 < ljharb> guor: phonegap is a cheap native wrapper around a web app that could normally run in the phone's web browser. plus, some extra JS hooks that let you access native APIs.
08:18 < guor> oh in that case you are probably fine, the ios/android part has nothing to do with node.js then
08:18 < ljharb> guor: node is a web server technology, and is something a phonegap app could talk to - but the phonegap content has to be local.
08:18 < guor> ljharb, i have experience with both, it's just not clear what mike029 needs from what he wrote
08:19 < mike029> node.js is definitely the solution, it's a drupal backend using REST to pull info down into a webpage with jquery doing some magic to trick people into thinking it's a native app....maybe i should ask this: is there an online tutorial series for node.js that people find useful??
08:19 < ljharb> mike029: if you're using drupal, that's your server. drop drupal and use node instead :-)
08:20 < sinclair|work> or use drupal and embrace its wealth of modules
08:20 < guor> sounds like you are going to create a webapp masked by phonegap. the server you use really doesn't matter unless you go more specific
08:21 < guor> you could use even use php \o/
08:21 < guor> -use
08:21 < mike029> the project is massive and drupal suits it perfectly haha
08:21 < sinclair|work> ljharb: so, you have only ever done PHP and nodejs
08:21 < sinclair|work> ljharb: ?
08:21 < mike029> basically i want it so when content is updated it will alert the phone somehow
08:21 < mike029> without me having to make it manually refresh every minute or so
08:22 < ljharb> sinclair|work: nope, i've done lots of things. but of course, how many languages or frameworks one can list is never an indicator of anything :-)
08:22 < GreenDog> ooooh
08:22 < mike029> i was under the impression that's what node.js functionality is
08:22 < ljharb> mike029: nope - node is a webserver. you'd use it OR drupal OR php OR ruby on rails OR something else.
08:22 < dileep> sha256 code needed in javasript
08:22 < sinclair|work> ljharb: but you have used streams before right?
08:22 < GreenDog> if i knew python i'd use drupal
08:22 < ljharb> sinclair|work: nope, i hadn't. everything i've used masks that concept behind simple read and write metaphors, so i've never had to grok it
08:23 < ljharb> dileep: please don't come in and bark orders. also, don't do encryption in client-side JS.
08:23 < mike029> i found a chat module that uses node.js now...but i was hoping to use drupal to fire something when content is updated
08:23 < sinclair|work> ljharb: erm...ok
08:23 < mike029> ljharb, ok cool, i know nothing about this
08:23 < dileep> ljharb,i want to do server side,
08:24 < guor> dileep, dig into the code of kraken crypto benchmarks
08:24 < ljharb> dileep: http://code.google.com/p/crypto-js/ first result in google
08:24 < guor> then contain a minified version of some crypto stuff :)
08:24 < guor> *they
08:24 < mike029> do you guys have a good starting place to learn about it (maybe it's not suitable for this project but just to familiarize myself for the future)??
08:24 < dileep> thankx buddy
08:24 < ljharb> next time google first :_)
08:24 < ljharb> :-)
08:25 < GreenDog> mike029, you might want to search sockets, websockets on the drupal website
08:25 < dileep> already done that found somethn else...dats y i put here
08:25 < ljharb> i typed "sha 256 javascript".
08:26 < mike029> greendog, thanks for the heads up :-)
08:26 < GreenDog> http://www.nodebeginner.org
08:28 < mike029> greendog, perfect, thanks for the help 
08:28 < mike029> greendog++
08:28 < GreenDog> YW
08:29 < dileep> well anyway thanks bussy
08:29 < dileep> buddy*
08:31 < GreenDog> hi [AD]Turbo!
08:31 < sinclair|work> ljharb: http://pastebin.com/kUMsKtWu
08:32 < ljharb> oh whoa, you can just read and encode into base64 right from the filesystem? nice
08:32 < [AD]Turbo> hi there
08:32 < bakadesu> has anyone used davis.js?  I can't figure out how to call a route on page load
08:33 < sinclair|work> ljharb: you can add the extra step of buffering the output, and error handling
08:33 < sinclair|work> ljharb: npm install typescript -g
08:34 < ljharb> thanks, i'll just read what you wrote and see if i can adapt it :-)
08:35 < GreenDog> you know waht i like most about npm?
08:35 < GreenDog> it stands for something
08:35 < GreenDog> unlike easy_install or pear
08:35 < GreenDog> actually, pear might, but who knows
08:36  * sinclair|work likes Nuget
08:36 < deoxxa> GreenDog: and what does it stand for?
08:36 < sinclair|work> but, i bet everyone else here hates it
08:36 < GreenDog> node package manager
08:36 < deoxxa> wrong!
08:36 < GreenDog> i guess
08:36 < GreenDog> hehe
08:36 < GreenDog> node something
08:36 < ljharb> you don't think easy_install means something?
08:36 < GreenDog> how would anyone know that relates to python?
08:36 < deoxxa> GreenDog: https://npmjs.org/doc/faq.html#If-npm-is-an-acronym-why-is-it-never-capitalized
08:37 < GreenDog> hahaha
08:37 < GreenDog> those funny npm guys
08:38 < GreenDog> i bet it stands for node package manager no matter what they like to claim
08:38 < deoxxa> wat
08:39 < GreenDog> and then they decide, "hey let's mess with people and tell em it doesn't stand for that!"
08:40 < GreenDog> node pkgmakeinsg
08:40 < GreenDog> node pkgmakeinst
08:40 < GreenDog> great, now i have to hate npm cuz it doesn't stand for something
08:42 < Doomground> anyone running their nodejs program from supervisor here?
08:42 < Doomground> if my nodejs app shuts down unexpectedly
08:43 < Doomground> what's the best way to see why? error logs are empty... is there a way to access process output while it's running under supervisor?
08:43 < GreenDog> there's a bunch of process.on() events
08:44 < GreenDog> let me ask, are you using socket.io or engine.io?
08:45 < Doomground> neither
08:46 < GreenDog> darn
08:46 < Doomground> why
08:46 < sinclair|work> ljharb: winning?
08:46 < GreenDog> because when i use socket.io or engine.io, my node.js app dies for apparently no reason what-so-ever
08:47 < sinclair|work> GreenDog: after how long ?
08:47 < GreenDog> it varies
08:47 < GreenDog> somewhere between 8 hours and 4 days :)
08:47 < sinclair|work> ive heard of this issue before
08:48  * sinclair|work really wants to like socket.io, but is fearful of random crashes
08:48 < GreenDog> maybe it was me mentioning it before
08:49 < GreenDog> i have socket.io 0.7 running on a low traffic site with no problems
08:49 < GreenDog> but the higher traffic site, kerplunk!
08:50 < GreenDog> maybe i should try 0.7 on the high traffic site
08:51 < NodeX> I've had similar issues on socket.io on a high traffic site
08:51 < NodeX> for no reason the CPU tops at 100% 
08:51 < sinclair|work> NodeX: how much traffic?
08:51 < GreenDog> i hadn't checked cpu
08:51 < GreenDog> but my issue is the app just stops
08:51 < sinclair|work> NodeX: memory leak, perhaps v8 is not collecting memory?
08:52 < GreenDog> no errors, no nothing
08:52 < GreenDog> memory was fine
08:52 < sinclair|work> i assume v8 has a garbage collector
08:52 < GreenDog> err, no out of memory error any way
08:52 < GreenDog> and my monitors showed ok memory
08:52 < GreenDog> and i even ran another process to use all the memory to see what would happen
08:53 < ljharb> sinclair|work: yeah, thanks for the tips. i'm making it work with strings. i just hadn't seen any examples using "data" and "end" like that.
08:53 < GreenDog> it didn't break socket
08:53 < sinclair|work> ljharb: wanna know how i knew what to do?
08:53 < ljharb> if you can tell me without being a dick, sure
08:54 < sinclair|work> ljharb: i read the docs
08:54 < ljharb> but i'm assuming you read the same docs i did, and the way they were written catered to you more than it did me.
08:54 < sinclair|work> ljharb: well, i also have the node.d.ts file
08:54 < ljharb> docs that aren't understandable by everyone aren't well written. :shrug:
08:55 < GreenDog> i'm always shocked at how well docs are written
08:55 < ljharb> but thank you, the docs, plus your example was enough to get me started
08:55 < GreenDog> like if you ever read any of my docs, forget it
08:55 < sinclair|work> ljharb: http://pastebin.com/RVna436j
08:56 < sinclair|work> ^ + docs == all you need you know
08:56 < ljharb> some of the best docs i've seen are mongoosejs.com and backbonejs.org
08:56 < sinclair|work> ljharb: and yes, the node docs can be misleading
09:00 < bakadesu> To answer my own question regarding davis.js:  https://github.com/olivernn/davis.js/issues/35
09:02 < GreenDog> mongoosejs does look like it has pretty nice docs
09:02 < GreenDog> i'm not thrilled with backbone's though
09:02 < sinclair|work> MSDN has some awesome docs
09:04 < sinclair|work> MSDN + stack overflow and you're set for whatever
09:04 < ljharb> unless you're working on a platform other than windows :-)
09:04 < ljharb> GreenDog: yeah, backbone's need some work. mongoose's are awesome tho.
09:05 < sinclair|work> ljharb: where are you hosting your apps?
09:05 < ljharb> mostly on my own servers, but i've used nodejitsu also
09:06 < sinclair|work> ljharb: how is nodejitsu, i pay for azure currently
09:06 < sinclair|work> ljharb: i pay for the node, .net and service bus back planes, but could replicate a lot of that with node and redis
09:07 < ljharb> so far it's great. their service is excellent.
09:08 < sinclair|work> ljharb: what does it offer besides node?
09:08 < GreenDog> owners that like node
09:08 < ljharb> autoscaling clusters, altho i haven't used that yet
09:08 < ljharb> but "nodejitsu" isn't for non-node things
09:08 < sinclair|work> ljharb: data storage?
09:09 < ljharb> nope, its just node hosting
09:09 < ljharb> AWS has data storage tho
09:09 < sinclair|work> what about mongo?
09:10 < sinclair|work> ljharb: its pricing isn't that much better than azures
09:10 < GreenDog> i run node on ec2
09:10 < GreenDog> and i've run it on a few other things as well
09:10 < ljharb> i use nodejitsu for mongo, but they get mongo from mongoHQ (which is free for me currently)
09:11 < sinclair|work> ljharb: what about native node modules?
09:11 < sinclair|work> ljharb: i really need to write some native stuff for node
09:11 < ljharb> if they have no other dependencies they'll definitely run. i'm sure it depends on what deps they have tho
09:12 < sinclair|work> ljharb: i am looking to have a go binding the .net clr to node
09:13 < sinclair|work> ljharb: also, it would be quite sweet to bind a GPU to node also
09:13 < ljharb> give it a shot, but i suspect that will be more than nodejitsu is designed to handle
09:13 < RLa> you guys still talk about .net bindings?
09:14 < sinclair|work> RLa: i am thinking about it
09:15 < sinclair|work> ljharb: at this stage, it appears to be more than azure is capable of :(
09:16 < sinclair|work> at least for its PAAS offerings
09:16 < ljharb> i will say tho that binding two different platforms together like that sounds crazy.
09:16 < ljharb> regardless of which platforms they are.
09:16 < ljharb> SOA means you should have a nice http api between them. preferably rest.
09:16 < sinclair|work> ljharb: and suffer network latency ?
09:17 < ljharb> a small price to pay for decoupled architecture.
09:17 < sinclair|work> ljharb: not everyone can afford thousands of compute servers
09:17 < ljharb> you don't even need 2. you can do http within the same machine.
09:17 < sinclair|work> ljharb: and suffer http overhead? why not just use a socket?
09:18 < ljharb> decoupling. if you've baked in the assumption that they're on the same machine, you're screwing yourself later.
09:18 < sinclair|work> ljharb: not for a windows enterprise
09:19 < ljharb> lol, true, but why use node if you're going to drag it back into that overly coupled mess?
09:19 < sinclair|work> ljharb: you just stumbled on nodes lack of image capibilities
09:19 < sinclair|work> ljharb: if i bind .net, i get GDI 
09:19 < ljharb> i had no trouble with image capabilities, i have imagemagick installed.
09:20 < sinclair|work> ljharb: its the same thing
09:20 < ljharb> you're worried about http overhead, but not about .net overhead?
09:20 < sinclair|work> ljharb: again, what's wrong with .net?
09:20 < ljharb> tying yourself to the windows OS is far more overhead than an inter-machine http request.
09:20 < ljharb> not everyone can afford thousands of .net licenses :-)
09:20 < sinclair|work> ljharb: .net is free
09:20 < ljharb> running it isn't.
09:21 < ljharb> because the OS isn't.
09:21 < sinclair|work> businesses are willing to pay for quality 
09:21 < ljharb> there's tons of other options for those willing to pay. with .net, your only option is to pay.
09:22 < sinclair|work> ljharb: established enterprises are likely to be using windows, active directory, sql server and other things
09:23 < sinclair|work> .net is capability of integrating with the above, binding node to .net means node can integrate with a windows enterprise too
09:23 < ljharb> indeed they are. "likely" has no correlation to "a good choice"
09:23 < ljharb> however, if instead of binding .net to node, you use an http interface, then *everyone* can integrate with it.
09:23 < sinclair|work> ljharb: and active directory and ldap?
09:24 < sinclair|work> ljharb: what about windows identities ?
09:24 < sinclair|work> exchange server
09:24 < ljharb> https://github.com/jeremycx/node-LDAP
09:24 < ljharb> and i understand, but that's why all those things are poor choices in the first place.
09:24 < ljharb> because those choices *force* other choices.
09:25 < sinclair|work> ljharb: they are not poor choices at all
09:25 < sinclair|work> ljharb: -1
09:25 < nathan7> sinclair|work: I thorougly disagree.
09:25 < ljharb> constraining choices are poor ones.
09:25 < nathan7> Yes.
09:25 < ljharb> no matter how good they otherwise are.
09:25 < sinclair|work> i think you have unrealistic expectations on how enterprises actually work
09:26 < nathan7> Unrealistically optimistic, perhaps.
09:26 < sinclair|work> its more than just a node module here and there
09:26 < ljharb> that doesn't mean i'm not right. it just means enterprises don't work well.
09:26 < ljharb> that's why google is a thing. because enterprises suck.
09:26 < sinclair|work> its the automation tools, the server management tools, the provisioning tools, the directory services, domain management
09:26 < ljharb> s/google/any successful disruptive company
09:27 < FrenkyNet> morning all, I found https://github.com/liamks/Delivery.js when looking for a file over scoket solution. It just says to be experimental… does anyone have any recommendations on what yo check out? Maybe another package?
09:27 < ljharb> that's why amazon does so well. because they built everything with SOA over http.
09:28  * nathan7 nods
09:28 < sinclair|work> ljharb: Microsoft spent a lot of time working out how a business enterprise should work, and it works really really well
09:28 < sinclair|work> (unless its been setup all wrong, which happens)
09:28 < ljharb> lol yeah microsoft hasn't been doing well the last 15 years
09:28 < sinclair|work> ljharb: in the enterprise ?
09:28 < ljharb> overall.
09:28 < nathan7> FrenkyNet: You mean plain HTTP?
09:29 < ljharb> indeed, enterprise software is one of their few money makers
09:29 < ljharb> but that's not because it's good. it's because their clients are locked in.
09:29 < sinclair|work> ljharb: they paid for a product
09:29 < ljharb> microsoft's customers can't switch even if they wanted to, because their choices are now forced.
09:29 < sinclair|work> they got a product
09:29 < sinclair|work> they could pay for node, and they would be locked into node also
09:29 < ljharb> unless they built a bunch of small, http-connected apps
09:29 < nathan7> Yes, node is a platform.
09:30 < ljharb> then it doesn't matter if some of them are node, some are .net, et cetera
09:30 < nathan7> You could build another platform that is compatible
09:30 < ljharb> that's why many small apps, connected over http, is always better.
09:30 < sinclair|work> ljharb: you don't even know what you are talking about
09:30 < sinclair|work> conversation over
09:30 < FrenkyNet> nathan7: nope, over sockets.io
09:30 < nathan7> sinclair|work: Well, that's an argument.
09:30 < ljharb> hahaha that's always how you end conversations
09:30 < nathan7> FrenkyNet: That makes no sense.
09:31 < ljharb> "i'm right and you're wrong so i don't want to talk about it anymore"
09:31 < FrenkyNet> nathan7: yes it does
09:31 < nathan7> ljharb: But it's an excellent argument! He has clearly demonstrated his ability to ignore your arguments!
09:31 < ljharb> ha
09:31 < sinclair|work> ljharb: i can't be bothered explaining things, you are stuck taking about http connected applications, who cares, thats not an enterprise
09:32 < sinclair|work> and has little bearing on anything
09:32 < ljharb> absolutely that's an enterprise. that's how AWS is built.
09:32 < ljharb> tons of enterprises run on it.
09:32 < nathan7> I'm building a largish app on everything-HTTP
09:32 < ljharb> "enterprise" and "microsoft" aren't synonymous.
09:33 < sinclair|work> ljharb: it is when im talking about integrating node with a windows enterprise
09:33 < sinclair|work> anyway, conversation over
09:34 < ljharb> what makes you think saying that actually ends a conversation, ever?
09:34 < nathan7> I.. I cannot fathom the stupidity of people :(
09:35 < ljharb> integrating with a windows enterprise is its own best punishment.
09:35 < guor> http://xkcd.com/386/
09:35 < sinclair|work> ljharb: why were you struggling with node streams ?
09:35 < ljharb> i'm not anymore. how is that relevant?
09:36 < freewil> stfu you assholes
09:36 < Zelest> obvious troll is obvious
09:37 < sinclair|work> http://channel9.msdn.com/Blogs/Charles/Introducing-RxJS-Reactive-Extensions-for-JavaScript
10:12 < gmg85> hi guys?
10:12 < gmg85> what would you say are the main features o f node.js?
10:13 < RLa> async i/o!
10:13 < kranius> it's javascript
10:13 < sinclair|work> gmg85: evented io
10:13 < NodeX> it uses language to achieve it's goal
10:15 < sinclair|work> gmg85: node is remarkably good and pushing data around, that's its main feature.
10:15 < NodeX> it also runs nicely on Linux and winblows and mac
10:15 < gmg85> thanks guys...
10:15 < gmg85> keep them coming though .. :)
10:18 < gmg85> can i use node.js for just any type of javascript application or are there bad use cases?
10:18 < deoxxa> node.js can vibrate through walls and burrow underground at incredible speeds
10:18 < guor> node.js doesn't have a dom, if that's what you are asking for
10:18 < sinclair|work> gmg85: nodejs wins at 95% of what a traditional programmer would do for a job
10:20 < AAA_awright> Node.js doesn't provide a DOM, not that you can't
10:20 < gmg85> i see
10:20 < AAA_awright> Or any data model for that matter, except streams
10:20 < gmg85> so i can use it anywhere i'd  use lets say GWT for example?
10:21 < gmg85> i meant use it for the back end
10:21 < radiodario> i use it to make apis
10:21 < sinclair|work> gmg85: at its core, node is basically JS on top of io interfaces (file, network), with a few baked in utilities, like a http server, encryption, path utils, a repl etc
10:22 < deoxxa> i use node.js to transport insects securely between treatment galleries
10:22 < sinclair|work> gmg85: outside of this, there are a multiple of JS written libs, that make up the majority of nodes functionality 
10:22 < sinclair|work> gmg85: thats about it really
10:23 < gmg85> sinclair|work, cool
10:24 < sinclair|work> gmg85: it is pretty cool, when you think about most applications written today basically shuffle data from disk to a network and back, node handles this incredibly well, the rest is implementation with JS
10:24 < gmg85> sinclair|work, lets say i had a web app that is javascript intensive and was using php for the back end..can i replace the backend with node.,js in all cases or are there times when it would not be suitable?
10:24 < sinclair|work> node works well as a web backend
10:24 < NodeX> anyone know of a jquery lib for node - one that will let me throw an html page at it and walk its dom tree?
10:25 < guor> NodeX, you would have need a DOM implementation for node.js first
10:25 < guor> -have
10:25 < deoxxa> NodeX: cheerio
10:25 < sinclair|work> gmg85: node uses a variety of html templating libs to do your pages
10:25 < NodeX> cheerio ?
10:26 < radiodario> NodeX: zombie.js
10:26 < sinclair|work> gmg85: i quite like swag, but there is jade and ejs for html templating
10:26 < NodeX>  finally removed coffeescript dependency thanks to @davidchambers [Matthew Mueller] 
10:26 < NodeX> oops
10:26 < radiodario> might be what you're looking for
10:26 < NodeX> https://github.com/MatthewMueller/cheerio
10:27 < NodeX> cool thanks, will look into that
10:28 < sinclair|work> gmg85: i think a lot of PHP developers migrated over to node, but i think if you had a background in python, node will be easier to learn
10:28 < NodeX> I use php and node on most of my sites
10:29 < NodeX> but I also know alot of javascript so that probably helps
10:29 < gmg85> sinclair|work, thanks
10:29 < sinclair|work> gmg85: np
10:30 < trankil> Marak === _Why - 15
10:30 < trankil> Substack === _Why - 5
10:30 < sinclair|work> gmg85: oh, if you have used PHP frameworks like slimPHP, then check out express()
10:31 < trankil> Ry === _Why - 2
10:31 < sinclair|work> gmg85: http://www.slimframework.com/
10:31 < trankil> Whos the next?
10:31 < sinclair|work> gmg85: express is very similar 
10:33 < Thomas`-> Hey everyone. In my application I have an array with integer code and string descriptions for error and success code and message combinations. Is it somehow possible for both server and client to use this array without having to maintain 2 seperate copies?
10:35 < sinclair|work> thomas: have the server send the client a single error ? i.e {code:123, message:'oh noes?' } 
10:35 < trankil> WHO IS THE NEXT ?
10:35 < NodeX> the one before the for
10:37 < sinclair|work> has anyone taken a look at rxjs and compared it to nodes async ?
10:37 < trankil> WHO IS THE NEXT ?
10:37 < RLa> what next
10:39 < sinclair|work> trankil: http://lmgtfy.com/?q=WHO+IS+THE+NEXT%3F
10:39 < gmg85> sinclair|work, cool
10:40 < Thomas`-> sinclair|work, so is it not possible to have a shared file?
10:41 < sinclair|work> thomas: i suppose you could, but im not sure why you would
10:42 < sinclair|work> thomas: not based on the example you gave at least, typically, the server knows about the errors, and sends them out, the client recv's the error, and handles it accordingly, 
10:42 < sinclair|work> thomas: that doesn't necessarily mean sharing the same errors and codes
10:43 < Thomas`-> sinclair|work, i gave a bad example. its for a reason todo with ajax and keeping consistant error and success code mechenisms across different platforms. e.g. data[Ajax.FAILURE_REASON]
10:43 < sinclair|work> thomas: rather, they understand the codes (agree on what they are, and how to interpret them)
10:44 < sinclair|work> thomas: why not introduce a thing called a network exception
10:44 < adammw111> HI all. I'm trying to write a node.js module to use the rtmp protocol but I'm coming a bit unhinged on my api design, I think partially due to my heavily OO design and stateful nature of the protocol. going back, should i be considering using Streams and if so, how do Streams that do processing work generally? Does anyone know of examples that do similar things?
10:44 < Thomas`-> like if node needs to talk to a php server or if javascript needs to talk php i want it all to use same conventions
10:45 < sinclair|work> thomas: and put all the information the client needs to know about the exception, in an exception object, if the client recv's an exception object, it throws the error ?
10:45 < sinclair|work> thomas: that doesn't warrent a sharing of code, rather it requires a protocol that all systems understand
10:46 < sinclair|work> thomas: for example, IE, Firefox, and Chrome all understand HTTP status codes, but none of them share any code
10:46 < sinclair|work> well, FF and chrome might, but you get the idea
10:46 < Thomas`-> yeah but i thought node and client js could somehow share same file
10:46 < Thomas`-> i get php needs to be seperate
10:47 < sinclair|work> well, it can only be shared over http
10:47 < Thomas`-> maybe symbolic link or smth?
10:47 < sinclair|work> no
10:47 < sinclair|work> thomas: create a file called 'errors.js'
10:48 < sinclair|work> have node load the file with fs.readFile()
10:48 < sinclair|work> have the browser load the file over http
10:48 < substack> Thomas`-: if you use browserify you can use many of the same node core libs
10:48 < sinclair|work> (which means you have to serve it through node)
10:48 < substack> including http.request()
10:49 < sinclair|work> substack: have you looked at meteor ?
10:49 < Thomas`-> sinclair|work, is there any reason why a link wouldnt work?
10:50 < substack> sinclair|work: I prefer to just use the same npm modules on the browser and in the server instead of the same code
10:50 < sinclair|work> substack: not a fan?
10:51 < sinclair|work> substack: i've only briefly looked at it, wasn't sure what to make of it
10:51 < substack> frameworks decide too many things up-front for my tastes
10:51 < substack> even express feels like too much
10:52 < substack> I like tiny libraries that do just 1 thing
10:52 < sinclair|work> substack: im not fond of express 
10:52 < sinclair|work> substack: but im not entirely against frameworks
10:53 < substack> frameworks make me angry and frustrated so I don't use them
10:53 < gmg85> "The truth is that while we are starting to see good frameworks for node.js, there is nothing as powerful as Rails, CakePHP or Django on the scene yet. If most of your app is simply rendering HTML based on some database, using node.js will not provide many tangible business benefits yet."
10:53 < gmg85> http://nodeguide.com/convincing_the_boss.html
10:54 < gmg85> what do you guys have to say about that?
10:54 < sinclair|work> gmg85: stop trying to justify node, and just use it. 
10:55 < sinclair|work> gmg85: i was all 'err' about node, until i actually started doing stuff with it, 
10:55 < substack> gmg85: don't use databases
10:55 < gmg85> sinclair|work, writing a document to try convince my boss
10:55 < gmg85> so am just doing all sorts of research into it
10:55 < sinclair|work> gmg85: replace the document with a socket.io app
10:55 < substack> just replicate state continuously instead
10:55 < gmg85> substack, like in RAM?
10:55 < substack> unimportant
10:56 < aandy> substack: what do you mean by that?
10:56 < gmg85> substack, what happens when the lights go off? :D
10:56 < substack> gmg85: you just replicate with the other nodes in the network when you come back online
10:57 < substack> and you use commutative operations so your states will merge cleanly and be eventually consistent
10:57 < substack> use these libs: https://github.com/dominictarr/scuttlebutt https://github.com/dominictarr/crdt
10:57 < adammw111> substack, just discovered node-binary, is that good for my use-case? I was using Buffers directly but seems like it lets you get some stream-based things going
10:57 < sinclair|work> substack: or use a queue ?
10:57 < Thomas`-> eval(require('fs').readFileSync('somefile.js', 'utf8'));  hm maybe this would be best?
10:58 < substack> dominictarr is working on a leveldb-based serialization layer to hook into scuttlebutt with .pipe()
10:58 < substack> and map/reduce
10:58 < Ophidian> hello i have a question about creating domains in a loop, some one can and want help me?
10:59 < substack> once those exist, you'll have all the features of a very capable database in tiny reusable pieces
10:59 < substack> and everything can be in-process
10:59 < sinclair|work> substack: rdbms isn't dead tho
10:59 < substack> so you can replicate whichever pieces of data you want to, in browsers or in node
10:59 < substack> sinclair|work: rdbms don't replicate well
11:00 < sinclair|work> substack: or "horizontally scale" well
11:00 < substack> I'm not interested in databases that don't have continuous peer-to-peer replication with eventually-consistent histories to handle network partitions
11:00 < sinclair|work> substack: but they do support transactions, and in some more capable systems, distributed transactions
11:00 < substack> and I'm certainly not interested in databases that don't let me subscribe to updates
11:01 < aandy> +1
11:01 < substack> the stuff in redis and couch feels really crude compared to crdt in that respect
11:01 < substack> and those are about the best you can get in an application-server style database right now
11:01 < substack> sinclair|work: eventual consistency has much more desirable properties than transactions
11:02 < substack> especially when you need a lot of nodes living on different networks and running on different platforms to cooperate
11:02 < sinclair|work> substack: consistency  is the 'c' in acid
11:02 < sinclair|work> substack: which transactions are based on acid
11:02 < sinclair|work> substack: also, you can get away with a lot using a queuing back plane
11:03 < substack> sinclair|work: if your operations are merely eventually consistent you get a lot of amazing features instead
11:03 < sinclair|work> substack: like data normalization, and advanced aggregates ?
11:03 < Ophidian> hello i have a question about creating domains in a loop, some one can and want help me?
11:03 < substack> like the ability to update your local copy of the state immediately and merge those changes cleanly by using commutative operations when you get back online
11:04 < trankil> I AM SPEECHLESS WHEN SUBSTACK IS WRITING ON THE INTERNETS
11:04 < daleharvey> dominic is just going to have to rewrite pouchdb to get any sane level of conflict resolution done
11:04 < daleharvey> so contribute to pouch yo :P
11:04 < substack> daleharvey: commutative operations let you not need to worry about conflicts
11:05 < sinclair|work> substack: they do the exact same thing as a serialized transaction
11:05 < substack> plus an eventually consistent history mechanism
11:05 < daleharvey> commutative operations means entire datasets have to replicate, which is about a worst case for browser / mobile sync
11:06 < daleharvey> s/entire datasets/entire deltas
11:06 < substack> daleharvey: you can parition them at the document level it seems
11:06 < substack> the next scuttlebutt version is going to have range queries
11:06 < substack> to get around this
11:06 < daleharvey> you can partition them by document, that doesnt really do anything, if you a constructing a document as a series of changes, you need to transfer all the changes
11:07 < substack> sinclair|work: in transactions you need to wait for a centralized authority node to make a decision about the updates
11:07 < sinclair|work> substack: a 'centralized authority node' 
11:07 < substack> or else some quorum of nodes which requires a lot of latency and communication
11:07 < sinclair|work> substack: seriously ?
11:07 < substack> daleharvey: only the changes up to a certain point
11:08 < adammw111> just found mtrude, seems to have done most of the hard work for me, thanks any way...
11:08 < sinclair|work> substack: you've been drinking the no-sql kool aid 
11:08 < substack> you can prune old changes with a lower monotonic timestamp that update the same values as changes that are newer which overwrite old values
11:08 < sinclair|work> substack: in english
11:09 < substack> sinclair|work: none of the "nosql" databases even let you do this stuff
11:09 < substack> maybe riak has elements of these ideas since it's based on dynamo but it's a big application, not a library that lets you do distributed replication
11:11 < substack> sinclair|work: I mostly just want more applications to have a /replicate url that can be used to glue the state of multiple servers and clients together
11:11 < sinclair|work> substack: i need to see a topology diagram to get what you are talking about
11:11 < substack> sinclair|work: https://github.com/substack/stream-handbook#scuttlebutt
11:11 < sinclair|work> substack: something about state, pushing writes out to shards, and having them quickly replicated
11:12 < substack> there's a network graph in that example
11:12 < sinclair|work> i see it
11:12 < substack> scuttlebutt just implements a gossip protocol to get eventually consistent updates
11:13 < sinclair|work> substack: do you cater for write contention ?
11:13 < RLa> how bandwidth efficient it is?
11:13 < substack> sinclair|work: it's just eventually consistent
11:14 < substack> using vector clocks and monotonic timestamps
11:14 < sinclair|work> so if a write fails, it tries again?
11:14 < sinclair|work> substack: why not queue?
11:14 < substack> sinclair|work: it's eventually consistent
11:14 < substack> you just write
11:14 < sinclair|work> substack: with a queue, you don't need any of that
11:14 < substack> and it updates your local copy immediately
11:15 < substack> like this: model.set(key, value)
11:15 < substack> and this: var value = model.get(key)
11:15 < substack> that's IT
11:15 < sinclair|work> substack: its an elegant interface
11:16 < substack> then if you want ordered sets with commutative operations, you can use crdt.Seq
11:17 < substack> so anyways this stuff is basically living in the future
11:18 < sinclair|work> substack: gotta ask, why the made up words ?
11:18 < substack> sinclair|work: this is the jargon used in the literature
11:19 < sinclair|work> can trim it down a bit tho
11:19 < sinclair|work> monotonic timestamp can be...timestamp
11:19 < substack> timestamp with the added property that it only ever increases
11:20 < substack> an important property if you're using timestamps to resolve conflicts
11:20 < sinclair|work> sure
11:21 < sinclair|work> anyway, its an interesting style you have thar mate
11:23 < sinclair|work> substack: ill have to spend some time looking at this project of yours, im not entirely sure i get the innards, or why you wouldn't have used other approaches, but it looks cool
11:23 < JohnMcLear> looks interesting
11:24 < substack> it's dominictarr's project
11:24 < substack> I'm just a heavy user
11:24 < JohnMcLear> That's how we all begin, it's a gateway drug
11:25 < sinclair|work> JohnMcLear: node has that effect on people
11:25  * sinclair|work sinks into node narcotics
11:26 < sinclair|work> substack: anyway, i have a hard time explaining away rdbms for the work i currently do
11:28 < sinclair|work> i think the rationals of object stores over normalized data doesn't ring as true to me as it would for someone who didn't have linq with all its data projection goodness
11:28 < Thomas`-> Is there no simple way to get a remote page such as the easy to use PHP function file_get_contents()?
11:28 < sinclair|work> substack: plus, the people i work with don't know anything else other than rdbms
11:29 < Thomas`-> actually nvm
11:29 < substack> Thomas`-: http://npmjs.org/package/request
11:30 < Thomas`-> thanks
11:30 < sinclair|work> substack: what do you work on btw?
11:31 < substack> sinclair|work: everything
11:31 < sinclair|work> what could possible warrant such a extreme key/value store?
11:31 < substack> extreme?
11:31 < Kakera> Thomas`-, I use superagent
11:32 < Kakera> it has much clearer documentation
11:32 < sinclair|work> substack: well, i figured you were doing something that was serving billions of clients
11:32 < substack> crdt/scuttlebutt just replicates and is partition tolerant and can do commutative eventually consistent merging
11:32 < substack> and it's easy
11:33 < Thomas`-> cool Kakera 
11:34 < sinclair|work> substack: i've not dealt a lot with nosql stores (outside of appengine), i figured you needed high / safe replication for reading and writing data from balanced shards
11:34 < sinclair|work> or something 
11:34 < sinclair|work> going off what ive been reading up on on say...mongo replication
11:34 < substack> sinclair|work: scuttlebutt just uses the gossip protocol technique documented in dynamo where it sends out updates randomly to neighbors
11:34 < substack> it's really a simple idea
11:35 < sinclair|work> substack: ah ok
11:35 < substack> you just pipe all the pieces together and it works
11:36 < sinclair|work> substack: so its a message bus
11:36 < sinclair|work> or sorts
11:36 < sinclair|work> *of
11:36 < substack> the eventually consistent history is a big part of the algorithm
11:38 < ins0mnia> substack: This might sound like a n00b question but what happens if a node goes down?
11:38 < ins0mnia> and then up again
11:39 < substack> ins0mnia: the histories merge together
11:39 < substack> in an eventually consistent way such that the histories will become the same in both places
11:39 < substack> even if updates happened on each side while the nodes were offline
11:40 < substack> or if a node has empty state it can replicate from its peers
11:40 < ins0mnia> cool
11:41 < ins0mnia> substack: Does this include some sort of heartbeat check to check if a node is down and streaming to it all together?
11:41 < ins0mnia> avoid streaming
11:43 < Thomas`-> How could I get the parent directory of the current working directory?
11:43 < Kakera> some time ago I submitted a pull request to node-protobuf to include the source code of protobuf to make it compatible with Windows and not rely on a system-wide protobuf install
11:43 < Kakera> it got accepted, but now node-protobuf takes forever to compile
11:44 < Kakera> as in, 1-2 minutes
11:44 < Kakera> perhaps there's a better way
11:46 < RLa> 1-2 minutes is hardly forever
11:46 < Kakera> are there other modules that take that long to install?
11:48 < Thomas`-> sorry disregard my question, got it working
11:48 < RLa> probably any native module that has significant amount of code
11:50 < trankil> RLa, kakera: all the substack modules take 5years to compile, in my head.
11:59 < sinclair|work> http://bartriemens.wordpress.com/2010/09/18/how-to-wrap-your-brain-around-rx-and-rxjs/ <-- for anyone curious
12:04 < RLa> haha
12:06 < trankil> sinclair|work: Please, don't post .NET links here anymore, thank you.
12:06 < sinclair|work> trankil: its javascript 
12:06 < RLa> it is
12:07 < sinclair|work> trankil: furthermore, there is a npm package for it
12:08 < sinclair|work> trankil: please don't ping me unless you have something valid to say, thank you.
12:18 < simong> Does anyone have a good tool/methodology to debug memory leaks?
12:19 < SomeoneWeird> simong, https://github.com/felixge/node-memory-leak-tutorial
12:20 < sinclair|work> simong: perhaps http://nodetime.com/
12:20 < trankil> omg
12:20 < trankil> some people are *really* stupid.
12:20 < trankil> sinclair|work: If all you found useful to say in this chan is "olook ihaz my npm module, its a .net port ololo im a nodejs guy ololl"
12:20 < trankil> what a shame.
12:20 < simong> thanks for the pointers, I'll read up a bit
12:21 < sinclair|work> trankil: grow up
12:21 < simong> One thing I've noticed though, my activity monitors tells me my node process uses ~250MB, while the heap in node-inspector is only about 18MB
12:24 < sinclair|work> simong: also, this looks interesting https://github.com/mape/node-profile
12:25 < Sonderblade> simong: linux?
12:25 < simong> os x
12:26 < simong> I've been using https://github.com/c4milo/node-webkit-agent
12:26 < simong> (so not node-inspector)
12:26 < simong> I don't get the profiles tab in node-inspector
12:26 < Sonderblade> simong: dunno about os x, but on linux it is very common that the os commits much more memory to a process than what it needs
12:26 < simong> Yea, but I see it going up when I'm executing my leaky code
12:27 < simong> (In this case uploading a file causes the memory usage to go up by the size of that file + a bit extra)
12:33 < sinclair|work> simong: is it limited to file uploads only?
12:34 < simong> Yes
12:34 < sinclair|work> are you storing a buffer of the uploaded file ?
12:34 < simong> I'm using a MemoryStream to keep the uploaded file in memory, but it looks like it's not being cleared
12:34 < sinclair|work> a memory stream as in a node package, or something you wrote?
12:35 < simong> It's a package, hang on
12:35 < simong> This one: https://github.com/JSBizon/node-memorystream.git
12:35 < SomeoneWeird> well, that's why
12:35 < SomeoneWeird> stream it to a tmp file
12:35 < SomeoneWeird> then delete it
12:35 < simong> I am deleting it, that's the weird thing
12:36 < simong> I'm plowing trough the code to see if anything else is referencing it, but I can't immediately find anything
12:36 < SomeoneWeird> from memory?
12:36 < sinclair|work> simong: you should just be able to stream it to disk
12:36 < SomeoneWeird> or "delete filestream;"
12:36 < simong> SomeoneWeird: I'm doing delete filestream;
12:37 < simong> sinclair|work: That's how we initially did it, but our app requires us to send uploads to S3
12:37 < sinclair|work> can you stream directly to s3?
12:37 < sinclair|work> using knox i suppose?
12:37 < SomeoneWeird> pretty sure you can
12:37 < simong> If you keep chunks of 5MB in memory you can
12:37 < simong> knox can't do it, but awssum can
12:38 < SomeoneWeird> and lol, sinclair|work can't see this so it must look weird
12:40 < sinclair|work> simong: are you sure knox can't stream for you?
12:40 < simong> last time I checked it couldn't
12:41 < simong> Well, perhaps if you're streaming from disk it can
12:41 < simong> not if the stream is in-memory and not fully read yet
12:41 < sinclair|work> simong: well, it looks like it has the WriteableStream implementation on the request it has
12:41 < sinclair|work> simong: req.pipe(knox.request) ?
12:42 < sinclair|work> or something
12:42 < sinclair|work> simong: im looking here.... https://github.com/LearnBoost/knox
12:42 < sinclair|work> simong: search for putStream on this page
12:43 < simong> Yes, but that requires you to know the filesize before you start streaming
12:43 < simong> which you can do when you want to upload a file on disk or an http response
12:43 < simong> not when you're dealing with file uploads
12:44 < sinclair|work> simong: hmmm,  perhaps streaming to disk first is reasonable option
12:44 < simong> I think it is, but I've been told to stream it directly :(
12:44 < sinclair|work> simong: there are not any http headers for the file upload
12:45 < simong> No, there is only a Content-Length header for the entire request
12:46 < sinclair|work> simong: subtracting the header length from the body length would yeild the body size 
12:46 < simong> Yea, that trick only works for a single upload though ;)
12:46 < sinclair|work> simong: multipart uploads...
12:46 < sinclair|work> well, ummm
12:47 < sinclair|work> simong: and you are certain than s3 requires a content length prior to uploading?
12:47 < simong> yes
12:47 < simong> Unless you use the streaming API
12:47  * sinclair|work notes its been a while since digging deep into aws
12:49 < sinclair|work> simong: the streaming api accepts 5mb chunks
12:49 < simong> Yes, that's what I'm doing right now
12:49 < simong> but I have a leak in there somewhere
12:49 < sinclair|work> but buffering tho
12:50 < sinclair|work> simong: one second
12:51 < sinclair|work> simong: it would be nice to have pipe support more options
12:52 < simong> how do you mean?
12:52 < sinclair|work> the ReadableStream has pause, resume, destroy on it, i wonder if its possible to pause the stream mid flight, flush the data, and resume
12:52 < simong> I *think* you can do that
12:53 < simong> So you might be able to throttly the upload
12:53 < simong> s/throttly/throttle
12:53 < sinclair|work> simong: well, pipe() is pretty much all or nothing, but there are events in there
12:53 < sinclair|work> ReadableStream is also of type NodeEventEmitter
12:54 < sinclair|work> simong: lets have a dig for an answer
12:54  * sinclair|work notes he was looking to do something similar
12:55 < sinclair|work> simong: is this helpful ? http://debuggable.com/posts/streaming-file-uploads-with-node-js:4ac094b2-b6c8-4a7f-bd07-28accbdd56cb
12:56 < simong> My work is based on the FelixGe's excellent formidable library
12:57 < sinclair|work> simong: i think its a bit rough node is stinging you in this way 
12:58 < sinclair|work> simong: you would hope it was capable of  half decent garbage collection, especially if you implicitly delete the obj
12:59 < simong> yea
13:00 < sinclair|work> simong: so, form on progress 
13:00 < sinclair|work> form.on('progress', function(bytesReceived, bytesExpected) { });
13:10 < sinclair|work> simong: the only way is to chunk the upload somehow, id be keen to see your solution
13:12 < sinclair|work> simong: i wonder what would happen if you PUT with a 5mb upload to s3, but only submit 1mb, and terminate the data somehow
13:13 < sinclair|work> then there wouldn't be an issue
13:13 < simong> I'm not sure if S3 allows that
13:14 < simong> It's probably a bit dense but here is the S3 code:
13:14 < simong> https://github.com/simong/Hilary/blob/filesStreaming/node_modules/oae-content/lib/backends/amazons3.js
13:14 < simong> Don't mind the log statements =)
13:14 < sinclair|work> simong: well, the only other option is to stream to disk, which is still ok 
13:14 < simong> https://github.com/simong/Hilary/blob/filesStreaming/node_modules/oae-util/lib/server.js - parseMultipart handles the form specific stuff
13:15 < simong> Yea, I think we'll probably have to go down that route and just make sure our disks are big enough
13:15 < simong> They were filling up our /tmp space when out under load
13:15 < simong> s/out/put
13:18 < sinclair|work> simong: can think of it as a queue of things to upload, then its not so bad :)
13:18 < sinclair|work> simong: pipe the file up there, once complete, unlink!
13:19 < simong> Yea, that's how our disks ran out of space :(
13:19 < simong> We're trying to run all our app servers on super small instances as node barely needs any memory
13:19 < sinclair|work> simong: how much upload traffic?
13:19 < simong> I'll need to check, but we're running automated performance tests and it fell over in wave 1 of 7
13:20 < simong> and each wave has progressively more requests
13:20 < sinclair|work> simong: if node would be good at anything, you think it would be this, tho the problem either is with memorystream or v8
13:20 < sinclair|work> simong: i would lean towards memory stream
13:21 < simong> Yea, same here
13:21 < monkeychops> hi :)
13:21 < sinclair|work> simong: i wonder if you can implement a dispose on it
13:21 < sinclair|work> or..
13:21 < sinclair|work> simong: emoryStream.prototype.destroy = function() {
13:21 < guor> run it with --trace-opt and --trace-deopt and see if you notice some correlation
13:21 < simong> .destroy is being called on it
13:22 < trankil> monkeychops: Please, change your pseudo.
13:22 < sinclair|work> monkeychops: don't listen to trankil 
13:22 < sinclair|work> simong: looking at destroy, its not doing a lot
13:22 < sinclair|work> this.queue = [];
13:23 < sinclair|work> this.end()
13:23 < simong> but that should at least dereference all the data that sat in the queue right?
13:23 < simong> or should it call: delete this.queue; this.queue = []; ?
13:23 < sinclair|work> simong: maybe, give it a try :)
13:24 < trankil> simong: v8 does the gc for you.
13:24 < monkeychops> does anyone know of any node community web site that has reviews and comparisons of different node ORMs, web frameworks, view engines etc?
13:24 < trankil> don't be that dumb.
13:24 < monkeychops> i am finding everything is kinda strewn over google
13:24 < monkeychops> :P
13:24 < simong> Yea that's what I thought trankil
13:25 < sinclair|work> trankil: leave troll
13:25 < trankil> monkeychops: npmjs.org
13:25 < guor> simong: do I understand correctly that you are noticing some unexpected performance drop?
13:26 < simong> My uploaded file remains in memory
13:26 < sinclair|work> guor: the v8 GC doesn't appear to collect
13:26 < guor> maybe it's being retained by something
13:26 < simong> but I'm fairly sure it's an error somewhere on my side
13:26 < simong> Can you recommend any good tooling?
13:26 < simong> node-inspector doesn't result in much tbh
13:27 < guor> well... what I do know is that in chrome you can make a heap snapshot and see what's retaining what
13:27 < guor> but I don't know anything similar for barebone v8
13:28 < monkeychops> trankil: I know npmjs.org of course, but thats just a huge list of modules, what I am looking for is a site that discusses and rates them etc, recommendations, best practice etc.
13:28 < monkeychops> could be it doesnt exist yet :P
13:28 -!- mode/#node.js [+o piscisaureus_] by ChanServ
13:28 < guor> what sort of file is it that you are uploading? it's not a js source file, right?
13:28 < sinclair|work> simong: did you try delete ?
13:29 < simong> 25MB apache log file :)
13:29 < simong> sinclair|work: yes
13:29 < sinclair|work> simong: no change?
13:29 < simong> well, delete memoryStream;
13:29 < sinclair|work> what abotu delete queue ?
13:29 < simong> no change
13:29 < sinclair|work> hmm
13:30 < simong> I'm using local storage now rather than amazon
13:30 < simong> same thing
13:30 < simong> code @ https://github.com/simong/Hilary/blob/filesStreaming/node_modules/oae-content/lib/backends/local.js starting from line 120
13:30 < trankil> monkeychops: what?
13:30 < trankil> monkeychops: the only best pratice we can give you is npmjs.org.
13:31 < trankil> monkeychops: and please, we don't want to repeat that: change your pseudo.
13:31 < monkeychops> what the heck is a pseudo?
13:32 < sinclair|work> simong: line 127
13:32 < sinclair|work> https://github.com/simong/Hilary/blob/filesStreaming/node_modules/oae-util/lib/server.js
13:32 < sinclair|work> req.streams[part.name] = {
13:32 < sinclair|work> 'stream': new MemoryStream(),
13:33 < simong> yes?
13:33 < sinclair|work> simong: without see much else of your code, its is possible that req is being references around the place ?
13:34 < trankil> monkeychops: whats your problem? you are disturbing us.
13:34 < simong> req obviously is
13:34 < sinclair|work> sure, thats holding a reference to an hash of memory streams
13:35 < monkeychops> ok... trankil you are clearly a troll so I am putting you on ignore :P
13:35 < simong> not really, that stream gets deleted so it's removed from the hash
13:35 < trankil> monkeychops: Please, logout and let we think about our stream memory leaks. Thank you.
13:36 < sinclair|work> simong: i can't see where you are deleting it
13:36 < sinclair|work> external from the module?
13:36 < simong> https://github.com/simong/Hilary/blob/filesStreaming/node_modules/oae-content/lib/backends/local.js
13:36 < simong> line 143
13:37 < simong> err
13:37 < simong> Yea 143
13:37 < simong> That delete opts.stream statement is unneeded
13:37 < sinclair|work> do you think it could be you are deleting here, as apposed to inside the module could be an issue?
13:39 < simong> Perhaps
13:39 < sinclair|work> simong: having the module dispose of itself would be tidier at least, i wonder if there is a bit of confusion going on with disposing from a different context/module/vm (at least i think modules are loaded as vms under the hood)
13:40 < sinclair|work> simong: aside from that mate, im out of ideas
13:40 < sinclair|work> :)
13:41 < trankil> sinclair|work: Ok, logout.
13:41 < trankil> simong: Thats where i can help you.
13:42 < sinclair|work> trankil: you are the epitome of a nodejs teenage stereotype, retarded
13:42  * monkeychops agrees with sinclair|work 
13:54 < stagas> is there a session middleware that doesn't suffer from race conditions?
13:56 < thomas`-> Is there any guidelines i should read before writing a protocol for use on a web application?
13:56 < stagas> meh, I should probably just never save anything important, and pray
13:57 < SomeoneWeird> thomas`-, define protocol
13:57 < thomas`-> stagas, "session middleware" - that sounds interesting, maybe its what i need - do you have any examples of this middleware implementations ?
13:57 < sinclair|work> thomas: you don't need session middleware
13:57 < thomas`-> SomeoneWeird, "The official procedure governing affairs of state or diplomatic occasions."
13:58 < thomas`-> oh ok
13:58 < SomeoneWeird> >.>
13:58 < Huzaa> Ni! I have a problem with printing text from json. I'm using node.js + express + ejs templating system. If I log the json, it doesn't have spaces in front of it but as soon as i use <%= postContent %> it creates a bundle of spaces infrotn the text causing the html tags to be shown as plain text
13:58 < Huzaa> any ideas?
13:58 < stagas> thomas`-: http://www.senchalabs.org/connect/middleware-session.html
13:59 < thomas`-> ahh its like PHP_SESSION
14:00 < thomas`-> ok, well im just going to read some tutorials on how people have made node.js servers for a while for some ideas regarding my protocol
14:02 < stagas> the problem is an earlier version might be saved after a newer one, which sucks. it needs a lock, or something
14:02 < fotoflo_> 有人在北京沫？
14:03 < khurram> hi guys
14:03 < khurram> i have went through a very good nodejs tutorial and feeling happy to successfully run in as well 
14:04 < khurram> just wanted to know how can i use the backbone.js with nodejs i.e. backbone , jquery on frontend and nodejs on backend as server
14:06 < radiodario> khurram: you can use node.js as the http server for your backbone.js static files, using something like express.js to serve a local directory (such as public)
14:06 < khurram> radiodario: thanks i install the npm install express 
14:06 < khurram> right ?
14:08 < SomeoneWeird> yes
14:09 < khurram> SomeoneWeird: i have just installed it 
14:09 < SomeoneWeird> http://expressjs.com/guide.html
14:10 < trankil> Please read this, everyone. http://blog.izs.me/post/30036893703/policy-on-trolling
14:10 < trankil> Trolls are not accepted here. This is not negotiable, Thank you.
14:12 < radiodario> khurram: ok now read this com
14:12 < radiodario> expressjs.com
14:12 < khurram> radiodario: thanks im on the page 
14:12 < killfill> hm.. deployd looks awavesome..
14:12 < radiodario> sweet
14:13 < trankil> sinclair|work, monkeychops: Please, read the link i posted before. This isnt an option.
14:14 < killfill> anyone uses it?..
14:15 < radiodario> killfill: i'm currently checking it out, looks pretty rad
14:16 < killfill> wondering if nyone is using it for a 'production' project...
14:16 < sinclair|work> trankil: you mean to say, i can ask a op to ban you from this channel ?
14:16 < sinclair|work> trankil: sounds like a win to me
14:17 < sinclair|work> trankil: "In the node community, we are not afraid of losing a few kids who can’t manage to behave in a grown-up manner. We’re more concerned with losing the people that those kids chase away."
14:18 < khurram> radiodario: how would i install express with backbone.js support  --backbone ?
14:18 < sinclair|work> trankil: im highly annoyed by your presence, and constant pinging me when im helping people elsewhere, 
14:18 < khurram>  express --sessions --css --[i want underscore template] --[backbone.js ] ?
14:19 < khurram> also jquery ?
14:20 < radiodario> khurram: i dont think you can install backbone.js on express
14:20 < radiodario> you might want something like yeoman.io
14:20 < khurram> radiodario: i want to use backbone + jquery for DOM events and other stuff where as i want node to serve as a server ?
14:21 < radiodario> i don't understand your question
14:22 < khurram>  i want to treat nodejs as apache  whose job is to serve  the page contents after rendering  so in a normal MVC i will have have view constructed from different files i.e.  header.js  [ have all the backbone, jquery etc ] body.js to have body and footer ?
14:23 < khurram> same way when the node receives the request it goes and concatenate all the html and then make one and send it accross am i right ?
14:24 < radiodario> if create a express.js public folder, you can put all your static mvc files in there
14:25 < khurram> radiodario: i installed the express but it did not created any folder called express in my directory ?
14:26 < radiodario> khurram: read the documentation for express please
14:26 < khurram> even i manually created express folder and issued the command npm installl -g express but still the files are copied to the previous node_modules directory ?
14:26  * khurram ahem ok 
14:28 < khurram> radiodario: thanks it works actually the express is installed inside the node_modules/express 
14:29 < khurram> radiodario: can i move this directory to somewhere else ? or now i have to work inside node_modules/express ?
14:29 < radiodario> ok you also need to read the manual for npm
14:30 < khurram> radiodario: npm again ?
14:30 < radiodario> if you pass the -g flag to npm it'll install it globaly (usually in /usr/local)
14:30 < radiodario> khurram: is what you used to install express.js
14:30 < khurram> ic
14:30 < radiodario> npm install express ? do you remember that
14:30 < radiodario> well npm is the package manager
14:30 < radiodario> read its manual
14:30 < khurram> yep
14:31 < khurram> manual yeah 
14:31 < ralphholzmann> substack: reading your stream handbook -- for writable streams destroy method, would it be a good idea to throw this.removeAllListeners() in that function?
14:32 < ralphholzmann> ^^^ or anyone else with an opinion
14:41 < Aikar_> .
14:48 < vmx> does anyone have an example using JSONStream that processes the json a bit and then outputs that to a file that doesn't leak memory, but works at a reasonable speed?
14:51 < EvilJordan> I am seeing some bizarre behavior with node-mysql and the debug message: "debug - client authorized for...". Basically, the first time a client connects, a db connection goes out and gets some stuff, but the debug log shows: "debug - client authorized for" which is incomplete, and communication between server/client doesn't work (though no errors are thrown.) If I then reload the client by refreshing the page, the db goes out, ge
14:51 < EvilJordan> debug - client authrorized for /NAMESPACE HERE
14:52 < EvilJordan> any ideas why it would only workon refresh once the client is already connected and somehow involves node-mysql?
14:52 < EvilJordan> i'm pretty stumped
14:53 < EvilJordan> It's like the trip to the DB server is interfering with socket.io's ability to establish it's connection
15:23 < tohoe> who moderates the mailing list ?
15:25 < jugatsu> hello guys
15:29 < eviljordan> Ok, more succinct question: one a client connection, i need to do a db lookup via node-mysql to make sure the namespace is authorized. this apparently takes too long and since it's asynchronous, causes problems. ideas? 
15:39 < pixie79> hi all - has anyone here used the node cloudfiles module? I am having issues working out how to get files to download from rackspace
15:43 < EyePulp> happy monday (assuming you're not 9 hours behind me or 15 hours ahead)!
15:53 < eviljordan> is there a way to read the namespace of a connected socket?
15:53 < MI6> joyent/node: isaacs v0.8.15-release * 5954a4a : v0.8.15 wip - http://git.io/KEnh7w
15:54 < bnoordhuis> eviljordan: define namespace?
15:54 < eviljordan> http://nodeserver/namespace
15:55 < bnoordhuis> eviljordan: is that a socket.io question?
15:55 < eviljordan> yes
15:55 < bnoordhuis> then i don't know. maybe try #socket.io
15:56 < eviljordan> thanks. they're silent.
15:56 < eviljordan> i'lljust sit around for a day or two
15:57 < bnoordhuis> someone's bound to stir eventually
16:00 < exit2> Any node book recommendations?  I'm going to take advantage of the oreilly 50% thing I think
16:01 < svnlto> exit2: http://nodejs.org/docs/latest/api/
16:01 < svnlto> free docs. there you go
16:01 < exit2> word
16:01 < EyePulp> exit2: I'd be surprised if there are any node books that are worth the investment, considering the fluidity of the platform (though maybe that's not such an issue in the last 6 months)
16:02 < wgolden> Good morning everyone :D
16:02 < jugatsu> guys, i wanna jump into world of node.js programming but i haven't any programming backgroud. i'm software and network engineer and wanna implement some network daemons like web-server serving POST requests or restful API. 
16:02 < jugatsu> what books do you suggest
16:02 < jugatsu> &
16:02 < jugatsu> ?
16:03 < exit2> jugatsu: I was just recommended to the node docs
16:03 < EyePulp> what does /25 after an IP represent?  I mean I know it's a block of IPs, but how do you calculate that?
16:04 < EyePulp> jugatsu: you have no programming background but you're a software engineer?
16:04 < jugatsu> yes, i am
16:04 < exit2> jugatsu: how do you have no programming background and you engineer software?
16:04 < bnoordhuis> EyePulp: /25 means 'mask of 25 high bits'
16:04 < jugatsu> network and infratsructer orchestration
16:05 < bnoordhuis> EyePulp: what remains is the subnet that the cidr applies to
16:05 < jugatsu> what's wrong?
16:05 < exit2> I wouldn't consider that a software engineer, but whatever :P
16:05 < bnoordhuis> EyePulp: with /25 that'd be a 128 addresses subnet
16:05 < exit2> jugatsu: what kind of stuff do you want to accomplish with Node?
16:05 < bnoordhuis> *mask off
16:06 < EyePulp> bnoordhuis: so a dotted quad ending in 128/25 means 128-254 ?
16:07 < bnoordhuis> EyePulp: yes, probably. convert the address to a 32 bits binary number and mask off the 25 high bits
16:07 < bnoordhuis> (i forgot to mention the conversion to binary :)
16:07 < EyePulp> jugatsu: I'm going to go out on a limb and say node might not be the easiest place to start learning to program.  Basic javascript might be a better jumping off point, and then node after you have a grasp of the fundamentals
16:08 < MI6> joyent/node: isaacs v0.8.15-release * 2adb516 : v0.8.15 wip - http://git.io/6fSwog
16:08 < bnoordhuis> jugatsu: python is a nice language to get started in
16:08 < _smf_> EyePulp: http://www.subnet-calculator.org/cidr.php
16:08 < jugatsu> exit2: dynamic user directory using mod_xml_curl http://wiki.freeswitch.org/wiki/Xml_curl, just serve POST request and make lookup @ sql/nosql and then proceed with xml file
16:08 < EyePulp> I would have suggested python as well if I hadn't feared the sacrilege… =)
16:08 < jugatsu> i like node )
16:09 < EyePulp> jugatsu: what do you like about node?
16:10  * EyePulp doesn't like to see people get excited about shiny new toys and then get mad about them because they're harder/poorly suited to the task you give them
16:10 < jugatsu> lightweight, event loop, in my usecase, apache with cgi would be a little overhead
16:13 < jugatsu> guys, so node .js is not better way to start programming
16:14 < EyePulp> jugatsu: your desire to build "network daemons" (REST or POST based) is not a skill that makes node a better/worse fit than literally a dozen or more other languages and platforms.  And lightweight is… pretty vague.   I think what I'm saying is that node doesn't make life any easier to start learning from scratch - other languages and platforms do...
16:15 < EyePulp> Node is great, but it's hard to appreciate that or enjoy it much with no base of reference from other solutions.  And in fact it can be harder to figure things out precisely because of the event loop and callback based programming.  Just my two cents. =)
16:16 < jugatsu> so my next turn is twisted :)
16:16 < EyePulp> jugatsu: good lord no.  
16:16 < jugatsu> why? )
16:16 < EyePulp> jugatsu: You need to learn basic programming in a simple environment.  Then start adding ridiculousness.
16:17 < jugatsu> EyePulp: thanks for suggestions! what book do you recommend?
16:18 < EyePulp> Otherwise you'll have no idea of how things are happening, and reading code will be meaningless.  you won't know what's the core language vs what's an added module/library/feature  -- which makes debugging kind of hard.
16:19 < EyePulp> jugatsu: I'd consider python as a great starting point, and for a book, I guess it depends on how new this all it.
16:19 < EyePulp> *all is
16:21 < EyePulp> Learn Python The Hard Way  is a very basic starting point, but very thorough.  If you do the work in it, you'll have a great foundation in the language.  It may be too simple for you, but that just means you can go through it faster.
16:21 < jugatsu> i have 1.2G of books about python :)
16:21 < exit2> jugatsu: I agree with the Python comment, there are a ton of online tutorials for free
16:21 < exit2> Yeah Learn Python The Hard Way is great
16:21 < jugatsu> but never read em
16:21 < EyePulp> http://learnpythonthehardway.org/book/
16:22 < EyePulp> free html, or a few buck for the PDF
16:22 < jugatsu> ok, thanks guys )
16:23 < EyePulp> jugatsu: Here's the key:  If you want to learn a skill, you'll have to work for it - not google the answer, or cut and paste it.  Do the work.
16:24  * EyePulp gets all serious and furrows his brow.  Music swells.  End scene.
16:24 < jrajav> Actually, more "skilled" programmers basically just know better Google queries
16:24 < jrajav> And they might cache some of the responses
16:24 < jrajav> :P
16:25 < jrajav> We're essentially very expensive database lookup machines
16:25 < EyePulp> jrajav: That's a different skill - and a nice one to have, but it doesn't make one a very good programmer.
16:27 < jrajav> I'm not sure I even want to go down this road again, but...... 1) What's "good" in this context? 2) Why not?
16:27 < jrajav> Those may or may not be the same question, depending on your perspective.
16:29 < EyePulp> jrajav: Finding stuff (via google we'll say) doesn't necessarily help you implement it in the context of your own code.  Finding is one skill, implementing is another, and quality of implementation a degree of that latter skill.  I'd say a "good" programmer should have both the discovery and implementation skills in high amounts.
16:30 < jrajav> But all of those implementation details can be recursively googled
16:30 < jrajav> All the way down to "what is a variable" and "how do I add two numbers in Javascript"
16:30 < jrajav> See where I'm going with this?
16:32 < trankil> jrajav: The 'add' operation is async in node. Its not like java or php.
16:32 < EyePulp> jrajav: I see, but irrespective of the massive inefficiency of having no baseline programming knowledge to draw on to form your search queries, you assume a level of caching on the part of the searcher, in order for them to do successive queries to answer their more complex questions, no?
16:32 < jrajav> Anyway, all you're doing is redefining "good programmer" with the same terms: "quality", "skill"
16:32 < shanse> add is part of jquery, afaik
16:32 < jrajav> That still doesn't get at what "good" is
16:33 < jrajav> EyePulp: Right, so we can define a programmer's efficiency based on the efficiency and breadth of their caching
16:33 < jrajav> EyePulp: That still doesn't necessarily define a good programmer (or does it?)
16:33 < jrajav> trankil: Was that meant for someone else?
16:34 < danmactough> Is it possible to have an Object with no constructor? I'm trying to debug a TypeError, and this is stumping me...
16:34 < _smf_> jrajav: I'd say some analytical skill is also required when things don't go quite right.
16:34 < jrajav> _smf_: Yes, good point.
16:34 < jrajav> When they do, too.
16:35 < _smf_> *nod*
16:35 < EyePulp> jrajav: _smf_ makes a decent point.
16:35 < trankil> jrajav: No. You should just know that everything in node is async, even the '+' operator and the while loop.
16:35 < jrajav> trankil: ...
16:35 < EyePulp> It also requires that anything you're attempting has been written up in terms you understand/expect in a search engine
16:36 < jrajav> I still don't think any of this approaches what "good" programmers or "good" code is.
16:36 < jrajav> We generally regard "good code" to be something approximating "clear, maintainable code"
16:36 < jrajav> Skill in googling/remembering and analytical both don't seem to touch on that
16:37 < trankil> If you write a node app, then your code is good.
16:37 < TheEmpath> each line of code is put into the htread pool
16:37 < jrajav> You can have the greatest analytical mind in the world, and a very technically skilled coder, and still produce what most would call bad code.
16:38 < EyePulp> jrajav: I suppose a good programmer is someone who can solve the problem and understand their solution and produce it in such a way that I (their boss) am satisfied with the results and the methods chosen to arrive at the results.  Relative, for sure, but hey, I sign the paychecks. =)
16:38 < nathan7> That's pretty much CS students
16:38 < jrajav> EyePulp: That's better than any definition up to this point, but it still doesn't answer what it is that you would be satisfied with
16:39 < EyePulp> jrajav: If they brought me decent coffee in the morning it would go a long way...
16:39 < jrajav> good coding = coffee
16:39 < jrajav> I'm willing to stop there
16:40 < EyePulp> ==
16:40 < TheEmpath> can you debug patiently and can you reconfigure the entropy of a project?  that's about all i care about when looking for programmers
16:40 < EyePulp> I can reconfigure the entropy - just not in the direction you're probably looking for
16:40 < jrajav> What exactly does "reconfigure the entropy" mean?
16:40 < jrajav> You want different entropy?
16:40  * EyePulp adds to it
16:40 < Shadow_S> i always find it funny when people just give up on debugging something when they haven't figured it out in a set period of time
16:41 < EyePulp> I just changed your DB PK schema to allow dupes.  Deal with that entropy.
16:41 < bnoordhuis> Shadow_S: apropos nothing, that's the prevailing mindset in the haskell world: don't debug, rewrite
16:42 < trankil> TheEmpath: you should rename your username, and put "TheEntropy"
16:42 < TheEmpath> lol
16:42 < EyePulp> bnoordhuis: nice
16:42 < jrajav> You let ordinary coders near your DB schemas?
16:42 < jrajav> You deserve duped PKs
16:42 < Shadow_S> bnoordhuis: is that mainly because it's hard to figure out what exactly is happening behind the scenes?
16:42 < MI6> joyent/node: isaacs v0.8.15-release * fdf91af : 2012.11.26, Version 0.8.15 (Stable)  * npm: Upgrade to 1.1.66 (isaacs)  * - http://git.io/-21htg
16:43 < jrajav> Oh I forgot, it doesn't matter because we all use NoSQL now
16:43 < EyePulp> jrajav: oh, okay, so DBAs are some rarified skilled beast, whereas programmers are mindless google monkeys.  Color me intrigued.
16:43 < trankil> So everybody is saying that Haskell prg >>>> Node prgramers???
16:43 < bnoordhuis> Shadow_S: yes, quite possibly
16:44 < jrajav> EyePulp: A huge portion of programming disciplines, programming businesses, and programming languages is geared specifically to mitigate the damages of mediocre programmers.
16:44 < EyePulp> "hipsterslapfight" is possibly my favorite nick ever.  Except maybe nick cave.
16:44 < TheEmpath> jrajav: http://en.wikipedia.org/wiki/Configuration_entropy
16:45 < TheEmpath> build a system that anticipates that (monadically, obviously) and you can adapt much easier
16:45 < jrajav> TheEmpath: This doesn't make "reconfigure the entropy" make much more sense, but I see what you're saying
16:46 < TheEmpath> No one codes monadically.  They code to spec.
16:46 < TheEmpath> So, you'll always have to gut their personal bias to make it more functional.
16:46 < trankil> jrajav: TheEmpath is a physicist, not you.
16:46 < trankil> jrajav: you are just a little programmer
16:47 < TheEmpath> shoo, peon
16:47 < TheEmpath> and fetch me a neutrino
16:47 < EyePulp> I read that as maniacally, which I do cackle and drum my fingertips together during coding sessions, so… yeah
16:47 < jrajav> Physicists get a free card for incomprehensible grammar?
16:47 < TheEmpath> no that's economists
16:47 < TheEmpath> they get paid better premiums via obfuscation.  it is the opposite of programming
16:47 < bnoordhuis> and linguists
16:48 < EyePulp> are they cunning?
16:48 < TheEmpath> very
16:49 < tohoe> who moderates the node.js mailing list ?
16:50 < TheEmpath> the middleware approach for node.js helps achieve maximum configuration entropy
16:51 < bnoordhuis> tohoe: i do
16:51 < TheEmpath> ideal if you can reduce your results to statistical values.
16:52 < trankil> tohoe: I will help you. Its bnoordhuis.
16:52 < EyePulp> anyone heard the rumor about bnoordhuis maintaining the node mailing list?  vicious stuff.
16:53 < bnoordhuis> let's not forget he also moderates #node.js and is very ban hammer-happy
16:53 < zot> is there something like 'atexit' for node?  or do i have to react on 'exit', signals and exceptions independently?
16:53 < bnoordhuis> zot: process.on('exit', cb)
16:54 < bnoordhuis> zot: you can catch (some, not all) signals with process('SIGNAME', cb)
16:54 < bnoordhuis> zot: err, process.on('SIGNAME', cb)
16:54 < trankil> bnoordhuis: Hi Ben.
16:54 < trankil> bnoordhuis: you are my idol.
16:54 < TheEmpath> that is all
16:54 < bnoordhuis> zot: there's process.on('uncaughtException', cb) but you should probably not use that
16:54 < bnoordhuis> trankil: your good taste is commendable
16:55 < TheEmpath> Also, the ECB, NYFR, and the BOE are all being ran by former employees of Goldman Sachs.  BIS is next!  See ya!
16:55 < trankil> bnoordhuis: Sincerely, but after _why
16:55 < trankil> and substack
16:55 < trankil> and ry.
16:55 < bnoordhuis> trankil: ry? i taught him everything he knows
16:55 < bnoordhuis> trankil: mind you, not everything *i* know
16:56 < bnoordhuis> zot: i'll refer you to the docs about why uncaughtException is, in general, a bad idea
16:58 < zot> bnoordhuis: that doesn't get callback if you simply ctrl-c the program (at least on my mac)
16:58 < zot> that was the first thing i tried
16:58 < zot> only when explicitly calling process.exit(), for me
16:58 < bnoordhuis> zot: you need to catch SIGINT
16:58 < zot> but that was my question.  is there something like atexit for best effort cleanups ;)
16:59 < bnoordhuis> zot: not a single function/event, no
16:59 < zot> kk, dankje
16:59 < bnoordhuis> graag gedaan :)
17:05 < trankil> bnoordhuis: Please, don't be so pretentious.
17:06 < trankil> so you did like this bnoord.pipe(ry)
17:06 < trankil> but you exit before the normal end()
17:07 < trankil> bnoordhuis: i'm shocked. And i remove you from my list, thats all.
17:09 <@isaacs> tohoe: most of the committers are mailing list mods, and a few other people.
17:10 <@isaacs> tohoe: the mailing list has a bunch of moderators, actually.  but bnoordhuis is probably the most active.
17:10 < devdazed> hmm, for some reason i cant install v0.8.14 via nvm
17:11 < devdazed> jsut says it fialed
17:11 <@isaacs> devdazed: try 0.8.15
17:12 <@isaacs> (in the process of finishing the release, about to send email etc.)
17:12 < devdazed> @isaacs: thanks, looks like it is installing now
17:13 < tohoe> isaacs: thanks for the info 
17:15 < MI6> joyent/node: isaacs created tag v0.8.15 - http://git.io/hOAwiw
17:15 < MI6> joyent/node: isaacs v0.8 * 9f51fd6 : Merge branch 'v0.8.15-release' into v0.8 (+1 more commits) - http://git.io/9QssRA
17:16 < devdazed> @isaacs: http://pastie.org/5438221
17:17 < devdazed> OSX Mountain Lion, latest fwiw
17:20 < MI6> joyent/node: isaacs v0.8 * e773be6 : Now working on 0.8.16 - http://git.io/HIsnKA
17:22 < MI6> joyent/node: isaacs v0.8 * 122ac4e : blog: Post for v0.8.15 - http://git.io/vkIpog
17:41 < st_luke> anyone in NYC looking for a node job?
17:42 <@isaacs> devdazed: that's a bug in nvm
17:42 <@isaacs> devdazed: it doesn't have to install npm.  just doing `make install` with node will do that.
17:43 <@isaacs> devdazed: if it's going to curl|sh npm, it needs to either use curl -L or use the https url
17:43 <@isaacs> devdazed: you should just use nave instead.
17:43 <@isaacs> it's faster :)
17:43 < devdazed> thanks @isaacs 
17:59 < pspeter3> Are there any good examples of a node.js module wrapping the functionality of a unix command line tool? I want to create a wrapper for arp-scan
18:04 < s5fs> pspeter3: yeah, there are a few, here's a ninjablocks example https://github.com/ninjablocks/wifi
18:04 <@mbalho> pspeter3: heres a pretty straightforward one i wrote https://github.com/maxogden/node-mdb
18:05 < zkirill> Has anyone here ever used SQLite with Node.js and can share their experiences?
18:05 <@mbalho> pspeter3: key parts of mine a) it uses and exposes streams and b) it uses the concat-stream module
18:06 < s5fs> yeah, maybe so, but it doesn't have any mention of "ninjas" that i can see ;-)
18:06 < pspeter3> zkirill: The development seed one is the best
18:07 < zkirill> pspeter3: wow, thanks so much, for some reason I couldn't find this after looking for sqlite + node.js all morning
18:09 < pspeter3> zkirill: Yeah, it's strangely low down but it works great. db-migrate is also an awesome tool for database migrations. It actually stores the migrations in the database as another table so you don't need to worry about some hidden state file
18:09 < zkirill> pspeter3: thank you kindly
18:18 < trankil> I miss Marak :(
18:22 < st_luke> :(
18:24 < trankil> st_luke: dont be sad anyway, im sure ill be back soon :)
18:24 < trankil> before 2013 \o/
18:28 < sosnon> besides express and greedy are there any other web frameworks in the same vein that i should consider using for fast application development?
18:28 < sosnon> *geddy, not 'greedy'
18:28 < trankil> sosnon: don't use flatiron.
18:29 < trankil> sosnon: marak is not working on it anymore :(
18:29 < sosnon> oh, i ttoally left out flatiron
18:30 < st_luke> sosnon: there's tower
18:30 < st_luke> lance pollard does some awesome work on it
18:31 < sosnon> well, tower is express-based
18:33 < trankil> sosnon: your question is pretty good.
18:33 < trankil> So be happy, i will help you.
18:33 < eydaimon> sosnon: it's express-based, but may help you get up and going faster. also seems to have nice conventions
18:33 < trankil> The full list here :) https://github.com/joyent/node/wiki/modules#wiki-web-frameworks-full
18:34 < sosnon> trankil: i don't get it; is this irony in your words or actually meant like that?
18:34 < Blkt> good evening everyone
18:34 < sosnon> eydaimon: sure, i'm very thankful that you guys mentioned it, i did forget it, just like flatiron
18:35 < mmalecki> trankil: whole nodejitsu team is maintaining flatiron, there's no reason not to use it :)
18:35 < trankil> sosnon: no, im not ironic all the time :)
18:35 < eydaimon> that's a lot of frameworks
18:35 < deoxxa> frameworks bad! libraries good!
18:35 < sosnon> maybe i should extend my question and just ask: what is a good way to get up and running with a node app MVC-app; my answer would have been: use a web framework (or write one yoursel); so i just asked for web frameworks
18:37 < sosnon> i like the execute-command-and-get-a-scaffold approach of express and geddy, though; also having an ORM is pretty useful
18:37 < deoxxa> both of those things are bad
18:38 < sosnon> deoxxa: ok, i'm open to suggestions; why?
18:38 < deoxxa> i don't see the point of generating files in a language as wildly dynamic as javascript
18:38 < deoxxa> if you want to generate a default controller, why not just make that default behaviour?
18:38 < deoxxa> stuff like that
18:38 < sosnon> i see
18:39 < deoxxa> also ORM is bad for the same reasons it's been bad forever
18:39 < deoxxa> ...for
18:39 < sosnon> sure, if you provide it to me in form of a library, i will use that instead, sure; still i'm used to the scaffold coming from django
18:39 < deoxxa> but what does that "scaffolding" contain?
18:40 < deoxxa> default configuration, default templates, etc
18:40 < deoxxa> why not just make that your default behaviour when no other behaviour is specified
18:40 < deoxxa> scaffolding generators pretty much always end up becoming really really complex and generating ream after ream of code after a while
18:41 < deoxxa> the first generation of a scaffolding generator might be really neat and easy to use
18:41 < deoxxa> then the next one they add some examples about how to use x
18:41 < deoxxa> then some examples of how to use y
18:41 < sosnon> sure, you could do that; glimpsing at geddy, however, scaffolds are very slick when implementing your model manually; so it boils down to file names with boilerplate
18:41 < trankil> Im pretty ok with deoxxa.
18:41 < trankil> on this scaffold point.
18:42 < sosnon> makes sense, true
18:42 < deoxxa> basically the scaffolding becomes a replacement for documentation
18:42 < deoxxa> "how do i set up x to work like y"
18:42 < deoxxa> "run the scaffolding generator"
18:42 < deoxxa> just look at rails these days
18:43 < deoxxa> i doubt there's anyone around who can manually set up a rails project anymore
18:43 < trankil> dhh
18:43 < nathan7> deoxxa: I did the rails boilerplate thing
18:43 < sosnon> anyway; i find it a lot more comfortable then dispatching/routing requests manually, popping in a templating package and connecting to mongo/redis/... manually if i 'just' need an app running in the browser
18:44 < nathan7>  ~/rails-boiler ⮀ find -type f | wc -l
18:44 < nathan7> 50
18:44 < nathan7>  ~/rails-boiler ⮀ find -type f -exec cat {} \; | wc -l
18:44 < nathan7> 1231
18:44 < nathan7> deoxxa: Isn't it grand?
18:44 < deoxxa> sosnon: definitely you should use something that provides you with a router and such, but beware anything that tries to sell you on being an all-in-one package
18:44 < deoxxa> nathan7: lol, 1231 lines of code
18:44 < deoxxa> that's crazy
18:44 < nathan7> deoxxa: This is just a completely empty Rails app, generated by 'rails' utility.
18:45 < nathan7> deoxxa: rails generate $appname
18:45 < sosnon> deoxxa: still i doubt there is anyone who could handle forgery/routing/request handling/... manually if not by using rails or django or ... not in the same amount of time
18:46 < deoxxa> sosnon: i don't want to try to push my opinions of what to use onto you, but i use express and all that stuff is really easy to set up manually
18:46 < sosnon> deoxxa: ok, what would you suggest besides things like express/geddy/flatiron/tower/... ?
18:46 < deoxxa> sosnon: usually <50 lines of code to get started
18:46 < deoxxa> besides them?
18:46 < deoxxa> how do you mean?
18:47 < sosnon> ok, express is solid, though very lightwight... but usable, that's true
18:47 < Kakera> Steam allows you to send private messages and chat messages. Should I have two separate events for each message type, or one for both and let the user figure it out from the 'source' argument if he wants to?
18:47 < Marak> who wants a job? remote telecommute, 30+ hours a week. all open-source work
18:47 < sosnon> 'besides them' as in 'just let me know if i'm doing it all wrong if i just use express/geddy'
18:48 < deoxxa> sosnon: well again, this is just me, but i basically exclusively use express for web stuff
18:48 < sosnon> mandric: no, but what does the job include, i'm curious anyway
18:48 < Kakera> or do what node-irc did, have one event to catch-'em-all and separate events for different message types
18:48 < trankil> :o
18:48 < Kakera> 'message', 'chatmsg', 'friendmsg'
18:49 < trankil> Marak \o/
18:49 < trankil> oh noes, marak is Jimbastard :(
18:53 < sosnon> trankil: Jimbastard? i though Marak was https://github.com/Marak
18:53 < trankil> on irc
18:54 < sosnon> ?
18:54 < sosnon> never mind, i'm unfamiliar with #node.js internals
18:55 < kenperkins> why wouldn't you want a load balancer to use Keep-alives between the LB and some backend servers?
19:03 < nicolaas> Is it possible to specify a git dependency in package.json that will cause the dependency to be checked as a git repo, so I can commit back to it when making changes in its folder ?
19:04 < nicolaas> This is for a repo that I control, and that's used in a different project I'm working on
19:04 <@isaacs> deoxxa: ++
19:04 <@isaacs> deoxxa: re "frameworks bad! libraries good!"
19:05 < deoxxa> and here i was thinking you just liked me as a person
19:05  * deoxxa cries
19:06 <@isaacs> deoxxa: well, NOW i do ;)
19:06 < deoxxa> yay!
19:06 <@isaacs> deoxxa: i se the `npm init` function as something of a failure of npm
19:06 <@isaacs> deoxxa: otoh, "convention over configuration" is also fraught with difficulty
19:07 < deoxxa> this is true
19:07 < deoxxa> see i don't mind `npm init' - it's not like it does a whole lot
19:07 < sosnon> >> this == true
19:07 < purr> sosnon: (boolean) false
19:07 < sosnon> ;)
19:07 < deoxxa> maybe if it went and created a readme etc, then it'd be going too far
19:08 < kenperkins> isaacs: ++ for pulling out the old napster metallica reference
19:08 <@isaacs> kenperkins: ?
19:08 <@mbalho> perhaps said reference was a reference
19:09 < sosnon> deoxxa: someone will eventually ask 'How much is too much? And who decides?' if i look at the complexity of the djangoverse then i have to say it does not really generate too much (for my taste)
19:09 < kenperkins> [11:03:00] <@isaacs> deoxxa: re "frameworks bad! libraries good!"
19:09 < sosnon> ... and deleting is always an option, too ;)
19:09 <@isaacs> mbalho: i think it's not a reference so much as a null pointer.
19:09 < kenperkins> isaacs: obligatory https://www.google.com/search?q=napster+bad+metallica+good&aq=1&oq=napster+bad+met&aqs=chrome.2.57j0l3.4794&sugexp=chrome,mod=14&sourceid=chrome&ie=UTF-8
19:09 <@isaacs> mbalho: the minute someone tries to figure it out, they'll segfault
19:09 <@mbalho> hah
19:09 < trankil> huhu
19:09 < trankil> npm init doesnt generate code. thats not a problem.
19:10 < deoxxa> i was actually thinking animal farm...
19:10 < trankil> Im gone.
19:10 < kenperkins> :(
19:10 < kenperkins> maybe I busted up the reference then :P
19:10 < sosnon> 'I'm gone.', isn't that a pradox?
19:11 <@isaacs> kenperkins: oh, i think "$thing GOOD. $otherthing BAD." predates metallica by a bit
19:11 < trankil> sosnon: i'm gone in the previousTick of my life.
19:11 <@isaacs> trankil: no, it does generate config though, which is just code in another format.
19:12 <@isaacs> trankil: in an ideal world, it'd be unnecessary
19:12 < deoxxa> what's up with single-file modules, btw
19:12 < deoxxa> are they a thing yet
19:12 < sosnon> trankil: scary, i hope there is a currentTick.. and again, you're still here, even more paradoxic
19:13 < sosnon> traph: actually, you were never here sitting next to me, so... wait... world's falling apart!
19:13 <@isaacs> oh, god, the lack of context with @horse_js is always so awesome.
19:13 <@isaacs> totally changes meaning
19:13 < nicolaas> can anyone help with this?
19:13 < nicolaas> Is it possible to specify a git dependency in package.json that will cause the dependency to be checked as a git repo, so I can commit back to it when making changes in its folder ?
19:13 <@mbalho> yea excellently curated
19:13 < sosnon> here = this; here.isWorld = true;
19:15 < deoxxa> isaacs: single file modules - fact or fiction?
19:16 < sosnon> every dependency adds a file, technically speaking, doesn't it?
19:25 < Mr_Grim> hey hey
19:25 <@isaacs> deoxxa: it works
19:26 < deoxxa> how2do
19:26 <@isaacs> deoxxa: add /**package {jsons here} **/ in your file
19:26 < deoxxa> niiiiiice
19:26 <@isaacs> deoxxa: and then you can npm install just that thingie
19:26 < deoxxa> that's win
19:26 <@isaacs> deoxxa: if that file is named "index.js", then it'll work instead of a package.json file in a folder
19:26 <@isaacs> deoxxa: but you can also do stuff like `npm install http://raw.gist.github.com/blahblahblah/foo.js` if it has that comment in it
19:27 < deoxxa> yep
19:27 < deoxxa> that's what i'm after
19:27 <@isaacs> deoxxa: when it installs, i think it'll still put a package.json there
19:27 < deoxxa> ah
19:27 < deoxxa> that should be ok
19:27 <@isaacs> deoxxa: but the logic now detects a file, a tarball, a gziped tarball, etc.
19:27 < deoxxa> neat
19:27 < sosnon> what is the introductory material/tutorial/getting started guide for node in #node.js at the moment?
19:28 <@isaacs> deoxxa: it installs with a folder with package.json, just for consistency more than anything else.  makes other stuff easier later.
19:28 <@isaacs> deoxxa: ls, update, etc.
19:28 < deoxxa> that actually makes things way easier for me at work - we generate js from project definitions and until now i've had to do some hacky `tar cvzf -' stuff to get that into a form i can install via npm
19:28 < sschultz> Anyone in the NYC-area who would be willing to spend some hours tutoring two node newbies? Or know anyone who you think would be willing to?
19:30 < context> wth is this horse_js suppose to be
19:30 < sosnon> so there is no official node guide in #node.js?
19:30 < Mr_Grim> i installed npm for os x, then i installed uglify-js. in windows i didnt have to bother to setup my path, but it os x, it seems like i dont have the paths setup. is there an additional step to configure path information for npm on mac os x?
19:30 < Mark___> Hey @isaacs I'm trying to install a module via npm install module-name and while it works fine on my Mac, it fails on Solaris
19:31 < Mark___> @isaacs my path is git+ssh://git@github.com:mycompany/module-name.git#v0.1.1" and it tries to get the package from the npm registry, if I remove the "+ssh" then it tries to do a git clone, but with the wrong params
19:32 < Mr_Grim> got it
19:32 < Mr_Grim> had to use the -g flag
19:32 < Mark___> @isaacs any idea what's up? I just updated to npm 1.1.66
19:37 < jamespollack> hey all -- quick question:  any tips on how to get npm modules installed globally on OS X (mountain lion) into my path?  i can't use grunt or forever in my local environment... node@0.8.14, NPM@1.1.65
19:38 < othiym23> jamespollack: npm install -g?
19:38 < deoxxa> jamespollack: i have node installed locally to my user, so i never have to sudo or anything to use -g
19:38 < jamespollack> i used homebrew to install...?
19:38 < deoxxa> haha
19:38 < deoxxa> oops
19:39 < jamespollack> everything with any local npm module works fine, will a reinstall of npm with the -g flag change things?
19:39 < deoxxa> doubtful
19:40 < jamespollack> if i do an npm ls -g
19:40 < deoxxa> if i were you, i'd remove the homebrew-installed version and install it locally
19:40 < jamespollack> kk just compile node and npm from source?
19:40 < jamespollack> then install with the -g flag...
19:40 < jamespollack> ?
19:41 < deoxxa> don't have to compile it if you don't want, there's binaries at http://nodejs.org/dist/latest/node-v0.8.15-darwin-x86.tar.gz
19:41 < deoxxa> download, untar into somewhere local (i use ~/local/node/versions/0.8.15), add to path
19:41 < deoxxa> and that's literally it
19:41 < jamespollack> kk
19:41 < deoxxa> you're ready to go after that
19:41 < jamespollack> thought the pkg manager was the way to go but kinda wonky
19:42 < jamespollack> i'll give that a go
19:42 < deoxxa> with something that's updated as often as node, a package manager is always going to mean a bit of friction
19:42 < jamespollack> @deoxxa thanks i'll give this a try
19:42 < deoxxa> but yeah, the main advantage to keeping your install local and isolated like that is that you don't have to worry about permissions or whatever
19:43 < deoxxa> and you always know your entire node install is in this one location
19:43 < deoxxa> so you can get rid of it all and start again if you break it bad enough
19:50 < trypwire> hey guys. I'm writing a node cli tool that outputs a javascript file using a doT template
19:50 < joemccann> getting ECONNREFUSED on dillinger.io
19:50 < trypwire> is there a way to minify the template string before writing to file?
19:50 < joemccann> ???
19:51 <@mbalho> joemccann: is that on nodejitsu?
19:51 < joemccann> yeah
19:51 < joemccann> dill.jit.su
19:51 < joemccann> cname'd to dillinger.io
19:51 <@mbalho> joemccann: they have a support channel #nodejitsu but usually they just tell you to do jitsu restart <appname> or wahtever
19:51 < joemccann> oh shit
19:51 < joemccann> whoops
19:51 < joemccann> heh
19:51 < joemccann> wrong window!
19:52 < Mark___> heh
19:52 < Mark___> Does anyone here use private modules?
19:52 < Mark___> and install them via npm from package.json?
19:55 < polyrhythmic> Mark___: yes, check out npm link https://npmjs.org/doc/link.html , or if you have a private git repo npm takes git:// urls as well
19:57 < Mark___> @polyrhythmic Thanks, is Solaris a supported npm platform? Because it works fine on Mac, but not on Solaris (private github module that is)
19:58 < Mark___> Even something as simple as this doesn't work on Solaris: https://un:pw@github.com/mycompany/module-name/archive/v0.1.1.tar.gz
19:59 < Mark___> npm says, package is not in the npm registry
20:01 < deoxxa> Mark___: what version of npm are you using?
20:02 < CoverSlide> Mark___: does un:pw work on other platforms?
20:03 -!- mode/#node.js [+o piscisaureus_] by ChanServ
20:04 < Mark___> @deoxxa: 1.1.66
20:04 < deoxxa> well that should definitely support http links
20:04 < Mark___> @CoverSlide: Yes, I can use the same command on my Mac and it works fine
20:04 < CoverSlide> hmm
20:05 < CoverSlide> command? what command? maybe there is a shell character escaping issue
20:05 < Mark___> @CoverSlide @deoxxa I believe it's a solaris bug, since it keeps telling me the package is not in the npm registry
20:05 < CoverSlide> or does that not work in your package.json?
20:05 < Mr_Grim> dumb question... anyone still use a sparc workstation for fun?
20:06 < Mark___> @CoverSlide: it doesn't work from command line or package.json
20:06 < Mr_Grim> i have not used one for YEARS
20:06 < CoverSlide> noone uses a sparc workstation for fun
20:06 < Mark___> @CoverSlide: Here's an example of a command that installs just fine on Mac, but tells me that package is not in NPM on Solaris: npm install git+ssh://git@github.com:mycompany/module-name.git
20:06 < Mr_Grim> there is probably not too much you can do with one these days
20:07 < CoverSlide> if they use it at all it's because their company won't pay to upgrade it, and they can't get a better job
20:07 < CoverSlide> hmm
20:08 < Mr_Grim> the RISC glory days are gone and over
20:08 < Mr_Grim> i have macbook pro for 90% of work now
20:08 < joaojeronimo> is it a good idea to use dnode across the network on top of shoe and sockjs ? (clients being node.js)
20:09 < CoverSlide> i have a 10" Eee PC for 90% of my work
20:09 < CoverSlide> i think it would be better to use dnode clients for dnode
20:09 < CoverSlide> dnode has shoe as a dependency for browsers
20:10 < joaojeronimo> CloverSlide: yes
20:10 < CoverSlide> if you add more cogs, you will inevitably get some performance penalties
20:10 < CoverSlide> is it a bad idea? no. is it a good idea? definitely not
20:10 < joaojeronimo> CloverSlide: but the thing is that dnode is a duplex stream, and over TCP... humm.. maybe that will work better yes
20:13 < context> wtf. ec6 modules look crazy
20:13 < context> and i bit.... pythonish
20:13 < context> or perlish
20:14 < context> err, es6
20:16 < nathan7> context: they are crazy indeed
20:16 < phrearch> ec6 is in nodejs?
20:16 < phrearch> err es6
20:17 < nathan7> phrearch: It is as much in node.js as it is in V8
20:17 < nathan7> phrearch: which is 'slightly'
20:18 < phrearch> nathan7: would be nice to use the class syntax in es6
20:18 < nathan7> meh.
20:18 < nathan7> Fuck that shit.
20:19 < phrearch> not that important?
20:19 < nathan7> JS's object model is fucking awesome.
20:19 < nathan7> I think classes are an abomination.
20:19 < phrearch> hm, es6 support in the browser also allowes imports right?
20:20 < phrearch> i still have a lot to learn :)
20:20 < superjoe> nathan7, when you say js's object model, are you including prototype stuff? or are you saying prototypal inheritance is an abomination?
20:22 < CoverSlide> classes in JS use prototypal inheritance, but that doesn't mean that prototypal inhertiance uses classes
20:24 < MitchW> mildly off topic: anyone here use bootstrap by any chance? (and maybe know why navbar menus don't show up in responsive design sites / what the intended alternative is)
20:25 < nathan7> superjoe: I'm saying classes are an abomination. Prototypal inheritance <3
20:27 < CoverSlide> the solution to bootstrap problems is to not use bootstrap
20:28 < superjoe> MitchW, s/^m/w/
20:29 < MitchW> incredibly unhelpful, thanks...
20:30 < gf3_> deoxxa: It works great
20:30 < gf3_> deoxxa: It's in anticipation of Harmony
20:30 < superjoe> MitchW, ##twitter-bootstrap
20:30 < deoxxa> i get that
20:30 < deoxxa> but right now, broken
20:30 < deoxxa> @ gf3
20:31 < MitchW> superjoe: ta (and not that wildly offtopic, I'm doing it all in node ;))
20:31 < chilts> CoverSlide: yup, so true. So long as you know what you're getting or not getting, or getting yourself in for. :)
20:31 < deoxxa> unfortunately i live in this dimension that has gravity and solar flares and all sorts of messed up things like time being relatively immutable from my perspective
20:31 < bitfed> Code Academy could really use some node courses
20:31 < deoxxa> so i can't just nudge it along to the point where harmony is a properly working thing
20:31 < deoxxa> :<
20:32 <@mbalho> bitfed: these people with the basically nonexistent open source presence?? https://github.com/codecademy
20:32 < gf3_> deoxxa: Did you use the --harmony flag w/ node?
20:32 < deoxxa> i did, and that worked fine
20:32 < deoxxa> but like, harmony as a requirement for an irc parser?
20:32 < deoxxa> seems a bit overkill
20:34 < gf3> deoxxa: I guess I just don't see the issue
20:34 < gf3> deoxxa: It's available in virtually all current node versions, and it'll be standard soon
20:34 < nathan7> deoxxa: time immutability needs work, yeah
20:36 < Kakera> is there any difference between {read,write}UInt8 and the [] operator?
20:36 < bitfed> hmm, I hadn't considered the non open source nature of it.  I, however, may still donate my own time to teaching others to code node.js regardless of the open source status of the site.
20:37 <@mbalho> bitfed: i wrote jsforcats.com and started working on nodeforcats https://gist.github.com/4011336 in case you wanna contribute some
20:37 < bitfed> oic, I hadn't realised it was a learning resource
20:41 < bitfed> Rad beard btw.
20:41 < Kakera> looks like writeUInt8 actually uses []
20:41 < CoverSlide> Kakera: yes
20:41 < Kakera> so they are equivalent
20:41 <@mbalho> bitfed: thanks
20:41  * mscdex shakes a fist at OpenSSH!
20:42 < nathan7> mbalho's beard kind of looks out of place
20:42 < CoverSlide> Kakera: it does some assertion checking as well
20:42 < Kakera> writeUInt8 does bound checks though
20:42 < nathan7> like it was shopped in
20:42 <@mbalho> nathan7: do a google image search for maxogden to see overwhelming evidence in support of my beards realness
20:42 < nathan7> Kakera: ~and I do a bounds check / before I write to an array~
20:42 < CoverSlide> mbalho: when did you start your beard?
20:42 < nathan7> mbalho: I believe in its realness
20:43 <@mbalho> CoverSlide: 2008 or so
20:43 <@mbalho> when i quit ruby
20:43 < nathan7> mbalho: It's just that it looks entertaining to me
20:43 < CoverSlide> cool
20:43 <@mbalho> too much face manicuring in that community
20:43 < bitfed> Entertaining is but one of the many benefits of having such an awesome beard
20:44 < nathan7> peer pressure made me shave :(
20:44 < bitfed> It's a lot of pressure to be such a man
20:44 < nathan7> Now I'm waiting for my beard to reappear
20:44 < bitfed> But, beards grow back
20:44 < CoverSlide> itchiness made me shave
20:44 < nathan7> I miss my beard
20:44 < nathan7> I once had an English test and I'm done with those pretty quickly
20:45 <@mbalho> i hear a lot about itchiness but never had any issues
20:45 < nathan7> and noise was coming from upstairs
20:45 < nathan7> so I was sent up to complain by the teacher
20:45 < nathan7> I find an entire class of kids who are convinced the hammer is the Ultimate Tool.
20:45 < bitfed> A good animal hair brush will help move the oils around so it doesn't get dry enough to itch
20:46 < nathan7> That day my normal hat had disappeared for a bit, so I was wearing a purple hat
20:46 < nathan7> I poke their teacher to complain, and she tells all the kids to stop
20:46 < nathan7> And starts saying "There's a *teacher* coming from downstairs, saying he can't give his lesson because of the noise you're making!"
20:46 <@mbalho> haha
20:47 < nathan7> I'm 17. I was wearing a purple fucking hat.
20:47 < nathan7> And nothing tips her off that I'm not a teacher, apparently.
20:47 <@mbalho> purple is a royal color
20:47 <@mbalho> and beards imbue you with natural confidence
20:47 < CoverSlide> you can't wear hats in the classroom
20:47 < CoverSlide> i call bullshit
20:47 < nathan7> Yeah, the beard did the trick.
20:47 < nathan7> CoverSlide: My teachers are surprisingly okay with it
20:47 < chovy> that's when you go buy beer
20:47 < nathan7> CoverSlide: This was on a monday
20:47 < bitfed> haha
20:47 < jamespollack> @deoxxa -- the reinstall of node works great globally, thanks for the advice
20:48 < nathan7> CoverSlide: The thursday before I forgot my hat. I thus had no hat on friday.
20:48 < deoxxa> jamespollack: :D
20:48 < nathan7> CoverSlide: My biology teacher jokingly told me he couldn't let me into the lesson without a hat
20:48 < nathan7> CoverSlide: I didn't have my hat back yet on tuesday, so I wore the purple one to school
20:48 < CoverSlide> i wish to wear a motorcycle helmet at all times
20:48 < CoverSlide> but it seems banks and convenience stores have a problem with it
20:48 < jamespollack> i'm a work and someone else was having the same problem and just writing custom path variables all the time in their scripts.  so now it's fixed for us both
20:49 < nathan7> My chemistry teacher demans I tip my hat at the start of every lesson
20:49 < deoxxa> GOOD DAY, SIR
20:49 < nathan7> and today my classmates decided my hat was the ideal object for playing catch with
20:50 < deoxxa> ohhh dear
20:51 < DukeofUR_> howdy, I have a question, how do I setup like environment variables so node knows which ports to listen on? of even if this is an environment variable thing
20:51 < CoverSlide> I AM NOT A HAT! http://www.youtube.com/watch?v=YhtD0ixuEFg
20:51 < chovy> NODE_ENV=test node ./app.js
20:52 < chovy> process.env.NODE_ENV
20:52 < chovy> DukeofUR_
20:52 < DukeofUR_> where am i putting that?
20:52 < chovy> well you need to check process.env.NODE_ENV 
20:52 < chovy> to see which env you are in
20:53 < DukeofUR_> ok 
20:53 < chovy> typically i have a ./config/index.js that will load ./config/config.test.js based on the env.
20:53 < nathan7> CoverSlide: Fezzes!
20:53 < chovy> or ./config/config.production.js
20:53 < nathan7> I ought to print me a fez.
20:53 < chovy> etc
20:53 < bingomanatee> afternoon node
20:53 < chovy> and a ./config/config.global.js for everything else.
20:53 < DukeofUR_> i see
20:53 < nathan7> The world is like lacking in fezzes and stuff.
20:54 < nathan7> open-source fezzes
20:54 < chovy> DukeofUR_: maybe i can do a post about this
20:54 < bingomanatee> Does anyone have a multi-column sort algorithm for sorting arrays of objects they like?
20:54 < DukeofUR_> chovy: do u have a blag?
20:54 < bingomanatee> given values can be string or numeric
20:55 < chovy> DukeofUR_: yeah, but i haven't written anythign yet
20:55 < CoverSlide> i like ruby's sort_by method
20:56 < CoverSlide> it wouldn't be too hard to implement that into javascript
20:56 < DukeofUR_> chovy: like u havent written anything at all or just on this subject?
20:56 < chovy> not on node
20:56 < chovy> i have some really old stuff i posted a few years ago 
20:57 < DukeofUR_> chovy: what kind of stuff do you write about?
20:57 < chovy> webdev in general
20:57 < DukeofUR_> linky
20:57 < chovy> chovy.com
20:57 < DukeofUR_> ahah
20:57 < CoverSlide> bingomanatee: schwartzian transforms are great for this type of thing, but I know of no lib that does it, ususally people just write their login in their .sort callback
20:57 < CoverSlide> *logic
21:01 < nathan7> CoverSlide: inspired by your link.. https://github.com/nathan7/scadstuff/commit/a756 :D
21:02 < CoverSlide> neat!
21:03 < nathan7> It even has a hole for the string!
21:04 < bingomanatee> what is OpenSCAD
21:06 < CoverSlide> it's a scad, but open
21:06 < bingomanatee> go on...
21:06 < nathan7> http://url.nathan7.eu/openscad
21:07 < bingomanatee> ok
21:30 < chovy> DukeofUR_: http://www.chovy.com/node-js/managing-config-variables-inside-a-node-js-application/
21:31 < DukeofUR_> chovy: cool!
21:33 < Leeol> Anyone ever have trouble with Stylus Middleware behaving.. oddly? Sometimes it renders, sometimes it does not.. and i can't make sense of it
21:43 <@mbalho> bingomanatee: i like crossfilter.js
21:47 < chovy> DukeofURL: let me know if there's anything else you want me to write about
21:48 < TheDracle> I'm trying to use a Future from the fibers package, and having some strange issues. I'm wrapping a function that expects two arguments an a callback: sync_create_session = Fiber.wrap(api.createSession); sync_create_session.call(api, 'arg1', 'arg2').wait();
21:48 < DukeofURL> fo sho, i will hit u up when i smash my head against the next wall
21:48 < TheDracle> Normally the third argument would be the callback. In the api.createSession it calls the callback, but for some reason it's set to function(){[native code]}, with no params.
21:49 < peterschmidler> Is there a recent book that covers node express etc?
21:51 < tracker1> Does anyone know of a script that essentially does what util.inspect does, but returns valid JSON?  (this is for logging error data, which is sometimes circular)
21:52 < bingomanatee> thx
21:53 < tracker1> right now, using a try/catch block that will prefer JSON.stringify, and in the catch, to util.inspect ... but would really prefer to always have valid JSON
22:08 < brianc1> tnorris: you there?
22:10 < tnorris> brianc1: yo yo
22:12 < Kakera> guys, what's less unreadable
22:12 < Kakera> return this.buffer.readInt16LE(4) + ((this.buffer[6] & 0x0F) << 16);
22:12 < Kakera> or
22:12 < Kakera> return this.buffer[4] + this.buffer[5] << 8 + ((this.buffer[6] & 0x0F) << 16);
22:12 < Kakera> forgot the brackets in the second line -.-
22:13 < brianc1> tnorris: i see you still using githubhub
22:13 < brianc1> tnorris: you notice ever when you log in that the UI doesn't "work" and you have to log out and log back in?
22:13 < tnorris> brianc1: nope. works every day for me.
22:13 < Topcat> has anyone had any joy with usb node and windows ?
22:14 < brianc1> tnorris: okay strange
22:14 < brianc1> tnorris: well if you notice anythang lemme know. :) I didn't get a chance to work on it over the t-day break. I took a computer break. decorated mah xmas tree.
22:14 < tnorris> Kakera: the first one is better, imho.
22:15 < tnorris> brianc1: will do. but it's been working great for me.
22:15 < brianc1> tnorris: yay!
22:20 < flazz> did npm stop linking to prefix/bin and now links to prefix/share/npm/bin ?
22:27 < Kakera> hm, I think return this.buffer.readUInt32LE(4) & 0x000FFFFF; should be equivalent to what I pasted above
22:49 < styol> Does anyone know how to write mocha tests while utilizing the core cluster module? It seems that some unexpected behavior occurs due to clustering. In particular, the tests are only ran against the master process and the child process(es) are initialized after the test(s) are ran
22:50 < styol> Well, perhaps it is expected behavior, but I'm not sure how to work around it and write tests against the child process(es) where a majority of the functionality exists in this instance
22:56 < styol> I guess I could separate out the contents of the child process into a separate module that gets included into the usual place and unit test accordingly
22:58 < jerrysv> Nexxy: ping?
23:01 < context> id avoid child process if at all possible otherwise the logic for that process id seperate out into its own module[s]
23:01 < Topcat> node-gyp, windows, python27 and 33 install how can i tell what version node-gpy is trying to use?
23:02 < TooTallNate> Topcat: it tries to use the "python" in your PATH first
23:02 < TooTallNate> otherwise checks manually in C:\Python27
23:02 < Topcat> ah ok
23:02 < TooTallNate> Topcat: you can force one with --python=python27 or whatever though
23:11 <@isaacs> mbalho: I HAVE A PRESENT FOR YOUUUUUUUUUUUUUU!!!!
23:12 < ohnoimdead> can someone point me in the right direction for best practices packaging a node application (along with npm package deps) as a debian package? specifically i'm trying to figure out how to freeze the npm packages our app requires and "npm install" them from a build machine with no access to the public npm repo (or the internet for that matter).
23:13 < othiym23> isaacs is uncharacteristically effusive
23:14 < othiym23> ohnoimdead: I think you can choose to make a Debian package containing all the deps, a set of Debian packages that each contain a single dep, or do the whole thing via npm, but be careful of combining the two packaging systems
23:14 < othiym23> there are many pitfalls for the unwary in there
23:15 < othiym23> if you're making a deb for local use, I'd suggest just creating the whole thing as a single deb with all the node_modules in there as a unit
23:15 < niggler> just stick the deps in the repo
23:15 <@mbalho> isaacs: i just got to xolo, bout to start hackin (with jlord)
23:16 < mscdex> i'm so glad making a public key from a private key is so simple
23:17 < niggler> hey mscdex came across jszip and zip.js  familiar with either?
23:17 < mscdex> nope, sorry
23:18 < CoverSlide> jszip is cross browser, zip.js needs typed arrays
23:19 < CoverSlide> zip.js uses web workers to avoid blocking
23:19 < CoverSlide> i'm assuming jszip does not
23:19 < ohnoimdead> othiym23: i would do that (just include the full node_modules folder in the package) however we have two npm packages that compile native code which i believe is done as a part of npm install. so i think what i need is to have the source of each npm package available in the debian package bootstrap chroot thingy and do the npm install from there. unless i'm just thinking about this all wrong (which is entirely possible and i would be delighted to hear).
23:22 < othiym23> ohnoimdead: just package up the compiled modules and don't run npm install once you install the dep
23:23 < deoxxa> ohnoimdead: you might be able to just run the `postinstall' script inside those modules
23:23 < deoxxa> ohnoimdead: likely it's `node-gyp rebuild' or `node-waf configure build'
23:24 < deoxxa> ohnoimdead: that'll work for nearly all cases except where a specific, different module is required for a certain platform (which is very rare, i think the only one i've seen so far is appjs)
23:25 < othiym23> the whole point of using the Debian packaging system is that you have control over what binary dependencies are installed before your package, so you shouldn't have to do any rebuilding after install
23:25 < deoxxa> othiym23: he's asking about building the package itself
23:25 < othiym23> gotcha
23:25 < deoxxa> othiym23: building packages for debian is a huge runaround, surprisingly enough
23:25 < styol> context: yeah, seems that might be my only option.. gracias. Seems to be working, but getting some oddness on paths to require'd modules. require(__dirname + '/../dir/file.js') seems to be an odd fix ;)
23:26 < deoxxa> you have to build them in a special chroot for each architecture you want to support
23:26 < ohnoimdead> deoxxa: we are using node-gyp. and yeah, the trouble is with the packaging itself. these build boxes are internetless.
23:26 < deoxxa> ohnoimdead: that should be fine - run the install and package it up on a box with internet, then run `node-gyp clean configure build' in each dependency that needs to be compiled
23:27 < deoxxa> ohnoimdead: i reckon that'll get it building cleanly first go for you, but if it doesn't... it'll at least get you closer
23:27 <@isaacs> mbalho: there is no way i'll make ti to xolo
23:27 < ohnoimdead> deoxxa: i think that would work!
23:27 <@isaacs> mbalho: but you'll like it.
23:27 < niggler> or if the kernel/distro on the deployment targets are the same, you could just tarball it
23:27 <@isaacs> mbalho: everyone who's seen it has said "You have to give that to Max"
23:27 < deoxxa> niggler: debian supports like 10 different architectures, lol
23:27 <@mbalho> isaacs: whaaaaaaa
23:27 <@isaacs> mbalho: yeah
23:28 < niggler> deoxxa: it sounds like ohnoimdead had a specific deployment in mind
23:28 -!- mode/#node.js [+o piscisaureus_] by ChanServ
23:29 < JohnMcLear> anyone got any opinion on the cleanest way to detect if a file is encoded as ASCII and convert to utf-8?
23:29 < chilts> ascii is a subset of utf-8 so it's automatically urf-8
23:29 < JohnMcLear> this file is being uploaded, btw..
23:29 < chilts> *utf-8 :)
23:30 < JohnMcLear> okay let me rephase, a .txt file that contains "£" when uploaded and converted to utf-8 shows the "£" as "?"
23:30 < JohnMcLear> Further details: https://github.com/ether/etherpad-lite/pull/1205#issuecomment-10738428
23:30 < ohnoimdead> well, part of the whole debian bootstrap chroot thingy is to provide an environment identical to the target install environment. so building from there totally works. i just can't actually get the npm packages from there. in sad, but maybe humorous news, i tried using npm cache add then include that in the source for our build thinking that doing an npm install would try to install from the cache and maybe not care it has no internets. but no. not the way
23:31 < deoxxa> ohnoimdead: it... can... kind of... if you specify the exact version
23:31 < deoxxa> ohnoimdead: i think.
23:31 < deoxxa> ohnoimdead: but i might be making that up
23:32 < ohnoimdead> deoxxa: i really tried. my next step was to fake the public repo. that's when i was like STOP THE INSANITY
23:32 < deoxxa> haha
23:32 < deoxxa> "hoooooold on"
23:32 < deoxxa> i know that feeling :P
23:32 < ohnoimdead> i think doing just the gyp building from the package server will work though
23:35 < ohnoimdead> thanks othiym23 and deoxxa!
23:36 < deoxxa> good luck!
23:36 < ohnoimdead> O.o <- how i feel about packaging
23:36 < othiym23> ohnoimdead: yeah, hope it works out
23:37 < othiym23> APT / dpkg are awesome to use, but they're not so fun to build for
23:37 < ohnoimdead> othiym23: you ain't just whistling dixie
23:46 < troy--> bros, when doing console.log on a post request, does it still go to the same stdout?
23:47 <@mbalho> dont bro me if you dont know me
23:48 < troy--> , when doing console.log on a post request, does it still go to the same stdout?
23:48 <@mbalho> thanks. a node process has 1 stdout and console.log will write to it
23:49 <@mbalho> you can also pipe streams to process.stdout, e.g. request('foobar').pipe(process.stdout)
23:49 < chovy> troy--: i usually do something like 'node app.js > app.log 2>&1
23:49 < chovy> then i tail app.log
23:49 < troy--> ok thanks
23:50 < Valorr> Hey, kind of random, but I'm trying to navigate the file picker api and only allow ePubs to be uploaded. The mime type is epub+zip, but it keeps skipping the '+' and not validating proper epub files. Anyone have any stabs at how to fix this?
23:52 < CoverSlide> epub+zip doesn't look like a valid mime type
23:52 < CoverSlide> you're sure it's not application/epub+zip ?
23:53 < Valorr> Might be. Lemme give that a try!
23:53 < Valorr> it's weird, because the api was validating images as image/*
23:53 < CoverSlide> hmm
23:53  * CoverSlide shrugs
23:54 < Valorr> thanks regardless 
--- Log closed Tue Nov 27 00:00:20 2012