[04:32] Aria: That's HTML, not Javascript. Not Node, anyway.
[04:33] singalong: Aria: well just thought people here could help with some javascript.
[04:33] Aria: Yeah, but it's not actually a javascript question terribly much. The Javascript part is 'do the math'.
[04:33] Aria: Get the bounding box, check for overlap.
[04:34] singalong: Aria: should that have to be in a loop? or are there events for that?
[04:35] Aria: Events for doing math?
[04:35] Aria: I think you're probably asking for rather browser-centric questions.
[04:35] Aria: DOM stuff.
[04:35] singalong: Aria: ya suppose I have to detect collision, would I have to detect them in a for loop etc?
[04:35] Aria: Well, you have to test against every possible object...
[04:36] Aria: Where hopefully 'possible' is a carefully groomed list so you don't do much work for each check.
[04:37] singalong: Aria: My idea was to make a 2d game (top view). but sounds like it's going t be tough if it's top-view. detecting every object on the screen,
[04:38] blogometer: ryah: Found the bug.
[04:38] Aria: Shouldn't be too bad ... people do this all the time, singalong.
[04:39] Aria: I mean, the DOM is an ugly API and painful, but not too terrible, really.
[04:40] singalong: Aria: I only wish there was something like onCollide() which returns the id of the element it collided with :)
[04:40] Aria: Hehe. So write that.
[04:40] Aria: Though the real trick is pruning the list of possible objects.
[04:40] blogometer: singalong: Yeah. Add it to the prototype.
[04:40] singalong: Aria: ya was just thinking about it
[04:40] ivong has joined the channel
[04:41] blogometer: singalong: R-tree?
[04:41] singalong: blogometer: well i dont understand that (just googled. will have to read the wikipedia page to know what it is)
[04:41] dnyy: /w/win12
[04:41] dnyy: fucking wireless, ignore that
[04:42] blogometer: singalong: It is a solved problem.
[04:43] Validatorian: is there any way in node to add and remove content to a single line in console... for example a loading indicator, counting up from 1 to 100%, while taking up the same space...
[04:43] singalong: seems like there's another nice way
[04:43] blogometer: I'm tired.
[04:44] singalong: Aria, blogometer: elementFromPoint() https://developer.mozilla.org/en/DOM/document.elementFromPoint
[04:44] Aria: Nice if you have a direct hit.
[04:44] davidwalsh has joined the channel
[04:44] blogometer: But...
[04:44] Aria: But it's a single point and not a bounding box.
[04:44] Aria: If your objects are all such that they'd always hit a corner, you could just test the corners of your objects.
[04:45] Aria: But if not. ..
[04:45] blogometer: Try just seeing if you have an overlap on the x-axis.
[04:45] singalong: it reduces the possibilities
[04:45] JosephMoniz: whats all this fast buffer hype?
[04:45] Yuffster has joined the channel
[04:46] nefD has joined the channel
[04:46] Aria: JosephMoniz: A buffer pool to reduce the number of small allocations.
[04:46] blogometer: Create an array (list) of left bounds and right boutnds in an array.
[04:46] JosephMoniz: so like a mini-malloc in JS?
[04:46] blogometer: Easy to move around and test.
[04:46] singalong: and woot! http://code.google.com/p/doctype/wiki/DocumentElementFromPointMethod
[04:46] singalong: supported browser list
[04:47] singalong: supported in IE9 too according to quirksmode
[04:47] Aria: JosephMoniz: No, like a pool of buffers that are allocated, returned to the pool when not used anymore. Since it turns out that allocating small buffers is a part of the performance hit.
[04:47] blogometer: http://www.codeproject.com/KB/GDI-plus/PolygonCollision.aspx
[04:48] JosephMoniz: ooooo, hey thats pretty creative
[04:49] blogometer: singalong: Meh.
[04:49] blogometer: You're going to need to create your own structure of bounding boxes.
[04:49] singalong: :) well my app just needs to detect collision between images.
[04:50] blogometer: But, good luck with it. Does SVG have the right function I wonder?
[04:50] singalong: that way i can keep it simple too.
[04:50] singalong: blogometer: I'll have to skip canvas and svg and use a simple div with images or whatever simple stuff available
[04:50] jamescarr_: so many good books spend eternity in Manning's MEAP :(
[04:51] blogometer: So simple it doesn't work.
[04:51] jamescarr_: I was for sure Secrets of a Javascript Ninja would be out by now :(
[04:51] blogometer: But, good luck. If it doesn't work, then model it yourself.
[04:55] Tim_Smart: ryah: I have to go through any api method that uses buffers, and insert checks for buffer.parent and alter the offsets etc as needed.
[04:56] blogometer: clear
[04:59] joshbuddy has joined the channel
[05:02] blogometer: ryah: I'm a winner!
[05:03] blogometer: Found the bug. Created a fix.
[05:03] MikhX_ has joined the channel
[05:03] blogometer: Will let Felix know.
[05:03] _announcer: Twitter: "How expensive are function calls in JS running under Node? I hope not very, because Fab.js looks like it must be SLOW." -- Ryan W Tenney. http://twitter.com/ryantenney/status/21724450461
[05:05] micheil: blogometer: bug?
[05:05] blogometer: micheil: Bug in node-mysql.
[05:05] micheil: k
[05:06] blogometer: Forking, patching now.
[05:11] jimmybaker has joined the channel
[05:17] kreyman has joined the channel
[05:21] mattly has joined the channel
[05:24] ryah: blogometer: good !
[05:25] blogometer: Node.js is soooo much more fun than Java.
[05:26] singalong: now I know why Fab.js isnt popular.
[05:26] singalong: Fab.js syntax is similar to lisp :P
[05:29] saikat: has anyone tried using Connect with Socket.io?
[05:29] saikat: is there a way to maket he connect logger work behind socket.io? i talked to creationix earlier today and he said "sort of" but i forgot to ask more about it
[05:36] mjr_ has joined the channel
[05:38] schwa23 has joined the channel
[05:39] schwa23: Can anyone tell me why http requests made through node wouldn't show up in charles proxy?
[05:39] bdha has joined the channel
[05:40] jetienne has joined the channel
[05:41] saikat: does connect still come with an executable that can daemonize my process?
[05:43] benburkert has joined the channel
[05:44] marek_z: christ, *anything* is more fun than java
[05:44] schwa23: I'm trying to see what's happening to a request I'm making (using restler at the moment) but seems like node doesn't use the MacOS proxy settings…
[05:44] schwa23: anyway to force it to go through charlesproxy?
[05:45] satori_ has joined the channel
[05:47] siculars has joined the channel
[05:47] mjr_: schwa23: node does not use the proxy settings
[05:47] mjr_: node is very low level
[05:48] holydevil has joined the channel
[05:48] mjr_: However, you can force things through your proxy in node by using the host/port of the proxy
[05:49] schwa23: mjr_: that kind of makes sense -- so set up some kind of reverse proxy through charles so my node requests go through localhost?
[05:50] mjr_: schwa23: basically, yes. That is what Safari does when you change the proxy settings. Instead of doing a DNS lookup and connecting to the resulting IP, it sends all traffic to your proxy server.
[05:51] mjr_: Another thing you could do is use my http_trace program, which just pulls bits off the wire and decodes them.
[05:51] schwa23: mjr_ that sounds like what i need
[05:52] schwa23: where do I get that?
[05:52] mjr_: npm install pcap
[05:52] mjr_: Then you'll get a program called http_trace
[05:52] mjr_: http://github.com/mranney/node_pcap
[05:53] schwa23: mjr_: sweet, I'll give it a shot! Thanks!
[05:53] mjr_: charles proxy looks fantastic though. I hope http_trace is that good some day.
[05:54] Egbert9e9 has joined the channel
[06:02] amuck has joined the channel
[06:02] jakehow has joined the channel
[06:03] schwa23: yeah charles is pretty awesome. More power than I need for this simple debugging task, but overall a great tool.
[06:04] schwa23: mjr_: http_trace working like a champ. thanks for the tip & the tool!
[06:04] mjr_: nice, glad to hear it
[06:04] mjr_: I've got big plans for it with various options and shit, which I hope to work on this weekend.
[06:11] _announcer: Twitter: "Established on a machine node.js I did not think it would be so easy:) But when doing. / Configure I found that some libraries do not" [ru] -- Pavel Pavlenko. http://twitter.com/pavlik/status/21728191082
[06:11] stellar has joined the channel
[06:19] ryah: how can we get node's libraries up to speed...
[06:19] ryah: 3rd party modules that is...
[06:19] mjr_: what bits are not up to speed?
[06:19] mjr_: Like, not up to node knockout speed?
[06:19] satori_: Speed? performance? reliability? conformance?
[06:19] ryah: no - usablity
[06:20] ryah: general quality
[06:20] satori_: k
[06:20] mjr_: Yeah, we need more quality for sure.
[06:20] ryah: how can we avoid the ruby gem's trash dump?
[06:20] satori_: cream will always rise to the the top
[06:20] ryah: there's something to this app store idea.
[06:20] blogometer: More collaboration.
[06:20] ryah: someone who looks over your work
[06:20] blogometer: Also, a web site that has a place to vote up and down.
[06:20] blogometer: Yes.
[06:20] ryah: approves or rejects
[06:20] blogometer: Peer review.
[06:20] blogometer: No.
[06:21] blogometer: Because, then it becomes Apache's trash dump.
[06:21] ryah: :)
[06:21] Tim_Smart: ryah: The package managers need well defined rules.
[06:21] blogometer: Bike sheds galore.
[06:21] ryah: agreed. we want something between apache and ruby gems
[06:22] ryah: there needs to be structure to the module system.
[06:22] ryah: automated tests or something..
[06:22] Tim_Smart: ryah: Btw, getting FastBuffers working with the old api has more to it than expected...
[06:22] ryah: Tim_Smart: well, there's the whole c++ side of things
[06:22] ryah: you need to reimplment the c++ side to call js
[06:22] mjr_: ryah: what are you noticing right now that has you bring this up?
[06:23] blogometer: The Apache process leads to a communal refactoring.
[06:23] ryah: that last tweet: http://twitter.com/pavlik/status/21728191082
[06:23] blogometer: Which is really, really awful to watch.
[06:23] mjr_: I don't speak unicode
[06:23] satori_: ...or russian
[06:23] blogometer: GitHub makes for better code.
[06:24] mjr_: I see the translation above.
[06:24] ryah: http://translate.googleusercontent.com/translate_c?hl=en&ie=UTF-8&sl=ru&tl=en&u=http://twitter.com/pavlik/status/21728191082&rurl=translate.google.com&twu=1&usg=ALkJrhikYGOf0zkkTNchXOQP8HulpMZNPw
[06:24] ryah: well, whatever
[06:24] ryah: i mean who knows what it says :)
[06:24] mjr_: some libraries do not have
[06:24] satori_: heh
[06:24] blogometer: Looks like he can't build.
[06:24] blogometer: Maybe he pulled it onto Windows.
[06:24] mjr_: some libraries do not have hidden pornography
[06:25] mjr_: Although, to be fair, some do have hidden pornography.
[06:25] blogometer: Do you want to discourage people from creating libraries?
[06:25] ryah: a wiki page is not enough
[06:25] bdha: blogometer: Lots of libraries on the CPAN. Lots of awful ones.
[06:25] mjr_: ryah: I think we just need way, way better docs.
[06:25] bdha: You have to Know which to use.
[06:25] satori_: yeah. A wiki page is alittle too amatuer hour on it's own
[06:25] ryah: yeah
[06:25] ryah: maybe creating a system for doing docs is the way
[06:26] bdha: cpantesters/cpanratings helps.
[06:26] ryah: i mean, cpan is pretty good.
[06:26] mjr_: We probably have to build up some official documentation that isn't restricted to fitting into a man page.
[06:26] bdha: I love CPAN, but it's a stressful relationship. ;)
[06:26] ryah: yeah, the docs have to break out of the man page
[06:27] blogometer: Maybe docs, but Node.js is open source.
[06:27] blogometer: Don't go the Javadoc route, where people think they are not *supposed* to read the source code.
[06:27] mjr_: If you look at the python or php docs, the shit is exhaustive.
[06:27] ryah: i sure do love man pages though...
[06:27] mjr_: Completely covering all possible angles.
[06:27] blogometer: I like man pages too.
[06:27] blogometer: I like the UNIX nature of Node.js
[06:27] ryah: i never use it for node, admittedly
[06:28] ryah: i always go to the website to look up api
[06:28] satori_: non trivial, commented, and up to date examples go a long, long way.
[06:28] mjr_: I love the idea of man pages, but I use the web pages too.
[06:28] blogometer: http://expressjs.com/
[06:28] mjr_: Like, I used the man pages just to make sure that the ronn conversion was working.
[06:28] blogometer: The Ruby world has a lot of this stuff, pretty landing pages at the GitHub page.
[06:28] mjr_: And never again.
[06:28] zomgbie has joined the channel
[06:28] blogometer: That is the sort of thing that builds confidence for me.
[06:29] ryah: i think the man pages can be split up
[06:29] mjr_: If man pages are getting int he way of good docs, and I'm not actually sure they are, but if they are, then we should probably just lose the ronn/man stuff.
[06:29] ryah: man node-http
[06:29] ryah: man node-net
[06:29] ryah: etc
[06:29] blogometer: Actually, before I abandoned Java, I was working on a build system for Java that worked closely with GitHub.
[06:29] blogometer: To try to get Javans to see the value of Github.
[06:30] ryah: the problem is - people make a little library - they put it on a website
[06:30] ryah: then it's abandoned
[06:30] blogometer: Encourage people to use GitHub.
[06:30] ryah: and everyone for the next two years has to sort through that
[06:30] blogometer: So that it can be forked and perpetulated.
[06:30] bdha: ryah: You need an Andy Lester. :)
[06:30] ryah: when they want a mysql module - they have to go find out which one to use
[06:31] blogometer: Also, with the new Wiki feature, maybe there's a way to build man pages from Wiki pages and vice versa.
[06:31] ryah: one simple solution is to add an appendex to the docs
[06:31] ryah: simple tell people which modules to use
[06:31] jetienne: http://www.symfony-project.org/plugins/ <- ryah here is how symfony (php framework handle it)
[06:31] streampunk has joined the channel
[06:32] ryah: bdha: i dont know hwo that is
[06:32] blogometer: Gah!
[06:32] bdha: ryah: petdance. Leader of the CPAN testing movement.
[06:32] jetienne: ryah: they host, provide a well defined set of info to put in there. and you got user votes.
[06:32] bdha: jetienne: That's.. not so great.
[06:32] blogometer: That's messy.
[06:32] bdha: It's a lot of noise.
[06:32] blogometer: Hey, as the resident newbie and fanboi...
[06:32] bdha: Appendex isn't a bad idea, but then why not just include them in the dist?
[06:32] mjr_: ryah: perhaps a node distribution can address this library quality issue.
[06:32] blogometer: ...I'd have to say, the process of asking the IRC channel for recommendations ...
[06:32] jetienne: ryah: user vote allows to get user usage. so a way to estimate quility
[06:32] blogometer: ... has been quick and rewarding.
[06:33] mjr_: blogometer: yeah ,but it doesn't sacle
[06:33] blogometer: I don't like the idea of a node distribution.
[06:33] blogometer: Now done is a big heap of shit to install just like Java.
[06:33] blogometer: s/done/node/
[06:34] jetienne: node distro will happen, but it is too early
[06:34] ryah: well - maybe it's not urgent - but in the coming months the issue of how to organize these modules is going to be a concern
[06:34] ryah: now that it's clear this project is marching forward :)
[06:34] satori_: you sure? :P
[06:34] blogometer: ryah: Why not make GitHub the default repository for Node.js.
[06:34] blogometer: ?
[06:34] ryah: blogometer: eh.
[06:35] blogometer: Well, then you can develop conventions, if you wed 3rd party development to GitHub.
[06:35] ryah: satori_: well, at least for the next couple months - yes
[06:35] Validatorian: is there a consensus on the fastest flow/parallel script? Step, Parallel, async, flow, and probably others I don't know of
[06:35] blogometer: You can use the APIs to gather statistics and report on relative health.
[06:35] blogometer: Health of projects.
[06:36] blogometer: Which are active, trending, commonly downloaded.
[06:36] ryah: what if i just choose modules and put them in the docs
[06:36] ryah: that sounds easier
[06:36] bdha: haha.
[06:36] bdha: Social solution > technical solution
[06:36] jetienne: ryah: doesnt scale.
[06:37] blogometer: Apache.
[06:37] ryah: i've heard larry elison makes every decision at oracle
[06:37] ryah: every hire
[06:37] ryah: every price change
[06:37] blogometer: Oh, shit.
[06:37] Tim_Smart: Validatorian: They are roughly the same
[06:37] blogometer: Well, I'm not working for Larry Ellison and I'm certianly not volunteering time.
[06:37] ryah: i think i can pick the top 20 modules each month
[06:38] bdha: ryah: And look what happened to Solaris? ;)
[06:38] ryah: and by me, i mean the community
[06:38] teemow has joined the channel
[06:38] ryah: but committed by me :)
[06:38] blogometer: ryah: You're being Socratic.
[06:38] blogometer: What is the problem?
[06:38] mjr_: why does it matter if people can find the right mysql module?
[06:38] blogometer: Are you getting stressed out by inquiries or something?
[06:38] mjr_: Or if people abandon their little library projects?
[06:39] Validatorian: Tim_Smart: alright, thank you
[06:39] blogometer: Really. I'd rather have an ecosystem.
[06:39] ryah: mjr_: makes it harder for newbies
[06:39] blogometer: And GitHub is already the repsoitory.
[06:39] blogometer: Bullshit.
[06:39] bdha: mjr_: People download a random module from the CPAN. It's utter garbage. It eats their faces and kills their dog.
[06:39] blogometer: I'm a newbie and this is much easier than Haskell.
[06:39] ryah: direction is good
[06:39] bdha: At the hospital and vet, all they do is say how much Perl sucks.
[06:39] blogometer: Landing at the Node.js page was reassuring, because there was so little to consdier to get started.
[06:39] blogometer: Landing at Erlang, I don't know what the fuck I was supposed to do.
[06:39] ryah: if the nodejs.org docs say "use mysql module X"
[06:39] ryah: then people will
[06:39] jetienne: what about something ala jquery ? several related projects but independant... jquery core, jquery ui, jquery mobile
[06:39] ryah: and they won't have to worry
[06:40] mjr_: Yeah, I guess that's all true.
[06:40] blogometer: Again, who's worried.
[06:40] mjr_: Is node.js a system for newbies?
[06:40] blogometer: ?
[06:40] ryah: and perhaps mysql module X maintainer will feel a bit resonsible
[06:40] ryah: mjr_: everythign should be a system for newbiews
[06:40] deadlyicon has joined the channel
[06:40] blogometer: Again, I had no problem finding a mysql module.
[06:40] ryah: perhaps they're experienced programmers like blogometer - but they're new to the framework
[06:41] blogometer: Okay. Well listen to jetienne then.
[06:41] blogometer: Don't make me feel like I need to deal with all that other stuff.
[06:41] ryah: jetienne: eh
[06:41] blogometer: Break it up into projects.
[06:41] ryah: jetienne: jquery side projects haven't done really well though
[06:41] ryah: it's not a success story
[06:41] blogometer: No idea what the problem is.
[06:42] blogometer: ryah: Based on what?
[06:42] jetienne: ryah: true but it doesnt mean all is bad
[06:42] ryah: I imagine somoene sending me a patch "I think the docs should have mysql module Y instead of mysql module X"
[06:42] ryah: and then a huge thread ensuing on the mailing list
[06:42] ryah: that's healthy
[06:42] blogometer: That's bike shedding.
[06:42] ryah: not when an answer is ultimately decided
[06:42] jetienne: i think it is too early
[06:42] blogometer: And a lot of argument in the list about which is the "blessed" mysql package.
[06:42] bdha: ryah: Or you end up with a tar vs. star explosion. :P
[06:43] mjr_: I think you can either do Batteries Included, or you can be a low level system that people can implement lots of different stuff on top of. Part of the benefit of a low level system is that it encourages experimentation and people duplicate effort for the greater good.
[06:43] jetienne: node.js as it is now still need a lot of work
[06:43] holydevil has joined the channel
[06:43] ryah: mjr_: i want to be low-level with batteries included
[06:43] ryah: without including batteries
[06:43] blogometer: Every bike shed now become a matter of ego and politics. It will be ugly.
[06:43] mjr_: tradeoff: failed
[06:43] jetienne: to work on a common library is usefull, but should not be that correlated to node for now
[06:43] ryah: i want node to be on the store shelf, and the batteries next to it
[06:44] jetienne: common lib = more standard = more used then less bugs = better doc
[06:44] ryah: i want to document which batteries to use
[06:44] blogometer: Node.js should stay small, things like Express and Connect can vie for the framework space.
[06:44] jetienne: there is time line tho
[06:44] ryah: blogometer: software is a social problem to begin with
[06:45] ryah: it's just stuff interacting with other stuff
[06:45] ryah: all human made
[06:45] blogometer: So, how big does Node.js become?
[06:45] ryah: all relatively arbitrary
[06:45] blogometer: ryah: None of that is useful.
[06:45] ryah: node itself certainly does not grow much bigger than it is now
[06:45] _announcer: Twitter: "Had a good night. But now its time for some REAL fun with node.js. WOW I SOUND COOL" -- Chirag Davé. http://twitter.com/ChiragRules/status/21729895237
[06:46] jetienne: i like the node.core (lean/clean) + node.lib
[06:46] Aria: I think there's an alternative to the batteries-included vs just-a-core philosophies. Build a library set as a separate project.
[06:46] ryah: but clearly it needs modules
[06:46] jetienne: node.core would be node.js as it is noz
[06:46] blogometer: No one is arguing that software is not social.
[06:46] Aria: ++
[06:46] blogometer: I believe, I'm arguing that the nature of open source, of a vibrant community, is one that has an ecosystem.
[06:46] jetienne: node.lib would be all the libraries requomended by node people
[06:46] ryah: blogometer: indeed
[06:46] jetienne: thus node.core and node.lib are clearly distinct
[06:46] ryah: blogometer: we wouldn't want to kill the ecosystem
[06:47] blogometer: The successful enviornments of the past have been Perl, Python, Ruby.
[06:47] blogometer: The anti-social enviornments would be PHP and Java.
[06:47] ryah: hm
[06:47] ryah: php and java are far more successful than those other three
[06:47] blogometer: PHP is all newbies all the time.
[06:47] ryah: whose names i've already forgotten
[06:47] jetienne: blogometer: i disagree php has the best doc i have ever seen
[06:47] mjr_: newbies like Facebook.
[06:48] bdha: Some of the biggest social networking sites on the web are written in PHP, dude.
[06:48] Aria: jetienne: But outside the core library, it's all 'not invented here'
[06:48] blogometer: jetienne: Ah, well, I haven't read it.
[06:48] bdha: ..
[06:48] ryah: node can hardly be newbies all the time - no matter how my flowers we tack on
[06:48] ryah: it's pretty low-level
[06:48] Aria: It's mostly NIH that pervades both PHP and Java cultures.
[06:48] blogometer: I suppose, I'm just considering the environment I want to work in.
[06:48] jetienne: Aria: yep and this is what we try to fix
[06:48] blogometer: If Node.js becomes not that enviornment, that is fine.
[06:48] Aria: (Ruby to a degree, but mostly it's just polyphony)
[06:48] jetienne: Aria: nih is common no matter the language
[06:49] mjr_: PHP is an nice templating language but ugly way to write programs. Their docs are pretty excellent, and they include all of the batteries most people ever need.
[06:49] jetienne: tss tss let not turn that in a language war
[06:49] SubStack: right into the core
[06:49] blogometer: If Java is the measure of success and Ellison is the model leader, then I'm sure there's a community that has a direction that is going to fit better with my plans.
[06:49] Aria: jetienne: Not so! Look at CPAN, look at TeX, look at Ruby, even. Things work together, and people don't build such tall silos of their own stack.
[06:49] ryah: blogometer: that's why i think the appendix idea is okay - it somehow elevates some - but in no way degregates others
[06:49] SubStack: php's namespace pollution might even be partly responsible for its success, as difficult as that is to accept
[06:49] Aria: appendix ++
[06:49] ryah: the lonely planet of module systems
[06:50] jetienne: lets focus on how other project handle modules/lib
[06:50] mjr_: I like the idea of this curated collection being a separate project.
[06:50] Aria: Heh, yeah, SubStack. Though I think commonjs modules have most of the positive features of that and not so much of the downsides.
[06:50] jetienne: Aria: lets just say i disagree and focus on node
[06:50] blogometer: ryah: Maybe, just a separate identity.
[06:50] blogometer: Because, I believe it is really valuable for people to see that Node.js is not just another framework.
[06:51] ryah: SubStack: i agree
[06:51] SubStack: require() could pull down missing libs from npm >_>
[06:51] ryah: blogometer: perhaps
[06:51] ryah: i could see howtonode.org making a list of libs
[06:51] Aria: SubStack: Indeed. Or support HTTP with enough caching that it's only slow once.
[06:51] blogometer: There is a lot of value in showing experienced developers that they are not going to be in a framework space that has no creative space.
[06:52] jetienne: ryah: this node.lib could be handled by people you trust
[06:52] jetienne: but good lib + good doc is needed to make ndoe.js mainsream
[06:52] jimt_ has joined the channel
[06:52] jetienne: it is all in how and when
[06:53] blogometer: I still don't understand the problem or the App Store / Ellison inspration.
[06:53] ryah: ellison is just mentioned because - i think a single maintainer can scale
[06:53] Aria: Indeed. If there's nothing that will make them a bottleneck.
[06:53] blogometer: Well, you should split identities.
[06:54] jetienne: but this maintainer will have to test and document all those lib ?
[06:54] Aria: (Linus telling people off loudly and to come back with better ideas works out well -- lots of people who do good stuff under him, but he'll still be the one to bitch people out who need it)
[06:54] jetienne: am i missing something?
[06:54] blogometer: Because the big selling point for Node.js at this point, is that it is bound so tightly to async I/O.
[06:55] Aria: jetienne: Tests should be automated, and docs should be included. Or people yelled at until they make some.
[06:55] mjr_: tests and docs for modules doesn't really address the "abandoned module" issue
[06:55] blogometer: jetienne: I'm all for a preferred Node.js build system.
[06:55] ryah: jetienne: maybe the node.modules project requires people to do their docs in markdown
[06:55] mjr_: or the overall quality issue.
[06:55] Somebi has joined the channel
[06:55] ryah: jetienne: in order for projects to get into node.modules, they have to provide docs in markdown.
[06:55] ryah: then we can generate a website with them
[06:55] blogometer: I'd like to have a way to do coverage and testing of both C/C++ and Javascript.
[06:56] Aria: Well, for overall quality, treat bugs against the individual things as bugs against the whole lib module.
[06:56] ryah: i worry about too much automation
[06:56] blogometer: ryah: Which is why you can tell them to just use the GitHub wikis.
[06:56] jetienne: Aria: node.lib should be a community. aka i know how to write doc, you know how to write code, i do you doc
[06:56] blogometer: Which are all markdown and versioned.
[06:56] Aria: jetienne: There's nothing that would ever stop that.
[06:56] blogometer: (Who wants to write doc for someone else's code.)
[06:56] jetienne: Aria: yep but i think node.lib should help them
[06:56] Aria: (I like writing API docs, if the barrier to entry is low.)
[06:57] Aria: jetienne: 'help' how?
[06:57] blogometer: ACTION Note to self. Befriend Aria.
[06:57] jetienne: Aria: puting them in contact, pointing that this module has no doc, so people collaborate for a better node.lib
[06:57] SubStack: I like examples more than docs
[06:57] SubStack: so long as there are enough of them
[06:57] Somebi: good morning to everyone ^_^
[06:57] Aria: ACTION nods. Indeed -- tools and people who do that, both are good, jetienne.
[06:57] blogometer: I like the five minute intro.
[06:58] ryah: ive been told that drupel is has the best community ever
[06:58] ryah: and it's largely because of its website
[06:58] Aria: I like docs more than examples, at least as far as reference goes. but getting into something from docs alone is hard.
[06:58] blogometer: Nice sites:
[06:58] blogometer: http://jashkenas.github.com/coffee-script/
[06:58] ryah: or whatever that project is called...
[06:58] ryah: the cms
[06:58] mjr_: and maybe that the point of drupal is make collaboration tools. :)
[06:58] jetienne: drupal
[06:59] blogometer: http://yardoc.org/
[06:59] SubStack: I should get around to writing docs for my modules
[06:59] amerine has joined the channel
[06:59] bpot has joined the channel
[07:00] Somebi: heard about coffee
[07:00] Somebi: how is it going?
[07:00] blogometer: http://g.raphaeljs.com/
[07:00] ryah: the problem with trusted people, is not that i don't trust them
[07:00] jetienne: Aria: node.lib should try to build a common lib for node.core. a common lib = feature coverage, good doc, good testing etc...
[07:00] ryah: it's that im the only one here paid to work on node - other people have less time
[07:01] Aria: jetienne: Of course. (feature coverage being the only part people would debate about that.)
[07:01] jetienne: there is no reason node.js not to have a library of ruby/python quality. this is more "how to reach this"
[07:01] ryah: i think a bit of "blessing" is required...
[07:01] ryah: mroe so than in jquery
[07:02] jetienne: not up to you doing it all
[07:02] Somebi: yea good doc with examples will help project to grow
[07:02] Somebi: +1
[07:02] Aria: Yeah. Picking best-of-breed projects -- and more so, ones that play nice together -- and blessing them.
[07:02] ryah: and that work with a particular version of node...
[07:03] jetienne: to highlight good lib could be a first step
[07:03] Aria: .oO(Can you plumb, say, a POP3 library and SMTP library together in a few lines of code, and send all messages in a mailbox to someone? That kind of play together)
[07:03] Aria: Yeah. Keeping a lib roughly in sync with node is important.
[07:03] Aria: (I'd love to see a lib module that has the same version numbering)
[07:03] MikhX has joined the channel
[07:04] jetienne: what about something like bsd kernel ?
[07:04] ryah: there's a lot of merit in drupal like site
[07:04] Aria: Where the kernel and userland are synced, ish?
[07:04] Aria: (node-lib 0.2.0 works with node 0.2.0, etc.)
[07:04] ryah: but i don't think we've got the gas to build it
[07:04] jetienne: like there are the people, and the "core team"
[07:04] ryah: not yet anyway
[07:05] jetienne: ryah: what about howtonode to make a "libraries of the week" ?
[07:05] Aria: Yeah. One thing to avoid is becoming gnustep.{org,it,com} ... just so much website fragmentation, everyone being almost-the-official-site.
[07:05] jetienne: and see how it goes for a few months
[07:05] blogometer: Oh, man. Drupal is nausea.
[07:05] Aria: blogometer: The community or the actual code?
[07:06] blogometer: I just associate it with a lot of people who can't do more than unzip Drupal.
[07:06] konobi: i'm a fan of the private wiki that gets generated into a website kinda thing
[07:06] Tim_Smart: ryah: What method do you use to check if a object has a property in C++>
[07:06] Tim_Smart: ?
[07:06] mjr_: I think node has the best community of any project I've seen.
[07:06] mjr_: It's a big reason why I wanted to get involved.
[07:06] Aria: Me too. It reminds me of the early days of ruby.
[07:06] Tim_Smart: !object->Get('prop')?
[07:06] konobi: Tim_Smart: object->Get('property')
[07:06] Aria: Like pre-rails Ruby.
[07:06] blogometer: mjr_: yes.
[07:06] Aria: Me too.
[07:06] blogometer: I agree.
[07:06] blogometer: And all of this sounds really awful.
[07:06] mjr_: So I dunno, I think we are doing a pretty good job with the chaos.
[07:06] ryah: Tim_Smart: obj->Has(String::New("blah"))
[07:07] konobi: hhhmmm... good to know
[07:07] ryah: blogometer: :)
[07:07] mjr_: The API used to change about once a week anyway, so it's been kind of hard to expect a ton of quality for people contributing libraries.
[07:07] Tim_Smart: ryah: So why do you have to chuck in a v8 string?
[07:07] ryah: Tim_Smart: that's how it is
[07:07] blogometer: Like Apache. An incredible collection of really bloated committee driven software.
[07:07] Tim_Smart: hmm
[07:07] blogometer: Why not use Linux as a model.
[07:07] Tim_Smart: ryah: You don't have to for Get?
[07:07] blogometer: ?
[07:07] ryah: blogometer: yes
[07:08] konobi: you would have thought it would have cast it
[07:08] ryah: that's what i think too
[07:08] Aria: Linux is a good model as far as it's analogous.
[07:08] ryah: linux is a good model
[07:08] Aria: But Linux also is something else, so watch out where the differences lie.
[07:08] Aria: (It's a single product, the kernel)
[07:08] blogometer: Node.js is more like Linux than it is like Drupal.
[07:08] konobi: freebsd model is also good
[07:08] Aria: Yeah. It really is.
[07:08] ryah: i hate how ruby includes libraries
[07:08] konobi: it's a single product... kernel and userland
[07:09] ryah: like a yaml parser
[07:09] blogometer: Also, Node.js is the next thing.
[07:09] Aria: I like the freebsd structure in some ways: A core kernel (~node...), a userland (... node-lib...) and a community of other stuff that runs on it (everything else.)
[07:09] mjr_: I tried to do Linux networking way back in the "net2" days, I think they called it, and everybody was mean. It sucked.
[07:09] ryah: and like - years down the road you realize that ruby's yaml parser isn't the best
[07:09] mjr_: So hopefully that's not the "linux model"
[07:09] blogometer: A small executable that is easily redistributable with an interpreter that can work with many small objects.
[07:09] jetienne: mjr_: this "you are not part of the group" came later, early 2000
[07:10] blogometer: Node.js and Google V8 are evolutions.
[07:10] ryah: blogometer: no fear for the size of node itself - it's not going to get big
[07:10] Aria: ryah: Yeah. And, because the core devs don't speak English, there's huge NIH syndrome going on, everyone just starts using an addon library with a slightly different API, and the one included with Ruby rots.
[07:10] blogometer: ryah: Okay.
[07:10] ryah: blogometer: or - acutally do fear - it keeps me honest
[07:10] blogometer: ryah: Okay.
[07:10] ryah: but i have no intention of letting this thing grow much more
[07:10] konobi: more primitives, less abstractions
[07:11] blogometer: ryah: If you want to think like Ellison, that's fine, but think like Stallman.
[07:11] jetienne: ryah: do you have an explicit limit ?
[07:11] blogometer: A tiny performant, redistribuatble, scripted server platform.
[07:11] ryah: jetienne: i want to have explicit startup time contraints
[07:11] blogometer: It shall be a destroryer of worlds.
[07:11] ryah: jetienne: startup size, etc
[07:11] ryah: but i dont have the CI stuff to measure that yet
[07:11] jetienne: ok
[07:11] konobi: we'll get there
[07:12] ryah: anyway
[07:12] blogometer: Anyone written a mail server in Node.js yet?
[07:12] ryah: appendix doesn't add to bloat
[07:12] ryah: it just needs to be maintained
[07:12] ryah: it would be silly if it was just a list of what i thought was cool in august 2010
[07:12] SamuraiJack has joined the channel
[07:13] Aria: blogometer: funny you should ask that... http://github.com/aredridel/node-smtp/
[07:13] Aria: Yeah, indeed, ryah.
[07:13] blogometer: bad link.
[07:13] ryah: with that. i write an appendix
[07:13] Aria: Oh, bah.
[07:13] blogometer: Aria: LOL.
[07:13] _announcer: Twitter: "The-d Pushonhoshiikedoarunokana node.js us" [ja] -- Yohei Sasaki. http://twitter.com/yssk22/status/21731180596
[07:13] blogometer: Node.js code like that fills my heart with joy.
[07:13] Aria: Oh geez. I renamed the project twice and they applied in reverse order.
[07:14] jetienne: ok it seems clear that current community is showing its limit. i think it should become more organized
[07:14] blogometer: It is just sick and wrong what you can do in 108 lines of Node.js.
[07:14] Aria: try now, blogometer
[07:14] blogometer: I'm reading it.
[07:15] konobi: Aria: mmm... pretty basic... but kinda cool actually
[07:15] Aria: Yeah. This is what I did in the past, oh, two hours. Mostly involving writing a line-buffering stream.
[07:16] Tim_Smart: Does this seem right? https://gist.github.com/50c0501d3b67e018c1ee
[07:16] blogometer: Aria: Looks great. Just fill in the blanks.
[07:16] blogometer: I'd written the start of SMTP and IMAP in Java.
[07:16] Tim_Smart: I have a feeling I'm doing too much.
[07:16] blogometer: To learn NIO.
[07:16] Aria: Yeah. I'm about to make it emit a bunch of events, blogometer, but the rough shape is coming together now.
[07:17] blogometer: It was thousands of lines of code that didn't do as much as this does.
[07:17] Aria: Tim_Smart: That's about how V8's code looks internally. Looks about right.
[07:17] konobi: Aria... then the fun with timing, scanning, auth, delays, greylisting, etc.
[07:17] Aria: konobi: http://aria.blogs.theinternetco.net/ ;-)
[07:17] blogometer: Actually, I'm really curious to see Node.js do SMTP.
[07:17] Tim_Smart: Aria: OK, well I'll continue with the patch and let ryah review it later.
[07:18] blogometer: Because then the community will get wise to security in a big way.
[07:18] Aria: Hehe, yeah.
[07:18] konobi: Aria: ah... heh... i'm an old mail security bod
[07:18] Tim_Smart: Aria: I'm not sure if line 10 will work as I expect it to.
[07:18] jetienne: going for mac packaging
[07:18] Aria: Yeah, not sure how that works, Tim_Smart.
[07:19] jetienne has joined the channel
[07:19] benburkert has joined the channel
[07:20] zomgbie has joined the channel
[07:20] Aria: Now, I think http://github.com/aredridel/ruby-linebuffer/blob/master/lib/linebuffer.js is ugly, and it's making me wonder what can be done to make stream objects simpler to implement.
[07:20] blogometer: Every day now, this Node.js thing is blowing my mind, the community is constantly cranking out terse, performant code, that I can read and understand and patch if it breaks.
[07:21] blogometer: I do not want to see this community change.
[07:21] Aria: Hehe. Just gotta enshrine the parts that rock.
[07:21] Aria: And keep those patches flowing. Even if the original maintainers screw it up.
[07:21] blogometer: Aria: I want smtp because I want to build a mailserver with moderation.
[07:21] jetienne: ryah: email linus
[07:21] blogometer: For group discussions.
[07:22] Aria: Huh. Wow, and you'd do that at the SMTP layer? Rather than further down a mail processing chain?
[07:22] blogometer: You'd need an SMTP layer to start, though.
[07:22] Aria: Yeah, but those /do/ exist ;-)
[07:22] Aria: exim, postfix, etc.
[07:23] blogometer: Yeah, but the wheel that your reinventing will be *rounder*. I just know it.
[07:23] Aria: ACTION laughs.
[07:23] Aria: It /is/ amazing how many sharp corners mail servers grow.
[07:23] konobi: ACTION never wants to see SMTP again
[07:23] blogometer: I'd like to see SMTP that I can read and understand.
[07:23] Aria: SMTP's a spiffy protocol. Problem is, the people who use it...
[07:23] bdha: SMTP is awful.
[07:24] Aria: SMTP is the original MQ ;-)
[07:24] konobi: that protocol is a PITA, not because of the protocol, but because of how people implement it
[07:24] blogometer: Like with postfix, I don't need to deliever to the local filesystem. It can all be in memory or in a database.
[07:24] bdha: konobi: It's not the people who implement it... it's the specs they have to implement.
[07:24] Aria: Remind me to send you a 2GB email sometime.
[07:24] konobi: then atop that you have MIME which is a PITA
[07:24] Aria: bdha: like which?
[07:24] konobi: multi-part mime at that
[07:24] blogometer: It's all pretty easy to understand.
[07:24] bdha: Aria: http://video.google.com/videoplay?docid=7054401183589794595#
[07:24] Aria: ACTION laughs. I kinda like mime, too. Even if boundaries are a pita slightly.
[07:24] bdha: Enjoy. :)
[07:24] ryah: what's the best couch binding?
[07:25] blogometer: The hard part has always been delivery mail as the UNIX user.
[07:25] blogometer: And running his scripts.
[07:25] blogometer: ryah: Check you appendix. Don't ask us.
[07:25] jetienne: :)
[07:25] blogometer: ACTION Goes to room and slams door.
[07:26] ryah: :)
[07:26] Aria: ACTION smirks at bdha. Oddly, none of that has ever bugged me.
[07:26] bdha: Aria: Well, ...yay?
[07:27] bdha: Having worked at an email company (Pobox) since 2006, I can say: Email sucks.
[07:27] bdha: How people use email sucks.
[07:27] Aria: Hehe.
[07:27] Aria: That last part especially.
[07:27] bdha: Unfortunately Wave didn't replace email, so.
[07:27] bdha: I guess it's here to stay. :P
[07:27] Aria: (and yay Pobox. All the discussions I've had with people who've worked there have been super smart and sensible.)
[07:28] CIA-77: node: 03Ryan Dahl 07master * r94d1708 10/ (5 files in 3 dirs): Add appendix to docs - http://bit.ly/d9iD99
[07:28] bdha: Thanks, we try. ;)
[07:28] bdha: ACTION is the SA, for whatever that's worth.
[07:28] ryah: oops
[07:28] konobi: bdha: ah... i know a few of the pobox crew former and current
[07:29] blogometer: ACTION Resolves to be in the appendix.
[07:30] bdha: blogometer: Better than _being_ the appendix. Then you just go bad and get removed. Like in Evil Dead.
[07:30] bdha: konobi: Wel. hai fren. :)
[07:30] CIA-77: node: 03Ryan Dahl 07master * r33e919a 10/ doc/api.markdown : Add appendix to docs - http://bit.ly/9IVe94
[07:30] konobi: mainly the perl bods
[07:31] bdha: konobi: We're all Perl people to some extent. :)
[07:31] blogometer: Wave?
[07:31] bdha: blogometer: Google Wave.
[07:31] bdha: It was this thing, right? This thing that did some undefinable stuff, but it was totally awesome, somehow.
[07:31] bdha: ACTION waves hands around.
[07:31] blogometer: I know. It was awful.
[07:31] bdha: Yes, I know.
[07:31] bdha: That "unfortunately" was sarcasm. :)
[07:32] blogometer: Sarcasim doesn't come thorugh IRC to good.
[07:32] blogometer: s/to/too/
[07:32] blogometer: How much money did Google spend to get that week's worth of attention?
[07:33] bdha: Better question: Did their bottom line notice?
[07:33] bdha: Answer: Haha.
[07:33] Tim_Smart: ryah: Should I allow both normal Buffers and FastBuffers?
[07:33] blogometer: bdha: LOL.
[07:33] sveimac has joined the channel
[07:33] ryah: Tim_Smart: no
[07:33] ryah: Tim_Smart: old buffers houdl be renamed to
[07:33] blogometer: ACTION Busy deleting all the Java projects from his GitHub account (github.com/bigeasy).
[07:33] ryah: SlowBuffers
[07:34] ryah: and the FastBuffers should become just Buffer
[07:34] Tim_Smart: OK, and SlowBuffers are only accessible through process.binding?
[07:34] ryah: yeah
[07:34] Tim_Smart: Sounds good to me.
[07:36] Tim_Smart: args[0]->IsObject() && args[0]->ToObject()->Has(String::New("parent")) && Buffer::HasInstance(args[0]->ToObject()->Get(String::New("parent")))
[07:36] konobi: would be nice to standardize a little on how to run tests, so that we can do a "fastbuffers" branch then run a whole bunch of libraries tests through it
[07:36] Tim_Smart: Is that too much validation, or is it fine?
[07:37] ryah: Tim_Smart: *shrug*
[07:37] ryah: i prefer to take it easy on the type checking
[07:37] Tim_Smart: OK. I'll just mess around with it for a while.
[07:38] ryah: Tim_Smart: note that that c++ buffer class cannot change interface
[07:38] Tim_Smart: Yup.
[07:38] ryah: cool
[07:38] Tim_Smart: ryah: Oh wait, I can just use the isBuffer function right?
[07:39] ryah: you can use it for slowbuffers
[07:39] ryah: hm
[07:39] ryah: you know that c++ buffer class
[07:39] Tim_Smart: Ya
[07:39] ryah: that should be renmaed to SlowBuffer
[07:39] ryah: also be a c++ class
[07:40] Tim_Smart: OK.
[07:40] ryah: bt then a new c++ Buffer class should be made
[07:40] ryah: which is implemented new
[07:40] ryah: just as a shim for the old APIs
[07:40] ryah: the SlowBuffer class should be private to node_buffer.cc and defined there.
[07:40] Tim_Smart: http://github.com/Tim-Smart/node/commit/a5972040387ee38f77e75190c2e19de55fd07b7a#L0R162
[07:40] ryah: i think
[07:41] Tim_Smart: ryah: I thought you wanted fast buffers JS only?
[07:41] ryah: this is becoming a bit too much for me to think about without acutally doing it
[07:41] ryah: so far that looks fine
[07:41] ryah: you should %s/FastBuffer/Buffer/
[07:41] jsilver has joined the channel
[07:41] ryah: but dealin with the c++ side is going to be tricky.
[07:42] ryah: forget what i just said about the SlowBuffer c++ class..
[07:42] ryah: just keep it Buffer
[07:42] Tim_Smart: Yup. I think the C++ should stay as is, just need to change how the function handle the buffer argument,
[07:42] ryah: Buffer::New will just return a Buffer as it always did
[07:42] ryah: it'll just be a slow buffer
[07:42] ryah: we can add a new function for allocing a FastBuffer in c++ later
[07:43] Tim_Smart: functions*
[07:43] jsilver: hi guys++
[07:43] Tim_Smart: Anyway. Dinner time for me.
[07:43] ryah: Tim_Smart: bed for me
[07:43] ryah: ttyl
[07:43] ryah: hi jsilver
[07:43] jsilver: i BedDinner
[07:43] jsilver: hi
[07:43] ryah: ACTION away
[07:43] jsilver: bye
[07:45] blogometer: Actually, I'm going to suggest a knitting circle.
[07:45] blogometer: I'd like to build some stuff in Node.js.
[07:45] blogometer: And I'd like to learn how to build performant servers quickly.
[07:46] blogometer: And the Javascript idioms.
[07:46] blogometer: So, maybe, community wise, there's some way to establish a Node.js stylebook.
[07:52] jsilver: does anyone wanna have randy node js time with me yeah>
[07:52] jsilver: blogometer, google: DOUGLAS CROCKFORD js videos
[07:52] jsilver: wil not disappointed
[07:53] mattikus has joined the channel
[07:53] blogometer: Noted.
[07:54] joshbuddy has joined the channel
[08:01] felixge has joined the channel
[08:01] felixge has joined the channel
[08:01] twoism has joined the channel
[08:01] pdelgallego has joined the channel
[08:02] rnewson has joined the channel
[08:04] astrolin has joined the channel
[08:05] Aria: Night, all.
[08:15] FND has joined the channel
[08:16] FND: hi - `echo "var i = 1;" | node` raises ENOTTY - where can I find documentation on how to do this right?
[08:18] FND: same for `node < foo.js`
[08:18] halorgium: FND: by default node loads the repl when invoked with no args
[08:19] halorgium: i don't believe you can specify to read from stdin
[08:19] FND: I've looked at node --help, but there aren't many options there
[08:19] FND: I'll use a temporary file then, I guess
[08:19] zomgbie has joined the channel
[08:20] FND: is there a way to specify multiple files? as in `node foo.js bar.js`?
[08:20] rnewson has joined the channel
[08:20] rnewson has joined the channel
[08:20] halorgium: nope
[08:21] FND: I'm asking is because I used to use Spidermonkey as terminal-based JS engine, so now learning about the slight differences
[08:21] FND: s/is //
[08:21] dnolen has joined the channel
[08:21] FND: so I'd have to concatenate foo.js and bar.js into a temporary file before feeding it to Node?
[08:21] FND: actually
[08:21] FND: I could create a temporary file with require("foo"
[08:22] FND: ACTION tests
[08:22] halorgium: remember that each file has its own scope
[08:22] FND: yeah, this might be trickier than I expected
[08:23] FND: I wrote a few JavaScript shims to make QUnit run on Spidermonkey, now trying to get this to work on Node
[08:24] CIA-77: node: 03Ryan Dahl 07master * r488aff0 10/ (5 files in 3 dirs): Improve appendix markdown - http://bit.ly/cRRSyi
[08:35] FND: the dispatch.js with require("foo"); require("bar"); works - also allows me to deal with variable scope
[08:37] pdelgallego: Hi what is a good node.js project to do in a weekend? I already have done a chat.
[08:44] hellp has joined the channel
[08:45] sveimac has joined the channel
[08:49] jimmy203 has joined the channel
[08:51] FND: `node --help` mentions `node --v8-options` -- but how do I invoke V8 options?
[08:52] jimmy203: Whats the easiest wat to get the property name from a json object?
[08:52] jimmy203: way*
[08:53] frza has joined the channel
[08:54] femtoo has joined the channel
[08:55] jimmy203_ has joined the channel
[09:05] _announcer: Twitter: "Perhaps because of libevent libev trumpet, node-js, perhaps because of, node-dht is probably because of one. Well not libevent, libev going to be enough or do I support" [ja] -- Yuuki Takano. http://twitter.com/ytakanoster/status/21735921752
[09:05] hzin has joined the channel
[09:05] sveimac has joined the channel
[09:14] matthewford has joined the channel
[09:16] Throlkim has joined the channel
[09:25] SamuraiJack_ has joined the channel
[09:26] claudiu__ has joined the channel
[09:39] mattly_ has joined the channel
[09:40] _announcer: Twitter: "@ryah let's hear it for RPC/RMI !! Who needs REST APIs anyway??? ;-) /me goes back to building and RMI framework with node.js...." -- Michael Bradley. http://twitter.com/michaelsbradley/status/21737382988
[09:44] matthewford has joined the channel
[09:50] mAritz has joined the channel
[09:55] felixge has joined the channel
[09:55] felixge has joined the channel
[09:57] robotarmy has joined the channel
[09:57] Tim_Smart: Does somewhere here know how to call a function on a object from C++?
[10:00] rnewson has joined the channel
[10:00] rnewson has joined the channel
[10:04] _announcer: Twitter: "JSSP - JavaScript Server Pages Engine on top of NodeJS (Alligator): http://wp.me/puFuL-2t" -- MrOhad. http://twitter.com/MrOhad/status/21738366784
[10:04] aho has joined the channel
[10:12] ay has joined the channel
[10:13] ay: Just a quick question. From a old Perl-coder... Using http.client. What is the prefered way to extract the title of the document you retrive...
[10:15] Tim_Smart: ay: Regex or parse with DOM.
[10:16] Tim_Smart: s/with/into/
[10:18] ay: Ok. Regexp is easy. But does node.js provide som cool tools to dive into the dom or is plain javascript parsing..
[10:18] tobiassjosten has joined the channel
[10:19] felixge: ay: node does not provide html parsing, but there are people working on libraries for it
[10:19] tobiassjosten: Anyone here using mongoose? I need some practical examples on how to use its models.
[10:19] ay: felixge: So regexping the complete document is the easiest way?
[10:19] ay: I'm fine with that.
[10:20] felixge: ay: yeah, I mean this seems like anything else might be overkill
[10:20] felixge: ay: if you need to parse more than just the title tag, I'd look into finding a decent html parsing lib for node
[10:21] ay: Ok. I Will need that after a while.
[10:21] astrolin has joined the channel
[10:21] ay: The examples I've seen so far have all be using exec("curl -e"
[10:21] ay: Instead of http.client...
[10:22] felixge: ay: well, the http client is a little low level. You have to take care of redirects yourself for example
[10:23] ay: Ok.
[10:23] felixge: ay: http://github.com/silentrob/Apricot this looks really cool
[10:24] ay: felixge: Cool. Thank You
[10:24] felixge: ay: in fact, it looks so cool, I need to play with it myself for a minute :)
[10:31] bradleymeck has joined the channel
[10:32] felixge: ay: man, that lib is insanely cool :)
[10:33] _announcer: Twitter: "Apricot - an insanely cool html scraping tool for #nodejs. http://bit.ly/bQUTgk" -- Felix Geisendörfer. http://twitter.com/felixge/status/21739597420
[10:36] jetienne: felixge: you got aware of it from ryah recomendation in the doc ?
[10:36] felixge: jetienne: no, just looked for a lib like that because ay asked
[10:36] jetienne: ok
[10:37] felixge: jetienne: where does ryah recommend it?
[10:37] jetienne: felixge: a new part of the document is "good libs". in order to stop the big mess of unmaintained code
[10:38] felixge: jetienne: is that published yet?
[10:38] jetienne: http://github.com/ry/node/commit/33e919ae6f365c81f345f70e2fdb406a2268756a
[10:38] felixge: jetienne: ah ok
[10:38] felixge: jetienne: yeah, that's not live yet
[10:39] felixge: that's a strange patch btw.
[10:39] felixge: looks like ryan commited some other stuff along with it
[10:40] _announcer: Twitter: "http://tinyurl.com/ylkfkeu Deploying Node.js With Upstart and Monit - How To Node" -- catch_fish. http://twitter.com/catch_fish/status/21739874409
[10:41] sztanpet has joined the channel
[10:45] rubydiamond: Anybody using node-oauth?
[10:47] rubydiamond: does node-oauth support Oauth 1.0a ?
[10:49] jkreeftmeijer_ has joined the channel
[10:50] jkreeftmeijer_: Hi! Did anyone here get Socket.IO working with the iPhone?
[10:50] jkreeftmeijer_: It connects (using xhr-polling), but disconnects after a couple of seconds:
[10:50] jkreeftmeijer_: 21 Aug 10:37:49 - Initializing client with transport "xhr-polling"
[10:50] jkreeftmeijer_: 21 Aug 10:37:49 - Client 5455555797088891 connected
[10:50] jkreeftmeijer_: 21 Aug 10:37:57 - Client 5455555797088891 disconnected
[10:51] rubydiamond: jkreeftmeijer_: hey
[10:52] jkreeftmeijer_: hey rubydiamond :)
[10:52] rubydiamond: jkreeftmeijer_: did you use node-oauth?
[10:52] jkreeftmeijer_: Eh, no.
[10:53] rubydiamond: okay
[10:58] jkreeftmeijer_: hmm... the chatbox example seems to work...
[10:58] jkreeftmeijer_: Anything I should keep in mind? :)
[11:00] _announcer: Twitter: "Have to look at Heroku nodejs docs for @node_knockout . Time passes :)" -- Kadir Pekel. http://twitter.com/kadirpekel/status/21740735359
[11:02] sveimac has joined the channel
[11:03] AAA_awright_ has joined the channel
[11:03] sveimac has joined the channel
[11:03] femtoo has joined the channel
[11:04] sveimac_ has joined the channel
[11:05] _announcer: Twitter: "0.2.0 Вышел nodejs nodejs http://nodejs.org/ # # download" [pl] -- bnetteam. http://twitter.com/bnetteam/status/21740946174
[11:07] _announcer: Twitter: "@ Nerdson Node.js What is this?" [pt] -- Laercio Motta. http://twitter.com/laerciomasala/status/21741051238
[11:14] tuxsbro_ has joined the channel
[11:15] _announcer: Twitter: "Since the dawn of the server configuration I'm looking for a netty documents, sometimes for a taste if you need to configure a server that thought out there. It seems to consist node.js interesting ... It's also a time to study." [ko] -- Qoo. http://twitter.com/garangnip/status/21741370226
[11:19] sveimac_ has joined the channel
[11:22] sveimac_ has joined the channel
[11:25] feroz_ has joined the channel
[11:27] mertimor has joined the channel
[11:31] rubydiamond: Getting this error http://pastie.org/1105783 for npm install macaddr ... am I missing something here? I am using npm version 0.1.25 and node version v0.2.0
[11:31] rubydiamond: ryah: you there?
[11:31] jashkenas has joined the channel
[11:33] Tim_Smart: rubydiamond: macaddr isn't in the registry
[11:34] rubydiamond: Tim_Smart: okay thanks
[11:37] Tim_Smart: Grr... Does someone around here that knows node's C++ side of things, know why this might make node silently exit? http://github.com/Tim-Smart/node/compare/master...fast-buffer
[11:42] thedjinn has joined the channel
[11:52] jacobolus has joined the channel
[11:53] herbySk has joined the channel
[11:54] confoocious has joined the channel
[11:58] caolanm has joined the channel
[11:58] caolanm: hey
[11:59] caolanm: here's an odd one... if I eval a function it is an instanceof Function, if I use Script to run it its not an instanceof Function, but both are typeof 'function'
[12:00] caolanm: sometimes js is strange
[12:03] dipser has joined the channel
[12:04] polotek has joined the channel
[12:06] junkee[] has joined the channel
[12:10] claudiu__ has joined the channel
[12:13] ker2x: Mmm... i'm sure i have a memory leak. and i don't understand how i could possibly have a memory leak in my js code
[12:14] ker2x: it could be in one of the module
[12:14] polotek: ker2x: js can have memory leaks if you're leaving references around after you're done with them
[12:14] polotek: but more likely it's a C addon
[12:15] polotek: what are you doing?
[12:15] ker2x: my js code is simple. but i also use dgram, http, mongodb
[12:15] herbySk: caolanm: if you use it to run in itsd own context, then it's as it should be: each context has its own, distinct "Function"
[12:16] ker2x: it probably is in the mongo driver
[12:16] herbySk: (as well as Object, Number, String, thier prototypes, etc. ... all startup data for js)
[12:16] polotek: never used dgram
[12:17] ker2x: dgram is very easy and simple to use :)
[12:17] ker2x: udp is <3
[12:17] Tim_Smart: herbySk: He left :/
[12:17] Tim_Smart: I was about to answer his question.
[12:18] d0k has joined the channel
[12:18] Tim_Smart: ker2x: node-inspector can take heap snapshots.
[12:18] Tim_Smart: I personally haven't got it working, but it looks cool.
[12:19] nerdEd has joined the channel
[12:19] ker2x: i use --trace_gc
[12:19] stellar has joined the channel
[12:20] ker2x: Scavenge 34.9 -> 31.5 MB, 5 ms.
[12:20] ker2x: Scavenge 36.1 -> 33.2 MB, 8 ms.
[12:20] ker2x: Scavenge 37.8 -> 34.9 MB, 7 ms.
[12:20] ker2x: Scavenge 39.5 -> 36.6 MB, 8 ms.
[12:20] ker2x: Mark-sweep 46.6 -> 30.5 MB, 2 / 89 ms.
[12:20] ker2x: 12h ago it was ~5MB
[12:22] konobi: closures are generally pretty leaky
[12:23] ker2x: humm
[12:23] ker2x: ok
[12:23] konobi: http://www.ibm.com/developerworks/web/library/wa-memleak/
[12:23] ker2x: ha ! thx a lot :)
[12:27] rnewson has joined the channel
[12:30] frza has joined the channel
[12:34] tisba has joined the channel
[12:42] rkieffer has joined the channel
[12:43] feroz_ has joined the channel
[12:50] hsuh has joined the channel
[12:55] _announcer: Twitter: "@ Ba78 you can say a few peldafeladatot node.js to? beleasnam myself." [hu] -- deejayyhu. http://twitter.com/deejayyhu/status/21746324531
[12:55] mAritz: theoretical question: i'd need an idempotent way to generate a sector in 3d-space given the sectors coordinates and side length (cube). what could i use? :D
[12:55] mAritz: as in what kind of algorithm
[12:57] jashkenas: idempotent? generating what exactly?
[12:58] _announcer: Twitter: "@danrough we were going do nodejs but he screwed his nose up at modern garbage collection, wants to "cut code old school like." Obj-C then." -- Tim Macfarlane. http://twitter.com/timmacf/status/21746499595
[12:58] sveimac has joined the channel
[13:00] _announcer: Twitter: "UglifyJS works well for me, it produces a view bytes more then closure compiler, but passes well to nodejs ecosystem" -- Oleg Slobodskoi. http://twitter.com/oleg008/status/21746628138
[13:01] _announcer: Twitter: "Good time spent at @arnabc's place. hacked Node.js a little." -- Aniλ Wadghuλe. http://twitter.com/anildigital/status/21746703809
[13:17] rnewson has left the channel
[13:18] hsuh_ has joined the channel
[13:23] davidwalsh has joined the channel
[13:25] jimmy203_: anyone here good with mongoose?
[13:25] polotek: jimmy203_: never used it. but I'll try to help out
[13:25] jimmy203_: I`m trying to understand how to do an update after a find?
[13:26] polotek: jimmy203_: post some code?
[13:26] jimmy203_: hold on.
[13:27] polotek: unrelated, anybody know about v8 classes with Persistent Handle properties?
[13:27] polotek: if I have a class and one of it's properties is Persistent
[13:28] polotek: will it get cleaned up automatically in the classes destructor
[13:28] polotek: I'm guessing no
[13:29] jashkenas has left the channel
[13:29] jimmy203_: http://pastebin.com/xMD3q0BN
[13:30] Tim_Smart: polotek: I know a little about Persistent scope. "Persistent handles are not held on a stack and are deleted only when you specifically remove them."
[13:30] Tim_Smart: From http://code.google.com/apis/v8/embed.html
[13:30] polotek: Tim_Smart: yeah that's what I'm reading
[13:30] polotek: how would I test this?
[13:30] sideshowcoder has joined the channel
[13:30] polotek: I'm running --trace_gc
[13:30] polotek: but it looks to be stable
[13:31] polotek: well the other thing is that the property isn't always initialized
[13:31] polotek: I guess I should run a test that makes sure to initialize it
[13:31] mAritz: jashkenas: idempotent = given the same input (coordinates and side length) it always has to have the same output. generating a bunch of points in that quadrant.
[13:33] robotarmy has joined the channel
[13:34] jimmy203_: Do I have to retrieve all the ids from find and then use model.update and loop through and update each field?
[13:37] polotek: jimmy203_: I thought you could just update the properties right on the object. and then do obj.save()
[13:37] polotek: is that not working?
[13:38] zomgbie has joined the channel
[13:39] _announcer: Twitter: "String.prototype.format, Date.prototype.strftime such as adding a node.js." [ja] -- Yohei Sasaki. http://twitter.com/yssk22/status/21749001171
[13:40] rkieffer: How do you start the event loop w/out creating a server?
[13:40] jimmy203_: let me try that..
[13:40] rkieffer: e.g. if I run a program that is nothing more than "setTimeout(function() {process.exit();}, 1000);" ... it never exits.
[13:40] Tim_Smart: rkieffer: You don't need to worry about the event loop.
[13:41] rkieffer: Tim_Smart well, I do if I want setTimeout() or nextTick() to work. :)
[13:41] rkieffer: If I prefix that with an http.createServer().listen() call... the timeout gets called as expected.
[13:41] polotek: rkieffer: the event loop starts automatically when node starts
[13:42] polotek: it's not something you need to handle
[13:42] maushu has joined the channel
[13:42] polotek: rkieffer: your example works fine for me
[13:42] Tim_Smart: rkieffer: Tried that code and it exiting after 1 second.
[13:42] rkieffer: polotek: hmm... then I'm confused by why setTimeout isn't getting called.
[13:43] jimmy203_: polotek, that didn`t seem to work :(
[13:43] rkieffer: I'm running latest build.
[13:43] rkieffer: let me git pull just to make sure
[13:44] mau has joined the channel
[13:44] polotek: jimmy203_: what does it do? are you getting errors? you could also check the mongo logs to make sure it's even receiving anything
[13:44] polotek: rkieffer: when I updated to the latest node this morning, I had an issue or 2
[13:44] polotek: I had to do make clean && make distclean
[13:45] polotek: and then ./configure and make again
[13:45] polotek: you might want to try that
[13:45] rkieffer: k, one sec.
[13:46] Tim_Smart: polotek: Do you have any idea why node might silently exit here: http://github.com/Tim-Smart/node/compare/fast-buffer
[13:47] polotek: Tim_Smart: what am I looking for? this is a big diff
[13:47] Tim_Smart: polotek: Good question. I have looked over it a few times and could find any obvious bugs.
[13:48] Tim_Smart: ryah would probably the best person to ask...
[13:48] Tim_Smart: be the*
[13:49] polotek: Tim_Smart: is it segfaulting or just not doing anything?
[13:49] Tim_Smart: Just not doing anything
[13:50] jimmy203_: I`m just getting errors.
[13:50] rkieffer: polotek: jimmy203_ rebuilding did the trick. thx polotek.
[13:51] polotek: jimmy203_: errors from node? post them
[13:52] jimmy203_: no sorry.. I`m still talking about mongoose and trying to save from the find.
[13:52] polotek: Tim_Smart: I'm running to get breakfast. but if you want I'll load it up and help you run some tests when I get back.
[13:52] polotek: jimmy203_: I mean you're getting javascript errors and not errors from mongo
[13:53] Tim_Smart: polotek: I'm about to go to sleep, so no worries.
[13:53] polotek: Tim_Smart: ok
[13:53] Tim_Smart: polotek: I'll just try rebuilding everything once more, then wait until I catch up with ryah again.
[13:54] polotek: back in a bit
[13:54] jimmy203_: I must be doing something simple wrong..
[13:54] maushu has joined the channel
[13:54] Tim_Smart: jimmy203_: Chuck your code on http://pastie.org/ or something.
[13:55] Tim_Smart: Otherwise someone will have to take a really good guess :)
[13:55] jimmy203_: http://pastebin.com/xMD3q0BN
[13:56] jimmy203_: I`m just trying to figure out how to update values :)
[13:56] jimmy203_: after you do a find..
[14:00] hsuh_ has joined the channel
[14:00] Akufen has joined the channel
[14:00] Akufen has joined the channel
[14:03] mau2 has joined the channel
[14:12] icy: is the src of _announcer public?
[14:12] megan: good morning node.js
[14:13] hsuh_ has joined the channel
[14:16] _announcer: Twitter: "Why You Should Pay Attention to Node.Js http://is.gd/esXhS <- old post but valid reasons" -- Hernan Garcia. http://twitter.com/theprogrammer/status/21751453311
[14:16] nerdEd_ has joined the channel
[14:16] Blackguard has joined the channel
[14:21] hsuh_ has joined the channel
[14:29] pbunbun has joined the channel
[14:29] _announcer: Twitter: "uptake while intently at the node.js chirpstream CouchDB evently browser discharge in volume." [ja] -- Yohei Sasaki. http://twitter.com/yssk22/status/21752391622
[14:29] _announcer: Twitter: "@ Epoch__ chirpstream to capture while intently at CouchD http://9zqt8sm.itter.ch/ node.js" [ja] -- みにえぽっく. http://twitter.com/miniepoch__/status/21752391856
[14:30] pbunbun: Yo, have question
[14:31] pbunbun: I have a main script (say "main.js") and it requires "quote.js", I get that main.js can access anything quote.js puts in the export object (and it works)
[14:31] pbunbun: But is there a way for a function in quote.js to access data/functions in main.js?
[14:32] sideshowcoder_ has joined the channel
[14:34] _announcer: Twitter: "Wow. Implementing a node.js (async) version of Ruby's Dir.glob (http://bit.ly/byotsO) is hard." -- Tobie Langel. http://twitter.com/tobie/status/21752756620
[14:37] junkee[] has left the channel
[14:39] zomgbie has joined the channel
[14:40] mscdex: pbunbun: no
[14:40] mscdex: pbunbun: not unless you pass something in to one of your exported functions
[14:44] pbunbun: Ah, I'll consider that then
[14:44] polotek: pbunbun: that's the cleaner way to do it anyway. the point of modules is to avoid mixing scopes
[14:44] polotek: well one of the points anyway
[14:45] megan: I want my node app to run on port 80, but that means only root can start the process right???
[14:45] polotek: megan: correct, node needs root privileges to bind to low ports
[14:45] Tim_Smart: polotek: What is the best way to cast a v8 string to a c++ string?
[14:45] polotek: Tim_Smart: use Utf8Value
[14:45] megan: but polotek, I am using an svn hook to check it out and then I want to run it
[14:45] polotek: lemme find and example
[14:45] Tim_Smart: ok
[14:45] polotek: megan: use sudo
[14:46] megan: okay, but won't that prompt for a password?
[14:46] megan: I don't even have sudo ;D
[14:46] megan: so there
[14:47] sideshowcoder_ has joined the channel
[14:47] polotek: Tim_Smart: http://github.com/polotek/libxmljs/blob/master/src/xml_element.cc#L29-30
[14:47] polotek: so basically you take a handle and pass it in to construct a Utf8Value
[14:47] polotek: Then when you dereference that Utf8Value it'll give you the underlying char *
[14:48] polotek: but it's const so if you want to keep it, you'll have to copy it
[14:48] Tim_Smart: OK thanks.
[14:48] herbySk: megan: chown root, chmod u+s ?
[14:50] jMCg: brrrrr
[14:51] jMCg: megan: http://yaws.hyber.org/privbind.yaws
[14:53] rnewson has joined the channel
[14:55] polotek: jMCg: that's cool
[14:55] polotek: usually what happens is you start as root and then the program explicitly downgrades it's user privileges
[14:55] polotek: that's what apache does right?
[14:55] polotek: I'm sure i'm not understanding it entirely correctly
[14:56] polotek: But I do like the idea of granting the non-root user only the specific privileges it needs
[14:57] jMCg: polotek: take a look at mongrel2 daemonization code to get the idea. It's really.. simple.. Once you figure it out and get it right ;)
[14:57] sideshowcoder_ has joined the channel
[14:57] pgriess has joined the channel
[14:57] jacobolus has joined the channel
[14:58] polotek: jMCg: word
[15:04] jacobolus has joined the channel
[15:07] sideshowcoder has joined the channel
[15:07] nerdEd has joined the channel
[15:09] jMCg: Then again.. mongrel2 is tracked in fossil scm, and actually looking at the code would require more innitial effort than most people can bother to put in
[15:10] mscdex: is it called fossil because it uses outdated concepts/technology? :p
[15:12] _announcer: Twitter: "Playing with NodeJS today. Attempting to build a better chat server for one of my the projects I am working on!" -- Jamund Ferguson. http://twitter.com/xjamundx/status/21755570125
[15:12] jMCg: http://mongrel2.org/artifact/d52f279001bc435282fa75df9424c575f33173c6
[15:12] jMCg: I don't know or care. I try not to get involved into wars about software, editors, OSes, SCMs, because it just makes me think of genozide.
[15:14] c4milo has joined the channel
[15:16] nerdEd_ has joined the channel
[15:21] rubydiamond: We can haz Ship It! Node http://ship-it-node.heroku.com/
[15:22] mscdex: :S
[15:26] mscdex: jMCg: have you tried mongrel2's websocket support?
[15:28] mscdex: hrm... reading about websocket support in their docs leads me to believe it's not implemented
[15:29] c4milo has left the channel
[15:30] Tim_Smart: polotek: OK well the fast-buffers branch can actually run stuff, but it needs a bit of work.
[15:31] polotek: Tim_Smart: so I took a closer look
[15:31] Tim_Smart: yeah buffer.js had some silly mistakes.
[15:31] polotek: are you basically using the pixel functions to store the data because it's faster to store and retrieve them?
[15:32] polotek: I'm not sure I understand why fast_buffer is faster
[15:32] softdrink has joined the channel
[15:32] Tim_Smart: polotek: Ryah implemented the conceptual magic.
[15:32] Tim_Smart: Basically it uses the old buffer as a memory pool for small buffers.
[15:33] rubydiamond: guys Ship It! Node style!
[15:33] rubydiamond: http://shipitnode.heroku.com
[15:33] Tim_Smart: So you aren't doing as many Buffer creations as you usually would.
[15:34] rubydiamond: is it possible to stream image data
[15:34] rubydiamond: instead of waiting
[15:34] polotek: i see. so it's the constant buffer allocation that is the main speed drag?
[15:34] rubydiamond: basically create a websocket and stream the data to website
[15:34] polotek: rubydiamond: yes, but only if you stop spamming the channel with ship it links
[15:34] rubydiamond: want it for Ship it node app
[15:34] polotek: :P
[15:34] rubydiamond: polotek: ah sorry
[15:35] Tim_Smart: polotek: ObjectWrap is the main speed drag - something to do with v8's external stuff.
[15:35] polotek: rubydiamond: you can stream binary image data just like anything else
[15:35] Tim_Smart: the malloc wasn't half bad
[15:35] rubydiamond: polotek: I am assuming it browser will render it runtime
[15:35] polotek: I admit I don't know how difficult it would be to assemble it and get a an img tag in the client
[15:36] polotek: rubydiamond: if you want it rendered into an
![]()
tag then you don't need websockets
[15:36] rubydiamond: polotek: basically I want it to stream
[15:36] polotek: just load the url in your im age tag
[15:36] rubydiamond: than getting the file
[15:36] rubydiamond: polotek: but will it stream ?
[15:37] pr0cy0n has joined the channel
[15:37] Tim_Smart: rubydiamond: Yes it will. Just not over websockets.
[15:37] Aria has joined the channel
[15:37] polotek: rubydiamond: have you tried it. I don't know if img tags handle chunked data
[15:37] Tim_Smart: Or whatever you are doing.
[15:37] polotek: never tried it
[15:37] polotek: but yes, node will stream it
[15:37] [[zz]] has joined the channel
[15:37] Tim_Smart: polotek: Considering all the static file servers in node usually use chunked, it will be fine.
[15:37] rubydiamond: polotek: you mean if I put filepath in img tag.. node will stream ?
[15:38] megan: should I be running my node app as root?
[15:38] Tim_Smart: rubydiamond: streaming is up to the server, not the client.
[15:38] polotek: rubydiamond: as long as you don't set the content-length and set transfer-encoding: chunked
[15:38] polotek: it will stream
[15:38] rubydiamond: If I put base64 encoded data in img tag..
[15:38] Tim_Smart: megan: Try not too...
[15:38] megan: but then I don't get port 80
[15:38] rubydiamond: will node.js stream it or will it do something else?
[15:39] herbySk: megan: fo you really need 80? wouldn't 8080 or 8008 be enough?
[15:39] polotek: rubydiamond: it doesn't matter what kind of data you send
[15:39] Tim_Smart: megan: OK run as root, then drop privileges as soon as you are bound to the port.
[15:39] megan: no herbysk, not for the web
[15:39] polotek: megan: you should plan to put node behind a proxy anyway
[15:39] herbySk: megan: I see
[15:39] Aria: Yeah, drop privileges.
[15:39] Tim_Smart: megan: http://nodejs.org/api.html#process-setuid-68
[15:39] Aria: Or, if you want, adjust your server to allow it without being root, if you're running Linux.
[15:40] megan: well how does apache do it??
[15:40] Aria: Just like that.
[15:40] megan: I want it to come up with the server
[15:40] Aria: Well, THAT is controlled by a startup script. Usually SysV under linux, varies though.
[15:40] Aria: Launchd on macos.
[15:41] megan: I will put it in init.d
[15:41] Aria: Usually that starts it as root, and the daemon itself setuids.
[15:41] megan: that sounds like a good idea
[15:41] Aria: (and fork, setsid, forks and setgid...)
[15:42] megan: now to figure out how to get svn hooks to be able to restart the daemon
[15:43] dylang has joined the channel
[15:45] rubydiamond: 'Transfer-Encoding': 'chunked' ? is this right?
[15:45] Aria: Interesting choice.
[15:45] jashkenas has joined the channel
[15:46] polotek: rubydiamond: write some test code, and if it doesn't work, post it here in a gist or pastebin so we can take a look
[15:46] rubydiamond: http://shipitnode.heroku.com/ now supports chunked response
[15:47] polotek: rubydiamond: ah, you were trying to stream it out as a data image
[15:47] polotek: makes sense now
[15:47] polotek: nice
[15:50] pbunbun: Stupid question, but how do I tell if an object has an attribute, like if (commands.has("whatever")) returns true if commands["whatever"] exists and false otherwise?
[15:51] jashkenas: pbunbun: that's actually an interesting question in JavaScript.
[15:51] jashkenas: Usually, you can use "if (commands.whatever)"
[15:51] jashkenas: But that will fail if the value of "whatever" is 0 or and empty string.
[15:52] jashkenas: You can use "if ('whatever' in commands)"
[15:52] jashkenas: But that will return true if "commands.whatever == null"
[15:52] pbunbun: Ah, well commands only has objects, so either of those SHOULD work I think
[15:52] pbunbun: TY
[15:53] pbunbun: And if it has anything other than an object for some reason I don't want to know anyway :P
[15:53] ker2x_ has joined the channel
[15:53] jashkenas: pbunbun: no problem. It's the reason we have the existential operator in coffeescript: http://jashkenas.github.com/coffee-script/#existence
[15:53] jashkenas: It's fine with most things, but deadly with numbers, because of zero.
[15:56] Aria: That's one thing that I think Ruby got really right.
[15:56] Aria: Also, confuses the hell out of people that 0 is true.
[15:57] polotek: I think checking for false-y values is usually what you want 99% of the time
[15:57] polotek: of if not then typeof or === fills the gap
[15:57] jashkenas: polotek: absolutely, except if they can be 0 or "", or NaN, which should all be truthy.
[15:58] Aria: Yeah, but it's often a pain. Ruby's consistent about it, and it works nicely.
[15:58] jashkenas: only undefined and null and false should be falsy in JS.
[15:58] Aria: (and it has functions that return number-or-null)
[15:58] mu-hannibal has joined the channel
[15:58] Aria: (And things that in other languages might return the count of something, instead they return the thing or null)
[15:58] markwubben: mde: looking at your node-jake, doesn't really seem to support async tasks combined with dependencies?
[15:58] herbySk: pbinbun: "whatever" in commands if you want take also inherited thisngs into account or commands.hasOwnProperty("whatever") then
[16:00] polotek: jashkenas: I disagree. 0 is often considered "nothing"
[16:00] polotek: and you can't do much with ''
[16:00] steadicat has joined the channel
[16:00] polotek: if you know you're dealing with numbers then you should be using logic operators
[16:00] polotek: === < > etc
[16:01] aliem has joined the channel
[16:01] polotek: otherwise they are false-y
[16:01] polotek: that makese sense to me
[16:01] polotek: and usually saves me typing
[16:01] jashkenas: polotek: glad you're comfortable with it. There's still a lack of a built-in operator that tells you if a variable "exists" ... meaning "has a value" or not.
[16:01] polotek: which is what I assume Aria means by "a pain"
[16:02] polotek: jashkenas: yes, that is typeof
[16:02] herbySk: rubydiamond: I don't know what that page is trying to do, but it eats my CPU 100% fot a minut and then shows one img???
[16:02] polotek: typeof val != 'undefined' is exactly that
[16:02] Aria: Yeah, though I definitely prefer Ruby's way.
[16:02] jashkenas: Or null, which means "not a value".
[16:02] polotek: it's more verbose sure
[16:02] polotek: but how often does that even come up?
[16:02] tisba_ has joined the channel
[16:02] Aria: Heh. How often do you work with code you didn't write?
[16:03] polotek: typeof null === 'object'; Yes some consider it a wtfjs but that's because they are thinking of typeof the wrong way IMO
[16:03] polotek: yes the scheme is a little unintuitive, but it's consistent and it works for the vast majority of common cases
[16:03] jashkenas: polotek: the check you want, to be clear, is: typeof x !== 'undefined' && x !== null right?
[16:04] polotek: jashkenas: it really doesn't matter if you think about it
[16:04] polotek: sure the identity operator is a tad bit faster
[16:04] polotek: but typeof always returns a string
[16:04] polotek: so you don't have to do any type conversions anyway
[16:04] jashkenas: huh? what about null?
[16:04] polotek: and nothing else can == 'undefined' accept for .. undefined
[16:04] herbySk: what is the problem? x === / !== undefined, x !== / === null and x != / == null covers all cases you want to cover, don't they?
[16:05] polotek: typeof null == 'object'
[16:05] jashkenas: so how does that help you determine if it exists?
[16:05] mu-hannibal has left the channel
[16:06] polotek: jashkenas: what I'm saying is that if you're checking for "existence" meaning "is not undefined". then just do typeof x != 'undefined'
[16:06] rubydiamond: herbySk: Sorry for that
[16:06] polotek: even null will pass this check
[16:06] rubydiamond: basically want to enable streaming
[16:06] rubydiamond: hope I can do that
[16:06] polotek: if you want "not undefined and not null"
[16:06] polotek: then you need two checks
[16:06] polotek: but that is the case everywhere
[16:06] herbySk: polotek: x !== undefined does the same without typeof
[16:07] polotek: herbySk: unless x isn't defined at all in the scope
[16:07] polotek: then you get an error
[16:07] polotek: if you know it's there then sure
[16:07] herbySk: and you do not need two checks, x == null / x == undefined catch both values
[16:08] jackish has joined the channel
[16:08] herbySk: polotek: hm, I did not check if each scope has its own undefined...
[16:08] jashkenas: I think this debate is a pretty good example of why there should be a built-in operator for it ;) The fact that there's disagreement about the correct way to check for the presence of a value is telling... no?
[16:09] polotek: jashkenas: maybe. but what I'm saying is that it doesn't come up very often at all
[16:09] polotek: I think two checks is good to be explicit
[16:09] polotek: if you're actually checking for existence
[16:09] jashkenas: polotek: it just did. someone asked the question. ... and I think two checks is good too, so that the variable doesn't even have to be defined: http://jashkenas.github.com/coffee-script/#existence
[16:09] polotek: but usually you're just checking whether something truth-y is there
[16:09] polotek: which works
[16:09] jashkenas: yes. agreed.
[16:10] polotek: jashkenas: okay I guess I get your point. it could be simplified
[16:10] polotek: I admit I'm biased because ruby syntax bugs the crap outta me
[16:11] jashkenas: Ruby isn't the best ... obj.nil? doesn't work on "undefined" local variables.
[16:11] jashkenas: But does on instance variables... So it's a bit problematic.
[16:11] polotek: see, nothing is perfect
[16:12] felixge has joined the channel
[16:13] yoni has joined the channel
[16:13] shimondoodkin has joined the channel
[16:14] blogometer has joined the channel
[16:14] rubydiamond: is it possible to send base64 encoded string back as content type image/png ?
[16:14] dylang_ has joined the channel
[16:15] polotek: rubydiamond: no
[16:15] saikat has joined the channel
[16:15] rubydiamond: okay..
[16:16] rubydiamond: thanks
[16:17] softdrink has joined the channel
[16:17] shimondoodkin: rubydiamond: but you can send chanked mime with transfer encoding base64 + content type image/png
[16:18] shimondoodkin: rubydiamond: (it was a guess answer, just a clue for you)
[16:19] jacobolus has joined the channel
[16:19] rubydiamond: shimondoodkin: trying :) thanks
[16:21] rubydiamond: shimondoodkin: res.writeHead(200, {'Content-Type': 'image/png', 'Transfer-Encoding': 'base64'})
[16:21] rubydiamond: how to force it as chuncked ?
[16:21] rubydiamond: chunked*
[16:22] markwubben has joined the channel
[16:23] mscdex: rubydiamond: if you don't specify a content-length it should be set to chunked encoding
[16:23] rubydiamond: lol nice
[16:23] mscdex: but if you know the content length of the image, it's probably better to set the content-length header
[16:25] _announcer: Twitter: "Getting ready for my next #node.js toy project. We want to use #websockets and #socket.io to push images to several connected clients ..." -- pfleidi. http://twitter.com/pfleidi/status/21761065669
[16:26] pr0cy0n has left the channel
[16:26] rubydiamond: shimondoodkin: it says "Transfer-Encoding: base64 "
[16:26] rubydiamond: not 'chunked'
[16:27] shimondoodkin: oh ok, thnaks
[16:28] rubydiamond: shimondoodkin: any bug ?
[16:28] rubydiamond: res.writeHead(200, {'Content-Type': 'image/png', 'Transfer-Encoding': 'base64'})
[16:28] shimondoodkin: i guess no
[16:28] shimondoodkin: have you tried it?
[16:28] rubydiamond: shimondoodkin: my browser just does not load anything
[16:28] rubydiamond: shimondoodkin: yes
[16:28] rnewson has joined the channel
[16:28] rnewson has joined the channel
[16:29] shimondoodkin: anyways i think it is probably best to decode it to best match browser competability
[16:29] polotek: rubydiamond: I think that's the wrong syntax for Transfer-Encoding
[16:29] _announcer: Twitter: "@pfleidi If you use a helping hand I'm also interested in #nodejs and I have some experience with #websockets..." -- Markus Knittig. http://twitter.com/mknittig/status/21761347145
[16:29] polotek: I think you have to do more but I'm not sure because the only one I ever use is just chunked
[16:30] shimondoodkin: it might be : res.writeHead(200, {'Content-Type': 'image/png', 'Content-Transfer-Encoding': 'base64'})
[16:30] _announcer: Twitter: "@mknittig You could fork our (now emtpy) repository http://bit.ly/9NVZda #nodejs #websockets" -- pfleidi. http://twitter.com/pfleidi/status/21761446770
[16:31] polotek: rubydiamond: try Transfer-Encodng: base64 ; chunked
[16:31] polotek: no that's not right either
[16:31] polotek: http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.6
[16:32] rubydiamond: Transfer-Encodng: base64 ; chunked ?
[16:32] rubydiamond: not right?
[16:32] shimondoodkin: the correct header name probably is : 'Content-Transfer-Encoding'
[16:33] rubydiamond: shimondoodkin: okay trying
[16:33] nateanderson has joined the channel
[16:33] shimondoodkin: il watch it in firebug what it accepts
[16:33] felixge has joined the channel
[16:33] felixge has joined the channel
[16:34] polotek: "One is that
[16:34] polotek: HTTP, unlike MIME, does not use Content-Transfer-Encoding, and
[16:34] polotek: does use Transfer-Encoding and Content-Encoding."
[16:35] rubydiamond: shimondoodkin: hmm
[16:35] polotek: try Transfer-Encoding: chunked, Content-Encoding: base64
[16:35] rubydiamond: polotek: okay
[16:36] shimondoodkin: polotek: humm
[16:36] polotek: I know right? My http fu is still lacking
[16:37] jakehow has joined the channel
[16:38] MikhX has joined the channel
[16:38] rubydiamond: polotek: browser does not show anything but curl shows this http://pastie.org/1106221
[16:39] polotek: rubydiamond: do you get the data without the -I ?
[16:39] rubydiamond: polotek: yes, I get the data without -I
[16:39] rubydiamond: but browser is not rendering it
[16:40] polotek: are you trying to load it as the src of an image tag?
[16:45] _announcer: Twitter: "Combination of #nodejs + #couchdb + #couchapp is pretty good!" -- Yohei Sasaki. http://twitter.com/yssk22/status/21762475952
[16:46] mscdex: i've never seen base64 encoding used in http, except for http auth stuff
[16:47] mscdex: why are you wanting to send it base64 encoded anyway?
[16:47] rubydiamond: mscdex: hax
[16:47] mscdex: eh?
[16:49] shimondoodkin: as i see it firefox supports only thouse: Accept-Encoding gzip,deflate , see this how it is done with headers, http://www.websiteoptimization.com/speed/tweak/compress/
[16:50] polotek: shimondoodkin: ah, yeah, accepts headers
[16:50] femtoo has joined the channel
[16:51] shimondoodkin: internet explorrer only supports also displaying of mhtml files
[16:51] rubydiamond: mscdex: for hacking purpose
[16:51] mscdex: :S
[16:52] polotek: what's the module that lets you expect the heap memory?
[16:52] polotek: or maybe it's a v8 thing
[16:53] herbySk: felixge: do you know some values where the optimized multiplier is not giving the right values?
[16:54] benburkert has joined the channel
[16:56] rubydiamond: polotek: see this http://pastie.org/1106221
[16:57] rubydiamond: it does show
[16:57] rubydiamond: Transfer-Encoding: chunked
[16:57] rubydiamond: Content-Encoding: base64
[16:57] rubydiamond: but data it returns is base64 data encoded only
[16:57] polotek: rubydiamond: I believe you. I just think browsers don't generally accept lots of different encodings
[16:57] rubydiamond: I think data should automatically get decoded
[16:59] jakehow has joined the channel
[16:59] rubydiamond: shimondoodkin: any idea about http://pastie.org/1106221
[17:00] shimondoodkin: not yet iam stil researching it alittle bit
[17:00] shimondoodkin: in general i think that browsers do not support many encodings
[17:05] shimondoodkin: rubydiamond: ok, its not possible , it might work specific of a certian browser, https://developer.mozilla.org/en/How_Mozilla_determines_MIME_Types , for ecxample in mozila there is unknown decoder in tother browser might be another bug
[17:06] shimondoodkin: you might try image/gif;base64 http://www.sweeting.org/mark/blog/2005/07/12/base64-encoded-images-embedded-in-html
[17:07] _announcer: Twitter: "Countdown to Knockout: Post 6 - Node.js and MongoDB http://post.ly/t19M" -- Gerad Suyderhoud. http://twitter.com/gerad/status/21763988628
[17:08] polotek: shimondoodkin: actually he already pulled that off. works well.
[17:11] shimondoodkin: polotek: do rubydiamond succeed?
[17:12] rubydiamond: shimondoodkin: nope
[17:12] polotek: I mean rubydiamond already has the embedded base64 image working
[17:12] rubydiamond: I am now trying to
[17:12] rubydiamond: var fs = require('fs')
[17:12] rubydiamond: var shipitnode = fs.read('shipitnode.png');
[17:12] rubydiamond: but that is not working
[17:12] shimondoodkin: i ment maybe the same mime of embeding images might work with http
[17:13] rubydiamond: buffer = new Buffer(length); bad argument error
[17:13] polotek: rubydiamond: fs.read is asynchronous
[17:13] polotek: you need callbacks
[17:13] polotek: you should probably do fs.createReadStream instead
[17:13] polotek: it's a little nicer to work with
[17:14] jacobolus has joined the channel
[17:15] joshbuddy has joined the channel
[17:15] joshbuddy has joined the channel
[17:15] shimondoodkin: today i want to develop a a pango-
[17:15] rubydiamond: polotek: okay
[17:15] shimondoodkin: cairo basic exctenion
[17:16] shimondoodkin: is there already something like this?
[17:16] shimondoodkin: except the iamgemagistick
[17:16] shimondoodkin: which is bas with international text
[17:18] rubydiamond: polotek: what is problem here http://pastie.org/1106264
[17:18] rubydiamond: it does not show my image
[17:18] shimondoodkin: also i have an idea to use py++ or gcc-to-xml to create half automatic binding generation
[17:19] JimBastard has joined the channel
[17:19] JimBastard: its node.js time! yaaaaaa
[17:19] polotek: fs.readFile is async
[17:19] polotek: you are calling res.end() before shiptinode comes back
[17:19] programble has joined the channel
[17:19] rubydiamond: polotek: hmm
[17:20] polotek: just move res.end inside your readFile callback
[17:20] shimondoodkin: rubydiamond, also you might want to remove the 'Transfer-Encoding': 'chunked'
[17:20] polotek: rubydiamond: also do res.setEncoding('utf8')
[17:20] polotek: actually
[17:21] polotek: yes, do that since you are writing a string to it
[17:21] _announcer: Twitter: "Step by step instructions to install NodeJS on Windows http://is.gd/eubRB <- using Cygwin and node.js v 0.2.0" -- Hernan Garcia. http://twitter.com/theprogrammer/status/21764822803
[17:21] _announcer: Twitter: "@ Ba78 node.js good in this, but first you should learn to think about asynchronous:)" [hu] -- deejayyhu. http://twitter.com/deejayyhu/status/21764824550
[17:21] shimondoodkin: polotek: i think if you dont set it it comes back as buffer whitch is faster
[17:21] shimondoodkin: you can stream.write(buffer)
[17:21] blogometer has joined the channel
[17:21] polotek: shimondoodkin: yeah you're right, readFile comes back as a buffer
[17:22] polotek: rubydiamond: so don't bother with setEncoding
[17:22] rubydiamond: TypeError: Object #
has no method 'setEncoding'
[17:22] rubydiamond: okay
[17:22] polotek: hmm, interesting
[17:22] shimondoodkin: hehe, i had this too some days ago
[17:22] polotek: it bothers me when some parts of http don't follow the stream api
[17:23] shimondoodkin: i guess it is of a readstream
[17:23] rubydiamond: polotek: done
[17:23] rubydiamond: :)
[17:23] polotek: yeah, I was just looking it up
[17:23] rubydiamond: now will shit it
[17:23] polotek: readStream has setEncoding
[17:24] jackish has joined the channel
[17:24] polotek: writeStream has stream.write(data, encoding)
[17:24] polotek: it should really be on both
[17:24] rubydiamond: :)
[17:25] mu-hannibal has joined the channel
[17:25] rubydiamond: also now I want to server my image when request is '/imagename.png'
[17:25] rubydiamond: and for "/" I want to server html page
[17:26] Yuffster has joined the channel
[17:26] rubydiamond: see streaming live here
[17:26] rubydiamond: http://shipitnode.heroku.com/
[17:26] rubydiamond: streaming for real
[17:27] jakehow has joined the channel
[17:27] rubydiamond: it looks quite cool
[17:27] rubydiamond: like I am living in 90's
[17:27] shimondoodkin: there is also anotehr trange of read stream is that it creates a stream decoder and you cant unset it with documented api
[17:28] rubydiamond: shimondoodkin: can I stream it downside up
[17:28] polotek: shimondoodkin: the stream decoder is necessary is most cases
[17:28] polotek: because you're expecting multi byte data
[17:29] polotek: but there's no guarantee the tcp stream won't break in the middle of a multi byte char
[17:29] polotek: that's what the stream decoder does, it handles that case and waits for the rest of the char
[17:29] sudoer has joined the channel
[17:29] polotek: doesn't add much overhead
[17:30] shimondoodkin: polotek: hummm(about waits for two chars)
[17:31] shimondoodkin: polotek: i have found a useful way of not useing stream decoder: buffer is good for binnary data , a buffer is automaticaly converteed to utf 8 when it is converted to string, if i use buffer it is good for utf98 and binary data
[17:32] polotek: shimondoodkin: that's not what I'm saying. it doesn't need to be converted. you need to make sure that you don't end up with two buffers and a multibyte character is split between them
[17:32] rubydiamond: polotek, shimondoodkin chrome streams image correctly.. firefox shows it's name
[17:32] rubydiamond: it downloads it fully then shows it
[17:32] rubydiamond: that's quite wrong with FF
[17:33] rubydiamond: polotek, shimondoodkin Can I stream image from it's downside
[17:33] polotek: ru it streams for me on FF too
[17:33] polotek: 3.6 on mac os x
[17:33] rubydiamond: hmm I am running 4.0beta
[17:35] rubydiamond: polotek: btw what is the default behaviour
[17:35] sudoer has joined the channel
[17:35] rubydiamond: I am assuming that streaming images is not default behavior is other frameworks
[17:36] _announcer: Twitter: "http://tinyurl.com/24sjse7 NodeJS Tutorial with CouchDB and Haml – ErdNodeFlips" -- catch_fish. http://twitter.com/catch_fish/status/21765796664
[17:36] polotek: rubydiamond: depends
[17:36] brucespang has joined the channel
[17:36] polotek: most frameworks don't recommend serving static images through them
[17:37] polotek: you use something linke nginx for that
[17:37] davidwalsh has joined the channel
[17:37] rubydiamond: polotek: that's right
[17:37] polotek: node is actually pretty decent at serving static images though.
[17:37] rubydiamond: polotek: can I stream image chunks by chunks randomly (not from upside down)
[17:39] polotek: rubydiamond: yes
[17:39] rubydiamond: polotek: hmm how ?
[17:39] polotek: you can write anything you want to the response
[17:39] polotek: res.write(data)
[17:39] polotek: you can call it multiple times
[17:40] polotek: the response doesn't end until you call res.end()
[17:41] rubydiamond: polotek: cool
[17:41] rwaldron has joined the channel
[17:41] rubydiamond: how do I assure suppose all corners are rendered first
[17:41] rubydiamond: then center
[17:41] rubydiamond: then other data
[17:41] mu-hannibal has joined the channel
[17:42] rwaldron has joined the channel
[17:42] maushu has joined the channel
[17:43] polotek: rubydiamond: you can't render an image like that
[17:43] rubydiamond: :)
[17:43] jashkenas has left the channel
[17:43] polotek: rubydiamond: maybe you could use a canvas
[17:44] rubydiamond: polotek: no .. i just want to take full advantage of http
[17:44] rubydiamond: :)
[17:44] rubydiamond: polotek: see this site http://blog.evanweaver.com/
[17:44] rubydiamond: see how image got rendered on it
[17:44] rubydiamond: love it
[17:45] gerad has joined the channel
[17:46] polotek: rubydiamond: looks like it's usually multiple http elements loading multiple images
[17:46] _announcer: Twitter: "I see the amazing Seesmic Desktop 2 Beta also ate the first Tweet... "Time to finally try out #nodejs http://nodejs.org" #uninstall" -- Tom Dancer. http://twitter.com/tdous/status/21766426297
[17:46] rubydiamond: polotek: hmm
[17:47] polotek: http://blog.evanweaver.com/images/leaves_top_left.gif
[17:47] polotek: http://blog.evanweaver.com/images/leaves_bottom_right.gif
[17:47] rubydiamond: you are right
[17:47] rubydiamond: yeah
[17:47] polotek: wow, colloquy dropped those into my window. that sucks
[17:48] rubydiamond: it's beauty though
[17:48] mechanate has joined the channel
[17:48] rubydiamond: I can see here in limechat too
[17:48] gerad_ has joined the channel
[17:49] rubydiamond: I hope I could render shipitnode image using socket.io and tweak it to render corners first then center then rest
[17:51] rwaldron_ has joined the channel
[17:52] brucespang has joined the channel
[17:52] rubydiamond: polotek: can you tell me how http://shipitnode.heroku.com/ uses node.js and is good compared to other web frameworks
[17:52] rubydiamond: what is awesome thing done by node.js here
[17:53] rubydiamond: is it just streaming or more?
[17:55] schwa23 has joined the channel
[17:56] brucespang has joined the channel
[17:56] rubydiamond: polotek: unfortunately app automatically broke http://shipitnode.heroku.com/
[17:56] sveimac has joined the channel
[17:57] rubydiamond: throw new Error("This type of response MUST NOT have a body.");
[18:02] [[zz]] has joined the channel
[18:02] rubydiamond: shimondoodkin: you there?
[18:03] shimondoodkin: is anyone here?yes
[18:03] shimondoodkin: yes
[18:03] ajpiano has joined the channel
[18:03] shimondoodkin: s/is anyone here?yes/yes
[18:04] shimondoodkin: rubydiamond:is it should be slow?
[18:04] rubydiamond: shimondoodkin: http://pastie.org/1106348 look at response
[18:04] rubydiamond: there is no chunked there
[18:05] shimondoodkin: yes it chould noy be here
[18:05] rubydiamond: does it mean nginx is making it non-chunked
[18:05] ceej has joined the channel
[18:05] shimondoodkin: maybe
[18:05] rubydiamond: shimondoodkin: how?
[18:05] shimondoodkin: there are two reverse proxies there
[18:05] shimondoodkin: varnish and nginx
[18:05] MikhX has joined the channel
[18:05] rubydiamond: okay.. on browser it looks like it's streaming .. but I am doubful
[18:06] rubydiamond: doubtful*
[18:06] shimondoodkin: it is not
[18:06] saikat has joined the channel
[18:06] rubydiamond: shimondoodkin: it is not streaming ?
[18:07] shimondoodkin: actualy it might be bercause nginx is event based with small memory foot print so it might be that it translates as it comes
[18:07] visnup has joined the channel
[18:07] shimondoodkin: what doy want to do, (why streaming is importent to you?)
[18:07] polotek: rubydiamond: varnish is a cache proxy
[18:07] rubydiamond: shimondoodkin: just for proof of concept
[18:07] polotek: it has the whole response in memory
[18:07] shimondoodkin: of what?
[18:08] polotek: so it can set a content-length header and serve the whole thing
[18:08] shimondoodkin: of sending an image?
[18:08] polotek: streaming is really only important if you don't know the size of the data yet
[18:08] polotek: rubydiamond: or if the data is very large and you don't want to read it all into memory
[18:08] rubydiamond: so looks like it's not streaming
[18:09] shimondoodkin: yes it is just sending as usual
[18:09] rubydiamond: :(
[18:09] shimondoodkin: you canty stream behind nginx
[18:10] rubydiamond: whole point of
[18:10] shimondoodkin: ngix is simle http1.0
[18:10] shimondoodkin: nginx does for you http1.1 and keep alive
[18:11] steadicat has joined the channel
[18:11] rubydiamond: shimondoodkin: does that mean streaming?
[18:11] shimondoodkin: the idea of streaming is keeping connection conneced
[18:12] mu-hannibal has joined the channel
[18:12] _announcer: Twitter: "Node.js experts: lots of beginner questions on the #node.js irc channel before @node_knockout. Could use your help answering!" -- node knockout. http://twitter.com/node_knockout/status/21767985033
[18:12] rubydiamond: Connection: keep-alive is kinda common
[18:12] shimondoodkin: because it takes time to setup a connection from the other side of the world it is like a seconds an a half
[18:12] shimondoodkin: so you keep connection alive
[18:13] shimondoodkin: also you might want two way connection
[18:13] qschzt has joined the channel
[18:13] shimondoodkin: and so you have fast two way communication
[18:13] chrischris has joined the channel
[18:13] shimondoodkin: because you do not reconnect every time
[18:14] shimondoodkin: keep alive is like streaming
[18:14] shimondoodkin: but it is one way streaming
[18:14] rubydiamond: hmm
[18:14] polotek: keep alive is not the same as streaing
[18:14] polotek: nginx does chunked encoding
[18:14] rubydiamond: that's the take
[18:14] polotek: it doesn't do keep alive
[18:15] rubydiamond: polotek: http://pastie.org/1106348 look at response
[18:15] polotek: it is either some setting in nginx that's keep it from sending chunked
[18:15] polotek: or it is varnish
[18:15] polotek: as I mentioned before
[18:15] polotek: varnish sits in front of your server and caches the entire response
[18:15] polotek: it might also be setting the content-length and not bothering to stream it because it's already in memory
[18:16] qschzt has left the channel
[18:17] polotek: rubydiamond you should check the heroku docs and see
[18:17] polotek: what's happening with varnish
[18:17] polotek: from what I've heard
[18:17] polotek: if nginx receives an http 1.1 keep alive request
[18:17] polotek: it will just downgrade it to a normal one way http 1.0 request
[18:18] shimondoodkin: yes nginx , comunicates with your application as http1.0 with the user it comunicates http1.1
[18:18] rubydiamond: I think i should explicitly set "Transfer-Encoding: chunked"
[18:18] rubydiamond: that would make response as chunked
[18:19] shimondoodkin: you should dont cear about it ad let the server choose the corect transfer method
[18:19] rubydiamond: oh I have already done that
[18:19] rubydiamond: it's there (Transfer-Encoding: chunked)
[18:19] rubydiamond: varnish is overriding I think..
[18:19] shimondoodkin: maybe varnish is in front
[18:20] shimondoodkin: waht varnish does as i understand is when you put expiers headers it blockes access to your application until it expiers
[18:21] shimondoodkin: and varnish serves the file instead
[18:21] jacobolus has joined the channel
[18:23] gryzzly_ has joined the channel
[18:23] gryzzly_: hi, should I use kiwi or npm for package management?
[18:23] rubydiamond: npm
[18:23] rubydiamond: that I use
[18:25] gryzzly_: :-)
[18:25] gryzzly_: is there any difference? couldn't google anything on the subject
[18:25] gryzzly_: they are both being mentioned in tutorials on node
[18:25] mscdex: kiwi has been deprecated
[18:25] gryzzly_: oh, thanks
[18:26] gryzzly_: :( their git page doesn't suggest that
[18:26] gryzzly_: i'll use npm then
[18:26] mscdex: there was an announcement on the mailing list some time back
[18:26] mscdex: ACTION shrugs
[18:28] pgriess has joined the channel
[18:28] luddep_ has joined the channel
[18:29] mu-hannibal has joined the channel
[18:29] gryzzly_: it actually reads: NOTE: kiwi is no longer actively developed, however it will remain available until november 2010.
[18:29] gryzzly_: so sorry for not noticing this
[18:32] polotek: anybody know the proper way to pass v8 handles around as void pointers?
[18:34] matt_c has joined the channel
[18:34] MikhX has joined the channel
[18:34] rubydiamond: TypeError: Cannot read property 'path' of undefined
[18:34] rubydiamond: for switch (req.uri.path) {
[18:35] mu-hannibal has joined the channel
[18:35] voxpelli has joined the channel
[18:36] rubydiamond: lol it says req is undefined
[18:36] rubydiamond: polotek: you there?
[18:37] rubydiamond: for var server = http.createServer(function (req, res) {
[18:37] rubydiamond: it say's req is undefined when I dod " sys.puts(req.inspect)"
[18:39] DriverDan has joined the channel
[18:39] polotek: rubydiamond: need more context than that
[18:39] rubydiamond: polotek: ok
[18:39] polotek: also
[18:39] polotek: req.url is the url propery. it's a string and you have to parse it yourself
[18:39] polotek: use require('querystring')
[18:39] rubydiamond: polotek: http://pastie.org/1106392
[18:40] luddep has joined the channel
[18:40] bsstoner has joined the channel
[18:43] mertimor has joined the channel
[18:43] rubydiamond: polotek: is there anything wrong in code?
[18:44] polotek: rubydiamond: you don't have to do parseInt on port
[18:44] polotek: and like i said req.url not req.uri
[18:44] polotek: I have no idea what req.inspect is
[18:45] polotek: you are probably looking sys.puts(sys.inspect(req))
[18:45] polotek: or you can do it the easy way
[18:45] polotek: console.log(req)
[18:46] polotek: rubydiamond: check out the querystring module in the docs for parsing req.url
[18:47] rubydiamond: polotek: hmm
[18:47] _announcer: Twitter: "@node_knockout hey I joined to the channel and nobody was there , I tried #node.js #nodejs" -- alejandro. http://twitter.com/ramonlechuga/status/21769995982
[18:47] rubydiamond: okay
[18:48] shimondoodkin: polotek: did you used some good examples for c++ node modules?
[18:48] polotek: shimondoodkin: what do you mean?
[18:48] shimondoodkin: i want to develop a c module for v8 for node, i never did it
[18:49] stepheneb has joined the channel
[18:49] mtodd has joined the channel
[18:49] stepheneb has joined the channel
[18:49] shimondoodkin: you asked a question like you already did some c++ modeules, you maybe you pass over some example that is a good to understand how to do bindings for node
[18:50] polotek: shimondoodkin: It's defiinitely non trivial
[18:50] polotek: you have to learn the v8 api
[18:51] polotek: http://code.google.com/apis/v8/embed.html
[18:51] ramonlechuga has joined the channel
[18:51] shimondoodkin: there are v8 proxy/juce
[18:51] polotek: you can start there
[18:51] shimondoodkin: is it the same with node?
[18:51] polotek: shimondoodkin: I'm pretty sure you can use one of those with a node addon
[18:51] polotek: but I don't know anyone who has tried
[18:52] shimondoodkin: thanks ill read the api it is probably the correct start point
[18:52] polotek: C++ frameworks are somewhat frowned upon :)
[18:52] polotek: shimondoodkin: also check out http://github.com/isaacs/node-glob
[18:52] polotek: I think it's a pretty straight forward addon
[18:52] polotek: without being too trivial
[18:53] polotek: there is also http://github.com/polotek/libxmljs
[18:53] polotek: which I maintain
[18:53] polotek: but it is moderately complex
[18:53] polotek: not really good for learning
[18:54] polotek: also http://github.com/ry/node_postgres
[18:54] polotek: also pretty simple
[18:54] codetonowhere has joined the channel
[18:54] icozzo has joined the channel
[18:55] shimondoodkin: thanks it is already more then i able to read
[18:56] Aria has joined the channel
[18:57] rubydiamond: shimondoodkin, polotek for what you use node.js do you guys have any work flow / use case for using it as everyday purpose
[18:57] ramonlechuga: Hello, I already installed node.js, I tried the hello world and it works .. but I don't any I Idea what else to do, does anyone here knows a good guide for beginners?
[18:58] rubydiamond: here is updated version of http://shipitnode.heroku.com/
[18:58] robotarmy has joined the channel
[18:58] shimondoodkin: what app do you want to create?
[18:58] Aria: rubydiamond: I'm writing a spam filter SMTP daemon.
[18:58] Aria: rubydiamond: And I'm writing a filtering and HTML-reparsing proxy that will eventually run greasemonkey scripts.
[18:58] shimondoodkin: i am writing a super fast scalable website
[18:58] JimBastard: ship it squirrel
[18:58] blogometer has joined the channel
[18:59] blogometer: Mongo DB: Love it or hate it?
[18:59] shimondoodkin: you will hate no sql there are no joins (LOL)
[19:00] Aria: Sure there are. You just have to DIY, preferably in a batch, in advance.
[19:01] rubydiamond: Aria, shimondoodkin cool
[19:01] shimondoodkin: i have a guy that helps me he as a begginer like 3 years in programming , and he finds it hard that there are no joins
[19:01] Aria: Well, yeah, it's different. But instead you have structured documents, which is often what people use joins for.
[19:01] rubydiamond: omg no joins is no sql ?
[19:01] shimondoodkin: rubydiamond: you can do real time colaboration one page apps in node js
[19:01] JimBastard: rubydiamond: you can do relationships in nosql
[19:01] JimBastard: why not
[19:02] rubydiamond: http://search.twitter.com/search?q=shipitnode is not not showing anything yet .. come on guys .. show some love
[19:02] rubydiamond: shimondoodkin: actually I have an idea in mind about real time collaboration app with node.js
[19:03] rubydiamond: but I heard that it requires a hard algorithm to know about real time collaboration
[19:03] gryzzly_: guys, i'm really new to node. how can I get the express within app.js, that is located here: '~/nodecode/blog/app.js' if I installed express with npm?
[19:04] shimondoodkin: JimBastard> why not
[19:04] shimondoodkin: [22:01] http://search.twitter.com/search?q=shipitnode is not not showing anything yet .. come on guys .. show some love
[19:04] shimondoodkin: oops
[19:04] gryzzly_: is there a way to do like var npm = require('npm');
[19:04] rubydiamond: what was that
[19:04] gryzzly_: and then somehow to ask npm about express's location
[19:04] shimondoodkin: i tried to copy the name of JimBastard
[19:05] shimondoodkin: JimBastard: can you show me how to do join in mongo db in a one query
[19:05] mattly has joined the channel
[19:05] polotek: gryzzly_: npm puts your modules on your require path
[19:05] polotek: you should just be able to do require("express")
[19:07] mau has joined the channel
[19:08] cjm has joined the channel
[19:09] shimondoodkin: JimBastard: i have searched a whole day and decided that it is imposible to do joins in mongodb and i have to do them in code
[19:09] gryzzly_: @polotek thanks!
[19:09] cjm: He
[19:09] cjm: hey
[19:10] JimBastard: shimondoodkin: sure, the tutorial is up @ http://couch.io
[19:10] cjm: I'm having issues building node.js on a mac os x
[19:10] JimBastard: shimondoodkin: but i gotta assume mongo has that query thingy, surely it can join collections
[19:11] shimondoodkin: JimBastard: i dont understand what do you mean
[19:12] JimBastard: me neither man, sorry. i dont really use mongo
[19:12] shimondoodkin: JimBastard: is couch is good?
[19:12] brucespang has joined the channel
[19:12] JimBastard: shimondoodkin: we are using it on our production stack @ nodejitsu
[19:12] cjm: When I run ./configure
[19:12] JimBastard: and the support has been great
[19:13] JimBastard: mikeal is one of the core couch guys too, hes usually aruond
[19:13] cjm: I'm told I need library dl, execifo, openssl, ...
[19:13] JimBastard: around*
[19:13] cjm: what is dl?
[19:13] JimBastard: it depends what you need to do shimondoodkin , there are many database solutions
[19:13] polotek: cjm: do you have XCode installed?
[19:13] polotek: you need it
[19:13] blogometer: Okay. So, I'm hearing you say good, sensible things about Mongo DB. Thank you.
[19:14] shimondoodkin: i need fast retrivel of large lists
[19:14] blogometer: Thank you: JimBastard, Aria.
[19:14] polotek: You definitely can't do joins across collections in mongo
[19:14] cjm: I don't
[19:14] cjm: by I have openssl
[19:14] cjm: and It can't find it
[19:14] polotek: within a collection you can simulate a join by doing a map function
[19:15] JimBastard: polotek: server-side right?
[19:15] JimBastard: err, database side?
[19:15] shimondoodkin: how to do map function to do joins, map recuse seems to me only inside a collection
[19:15] polotek: JimBastard: yeah you can write the function in js and send it to mongo to execute
[19:15] polotek: shimondoodkin: yeah that's what I said, not across collections
[19:15] JimBastard: exactly
[19:15] JimBastard: map/reduce
[19:15] benburkert has joined the channel
[19:15] polotek: shimondoodkin: but there's not reason not to have all your objects in one collection
[19:16] polotek: mongodb is schemaless
[19:16] polotek: it doesn't matter what you put it in it
[19:16] JimBastard: http://browsertoolkit.com/fault-tolerance.png
[19:16] JimBastard: i think that sums it up
[19:16] shimondoodkin: it would nece to have expanded mongo ids
[19:16] badaxx has joined the channel
[19:16] polotek: JimBastard: lol, yeah I've seen that
[19:16] polotek: awesome
[19:17] polotek: shimondoodkin: how do you mean?
[19:17] badaxx: hello, I've got a problem with mysql / node-mysql. when I do the first query with client.query(...) it works fine but on the second query i get the folowing error: http://pastebin.com/bYXZTB5k
[19:18] badaxx: i'm using the same mysql.Client instance for both queries
[19:18] benburkert: does npm work with node 0.2.0?
[19:19] shimondoodkin: i have collection with references like {_id:objectid(213123fgsda),name:"asdfas",sub_object:objectid('asdfasdf')} i would be happy to have the sub object expanded to its value from other collection of sub objects
[19:19] benburkert: i've been following this websockets post here and i can't require the websockets-server module: http://static.brandedcode.com/nws-docs/
[19:19] mscdex: badaxx: already been reported: http://github.com/felixge/node-mysql/issues#issue/1
[19:19] polotek: shimondoodkin: I see you've been editing the readme for libxmljs
[19:19] polotek: you wanna check out the new wiki
[19:19] polotek: http://github.com/polotek/libxmljs/wiki
[19:20] badaxx: mscdex: ouch...
[19:20] polotek: shimondoodkin: you can next objects in mongodb
[19:20] polotek: nest*
[19:20] polotek: put your subobjects directly inside
[19:20] badaxx: what about this one, any experiences with that? http://github.com/masuidrive/node-mysql
[19:20] polotek: you can still query them
[19:21] mscdex: badaxx: not in awhile, no
[19:21] dnolen has joined the channel
[19:21] mscdex: badaxx: i don't believe it is maintained anymore
[19:21] cjm has left the channel
[19:22] badaxx: any other mysql modules for node? :(
[19:22] benburkert: how does node/npm manage the load path? is there something i can inspect?
[19:23] shimondoodkin: polotek: i saw the wiki :)
[19:23] gerad has joined the channel
[19:23] polotek: shimondoodkin: let me know your feedback
[19:23] mscdex: badaxx: you can give this one a try: http://github.com/sidorares/nodejs-mysql-native
[19:24] polotek: this is going to replace the other site this week
[19:24] badaxx: think I'm gonna fix that bug myself
[19:24] shimondoodkin: its awsome
[19:24] shimondoodkin: polotek: it is awsaome i whish i had it when i had stqarted
[19:24] polotek: badaxx: yeah I believe mysql-native has also been updated recently
[19:24] micheil: benburkert: problem?
[19:25] badaxx: can i do asynchronous queries with it?
[19:25] micheil: ACTION is the developer.
[19:25] pbunbun: Hmmm.
[19:25] pbunbun: I have "var quoteFile = fs.createReadStream(filename);"
[19:25] gerad_ has joined the channel
[19:25] gerad has joined the channel
[19:25] pbunbun: I'm trying to do quoteFile.addListener("end", function() {quoteFile.destroy()});
[19:25] polotek: badaxx: yes, I believe that was the recent update
[19:25] micheil: benburkert: also, those docs won't work for anything anyone currently has
[19:25] pbunbun: But getting Error: EBADF, Bad file descriptor
[19:26] polotek: it uses the node async io thread pool to take things out of band
[19:26] benburkert: micheil: hi, i'm trying to understand why require('websocket-server') is failing after i've run 'npm install websocket-server'
[19:26] micheil: benburkert: they target a currently unrelease version, and I'm surprised someone's linked you to them
[19:26] mscdex: badaxx: i think your problem is that you're ending the connection before the queries can take place
[19:26] micheil: benburkert: it shouldn't be failing.
[19:26] badaxx: mscdex: no, the connections are still open
[19:26] mscdex: badaxx: paste your code?
[19:27] badaxx: compile coffeescript or coffeescript?
[19:27] polotek: benburkert: did you install anything through homebrew?
[19:27] polotek: people have been having problems with node + homebrew + npm lately
[19:27] mscdex: badaxx: eh.. javascript? :-)
[19:27] benburkert: micheil: yeah, i think my homebrew setup might be a bit off, is there someway of printing out the module path?
[19:27] micheil: umm.. require.paths
[19:27] polotek: benburkert head into the node repl and type
[19:27] polotek: yep require.paths
[19:27] polotek: that shows the module load path
[19:28] polotek: I haven't seen a fix for the homebrew issue yet
[19:28] dylang has joined the channel
[19:28] polotek: but a workaround is to exlicitly set your $NODE_PATH env variable
[19:28] badaxx: mscdex: http://pastebin.com/4SU1kgTn << JavaScript, http://pastebin.com/g1LYwiTi << CoffeeScript
[19:28] polotek: to /usr/local/lib/node
[19:28] micheil: benburkert: at any rate, that documentation won't help you much, as it's for a newer version of the server, iirc.
[19:28] rubydiamond: polotek: is there any simple tutorial / link / video explaining socket.io functionality
[19:28] badaxx: mscdex: getConfig is called before getServerInfoById. So the getServerInfoById query fails
[19:28] polotek: rubydiamond: never used socket.io
[19:29] rubydiamond: ok
[19:29] marek_z has joined the channel
[19:30] benburkert: ahh, ok: http://pastie.org/private/q7h2pdzvmfloje0pm2w
[19:30] ajpiano has joined the channel
[19:31] benburkert: polotek: so i should set $NODE_PATH to ~/Developer/lib/node? http://pastie.org/private/alfojl3ho7hupaxlowjxw
[19:31] mscdex: badaxx: do you have the sample code that uses the functions you showed?
[19:33] badaxx: mscdex: sample code? from node-mysql?
[19:34] polotek: benburkert: if you run npm config get root
[19:34] polotek: it will tell you where npm is putting your modules
[19:34] mscdex: badaxx: how are you using the code you showed? I don't see anything that calls those get* functions
[19:34] polotek: set your NODE_PATH to that
[19:34] benburkert: polotek, micheil: awesome, thanks for the help guys: http://pastie.org/private/b32ws5bhkhomu7yfl5i0w
[19:34] badaxx: mscdex: ah one second
[19:34] mAritz: fictorial: does redis-node-client support multi/exec?
[19:35] badaxx: mscdex: http://pastebin.com/tQQ2taJT
[19:35] badaxx: mscdex: http://pastebin.com/0dk9wUsq << js version
[19:38] blogometer: mscdex: You all get my patch yet?
[19:38] blogometer: ACTION Taps feet.
[19:38] mscdex: blogometer: huh?
[19:39] blogometer: Grr...
[19:39] blogometer: http://github.com/bigeasy/node-mysql
[19:39] blogometer: ACTION Cross arms taps feet faster.
[19:39] mscdex: blogometer: i told you yesterday i am not the developer
[19:39] mscdex: i don't even have a fork
[19:39] blogometer: Oh, well.
[19:39] mscdex: :S
[19:40] isaacs has joined the channel
[19:40] mephux has left the channel
[19:40] joshbuddy has joined the channel
[19:40] mscdex: blogometer: felixge is the man you want
[19:40] blogometer: I was too lazy to write a test.
[19:41] blogometer: I'm sure he'll get to it.
[19:41] twoism has joined the channel
[19:41] mscdex: badaxx: i think the problem is you are expecting the get* functions to be synchronous
[19:42] blogometer: mscdex: Can you recommend a Javascript style guide?
[19:42] mscdex: blogometer: sure: no comma first and end statements with a semi-colon
[19:42] mscdex: blogometer: :p
[19:43] pbunbun: If I do: var test = require("./test"), then later on delete test
[19:43] pbunbun: And do var test = require("./test")
[19:44] pbunbun: It doesn't seem to reload the file
[19:44] pbunbun: It just provides it as it was at the beginning?
[19:44] mscdex: badaxx: consider having getConfig call getServerInfoById inside getConfig's callback
[19:44] shimondoodkin: pbunbun: it has cache
[19:44] mscdex: badaxx: and then just only call getConfig in your main program there
[19:45] shimondoodkin: pbunbun: this might give you sort of a solution http://github.com/shimondoodkin/node-hot-reload
[19:45] mscdex: badaxx: it's still not synchronous, but it will execute them one after the other
[19:45] pbunbun: Ah ty
[19:46] mscdex: badaxx: ideally you would pass in a callback to getConfig that gets executed when the results come in, so you know when to continue
[19:46] rnewson has joined the channel
[19:46] rnewson has joined the channel
[19:47] siculars has joined the channel
[19:53] _announcer: Twitter: "Node.js: JavaScript on the Server http://bit.ly/cvd2Sr" -- Lyla Leibowitz. http://twitter.com/LylaLeibowitz/status/21773617441
[19:55] probablyCorey has joined the channel
[20:01] saikat has joined the channel
[20:04] voodootikigod has joined the channel
[20:06] sephr has joined the channel
[20:06] nerdEd has joined the channel
[20:06] sephr: Does anyone here have a burning desire to be able to use E4X in node.js?
[20:07] derferman has joined the channel
[20:07] inimino: ACTION hates E4X
[20:07] shimondoodkin: i dont know the need for e4x
[20:07] shimondoodkin: seems like a simple xml parses does the solution
[20:08] sephr: JavaScript lacks multi-line strings, which E4X makes up with
[20:08] sephr: and inline logic is nice
[20:08] sephr: though php-y
[20:09] mmetzger has joined the channel
[20:10] Aria: I'd love E4X, though even more, I'd love E4HTML
[20:11] sephr: Aria: please elaborate
[20:11] sephr: what'd E4HTML be that E4X isn't?
[20:11] shimondoodkin: i dont like it i prefer normal strings
[20:11] Aria: HTML not XML.
[20:11] Aria: They /are/ different languages.
[20:11] ramonlechuga has joined the channel
[20:12] sephr: yes, but why would you want to use HTML instead of XML in a programming language?
[20:12] sephr: XML is messy enough as is
[20:15] Throlkim_ has joined the channel
[20:15] polotek: sephr: I think what he means is being able to get a decent e4x structure from potentially crappy html
[20:15] polotek: needs to have a good parser
[20:16] polotek: also things like singleton attributes and stuff
[20:16] polotek: html is different from well-formed xml
[20:16] polotek: unary attributes? is that the appropriate phrasing?
[20:16] jamescarr_ has joined the channel
[20:17] polotek: anyway, I like e4x too.
[20:17] polotek: if you're just building up a string to output, maybe you don't need it
[20:17] rubydiamond: ping
[20:17] polotek: but if you have to do processing and manipulation, having an object with a decent api is nice
[20:19] softdrink has joined the channel
[20:19] drudge: when trying to compile node on OSX, it tells me openssl is not found but SSL_library_init and openssl/crypto.h are found. is this normal?
[20:19] drudge: http://pastie.org/1106521
[20:20] shimondoodkin: yes
[20:20] shimondoodkin: configure' finished successfully (5.116s)
[20:22] sephr: polotek: you can already convert malformed HTML to E4X using the DOM
[20:23] mscdex: drudge: yes, the openssl line is if you have pkgconfig installed
[20:23] sephr: but why you would want to do that on the server side, where you're writing the XHTML to begin with, is beyond me
[20:23] badaxx: mscdex: okay that works fine for the first request to the http server
[20:23] mscdex: drudge: if the other lines show yes, then you're fine
[20:23] badaxx: mscdex: but on the second request I've got the same error
[20:23] drudge: mscdex: i thought so but i'm having odd ssl issues with http.Client
[20:23] mscdex: badaxx: code?
[20:23] mscdex: drudge: like?
[20:24] drudge: net:992 this.secureStream.close();
[20:24] drudge: TypeError: Cannot call method 'close' of undefined
[20:24] badaxx: mscdex: http://pastebin.com/z7DFYydw
[20:25] mscdex: drudge: gist the code you're using
[20:25] _announcer: Twitter: "brequire, commonjs loader in browser, simple and clean. excelent to share between node.js and web browsers http://github.com/weepy/brequire" -- Jerome Etienne. http://twitter.com/jerome_etienne/status/21775238182
[20:25] badaxx: mscdex: http://pastebin.com/py85cES3 the get* functions
[20:25] drudge: mscdex: moment i'm trying something
[20:25] hellp has joined the channel
[20:25] polotek: sephr: why would you want to go through the intermediate DOM?
[20:26] polotek: you would want to take your html string parse it and get an e4x object back directly
[20:26] mscdex: badaxx: i think you're still running into the same problem
[20:27] polotek: sephr: also you're not always in charge of the content on the server. you could be reading it from some other source
[20:27] sephr: polotek: by DOM, I mean the parsers provided by the DOM
[20:28] mscdex: badaxx: one moment
[20:28] sephr: you can parse them into nodes and then serialize them into strings
[20:28] sephr: and then parse those corrected serializations with E4X
[20:28] polotek: sephr: can you use those directly in v8?
[20:29] sephr: the DOM and E4X?
[20:29] polotek: sephr: wow, that's a lot conversion overhead
[20:29] shimondoodkin: and then eval() to have the local scope
[20:29] polotek: if you can access the parser in a sax context you could have it create e4x directly
[20:29] sephr: polotek: I'm _not_ recommending this
[20:29] polotek: that is if you have e4x available
[20:29] sephr: I'm saying that's what it looks like you're saying
[20:29] sephr: oh wait not you
[20:29] polotek: sephr: I was asking if you could access the html parser functions directly
[20:30] sephr: whoever said the E4HTML thing
[20:30] polotek: but actually I don't know if those are event included in v8
[20:30] polotek: they are probably in webkit/chrome
[20:30] inimino: That was Aria.
[20:30] sephr: why would you want to parse HTML on the server side anyways?
[20:30] sephr: I don't see the point, when you are the one generating said html
[20:30] Aria: Because it's a much neater way of generating it.
[20:30] inimino: But the idea was that you'd write e.g. "foo
bar" and get a correct DOM, not a parse error.
[20:30] sephr: huh?
[20:31] sephr: Aria: I'm absolutely lost
[20:31] mscdex: badaxx: to be honest, there's several things wrong. i'm not sure if the coffeescript compiler has anything to do with it or not. i'll give you an example of what it should look like
[20:31] inimino: HTML is neater than XML.
[20:31] Aria: Ever dealt with string interpolations? Escaping?
[20:31] TooTallNate has joined the channel
[20:31] Aria: It's SO EVIL.
[20:31] sephr: inimino: heh. Not even mozilla goes far enough to implement that part of E4X
[20:31] inimino: sephr ⋱ It wouldn't be E4X anymore, it's be E4HTML
[20:31] sephr: it'd be E4DOM
[20:32] nerdEd has joined the channel
[20:32] inimino: (because "
foo
bar" isn't XML, but it is HTML.)
[20:32] sephr: but what you explained is actually valid E4X stuff
[20:32] TooTallNate: ping: isaacs
[20:32] isaacs: hi
[20:32] sephr: oh I didn't see that the second p was a start tag not end tatg
[20:32] mAritz: anyone here who has used redis-node-client with multi/exec enabled?
[20:32] TooTallNate: I just tried to publish my "applescrip" module
[20:32] inimino: sephr ⋱ Take E4X, rip out the XML parser, add an HTML parser, and you have E4HTML.
[20:33] sephr: inimino: I thought you said
foo
bar, which is perfectly valid E4X in an XMLList literal (<>)
[20:33] TooTallNate: It worked the first time
[20:33] sephr: in other terms, you'd have PHP
[20:33] TooTallNate: but then I "unpublish"ed it (bad I know) cuz I wanted to add a license
[20:33] sephr: but with the DOM I guess
[20:33] inimino: sephr ⋱ And E4HTML might actually be useful.
[20:33] sephr: I'm just lost as to why you'd want to generate malformed HTML purposely?
[20:33] TooTallNate: then republished it and it seemed to work (it's there in "npm ls")
[20:33] isaacs: k
[20:33] sephr: s/\?//
[20:34] inimino: sephr ⋱ Well, you can't manipulate the chunks of HTML in PHP, they're just dumped out. E4X creates variables.
[20:34] Aria: Only problem with E4HTML is that end-of-parse conditions are harder.
[20:34] TooTallNate: but "npm install applescript" doesn't work from my other computer
[20:34] isaacs: TooTallNate: works for me
[20:34] inimino: sephr ⋱ "foo
bar" isn't malformed HTML, it's correct HTML.
[20:34] isaacs: TooTallNate: does that other computer have it installed already? what does "doesn't work" mean?
[20:34] sephr: I mean why you'd generate malformed XHTML purposely (even if you're serving it as HTML)
[20:35] isaacs: sephr:
a
b is not malformed, though, it's fine. some sites care about every byte
[20:35] Aria: Because it's shorter, easier to read, happens anyway ... perhaps comes partially from user content.
[20:35] inimino: sephr ⋱ Because some people think that HTML syntax is much more usable, and XML was a failure.
[20:35] sephr: those people aren't smart enought o use E4X + node.js heh
[20:35] TooTallNate: https://gist.github.com/836068acf0c8aad52dba
[20:35] inimino: (Or at least, XHTML was a failure.)
[20:35] sephr: those are the "HTML programmers", or so they say on their CV
[20:36] inimino: sephr ⋱ I'm one of those people, actually :-)
[20:36] sephr: program me up some HTML ;)
[20:36] inimino: And I think XML wasn't the greatest idea either.
[20:36] fod has joined the channel
[20:36] shimondoodkin: i think i have missed the point about e4x, it is a new way of building an object instead of {bla:blabla,foo:{}} and make it xml object type you just enter it as xml and it returns an object... pretty neat idea but who needs it anyways...
[20:36] polotek: xhtml was a regrettable failure
[20:36] sephr: well I guess boolean attributes are alright
[20:36] ramonlechuga has left the channel
[20:36] sephr: all of these markup languages suck
[20:36] polotek: shimondoodkin: e4x gives you a DOM
[20:37] polotek: it's different from the browser DOM
[20:37] polotek: has it's own api
[20:37] shimondoodkin: yes i get it
[20:37] polotek: but the really cool thing about e4x is it's supported as native syntax in js
[20:37] inimino: unquoted attributes, shortcut attributes, omitted tags... I like all of these.
[20:37] sephr: though it's fairly underpowered
[20:37] drudge: mscdex:
[20:37] drudge: http://pastie.org/private/hshgbgqhsinte5xemajg
[20:37] sephr: you can't mess with comments and processing instructions with ease in E4X
[20:37] isaacs: TooTallNate: i dunno
[20:37] TooTallNate: isaccs: did you get that? https://gist.github.com/836068acf0c8aad52dba
[20:37] sephr: nor doctypes
[20:37] isaacs: TooTallNate: yeah
[20:37] isaacs: TooTallNate: works for me, no problems.
[20:37] polotek: sephr: the spec is certainly incomplete
[20:37] TooTallNate: huh
[20:37] isaacs: TooTallNate: npm rm applescript and then try again?
[20:38] isaacs: i dunno
[20:38] polotek: and the reason is that it didn't really generate much excitement
[20:38] inimino: E4X consumed too much syntax real estate for too little benefit.
[20:38] polotek: probably no reason to really improve it
[20:38] drudge: mscdex: and this is output:
[20:38] drudge: http://pastie.org/1106545
[20:38] polotek: inimino: syntax real estate?
[20:39] inimino: polotek ⋱ I.e. killing future extension possibilities.
[20:39] polotek: inimino: hmmm, I don't now much about language development.
[20:39] inimino: I guess I really meant "future syntax entension possibility real estate".
[20:40] TooTallNate: isaacs: "npm rm applescript" didn't help
[20:40] TooTallNate: I'll try reinstalling npm?
[20:40] isaacs: TooTallNate: sure, that'll keep you busy for a bit
[20:40] TooTallNate: haha
[20:40] polotek: this is all avoiding the real issue though
[20:40] isaacs: TooTallNate: what does "npm ls applescript" show on that other computer?
[20:40] inimino: Well, all sorts of cool things that have been proposed for future versions of ECMAScript haven't been possible or haven't had ideal syntax because E4X squatted the syntax.
[20:40] polotek: xml and html are useful to handle as DOM
[20:41] polotek: browsers have the best parsers
[20:41] sephr: inimino: didn't I propose a binary markup format before in ##paws that also had a common string serialization that humans can easily modify?
[20:41] polotek: why don't they give them to us?
[20:41] inimino: That's why I'd love to see E4X die completely.
[20:41] inimino: sephr ⋱ Maybe?
[20:41] inimino: sephr ⋱ Like binary XML?
[20:41] sephr: lol, no
[20:41] sephr: it'd replace XML/HTML for sure
[20:42] sephr: but it wouldn't be a binary serialization of a doomed format
[20:42] sephr: s/doomed/bad/
[20:42] inimino: polotek ⋱ You could port one of the browser's parsers to node, but it'd be a lot of work.
[20:42] sephr: sadly I don't think it'll ever die
[20:42] inimino: Oh, it'll die eventually.
[20:42] polotek: inimino: I know.
[20:43] inimino: Like all human works ^_^
[20:43] polotek: inimino: the one in webkit has feelers through the rest of hte code like a spider
[20:43] Aria: inimino: That's what the html5 library is. Already ported to node.
[20:43] Aria: It's almost exactly the new parser Mozilla just got.
[20:43] m64253 has joined the channel
[20:43] gryzzly_: hey you guys. is there any nice express tutorials out there?
[20:43] polotek: there are so many things that have to change depending on the browser environment
[20:43] inimino: Aria ⋱ Yeah, I know, but it's written in JavaScript, right?
[20:43] Aria: And yes, it was a lot of work ;-)
[20:43] inimino: I know :-)
[20:43] sephr: we'd have to kill the web first and then reinvent everything before we can kill html
[20:44] Aria: Yeah, it is. But that could be fed back into making an E4HTML parser ;-)
[20:44] sephr: s/web/http?/
[20:44] gryzzly_: express node web framework
[20:44] inimino: I meant you could rip out the C++ implementation and port that (which would also be lots of work).
[20:45] polotek: sephr: http has nothing to do with it
[20:45] polotek: gryzzly_: have you looked at hte site?
[20:45] polotek: http://expressjs.com/
[20:45] badaxx: mscdex: guess it's me who did things wrong, I'm new to node.js and coffeescript.but i'll wait for your solution
[20:45] gryzzly_: sure, i've read the guide part also
[20:45] sephr: polotek: well I mean the whole web
[20:45] sephr: everything except TCP just being redone
[20:46] gryzzly_: but as I am client-side developer, I'd need something more step-by-step in order to get into things.
[20:46] polotek: gryzzly_: also this http://nodeknockout.posterous.com/countdown-to-knockout-post-2-hello-world-with
[20:46] polotek: sephr: the web has issues sure
[20:47] TooTallNate: isaacs: Now this is what I get when trying to uninstall: https://gist.github.com/f8117d21bf4130a7e418
[20:47] polotek: but it could be worse
[20:47] isaacs: TooTallNate: what version of node do you have?
[20:47] TooTallNate: v0.1.104
[20:48] gryzzly_: @thanks for that last link. I've managed to do that myself! exactly as it is in the tut from nodeknockout but without the console.log line
[20:48] gryzzly_: :-)
[20:48] gryzzly_: my question would be how would I connect couch-db to the business
[20:48] gryzzly_: i am trying to write a simple blog app, just for learning it
[20:48] isaacs: TooTallNate: there is something strange with your setup
[20:49] TooTallNate: damnit
[20:50] isaacs: TooTallNate: rm -rf $(npm config get root)/{.npm/npm,npm}* $(npm config get binroot)/npm*
[20:50] isaacs: TooTallNate: then install npm again
[20:50] mape: isaacs: http://npm-gen.mape.me/mape/connect-assetmanager
[20:51] isaacs: mape: nice
[20:51] mape: not sure how "edge case" heavy it should be
[20:51] TooTallNate: isaacs: K, did that, "npm ls applescript" shows "applescript@0.1.0 =TooTallNate latest remote"
[20:52] mape: and should probly add helt texts
[20:52] isaacs: TooTallNate: and what happens when you install it?
[20:52] TooTallNate: but "npm install applescript" throws the same error
[20:52] isaacs: TooTallNate: how did you install npm?
[20:52] isaacs: TooTallNate: and what does node -v show?
[20:52] TooTallNate: git clone; make install
[20:52] mscdex: badaxx: something like this: http://gist.github.com/542833
[20:52] isaacs: TooTallNate: can you instal other things?
[20:52] mscdex: badaxx: it can take awhile to get used to thinking asynchronously
[20:53] isaacs: TooTallNate: maybe clear the cache and try again? i have literally no idea whatsoever. what you're seeing is not what it should be doing.
[20:53] bradleymeck2 has joined the channel
[20:53] TooTallNate: isaacs: I was able to update connect just now
[20:53] isaacs: update to node 0.2.0? i don't know.
[20:53] bradleymeck2: for knockout can we have pre downloaded libraries / icon sets (no code for project itself)
[20:53] visnup: bradleymeck2: yes
[20:54] polotek: gryzzly_: most people access couchdb using the REST interface over http
[20:54] bradleymeck2: cool, just checking, starting to dl assets then
[20:54] visnup: bradleymeck2: but don't commit anywhere yet
[20:54] TooTallNate: isaacs: well I'll reboot and try on Ubuntu, and try updating node as well
[20:54] bradleymeck2: im not going to :) read the email
[20:54] mscdex: drudge: one thing i see right away is that your usage of createClient is wrong
[20:54] drudge: yeah i noticed that too :/
[20:54] drudge: fixing it up now with the headers
[20:55] isaacs: TooTallNate: ok. i'm gonna be offline, but you can email npm-@googlegroups.com or post an issue.
[20:55] TooTallNate: isaacs: thanks
[20:55] isaacs: TooTallNate: this "set before loading" thing is very very odd.
[20:55] mscdex: drudge: i could see how that could easily make secure mode not work
[20:55] MaSch has joined the channel
[20:56] mscdex: drudge: let me know if that fix doesn't help
[20:56] sonnym has joined the channel
[20:58] gryzzly_: polotek: so say there is a mongod that is running on my mac os, and it's accessible via http://127.0.0.1:27017 so I just make a call there and get the data?
[20:59] twoism has joined the channel
[20:59] polotek: gryzzly_: you said couchdb not mongo :)
[20:59] polotek: gryzzly_: for mongo check out this http://github.com/christkv/node-mongodb-native
[20:59] gryzzly_: sorry, it's 4 am in Siberia :-)
[20:59] gryzzly_: thanks
[21:02] ph^ has joined the channel
[21:04] pgriess has joined the channel
[21:06] satori_ has joined the channel
[21:09] drudge: mscdex: getting 411 status back now
[21:09] drudge: http://pastie.org/private/tc7xb23rto89h4b0lq
[21:11] shimondoodkin: http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=http+411
[21:11] rnewson has joined the channel
[21:11] mscdex: drudge: try setting the Content-Length header
[21:12] shimondoodkin: if you do post it needs a content length
[21:12] mscdex: brb
[21:13] drudge: bingo
[21:13] drudge: works perfectly
[21:13] jackish has joined the channel
[21:14] drudge: badass.
[21:15] feroz_ has joined the channel
[21:19] ajpiano has joined the channel
[21:19] _announcer: Twitter: "Websocket en python backend is absurdly simpler than in Node.js, comes to be cowardice." [pt] -- Rafael Barros. http://twitter.com/rafaelhbarros/status/21778100555
[21:25] TooTallNate has joined the channel
[21:27] JimBastard has joined the channel
[21:28] mattly has joined the channel
[21:29] _announcer: Twitter: "Already have mega problems to squeeze four million old MySQL data on the # # NodeJS in CouchDB. Starts at so great." [de] -- Falco. http://twitter.com/ulmerleben/status/21778641284
[21:30] luddep_ has joined the channel
[21:31] steadicat has joined the channel
[21:35] codetonowhere has joined the channel
[21:35] Validatorian: I think I found a node bug... can someone verify? http://gist.github.com/542884
[21:39] MaSch: same behavior here
[21:54] softdrink has joined the channel
[21:57] herbySk: Validatorian: what's the bug on this?
[21:57] Validatorian: unexpected/undesired sorting
[21:57] ezmobius has joined the channel
[21:58] Validatorian: it's possible that node does it right and firefox doesn't, I'm not sure if there are specs on this, and what they are if so
[21:58] herbySk: Validatorian: it's an object, you never had the order of properties set in any way / stable / whatever
[21:59] herbySk: the fact that firefox keeps it is just an implementation detail
[22:00] herbySk: specs, if they say something, say "there is no guaranteed ordering of properties, with the exception of numbered properties in an array (they go first and ordered)"
[22:00] shimondoodkin: yes the behavior is different but it is ingnoreful because the app would work anyways
[22:00] shimondoodkin: other thing is that the numbers become strings
[22:01] herbySk: shimondoodkin: property keys are always strings
[22:01] shimondoodkin: yeah... i never tested it
[22:02] Validatorian: okay, so if I wanted to retain the order that I place them in, I need to not have them as numbers? if I make them 1x, 2x, 3x, the order is preserved.
[22:02] shimondoodkin: Validatorian: you could put it in an array to preserve order...
[22:02] herbySk: in fact, the spec says a[x] is done so that first x.toString() is executed and the result is used as key... so {'foo':'bar'}[{toString:function(){return 'foo'}}] should return 'bar'
[22:04] sephr has left the channel
[22:05] joshbuddy has joined the channel
[22:05] Validatorian: alright, I'll look into that. Thanks
[22:05] grahamalot has joined the channel
[22:06] herbySk: Validatorian: yes, to preserve order use an array: ["first", "second", "third"]
[22:06] inimino: Validatorian ⋱ Node leaves them in insertion order, it's only the pretty-printer that would sort them.
[22:06] inimino: Validatorian ⋱ But yes, when order is significant an Array should be used.
[22:07] shimondoodkin: does anybody knows a good routing thing, i couild copy its syntex from or use?
[22:08] LionMadeOfLions has joined the channel
[22:11] sh1mmer has joined the channel
[22:13] amerine has joined the channel
[22:17] shimondoodkin: found one i liked:http://github.com/defrex/node.routes.js/
[22:21] _announcer: Twitter: "Spent the last 2 days playing with #WebSockets and #NodeJS. Can't wait to write about it. Here is what I did http://bit.ly/cRxLbL" -- superfeedr. http://twitter.com/superfeedr/status/21781513390
[22:23] _announcer: Twitter: ".@superfeedr Ha. Cool. #WebSockets #NodeJS" -- Steve Malloy. http://twitter.com/sophware/status/21781622218
[22:31] mattly has joined the channel
[22:32] MaSch: damn >.< got no really great idea for the knockout
[22:33] ryah: MaSch: tank wars
[22:33] streampunk: MaSch: games in general are a good bet. ;-)
[22:34] streampunk: I'm not competing this year but I can't wait to see some of the fun stuff.
[22:34] TooTallNate: What are some other signals to listen for other than Ctrl+C?
[22:34] MaSch: i thought about an MMO Proof of Conecpt
[22:34] _announcer: Twitter: "Hats off to everyone able to ship node.js apps. Just spent another Saturday afternoon fighting with versions and documentation." -- Bryan Woods. http://twitter.com/bryanwoods/status/21782225824
[22:34] ryah: TooTallNate: man signal
[22:34] MaSch: man 7 signal !
[22:35] streampunk: MaSch: I wouldn't scale it too big. Just a few people interacting is good for starters.
[22:36] Pilate: Node seems to be dropping multiple set-cookie headers on the httpClient response... is there a way to keep them in tact?
[22:36] TooTallNate: Which ones can be invoked from the keyboard like Ctrl+C?
[22:36] MaSch: yes i know
[22:36] ryah: Pilate: it should put them together into one string
[22:36] ryah: with a comma
[22:37] MaSch: TooTallNate: http://en.wikipedia.org/wiki/Control-C google is your friend
[22:37] mmetzger has left the channel
[22:37] Pilate: ryah: i figured... but it isnt
[22:38] mikeal1 has joined the channel
[22:40] ryah: Pilate: hm
[22:40] Pilate: http://pastebin.com/xPBPMLEq
[22:40] Pilate: the top is what node prints, the bottom is from wireshark
[22:40] Pilate: its actually 4-5 set-cookie lines but it doesnt do more than the first
[22:40] teemow has joined the channel
[22:40] ryah: Pilate: loka
[22:40] ryah: er
[22:40] ryah: okay
[22:45] Pilate: ah it looks like it only combines 'cookie' fields not set-cookie?
[22:47] softdrink has joined the channel
[22:47] ryah: yeah
[22:47] ryah: fix coming
[22:48] Pilate: thanks a ton ;)
[22:53] warfangle has joined the channel
[22:53] warfangle: Hi there, just installed node and tried to start up the example.js to make sure everything was working correctly.
[22:54] warfangle: But it doesn't output the server running notification to my command line, and any reqs to 127.0.0.1 at the specified port timeout.
[22:54] warfangle: any hints?
[22:56] mattikus has joined the channel
[22:57] ryah: which example
[22:58] jacobolus has joined the channel
[23:04] _announcer: Twitter: "@davidcodes because learning, keeping up and developing in 1 language is a lot easier than coding in 2 or 3! And... #nodejs is blazing fast!" -- Jeroen van Duffelen. http://twitter.com/jeroenvduffelen/status/21783857379
[23:06] Tim_Smart has joined the channel
[23:07] MikhX has joined the channel
[23:09] mscdex: node.js rules!
[23:14] mtodd has joined the channel
[23:16] drudge: mscdex: my first attempt, mostly working https://gist.github.com/52fb6ec7a228e5c8bfdb
[23:17] MikhX has joined the channel
[23:18] _announcer: Twitter: "Hello world by #nodeJs on #heroku. Ninja at home!" -- Kadir Pekel. http://twitter.com/kadirpekel/status/21784635198
[23:19] rhymes_ has joined the channel
[23:20] rhymes_ has left the channel
[23:20] rhymes has joined the channel
[23:22] m64253 has joined the channel
[23:22] jackish has joined the channel
[23:23] trotter has joined the channel
[23:29] _announcer: Twitter: "Nice that he nodejs #: Learning Server-Side JavaScript with Node.js | Nettuts + http://forr.st/ ~ y1G" [cs] -- lethargicpanda. http://twitter.com/lethargicpanda/status/21785223482
[23:30] _announcer: Twitter: "execute applescript in node.js http://github.com/TooTallNate/node-applescript" -- aki. http://twitter.com/aki_xavier/status/21785299528
[23:31] TooTallNate: Sweet! I got tweeted!
[23:33] _announcer: Twitter: "A REST(HTTP) API for interacting with iTunes written in node.js http://github.com/TooTallNate/nTunes.git" -- aki. http://twitter.com/aki_xavier/status/21785424913
[23:33] mikeal1 has joined the channel
[23:34] dylang has joined the channel
[23:34] CIA-77: node: 03Ryan Dahl 07master * r2982e75 10/ (lib/http.js test/simple/test-http-set-cookies.js): set-cookies header is an array of values. always - http://bit.ly/ajHdMr
[23:34] ryah: Pilate: there--^
[23:37] steadicat has joined the channel
[23:38] mscdex: yay!
[23:38] mscdex: drudge: "Gist has been deleted"
[23:38] mscdex: drudge: good to hear though :-)
[23:39] drudge: mscdex: eek, https://gist.github.com/800275b894d8c8bfb645
[23:40] ryah: that set-cookie fix probably can't be landed in the 0.2 branch
[23:40] mscdex: why not?
[23:40] ryah: api cahnge
[23:40] mscdex: oh
[23:40] mscdex: first 0.3.x commit!
[23:40] mscdex: :-D
[23:42] mscdex: drudge: instead of relying on base64_encode, you could create a new buffer and do a .toString('base64') on it
[23:42] drudge: ok
[23:43] mscdex: you could probably also do the same for utf8_encode
[23:43] drudge: it's only used in the base64_encode :P
[23:43] mscdex: oh
[23:43] mscdex: well much easier then :-p
[23:44] drudge: yeah that worked great, thanks. didn't even think of the Buffer class
[23:45] mscdex: buffers to the rescue!
[23:45] robotblake has joined the channel
[23:46] Pilate: thanks ryah =)
[23:46] jackish has joined the channel
[23:49] TooTallNate: What's the best way to listen for an HTTP client closing the connection early (i.e. XHR.abort())?
[23:49] jtoy has joined the channel
[23:49] mscdex: TooTallNate: probably the 'end' event
[23:50] TooTallNate: on the HTTP or Stream instance though?
[23:50] ryah: TooTallNate: req.connection.on('close', fn)
[23:50] TooTallNate: ryah: Thanks, that was more like what I was expecting
[23:50] ryah: TooTallNate: or maybe req.connection.on('end', fn)
[23:51] mscdex: TooTallNate: i've noticed that IE sends RST which triggers 'close', but all other browsers send FIN which triggers 'end' quicker
[23:51] chuhnk has joined the channel
[23:51] TooTallNate: mscdex: interesting, i'll have to experiment
[23:51] chuhnk: hi, im trying to compile node.js on mac i386 but it keeps detecting as x86_64, i've tried archflags but it doesnt help. any ideas?
[23:52] mscdex: i'm not sure if it's the same with all versions of IE, but that's what i saw happen with IE6
[23:53] mscdex: there are 32-bit x86 macs?
[23:53] mscdex: :S
[23:53] chuhnk: macs like 4 years old with snow leopard on it
[23:54] Aria: Yeah, rather a lot of 32-bit X86 macs.
[23:55] CIA-77: node: 03Ryan Dahl 07eventsource * rabc77b6 10/ (lib/fs.js src/node_stat_watcher.cc src/node_stat_watcher.h): StatWatcher to use EventSource (+8 more commits...) - http://bit.ly/cXpWlo
[23:56] mscdex: heh shows how much i know about macs :p
[23:58] davidwalsh has joined the channel
[23:58] davidwalsh has joined the channel
[23:58] chuhnk: any ideas on the compiling?
[23:59] mscdex: sorry, i only use linux and bsd :-\
[23:59] ryah: chuhnk: hm - waf makes that decision
[23:59] ryah: chuhnk: uname -a ?