--- Log opened Thu Feb 21 00:00:02 2013
--- Day changed Thu Feb 21 2013
00:00 < cianomaidin> ralphtheninja: 2014 will be the real year of p2p
00:00 < cianomaidin> ralphtheninja: 2013 will be the year of node, + a little p2p
00:00 < ralphtheninja> cianomaidin: I stand corrected :)
00:00 < cianomaidin> ralphtheninja: hey don't let me correct you;
00:01 < cianomaidin> ralphtheninja: I'm the person who is most often wrong
00:01 < cianomaidin> ralphtheninja: we can interchange dates and titles
00:01 < cianomaidin> ralphtheninja: but one way or the other the shit is going down
00:32 < dominictarr> Raynos: anyway
00:32 < Raynos> oh hi
00:32 < Raynos> cianomaidin: o/
00:32 < dominictarr> so we found this guy, and nearform is organizing a workshop where he's gonna come to ireland and teach us everything he knows
00:33 < Raynos> nice.
00:33 < cianomaidin> Raynos: hey yo
00:33 < Raynos> I want to be there to build topologies
00:33 < Raynos> and stuff
00:33 < Raynos> if we can have proper infrastructure that scales
00:33 < Raynos> then the rest is building js apps
00:35 < dominictarr> yes!
00:35 < dominictarr> we've got to show him that node.js has lots of crazy people who are all building next-level stuff
00:36 < dominictarr> that it's a dynamic systems language
00:36 < dominictarr> mbalho: https://github.com/dominictarr/semver-ftw
00:37 < dominictarr> mbalho: http://www.youtube.com/watch?v=MNzMNjW8oGs
00:40 < dominictarr> Raynos: so, you are coming to ireland then? it's at the end of march!
00:41 < Raynos> I want to
00:41 < Raynos> not sure whether I can
00:41 < Raynos> have a friend visiting SF 14th - 21st march
00:42 < ralphtheninja> oh bummer
00:42 < dominictarr> oh, it's not gonna clash with that
00:42 < ralphtheninja> it's on the 31st of March and 1st of April
00:42 < cianomaidin> Raynos: but 30th and 31st march is not quite a clash
00:42 < Raynos> oh then it may work
00:42 < Raynos> I just need to mentally condition myself
00:42 < ralphtheninja> make it work! :)
00:42 < cianomaidin> Raynos: in fact it leave a good buffer for guinness
00:43 < Raynos> to fly to ireland
00:43 < dominictarr> Raynos: also 1st April is my birthday!
00:43 < cianomaidin> Raynos: Dominic is 30
00:43 < ralphtheninja> we're going to have the workshop on dominics boat
00:43 < ralphtheninja> :p
00:43 < dominictarr> I'm getting a boat for my 30th? thanks nearform!
00:44 < ralphtheninja> hehe
00:44 < cianomaidin> Raynos: 31st could be an interesting day ; p2p in a boat
00:44 < cianomaidin> boat peer
00:44 < ralphtheninja> it's a submarine
00:44 < cianomaidin> ?
00:44 < cianomaidin> day
00:44 < cianomaidin> birth boat peer day
00:44 < ralphtheninja> that would be cool though, to have a conference under the sea
00:45 < cianomaidin> subconf
00:45 < ralphtheninja> hehe
00:45 < ralphtheninja> pubconf
00:45 < cianomaidin> we could get substack to be the ship captain
00:46 < ralphtheninja> it would be awesome if you could make raynos, we need some bad ass dudes there
00:46 < ralphtheninja> make it*
00:46 < Raynos> are we really doing this
00:46 < Raynos> on a submarine
00:46 < Raynos> in ireland?
00:47 < dominictarr> on a submarine in international waters
00:47 < Raynos> send me some details by gist or email to get me really pumped and ill fly down
00:47 < Raynos> but first
00:47 < ralphtheninja> by gist? lol
00:47 < Raynos> RAGEQUIT @ STREAMS 2
00:47 < LOUDBOT> AND THEN IT KINDA JUST SNOWBALLED FROM THERE
00:51 < Raynos> I recommend you just never use node streams, ever. It's easier that way. Just use ruby on rails
00:52 < Raynos> you can scale horizontally by buying more dynos from heroku
00:52 < mbalho> you laugh but that is a more viable model for lots of people :)
00:57 < Raynos> I know
00:58 < dominictarr> that is only if you are building CRUD apps
00:58 < substack> Raynos: I just use the old streams because they work and I don't need to upgrade anything
00:58 < substack> I'll use streams2 when it's easy and everyone is already using them
00:58 < dominictarr> that is a seam that hasn't been totally mined yet, but it has been worked over
00:59 < Raynos> substack: I have zero confidence in that back pressure works
00:59 < Raynos> I hate back pressure. it's hard.
00:59 < dominictarr> substack: if you use Through#queue then it will just drop into streams2
00:59 < dominictarr> you'll just go require('through/streams2')
00:59 < dominictarr> and it will be exactly the same API but give you streams2
01:00 < dominictarr> duplex will do that as well.
01:00 < substack> Raynos: just use through and other libs to handle all that crap
01:00 < Raynos> no
01:00 < substack> core libs are low-level and hard
01:00 < Raynos> thats called buffer all the things in memory
01:00 < substack> find a module to set watermarks
01:00 < Raynos> I am, it's called readable-stream
01:00 < Raynos> and it's annoyingly difficult to get right
01:04 < Raynos> streams2 are a good idea. they are just hard. which is frustrating
01:04 < substack> pesky
01:04 < Raynos> once they are "done"
01:04 < Raynos> then everything is good
01:04 < Raynos> and nice
01:04 < dominictarr> Raynos: streams are hard. it took me 6 months before I realized I could just make through
01:05 < Raynos> btw I wrote https://gist.github.com/Raynos/d76ef926edb849cd10a1 earlier
01:09 < Raynos> as documentation of what streams2 readable's look like
01:09 < substack> I just use through because I can just write 1 or 2 functions and it all pretty much works
01:09 < Raynos> dominictarr: queue will need to return a boolean to say whether the buffer is full and whether you should stop queuing data
01:10 < Raynos> thats what push() does
01:10 < Raynos> otherwise through's buffer fills up forever
01:10 < dominictarr> you can just check the length
01:10 < dominictarr> if you can't just stick things in, then you have the same problem
01:11 < dominictarr> through is for small streams
01:11 < dominictarr> it's the readable at the start of the pipeline that needs to be responsible for not bringing too much into memory
01:12 < Raynos> I guess your right
01:12 < Raynos> unless one of your through's is an expand
01:12 < Raynos> in which case through's later down the pipeline may care
01:13 < dominictarr> mostly through isn't used like that.
01:13 < dominictarr> I have never actually written an expand stream
01:17 < Raynos> json parse is expand
01:17 < Raynos> conceptually
01:27 < mbalho> dominictarr: sent you a pull req
01:36 < dominictarr> mbalho: whats a story about the bug in pepperoni()?
01:38 < mbalho> dominictarr: the chef got fired
01:38 < mbalho> for not using semver
01:41 < mbalho> seagulls in wellington are pretty aggressive towards one another
01:43 < dominictarr> maybe the issue that it was pepperoni supplier turned out to be from a factory farm, and then it's upgraded to an organic artisanal salami supplier.
01:47 < Raynos> dominictarr: semantically can you change a feature without breaking back compat
01:47 < Raynos> and should that be major or minor
01:48 < dominictarr> that should be a patch
01:48 < dominictarr> if it doesn't change the surface api
01:48 < dominictarr> like if you go through and refactor it's a patch
01:50 < Raynos> i mean change the semantics
01:50 < Raynos> but not in a back compat breaking fashion
01:50 < Raynos> i guess thats a new feature
01:50 < Raynos> or your wrong about "not breaking backcompat"
01:51 < Raynos> does anyone have a shim for 0.9 setImmediate ?
01:54 < mbalho> Raynos: if i depend on a module by doing ~0.0.1 and the maintainer ships 0.1.0 after they 'change semantics' but it has the exact same API then i would say that is less ideal because my module wont get the refactor until i upgrade my package.json
01:54 < Raynos> thats abuse
01:54 < Raynos> then depend on ~0.0
01:54 < mbalho> which part
01:54 < Raynos> which means you get 0.0 -> 0.9.9
01:55 < mbalho> i only want 0.0.0 - 0.0.infinity though since that is the range that the api will still work as per semver
01:56 < mbalho> what is an example of changing semantics withotu adding features
01:56 < mbalho> that isnt a bug fix/refactor
01:56 < Raynos> changing semantics is adding or removing features
01:56 < Raynos> im too tired for this
01:58 < dominictarr> Raynos: what do you mean 'change semantics'?
01:58 < mbalho> Raynos: do you have a concrete example of changing semantics without introducing any new apis
01:58 < Raynos> i dont know
01:58 < Raynos> no
01:58 < Raynos> no examples
01:58 < Raynos> im wrong
01:58 < Raynos> i retract whatever I said
01:58 < dominictarr> if the api has changed, it's but is backcompat, that is minor
01:59 < dominictarr> if it's changed and not backcompat, that is major
01:59 < dominictarr> if the api has not changed, it's a patch.
01:59 < mbalho> i think maybe there are 2 definitions of backcompat here
02:00 < mbalho> one where the program doesnt break and one where the program works as expected
02:00 < mbalho> subhtle but imporotant difference
02:01 < dominictarr> that is a blury difference, I think
02:05 < mbalho> basically if you push a new version and it creates a new bug in my module then it should be a major version, e.g. the 'api' is the same but the usage (semantics?) are different
02:16 < Raynos> backcompat means a test suite with 100% coverage
02:17 < Raynos> still passes
02:17 < Raynos> or whatever
02:17 < Raynos> Who gives a fuck
02:17 < Raynos> we know what it means
02:17 < Raynos> well actually backcompat is blury
02:17 < Raynos> who cares.
02:19 < mbalho> Raynos: ... you brought it up
02:20 < Raynos> :D
02:20 < Raynos> shoot me.
02:20 < dominictarr> so, this is "BEST EFFORT"
02:20 < dominictarr> 100% code coverage might not cover all states / code paths
02:20 < mbalho> unless you use haskell
02:20 < dominictarr> yeah, or ada2012
02:21 < dominictarr> but if you publish a version that was a breaking change, just fix that and publish a patch
02:22 < dominictarr> like, if you publish a bug, or a syntax error, just publish a new patch immediately
02:44 < dominictarr> mbalho: http://dominictarr.github.com/semver-ftw/
02:44 < dominictarr> http://rumoursdb.com/examples/css-colab/editor.html#semver-ftw
02:45 < dominictarr> ^ working on the css
02:49 < defunctzombie> ^ this is why you MUST ALWAYS PIN modules
02:49 < defunctzombie> because no one will do it right and no one can be trusted.. all changes are potential breaking changes :)
02:49 < defunctzombie> substack: https://github.com/shtylman/node-required/issues/6
02:49 < defunctzombie> substack: I think the resolve function will do what he wants.. but I want to tell him he should stop hahah
02:51 < Raynos> why do we even need node-required
02:51 < Raynos> why not just parse package.json recursively
02:51 < Raynos> and fuck relative requires
02:51 < Raynos> if your doing relative requires you should put it in npm
02:51 < Raynos> ONE PACKAGE. ONE FILE.
02:51 < LOUDBOT> JUST CALL MY PENIS RUM TUM TUGGER AND MY BALLS SIMBA!
02:58 < defunctzombie> Raynos: what you said makes no fucking sense :)
02:59 < defunctzombie> Raynos: I want my deps in my source so I don't have to go hunting in 300 places for them
03:00 < Raynos> it was abad joke
03:01 < Raynos> just have 5000 line files
03:04 < defunctzombie> bad joke cat does not approve
03:04 < defunctzombie> I will concatenate all my files into 1 giant file
03:33 < dominictarr> Raynos: "it's too late to change now"
03:33 < dominictarr> that is what isaacs will say.
03:34 < dominictarr> but we can make node^2 like that
03:34 < dominictarr> it would be cool if require('module')
03:35 < dominictarr> returned the index if it required the current module
03:35 < dominictarr> then examples could just do require('me')
03:35 < dominictarr> instead of require('..')
03:35 < defunctzombie> dominictarr: that would work if requires looked up the tree for package.json as well as node_modules
03:36 < dominictarr> yeah, exactly
03:36 < dominictarr> and then require('me/file2')
03:37 < dominictarr> so every require would be relative to the current file, although, if you changed the module name, you'd have to update all of that stuff.
03:38 < dominictarr> but it would be great for things like tests and examples, which depend on the module, but arn't depended on.
03:38 < dominictarr> because that code would look the same as normal user code.
03:39 < dominictarr> so, require(X) would look in package.json (if name==X) and then node_modules/X then ../package.json, then ../node_modules/X etc....
03:40 < dominictarr> or, should it be node_module/X first?
03:41 < dominictarr> adding this could beak a few rare cases, like if you used a symlinked node_modules/X -> ../../
03:43 < dominictarr> defunctzombie: even if you do pin everything, you should still TRY to use semver correctly
03:43 < defunctzombie> dominictarr: 100% agree
03:43 < dominictarr> because it shows intension.
03:44 < defunctzombie> dominictarr: semver is for humans
03:44 < defunctzombie> pinning is for history
06:29 < Raynos> WHY ARE STREAMS SO HARD >:(
06:29 < LOUDBOT> MY WIFE LEFT ME AND ALL I GOT WAS THIS BOX OF CHOCOLATES
06:29 < Raynos> ragequit forever.
07:30 < Raynos> substack: +1 for 0.8 streams
07:32 < dominictarr> Raynos: streams are just hard. they are hard because the amount of state they have.
07:32 < substack> Raynos: I think we just need to let the dust settle for a bit longer
07:33 < Raynos> oh for sure
07:33 < substack> we'll come up with good idioms but it takes time
07:33 < dominictarr> it's not many states, but it's enough to make it challenging.
07:33 < Raynos> https://github.com/joyent/node/issues
07:33 < Raynos> I started _really_ using streams2 and had to open 5 issues just today
07:33 < dominictarr> --and the amount of states is fundamental to streams, whether they are streams1 or streams2
07:34 < Raynos> dominictarr: streams2 have crazy states
07:34 < Raynos> streams2 do too much, simply put. we need to figure out what the 50% of bullshit is and bin it
07:35 < dominictarr> streams 1 was nice.
07:36 < substack> vintage
07:36 < substack> heirloom streams
07:36 < Raynos> well I like that buffering is build into streams2
07:36 < dominictarr> "classic streams"
07:36 < Raynos> and you can configure it watermarks
07:36 < dominictarr> like "classic coke" and "new coke"
07:36 < substack> userspace can handle buffering pretty easily though
07:37 < dominictarr> Raynos: but you could just use a streams1 base class that does buffering with watermarks
07:37 < dominictarr> you need to turn watermarks off for protocol streams, anyway
07:37 < Raynos> why?
07:38 < Raynos> dominictarr: the nice thing about buffering in core is that all 3rd party streams I use respect it
07:38 < Raynos> like req
07:38 < Raynos> & res
07:38 < dominictarr> because if I connect to a scuttlebutt and send a clock {id: id, clock: {}}
07:38 < dominictarr> because it's a new document, but is less than the water mark it will get buffered
07:39 < dominictarr> it won't send any more because it's waiting for more data from the other end.
07:39 < Raynos> in fact it means you can use streams with express
07:39 < Raynos> i hope
07:39 < Raynos> oh wait no
07:39 < Raynos> bodyParser fucks you over
07:39 < dominictarr> it would be silly to send {id: id, clock: {}} \n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
07:40 < dominictarr> just to get past a watermark
07:40 < dominictarr> I made a connect like thing that worked with streams
07:40 < dominictarr> but tj didn't like it
07:41 < dominictarr> http://github.com/dominictarr/mw-pipes
07:42 < Raynos> dominictarr: but just configure the lwm
07:42 < Raynos> i like readable watermarks
07:42 < Raynos> which are different from writable watermarks
07:42 < Raynos> / if read() and then below hwm then _read
07:42 < Raynos> // if push() and then below lwm then _read
07:43 < dominictarr> sure, just inherit from a WaterMarkedStream
07:43 < Raynos> yeah sure :D
07:43 < Raynos> i agree streams2 is complex
07:43 < Raynos> but fuck it
07:43 < Raynos> we'll see how it goes
07:45 < dominictarr> RAGEUNQUIT
07:46 < substack> SIGRAGE
07:47 < dominictarr> ERAYNOS
07:48 < substack> http://www.sydneyoperahouse.com/priority/Kraftwerk_Rego.aspx?utm_source=wordfly&utm_medium=email&utm_campaign=Inhouse20Feb&utm_content=version_A
07:48 < substack> left .au TOO SOON
07:54 < dominictarr> oh wow
09:23 < Raynos> https://github.com/Raynos/read-stream
09:23 < Raynos> I can finally sleep!
12:36 < dominictarr> substack: is it possible to put two browserify scripts on the same page?
13:11 < ralphtheninja> dominictarr: is it possible to setup rumours in the client without using browserify?
13:12 < dominictarr> yes
13:12 < dominictarr> point a script tag at /rumours.js
13:13 < ralphtheninja> mind joining #nearform?
13:18 < ralphtheninja> dominictarr: ✔
13:19 < dominictarr> oh, oops
13:22 < ITpro> Peter?
16:18 < defunctzombie> is anyone else experiencing npm publish issues on 0.8.20/1.2.11 ?
16:26 < jesusabdullah> 0.8.20? Dang open sauce moves fast
16:27 < CoverSlide> true story bro
16:29 < jesusabdullah> oh siiick no git tag
16:41 < defunctzombie> isaacs: weird npm publish issue
16:41 < defunctzombie> isaacs: I had a package.json inside one of my example folders
16:42 < tanepiper> AvianFlu: have you abandoned ntwitter?
16:42 < defunctzombie> isaacs: it was malformatted (not evne part of my repo) and npm publish would not publish the main project as a result
16:52 < isaacs> defunctzombie: sounds like a bug in fstream-npm
16:53 < isaacs> defunctzombie: patch welcome
16:53 < defunctzombie> isaacs: don't really know what that means
16:55 < isaacs> defunctzombie: fstream-npm is a node module that walks over a dir tree applying npm's ignore/include rules
16:55 < isaacs> defunctzombie: sounds like it should just not treat any non-parsing package.json files as anything special.
16:56 < isaacs> defunctzombie: but, really, it shouldn't be a surprise that a file called "package.json", which isn't json, will confuse npm.
17:02 < defunctzombie> isaacs: yea, it should be a surprise
17:02 < defunctzombie> that a non root file breaks it
17:04 < isaacs> defunctzombie: well, what if it's the root of some new thing?
17:05 < isaacs> defunctzombie: but sure, it should only treat them as special perhaps in ./(node_modules/)*package.json paths
17:05 < defunctzombie> isaacs: when I am publishing a module
17:05 < defunctzombie> there is only one root tho
17:05 < isaacs> er, ./(node_modules/[^/]+/)*package.json
17:05 < defunctzombie> everything else would be published under that module
17:05 < isaacs> defunctzombie: bundledDependencies
17:05 < defunctzombie> which are under node_modules tho right?
17:05 < isaacs> yeah
17:06 < isaacs> ./(node_modules/[^/]+/)*package.json
17:06 < isaacs> but what if you put a package in ./lib/foo./
17:06 < isaacs> and then you have a build script in there
17:06 < isaacs> and you have in there a "files":["blah/blerg"]
17:06 < isaacs> but your root .npmignore says "ignore all blergs"
17:06 < isaacs> and now your bundled lib/foo/ package is broken
17:06 < isaacs> because its package.json file isn't respected.
17:07 < defunctzombie> run that by me again?
17:07 < defunctzombie> if you put a package in ./lib/foo
17:07 < defunctzombie> that isn't a module as far as npm is concerned tho
17:07 < isaacs> sure it is
17:07 < isaacs> you can do require('./lib/foo/') and you'll load it
17:07 < defunctzombie> I cannot require('foo')
17:08 < isaacs> it's not a dependency
17:08 < isaacs> but it's still amodule
17:08 < isaacs> and npm shouldn't destroy it
17:08 < defunctzombie> sure, but it isn't an npm root level module
17:08 < defunctzombie> correct
17:08 < defunctzombie> npm should not do anything with it
17:08 < defunctzombie> npm install will do nothing
17:08 < defunctzombie> npm publish will do nothing
17:08 < defunctzombie> as far as top level npm is concerned it is just some files
17:08 < defunctzombie> for my overall module
17:52 < defunctzombie> rowbit: kill all the dissenters!
18:31 < Ikeyman> Hello!
19:16 < elliottcable> pkrummins, substack ⑊ er. Does t-CI not operate on any branch except Master, or what?
19:16 < elliottcable> can I change that?
20:01 < Ikeyman> Hey i need help!
20:09 < tanepiper> http://scotlandjs.com/proposals.html
20:09 < Raynos> isaacs, dominictarr: https://github.com/Raynos/read-stream/issues/4 when you have some time, I would love some peer-review on whether I interpreted streams2 correctly
20:15 < Ikeyman> any way to get a image of linux running stackvm?
20:41 < defunctzombie> substack: do you have something on top of dnode that does events?
20:41 < defunctzombie> like if I want the remote side to emit events?
20:43 < Raynos> just emit events ;)
20:43 < Raynos> events are just functions
20:44 < Raynos> dominictarr: were you serious about the submarine?
20:44 < defunctzombie> Raynos: can I emit on a remote?
20:44 < defunctzombie> I guess I can just try it haha
20:44 < defunctzombie> but I wasn't sure if all that magic was hooked up too
20:45 < dominictarr> Raynos: I am flattered that you think I may be that crazy.
20:45 < Raynos> cmon
20:45 < Raynos> I'm not going to fly to ireland
20:45 < defunctzombie> dominictarr: I have an idea
20:46 < Raynos> unless we do the entire thing on a submarine
20:46 < Raynos> preferably with a map of the internet
20:49 < Raynos> defunctzombie: when are you in SF again? You are coming to nodeconf right?
20:49 < defunctzombie> Raynos: google botched the scheduling, so I think I will be in SF mid march
20:49 < Raynos> cool
20:49 < Raynos> need a place to crash?
20:51 < defunctzombie> Raynos: yea, I think I will
20:51 < defunctzombie> Raynos: I want to bounce between a few places and see what I can
20:52 < substack> elliottcable: yep, only master works on testling-ci right now
20:52 < Raynos> cool
20:52 < Raynos> let me know when you arrive
20:52 < Raynos> i'm booked between 14th & 21st
20:53 < defunctzombie> Raynos: cool, will do
20:57 < defunctzombie> substack: how goes the quest for browserify v2? any new monsters slain?
21:39 < defunctzombie> Raynos:  I don't follow what you mean about emitting over dnode
21:40 < Raynos> MANY CATS
21:40 < LOUDBOT> THOUGH YELLOWFIN DOES DO THAT AS WELL AS AHI DOES
21:40 < defunctzombie> cats are not helping me at this time
21:40 < Raynos> client calls remote.on(...)
21:40 < Raynos> remote calls self.emit(...)
21:40 < Raynos> client's callback get's called
21:41 < defunctzombie> Raynos: and that will jsut hapen to work because remote is a stream
21:41 < defunctzombie> which is an event emitterish
21:41 < defunctzombie> type object?
21:41 < defunctzombie> I think I follow
21:41 < Raynos> it may or may not work
21:41 < Raynos> it only works
21:41 < Raynos> if dnode allows RPC calls of `on` and `emit`
21:42 < Raynos> and doesnt treat them as special reserved methods that can't be used for RPC
21:42 < Raynos> if it does the latter you can just ragequit forever
21:42 < defunctzombie> Raynos: seems like I need to add the methods myself
21:43 < defunctzombie> so doesn't appear to be anything special
21:43 < defunctzombie> is there a "correct" way to do this tho
21:43 < defunctzombie> so the object I pass to dnode
21:43 < defunctzombie> the properties are exposed
21:44 < defunctzombie> I need to make that object an event emitter, I hope that is possible heh
21:45 < Raynos> i dont know
21:45 < Raynos> what the correct way is
21:45 < Raynos> btw
21:45 < Raynos> RPC over events
21:45 < Raynos> is stupid
21:45 < Raynos> use methods
21:46 < Raynos> or streams
21:46 < defunctzombie> the whole point is I have some event on the server
21:46 < defunctzombie> that I want to notify the client of
21:46 < defunctzombie> event based interface
21:46 < Raynos> just use pubsub :3
21:46 < Raynos> pubsub all the things.
21:47 < defunctzombie> that is what .on .emit are
21:47 < Raynos> </end-shitty-advice>
21:50 < defunctzombie> wooo, got what I wanted to work
21:50 < defunctzombie> fuck yea
22:07 < st_luke> defunctzombie: are you in the city?
22:07 < defunctzombie> st_luke: yep
22:09 < st_luke> i think just sweet is the best public hacking spot in the city currently
22:09 < st_luke> after surveying a lot of places
22:10 < defunctzombie> st_luke: hahah nice
22:11 < defunctzombie> st_luke: is that where you work out of now?
22:11 < st_luke> defunctzombie: haha I should
22:12 < st_luke> I've been working out of GA lately but it's kind of a shitty environment
22:12 < defunctzombie> yea
22:12 < defunctzombie> wow
22:15 < st_luke> i mean the company im working for is great but they just happen to be renting GA space because it's convenient right now
22:16 < defunctzombie> sure
22:16 < st_luke> can usually only do ~3 hours of GA per day though because it's pretty insufferable
22:22 < substack> defunctzombie: I would probably build a thing to control a mux-demux stream alongside of dnode
22:22 < substack> mdm -> [ dnode, ...streams... ]
22:22 < defunctzombie> substack: I was able to do it with dnode quite simply actually
22:23 < substack> small amounts of data will work find using only dnode but it's not ideal for lots of data
22:24 < defunctzombie> substack: how large?
22:24 < defunctzombie> and do you mean as callback results?
22:25 < defunctzombie> I am basically using it to replace some of my REST calls
22:25 < defunctzombie> or want to try it for that
22:31 < substack> should be ok for that
22:53 < substack> dominictarr: so the bug I'm getting with browserify v2 is that my through streams are getting paused and not being resumed
22:54 < substack> and then the event loop empties out when there are paused streams
22:54 < chrisdickinson> i may have been having a similar problem in glslify-stream, iirc: https://github.com/chrisdickinson/glslify-stream/blob/master/index.js#L61-L66
22:55 < chrisdickinson> (also, the linked section is a massive hack ): ): )
22:55 < dominictarr> substack: are you using all through streams?
22:55 < dominictarr> what is the pipeline?
22:57 < substack> one sec
23:00 < substack> dominictarr: https://github.com/substack/node-browserify/blob/v2/index.js#L77
23:01 < dominictarr> are they all through streams?
23:02 < dominictarr> it might be that one is not emitting drain
23:02 < dominictarr> that would get them stuck
23:02 < substack> yep they are all through streams
23:06 < elliottcable> substack ⑊ is there something I can do about that? It works with “Master” (notice my branch is capitalized), as well.
23:06 < elliottcable> substack ⑊ do I have to set it as my GitHub primary-branch or something?
23:09 < dominictarr> elliottcable has to be Different
23:10 < defunctzombie> hahaha
23:11 < elliottcable> lol.
23:11 < defunctzombie> he has abandoned us!
23:12 < elliottcable> it already works with my “different”
23:12 < elliottcable> but I need it to work with my development branches as well. I don't want to push these changes to Master just to test them. >:
23:27 < dominictarr> that is it, I'm gonna start using the M4s73R!!!1 branch, to out do elliottcable
23:28 < elliottcable> dominictarr ⑊ Enjoy yourself.
23:42 < pkrumins> hey guys, we've added iphone 6, ipad 6, and safari 6 to testling-ci. Check out this badge Raynos just got for his project: http://ci.testling.com/Colingo/deep-merge.png
23:43  * mint_xian feels bad for IE8
23:43 < defunctzombie> pkrumins: is that an https badge? :p
23:44 < pkrumins> sorry not yet
23:44 < defunctzombie> I like it :)
23:44 < pkrumins> promise to get https the next week
23:44 < pkrumins> i dont know how to add https to nodejs that's why there is a delay
23:44 < defunctzombie> gotta get that shit so people can add to github
23:44 < defunctzombie> pkrumins: I wouldn't do it in node
23:44 < substack> I drew those images from the queensland state library in brisbane
23:44 < defunctzombie> use some other handler
23:44 < elliottcable> exciting
23:45 < elliottcable> substack ⑊ I want to make sweet, sweet love to your art.
23:45 < substack> pkrumins: browserling.com is on https already so could just see what webapp/server.js does
23:45 < substack> or bouncer/server.js rather
23:46 < pkrumins> it doesnt do it correctly though
23:46 < Raynos> we have a proxy for this stuff
23:46 < elliottcable> substack ⑊ I'd be happy to pay for a “PROFESH EDITION” Testling-CI account, by the way.
23:46 < pkrumins> or maybe it does.
23:46 < elliottcable> especially much so, if you could work in branches sometime this weekend. ;)
23:46 < substack> elliottcable: awesome!
23:46 < Raynos> thats starts two HTTP servers and has HTTPS forward to HTTP
23:47 < Raynos> `httpProxy.createServer(proxy_options, httpProxyPort, "localhost")
23:47 < Raynos>     .listen(httpsProxyPort)`
23:47 < elliottcable> I don't need any private repos at the moment; but priority on execution so the CI returns faster would be quite nice, if that's wrangle-able
23:48 < pkrumins> Raynos: yes but we use bouncy
23:48 < pkrumins> i think it will work the way browserling https works
23:49 < Raynos> maybe dont render partial results
23:49 < Raynos> to discourage me from hitting f5
23:49 < Raynos> waiting for it to be done
23:49 < pkrumins> looks great http://ci.testling.com/Raynos/read-stream.png
23:49 < Raynos> https://github.com/Raynos/read-stream also has all the badges
23:50 < Raynos> someone PR ie8 support for readable streams :p
23:50 < dominictarr> substack: I'm super busy right now, can you make a script to reproduce the bfy2 issue?
23:55 < pkrumins> elliottcable: we'll have the branches working soon
23:56 < elliottcable> love you lots.
23:56 < elliottcable> where do I give you more money?
23:56 < pkrumins> don't you have a browserling plan already?
23:56 < elliottcable> yeah, streaming results so I don't have to reload, would be great
23:56 < elliottcable> pkrumins ⑊ mmhmm
23:56 < elliottcable> can't afford one of your higher plans, so I can't just up the plan; but I'd happily pay twice or three times what I'm paying now.
23:56 < pkrumins> we don't have other payment options at the moment
23:56 < elliottcable> mmmm okay
23:57 < substack> elliottcable: yep streaming results would be pretty great
23:57 < elliottcable> before that, and far more useful, and probably easier:
23:58 < elliottcable> indication of which browsers are *currently running*, when you load the page
23:58 < substack> I just need to hook the page into a results stream
--- Log closed Fri Feb 22 00:00:14 2013