00:00:50  <isaacs>workign on http.js is so tedious. change one line, make test. change another line, make test. change another line, make test, 8000 FAILURES OMG. undo that change, read a hundred lines of code...
00:01:41  <bnoordhuis>it's code that's reached critical mass
00:01:45  <mjr_>isaacs: I can out-pain you on that one. Repeated make test on our SmartOS production machines often leads to a lockup of the whole system for a few minutes.
00:02:08  * perezdquit (Quit: perezd)
00:02:11  * piscisaureus_quit (Ping timeout: 260 seconds)
00:02:30  <isaacs>bnoordhuis: it reached critical mass a year ago
00:02:38  <isaacs>it's full on grimy now
00:03:20  * perezdjoined
00:08:18  <dap>bnoordhuis: I figured out the bug in my code that was causing accept() to return EINVAL, thanks to https://github.com/bnoordhuis/node/commit/83e9ce5#diff-0. I'd love to see that integrated in case it happens again.
00:09:28  <bnoordhuis>dap: i've already landed it, it'll be in the next 0.6 release
00:10:16  <dap>great! thanks
00:25:04  * mikealquit (Quit: Leaving.)
00:27:04  * perezdquit (Quit: perezd)
00:27:34  * TooTallNatejoined
00:28:20  <bnoordhuis>ircretary: tell piscisaureus https://github.com/bnoordhuis/node/compare/v0.6...accept-backoff
00:28:21  <ircretary>bnoordhuis: I'll be sure to tell piscisaureus
00:29:06  <CIA-155>libuv: Ben Noordhuis master * rad02404 / test/runner.c : test: prevent accidental division by zero - http://git.io/XtzCwQ
00:30:59  * travis-cijoined
00:31:00  <travis-ci>[travis-ci] joyent/libuv#272 (master - ad02404 : Ben Noordhuis): The build is still failing.
00:31:00  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/985b961...ad02404
00:31:00  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1300995
00:31:00  * travis-cipart
00:34:49  <tjfontaine>mjr_: I have a replacement dns module (in js) that only reads /etc/hosts once (or optionally enables a watch on it for auto reloading), but I'm not sure it's ready for your scale
00:35:20  <mjr_>tjfontaine: sounds cool
00:36:12  <tjfontaine>but if /etc/hosts ever actually becomes your bottleneck there is at least something to try :)
00:38:07  <tjfontaine>still finishing up the reading on backlog, it would also be trivial to add an in process cache to it
00:38:53  <mjr_>yeah, I'll put a simple app cache in there. It should be really easy.
00:39:26  <tjfontaine>I mean wrt my dns stuff, you get the real ttl information for free there without having to consult c-ares internals
00:39:49  <mjr_>Oh cool.
00:41:07  <mjr_>Here is an exciting thing, apropos of nothing: certain proxies at mobile network operations in south and central america are corrupting HTTP data. We see certain characters in our JSON that have the 7th bit flipped.
00:41:27  <mjr_>Sometimes this results in a JSON.parse exception, but sometimes it still makes valid JSON. Hilarious.
00:42:08  <tjfontaine>very cute
00:42:08  * loladiroquit (Read error: Connection reset by peer)
00:42:13  * loladiro_joined
00:43:25  * brsonquit (Ping timeout: 252 seconds)
00:48:16  * pieternquit (Quit: pietern)
00:56:22  * ericktquit (Ping timeout: 244 seconds)
00:57:18  * mikealjoined
00:57:54  <isaacs>mjr_: ok
00:58:00  <isaacs>mjr_: got that nextTick split up to a reasonable size: https://github.com/isaacs/node/compare/joyent:v0.6...isaacs:issue3236
01:02:02  <isaacs>bnoordhuis, TooTallNate, piscisaureus: review? https://github.com/isaacs/node/commit/e04de2de165eaeb3685a340a86971b6383214fce
01:02:25  <tjfontaine>you're going to be all over the blame now :)
01:02:37  * dapquit (Quit: Leaving.)
01:16:28  <bnoordhuis>isaacs: lgtm, assuming it works :)
01:16:44  <isaacs>bnoordhuis: all tests pass.
01:18:09  <isaacs>benchmarks/http.sh didn't change.
01:18:26  <isaacs>this is also a somewhat cleaner fix to the memory leaks that we fixed in 0.6.17.
01:19:01  <isaacs>i'd avoided doing it this way because it felt less safe to touch so many lines, but i think it's actually a bit nicer.
01:27:59  <bnoordhuis>yes, looks good - i don't object
01:28:05  <mjr_>Does that change make this bug look any different, or is it just some necessary organization work before digging deeper?
01:30:50  <isaacs>mjr_: a little of both
01:31:17  * TooTallNatequit (Ping timeout: 240 seconds)
01:31:18  <isaacs>mjr_: it *might* fix it
01:31:23  <isaacs>mjr_: since i can't reproduce it, i can't be sure.
01:31:42  <isaacs>so, here's how the flow happens, as best as I can determine:
01:31:55  <isaacs>socket.ondata calls parser.execute
01:32:21  <isaacs>parser.execute calls parserOnHeadersComplete
01:32:27  <isaacs>then that calls parser.onIncoming
01:32:50  <isaacs>in order for parser.onIncoming to NOT be a function, one of two things have to be true:
01:33:12  <isaacs>1. the parser was never assigned to a socket in either the client or server
01:33:32  <isaacs>2. the parser was improperly freeParser()'d, but the socket kept writing stuff to it
01:33:42  <isaacs>1 is impossible because the parser wouldn't be attached to a socket to parser.execute() it.
01:34:00  <isaacs>2 is impossible because we remove the socket/parser connection in freeParser (and socket.ondata)
01:34:04  <bnoordhuis>i did a lot of monkey-patching to force #2 but i couldn't get it to happen
01:35:21  <isaacs>so! the only slightly less impossible theory i have right now is that the nextTick is somehow to blame, and there is a data event sneaking through in that windo.
01:35:23  <isaacs>*window
01:35:35  * mmaleckiquit (Quit: leaving)
01:36:34  * abraxasjoined
01:36:50  <mjr_>That's the best theory I've heard so far.
01:37:15  <mjr_>I can try this tonight and we'll know in a few hours.
01:39:23  * ericktjoined
01:40:06  <isaacs>k
01:40:13  <isaacs>i do think this nextTick has to be remove.d
01:40:32  <isaacs>but ClientRequest.onSocket is called in a dozen different places, and just ripping out the nextTick breaks all kinds of stuff.
01:40:56  <isaacs>my rule of thumb is: if you can't reliably replace it wiht a setTimeout(fn, 1000), then you can't use nextTick
01:41:06  <isaacs>and if you replace this with a 1s timeout, it is very bad.
01:41:26  <isaacs>i mean, obviously, 1000ms is going to be much *slower* than nextTick
01:41:33  <isaacs>but it should still be *correct* behavior, jsut slower.
01:41:36  <isaacs>otherwise there's a problem.
01:42:04  <isaacs>mjr_: so, yes, please do try this tonight
01:42:06  <isaacs>that'd be stellar
01:44:33  * mikealquit (Quit: Leaving.)
01:54:57  * TooTallNatejoined
01:55:08  * TooTallNatequit (Client Quit)
01:55:23  <bnoordhuis>AssertionError: "7ScP3AvmU4QsLP/UQ5QNbSF6yCyDI8IGUJck6JbrDF0=" == "7ScP3AvmU4QsLP/UQ5QNbSF6yCyDI8IGUJck6JbrDDs="
01:55:23  <bnoordhuis> at Object.<anonymous> (/home/bnoordhuis/src/nodejs/master/test/simple/test-crypto.js:475:8)
01:55:37  <bnoordhuis>very sporadic, quite worrying
01:56:48  <isaacs>bnoordhuis: in master or v0.6?
01:57:13  <isaacs>bnoordhuis: i recall that the DiffieHellman test used to fail occasionally in a similar fashion.
01:57:19  <isaacs>it went away at some point, though
01:57:47  <bnoordhuis>master
01:57:56  <bnoordhuis>and it almost never happens, maybe once every 10,000 runs
02:20:18  * mattpardee_joined
02:27:55  * mikealjoined
02:35:37  * mikealquit (Quit: Leaving.)
02:37:06  * mattpardee_quit (Quit: Computer has gone to sleep.)
02:46:25  * mikealjoined
02:46:53  * hij1nxjoined
02:47:13  * mikealquit (Client Quit)
02:59:13  * mikealjoined
03:37:21  <CIA-155>node: Ben Noordhuis master * r12fc9fa / test/simple/test-dgram-ref.js :
03:37:21  <CIA-155>node: test: add failing dgram refcount test
03:37:21  <CIA-155>node: - http://git.io/Etw0yA
03:46:28  * benviejoined
03:48:10  * igorziquit (Ping timeout: 245 seconds)
03:59:58  * loladiro_quit (Quit: loladiro_)
04:14:06  * mikealquit (Quit: Leaving.)
04:40:14  * c4miloquit (Remote host closed the connection)
05:01:26  * deoxxajoined
05:01:44  <deoxxa>GENTLEMEN, BEHOLD
05:02:00  * deoxxasettles in
05:04:03  * TheJHjoined
05:04:03  * TheJHquit (Changing host)
05:04:03  * TheJHjoined
05:08:42  * TheJHquit (Ping timeout: 245 seconds)
05:14:17  * ericktquit (Quit: erickt)
05:23:47  * avalanche123quit (Read error: Operation timed out)
05:26:34  * avalanche123joined
05:30:40  * bulatshakirzyanojoined
05:32:47  * CoverSlidequit (Read error: Connection reset by peer)
05:34:57  * CoverSlidejoined
05:51:18  * paddybyersjoined
05:53:01  <mjr_>Fine, don't take the meat bridge.
05:59:36  <mjr_>isaacs: I'm rolling our your issue3236 branch right now. Hope it works.
05:59:56  <isaacs>meat... bridge?
06:00:26  <isaacs>mjr_: just to confirm, we expect the same errors to happen, but no new ones.
06:00:43  <isaacs>mjr_: if the errors go away, we can be delighted, for truly, a miracle will be before us.
06:00:56  <mjr_>isaacs: deoxxa made an Aqua Teen Hunger Force reference, and I followed up with an even more obscure one.
06:01:07  <isaacs>ohhh, i get it
06:01:15  <isaacs>it's funny, i was thinking about athf earlier today
06:01:31  <isaacs>i used to like that show when i was half-asleep, back when i had tv
06:01:33  <mjr_>How do I know so much about cartoon shit, having spent nearly no time watching anything animated for about 5 years?
06:01:53  <isaacs>i was wondering how much of my enjoyment of the show came from being in a weakened mental state whenever adult swim was on.
06:01:54  <deoxxa>because for the 10 years previous you spent nearly all your time watching?
06:02:04  <deoxxa>that'd be my reason, anyway
06:02:18  <deoxxa>isaacs: i'm pretty sure that's their whole deal
06:02:22  <mjr_>Yeah, before I had kids, I really wasted a lot of time doing fuck all.
06:02:58  <isaacs>not wasted
06:03:00  <isaacs>"used"
06:03:08  <mjr_>Consumed.
06:03:26  <deoxxa>creatively expended
06:03:45  * isaacsis rather immersed at the moment in anarchist philosophy, preparing his tacoconf and jsconfar talks
06:03:53  <isaacs>work is the waste of time!
06:04:01  <mjr_>Tell me about it
06:04:02  <deoxxa>+1
06:04:18  <deoxxa>i only got back from a holiday 2 days ago and i already want to leave again
06:04:29  <mjr_>Although, speaking of work, I did just tell 95 computers to go fetch your branch, build node, and install. All at the same time.
06:04:32  <deoxxa>feels like i'm wasting away under these fluorescent lights!
06:04:32  <mjr_>Kind of cool.
06:04:38  <isaacs>mjr_: yeah, it is!
06:04:49  <isaacs>mjr_: we need more of that. one person commanding an army of robots to go do his bidding.
06:05:06  <isaacs>mjr_: then, instead of all the time that you would have spent doing that by hand, you should spend several hours doing whatever the hell you want.
06:05:22  <isaacs>mjr_: or, since you probably wouldn't have done it by hand *yourself*, hire someone to go play for a day.
06:05:36  <mjr_>In my case, it's just slowly debugging why the army of things doesn't QUITE work the way it should always.
06:05:42  <isaacs>yes
06:05:56  <isaacs>the pattern of colored lights on the screen is ALL WRONG
06:06:00  <mjr_>And praying that the army of things doesn't get confused and destroy itself.
06:06:02  <isaacs>push more buttons!
06:14:00  * mattpardee_joined
06:25:53  * felixgejoined
06:25:53  * felixgequit (Changing host)
06:25:53  * felixgejoined
07:04:17  * txdvquit (Read error: Connection reset by peer)
07:04:42  * txdvjoined
07:18:32  * mikealjoined
07:19:31  * isaacsquit (Remote host closed the connection)
07:30:57  * mikealquit (Quit: Leaving.)
07:34:53  * orlandovftwjoined
07:36:40  * mjr_quit (Quit: mjr_)
07:38:50  * hij1nxquit (Quit: hij1nx)
07:53:29  * mikealjoined
08:17:26  * AvianFluquit (Quit: Leaving)
08:31:44  * bnoordhuisquit (Ping timeout: 272 seconds)
08:32:52  * mmaleckijoined
08:51:50  * orlandovftwquit (Ping timeout: 260 seconds)
09:02:47  * mmaleckiquit (Ping timeout: 240 seconds)
09:03:28  * mmaleckijoined
09:08:58  * rendarjoined
09:23:44  * mattpardee_quit (Quit: Computer has gone to sleep.)
09:50:42  * bnoordhuisjoined
09:55:47  * mjr_joined
10:10:32  * bnoordhuisquit (Ping timeout: 272 seconds)
10:13:53  * felixgequit (Quit: felixge)
10:19:20  * mmaleckiquit (Quit: Reconnecting)
10:19:29  * mmaleckijoined
10:23:27  * mmaleckiquit (Client Quit)
10:29:28  * mmaleckijoined
10:42:28  * abraxasquit (Remote host closed the connection)
10:43:01  * abraxasjoined
10:46:42  * abraxasquit (Read error: Operation timed out)
11:11:49  <CIA-155>node: Shigeki Ohtsu v0.6 * rcc8cfb1 / doc/api/buffer.markdown : doc: fix typo in buffer documentation - http://git.io/ZfVq3g
11:34:24  * mmaleckiquit (Quit: Reconnecting)
11:34:33  * mmalecki_joined
11:36:35  * mmalecki_quit (Client Quit)
11:36:42  * mmaleckijoined
11:42:43  * mmaleckiquit (Quit: Reconnecting)
11:43:14  * mmaleckijoined
11:43:44  * mmaleckiquit (Client Quit)
11:46:54  * mmaleckijoined
11:50:28  * mmalecki_joined
11:50:43  * mmaleckiquit (Client Quit)
11:51:06  * mmalecki_quit (Client Quit)
11:51:16  * mmaleckijoined
12:04:32  * indexzero_quit (Quit: indexzero_)
12:49:35  * bnoordhuisjoined
12:50:28  * piscisaureus_joined
12:50:41  * janjongboomjoined
12:50:49  <janjongboom>hoi @bnoordhuis
12:51:17  <janjongboom>bnoordhuis: do you have the status on node cluster shizzle work thingy
12:51:45  <janjongboom>bnoordhuis: or can you join skype?
12:51:59  * piscisaureus_quit (Client Quit)
12:52:13  <bnoordhuis>janjongboom: hey jan
12:52:41  <bnoordhuis>it's done, you can find a link to the commit in the issue
12:52:54  <janjongboom>where is the issue
12:53:16  * piscisaureus_joined
12:53:29  <piscisaureus_>hey janjongboom
12:53:55  <bnoordhuis>janjongboom: https://github.com/joyent/node/issues/3241
12:54:13  <janjongboom>is it going to land in node 0.6.x?
12:54:25  <bnoordhuis>yes, it's against 0.6 HEAD
12:54:36  <janjongboom>.6.18?
12:54:38  <janjongboom>.6.19?
12:54:40  <janjongboom>or later?
12:55:00  <bnoordhuis>probably 0.6.18
12:55:17  <janjongboom>alrighty
12:55:24  <piscisaureus_>bnoordhuis: does it need review? I saw you upgraded libuv to a nonexistent commit :-)
12:55:36  <bnoordhuis>piscisaureus_: check your logs :)
12:55:42  <bnoordhuis>but yeah, please test it
12:56:05  <mmalecki>FUCK YEAH, I found a house
12:56:40  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/libuv/commit/580deb1
12:56:56  <bnoordhuis>mmalecki: congrats :)
12:57:17  <mmalecki>bnoordhuis: thanks! :)
12:57:25  <mmalecki>I just have to learn how to cook and stuf
12:57:28  <mmalecki>*stuff
12:57:43  <bnoordhuis>no need, just marry
12:58:15  <mmalecki>or instant soups!
12:58:42  <bnoordhuis>i wouldn't have survived college if it weren't for ramen instant noodles
12:58:57  <mmalecki>I wouldn't have survived last week
12:59:29  <piscisaureus_>bnoordhuis: ha! so the next_watcher just sits there but it was completely unused?
12:59:41  <bnoordhuis>piscisaureus_: it's only used when closing the handle
12:59:58  <bnoordhuis>might as well put it to good use
13:00:24  <piscisaureus_>bnoordhuis:
13:00:26  <piscisaureus_>D:\libuv\src\unix>grep next_watcher -r -i *
13:00:28  <piscisaureus_><nothing>
13:00:47  <bnoordhuis>put the switches before the keyword :)
13:00:55  <bnoordhuis>also git grep ftw
13:01:29  <piscisaureus_>bnoordhuis: ah, ok, it's gone in master only
13:01:37  <bnoordhuis>yes
13:01:40  <piscisaureus_>bnoordhuis: the patch lgtm but there is no switch for it
13:01:54  <bnoordhuis>why would you want a switch?
13:02:31  <piscisaureus_>bnoordhuis: because probably it's preferrable to *not* use this mode when not running in a cluster setup
13:03:53  <bnoordhuis>what downside is there? i'm not that worried about running into the backlog limit
13:04:26  <piscisaureus_>bnoordhuis: because of the overhead of putting the fd in the epoll set and removing it again
13:04:31  <piscisaureus_>maybe that's moot
13:04:39  <piscisaureus_>did you try with http_simple?
13:04:41  <bnoordhuis>yes
13:07:42  <piscisaureus_>janjongboom: proxy is still failing for mer
13:07:45  <piscisaureus_>*me
13:16:37  * c4milojoined
13:20:51  * janjongboomquit (Quit: janjongboom)
13:22:46  * felixgejoined
13:22:47  * felixgequit (Changing host)
13:22:47  * felixgejoined
13:35:06  * irajoined
13:49:31  * felixgequit (Quit: http://www.debuggable.com/)
14:20:36  * loladirojoined
14:25:19  * isaacsjoined
14:28:39  * mmaleckiquit (Quit: Reconnecting)
14:28:48  * mmaleckijoined
14:31:54  * theColejoined
14:32:28  * felixgejoined
14:32:28  * felixgequit (Changing host)
14:32:28  * felixgejoined
14:33:23  <felixge>isaacs: hey
14:33:30  <felixge>how is npm install related to publishing?
14:33:39  <felixge>or did you mean 'publish' instead of install?
14:35:40  * txdvquit (Read error: Connection reset by peer)
14:35:58  <isaacs>felixge: well, i mean, you can publish --tag=preview
14:36:06  <isaacs>felixge: then your users won't get that version by default
14:36:09  <felixge>so npm publish --tag ?
14:36:10  * txdvjoined
14:36:27  <felixge>is 'preview' the magic here or can that be any string?
14:36:45  <isaacs>felixge: oh, whoops, yeah, typeo
14:36:56  <isaacs>felixge: publish --tag=blergbloop
14:37:07  <felixge>sweet
14:37:09  <felixge>that will work
14:37:09  <felixge>thanks
14:37:11  <isaacs>felixge: then have your beta user do: npm install [email protected]
14:37:18  <isaacs>felixge: bonus!
14:37:31  <isaacs>felixge: you can put "publishConfig": { "tag": "preview" } right in your package.json
14:37:31  <felixge>unfortunately I'm too drunk now to release something, friday … :)
14:37:40  <felixge>nice
14:37:42  <isaacs>felixge: npm publish --tag=drnuk
14:37:45  <felixge>haha
14:37:48  <felixge>drnuk
14:37:49  <felixge>nice
14:37:50  <felixge>:P
14:37:57  <isaacs>--tag=fitshaced
14:38:01  <felixge>was that a conscious typo?
14:38:04  <isaacs>felixge: yeah
14:38:06  <isaacs>:)
14:38:09  <felixge>: p
14:38:21  <felixge>good, it's a little early for you to be drinking :)
14:38:26  <isaacs>i'm not drunk, but i am still pre-coffee
14:38:34  <felixge>actually, that's worse
14:41:50  * theColequit (Quit: theCole)
14:51:17  <mmalecki>real programmer is never pre-coffee
14:52:29  <felixge>mmalecki: how do you consume coffee during sleep / skip sleep?
14:53:36  <mmalecki>felixge: i.v.
14:53:51  <mmalecki>don't tell me you still *drink* coffee...
14:55:51  <felixge>mmalecki: how do you deal with your body adjusting to the coffeine level
14:55:53  <felixge>?
14:57:05  <mmalecki>felixge: give it more caffeine, of course. at this point I have like 5 % of blood in my caffeine
14:58:56  <felixge>mmalecki: I wouldn't go to a programming conference if I was you
14:59:11  <felixge>you may think vampires are a joke … but you may easily become a target
14:59:25  <felixge>:)
14:59:45  <mmalecki>lol
15:03:09  * hij1nxjoined
15:04:15  * creationixis the weirdo who has never had coffee, ever
15:04:55  <mmalecki>that's not weird, just the matter of preference
15:05:02  <mmalecki>I never had tomatoes, for example
15:05:07  <mmalecki>now that's weird.
15:05:13  <creationix>I'm have enough addictions as it is. If I don't get my fix writing open source at least a few hours a week, I feel withdrawal symptoms
15:05:36  <deoxxa>mmalecki: you mean never, never?
15:05:53  <deoxxa>or like "stopped eating them at some point"
15:06:00  <mmalecki>deoxxa: yeah. like, I know I hate them, but I never actually tried
15:06:10  <deoxxa>heh
15:06:24  <deoxxa>what if somebody tricked you into eating it
15:06:31  <deoxxa>would your entire world come crumbling down?
15:06:49  <deoxxa>"but... this is who i am..."
15:06:50  <mmalecki>creationix: we should have an Anonymous Open Sourcers meetup
15:06:54  <mmalecki>deoxxa: lol
15:07:05  <creationix>mmalecki, nah, it's an addiction I don't mind
15:07:18  <mmalecki>I bet I wouldn't even know I actually ate them, but I prefer not to try XD
15:07:20  <creationix>though it does mean I have way more projects than I can possibly maintain
15:07:32  <deoxxa>"hey what is this? it's actually alright!" "that's... TOMATO" "NOOOOOOOOOOOOOOO"
15:08:07  <mmalecki>creationix: community man :)
15:08:08  <creationix>deoxxa, I accidentially had tea a couple times. At first I thought it was stale apple juice, but then the caffeine made my head spin
15:08:18  <creationix>but never coffee
15:08:58  <mmalecki>creationix: you're missing the taste of coffee. can you try decaffed one, maybe?
15:09:13  <creationix>unless you count the accidental mocha jelly bean or chocolate
15:09:15  <deoxxa>creationix: interesting!
15:09:58  <creationix>mmalecki, I have ground vanilla bean I put in smoothie, it's great and smells somewhat like coffee grinds
15:11:58  <mmalecki>creationix: I never had a vanilla bean. btw, you must be really good at making smoothies
15:12:30  <creationix>the thing I miss about living in Silicon Valley is being able to host node meetups at my house and serve smoothies
15:13:39  <creationix>I will say, if anyone makes the effort to come all the way to my house now, I'll put you up for a day or two and serve delicious smoothies :)
15:14:12  <creationix>I'd love to host a mini conf here on my land, but there are no direct flights here from anywhere except Dallas
15:15:38  <deoxxa>where are you at?
15:16:42  <creationix>Red Lick, Texas
15:17:16  <deoxxa>i hear texas is hot
15:17:26  <deoxxa>(i live in melbourne, australia)
15:18:48  <creationix>33.482634,-94.181661 to be exact
15:18:57  <deoxxa>heh
15:19:03  <creationix>that should put you in the front of my back yard
15:31:18  <mmalecki>hij1nx: hey, you down for visiting creationix and getting some delicious smoothies on our way to NY? ;)
15:32:18  <mmalecki>creationix: hah, your house is on street view
15:33:50  * loladiroquit (Quit: loladiro)
15:34:33  <bnoordhuis>piscisaureus_: turns out relaxed accept does impact performance
15:35:00  <bnoordhuis>in the single process case, thatis
15:35:14  <piscisaureus_>bnoordhuisL: yeah, so we should probably make it optional
15:36:03  <bnoordhuis>too bad but yes
15:36:26  <piscisaureus_>bnoordhuis: we already do it in windows so it's probably easy to get going
15:36:57  <hij1nx>creationix lives in rural texas, no?
15:37:05  <bnoordhuis>rural doesn't do it justice
15:37:10  <mmalecki>hij1nx: DETAILS.
15:37:25  <mmalecki>hij1nx: we can travel by foot.
15:37:26  <hij1nx>bnoordhuis: "middle of nowhere?" ;)
15:37:41  <bnoordhuis>"the land that time forgot"
15:37:50  <hij1nx>i've seen pictures though, it really is pretty.
15:37:53  <creationix>mmalecki, you do know that Texas alone is the size of Germany
15:38:40  <creationix>last fall from the back http://creationix.com/panoramas/backyard.jpg
15:38:40  <mmalecki>creationix: is this a challange?
15:39:05  <creationix>mmalecki, you're in europe right?
15:39:12  <bnoordhuis>*envy*
15:39:38  <mmalecki>creationix: well, for some time, yeah. I'll probably be in USA in 3 months
15:40:04  <creationix>well, a round-trip here from somewhere in the states is about $600 right now
15:40:18  <creationix>with atleast one layover
15:40:28  <mmalecki>that's cheap-ish
15:40:46  <creationix>driving here from san francisco or new york is something like 30 hours
15:41:04  <hij1nx>creationix: i've done it before, its exhausting =P
15:41:37  <hij1nx>there is a 9 hour stretch of west texas that is beyond desolate
15:41:42  <mmalecki>if there are no cops on the way, I could be driving ;)
15:42:45  <hij1nx>these days i fly everywhere. i've road tripped across the US 3 times.
15:43:00  <creationix>bnoordhuis, don'e have envy, I made a great one of the view from the c9 office too http://creationix.com/panoramas/waterfront.jpg
15:43:18  <creationix>Amsterdam is pretty too, just not quite as rural
15:43:34  <hij1nx>creationix: aug, 9 streets <3 i love Amsterdam
15:43:43  <mmalecki>creationix: it's a photo from c9 office in Amsterdam?
15:43:49  <mmalecki>I've been there!
15:43:56  <mmalecki>(it'll take ages until it loads)
15:44:06  <creationix>yep, I got there a couple hours early one day and took pictures while waiting for someone to unlock the door
15:44:07  <mmalecki>oh, yeah, it is, sweet :3
15:45:11  <creationix>mmalecki, umpossible! It's a node server, it's web scale
15:45:31  <mmalecki>creationix: it's a hotel wifi :)
15:45:38  <creationix>that will do it
15:45:46  <creationix>at least I scaled the images down to 1080px tall
15:45:51  <mmalecki>lol
15:45:53  <creationix>they were larger originally
15:46:13  <creationix>and it should 304 once your browser has it in cache
15:46:36  <mmalecki>yeah, it does. what server is that, http-server?
15:46:54  <creationix>creationix on npm, the static middleware
15:47:05  <creationix>I need to replace it with the new one I wrote for c9 at c9/vfs
15:47:12  <mmalecki>talk about eating your own dogfood
15:47:15  <mmalecki>!gh c9/vfs
15:47:16  <kohai>mmalecki, https://github.com/c9/vfs
15:47:54  * hij1nxquit (Quit: hij1nx)
15:47:54  * iraquit (Quit: Computer has gone to sleep.)
15:48:02  * creationixpart ("Leaving")
15:48:23  * creationixjoined
15:48:41  <mmalecki>looks neat
15:49:08  <creationix>yeah, the current code is https://github.com/creationix/creationix/blob/a9919fd1b2c9384d4f021b6058650d5d8a5f86d0/static.js
15:50:02  <mmalecki>https://github.com/creationix/creationix/blob/a9919fd1b2c9384d4f021b6058650d5d8a5f86d0/static.js#L70 wut?
15:50:25  <isaacs>review? https://gist.github.com/2660566
15:50:26  <creationix>node 0.2.x compat
15:50:45  <creationix>or was that 0.4.x
15:50:57  <creationix>it was needed to run on webos phones
15:51:01  <mmalecki>creationix: it wasn't 0.4 for sure
15:51:04  <isaacs>creationix: that'd be 0.2
15:51:05  <mmalecki>webos had 0.2 before
15:51:13  <creationix>createReadStream sucked back then
15:51:17  <creationix>yeah, I think it was 0.2
15:51:19  <isaacs>in related news, you can make your carriages go faster with these new buggy whips.
15:51:29  <creationix>took me 6 months to get 0.4.x on webos
15:51:43  <creationix>isaacs, :)
15:53:43  <mmalecki>ok, last 2 hours on this crappy wifi
15:54:00  <mmalecki>I can hopefully get something better at my new place
15:56:25  <CIA-155>node: isaacs master * ra811a4a / (lib/querystring.js test/simple/test-querystring.js): Fix #3058 querystring: Fix incorrect handling of empty keys - http://git.io/75AfYg
15:59:05  <isaacs>piscisaureus_, bnoordhuis: thoughts on this? https://github.com/joyent/node/issues/3170
15:59:21  <bnoordhuis>isaacs: it's in the works
15:59:29  <isaacs>oh, i see, you commented on the pull req
15:59:31  <isaacs>kewl
16:00:07  <isaacs>did the cares update in libuv happen yet? or still waiting on that?
16:00:40  <tjfontaine>it was going and there was a small bump on windows, I don't remember if that was resolved
16:01:11  * hij1nxjoined
16:03:33  * TheJHjoined
16:03:54  * perezdjoined
16:04:59  <bnoordhuis>isaacs: what tjfontaine said, it compile fine except it broke a test on that that other os
16:05:05  <bnoordhuis>*compiles
16:05:22  <isaacs>which other os? Looney OS? the one with the W?
16:05:32  <bnoordhuis>that one
16:05:42  <bnoordhuis>the w one
16:05:47  <deoxxa>aw, not webos :(
16:05:55  <bnoordhuis>no, webos is dead, ask tim
16:06:20  <isaacs>lol
16:06:39  <isaacs>aw, it's sad, we shouldn't joke.
16:08:12  <bnoordhuis>isaacs: https://github.com/joyent/node/issues/3081#issuecomment-5653745 <- i don't the v8 team is going to fix that by themselves
16:08:21  <bnoordhuis>i don't...
16:08:23  <bnoordhuis>i don't think
16:08:33  <bnoordhuis>seeing that they don't test on anything sunos
16:08:35  <isaacs>bnoordhuis: well... a node user can send them a patch.
16:08:42  <isaacs>bnoordhuis: or a joyent engineer can do it.
16:08:52  <isaacs>bnoordhuis: i'm not opposed to floating a patch.
16:08:56  <bnoordhuis>good point
16:09:04  <isaacs>i AM opposed to caring overly much about something that is trivial to work around.
16:09:30  <isaacs>it's not worth wasting your time on, imo. if it'd be fun for you to fix, then don't let me stop you.
16:12:56  * ericktjoined
16:18:09  * orlandovftwjoined
16:19:20  * orlandovftwquit (Client Quit)
16:19:33  * hij1nxquit (Read error: Connection reset by peer)
16:19:39  * orlandovftwjoined
16:23:10  <creationix>I'd say webOS is about as dead/alive as beOS (http://haiku-os.org/) open webos is slowing trying to survive, but has no real hardware to run on.
16:23:43  <creationix>which is a shame, webOS was going to ship node to millions of consumers
16:24:10  * hij1nxjoined
16:24:13  <creationix>at least the beaglebone comes with node of sorts
16:24:21  <creationix>and even an old version of c9
16:26:00  <creationix>there were ~1M touchpads sold in the firesale all with node 0.4.12
16:27:43  * dapjoined
16:31:32  <mmalecki>creationix: haiku is actually pretty sweet
16:31:38  <mmalecki>I like the architecture
16:32:06  <mmalecki>sure, it'd need a revamp (it's a reasonably old project), but I'd totally use it if it was more mature
16:33:27  <creationix>I own a copy. I bought it along with gobe production suite
16:33:32  <creationix>still have the CDs
16:34:07  <mmalecki>creationix: gobe production suite?
16:35:06  <creationix>http://en.wikipedia.org/wiki/Gobe_Software
16:35:28  <creationix>ClarisWorks ported to BeOS
16:35:34  <mmalecki>ah, hm. I never used actual BeOS
16:38:59  * c4milochanged nick to c4milo|lunch
16:41:15  * hij1nxquit (Quit: hij1nx)
16:53:02  * pieternjoined
16:54:26  * c4milo|lunchquit (Remote host closed the connection)
16:59:33  <bnoordhuis>piscisaureus_: that relaxed accept thing is curious... http_simple is ~15% slower but tcp4_pound_100 doesn't seem affected
17:00:31  * hij1nxjoined
17:01:14  * mattpardee_joined
17:02:30  <CIA-155>node: Ben Noordhuis v0.6 * r884499d / wscript : build: fix cross-compiling - http://git.io/KwQ_Lg
17:03:47  * mikealquit (Quit: Leaving.)
17:04:25  * mikealjoined
17:07:02  * mattpardee_part ("Textual IRC Client: http://www.textualapp.com/")
17:17:00  * orlandovftwquit (Ping timeout: 260 seconds)
17:19:17  * mikealquit (Quit: Leaving.)
17:19:24  * AvianFlujoined
17:19:42  * brsonjoined
17:21:41  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/node/compare/v0.6...accept-backoff <- test?
17:22:01  <piscisaureus_>bnoordhuis: test where/what?
17:22:31  <bnoordhuis>piscisaureus_: where=probably on smartos system with many cores, what=the above patch :)
17:22:43  <bnoordhuis>export UV_TCP_SINGLE_ACCEPT=1 enables relaxed accept
17:22:56  <bnoordhuis>i can attest to the fact that it works
17:23:09  <bnoordhuis>but on my puny dual core laptop the results are less than impressive
17:23:11  <piscisaureus_>bnoordhuis: ok. I can do that but everybody is leaving here.
17:23:19  <bnoordhuis>tell them to stay
17:23:23  <piscisaureus_>bnoordhuis: so I will forward that to the people who need it :-)
17:23:41  <bnoordhuis>okay, good
17:23:59  <bnoordhuis>piscisaureus_: i've observed that it balances the load more evenly
17:24:00  <piscisaureus_>bnoordhuis: it'll be Zef and Rik
17:24:06  <piscisaureus_>bnoordhuis: cool
17:24:13  <bnoordhuis>the downside is that it's slower
17:24:27  <piscisaureus_>bnoordhuis: it'll probably be tested with the test proxy and then in production
17:25:12  <bnoordhuis>i suspect you need a fair number of cores for the results to show
17:25:14  <piscisaureus_>bnoordhuis: we have the same issue on windows
17:25:18  <piscisaureus_>bnoordhuis: yeah we have 32
17:25:44  <bnoordhuis>that should be enough :)
17:25:56  <piscisaureus_>bnoordhuis: ok, rik knows
17:26:14  <bnoordhuis>say hi to rik from me
17:26:19  <piscisaureus_>bnoordhuis: ehm so I think in node we could automagically enable that but only for cluster servers
17:26:25  <piscisaureus_>bnoordhuis: like we do on windows
17:26:37  <bnoordhuis>probably. it's trivial to add
17:26:39  <piscisaureus_>bnoordhuis: but let's first figure out if it works
17:26:46  <piscisaureus_>bnoordhuis: thanks
17:26:54  <bnoordhuis>no problem
17:27:02  <bnoordhuis>how's the ref refactor coming along?
17:27:11  <piscisaureus_>bnoordhuis: I was grabbed to do other stuff. That: https://github.com/piscisaureus/murmur3
17:27:14  <piscisaureus_>bnoordhuis: :-(
17:27:28  <bnoordhuis>hah
17:27:31  <bnoordhuis>looks like fun
17:27:48  <piscisaureus_>bnoordhuis: also, I figured out some other issues with my approach for the refcount refactor
17:27:55  <bnoordhuis>like what?
17:28:06  <piscisaureus_>bnoordhuis: we have "reqs" for reading and writing, but they can be pending even after uv_read_stop was called...
17:28:24  <piscisaureus_>bnoordhuis: so `handle->pending_reqs > 0` is not a good test for activeness...
17:28:45  <bnoordhuis>how does that work? for the read reqs, i mean
17:29:03  <bnoordhuis>naively i would assume that not reading means you're not reading
17:29:05  <piscisaureus_>bnoordhuis: a read req is just a read that's happening in the background
17:29:24  <piscisaureus_>bnoordhuis: so as soon as you start reading we start such a read and when it comes back we post another one
17:29:35  <piscisaureus_>bnoordhuis: but if you call read_stop there is still a read pending
17:29:42  <bnoordhuis>ah, right
17:29:56  <piscisaureus_>bnoordhuis: but when it comes back while read is stopped it is just ignored or buffered up (depending on whether it was a 0-read or not)
17:30:11  <piscisaureus_>bnoordhuis: so... more work, that's all
17:30:17  <bnoordhuis>duly noted :)
17:36:05  <isaacs>mjr_: Any results from that patch?
17:40:11  * mmaleckiquit (Quit: leaving)
17:54:42  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
17:57:54  * theColejoined
18:04:26  <txdv>"According to the developers of Vert.x, vert.x is faster and more officient compared to node.js "
18:04:34  <txdv>efficient
18:05:03  <indutny>that's great
18:05:32  <indutny>JVM is really competitive in web-development platforms
18:05:32  * AndreasMadsenjoined
18:05:36  <indutny>hahaha, just kidding
18:05:37  <indutny>:D
18:06:20  * isaacsquit (Remote host closed the connection)
18:07:01  * AndreasMadsenquit (Remote host closed the connection)
18:07:38  <txdv>it has one big downside
18:07:42  <txdv>multiple languages
18:09:56  <indutny>it has another one downside
18:10:06  <indutny>author that do not know how to benchmark things
18:10:18  <indutny>and neither he wants to listen others opinion
18:10:28  <indutny>especially if it states that he was wrong
18:10:39  <indutny>see discussion here: https://gist.github.com/2652991
18:15:41  <AvianFlu>lol
18:16:47  * theColequit (Quit: theCole)
18:19:55  * orlandovftwjoined
18:29:22  * TooTallNatejoined
18:29:34  * xaqjoined
18:31:40  * igorzijoined
18:31:55  * mattstevensjoined
18:33:08  * c4milo|lunchjoined
18:34:01  * c4milo|lunchchanged nick to c4milo
18:34:29  * indexzerojoined
18:38:28  * AndreasMadsenjoined
18:39:12  * mikealjoined
18:39:58  * indexzeroquit (Ping timeout: 250 seconds)
18:42:12  * AndreasMadsenquit (Remote host closed the connection)
18:55:34  * mikealquit (Quit: Leaving.)
18:57:35  * mikealjoined
18:59:50  * isaacsjoined
19:11:03  * mikealquit (Quit: Leaving.)
19:57:29  * piscisaureus_joined
20:01:55  * mikealjoined
20:11:05  * avsejquit (Quit: Quit)
20:11:29  * avsejjoined
20:17:36  * avsejquit (Quit: Quit)
20:18:30  * avsejjoined
20:19:45  * mikealquit (Quit: Leaving.)
20:44:22  * pieternquit (Quit: pietern)
20:45:50  * dapquit (Ping timeout: 260 seconds)
20:46:38  * isaacsquit (Ping timeout: 260 seconds)
20:46:49  * pieternjoined
20:47:38  * paddybyersquit (Quit: paddybyers)
20:49:25  * paddybyersjoined
20:53:16  * mikealjoined
20:55:40  * paddybyersquit (Quit: paddybyers)
20:58:32  * paddybyersjoined
21:09:25  * isaacsjoined
21:10:21  * dapjoined
21:14:46  <creationix>isaacs, what do you think about https://gist.github.com/4b71912f266133c69506
21:19:44  * c4miloquit (Remote host closed the connection)
21:25:25  <bnoordhuis>piscisaureus_: ping
21:25:32  <piscisaureus_>bnoordhuis: hey
21:25:57  <bnoordhuis>piscisaureus_: ho
21:26:04  <piscisaureus_>bnoordhuis: pung
21:26:15  <bnoordhuis>isn't murmur supposed to always generate a 16 byte hex hash?
21:26:40  <piscisaureus_>bnoordhuis: bugje. just fixed it
21:26:45  <bnoordhuis>ah okay :)
21:26:56  <piscisaureus_>bnoordhuis: https://github.com/c9/murmur3/commit/bd12da1e6faa6ac54887989a5264c1b24c647ec3
21:27:06  <mjr_>isaacs: ready to have your fucking mind blown?
21:27:18  <bnoordhuis>piscisaureus_: like 5 seconds after i cloned the repo
21:27:28  <mjr_>That's an unrelated question. But hey, look at this:
21:27:29  <mjr_>https://skitch.com/mranney/83i8i/search-search-splunk-4.3beta
21:27:42  <piscisaureus_>bnoordhuis: bored? Maybe you can fix this: https://github.com/c9/murmur3/issues/2
21:28:08  <piscisaureus_>mjr_: to me, that looks good
21:28:18  <piscisaureus_>mjr_: unless "more green is better" :-)
21:28:23  <bnoordhuis>piscisaureus_: i don't quite understand what cristoph means
21:28:30  <piscisaureus_>me neither
21:28:37  <mjr_>Switching to the issue3236 branch eliminated a TON of uncaught exceptions from that parser thing.
21:28:43  <bnoordhuis>oh, maybe i do
21:28:54  <mjr_>Actual progress
21:28:57  <piscisaureus_>bnoordhuis: And I am definitely *not* going to relativize paths
21:29:14  <piscisaureus_>bnoordhuis: it would be nice to be able to pass multiple filenames too. it may be that that is what he wants.
21:29:27  <piscisaureus_>bnoordhuis: so you can do `murmur3 foo.js bar.js`
21:31:03  <bnoordhuis>piscisaureus_: btw, that makefile looks familiar...
21:31:12  <piscisaureus_>bnoordhuis: yup
21:31:21  <piscisaureus_>bnoordhuis: just has to do some search and replace
21:31:31  <piscisaureus_>add -fno-strict-aliasing and katching
21:56:55  <bnoordhuis>piscisaureus_: assert(ctx->unparsed_len < MURMUR3_BLOCK_SIZE) <- why that assertion?
21:57:01  <isaacs>mjr_: ok... what am i looking at with tha?
21:57:20  <piscisaureus_>bnoordhuis: just an internal sanity check
21:57:20  <mjr_>Sorry, I should have annotated it
21:57:41  <isaacs>mjr_: you said it's pure win.
21:57:44  <isaacs>mjr_: so... that's rad.
21:57:45  <isaacs>:)
21:57:52  <isaacs>mjr_: but i still see some onIncoming is not a function
21:57:53  <mjr_>isaacs: green is bad. You can see how in the middle of the night when I upgraded node to your branch that it fixed all of the parser exceptions on this process.
21:58:02  <isaacs>mjr_: the green is bad? that's awesome!
21:58:02  <piscisaureus_>bnoordhuis: ctx->unparsed_bytes is used to store bytes that haven't been processed yet, but it only has space for one block
21:58:03  <bnoordhuis>piscisaureus_: shouldn't that be <= ?
21:58:07  <isaacs>cuz the green went way way down!
21:58:11  * isaacscheers!
21:58:17  <bnoordhuis>piscisaureus_: i.e. up to 16 bytes, instead of less than 16 bytes
21:58:17  <mjr_>I know, less green is amazing.
21:58:31  <piscisaureus_>bnoordhuis: well if there are 16 bytes in there they should've been processed already
21:58:39  <isaacs>ohh, i see, the onIncoming error is from the 10th, not the 11th
21:58:45  <mjr_>isaacs: so your theory on the nextTick turned out to be correct.
21:58:56  <isaacs>mjr_: well... no.
21:58:57  <bnoordhuis>piscisaureus_: i ask because i'm triggering that assertion :)
21:59:03  <isaacs>mjr_: it turned out to be overly cautious
21:59:12  <isaacs>mjr_: my theory is that the nextTick caused it -- i didn't take the nextTick out.
21:59:19  <piscisaureus_>bnoordhuis: huh well then something is wrong :-)
21:59:32  <isaacs>mjr_: i *did* move everything out of the closure, though, so it's not trapping references to anything in the handlers defined in the nextTick cb
21:59:34  <mjr_>isaacs: so what different functionality is in this branch?
21:59:37  <bnoordhuis>piscisaureus_: wait, let me push my test case
22:00:11  <isaacs>mjr_: instead of this: nextTick(function() { foo.bar = function () { ... } }); we have this: function fooBar() {}; nextTick(function() { foo.bar = fooBar })
22:00:19  <isaacs>mjr_: ie, the function definitions are pulled out
22:00:35  <isaacs>mjr_: my hope was that this would make the code more readable (readable at all, really)
22:00:59  <mjr_>isaacs: yeah, I'm moving all of our code to that shape, building up a lot of different prototypes with related functionality.
22:01:12  <bnoordhuis>piscisaureus_: https://github.com/c9/murmur3/commit/6e8d1ac
22:01:22  <mjr_>Having a named constructor function makes all of the heap analysis and flame graphery work better as well.
22:01:40  <isaacs>mjr_: ok. well, i'm convinced it's good. i'll land it. bnoordhuis already reviewed and lgtm'd it.
22:02:02  <mjr_>Somehow this also seems to have cut down on the EBADFs.
22:02:33  <isaacs>mjr_: coincidence, i think.
22:02:40  <isaacs>mjr_: unless the crashes were somehow helping to induce them
22:02:48  <isaacs>(which is possible, i suppose)
22:03:00  <mjr_>Actually no, zooming out a bit, the EBADF rate is unchanged. Just a local minimum.
22:07:31  <piscisaureus_>bnoordhuis: end = start + (len & ~(MURMUR3_BLOCK_SIZE - 1));
22:07:39  <piscisaureus_>bnoordhuis: it could be that it is stopping too early
22:07:56  <piscisaureus_>bnoordhuis: I think it should be "for ( ; pos < end; pos += MURMUR3_BLOCK_SIZE) {"
22:07:58  <piscisaureus_>er
22:08:04  <piscisaureus_>"for ( ; pos <= end; pos += MURMUR3_BLOCK_SIZE) {"
22:09:48  <bnoordhuis>piscisaureus_: end = start + (len & ~(MURMUR3_BLOCK_SIZE - 1)); <- isn't that wrong? what if e.g. start==0 and len==1
22:10:12  <bnoordhuis>no wait, it's okay
22:10:17  <piscisaureus_>bnoordhuis: well then it would never hit that code path :_)
22:10:39  <piscisaureus_>bnoordhuis: but I think it may sometimes not process all the blocks
22:11:49  <piscisaureus_>bnoordhuis: yeah, I am hitting the assertion too
22:12:04  <piscisaureus_>bnoordhuis: if you have a nicer way to do the comutation, be my guest :-)
22:12:09  <bnoordhuis>heh okay
22:13:35  <piscisaureus_>bnoordhuis: hmm it looks clobbered. unparsed_bytes is unparsed_bytes = 0x0024fcbc
22:17:42  <piscisaureus_>ah
22:19:29  <piscisaureus_>bnoordhuis: fix pushed
22:22:09  <piscisaureus_>but I don't know how to interpret the benchmark outputs
22:22:11  <piscisaureus_>since there are none
22:23:41  * mmaleckijoined
22:26:21  <CIA-155>node: isaacs v0.6 * r07d8a46 / lib/http.js : Break up huge function in ClientRequest.onSocket - http://git.io/BQOxoQ
22:26:22  <CIA-155>node: isaacs v0.6 * r9239088 / test/gc/test-http-client-timeout.js : 500 is a magic number for the GC for some reason - http://git.io/VJjinQ
22:27:32  <mmalecki>well, I'm not homeless anymore!
22:30:16  <AvianFlu>mmalecki, ++
22:30:17  <kohai>mmalecki has 90 beers
22:32:55  * mikealquit (Quit: Leaving.)
22:33:07  <bnoordhuis>piscisaureus_: time ./bench
22:33:17  * rendarquit
22:33:45  <piscisaureus_>bnoordhuis: I have nothing to compare to... It would be nice to see whether chunk sizes make a difference etc
22:34:11  * mikealjoined
22:37:23  * mmaleckiquit (Quit: leaving)
22:37:27  <isaacs>piscisaureus_, bnoordhuis: what is murmur3?
22:37:41  <bnoordhuis>isaacs: a digest
22:38:10  <isaacs>ic
22:39:04  * paddybyersquit (Quit: paddybyers)
22:48:51  <isaacs>bnoordhuis: do you have any thoughts on https://github.com/joyent/node/pull/2772/files?
22:49:48  <bnoordhuis>my first thought was "damn, big PR"
22:50:16  * mikealquit (Quit: Leaving.)
23:04:15  <bnoordhuis>isaacs: reviewed the docs, i'm saving the rest for later
23:04:20  <bnoordhuis>i kind of like the concept though
23:04:31  <isaacs>yeah, it's a lot of green
23:04:43  * avalanche123quit (Ping timeout: 244 seconds)
23:04:44  * bulatshakirzyanochanged nick to avalanche123
23:06:56  * avalanch_joined
23:14:08  * loladirojoined
23:29:52  * mattstevensquit (Quit: mattstevens)