00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:08  * ircretaryjoined
00:06:32  <MI6>nodejs-v0.10-windows: #31 UNSTABLE windows-ia32 (11/584) windows-x64 (11/584) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/31/
00:10:23  <MI6>nodejs-master-windows: #37 UNSTABLE windows-x64 (16/600) windows-ia32 (16/600) http://jenkins.nodejs.org/job/nodejs-master-windows/37/
00:15:22  * dapquit (Quit: Leaving.)
00:17:57  * mikealjoined
00:24:15  <MI6>nodejs-v0.10: #203 UNSTABLE linux-x64 (1/585) smartos-x64 (1/585) http://jenkins.nodejs.org/job/nodejs-v0.10/203/
00:27:46  * hz_joined
00:27:46  * hz_quit (Changing host)
00:27:46  * hz_joined
00:27:46  * hzquit (Disconnected by services)
00:27:54  * hz_changed nick to hz
00:30:59  * mikealquit (Quit: Leaving.)
00:32:32  <MI6>nodejs-v0.10-windows: #32 UNSTABLE windows-ia32 (10/585) windows-x64 (9/585) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/32/
00:46:54  * hzquit (Ping timeout: 264 seconds)
00:47:10  * timoxleyquit (Quit: Computer has gone to sleep.)
00:54:06  * qardquit (Quit: Leaving.)
01:05:18  * inolenquit (Quit: Leaving.)
01:16:15  * timoxleyjoined
01:22:03  * amartensquit (Quit: Leaving.)
01:23:19  * wmilquit (Remote host closed the connection)
01:26:13  * sblomquit (Ping timeout: 276 seconds)
01:27:40  * abraxasjoined
01:33:35  * brsonquit (Read error: Operation timed out)
01:38:42  <MI6>joyent/node: Nathan Rajlich v0.10 * 99b737b : configure: respect the --dest-os flag consistently - http://git.io/pUIWEQ
01:39:29  * kazuponjoined
01:46:59  * inolenjoined
01:48:56  <MI6>nodejs-v0.10: #204 UNSTABLE osx-ia32 (1/585) smartos-x64 (1/585) http://jenkins.nodejs.org/job/nodejs-v0.10/204/
01:53:14  * kazuponquit (Remote host closed the connection)
01:53:16  * dapjoined
02:13:12  <MI6>nodejs-v0.10-windows: #33 UNSTABLE windows-ia32 (6/585) windows-x64 (9/585) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/33/
02:24:56  * timoxleyquit (Quit: Computer has gone to sleep.)
02:30:38  * wmiljoined
02:46:31  * kazuponjoined
02:52:18  * icarotjoined
02:55:57  * icarotquit (Remote host closed the connection)
02:59:04  * kazuponquit (Remote host closed the connection)
03:10:05  * mikealjoined
03:19:15  * mikealquit (Quit: Leaving.)
03:24:28  * wmilquit (Remote host closed the connection)
03:34:09  * wmiljoined
03:42:34  * piscisaureus_quit (Ping timeout: 245 seconds)
03:45:34  * wmilquit (Remote host closed the connection)
04:05:24  * kazuponjoined
04:09:16  * TooTallNatequit (Quit: ["Textual IRC Client: www.textualapp.com"])
04:11:24  * timoxleyjoined
04:15:31  * normanmjoined
04:23:19  * sblomjoined
04:27:56  * sblomquit (Ping timeout: 276 seconds)
04:28:42  * kevinswiberquit (Remote host closed the connection)
04:29:16  * kevinswiberjoined
04:29:59  * timoxleyquit (Quit: Computer has gone to sleep.)
04:31:16  * timoxleyjoined
04:33:26  * kevinswiberquit (Ping timeout: 240 seconds)
04:37:01  * defunctzombie_zzchanged nick to defunctzombie
04:45:50  * brsonjoined
04:49:11  * wmiljoined
05:04:21  * bajtosjoined
05:06:12  <MI6>nodejs-master: #228 UNSTABLE smartos-x64 (5/600) osx-x64 (3/600) linux-x64 (3/600) linux-ia32 (3/600) smartos-ia32 (3/600) osx-ia32 (3/600) http://jenkins.nodejs.org/job/nodejs-master/228/
05:07:53  * wmilquit (Remote host closed the connection)
05:33:20  * timoxleyquit (Quit: Textual IRC Client: www.textualapp.com)
05:41:59  * kazuponquit (Remote host closed the connection)
05:42:59  * kazuponjoined
05:52:18  * defunctzombiechanged nick to defunctzombie_zz
05:53:36  <isaacs>trevnorris: that's the worst
06:31:14  * rendarjoined
06:34:16  * paddybyersjoined
06:37:20  * dshaw_quit (Quit: Leaving.)
06:39:59  * `3rdEdenjoined
07:03:39  * paddybyersquit (Ping timeout: 260 seconds)
07:05:35  * bajtosquit (Quit: bajtos)
07:19:08  * bajtosjoined
07:31:15  * csaohjoined
07:40:44  * paddybyersjoined
07:55:56  * brsonquit (Quit: leaving)
08:00:49  * stagasjoined
08:01:05  * dominictarrquit (Quit: dominictarr)
08:07:56  * dapquit (Quit: Leaving.)
09:31:19  * csaohquit (Ping timeout: 252 seconds)
09:36:12  * csaohjoined
10:01:47  * kazuponquit (Read error: Connection reset by peer)
10:01:49  * kazupon_joined
10:12:52  * bajtosquit (Quit: bajtos)
10:13:39  * dominictarrjoined
10:20:10  * dominictarrquit (Ping timeout: 256 seconds)
10:20:32  * kazupon_quit (Remote host closed the connection)
10:22:24  * stagasquit (Read error: Connection reset by peer)
10:22:35  * abraxasquit (Remote host closed the connection)
10:39:44  * benoitcquit (Excess Flood)
10:42:13  * benoitcjoined
10:56:55  * bnoordhuisjoined
11:00:29  * bajtosjoined
11:07:51  * hzjoined
11:16:50  * kazuponjoined
11:48:40  <MI6>joyent/node: Ben Noordhuis v0.10 * 89dcf22 : doc: the build requires gcc 4.2 or newer - http://git.io/EeY48Q
11:57:09  * bnoordhuisquit (Ping timeout: 245 seconds)
11:58:43  <MI6>nodejs-v0.10: #205 UNSTABLE linux-x64 (1/585) smartos-x64 (1/585) http://jenkins.nodejs.org/job/nodejs-v0.10/205/
12:06:39  <MI6>nodejs-v0.10-windows: #34 UNSTABLE windows-ia32 (8/585) windows-x64 (10/585) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/34/
12:22:53  * abraxasjoined
12:27:55  * abraxasquit (Ping timeout: 276 seconds)
12:35:24  * kazuponquit (Remote host closed the connection)
12:39:18  * piscisaureus_joined
12:42:41  * benoitcquit (Excess Flood)
12:45:45  * kevinswiberjoined
12:46:43  * benoitcjoined
12:47:08  * bajtosquit (Ping timeout: 252 seconds)
13:19:16  * icarotjoined
13:20:01  * icarotquit (Remote host closed the connection)
13:31:30  * pachetjoined
13:32:21  * bnoordhuisjoined
13:33:58  * hzquit (Ping timeout: 256 seconds)
13:35:55  * benoitcquit (Excess Flood)
13:41:45  * benoitcjoined
13:45:51  * kazuponjoined
13:46:28  * kevinswiberquit (Remote host closed the connection)
13:47:04  * kevinswiberjoined
13:51:01  * kazuponquit (Ping timeout: 246 seconds)
13:51:40  * kevinswiberquit (Ping timeout: 256 seconds)
13:58:07  * kazuponjoined
14:37:32  * hueniversequit (Quit: Leaving.)
14:37:58  * hueniversejoined
14:41:37  * mmaleckichanged nick to mmalecki[out]
14:42:38  * c4milojoined
14:50:43  * kazuponquit (Remote host closed the connection)
15:04:32  <MI6>joyent/libuv: Ben Noordhuis v0.10 * c53fe81 : unix: implicitly signal write errors to libuv user - http://git.io/pFlN2w
15:06:48  <MI6>libuv-v0.10: #69 UNSTABLE smartos (2/186) windows (3/187) http://jenkins.nodejs.org/job/libuv-v0.10/69/
15:11:32  <MI6>libuv-v0.10-gyp: #34 UNSTABLE smartos-x64 (2/186) windows-x64 (3/187) smartos-ia32 (2/186) linux-ia32 (1/186) windows-ia32 (4/187) http://jenkins.nodejs.org/job/libuv-v0.10-gyp/34/
15:12:11  <isaacs>good morning
15:13:12  <isaacs>tjfontaine: back when github used to go down way more often, i remember we had another copy of the node repo on nodejs.org
15:17:52  * mikealjoined
15:20:23  <MI6>libuv-node-integration: #64 UNSTABLE linux-ia32 (1/585) smartos-ia32 (1/585) osx-x64 (2/585) osx-ia32 (2/585) linux-x64 (1/585) smartos-x64 (1/585) http://jenkins.nodejs.org/job/libuv-node-integration/64/
15:21:23  <saghul>oh, write_queue_size is a public member, I need to expose that in pyuv
15:22:43  <piscisaureus_>yes it is very important even
15:23:00  <piscisaureus_>to implement backpressure properly
15:25:09  <saghul>yep
15:27:58  * `3rdEdenquit (Remote host closed the connection)
15:28:01  * AvianFlujoined
15:32:18  * mikealquit (Quit: Leaving.)
15:41:13  * saghul_joined
15:44:47  * bnoordhuisquit (Ping timeout: 252 seconds)
15:44:49  * saghul_quit (Client Quit)
15:46:48  * saghul_joined
15:47:16  * defunctzombie_zzchanged nick to defunctzombie
15:48:17  * saghul_quit (Client Quit)
15:49:41  * AvianFluquit (Ping timeout: 256 seconds)
15:50:00  * saghul_joined
15:50:14  * saghul_quit (Client Quit)
15:51:30  * Linusjoined
15:51:43  * dapjoined
15:52:32  <Linus>Hello people!
15:53:03  <Linus>Isaacs said this might be a good place to turn to if I'm looking for feedback on a couple of pull requests, might that be correct? :)
15:54:55  * kazuponjoined
15:56:59  <tjfontaine>this is the place, and patience is definitely a virtue :)
15:58:05  <Linus>Fair enough. :)
15:59:24  <Linus>It's a quintuplet of pull requests, one set with three pull requests that ensures Node and it
15:59:35  <Linus>it's dependencies* will build on android.
16:00:22  <Linus>The second set is basically an API whitelisting system
16:01:30  <Linus>The Android support PR can be found here: https://github.com/joyent/node/pull/5514
16:01:58  <Linus>And the whitelisting patch is here: https://github.com/joyent/node/pull/5531
16:05:26  <Linus>I think I managed to reduce the impact of both patch sets as much as possible, but would appreciate any feedback on usefulness and/or `improvability'. :)
16:06:13  <piscisaureus_>Linus: I'm fine with android support.
16:06:23  <Linus>...and, of course, if there's any errors with regards to your coding standards, since I have very little experience with your codebase. :)
16:06:34  <piscisaureus_>Linus: but nobody got around to reviewing your patch yet (and we need to land in libuv first)
16:06:46  <piscisaureus_>Linus: however I don't know about the whitelisting support.
16:06:47  * TooTallNatejoined
16:06:57  <isaacs>Linus: you're Asa Erikson Hedstram?
16:07:09  <Linus>Haha, no
16:07:52  <Linus>Or, wait, waht?
16:07:55  <Linus>What*?
16:08:06  <Linus>Could that perhaps be the name they registered in the CLA?
16:08:18  <piscisaureus_>Linus: it looks like whitelisting it would make node much harder to maintains, and possibly it'll affect performance for everyone. Also we're suddenly making security guarantees that were never there.
16:09:22  <Linus>I considered that in the patch, and the whitelisting was written to avoid any impact on the APIs
16:10:09  <Linus>Additionally, it will only run once, before an exported object is exposed to Javascript
16:10:47  <Linus>isaacs: I checked, that's our legal counsel, probably the one who effected the signing
16:11:15  <isaacs>Linus: great. so they signed for thesmelves, not for you.
16:11:36  <isaacs>Linus: can you just go click the buttons, and mention in the employer info section that you emailed a scan signed by your legal people?
16:11:54  <Linus>The intention was probably to sign for everyone... And yes, I'm tired of this, so I'll do that. :q
16:12:00  <isaacs>kewl
16:12:10  <isaacs>Linus: thanks, sorry for the hassle.
16:12:14  * isaacsoff to class for a few hours
16:12:16  * isaacs&
16:12:16  <LOUDBOT>HE IS NOT LOUD ENOUGH
16:14:52  <Linus>piscisaureus_: Can you explain how you see it increasing maintenance? If there's anything I can do to resolve that I'd be happy to do so.
16:17:02  <isaacs>fwiw, i don't mind if node builds on android, obviously, but i am not too keen on adding another build target that we keep up with. Can be treated like OpenIndiana or FreeBSD: Send us a patch, we'll take it, but we won't test/guarantee on that platform.
16:18:29  <Linus>Would it be possible to set up equivalent jenkins builds for it to ensure the build works?
16:18:46  * stagasjoined
16:18:48  <tjfontaine>Linus: what would the slave look like I guess
16:19:16  <Linus>i.e. such that you can see if the build breaks, even if no tests are run?
16:20:42  <Linus>No experience with Jenkins myself, but it uses the standard gyp build setup with a replacement configuration file and added dependency on the android NDK
16:21:10  <tjfontaine>ya, so my question was mostly do I need an arm device that runs android for this, or just compiling against the ndk
16:21:23  <Linus>NDK, of course. :)
16:21:59  <tjfontaine>ok so what host platform is interesting, linux?
16:22:20  <Linus>In my eyes, yes, Linux
16:22:53  <Linus>I'm not even sure android supports the NDK on windows...
16:23:26  <tjfontaine>I have other ancillary builds on my todo list, such as libuv-mingw, freebsd, rasberry pi
16:23:39  <tjfontaine>things that are interesting to know the state of, but don't factor into decision making
16:24:13  <Linus>Alright
16:24:37  <tjfontaine>so if/when stuff lands I am not opposed to adding it in my spare time
16:25:04  <Linus>Wonderful. :)
16:25:43  <tjfontaine>send me a reference device and I'll even have it run tests :P
16:26:10  <Linus>That might be possible to arrange
16:26:31  <Linus>If you're willing to commit to that. ;)
16:26:33  * mikealjoined
16:26:40  <tjfontaine>define "commit" :)
16:26:48  <Linus>Actually do it. :p
16:27:14  <tjfontaine>device in hand I would figure out how to do it, and if failed I would return it
16:27:49  <Linus>I can check if it's possible and get back to you on that. :)
16:27:54  <tjfontaine>heh ok
16:29:50  <piscisaureus_>Linus: I need to discuss that with bnoordhuis and isaacs
16:30:37  <Linus>Of course
16:31:03  * mikealquit (Client Quit)
16:32:11  <Linus>I should say, I've done what I can to avoid maintenance altogether on the patch, and as for performance impact, since it's only run during module initialization, it will be minimal.
16:33:32  <Linus>I've been in some discussion over a similar solution with bnoordhuis, and this is what I thought up as a result of those discussions
16:34:05  <piscisaureus_>I see
16:34:25  <piscisaureus_>So I'll hear what bnoordhuis has to say :)
16:35:01  <Linus>His initial stance will of course be that this should be done with seccomp2, that I unfortunately have no access to. :)
16:36:32  * TooTallNatequit (Read error: Connection reset by peer)
16:36:47  <Linus>I believe his standpoint on the matter was that "If you can design something that will not affect the existing system, nor performance, then /maybe/".
16:37:04  <Linus>But, I'll keep in touch, and see how it goes. :)
16:38:40  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
16:41:21  * TooTallNatejoined
16:41:58  <Linus>Anyhow, 18:40, I was supposed to quit an hour ago. :)
16:42:48  <Linus>As I said, I'll try to keep in touch
16:42:55  <Linus>G'bye for now. :)
16:42:59  <tjfontaine>enjoy
16:43:09  * Linusquit (Quit: Page closed)
16:43:23  <TooTallNate>that guy...
16:43:34  <TooTallNate>nice punctuation
16:43:51  <TooTallNate>not hating, im impressed :p
16:44:00  <TooTallNate>could be an iOS device though…
16:45:12  <tjfontaine>you'd rather him be working on an ios device?
16:45:26  * TooTallNatequit (Client Quit)
16:48:28  * inolenquit (Quit: Leaving.)
16:55:45  * csaohquit (Quit: csaoh)
16:58:01  * amartensjoined
16:58:01  * kevinswiberjoined
16:59:00  * TooTallNatejoined
17:11:20  <tjfontaine>hmm wonder what bert is doing
17:12:45  * qardjoined
17:14:02  <mbroadst>does anyone have v8-profiler working with 0.10.4? I'm having trouble building it on ubuntu 13.04
17:14:05  <trevnorris>morning
17:14:47  <qard>I have a port of it here, removing the few things that don't work anymore; https://github.com/NodeFly/nodefly-v8-profiler
17:16:13  <mbroadst>qard: cool, can I use that as a drop-in replacement for the intructions for using node-inspector by any chance?
17:16:39  * st_lukejoined
17:17:11  * dapquit (Quit: Leaving.)
17:17:43  <qard>I think so.
17:17:56  <qard>You'd need to replace the package name though.
17:18:26  * kevinswiberquit (Remote host closed the connection)
17:18:58  * kevinswiberjoined
17:23:31  * kevinswiberquit (Ping timeout: 252 seconds)
17:23:53  * brsonjoined
17:24:04  <mbroadst>qard: thanks I'll give it a shot
17:25:15  <qard>Oh, the npm version is right, but if you checkout the repo, it's not on master.
17:25:28  <qard>Probably easiest to just npm install nodefly-v8-profiler
17:25:31  <mbroadst>right
17:32:21  * inolenjoined
17:44:09  * dshaw_joined
17:46:50  * piscisaureus_joined
17:48:10  * bnoordhuisjoined
17:50:17  * dapjoined
17:58:35  * kevinswiberjoined
17:58:59  * brsonquit (Ping timeout: 252 seconds)
17:59:08  <tjfontaine>piscisaureus_: cross closing issues from new repositories day?
17:59:17  * brsonjoined
17:59:21  <piscisaureus_>tjfontaine: huh?
17:59:47  <tjfontaine>you closed a couple issues with commits in libuv-wercker
18:00:18  <piscisaureus_>huh
18:00:19  <trevnorris>tjfontaine: that's actually github
18:00:36  <trevnorris>tjfontaine: it's going back and reading all the commit messages and closing associated issues.
18:00:36  <piscisaureus_>that was just a test push to a "private" repo of mine
18:00:39  <piscisaureus_>stupid github
18:00:52  <tjfontaine>trevnorris: yes well, it github did it because it was told :)
18:00:59  <tjfontaine>I figured it was supposed to be a private repo
18:01:20  <piscisaureus_>tjfontaine: which ones?
18:01:38  <trevnorris>i'm hating the github issue tracker more and more everyday.
18:01:44  <tjfontaine> #4809
18:01:50  <trevnorris>all the auto closing, cross linking crap
18:01:57  <tjfontaine>and #3172 was what trevnorris reopened
18:02:33  <trevnorris>something similar happeend to bnoordhuis recently, so figured it wasn't supposed to be closed. :)
18:02:42  <piscisaureus_>so only 2? good
18:02:46  <tjfontaine>ya
18:03:02  <piscisaureus_>4809 should indeed be closed
18:06:15  <bnoordhuis>https://github.com/piscisaureus/libuv-wercker ?
18:06:49  <bnoordhuis>is that a private repo? are you holding out on us, bertje?
18:07:16  <piscisaureus_>bnoordhuis: yes where I rewrite libuv in C++
18:07:37  <piscisaureus_>dropping unix support (who uses that anyway)
18:08:08  <bnoordhuis>yeah, i agree
18:08:18  <bnoordhuis>you're refocusing on the freedos backend, i suppose?
18:09:54  * benoitcquit (Changing host)
18:09:55  * benoitcjoined
18:10:05  <trevnorris>anyone feel like reviewing 4964 :)
18:12:01  <piscisaureus_>bnoordhuis: netware first
18:13:27  * kazuponquit (Remote host closed the connection)
18:14:13  * hzjoined
18:14:14  * hzquit (Changing host)
18:14:14  * hzjoined
18:16:50  * sblomjoined
18:19:19  * `3rdEdenjoined
18:23:44  * kazuponjoined
18:24:48  * kevinswiberquit (Ping timeout: 256 seconds)
18:27:26  <bnoordhuis>https://github.com/joyent/node/issues/5532 :-/
18:27:56  <trevnorris>heh
18:28:30  <trevnorris>isaacs: test-deprecation-flags.js is failing
18:28:52  <trevnorris>isaacs: heh, it's because the message is missing the period at the end. :)
18:29:56  <tjfontaine>trevnorris: does test-util-debug fail for you atm?
18:30:17  <trevnorris>tjfontaine: yes. also couple others, but let me verify those.
18:30:42  <tjfontaine>ok the others should just be decrecation errors
18:31:01  <trevnorris>yeah it looks like it.
18:31:03  <tjfontaine>bnoordhuis: if you get a chance could you look at #5182
18:31:12  <trevnorris>hm. also test-util-debug
18:31:32  <tjfontaine>I'm guessing it's mad at the deprecated methods
18:32:09  <tjfontaine>not sure how to write a test for what I'm doing currently
18:33:26  <trevnorris>ok. fix for those is trivial (remove '.' from the string) going to throw it on master
18:40:18  * kevinswiberjoined
18:45:44  <isaacs>tjfontaine, trevnorris: I'll push a fix shortly
18:45:46  <isaacs>sorry bout that
18:45:51  <trevnorris>thanks :)
18:47:22  <isaacs>piscisaureus_: i got your pmsg, but it looks like you've logged out and back on since then
18:47:33  <isaacs>piscisaureus_: try again :)
18:52:28  <tjfontaine>#5533 I'm not sure what a test for that would look like
18:55:13  <trevnorris>tjfontaine: what is uv_now do? no comments in the header
18:55:54  <tjfontaine>trevnorris: returns loop->time which is updated usually once per tick, using the monotonic clock
18:56:39  <trevnorris>tjfontaine: is there another way to get that information? if not then you'd just have to depend on the call, right?
18:57:08  <MI6>joyent/node: isaacs master * 52adc0d : test: Fix test-deprecation-flags - http://git.io/0KkGhw
18:58:14  * `3rdEdenquit (Remote host closed the connection)
19:01:21  * kevinswiberquit (Remote host closed the connection)
19:01:56  * kevinswiberjoined
19:06:19  * kevinswiberquit (Ping timeout: 256 seconds)
19:07:50  <MI6>nodejs-master: #229 UNSTABLE smartos-x64 (5/600) linux-x64 (1/600) linux-ia32 (1/600) http://jenkins.nodejs.org/job/nodejs-master/229/
19:08:28  * kevinswiberjoined
19:17:04  <MI6>nodejs-master-windows: #38 UNSTABLE windows-x64 (10/600) windows-ia32 (10/600) http://jenkins.nodejs.org/job/nodejs-master-windows/38/
19:17:32  * brsonquit (Ping timeout: 240 seconds)
19:18:22  * normanmquit (Quit: Computer has gone to sleep.)
19:23:31  * kazuponquit (Remote host closed the connection)
19:28:00  * kenperkinsquit (Quit: Computer has gone to sleep.)
19:34:14  * indexzerojoined
19:34:29  <bnoordhuis>#5533 seems to cause a lot of jenkins errors...
19:34:48  <bnoordhuis>having a hard time opening the tap results however
19:35:20  <bnoordhuis>AssertionError: "Trace: util.p: Use console.error() instead" == "Trace: util.p: Use console.error() instead."
19:35:23  * bnoordhuissighs
19:35:45  <trevnorris>bnoordhuis: isaacs just pushed a fix for that to master.
19:35:59  <bnoordhuis>okay. the other failures all seem tls related
19:36:15  * txdv_joined
19:36:59  * stagasquit (Read error: Connection reset by peer)
19:37:48  <trevnorris>bnoordhuis: you mean all 240 something? or you looking at a specific arch
19:38:02  <trevnorris>bnoordhuis: nm.
19:38:38  * ral_joined
19:38:54  <trevnorris>forgot about https, and debugger is never too happy
19:40:26  <bnoordhuis>trevnorris: ia32 linux
19:40:38  * `3rdEdenjoined
19:40:52  * indutny_joined
19:41:04  <bnoordhuis>x64 as well
19:42:03  <trevnorris>bnoordhuis: so seems linux and smartos hate the change, but osx are just fine. that's strange.
19:42:44  * ralquit (*.net *.split)
19:42:46  * txdvquit (*.net *.split)
19:42:46  * indutnyquit (*.net *.split)
19:42:47  * indutny_changed nick to indutny
19:43:15  <tjfontaine>heh, I guess I'll look into it on linux
19:43:19  * icarotjoined
19:43:44  * icarotquit (Remote host closed the connection)
19:43:48  <tjfontaine>trevnorris: about the only other way to avoid the boundary jump would be something like the tickinfobox
19:44:24  <trevnorris>tjfontaine: swear i'm going to keep rebasing and pushing my patch until the magic pixy dust lands on jenkins and give me green.
19:44:40  <tjfontaine>haha
19:47:13  <tjfontaine>so these tls failures are most certainly from my change to lib/tls.js which I'm guessing may be caused by being on the same tick so time hasn't progressed
19:47:43  <trevnorris>tjfontaine: but why would osx not have a problem?
19:48:04  <tjfontaine>trevnorris: potentially from how/when a read() gets fired from kqueue I guess
19:49:44  <trevnorris>oh man. v8 r14725 has landed. time to go exploring.
19:51:20  * icarotjoined
19:51:39  * icarotquit (Remote host closed the connection)
19:51:53  * dominictarrjoined
20:01:42  * bnoordhuisquit (Ping timeout: 256 seconds)
20:01:46  * AvianFlujoined
20:01:50  <trevnorris>well. that was a lot less painful than I thought it'd be
20:02:31  <trevnorris>indutny: you know if they're working on the removing Local<> patch? can't find it in their tracker.
20:02:32  <tjfontaine>aha, overflow
20:08:01  <indutny>trevnorris: hi, no idea
20:08:22  <trevnorris>indutny: thanks.
20:08:28  <trevnorris>lgtm anyone? https://github.com/joyent/node/pull/5536
20:08:34  <isaacs>reading
20:09:33  <isaacs>trevnorris: so, for V8 upgrades, it's best to have one commit that drops just the pristine V8, then another that reapplies any floating patches.
20:09:48  <trevnorris>isaacs: ok. i'll throw that in. it'll just take a min
20:09:59  <isaacs>trevnorris: sure
20:10:17  <isaacs>trevnorris: this makes it a lot easier to maintain in the future, since we can diff and see exactly what we changed
20:12:03  * hzquit (Ping timeout: 256 seconds)
20:12:55  * leonvvjoined
20:18:41  * `3rdEdenquit (Remote host closed the connection)
20:19:24  <trevnorris>isaacs: fixed, tested and benchmarked. looks good :)
20:22:00  * icarotjoined
20:23:31  * sblomquit (Ping timeout: 260 seconds)
20:23:58  * sblomjoined
20:24:40  * mikealjoined
20:27:11  <trevnorris>ok, wtf is isaacbw?
20:27:42  * bnoordhuisjoined
20:27:50  <tjfontaine>someone who's helpful in #node.js
20:29:34  * mikealquit (Ping timeout: 276 seconds)
20:29:59  <isaacs>trevnorris: isaacbw is a person with a very dashing first name.
20:30:15  <isaacs>trevnorris: minor style nit. otherwise lgtm if tests pass.
20:30:45  * sblomquit (Ping timeout: 256 seconds)
20:31:07  * sblomjoined
20:32:42  <tjfontaine>silly numbers, and silly osx
20:34:24  * mikealjoined
20:35:54  <trevnorris>isaacs: i've pushed that line to 80. any further will overflow. want me to back it up and line up w/ "New"?
20:36:43  * bnoordhuisquit (Ping timeout: 276 seconds)
20:37:13  <tjfontaine>trevnorris: you could Persistent<Context> tmp;\n tmp = ...
20:38:31  <trevnorris>tjfontaine: i could. but that makes my styley sense tingle.
20:38:53  <trevnorris>:)
20:39:21  <tjfontaine>you could typedef Persistent<Context> into something shorter :)
20:40:06  <trevnorris>lol
20:40:32  <tjfontaine>it's not uncommon, depending on how often you reuse the type
20:41:02  <trevnorris>twice in node_script and once in node.cc
20:43:07  <tjfontaine>I didn't mean common in node, but in the world of C and C++ :)
20:44:20  <trevnorris>ah, ok. :)
20:44:31  <tjfontaine>especially in worlds with lots of templates
20:45:21  <tjfontaine>ok #5533 behaves better
20:45:40  <tjfontaine>silly mistake
20:46:28  * sblomquit (Ping timeout: 276 seconds)
20:47:07  * sblomjoined
20:53:28  * brsonjoined
20:53:51  * kazuponjoined
20:58:45  * kazuponquit (Ping timeout: 248 seconds)
20:59:26  * kenperkinsjoined
20:59:56  * mikealquit (Quit: Leaving.)
21:01:09  * leonvvquit (Remote host closed the connection)
21:01:13  * AvianFluquit (Remote host closed the connection)
21:01:56  * rendarquit
21:02:14  * leonvvjoined
21:02:28  * leonvvquit (Remote host closed the connection)
21:03:25  * Somebodyjoined
21:03:52  * Somebodyquit (Remote host closed the connection)
21:08:00  <trevnorris>isaacs: space issue/
21:08:06  <trevnorris>s/\/?
21:08:18  <isaacs>trevnorris: hanging args should be lined up with one anther
21:08:47  <isaacs>trevnorris: if that can't be done, create a separate variable, and use that.
21:09:43  * skebcio_quit (Ping timeout: 245 seconds)
21:09:53  <trevnorris>isaacs: wait. wtf. that code isn't even necessary. it's just extracting the Local<> out of the persistent anyways. :P
21:10:04  * defunctzombiechanged nick to defunctzombie_zz
21:10:18  <isaacs>trevnorris: even better :)
21:11:00  * skebciojoined
21:13:13  <trevnorris>isaacs: coolio. much better.
21:16:23  <trevnorris>isaacs: done. good to go?
21:16:46  <isaacs>one sec
21:17:53  <isaacs>trevnorris: lgtm
21:18:28  <MI6>joyent/node: Trevor Norris master * 2868bf9 : src: upgrade to new v8 api (+2 more commits) - http://git.io/wNUV0w
21:20:45  * AvianFlujoined
21:22:34  * sblomquit (Ping timeout: 245 seconds)
21:23:27  * sblomjoined
21:29:11  <MI6>nodejs-master: #230 UNSTABLE smartos-x64 (3/600) linux-x64 (1/600) http://jenkins.nodejs.org/job/nodejs-master/230/
21:31:34  * skebcioquit (Read error: Connection reset by peer)
21:31:41  * skebciojoined
21:32:27  * Benvie_quit (Remote host closed the connection)
21:32:44  * Benvie_joined
21:35:43  * c4miloquit (Remote host closed the connection)
21:37:34  * icarotquit (Ping timeout: 245 seconds)
21:39:38  <MI6>nodejs-master-windows: #39 UNSTABLE windows-x64 (13/600) windows-ia32 (10/600) http://jenkins.nodejs.org/job/nodejs-master-windows/39/
21:40:17  <tjfontaine>ooh new failures on winders
21:40:41  * pachetquit (Quit: leaving)
21:40:43  <trevnorris>tjfontaine: yeah. just noticed that.
21:40:59  <trevnorris>don't understand how test-buffer.js could have failed.
21:41:07  <tjfontaine>I wonder if thsee go back further, test-buffer is kinda scary and sounds like something that mgith be related to the stringbytes stuff
21:41:41  <tjfontaine>ya test-buffer is failing on older builds
21:42:12  * bnoordhuisjoined
21:42:14  <trevnorris>tjfontaine: where buffer test first failed: http://jenkins.nodejs.org/job/nodejs-master-windows/33/testReport/
21:42:49  <tjfontaine>trevnorris: right which is the merge from v0.10 containing the string bytes work
21:42:58  * bnoordhu1sjoined
21:43:20  <tjfontaine>bnoordhu1s: you're multiplying
21:43:45  <bnoordhu1s>i'm going exponential. it's time
21:44:01  <bnoordhu1s>there were two prs you wanted me to review, right?
21:44:08  * kevinswiberquit (Remote host closed the connection)
21:44:17  <trevnorris>sure. think one of those was 4964 ;-)
21:44:27  <tjfontaine>er, right #5533 is the Timer.now thing
21:44:42  * kevinswiberjoined
21:44:48  * kevinswiberquit (Read error: Connection reset by peer)
21:45:02  <bnoordhu1s>apropos nothing but it's almost midnight here
21:45:03  <tjfontaine>#5182 also, but less interesting
21:45:14  <bnoordhu1s>and i haven't even started on the thing that was supposed to be my primary objective today
21:45:17  * kevinswiberjoined
21:45:17  <bnoordhu1s>i suck at planning :(
21:47:40  * bnoordhuisquit (Ping timeout: 256 seconds)
21:52:53  <trevnorris>bnoordhu1s: isn't it almost midnight there?
21:53:02  <bnoordhu1s>trevnorris: yes
21:53:17  * bnoordhu1schanged nick to bnoordhuis
21:53:50  <trevnorris>heh, then don't worry about it. now it was your primary objective from yesterday. ;-)
21:54:49  * indexzeroquit (Quit: indexzero)
21:54:51  * kazuponjoined
21:58:01  <bnoordhuis>i've been benchmarking node's http client
21:58:31  * sblomquit (Ping timeout: 252 seconds)
21:58:44  <bnoordhuis>it's 150x slower to issue a POST request with a buffer as its body than a string
21:58:45  * sblomjoined
21:59:01  <trevnorris>say what?
21:59:04  <bnoordhuis>if these numbers are to be believed that is
21:59:15  <trevnorris>bnoordhuis: is there a specific test you're using. like to run it against my buffer branch
21:59:39  * kazuponquit (Ping timeout: 260 seconds)
21:59:42  <bnoordhuis>trevnorris: https://github.com/bnoordhuis/node/commit/8b7b3e7
22:00:00  <bnoordhuis>my patch isn't very effective so far :/
22:00:26  <trevnorris>i think http-parser could use some substantial improvements, if we just didn't set the object properties in cc.
22:01:28  <bnoordhuis>well, i don't think that's it
22:01:52  <trevnorris>oh yeah. that was just thinking out lout (chat?) :)
22:02:01  <trevnorris>s/lout/loud
22:03:04  <bnoordhuis>0.14user 0.01system 0:03.03elapsed 5%CPU <- type=buf
22:03:11  <bnoordhuis>2.61user 0.44system 0:03.04elapsed 100%CPU <- type=asc
22:11:31  <bnoordhuis>piscisaureus_: so, you'll be speaking at the quicken loans technology conference, i gather? :)
22:11:53  <isaacs>ugh, this http client/agent code is such shit.
22:12:08  <tjfontaine>at least it's not all in one giant file :)
22:12:24  <tjfontaine>though now youg et my fun circular dependencies
22:13:44  <isaacs>yeah
22:13:53  <isaacs>there's alot of duplicated code, and long arg lists.
22:14:06  <isaacs>where really there ought to be options objects being passed around.
22:14:27  <isaacs>but the options objects for http aren't compatible with the options obj for net.connect or tls.connect
22:14:52  <trevnorris>bnoordhuis: for buf: "4634 97.1% 97.3% syscall"
22:17:38  <trevnorris>bnoordhuis: and strace -c show "71.67 0.000382 0 1385 write"
22:18:17  <bnoordhuis>trevnorris: what does type=asc give you?
22:18:18  <trevnorris>bnoordhuis: oh, and my buffer branch shows 0 improvement. which is really strange for anything that requires a buffer.
22:18:43  <trevnorris>bnoordhuis: "60.92 0.000901 0 9325 write"
22:18:53  <bnoordhuis>trevnorris: the issue is that OutgoingMessage#write() calls _send() multiple times when the request body is a buffer
22:19:04  <trevnorris>ah ok
22:20:40  <trevnorris>bnoordhuis: oh, wait. hold on. that was from my buffer branch. here are the results from master
22:20:56  <trevnorris>type=buf - "83.85 0.004064 8 480 epoll_wait"
22:21:18  <trevnorris>type=asc - "63.31 0.000535 0 8943 write"
22:21:31  <trevnorris>(method=write)
22:21:56  <bnoordhuis>yeah, that's about what i'm seeing with v0.10
22:22:20  <bnoordhuis>what happens is that a) OutgoingMessage#write() calls _send() several times
22:22:41  <bnoordhuis>and b) each _send() ends up in its own tcp packet
22:22:50  <bnoordhuis>hence you get lots more roundtrip traffic
22:22:58  <trevnorris>ah ok
22:23:18  <bnoordhuis>that's why libuv spends so much time in epoll_wait, it's waiting for acks
22:23:40  <isaacs>bnoordhuis: is this better in master, where we use writev?
22:23:55  <bnoordhuis>i don't think so but let me check
22:24:20  <trevnorris>well, this issue goes all the way back to v0.8 branch
22:24:25  <bnoordhuis>no, same - 50 req/s and only 5% cpu utilization
22:26:03  <piscisaureus_>bnoordhuis: we thought you'd be a good candidate
22:26:22  <bnoordhuis>piscisaureus_: i'd go but i'm afraid of detroit
22:26:33  <tjfontaine>depends on where in detroit rock city you are
22:26:38  * mikealjoined
22:26:43  <bnoordhuis>you mean it has good parts?
22:26:43  <piscisaureus_>bnoordhuis: apparently a lot of family members will be watching. You are the family man.
22:27:00  <bnoordhuis>piscisaureus_: they specifically asked for you, it said so in the email :)
22:27:04  <tjfontaine>bnoordhuis: they are few and far between, but yes they still exist
22:27:47  <piscisaureus_>bnoordhuis: you know how nobody can tell bert from ben
22:28:01  <bnoordhuis>sometimes that's a good thing imo :)
22:29:21  <bnoordhuis>lib/http.js is turning out to be very tamper-resistant :/
22:29:49  <bnoordhuis>whatever i do, it refuses to pack the headers and the buffer into a single packet
22:35:38  * sblomquit (Ping timeout: 240 seconds)
22:35:50  * sblomjoined
22:36:29  <bnoordhuis>isaacs: why does OutgoingMessage#write() put binary encoding strings in the slow path?
22:37:24  <bnoordhuis>you added it in commit e4b716ef with the following commit log:
22:37:24  <bnoordhuis> We were assuming that any string can be concatenated safely to
22:37:25  <bnoordhuis> CRLF. However, for hex, base64, or binary encoded writes, this
22:37:25  <bnoordhuis> is not the case, and results in sending the incorrect response.
22:38:06  <bnoordhuis>what does (binary_string + '\r\n') mess up?
22:40:18  * sblomquit (Ping timeout: 264 seconds)
22:40:29  * sblomjoined
22:42:45  <trevnorris>TooTallNate: know some really commonly used native modules?
22:43:03  <TooTallNate>trevnorris: isaacs can probably make a definitive list
22:43:07  <TooTallNate>trevnorris: mine will be biased :p
22:43:19  <trevnorris>heh
22:43:35  <trevnorris>honestly i'm just looking to get an idea of how other devs use node natively.
22:43:36  <TooTallNate>trevnorris: for your Buffer branch, all i ask is that https://github.com/tootallnate/ref continues to work
22:43:45  <wolfeidau>trevnorris: ws is one
22:44:11  <TooTallNate>wolfeidau: ws *should not* be a native module…
22:44:15  <TooTallNate>i mean it is...
22:44:20  <TooTallNate>but that's an unfortunate fact
22:44:30  <tjfontaine>indeed I woudl say ffi/ref/ws are probably the current top, probably also the dtrace one
22:44:44  <wolfeidau>trevnorris: Is a dependency socketio and yeah not ideal eh
22:44:47  <TooTallNate>trevnorris: contextify is another big one
22:45:41  <wolfeidau>TooTallNate: Your weak one is also seen around
22:45:49  <wolfeidau>am i right?
22:45:49  <TooTallNate>ahh right
22:45:55  <TooTallNate>i think that's broken atm though :p
22:46:00  <TooTallNate>on master node that is
22:46:06  <TooTallNate>trevnorris sent a PR already
22:46:17  <tjfontaine>was what we ship with node patched?
22:46:34  <TooTallNate>trevnorris: ya it's patched already, haha
22:46:41  <tjfontaine>tj :P
22:46:49  <TooTallNate>damnit
22:47:09  <TooTallNate>now i have *two* IRC rooms where i have to watch out for the autocomplete :p
22:47:15  <TooTallNate>since we just hired another ma* :p
22:47:26  <tjfontaine>heh
22:48:01  * sblomquit (Ping timeout: 276 seconds)
22:48:15  * sblomjoined
22:48:33  <wolfeidau>TooTallNate: I used your cross compile nodejs post the other day btw, nice work
22:48:44  <TooTallNate>wolfeidau: thanks :D
22:48:58  <TooTallNate>wolfeidau: i just landed a patch to ./configure last night that will make that a little more reliable as well
22:49:42  <wolfeidau>sweet I am now working for a company that runs node on ARM so learning a lot about building and maintaining stuff on that platform
22:49:52  <trevnorris>TooTallNate: ok, so not completely sure. since all Buffer instances are now just the js object, which means it's a Local<>, I can't figure out how to hand back the Persistent handle instead.
22:50:16  <TooTallNate>trevnorris: so it's not compiling?
22:50:28  <trevnorris>TooTallNate: oh, i'm just looking at source.
22:50:33  <TooTallNate>oh ok
22:50:45  <trevnorris>TooTallNate: but no, right now it wouldn't compile. there'll be some changes.
22:50:55  <trevnorris>minor though
22:50:57  <TooTallNate>trevnorris: it that due to V8 changes or the Buffer API changes?
22:51:10  * icarotjoined
22:51:26  <trevnorris>TooTallNate: Buffer API changes. e.g. no more c++ Buffer class. The return value of Buffer::New() is, well, Local<Object>
22:51:46  <TooTallNate>i see
22:52:08  <TooTallNate>i can probably live with that
22:52:25  <icarot>Is that what that Buffer nonsense was?
22:52:37  <icarot>Trying to learn libuv and didn't understand it.
22:52:51  <icarot>I was expecting it all to just be vanilla C, so I guess that's partially why I was confused.
22:52:54  <bnoordhuis>icarot: buffers in node and libuv are different things
22:52:58  <TooTallNate>icarot: well Buffer is a node.js thing, libuv has a lower-level concept
22:53:01  <trevnorris>TooTallNate: but did add Buffer::Use() so you don't have to do the whole noop callback thing. :)
22:53:20  * piscisaureus_quit (Read error: Connection reset by peer)
22:53:25  <TooTallNate>trevnorris: so what, i can pass an arbitrary pointer to that?
22:53:35  <icarot>Well, I was speaking specifically of Buffers as they pertain either to uv's fs or network procedures.
22:53:41  <tjfontaine>there is a uv_buf_t, which is a wrapper around memory you allocate
22:53:50  <bnoordhuis>icarot: btw, what do you mean by vanilla c? libuv is c89, doesn't get much more vanilla than that, right?
22:53:51  <trevnorris>TooTallNate: yeah. it'll use whatever you pass to it. then auto cleanup on gc
22:54:07  <TooTallNate>trevnorris: cool… sounds like there's some new fun APIs to play with :)
22:54:10  * defunctzombie_zzchanged nick to defunctzombie
22:54:15  <bnoordhuis>icarot: libuv buffers are mostly just struct iovecs by another name
22:54:33  <trevnorris>TooTallNate: definitely. There's a new Alloc cc api that has you written all over it. :)
22:54:33  <icarot>bnoordhuis: Whoops, certainly of course. That's definitely vanilla. I just meant, I didn't realize that 'C++ API style' code was in use. Or did I understand incorrectly?
22:54:48  <bnoordhuis>icarot: we needed our own type because windows doesn't have struct iovec
22:54:59  <bnoordhuis>and i don't understand the comment about c++ api style...
22:55:25  * kazuponjoined
22:55:30  <icarot>bnoordhuis: ignore it. I'm being stupid. I just woke up and I'm rambling about nonsense.
22:55:31  <TooTallNate>trevnorris: so does Buffer.alloc() set up the indexed properties as well?
22:55:39  <bnoordhuis>okay :)
22:55:39  <TooTallNate>or is that still Buffer.makeFastBuffer()?
22:56:11  <trevnorris>TooTallNate: makeFastBuffer no longer exists. Neither does SlowBuffer. Buffer.alloc sets up the external array index.
22:56:18  <TooTallNate>trevnorris: awesome :)
22:56:27  <TooTallNate>trevnorris: that was the answer i was hoping for =
22:59:25  <icarot>bnoordhuis: am I correct in saying that most of the documentation is in the header file? So, since the libuv book is pretty out of date, and that's what I've been using, is that what would be the best reference?
22:59:41  <icarot>(one of the header files, I don't remember which).
22:59:54  <trevnorris>icarot: include/uv.h?
22:59:56  <bnoordhuis>icarot: yes, uv.h is the canonical documentation
22:59:58  <icarot>yes.
23:01:04  <icarot>Okay. Great, just checking. Any knowledge of code bases that build on libuv other than Node?
23:01:28  <trevnorris>icarot: https://github.com/joyent/libuv/wiki/Projects-that-use-libuv
23:01:42  <icarot>Saw that. Was too scarce.
23:01:48  <icarot>Okay, was just wondering.
23:02:19  * kazuponquit (Ping timeout: 276 seconds)
23:03:42  * paddybyersquit (Ping timeout: 264 seconds)
23:05:32  <bnoordhuis>icarot: node, rust, julia and luvit are the best known users
23:06:08  <trevnorris>bnoordhuis: ah, heh. totally missed that comment. think I should remove it, or just leave for a latter commit?
23:06:19  <bnoordhuis>trevnorris: i'd remove it
23:06:24  <trevnorris>ok
23:07:13  <icarot>bnoordhuis: I've read the contributing spiel, but any tips for helping to understanding a foreign code base?
23:07:25  <bnoordhuis>icarot: ask questions :)
23:07:53  <icarot>bnoordhuis: Perhaps along the lines of "This looks weird!" "Does this need fixing?!:
23:08:24  <bnoordhuis>well, i was thinking of 'how does x work?' or 'why does x do y when z?'
23:08:38  <icarot>Oh. Even lower-level. Heh, okay.
23:09:01  <bnoordhuis>as to 'does this need fixing?' questions, only when they're about non-trivial things
23:09:13  <bnoordhuis>things like minor style violations, well, *shrug*
23:09:24  <icarot>Will do then. I've never interacted with POSIX APIs before and my C is rusty, so I have a lot of work out for me.
23:10:35  <icarot>And I was peeping in on the conversation on the memory-leak, and while awesome that someone seems to have found it, I definitely don't think I'll be accomplishing anything of that magnitude anytime soon :O
23:10:48  <MI6>joyent/node: Trevor Norris master * f58eb8f : src: remove unused comment - http://git.io/aseMjw
23:11:01  <isaacs>This: https://github.com/joyent/node/pull/5531
23:11:05  <isaacs>I dont' think this is so good.
23:11:17  * jesusabdullahjoined
23:11:35  * Domenic_joined
23:11:37  <jesusabdullah>isaacs: ftr I was just a little taken aback by that pr, had no idea what to even make of it
23:11:46  <bnoordhuis>isaacs: i told him to submit it if he wanted to but that it was unlikely we'd take it
23:11:54  <jesusabdullah>isaacs: was just curious about your reaction to it
23:12:01  <isaacs>bnoordhuis: well, i mean... what are you doing running untrusted code in nodejs?
23:12:08  <isaacs>that's insane.
23:12:18  <bnoordhuis>yeah, so i told him to use seccomp2 or namespaces :)
23:12:30  <jesusabdullah>I mean
23:13:09  <jesusabdullah>I'd run untrusted code either in a require('vm') or in a triple-dogged chroot maximum security prisonJail
23:13:45  <bnoordhuis>require('vm') is not for running untrusted code either
23:13:46  <jesusabdullah>but that doesn't mean everyone's needs are the same I guess
23:14:26  <jesusabdullah>bnoordhuis: I mean you can at least limit the scope of what's exposed to that though right? Or am I way off base here?
23:14:53  <bnoordhuis>jesusabdullah: yes, but it's actually pretty hard to get it right
23:15:06  * mikealquit (Quit: Leaving.)
23:15:18  <bnoordhuis>unless you're very, very careful it's often easy to escape from the vm by walking prototype chains
23:15:34  <jesusabdullah>uh-huh, I see
23:16:02  * mikealjoined
23:16:08  <jesusabdullah>so you wouldn't really call that sufficient for malicious content XD
23:16:18  <bnoordhuis>no indeed.
23:16:32  <jesusabdullah>I'm pretty trusting I think :(
23:16:36  <bnoordhuis>that's why there's a big warning at the top of vm.markdown because 99 out of 100 coders are going to get it wrong
23:16:44  <bnoordhuis>including yours truly probably :)
23:16:56  <jesusabdullah>good to know
23:16:56  <jesusabdullah>I mean
23:17:04  <isaacs>the way to run untrusted code is to spin up a brand new instance of your OS in a virtual machine, run the code in the VM, then kill the VM
23:17:35  <jesusabdullah>I've never exposed a vm like that saying, "hey world run whatever the fuck there is NO WAY this can GO WRONG"
23:17:50  <jesusabdullah>I've done it more from the context of, say, making a dsl interpreter or something like that
23:18:15  <tjfontaine>fwiw there probably is a decent amount of sandbox'ing already happening from the android layer
23:18:22  <jesusabdullah>also isaacs yeah I'm with it
23:18:32  <tjfontaine>they may be able to get with just a simple process model there
23:19:03  <tjfontaine>*get by
23:20:10  <jesusabdullah>anyhoo, I think I get the idea
23:20:28  * jesusabdullahpart ("I'll always remember you.")
23:20:31  * stagasjoined
23:20:45  <MI6>nodejs-master: #231 UNSTABLE smartos-x64 (5/600) linux-ia32 (1/600) http://jenkins.nodejs.org/job/nodejs-master/231/
23:21:30  <tjfontaine>so I've investigated those smartos x64 debugger failures, the reason they're failing is kinda depressing
23:22:08  <isaacs>tjfontaine: oh?
23:22:20  <tjfontaine>basically we're doing stream.on('data', function () { //figure out how many lines hvae been received }); and failing to handle the buffering very well
23:22:31  <tjfontaine>which is a pattern we do all over the place in the test suites
23:22:49  <tjfontaine>some of these same failures are happening on windows as well
23:22:58  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/ccaa065
23:23:13  <tjfontaine>I had this lying around and using it in place makes it easier to predict https://gist.github.com/tjfontaine/10ddf5cf1ce44c6341b0
23:23:40  <bnoordhuis>isaacs: unless you can explain to me what the issue with binary strings and CRLF is :)
23:23:49  <tjfontaine>so we can just do it once and then .pipe() or way into the fun
23:26:10  <isaacs>bnoordhuis: hmm.... you're right about binary strings.
23:26:15  <isaacs>length should not be altered...
23:26:22  <isaacs>i wonder what i was thinking.
23:26:26  <isaacs>hex and base64, of course.
23:26:28  <trevnorris>bnoordhuis: about that. doesn't the data as a string, need to be converted to a char* anyways before being sent out?
23:26:30  <isaacs>and utf8
23:26:36  <bnoordhuis>or i. i think i was the one that reviewed your patch
23:26:56  <bnoordhuis>trevnorris: yeah, but that doesn't alter string.length in js land
23:27:07  <isaacs>bnoordhuis: you can remove the "revisit in 0.11" since we just do writev in 0.11
23:27:21  <isaacs>bnoordhuis: or change it ot "irrelevant for 0.11"
23:27:22  <isaacs>:0
23:27:36  <bnoordhuis>right, that's why i put it there
23:27:36  <trevnorris>bnoordhuis: hm. don't understand the specifics here, but just seems strange that taking data from a buffer, back into a v8 string, then back into a char* would be faster.
23:27:43  <bnoordhuis>i'll change it to 'remove'
23:28:03  <bnoordhuis>trevnorris: have you read the commit log? :)
23:28:44  <bnoordhuis>it's not fast in terms of cpu utilization. it is however fast in terms of network throughput
23:28:53  <bnoordhuis>it saves a ton of roundtrips
23:29:00  <trevnorris>bnoordhuis: ah, got it.
23:29:02  <isaacs>bnoordhuis: yeah, it's a bit weird, though, to be using binary strings like this
23:29:11  <isaacs>bnoordhuis: since they're one of the slowest to decode
23:29:24  <bnoordhuis>well, i'm not married to this approach. i'm open to suggestions
23:29:29  <isaacs>bnoordhuis: and they use double the space
23:29:30  <trevnorris>isaacs: as in it will eventually be fixed in master? since the performance is the same as v0.10 right now.
23:30:02  <bnoordhuis>alternatively, i can alloc a new buffer and copy over the data
23:30:59  <MI6>nodejs-master-windows: #40 UNSTABLE windows-x64 (11/600) windows-ia32 (11/600) http://jenkins.nodejs.org/job/nodejs-master-windows/40/
23:37:47  <trevnorris>bnoordhuis: hm. will be faster than it is, but probably not be as fast as your current patch
23:40:06  <bnoordhuis>trevnorris: no, indeed. it also makes it impossible to pack the headers and the body
23:40:28  <bnoordhuis>it doesn't even make much of a mem usage difference
23:40:52  <isaacs>bnoordhuis: as long as it doesn't break any tests or make the other http benchmarks slower, i'm ok with it, i guess
23:41:43  <bnoordhuis>i guess there's one more optimization possible
23:42:12  <bnoordhuis>iff headersSent==true and chunkedEncoding==false, then don't convert to string because it doesn't matter anymore
23:42:34  * AvianFluchanged nick to AvianFluuuuuu
23:43:42  * bnoordhu1sjoined
23:44:24  <bnoordhuis>^ that's my macbook waking up out of the blue
23:45:29  <trevnorris>why does this look hellishly unsafe to me: "static Local<Value>* current_buffer;"
23:45:38  <tjfontaine>bnoordhuis: ya we see that every once in a while, I just assume it's you sleep-irc'ing
23:46:26  * dapquit (Quit: Leaving.)
23:46:47  <bnoordhuis>tjfontaine: i don't know why it does that. i speculate it's moving to a deeper sleep state but who can say?
23:46:53  <tjfontaine>nod
23:47:27  * bnoordhu1squit (Read error: Operation timed out)
23:54:16  <trevnorris>isaacs: think we can get 5520 kicked out? seems kinda dangerous.
23:54:26  <trevnorris>isaacs: actually caused my computer to crash couple days ago.
23:57:34  * brsonquit (Ping timeout: 276 seconds)
23:58:09  * kazuponjoined