00:08:46  * travis-cijoined
00:08:46  <travis-ci>[travis-ci] joyent/node#514 (v0.6 - 692bcbe : isaacs): The build passed.
00:08:46  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/7f58d20...692bcbe
00:08:46  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/754858
00:08:46  * travis-cipart
00:20:01  * mmaleckichanged nick to mmalecki[zzz]
00:24:19  <TooTallNate>isaacs: https://github.com/isaacs/fstream/pull/4
00:24:28  <TooTallNate>found the problem :)
00:26:37  <TooTallNate>holy crap, that was painful debugging :O
00:26:51  <TooTallNate>you need a test suite for that guy ;)
00:27:11  * dapjoined
00:36:12  <isaacs>mmalecki[zzz]: you around?
00:36:17  <isaacs>TooTallNate: yes, agreed.
00:36:37  <isaacs>TooTallNate: fstream's mission in life is to support node-tar, which is pretty thoroughly tested.
00:36:55  <isaacs>TooTallNate: basically, i just yanked it out of tar when it got big enough
00:37:56  <isaacs>TooTallNate: oh, there is a test command, it looks like
00:38:01  <isaacs>TooTallNate: just lives in ./examples
00:38:18  <TooTallNate>oh ok
00:38:21  <TooTallNate>:)
00:38:31  <TooTallNate>isaacs: patch look good?
00:38:35  <isaacs>yeah, pulled in
00:39:07  <isaacs>pushing, publishing, 0.1.13
00:39:16  <isaacs>thanks!
00:39:30  <TooTallNate>isaacs: thank you!
00:39:36  * brsonquit (Quit: leaving)
00:39:48  * brsonjoined
01:13:55  <bnoordhuis>mmalecki[zzz]: you might dig this: https://github.com/bnoordhuis/node-http-parser
01:21:26  * CoverSli1ejoined
01:21:47  <isaacs>bnoordhuis: care to review? https://github.com/isaacs/node/commit/864a9b37e1b18fd992376f4708d308c1d7c96e72
01:22:06  <isaacs>bnoordhuis: down to only one test failure on os x (most of the time)
01:22:28  * CoverSlidequit (Read error: Connection reset by peer)
01:23:06  <bnoordhuis>isaacs: what does `Script.Script = Script;` do?
01:23:15  <isaacs>bnoordhuis: require('vm').Script needs to work
01:23:29  <isaacs>bnoordhuis: the line above i set it to module.exports
01:23:40  <bnoordhuis>ah right
01:23:41  <isaacs>bnoordhuis: similar to require('stream') === require('stream').Stream
01:24:57  <bnoordhuis>isaacs: lgtm, i think
01:25:06  <isaacs>kewl
01:25:07  <bnoordhuis>is there a reason to stuff the binding into `this`?
01:25:19  <bnoordhuis>i mean, can't it be captured in a closure?
01:25:28  <isaacs>bnoordhuis: oh, yeah, i guess so.
01:25:31  <isaacs>then it won't be exposed at all.
01:25:37  <isaacs>godo idea.
01:29:27  <isaacs>bnoordhuis: https://github.com/isaacs/node/commit/44daa9836b5bb4cfb8434b35bb1a12fc324223aa
01:30:22  <bnoordhuis>isaacs: lgtm
01:33:19  * mikealjoined
01:33:49  <CIA-99>node: isaacs master * r29463cb / test/simple/test-process-argv-0.js : Update argv0 test to verify correct behavior - http://git.io/UgCDgQ
01:33:50  <CIA-99>node: isaacs master * r44daa98 / lib/vm.js :
01:33:50  <CIA-99>node: Wrap NodeScript binding class in JavaScript layer
01:33:50  <CIA-99>node: This makes it easy to prevent errors where Script methods
01:33:50  <CIA-99>node: are called on non-script objects, resulting in Assertion failures. - http://git.io/nr78QQ
01:34:42  <AvianFlu>bnoordhuis, there's a stdio bug on unix in my detached child process code that you merged, it needs something like this to work properly https://github.com/AvianFlu/libuv/commit/be1cd35a8ca8413c798c973e1ff67fb332858fe0
01:35:24  <bnoordhuis>AvianFlu: i'll look at it tomorrow, i was just about to be off to bed
01:35:38  <AvianFlu>cool, np
01:35:42  <bnoordhuis>but thanks for pointing it out :)
01:36:16  * bnoordhuisquit (Quit: Leaving)
01:48:08  * Ariajoined
01:49:23  * travis-cijoined
01:49:23  <travis-ci>[travis-ci] joyent/node#515 (master - 44daa98 : isaacs): The build is still failing.
01:49:23  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/493a6bb...44daa98
01:49:23  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/755390
01:49:23  * travis-cipart
01:56:19  * dapquit (Ping timeout: 246 seconds)
02:02:57  * mikealquit (Quit: Leaving.)
02:03:52  * dapjoined
02:09:56  * mikealjoined
02:18:42  * dshaw_quit (Ping timeout: 260 seconds)
02:35:40  * mralephquit (Quit: Leaving.)
02:41:21  * dapquit (Quit: Leaving.)
02:45:17  * brsonquit (Quit: leaving)
03:07:11  * isaacschanged nick to isaacs[away]
03:09:05  * sh1mmerquit (Quit: sh1mmer)
03:10:26  * sh1mmerjoined
03:14:03  * mikealquit (Quit: Leaving.)
03:31:47  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
03:33:42  * perezdquit (Quit: perezd)
03:33:51  * TooTallNatejoined
03:42:06  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
04:04:49  * Ariaquit (Remote host closed the connection)
04:31:39  * ryahquit (Quit: leaving)
04:39:39  * dshaw_joined
04:55:43  * felixgejoined
04:55:43  * felixgequit (Changing host)
04:55:43  * felixgejoined
04:57:23  * mikealjoined
05:00:35  * mikealquit (Client Quit)
05:04:26  * AvianFluchanged nick to AvianusAsleepus
05:10:50  * isaacs[away]changed nick to isaacs
05:28:19  * mikealjoined
05:28:37  * mikealquit (Client Quit)
05:38:46  * isaacsquit (Remote host closed the connection)
05:39:46  * isaacsjoined
05:44:37  * isaacsquit (Ping timeout: 260 seconds)
05:54:54  * pfox___quit (Ping timeout: 248 seconds)
06:12:26  * felixgequit (Quit: felixge)
07:01:46  * sh1mmerquit (Quit: sh1mmer)
07:02:49  * mikealjoined
07:02:51  * sh1mmerjoined
07:03:46  * sh1mmerquit (Read error: Connection reset by peer)
07:03:49  * sh1mmer_joined
07:06:06  * mikealquit (Client Quit)
07:08:00  * sh1mmer_quit (Ping timeout: 240 seconds)
07:26:08  * mikealjoined
07:33:33  * mikealquit (Quit: Leaving.)
07:44:48  * dshaw_1joined
07:45:58  * dshaw_quit (Ping timeout: 245 seconds)
07:53:12  * stephankquit (Quit: *Poof!*)
08:05:19  * rendarjoined
08:09:06  * paddybyersjoined
08:11:36  * paddybyers_joined
08:13:52  * paddybyersquit (Ping timeout: 276 seconds)
08:14:06  * paddybyersjoined
08:16:29  * paddybyers_quit (Ping timeout: 260 seconds)
08:16:38  * paddybyers_joined
08:19:04  * paddybyersquit (Ping timeout: 276 seconds)
08:19:08  * paddybyers__joined
08:20:20  * mralephjoined
08:21:20  * paddybyers_quit (Ping timeout: 272 seconds)
08:23:32  * paddybyers__quit (Ping timeout: 240 seconds)
08:25:21  * paddybyersjoined
08:27:53  * paddybyers_joined
08:30:07  * paddybyersquit (Ping timeout: 276 seconds)
08:30:21  * paddybyersjoined
08:32:44  * paddybyers_quit (Ping timeout: 272 seconds)
08:32:56  * paddybyers_joined
08:33:58  * felixgejoined
08:33:58  * felixgequit (Changing host)
08:33:58  * felixgejoined
08:34:43  * paddybyersquit (Ping timeout: 245 seconds)
08:35:26  * paddybyersjoined
08:37:07  * paddybyers_quit (Ping timeout: 245 seconds)
08:37:56  * paddybyers_joined
08:40:27  * paddybyers__joined
08:40:31  * paddybyersquit (Ping timeout: 276 seconds)
08:40:32  * paddybyers__changed nick to paddybyers
08:42:44  * paddybyers_quit (Ping timeout: 260 seconds)
08:44:57  * paddybyersquit (Ping timeout: 255 seconds)
08:45:22  * paddybyersjoined
08:47:49  * paddybyers_joined
08:49:33  * paddybyersquit (Ping timeout: 244 seconds)
08:50:23  * paddybyersjoined
08:52:52  * paddybyers__joined
08:52:52  * paddybyers_quit (Ping timeout: 276 seconds)
08:53:12  * paddybyers__part
08:54:59  * paddybyersquit (Ping timeout: 260 seconds)
09:04:15  * mralephquit (Quit: Leaving.)
09:10:24  * mikealjoined
09:17:12  * dshaw_1quit (Quit: Leaving.)
10:24:43  <CIA-99>node: koichik v0.6 * r7f93861 / test/simple/test-tls-over-http-tunnel.js : test: fix test-tls-over-http-tunnel with v0.7 - http://git.io/I4wpLw
10:24:43  <CIA-99>node: koichik v0.6 * re0ab4ec / (lib/net.js test/simple/test-net-write-connect-write.js):
10:24:43  <CIA-99>node: net: fix race write() before and after connect()
10:24:43  <CIA-99>node: Fixes #2827. - http://git.io/7y_a_w
10:30:32  <CIA-99>node: koichik v0.6 * r6343179 / (lib/net.js test/simple/test-net-write-connect-write.js):
10:30:32  <CIA-99>node: net: fix race write() before and after connect()
10:30:32  <CIA-99>node: Fixes #2827. - http://git.io/WL-fVQ
10:35:59  * travis-cijoined
10:35:59  <travis-ci>[travis-ci] joyent/node#516 (v0.6 - e0ab4ec : koichik): The build passed.
10:35:59  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/692bcbe...e0ab4ec
10:35:59  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/757742
10:35:59  * travis-cipart
10:38:46  * travis-cijoined
10:38:46  <travis-ci>[travis-ci] joyent/node#517 (v0.6 - 6343179 : koichik): The build passed.
10:38:46  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/e0ab4ec...6343179
10:38:46  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/757753
10:38:46  * travis-cipart
11:03:02  <CIA-99>node: koichik master * rcd5d247 / (lib/net.js test/simple/test-net-write-connect-write.js):
11:03:02  <CIA-99>node: net: fix race write() before and after connect()
11:03:02  <CIA-99>node: Fixes #2827. - http://git.io/ieWp9g
11:17:33  * travis-cijoined
11:17:33  <travis-ci>[travis-ci] joyent/node#518 (master - cd5d247 : koichik): The build is still failing.
11:17:33  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/44daa98...cd5d247
11:17:33  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/757846
11:17:33  * travis-cipart
13:20:34  * piscisaureus_joined
13:33:56  * bnoordhuisjoined
13:34:29  * txdvjoined
14:02:00  * pfox___joined
14:03:28  * AvianusAsleepuschanged nick to AvianFlu
14:18:31  <bnoordhuis>AvianFlu: what was that stdio issue you mentioned yesterday?
14:18:49  <bnoordhuis>i'm looking at https://github.com/AvianFlu/libuv/commit/be1cd35 atm
14:19:51  <AvianFlu>bnoordhuis, without some kind of redirection in the last child, the stdio from that child comes through to the shell you're using
14:19:54  <AvianFlu>and doesn't detach right
14:19:59  <AvianFlu>pretty nasty, really
14:21:06  <AvianFlu>the other question here is, would it be a good idea to just go all the way and add configurable file redirection there?
14:21:31  <bnoordhuis>hm, maybe
14:21:50  <bnoordhuis>i'll revert the commits for now
14:21:53  <bnoordhuis>let's try again
14:29:57  <CIA-99>libuv: Ben Noordhuis master * rec0eff9 / (6 files in 4 dirs): (log message trimmed)
14:29:57  <CIA-99>libuv: Revert b3e0ad4, 149d32c, e99fdf0 and ea9baef.
14:29:57  <CIA-99>libuv: Detaching doesn't work yet, the setsid() call fails and leaves the child process
14:29:57  <CIA-99>libuv: attached to the parent's session.
14:29:57  <CIA-99>libuv: Revert "test: Add test case for spawning detached child processes."
14:29:58  <CIA-99>libuv: Revert "win: Implement options.detached for uv_spawn() for Windows."
14:29:59  <CIA-99>libuv: Revert "unix: Implement options.detached for uv_spawn() for unix."
14:31:24  <AvianFlu>bnoordhuis, we also saw an issue, after compiling node with that libuv branch, where timers on os x wouldn't keep the event loop up after the child detached
14:31:37  <AvianFlu>just on os x, from the look of it
14:31:49  * travis-cijoined
14:31:49  <travis-ci>[travis-ci] joyent/libuv#108 (master - ec0eff9 : Ben Noordhuis): The build is still failing.
14:31:49  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/57c5fa1...ec0eff9
14:31:49  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/758664
14:31:49  * travis-cipart
14:31:54  <bnoordhuis>AvianFlu: that's a libev bug
14:31:58  <bnoordhuis>it also happens on linux btw
14:32:12  <bnoordhuis>there's an issue of it somewhere...
14:32:13  <AvianFlu>I was seeing it work right locally, for whatever reason
14:32:36  <bnoordhuis>https://github.com/joyent/node/issues/2515 <- that's the one
14:32:59  <bnoordhuis>that was the day i lost faith in libev :-/
14:33:51  <AvianFlu>yeah, indutny mentioned this yesterday
14:33:57  <AvianFlu>although without the issue link
14:34:09  <AvianFlu>so, is there anything to be done? or is that just going to happen with timers sometimes?
14:35:33  <bnoordhuis>well, i have this fix that seems to work
14:35:43  <bnoordhuis>but i'm kind of afraid of merging it
14:35:48  <AvianFlu>also, bnoordhuis, after realizing I needed the /dev/null redirect, I made this branch too https://github.com/AvianFlu/libuv/compare/child-stdio-redirect
14:36:02  <AvianFlu>on unix, that has exposed stdout_file and stderr_file options
14:36:07  <AvianFlu>windows, I can get to
14:36:14  <AvianFlu>if you think it's a sane way to go with it
14:36:27  <AvianFlu>hard-coding to /dev/null would mimic the existing windows behavior from the old patch
14:37:37  <bnoordhuis>let me discuss it with piscisaureus_
14:39:26  <AvianFlu>sure
15:23:53  * AvianFluquit (*.net *.split)
15:24:39  * AvianFlujoined
15:54:15  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/libuv/compare/timers-fix <- review / mind if i land that?
16:06:55  * AndreasMadsenjoined
16:07:56  <mmalecki[zzz]>bnoordhuis: hi
16:07:59  * mmalecki[zzz]changed nick to mmalecki
16:08:30  <bnoordhuis>mmalecki: ho
16:08:32  <mmalecki>bnoordhuis: oh damn, some nice code you have there
16:08:37  <AndreasMadsen>yo
16:08:42  <bnoordhuis>that's what she said
16:08:48  <bnoordhuis>she didn't use the word 'code' though
16:10:50  <AndreasMadsen>^ but Ben interpreted it like that
16:11:16  <mmalecki>bnoordhuis: you're dating wrong girls, obviously
16:11:43  <bnoordhuis>oh, i don't know
16:11:56  <bnoordhuis>any girl that doesn't know the word 'no' is great in my experience
16:12:20  <mmalecki>it's hard to say "no" when you're gagged, you know
16:12:51  <mmalecki>or dead.
16:13:00  <AndreasMadsen>it's all about culture
16:13:16  <AndreasMadsen>and law
16:14:02  <piscisaureus_>bnoordhuis: lgtm
16:14:05  <mmalecki>bnoordhuis: I like your creative use of macros :D
16:15:26  <piscisaureus_>I don't really like creativity in general
16:15:49  <piscisaureus_>I don't think you should need SET_FLAG but src/unix is ben's territory
16:16:13  <mmalecki>piscisaureus_: I'm talking about https://github.com/bnoordhuis/node-http-parser/blob/master/src/http-parser.cc
16:17:12  <bnoordhuis>mmalecki: i don't like typing
16:17:16  <mmalecki>bnoordhuis: y u no publish to npm>
16:17:25  <bnoordhuis>mmalecki: it only works with master right now
16:17:30  <bnoordhuis>but i'll publish it eventually
16:17:50  <mmalecki>fair enough
16:17:53  <bnoordhuis>piscisaureus_: i was in dubio about SET_FLAG but it saves a few lines of code
16:19:36  * AndreasMadsenquit (Remote host closed the connection)
16:20:21  <bnoordhuis>eh, maybe i'll strip it
16:26:53  <CIA-99>libuv: Ben Noordhuis master * r2f886c8 / (src/unix/core.c src/unix/ev/ev.c src/unix/internal.h):
16:26:53  <CIA-99>libuv: unix: don't rely on libev to track timer state
16:26:53  <CIA-99>libuv: An obscure libev bug sometimes makes it miss state changes. Keep track of the
16:26:53  <CIA-99>libuv: state ourselves.
16:26:53  <CIA-99>libuv: Fixes joyent/node#2515. - http://git.io/vsdfrQ
16:28:51  <CIA-99>node: Shigeki Ohtsu master * r18acdff / vcbuild.bat : build: fix the help output of vcbuild.bat - http://git.io/EbXC2g
16:29:31  <piscisaureus_>bnoordhuis: how do I detect gcc? __GCC__ ? __GNUC__ ?
16:29:43  <bnoordhuis>piscisaureus_: the latter one
16:31:55  <mmalecki>bnoordhuis: can't you fix libev?
16:34:00  <piscisaureus_>yay
16:34:13  <piscisaureus_>my string stuff is now finally almost always faster than v8 :-)
16:35:51  <indutny>haha
16:35:54  <indutny>piscisaureus_: congrats
16:38:18  <bnoordhuis>mmalecki: well... yes
16:38:30  <bnoordhuis>but i've decided to move away from libev where possible
16:40:04  <bnoordhuis>mmalecki: specifically, libev gets in the way of optimizations we can do on linux
16:40:24  <bnoordhuis>so i've been thinking about writing a custom backend for linux and keeping libev as a backup for the other unices
16:41:12  <mmalecki>bnoordhuis: rewrite ALL the things
16:41:22  <bnoordhuis>why?
16:41:34  <mmalecki>bnoordhuis: dunno, it's fun.
16:41:56  <bnoordhuis>i've better uses of my time :)
16:42:25  <bnoordhuis>also, don't tell anyone i said this - but i don't really care about solaris or darwin
16:42:34  <bnoordhuis>all the world is linux
16:42:39  <bnoordhuis>or the part that matters anyway
16:43:01  <mmalecki>bnoordhuis: agreed
16:43:01  <bnoordhuis>the only reason to care about solaris is that joyent uses it
16:43:36  <mmalecki>bnoordhuis: solaris is for hipsters anyway
16:43:58  <bnoordhuis>hah, i wouldn't use the word 'hipsters'
16:44:17  <tjfontaine>solaris is for those who got EOLd from HP-UX and AIX
16:44:21  * travis-cijoined
16:44:21  <travis-ci>[travis-ci] joyent/node#519 (master - 18acdff : Shigeki Ohtsu): The build is still failing.
16:44:21  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/cd5d247...18acdff
16:44:21  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/759391
16:44:21  * travis-cipart
16:53:32  * brsonjoined
16:54:18  * sh1mmerjoined
16:55:35  * perezdjoined
16:59:45  * travis-cijoined
16:59:45  <travis-ci>[travis-ci] joyent/libuv#109 (master - 2f886c8 : Ben Noordhuis): The build is still failing.
16:59:45  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/ec0eff9...2f886c8
16:59:45  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/759495
16:59:45  * travis-cipart
17:05:09  <piscisaureus_>bnoordhuis: fix that thing :-)
17:05:32  <bnoordhuis>piscisaureus_: it's finicky, that macbook
17:06:30  * piscisaureus_quit (Read error: Connection reset by peer)
17:06:33  * piscisaureus__joined
17:07:56  * stephankjoined
17:08:03  <bnoordhuis>piscisaureus__: seen this? https://github.com/joyent/node/issues/2839
17:08:20  <bnoordhuis>and this: https://github.com/joyent/node/issues/2837
17:08:30  <bnoordhuis>by the same guy
17:09:12  <bnoordhuis>piscisaureus__: also this: https://github.com/joyent/libuv/pull/325
17:09:22  <bnoordhuis>we should fix uv_dlopen, it's annoying
17:13:59  <CIA-99>node: Ben Noordhuis master * raf7960b / (24 files in 9 dirs): uv: upgrade to 2f886c8 - http://git.io/1ej-Dg
17:14:56  * TooTallNatejoined
17:15:24  * isaacsjoined
17:15:30  <AvianFlu>bnoordhuis, I've got a clean branch now that detaches properly, if you guys decide you don't want the file redirection
17:15:42  * pieternjoined
17:17:17  <bnoordhuis>AvianFlu: color me interested. can you submit it as a PR?
17:17:23  <AvianFlu>sure
17:17:24  <isaacs>AvianFlu: file redirection should be done separately from detaching.
17:17:46  <isaacs>AvianFlu: "daemonizing" is a uniquely unixy thing, which is made mostly of other cross-platform bits we can provide. let's do that first.
17:17:55  <AvianFlu>isaacs, sounds good
17:23:14  <isaacs>bnoordhuis: hey, https://github.com/joyent/node/commit/8ad75f9 looks good to me, but doesn't seem to merge cleanly.
17:23:38  <bnoordhuis>isaacs: well, it's over 8 months old
17:23:46  <isaacs>bnoordhuis: i'm willing to port the concepts if you still stand behind the idea.
17:23:52  <isaacs>bnoordhuis: we should behave like the browser here.
17:24:20  <isaacs>i mean, it's not a terribly conflicting, just a few hiccups
17:24:29  <bnoordhuis>i think the concept is still sound
17:25:34  <isaacs>k, cool
17:25:48  <isaacs>i'll migrate it to master.
17:27:15  <AvianFlu>bnoordhuis, https://github.com/joyent/libuv/pull/329
17:28:19  * dapjoined
17:28:55  <pfox___>bnoordhuis: so, about #328
17:29:09  <pfox___>what do you mean by "the send happens almost immediately" .. you mean prior to uv_run() being called?
17:29:32  <bnoordhuis>AvianFlu: will review it tonight or tomorrow
17:29:45  <bnoordhuis>pfox___: yes, it's a synchronous write to a pipe
17:30:25  * travis-cijoined
17:30:25  <travis-ci>[travis-ci] joyent/node#520 (master - af7960b : Ben Noordhuis): The build is still failing.
17:30:25  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/18acdff...af7960b
17:30:25  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/759737
17:30:25  * travis-cipart
17:30:30  * brsonquit (Ping timeout: 252 seconds)
17:31:01  <pfox___>so the callback is going to be executed in the thread that calls uv_async_send() ?
17:31:29  <bnoordhuis>pfox___: no
17:32:23  <bnoordhuis>pfox___: it's the write that's immediate, that's not related to what the other end (the other thread) does with it
17:32:29  <pfox___>gotcha
17:32:54  <pfox___>so the issue is in the impl details of win v. unix
17:32:58  <bnoordhuis>yes
17:33:03  <pfox___>because in unix you'll read from the pipe before trying to enter th eloop
17:33:11  <pfox___>while windows relies on entering the loop to find pending sends
17:33:20  <bnoordhuis>more or less, yes
17:33:30  <pfox___>yeah. more or less :)
17:33:41  <pfox___>forgive my ignorance.. id love to learn more of the impl.
17:34:09  <pfox___>although i imagine ill be permanently scarred by the experience
17:34:10  <bnoordhuis>what happens on unix is that we always run an iteration of the event loop
17:34:20  <bnoordhuis>so events get processed, callbacks get fired, etc.
17:34:42  <pfox___>yeah. so i talked with piscisaureus__ about it yesterday.. and his recommend change was to go through the event loop once
17:34:43  <bnoordhuis>windows has a check that's along the lines of `if (loop->num_handles == 0) return`
17:34:49  <pfox___>and it didn't change the outcome of my test case
17:35:07  <pfox___>so there may be more to it. or im dumb. or both.
17:35:38  <bnoordhuis>i think what we need to do is that both unix and windows keep the event loop artificially alive for one more tick after the call to uv_async_send()
17:35:56  <bnoordhuis>we could pick the other option too, bail out if no reffed handles
17:36:04  <pfox___>i reiterate my statement that i think this is a valid and useful use case for anyone who is trying to insulate libuv from multithread madness
17:36:06  <bnoordhuis>it doesn't really matter as long as it's consistent and convenient
17:36:32  <bnoordhuis>right, that's why doing one more tick seems like the best solution to me
17:36:50  <bnoordhuis>tick == iteration == tick
17:55:50  * mikealquit (Quit: Leaving.)
17:58:35  * mikealjoined
17:58:53  * mikealquit (Read error: Connection reset by peer)
18:01:02  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
18:06:49  * mikealjoined
18:11:26  <igorzi>mikeal: looks like koichi is ok with moving this tunnel code to request (https://github.com/koichik/node-tunnel/issues/1#issuecomment-4170690).. are you planning get that in?
18:11:54  <mikeal>i saw
18:12:00  <mikeal>yeah, when i get some time
18:30:44  * TooTallNatejoined
18:31:10  * AvianFluquit (Quit: Leaving)
18:31:12  * brsonjoined
18:45:32  * mikealquit (Quit: Leaving.)
18:48:20  * brsonquit (Ping timeout: 272 seconds)
18:48:32  * brsonjoined
18:52:27  <isaacs>piscisaureus__: bnoordhuis: ping. call in 10
18:53:48  <mmalecki>isaacs: oh hey, you pinged me
18:54:55  <piscisaureus__>isaacs: yup
18:55:44  <isaacs>mmalecki: yeah, was looking into the argv0 stuff. fixed it on master.
18:55:47  * mikealjoined
18:55:51  <isaacs>mmalecki: the test was verifying incorrect behavior.
18:56:20  <mmalecki>isaacs: yeah, Avian pinged me about this tomorrow, I didn't have time to fix it, sorry
18:56:34  <mmalecki>I probably shouldn't, like, try to fix stuff at 4 am
18:56:48  <isaacs>mmalecki: hehe
18:56:53  <mmalecki>s/tomorrow/yesterday/
18:57:01  <isaacs>mmalecki: it's ok, i just wanted to see if there was some reason behind it.
18:57:28  <mmalecki>so do I!
18:58:26  * xaqjoined
19:02:58  <piscisaureus__>bnoordhuis: igorzi: skype call
19:19:16  * mikealquit (Quit: Leaving.)
19:23:38  * xaqquit (Remote host closed the connection)
19:30:37  * dshaw_joined
19:42:07  * mikealjoined
19:45:30  * mikealquit (Client Quit)
19:45:41  * AvianFlujoined
20:01:06  * brsonquit (Ping timeout: 255 seconds)
20:06:13  * txdvquit (Ping timeout: 246 seconds)
20:17:14  * brsonjoined
20:41:11  * isaacsquit (Remote host closed the connection)
21:12:35  * isaacsjoined
21:16:45  * saghulquit (*.net *.split)
21:16:46  * stephankquit (*.net *.split)
21:16:47  * chrisinajarquit (*.net *.split)
21:16:47  * creationixquit (*.net *.split)
21:16:47  * perezdquit (*.net *.split)
21:16:48  * pquernaquit (*.net *.split)
21:16:48  * indutnyquit (*.net *.split)
21:16:49  * dapquit (*.net *.split)
21:16:49  * sh1mmerquit (*.net *.split)
21:16:50  * piscisaureus__quit (*.net *.split)
21:16:50  * DrPizzaquit (*.net *.split)
21:16:51  * jcequit (*.net *.split)
21:16:51  * russell_hquit (*.net *.split)
21:16:51  * CIA-99quit (*.net *.split)
21:16:51  * tjfontainequit (*.net *.split)
21:16:51  * pfox___quit (*.net *.split)
21:16:52  * dshaw_quit (*.net *.split)
21:16:53  * leifquit (*.net *.split)
21:16:54  * sj26quit (*.net *.split)
21:16:54  * avsejquit (*.net *.split)
21:16:54  * mnemocquit (*.net *.split)
21:16:55  * elijah-mbpquit (*.net *.split)
21:16:56  * rphillipsquit (*.net *.split)
21:16:56  * toothrquit (*.net *.split)
21:16:58  * mmaleckiquit (*.net *.split)
21:18:53  * dshaw_joined
21:18:53  * dapjoined
21:18:53  * stephankjoined
21:18:53  * piscisaureus__joined
21:18:53  * perezdjoined
21:18:53  * sh1mmerjoined
21:18:53  * pfox___joined
21:18:53  * toothrjoined
21:18:53  * sj26joined
21:18:53  * chrisinajarjoined
21:18:53  * leifjoined
21:18:53  * mmaleckijoined
21:18:53  * tjfontainejoined
21:18:53  * CIA-99joined
21:18:53  * russell_hjoined
21:18:53  * jcejoined
21:18:53  * DrPizzajoined
21:18:53  * elijah-mbpjoined
21:18:53  * avsejjoined
21:18:53  * saghuljoined
21:18:53  * creationixjoined
21:18:53  * pquernajoined
21:18:53  * indutnyjoined
21:18:53  * mnemocjoined
21:18:53  * rphillipsjoined
21:21:15  * igorziquit (Ping timeout: 245 seconds)
21:32:53  * mralephjoined
21:36:04  * igorzijoined
21:42:53  * TooTallNatequit (Quit: Leaving...)
21:44:13  * AvianFluchanged nick to AF|LUNCHEMUNCHEN
21:50:58  * mikealjoined
21:52:12  <piscisaureus__>isaacs: bnoordhuis: mraleph: https://gist.github.com/1935258
21:52:52  * pieternquit (Quit: pietern)
21:54:01  <isaacs>piscisaureus__: reading now. <3
21:54:02  <isaacs>thanks!!
21:54:41  <isaacs>piscisaureus__: ansi - string contains only ascii characters <-- should be "ascii"?
21:54:50  <isaacs>s/ansi/ascii/
21:54:51  <piscisaureus__>isaacs: yes, probably
21:55:03  <mraleph>piscisaureus__: a factor of two improvement over hinted version. nice.
21:56:07  <piscisaureus__>mraleph: unfortunately WriteToFlat is unbeatable for strings that really contain only ascii. So we would have to retain the old iterator for those cases
21:56:22  <mraleph>haha
21:56:24  <mraleph>complexity
21:57:06  <piscisaureus__>mraleph: the reason is obviously that for ansi-only strings WriteToFlat can iterate over the cons pieces out of order. My iterator always iterates front to back.
21:57:17  <piscisaureus__>which is also the reason that it can be much faster for non-ascii strings
21:59:25  <isaacs>piscisaureus__: ansi or ascii? i'm so confused.
21:59:36  <piscisaureus__>I always mean ascii when I say ansi
21:59:37  <piscisaureus__>sorry
21:59:39  <isaacs>i didn't think that v8 had any special anything for ansi strings.
21:59:46  <isaacs>ok. i'll try to keep remembering that :)
21:59:52  <isaacs>it's just that ansi IS a character encoding
22:00:01  <isaacs>but a slightly different one, which also is a superset of ascii
22:00:43  <isaacs>like, † is 0x86 in ansi, but U+2020 in unicode, and nothing in ascii
22:00:59  <piscisaureus__>mraleph: if you feel like shooting holes in my benchmark: https://github.com/piscisaureus/node/commit/v8string
22:01:08  * pieternjoined
22:04:16  <mraleph>piscisaureus__: I am feeling like falling face first onto my table
22:04:50  <piscisaureus__>mraleph: why?
22:05:10  <mraleph>did not have much sleep :-0
22:05:34  <piscisaureus__>mraleph: aha. go get it then. Github will live tomorrow
22:05:43  <pfox___>piscisaureus__: hey, i hate to pester you on #328 , again.
22:05:51  * perezdquit (Quit: perezd)
22:06:08  <pfox___>i tried making the suggested change in the windows core.c you suggested .. change the while loop to a do..while
22:06:13  <piscisaureus__>and?
22:06:14  <pfox___>and it didn't change the outcome of my tests
22:06:21  <piscisaureus__>hmm
22:07:20  <piscisaureus__>pfox___: what if you remove https://github.com/joyent/libuv/blob/master/src/win/core.c#L230 ?
22:09:26  <piscisaureus__>pfox___: (that is dangerous territory btw)
22:09:44  * indutnychanged nick to indutny_sleeping
22:11:56  * TooTallNatejoined
22:16:32  * AF|LUNCHEMUNCHENchanged nick to AvianFlu
22:22:36  <pfox___>piscisaureus__: in addition to the previous change.. or instead of?
22:24:22  <pfox___>hm. no change in either case.
22:29:22  * felixgequit (Quit: felixge)
22:44:05  <piscisaureus__>hmm
22:44:05  <piscisaureus__>suck
22:56:10  * piscisaureus__quit (Quit: ~ Trillian Astra - www.trillian.im ~)
22:59:59  * piscisaureus_joined
23:04:31  * piscisaureus_quit (Ping timeout: 244 seconds)
23:58:20  * xaqjoined
23:59:21  * mmaleckichanged nick to mmalecki[zzz]