00:00:01  * ircretaryquit (Remote host closed the connection)
00:00:09  * ircretaryjoined
00:00:44  * toothrchanged nick to toothrot
00:01:22  * jmar777joined
00:02:00  * trevnorris&
00:02:00  <LOUDBOT>FAT BALLS AND WANGS IN YOUR ASS
00:02:15  <trevnorris>yeah...
00:02:58  <wolfeidau>lol
00:03:20  <wolfeidau>Someone has tainted LOUDBOT
00:07:02  <othiym23>is that even possible?
00:07:12  <othiym23>I thought LOUDBOT was the prime source of LOUDBOT
00:07:23  <tjfontaine>ALL HAIL THE CHURCH OF LOUDBOT
00:07:24  <LOUDBOT>WELL SOME COMPUTER SCIENTIST I AM
00:26:15  * TooTallNatequit (Quit: Computer has gone to sleep.)
00:27:58  * jmar777quit (Remote host closed the connection)
00:28:34  * bradleymeckjoined
00:32:44  * jmar777joined
00:34:15  <groundwater>trevnorris: what are the other 2 hottest code paths?
00:34:36  * bradleymeckquit (Read error: Connection reset by peer)
00:39:41  * philipsquit (Quit: http://ifup.org)
00:42:01  * philipsjoined
00:43:05  * philipsquit (Client Quit)
00:45:11  * philipsjoined
00:49:57  * TooTallNatejoined
00:58:35  * defunctzombiechanged nick to defunctzombie_zz
01:01:04  * brson_joined
01:01:18  * brsonquit (Read error: Connection reset by peer)
01:03:22  * EhevuTovquit (Quit: This computer has gone to sleep)
01:07:49  * TooTallNatequit (Ping timeout: 248 seconds)
01:17:41  * dapquit (Quit: Leaving.)
01:18:14  * jmar777quit (Read error: Connection reset by peer)
01:18:48  * jmar777joined
01:19:16  * dapjoined
01:23:09  * TooTallNatejoined
01:25:10  * st_lukequit (Remote host closed the connection)
01:31:28  * TooTallNatequit (Quit: Computer has gone to sleep.)
01:34:31  * inolenquit (Quit: Leaving.)
01:35:55  * amartens1quit (Quit: Leaving.)
01:42:21  * mikealquit (Quit: Leaving.)
02:09:39  * defunctzombie_zzchanged nick to defunctzombie
02:19:40  <tjfontaine>I will collect and post the results
02:20:15  <tjfontaine>but the regression in cpu for vm is not from vm2, as vm1 takes the same amount of time post the fixes that were done for the memory
02:20:47  <tjfontaine>I can shave a certain amount of time off by removing the proxy global, but it's not necessarily enough to warrant the change
02:21:27  <tjfontaine>Domenic_: ^
02:36:19  * amartensjoined
02:39:13  * amartensquit (Client Quit)
02:40:54  * jmar777quit (Remote host closed the connection)
02:45:34  * EhevuTovjoined
02:54:16  * defunctzombiechanged nick to defunctzombie_zz
02:54:48  * defunctzombie_zzchanged nick to defunctzombie
02:58:57  * defunctzombiechanged nick to defunctzombie_zz
03:05:59  * dominictarrjoined
03:07:58  * mikealjoined
03:15:54  * mikealquit (Quit: Leaving.)
03:19:13  * mikealjoined
03:25:47  * dshaw_joined
03:34:37  * mikealquit (Quit: Leaving.)
03:40:16  * mikealjoined
03:47:47  * dshaw_quit (Quit: Leaving.)
03:51:44  * dominictarrquit (Quit: dominictarr)
03:55:11  * kazuponjoined
03:59:46  * groundwaterquit (Quit: groundwater)
04:04:53  * inolenjoined
04:12:23  * julianduquequit (Quit: leaving)
04:12:41  * TooTallNatejoined
04:18:34  * dshaw_joined
04:34:31  <othiym23>tjfontaine: is there a way to use your manta-fueled Nodeniscience(tm) to figure out at what point in the 0.11 dev process that regression crept in?
04:35:44  * c4miloquit (Remote host closed the connection)
04:35:49  * dshaw_quit (Ping timeout: 248 seconds)
04:36:15  * c4milojoined
04:37:27  <tjfontaine>othiym23: funny you should mention
04:37:33  <tjfontaine>https://us-east.manta.joyent.com/NodeCore/public/builds/node/vm2times
04:37:36  <tjfontaine>I'm about to post that
04:37:47  <tjfontaine>that's a single run against every commit
04:39:16  <tjfontaine>othiym23: between eec43351c44c0bec31a83e1a28be15e30722936a and 0181fee411e217236c4ec0bf22c61466df5a56b5 -- which I'm guessing is going to be ... the v8 upgrade
04:39:37  <indutny>tjfontaine: huh?
04:40:04  <tjfontaine>so the last good working commit that is sub 10 seconds, was eec43351c44c0bec31a83e1a28be15e30722936a
04:40:07  <tjfontaine>the next commit is
04:40:10  <tjfontaine>704fd8f3745527fc080f96e54e5ec1857c505399
04:40:15  <tjfontaine>v8: upgrade to v3.20.2
04:40:17  <tjfontaine>so
04:40:18  <tjfontaine>game
04:40:18  <tjfontaine>set
04:40:19  <tjfontaine>match
04:40:27  * c4miloquit (Ping timeout: 248 seconds)
04:43:56  <tjfontaine>anyway, I've been at work too long today, for a meetup
04:44:09  <othiym23>go home!
04:44:16  <othiym23>I'm working still
04:44:18  <othiym23>but at home
04:44:20  <tjfontaine>hehe
04:44:22  <othiym23>and at least I'm bitter about it
04:44:28  <tjfontaine>clearly :P
05:04:01  * kazuponquit (Read error: Connection reset by peer)
05:04:07  * kazuponjoined
05:08:23  * Chip_Zeroquit (Read error: Connection reset by peer)
05:19:03  * brson_quit (Quit: leaving)
05:22:09  * dapquit (Quit: Leaving.)
05:25:08  * andujoined
05:28:56  * andupart ("Leaving")
05:31:54  * toothrotquit (Ping timeout: 264 seconds)
05:35:24  * toothrjoined
05:54:19  * inolenquit (Quit: Leaving.)
05:55:26  * inolenjoined
06:41:51  <MI6>nodejs-v0.10-windows: #223 UNSTABLE windows-ia32 (7/600) windows-x64 (7/600) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/223/
06:46:14  * dominictarrjoined
06:52:49  * `3E|ZZZchanged nick to `3rdEden
07:08:21  * Chip_Zerojoined
07:11:50  * rjequit (Ping timeout: 240 seconds)
07:13:19  * rjejoined
07:15:25  * contrahaxquit (Ping timeout: 245 seconds)
07:16:02  * contrahaxjoined
07:17:02  * bajtosjoined
07:22:18  * EhevuTovquit (Quit: This computer has gone to sleep)
07:25:57  * dsantiagoquit (Ping timeout: 248 seconds)
07:31:14  * dsantiagojoined
07:36:39  * rendarjoined
07:39:03  * contrahaxquit (Ping timeout: 260 seconds)
07:44:41  * dominictarrquit (Quit: dominictarr)
07:59:52  * felixgejoined
08:01:29  * EhevuTovjoined
08:17:40  * tooxiejoined
08:22:11  * tooxiequit (Client Quit)
08:29:31  * hzjoined
09:14:37  * hzquit (Disconnected by services)
09:14:41  * hzjoined
09:19:36  * zotjoined
09:21:45  * hzquit (Disconnected by services)
09:21:49  * hzjoined
09:21:58  <zot>mornin' :) anybody seen/used/written a mock'd interface to any parts of libuv?
09:22:21  <zot>i'm working up a gmock'd suite to a library that uses libuv, and was about to write my own… but of course would rather borrow/steal :)
09:36:08  * bnoordhuisjoined
09:48:22  * bajtosquit (Quit: bajtos)
09:51:15  * bajtosjoined
10:02:07  * bnoordhuisquit (Ping timeout: 260 seconds)
10:03:22  * EhevuTovquit (Quit: This computer has gone to sleep)
10:18:40  * bnoordhuisjoined
10:21:13  * kazuponquit (Remote host closed the connection)
10:21:25  * Damn3dquit (Ping timeout: 260 seconds)
10:25:24  * piscisaureus_joined
10:26:01  <piscisaureus_>I just came to the shocking conclusion that my macbook adaptor is really broken
10:26:52  <piscisaureus_>saghul: re At 2013-09-23T22:33:11.701Z, in #libuv, saghul said: tell piscisaureus_ if uv_want_endgame is called for a handle, is the endgame run unfit uv__handle_close is called?
10:26:58  <piscisaureus_>saghul: what do you mean?
10:27:07  <piscisaureus_>if uv_want_endgame is called then
10:27:19  <piscisaureus_>... "soon" the endgame handler will run
10:27:48  <saghul>piscisaureus_ yeah, I finally figured. Thanks!
10:27:51  <piscisaureus_>what happens after that depends on the endgame. You can do anything in the endgame (think zombo.com), the handle doesn't actually need to close
10:28:48  * abraxasquit (Remote host closed the connection)
10:29:05  <piscisaureus_>bnoordhuis: saghul: I think starting a started handle should just be okay
10:30:02  <saghul>piscisaureus_ you mean we should stop and restart it?
10:30:10  <saghul>if its active, that is
10:30:36  <saghul>because, what should happen if you change the path of an fs_event or fs_poll handle?
10:30:50  <piscisaureus_>saghul: semantically, yes
10:31:00  <piscisaureus_>saghul: uv_poll and uv_signal handles do that as well
10:31:10  <bnoordhuis>piscisaureus_: disagree. what if new_filename != old_filename?
10:31:27  <bnoordhuis>like saghul said :)
10:31:48  <piscisaureus_>bnoordhuis: well, you can uv_signal_start with a different signo
10:31:50  <saghul>bnoordhuis we could *_stop() if uv__is_active(handle)...
10:31:50  <piscisaureus_>it will just switch
10:31:56  <piscisaureus_>to another signal
10:32:06  <piscisaureus_>same for uv_poll_start, it can switch to another event mask
10:32:34  <saghul>piscisaureus_ but you can also change the callback, I guess we'd need to completely stop the handle
10:32:55  <bnoordhuis>yes, it does - but that's not the best behavior imo, easy to ahide application bugs that way
10:33:01  <bnoordhuis>s/ahide/hide/
10:33:09  <bnoordhuis>or a-hidin'
10:33:29  <piscisaureus_>saghul: bnoordhuis: https://github.com/joyent/libuv/blob/581e8c92666b61af791115d9efd83207890dee73/src/win/signal.c#L271-L285
10:33:34  <bnoordhuis>i'm with saghul on this one
10:34:23  <piscisaureus_>I guess saghul has a vote now
10:34:25  <piscisaureus_>well then
10:34:30  <piscisaureus_>I'll go have a beer
10:34:41  <saghul>:-)
10:35:14  <piscisaureus_>bnoordhuis: when are you coming to 020 btw? Thursday?
10:35:42  <saghul>bnoordhuis piscisaureus_ either way, this expands to all handles, so I'd leave fs_events like fs_poll is and then make a change adjusting the behavior globally
10:36:12  * Kakerajoined
10:36:13  <piscisaureus_>please...
10:36:41  <piscisaureus_>esp. for poll stop-start is much less efficient then just start
10:36:50  <bnoordhuis>piscisaureus_: yep, that's fine
10:37:03  <bnoordhuis>saghul: okay, as long as you don't forget :)
10:37:15  <piscisaureus_>saghul: ... ^-- see my poll comment
10:37:23  <saghul>piscisaureus_ noted!
10:37:26  <piscisaureus_>I'd be against for reasons of efficiency
10:37:33  <bnoordhuis>oh, you're right - that's why we did it like that. i remember now
10:37:46  <bnoordhuis>you should've added comments, bertje :)
10:38:32  <piscisaureus_>I think "starting a started handle is semantically equivalent to stopping it first" is not bad
10:38:44  <piscisaureus_>it's easy to grok
10:39:01  <piscisaureus_>indeed, it might hide bugs but most of the time it probably doesn't
10:39:27  <piscisaureus_>I'm fine if you make fsevents fail
10:39:33  <saghul>piscisaureus_ agreed, if we make the code actually do that :-) right now fs_poll ignores the new arguments to start
10:39:34  <piscisaureus_>but please don't go change everything else
10:40:06  <saghul>piscisaureus_ ok
10:40:12  <piscisaureus_>saghul: yes. It's better to explicitly fail then to just not do what the user wants :)
10:40:22  <saghul>indeed
10:40:37  <piscisaureus_>bnoordhuis: so you're coming thursday? Cool.
10:41:37  <bnoordhuis>piscisaureus_: what time? i have to do some things in gouda in the morning
10:41:45  <piscisaureus_>bnoordhuis: whenever suits you
10:41:51  <bnoordhuis>but i can probably be there around two PM-ish
10:41:52  <piscisaureus_>bnoordhuis: I'm awake all day :)
10:42:32  <bnoordhuis>okay, cool. i'll phone you when i get on the train. that'll give you an hour to get up and about
10:42:55  <piscisaureus_>bnoordhuis: I get up at 7 these days
10:43:38  <piscisaureus_>Unless you're heading out at 5:30 there's little chance for you to beat me :)
10:44:24  <bnoordhuis>hah, okay
10:46:51  <MI6>nodejs-v0.10: #1497 UNSTABLE linux-ia32 (1/600) smartos-x64 (2/600) osx-x64 (1/600) http://jenkins.nodejs.org/job/nodejs-v0.10/1497/
10:47:00  <piscisaureus_>bnoordhuis: but yeah call me anyway, so you won't find me doing awkward stuff
10:49:09  * raglejoined
10:49:38  <piscisaureus_>ok heading out, bb later this afternoon
10:51:11  <indutny>TooTallNate: hey man
10:51:12  <indutny>yt?
10:51:30  <indutny>bnoordhuis: hi
10:51:38  <indutny>bnoordhuis: I was thinking about tls inside tls
10:51:54  <indutny>bnoordhuis: I'm afraid there's no way to to do it like in "old times"
10:51:54  * TooTallNatequit (Quit: Computer has gone to sleep.)
10:52:02  <indutny>bnoordhuis: semantics would be totally messed up
10:52:14  <indutny>bnoordhuis: i.e. we're gaining speed from stopping emitting 'data' events
10:52:33  <bnoordhuis>indutny: ho
10:52:51  <indutny>bnoordhuis: so I can create API to allow chaining of TLSWraps
10:53:02  <indutny>by registering OnRead handlers in C++
10:53:02  <bnoordhuis>what would it look like?
10:53:10  <indutny>and exposing Write
10:53:17  <indutny>and using it from TLSWrap
10:53:20  <indutny>instead of calling MakeCallback
10:53:22  <indutny>and uv_write()
10:53:27  <indutny>bnoordhuis: is it clear?
10:53:31  <bnoordhuis>i think so
10:53:36  <indutny>but the problem will remain :)
10:53:44  <indutny>original socket won't emit anything
10:53:52  <indutny>and neither will do tls level 1 socket
10:53:57  * piscisaureus_quit (Ping timeout: 248 seconds)
10:53:58  <indutny>only tls level 2 socket will work as real socket
10:54:10  <indutny>I've another idea
10:54:18  <indutny>its based on creating socket pair
10:54:22  <indutny>and piping stuff in C++
10:54:35  <indutny>so every TLSWrap will be actually based on a new fd
10:54:48  <bnoordhuis>sounds workable
10:54:50  <indutny>but that has its own consequences
10:54:51  <indutny>:)
10:55:01  <indutny>and still won't emit 'data' events
10:55:08  <bnoordhuis>do we have a test case of what nathan was complaining about?
10:55:09  <indutny>by default
10:55:19  <indutny>yep
10:55:20  <indutny>https://github.com/joyent/node/issues/6204#issuecomment-24260311
10:55:28  <indutny>sorry, wrong comment link
10:55:31  <indutny>its a bit above
10:55:38  <bnoordhuis>okay, thanks
10:55:59  <bnoordhuis>ah right, that one
10:56:45  <indutny>so… I'm not really sure what to do here
10:56:49  <indutny>and asking for advice :)
10:57:02  <indutny>because all solutions seems to be a bit improper to me
10:57:06  * bajtosquit (Quit: bajtos)
10:58:04  <bnoordhuis>yeah, i'm looking
10:58:44  <indutny>on other hand
10:58:53  <indutny>this could be accomplished using legacy API
10:58:58  <indutny>createCryptoStream or whatever
10:59:05  <indutny>ah
10:59:07  <indutny>createSecurePair
10:59:11  <bnoordhuis>yes
10:59:26  <indutny>but its marked as deprecated right now
10:59:47  <bnoordhuis>doesn't matter if we're falling back to that internally
10:59:55  <indutny>oh
10:59:57  <indutny>internally
11:00:03  <indutny>interesting
11:00:03  <bnoordhuis>what happens in node core, stays in node core
11:00:13  <bnoordhuis>to appropriate a saying
11:00:16  <indutny>:)
11:02:15  <indutny>bnoordhuis: sounds like a plan?
11:02:32  <bnoordhuis>yes, if you think you can get it to work
11:05:03  <indutny>I think so
11:17:30  * zotquit (Quit: Leaving.)
11:39:11  * bnoordhuisquit (Ping timeout: 245 seconds)
12:00:33  * zotjoined
12:42:39  * AvianFlujoined
12:46:12  * Guest21838quit (*.net *.split)
12:46:12  * joshthecoderquit (*.net *.split)
12:46:13  * swajquit (*.net *.split)
12:48:30  * joshthecoderjoined
12:49:28  * swajjoined
12:51:15  * sindresorhusjoined
13:09:59  * jmar777joined
13:28:25  * mikealquit (Quit: Leaving.)
13:33:05  * c4milojoined
13:33:09  * AvianFluquit (Remote host closed the connection)
13:39:52  * mikealjoined
13:46:54  * mikealquit (Quit: Leaving.)
13:47:09  * mikealjoined
13:59:58  * raglequit (Quit: Leaving)
14:09:57  * kazuponjoined
14:18:50  * c4miloquit (Remote host closed the connection)
14:21:30  * bajtosjoined
14:23:25  * AvianFlujoined
14:30:06  * abraxasjoined
14:32:46  * c4milojoined
14:34:35  * abraxasquit (Ping timeout: 248 seconds)
14:36:02  * austoquit (Quit: Leaving)
14:37:30  * c4miloquit (Ping timeout: 245 seconds)
14:51:28  * Damn3djoined
14:54:05  * bradleymeckjoined
14:57:29  * M28quit (Read error: Connection reset by peer)
14:58:18  * zot1joined
14:58:37  * M28joined
15:00:59  * zotquit (Ping timeout: 260 seconds)
15:06:50  * c4milojoined
15:12:18  <saghul>http://venturebeat.com/2013/09/18/can-this-startup-steal-node-from-joyent-vcs-bet-8m-on-it/
15:12:59  <saghul>haters, haters everywhere!
15:15:19  <rendar>i think 'hijack' is a very strong word
15:17:32  <MI6>nodejs-master: #573 UNSTABLE osx-x64 (1/641) smartos-ia32 (1/641) smartos-x64 (6/641) linux-ia32 (1/641) http://jenkins.nodejs.org/job/nodejs-master/573/
15:18:47  <rendar>"Sources close to the Node core told VentureBeat Noordhuis and Belder have been difficult to work with on the project and have actually slowed its momentum at various points." --- WHAT? unbeliveable.
15:19:11  <saghul>I kind of stopped reading there
15:19:37  * zotjoined
15:19:39  <rendar>i wanna go deeper
15:19:58  <rendar>i just wanna know where human stupidity ends
15:19:59  <rendar>:)
15:20:07  <saghul>nodeception
15:20:24  <zot>rendar: at the edge of the earth, right where you can jump off into the void
15:21:01  <rendar>zot: lol
15:21:04  * zot1quit (Ping timeout: 246 seconds)
15:32:33  * dominictarrjoined
15:33:43  * mikealquit (Quit: Leaving.)
15:33:56  * mikealjoined
15:34:20  * hzquit (Read error: Connection reset by peer)
15:34:43  * hzjoined
15:34:52  * mikealquit (Client Quit)
15:35:20  * bnoordhuisjoined
15:36:45  <indutny>bnoordhuis: hey ben
15:37:12  <indutny>bnoordhuis: may I ask you to review this please https://github.com/joyent/node/pull/6261
15:38:16  * piscisaureus_joined
15:39:39  <indutny>piscisaureus_: hey bert
15:40:07  <indutny>nice article on venturebeat about strongloop
15:40:19  <indutny>is it your official position, guys http://venturebeat.com/2013/09/18/can-this-startup-steal-node-from-joyent-vcs-bet-8m-on-it/ ?
15:41:50  <piscisaureus_>indutny: official position?
15:41:57  <indutny>I mean
15:42:18  <indutny>is venture capital expressing its own feelings
15:42:28  <indutny>or does it match SL's position about joyent and node
15:47:15  <bnoordhuis>venture capital? you mean venturebeat?
15:47:40  * kazuponquit (Remote host closed the connection)
15:48:05  <indutny>err
15:48:06  <indutny>yeah
15:48:09  <indutny>certainly
15:48:57  <bnoordhuis>it's venturebeat trying to stir up drama and draw traffic to their site
15:49:05  <bnoordhuis>that's journalism in 2013 for you
15:50:00  <indutny>ok then :)
15:50:26  <indutny>I was a bit angry after reading it
15:50:32  <indutny>call in 10 minutes?
15:51:03  <bnoordhuis>i'm game. hope my mic works this time
15:51:08  <bnoordhuis>what were you angry about btw?
15:52:42  <indutny>well, about SL contribution to node and joyent licensing
15:53:04  <indutny>I think Isaac's opinion is a bit orthodoxal
15:53:53  * hzquit (Remote host closed the connection)
15:54:16  * hzjoined
15:54:16  <piscisaureus_>I'll join
15:55:21  <bnoordhuis>indutny: orthodoxal?
15:55:27  <bnoordhuis>and do you mean isaac or issac? :)
15:55:42  <indutny>isaac
15:55:51  <indutny>roth
15:55:52  <indutny>:)
15:55:55  <indutny>ah
15:55:56  <indutny>he's issac
15:55:58  <indutny>god :)
15:55:59  <bnoordhuis>yep :)
15:56:03  <bnoordhuis>confusing, i know
15:58:20  <indutny>ok, anyway
15:58:27  <indutny>2 minutes until the call
15:59:51  <tjfontaine>call?
16:00:22  <bnoordhuis>yep
16:00:25  <bnoordhuis>hangout link?
16:01:38  <tjfontaine>https://plus.google.com/hangouts/_/calendar/aXNhYWNzY2hsdWV0ZXJAZ21haWwuY29t._8gr32d1p611k8b9o6cr48b9k6op3eba18cr32b9k88s3cgpi8d2k8dq560
16:01:42  <tjfontaine>not sure if that works for you or not
16:01:46  <tjfontaine>that's from my calendar item though
16:01:47  <bnoordhuis>let's find out
16:03:21  <bnoordhuis>indutny: c'mon
16:03:24  <bnoordhuis>piscisaureus_: ditto
16:03:31  <trevnorris>jumping on now
16:15:43  * dominictarrquit (Quit: dominictarr)
16:18:16  * dapjoined
16:20:21  * TooTallNatejoined
16:20:23  * TooTallNatequit (Client Quit)
16:24:11  * dshaw_joined
16:25:00  * defunctzombie_zzchanged nick to defunctzombie
16:28:44  <bnoordhuis>trevnorris: i can probably remove the ObjectWrap deps from core tonight
16:28:49  <bnoordhuis>it's not that much work really
16:29:02  <indutny>bnoordhuis: https://github.com/joyent/node/pull/6261
16:29:04  <indutny>;)
16:29:22  <trevnorris>bnoordhuis: you are my hero. in the case of crypto and tls I just don't know the code.
16:29:46  <indutny>noone knows
16:29:53  <indutny>:)
16:29:57  <indutny>crypto is a total mistery
16:30:02  <indutny>and C++ histery
16:30:09  <indutny>histeria*
16:30:47  <trevnorris>haha.
16:30:49  * mikealjoined
16:31:12  * TooTallNatejoined
16:31:44  <bnoordhuis>TooTallNate: too late
16:32:07  <TooTallNate>bnoordhuis: TooLateNate?
16:32:24  * creationixquit (Ping timeout: 260 seconds)
16:32:54  <bnoordhuis>good one, i'm going to remember that one :)
16:34:35  <bnoordhuis>indutny: socket.fd? is that what i think it is?
16:36:56  * defunctzombiechanged nick to defunctzombie_zz
16:39:20  <piscisaureus_>socket.fd should be a string
16:39:29  * creationixjoined
16:39:31  <piscisaureus_>which can be one of the following:
16:39:36  * mikealquit (Ping timeout: 240 seconds)
16:39:37  <piscisaureus_>"probably %d"
16:39:48  <piscisaureus_>"it's %d by virtue of being on unix"
16:39:56  <piscisaureus_>"thing of it as %d"
16:40:04  <piscisaureus_>s/thing/think/
16:40:30  * inolenquit (Quit: Leaving.)
16:43:22  <indutny>bnoordhuis: I don't know
16:43:27  <indutny>bnoordhuis: I just copied that code from v0.10
16:44:30  * amartensjoined
16:45:48  * hzquit (Ping timeout: 240 seconds)
16:46:57  <bnoordhuis>really?
16:47:04  * hzjoined
16:47:30  <MI6>joyent/node: Fedor Indutny master * 42acbf8 : tls: wrap tls inside tls using legacy API - http://git.io/DPvStA
16:48:13  <indutny>bnoordhuis: yeah :)
16:48:16  <bnoordhuis>indutny: isn't that a bit rushed? ^
16:48:24  <indutny>well, you LGTMed it
16:48:38  <indutny>and Nathan seems to be ok with it
16:48:39  <bnoordhuis>yes... and i wanted TooTallNate to take it for a test drive
16:48:45  <indutny>aaaaah
16:48:46  <TooTallNate>bnoordhuis: i just did
16:48:48  <indutny>really missed it
16:48:53  <TooTallNate>indutny: it's good :)
16:48:56  <indutny>ok
16:48:57  <bnoordhuis>okay then
16:48:57  <TooTallNate>indutny: thank you!
16:48:59  <indutny>:)
16:49:01  <indutny>TooTallNate: you're welcome
16:49:10  <indutny>terrible hacks wasn't really hard to implement
16:49:16  <TooTallNate>hahah
16:49:18  <indutny>and its part of my duty
16:54:09  * AvianFluquit (Remote host closed the connection)
16:55:09  * defunctzombie_zzchanged nick to defunctzombie
16:55:55  * defunctzombiechanged nick to defunctzombie_zz
16:56:34  <MI6>nodejs-master: #574 UNSTABLE smartos-ia32 (1/642) linux-x64 (1/642) smartos-x64 (6/642) http://jenkins.nodejs.org/job/nodejs-master/574/
17:01:01  * AvianFlujoined
17:02:36  * brsonjoined
17:05:54  * defunctzombie_zzchanged nick to defunctzombie
17:12:04  <trevnorris>eh, well poop. the IsOneByte() patch won't work. oh well.
17:14:21  * bentkusjoined
17:14:44  * octetcloudjoined
17:15:10  <bnoordhuis>anyone else hitting random v8 assertions after the upgrade to 3.20.17.11?
17:15:31  <octetcloud>?
17:15:34  <trevnorris>just rebased. i'll let you know if I do
17:15:46  <bnoordhuis>they all seem to be heap allocation / garbage collector related
17:15:58  <bnoordhuis>rolling back and running some tests now
17:16:15  <bnoordhuis>octetcloud: btw, hi :)
17:16:16  <bentkus>so
17:16:22  <bentkus>i see you come around to include ipv6
17:16:33  <bnoordhuis>ipv-what?
17:17:01  <MI6>nodejs-master-windows: #366 UNSTABLE windows-x64 (21/642) windows-ia32 (22/642) http://jenkins.nodejs.org/job/nodejs-master-windows/366/
17:17:02  <octetcloud>hi ben. that ? was me trying to get weechat help... sorry for the noise
17:17:04  <bnoordhuis>kidding aside, yes. better ipv6 support was never that controversial, just the timing wasn't right
17:17:53  <bentkus>have you thought about renamning IPV6ONLY to DUALSTACK?
17:17:55  <bnoordhuis>octetcloud: oh nice. i still use irssi myself, can't wean myself from it
17:18:12  <bnoordhuis>bentkus: tbh, no :)
17:18:13  * mikealjoined
17:18:37  <bentkus>bentkus: because now we have that one can only bind on dualstack(if it is supported) or only on ipv6
17:18:51  <bnoordhuis>bentkus: you were having that discussion with piscisaureus_ last time around, right?
17:19:03  * TooTallNatequit (Read error: Connection reset by peer)
17:19:10  <bentkus>yes
17:19:21  * TooTallNatejoined
17:20:21  <bnoordhuis>bentkus: then i'd suggest to pick it up with him again, i wasn't really involved last time iirc
17:21:31  <bentkus>actually only you commented on it
17:22:17  <bentkus>hm
17:22:31  <bentkus>you went with a single function and a pointer to the sockaddr struct
17:24:39  <bnoordhuis>yep
17:26:15  <bnoordhuis>rolling back to 3.20.17.7 appears to fix the issues
17:26:45  <bnoordhuis>looking at the diff between .7 and .11, there's enough heap / gc changes that might be the culprit
17:27:15  <bnoordhuis>indutny, tjfontaine, trevnorris: any objections against reverting to .7?
17:27:39  <bnoordhuis>i'll reopen the uglifyjs issue
17:27:44  * AvianFluquit (Remote host closed the connection)
17:27:48  <trevnorris>bnoordhuis: not here.
17:27:52  * kenperkinsquit (Quit: Computer has gone to sleep.)
17:28:37  * kenperkinsjoined
17:33:00  * dominictarrjoined
17:33:53  * dshaw_quit (Quit: Leaving.)
17:38:34  <indutny>bnoordhuis: wut?
17:38:55  <indutny>bnoordhuis: lets backport this fix from v8 then
17:39:02  <indutny>bnoordhuis: if you really feel like reverting it
17:39:22  <bnoordhuis>indutny: do you know what commit the fix is?
17:39:48  <indutny>https://chromiumcodereview.appspot.com/23533012
17:39:50  <indutny>bnoordhuis: ^
17:39:52  <bnoordhuis>and yes, i really feel like reverting it - i always do when v8 starts aborting randomly :)
17:40:27  * inolenjoined
17:40:33  <bnoordhuis>ah, CanHoldValue() was one of the suspect changes i was looking at actually
17:41:46  <indutny>bnoordhuis: oook :)
17:43:13  * groundwaterjoined
17:44:52  * bajtosquit (Quit: bajtos)
17:46:35  * zotquit (Ping timeout: 248 seconds)
17:46:44  * bajtosjoined
17:47:02  <octetcloud>I'm having trouble finding jsstack (dtrace/mdb extension). can anyone point me to it? is it only solaris?
17:47:24  * AvianFlujoined
17:48:13  * bajtosquit (Client Quit)
17:48:15  * kazuponjoined
17:51:32  <bnoordhuis>octetcloud: it's solaris-only
17:51:43  <tjfontaine>the code is opensource though
17:52:04  <tjfontaine>https://github.com/joyent/node/blob/master/deps/mdb_v8/mdb_v8.c is the code for findjsobjects
17:53:06  <octetcloud>thanks tj. anybody tried to get this working on top of dtrace4linux or OS X's dtrace?
17:53:20  <tjfontaine>bnoordhuis: do you want to do this revert before we release another 0.11?
17:53:41  <tjfontaine>octetcloud: d4l probably doesn't have enough there yet for it to work, but oracle's dtrace port *might* work
17:54:28  <MI6>libuv-master: #256 UNSTABLE windows (3/195) smartos (2/194) linux (1/194) http://jenkins.nodejs.org/job/libuv-master/256/
17:54:47  <bnoordhuis>tjfontaine: i guess so. else we'll have a release that crashes randomly
17:56:04  <tjfontaine>bnoordhuis: ok I'll hold off on that
17:56:14  * kazuponquit (Ping timeout: 240 seconds)
17:57:40  <indutny>tjfontaine: btw, what do you think about "manta"ing `npm test` for top npm repos?
17:57:42  <indutny>tjfontaine: before 0.12 RC
17:57:50  <indutny>to see how much we're out of the form
17:58:06  <bnoordhuis>^ good idea
17:58:50  <MI6>joyent/node: Ben Noordhuis master * 14687eb : Revert "deps: update v8 to 3.20.17.11" - http://git.io/F8jf3g
17:59:40  <tjfontaine>indutny: I talked about it for a while, it's something I want to look into
17:59:49  <indutny>tjfontaine: thanks
18:01:16  <bnoordhuis>dinner time, biab
18:05:25  * bnoordhuisquit (Ping timeout: 245 seconds)
18:07:34  * dshaw_joined
18:08:28  * M28quit (Ping timeout: 245 seconds)
18:18:48  <MI6>libuv-node-integration: #238 UNSTABLE osx-x64 (1/642) linux-ia32 (1/642) smartos-x64 (11/642) http://jenkins.nodejs.org/job/libuv-node-integration/238/
18:18:52  <MI6>nodejs-master: #575 UNSTABLE osx-x64 (1/642) smartos-ia32 (2/642) smartos-x64 (5/642) http://jenkins.nodejs.org/job/nodejs-master/575/
18:19:08  <tjfontaine>ircretary: tell bnoordhuis have we reported or isolated a case directly of the instability?
18:19:08  <ircretary>tjfontaine: I'll be sure to tell bnoordhuis
18:21:03  * EhevuTovjoined
18:26:37  * dshaw_quit (Quit: Leaving.)
18:27:28  <MI6>nodejs-master-windows: #367 UNSTABLE windows-x64 (22/642) windows-ia32 (21/642) http://jenkins.nodejs.org/job/nodejs-master-windows/367/
18:28:38  <trevnorris>othiym23: ping
18:38:51  * dominictarrquit (Quit: dominictarr)
18:41:29  <othiym23>pong
18:41:40  <othiym23>trevnorris: ^^ :D
18:45:32  <trevnorris>othiym23: hey. just need someone to talk this through. hope you don't mind. so you can add a domain to an event emitter after the fact, which is a problem because then we still need to check for the domain object in MakeCallback. or else the async listeners won't be activated while the callback is run.
18:47:55  <trevnorris>I can't figure out a way to get around needing to re-add all the domain specific native implementation.
18:48:30  <MI6>nodejs-master-windows: #368 UNSTABLE windows-x64 (21/642) windows-ia32 (20/642) http://jenkins.nodejs.org/job/nodejs-master-windows/368/
18:48:35  <trevnorris>though like hell i'm going to do that.
18:54:37  <trevnorris>othiym23: so one idea i'm playing with is adding "addAsyncListener" to every *Wrap so you can add a set of callbacks to an existing ref.
18:55:16  <trevnorris>problem there is, say in the case of net, the _handle isn't setup until after .listen() is called, but you can call Domain#add() on the event of the server any time you want.
18:55:23  <othiym23>trevnorris: or you can keep a handle to a list of all the domains and all their members and reverse the relationship at the cost of some performance loss when using domains
18:56:09  <othiym23>you could probably do it with a WeakSet, or whatever its equivalent is in C++, if that exists
18:56:21  * mikealquit (Quit: Leaving.)
18:57:32  * dshaw_joined
18:57:53  <othiym23>that's O(n * m) lookup where n and m are the number of domains and the average number of members per domain
18:57:59  * mikealjoined
18:59:21  * dshaw_1joined
18:59:41  <trevnorris>othiym23: please explain more of the implementation
18:59:43  * mralephjoined
19:00:47  <othiym23>as you know, Trevoer, right now domain.add(ee) sets a property on the EE denoting to which domain it belongs, and also adds it to a list of members kept by the domain
19:01:05  <trevnorris>yeah
19:01:39  <othiym23>what I propose removes the property from ee, and ensures that there's a way the asyncListener code for domains can get at all those lists of domains to determine whether the EE belongs to a domain
19:01:47  * dshaw_quit (Ping timeout: 248 seconds)
19:02:03  * bradleymeckquit (Quit: bradleymeck)
19:02:10  <othiym23>it also complicates the implementation of domain.add, because there needs to be a uniqueness constraint that ensures ee belongs to at most one domain
19:03:12  <trevnorris>trust me, no way it could complicate things more than with my current attempts.
19:03:16  <othiym23>but it pulls that membership test completely out of band and all you need is a reference to an Object, regardless of its later mutation, to determine membership
19:03:25  * bradleymeckjoined
19:04:06  * dshaw_1quit (Ping timeout: 252 seconds)
19:04:17  <othiym23>if you have a WeakMap or a WeakSet, you can do that in such a way that ee can be GCd when it's otherwise out of scope without any explicit finalization on the part of the domain
19:04:58  <tjfontaine>that may not necessarily be feasible for a while though, right?
19:05:01  <othiym23>and depending on how those are implemented, you may get faster lookup than linear scan across an array
19:05:23  <othiym23>tjfontaine: the code's in V8 right now, behind a flag, which presumes Node can probably get at it from the C++ side, no?
19:05:27  <othiym23>(you may be right)
19:06:06  <trevnorris>WeakMap is a JS thing. no v8 native API for it yet
19:06:14  <tjfontaine>othiym23: well it's certainly possible, and so long as we only opt in when necessary, it's probably more doable, but it is a bit worrisome
19:06:25  <tjfontaine>trevnorris: we can always "eval" :P
19:06:31  <trevnorris>hahaha
19:07:01  <trevnorris>dude, i'm almost to the point of considering even that.
19:07:38  * AvianFluquit (Ping timeout: 245 seconds)
19:09:39  <othiym23>well, we could also create our own implementation of something similar, although I'll be honest that I don't totally understand all the ins and outs of the ephemeron mechanism it's based on
19:10:18  <trevnorris>ephemeron, cool word
19:10:20  <tjfontaine>ya, there are certainly ways we can do this, if necessary
19:10:47  <othiym23>without it, you'd need some kind of hook added to destroy or destroyNow to dispose of the reference the domain has to ee (which domain.dispose still does, altho I don't know how good core is about calling it)
19:11:31  <othiym23>i.e. a finalization step to ensure the domain members list is kept up to date
19:11:43  <trevnorris>issue is, domains work off event emitters. EE's are technically synchronous, but used for async event triggers. so domains had to be injected directly into the async callback mechanisms of things like MakeCallback to make up for the fact that EE's don't know about them.
19:11:47  * bnoordhuisjoined
19:12:12  * julianduquejoined
19:12:34  <trevnorris>and it doesn't help that the Wrap instances != domain instances != EE instances. so you have three objects all circularly attached.
19:12:34  <othiym23>yeah
19:13:00  <othiym23>it seems hard to do this without having at least ee.emit('error') being domains-aware
19:14:00  <othiym23>in retrospect, having a special 'error' event on all EEs was a victory of consistency over usability / extensibility
19:14:01  <trevnorris>lib/events.js must have domain specific code. that I now know for sure. but now figuring out how to keep domains out of everything else that's a pain.
19:14:43  <othiym23>where are the other places where you're finding domain code hard to remove, trevnorris?
19:14:55  <othiym23>maybe we should try to talk them out case by case and see if anything pops up
19:15:45  <trevnorris>othiym23: MakeCallback. right now you can add a "domain" object to any EE before the actual Wrap instantiation (usually attached via _handle).
19:16:01  <trevnorris>othiym23: so MakeCallback has to check at call time if the domain object is set on the request object.
19:16:12  <trevnorris>or it can't run enter/exit
19:16:23  * bnoordhuisquit (Ping timeout: 245 seconds)
19:16:47  <othiym23>trevnorris: _handle is on the JS side, right?
19:16:50  <trevnorris>yeah
19:17:23  <othiym23>so why wouldn't having that ee in the domain's members be workable?
19:17:55  <trevnorris>when would the domain's member be checked?
19:18:04  <othiym23>i.e. assume there's a call like Domain.get(ee).emit('error', err)
19:18:16  <othiym23>at throw time
19:18:37  <othiym23>"throw" -> error-raising
19:19:20  <trevnorris>but that would require a domain to be active, correct?
19:19:35  <trevnorris>or are you saying to keep a complete list of all EE's that have every received a domain and check those?
19:20:03  <othiym23>the latter, yes
19:20:06  * mikealquit (Quit: Leaving.)
19:20:29  * mikealjoined
19:20:53  <othiym23>domain.members is a list of every EE that has been marked as being on the domain, either implicitly (because it was created within the context of an active domain) or via domain.add(ee)
19:21:43  <trevnorris>so on error-raising we'd have to detect which EE threw, then emit error on that EE?
19:22:13  <othiym23>we'd need to at least know which EE threw, yeah
19:22:51  <othiym23>and then there would be a globalish map from ees to domains
19:23:20  * dshaw_joined
19:23:29  * kazuponjoined
19:23:32  <trevnorris>i'd be fine with something like that. since it'd move the performance hit out of the hot path.
19:24:48  * octetclo1joined
19:24:52  <trevnorris>how would we handle the gc issue?
19:24:59  * octetclo1quit (Client Quit)
19:25:27  * octetcloudquit (Quit: WeeChat 0.4.1)
19:25:35  * jmar777quit (Remote host closed the connection)
19:26:25  * octetcloudjoined
19:26:49  * octetcloudquit (Client Quit)
19:27:06  * jmar777joined
19:27:55  * kazuponquit (Ping timeout: 245 seconds)
19:27:57  * dshaw_quit (Ping timeout: 268 seconds)
19:28:43  <trevnorris>othiym23: i'm liking the idea, but it would botch support for process.domain, since we wouldn't be setting the active domain at MakeCallback time.
19:28:45  <othiym23>it's tough without guaranteed finalizers, which is why I suggested Weak{Map,Set}
19:29:08  <othiym23>hmmmm
19:29:40  <othiym23>yeah, I think I'm the preeminent advocate for the ongoing survival of process.domain as a useful construct, too
19:30:22  <trevnorris>well, i'm digging the idea. it would allow full separation of domains from the EE.
19:30:54  <trevnorris>since I'd only have to set a single error handler when domains are added that will always run and simply check if the ee exists in the set.
19:31:01  <trevnorris>that would make this much simpler.
19:31:36  <trevnorris>ok. not "full" separation. .domain would still have to be set.
19:31:47  <othiym23>this is a total swag / spitball, but if process.domain were a getter, and we had some way of testing which domain was active from context, that would be a way to preserve it
19:32:10  <othiym23>i.e. it doesn't matter if process.domain is slow because it would only be used outside of core
19:32:17  <trevnorris>well, it's currently a getter.
19:32:26  <othiym23>a more complicated getter ;)
19:32:32  <trevnorris>heh, yeah :)
19:33:41  * dshaw_joined
19:33:44  <othiym23>it seems to me that if you can figure out what the domain is at error time, you should be able to figure it out at other times, but that may be acting on assumptions built up by having enter / exit ingrained in my brain
19:34:12  * octetcloudjoined
19:36:50  <othiym23>so to do that, you need to know 1) if the current code is executing inside domain.run / domain.bind, or 2) if we're inside EE code where errors will go to a different domain
19:37:20  <othiym23>trevnorris: how would _processFatalException know which EE caused a throw?
19:37:45  * mikealquit (Quit: Leaving.)
19:38:12  <trevnorris>othiym23: no freakin idea. only thing I can think is to set, say, require('events').active == currently executing event emitter
19:38:22  <trevnorris>but the overhead there would be ridiculous
19:39:08  * mikealjoined
19:39:34  <othiym23>can you think of a reason why whatever mechanism that uses wouldn't be accessible to the process.domain getter?
19:41:46  <trevnorris>nope
19:42:24  <othiym23>trevnorris: could you maybe do something like the tickinfobox that has two indexed properties, 0 being whatever domain is currently running and 1 being the currently "active" EE?
19:43:55  <trevnorris>well, tickInfoBox shares numeric status flags with the native layer. and the current domain getter sets the domain in index 0 of an array that's been persisted in native layer so it has easy access in MakeCallback.
19:44:21  <trevnorris>so there's no savings using a technique like that in this case.
19:44:25  <othiym23>OK
19:44:37  <othiym23>dynamic languages are such a PITA
19:44:41  <trevnorris>seriously.
19:45:17  <othiym23>I just want a little singularity where we can poke stuff and get it back from both C++ and JS with constant-time access
19:45:45  <othiym23>constant-time fast access, before any weisenheimers butt in
19:46:01  <trevnorris>well, you can return non-weak objects from C++ in sub 10ns. if you know beforehand exactly what you're returning.
19:46:18  * dshaw_quit (Quit: Leaving.)
19:47:19  <trevnorris>but, doing something like map.add(ee, domain) then somehow getting the running ee when _fatalE.. runs. that would simplify some stuff.
19:47:37  <trevnorris>let me run a benchmark to see how much overhead it'd add to set the active ee.
19:50:26  <trevnorris>tjfontaine: would you be vehemently opposed to adding --harmony-collections to our startup? (if I had a really really good reason :)
19:51:56  * felixgequit (Quit: felixge)
19:52:43  <trevnorris>hey i'm getting the segfaults ben was talking about
19:52:48  <trevnorris>on v8::internal::SemiSpace::Swap
19:52:56  * julianduquequit (Quit: leaving)
19:55:52  * bnoordhuisjoined
19:56:06  * dshaw_joined
19:56:48  * bradleymeckquit (Quit: bradleymeck)
19:57:15  <trevnorris>groundwater: hey, so process.nextTick(), EventEmitter#emit() and Buffer() are the three hottest paths. it'll vary between the three depending on what you're doing, but it's always one of those.
19:58:12  * felixgejoined
19:58:12  * felixgequit (Changing host)
19:58:12  * felixgejoined
19:58:22  <othiym23>awesome, creationix's addAsyncListener polyfill only wraps two of those, we're doing great
19:58:38  * bradleymeckjoined
20:00:25  * AvianFlujoined
20:02:02  <bnoordhuis>tjfontaine: re v8 issue, i don't have a truly isolated test case yet but e.g. the test from #6214 triggers it
20:02:35  <bnoordhuis>tjfontaine: i was thinking of running a bisect over v8 commits but that will take a while
20:02:53  <trevnorris>bnoordhuis: was getting segfault when running benchmarks on v8::internal::SemiSpace::Swap
20:03:35  <bnoordhuis>trevnorris: with .7 or .11?
20:03:43  <trevnorris>.11
20:03:48  <trevnorris>rolling back didn't
20:03:54  <bnoordhuis>right. yeah, they're all over the place
20:04:13  <bnoordhuis>btw, what do you need WeakSet for?
20:04:34  <trevnorris>a way to get around domain injection into core.
20:05:16  <trevnorris>since EE's have a many to one relationship with domains it'd be easy to keep a WeakMap of them, then on fatal error check which domain error handler needs to run
20:05:17  <bnoordhuis>okay. but what do you need WeakSet for in particular? :)
20:05:33  <bnoordhuis>ah, like that
20:05:42  <trevnorris>yeah. otherwise we'll have gc issues.
20:05:49  <othiym23>bnoordhuis: WeakSets or reliable EE finalizers, pick one
20:06:24  <bnoordhuis>is this purely at the js level?
20:06:25  <trevnorris>advantage is it would push a lot of the performance impact to when an error actually occurs
20:06:36  <trevnorris>or when an event is being added to a domain.
20:06:51  <trevnorris>but not on the execution of the code itself. which is what i've been really trying to avoid.
20:08:20  <trevnorris>bnoordhuis: if it worked, would you be very against adding --harmony-collections flag?
20:09:03  <bnoordhuis>well... i'm not sure Weak{Map,Set} are all that fast to start with
20:09:24  <bnoordhuis>but let me ask again, is this purely at the js level or is c++ involved too?
20:09:42  <trevnorris>it's to prevent needing to check if the domain object is set in MakeCallback
20:10:12  <bnoordhuis>okay, so everything's in c++ land?
20:10:43  <trevnorris>well, the use of WeakMap would only be in lib/domain.js, but it's to prevent needing domain specific code in c++ land.
20:11:07  * piscisaureus_quit (Ping timeout: 248 seconds)
20:11:45  <bnoordhuis>hrm, the issue with --harmony-collections is of course that they add globals
20:12:15  <trevnorris>yeah, and unfortunately there's no c++ api for use of weakmaps
20:12:33  <bnoordhuis>i think you can access them through the regular Object::Get/Set/etc. methods
20:12:36  <bnoordhuis>but that aside
20:13:30  <bnoordhuis>oh, i may be wrong here
20:13:40  <tjfontaine>bnoordhuis: ok, I'll spend some time on it
20:14:14  <tjfontaine>trevnorris: enabling harmony collections is interesting, but needs to come with excellent data indicating it will be ok :)
20:14:48  <trevnorris>tjfontaine: whatever it takes to prevent needing to inject domain code back into core after taking 2 weeks working it out.
20:17:00  <trevnorris>bnoordhuis: maybe you can pull out some wisdom. since *Wrap is set as the _handle on the actual EE instance, domains have a ridiculous amount of flexibility that allow them to be added even before the instantiation of the Wrap itself.
20:17:58  <trevnorris>bnoordhuis: so replicating Domain#add() is so far impossible since MakeCallback will need to know that the domain property has been set at callback execution time.
20:18:14  <trevnorris>well, impossible w/o needing to re-add the domain specific code.
20:18:24  <bnoordhuis>ah, don't expect deep insights from me tonight. i can barely stay awake :)
20:18:35  <trevnorris>haha, ok :)
20:21:46  <trevnorris>othiym23: so getting the process.domain from the active EE is simple, but what if someone tries to set it?
20:22:12  <othiym23>trevnorris: they should be taken out and soundly mocked
20:22:33  <trevnorris>othiym23: ok. so setting process.domain just shouldn't be allowed?
20:22:36  <othiym23>it has never been my understanding, nor have I knowingly communicated to others, that that's a mutable property
20:22:43  <trevnorris>ok cool
20:22:46  <othiym23>yeah, it shouldn't have a setter
20:23:06  <trevnorris>coolio.
20:23:07  <othiym23>better yet, trying to mutate it should throw
20:23:22  <othiym23>but that's me being overly prescriptivist
20:23:27  <trevnorris>heh
20:24:14  * kazuponjoined
20:25:21  <trevnorris>othiym23: though, i'm having a hard time seeing how this new functionality will fit into the domain stack.
20:26:38  <othiym23>yeah hmm
20:27:02  <trevnorris>also, what will enter/exit do if the domain is always controlled by the active event emitter?
20:27:04  <othiym23>the idea of an "outer" domain exiting "inner" domains is pretty key to the semantics of domains
20:27:24  <othiym23>it's what gives them their qualities as an asynchronous try-catch
20:27:50  <trevnorris>which again is a sugar shim. ;P
20:28:39  <othiym23>if the current EE belongs to a domain, conceptually you're entering the domain before any methods on it are called
20:28:54  <trevnorris>yeah. so enter/exit would technically be noop's right?
20:28:54  * kazuponquit (Ping timeout: 264 seconds)
20:29:08  <othiym23>if it's an emit, and the handler is bound to another domain, then that domain gets entered before execution
20:29:36  <tjfontaine>TooTallNate: hey
20:29:47  <TooTallNate>tjfontaine: yo
20:30:10  <tjfontaine>TooTallNate: that control character PR, you were lgtm on the issue, but not on the PR, s'ok for me to land on v0.10?
20:30:23  <trevnorris>othiym23: what i'm doing is setting EventEmitter.active at the top of emit(), which allows the domain module to know what is the currently running event. so even if the other handle is run, then it'll still need to "emit" right?
20:30:32  <TooTallNate>tjfontaine: ya i think so, 'slong as the tests pass
20:30:44  <tjfontaine>TooTallNate: nod
20:31:58  <othiym23>trevnorris: parsing, plz hold
20:33:15  <othiym23>trevnorris: when you say "it'll still need to emit", do you mean "it" as the original EE?
20:33:24  <othiym23>I'm having a tough time figuring out what you meant there ;)
20:33:32  <MI6>joyent/node: Eric Schrock v0.10 * 35ae696 : readline: handle input starting with control chars - http://git.io/UugpWQ
20:33:33  * felixgequit (Quit: felixge)
20:33:43  <trevnorris>othiym23: no worries. so am I. :P
20:34:14  <trevnorris>othiym23: so take http for example. and the connection listener is bound to one domain, then you bind req/res to another domain, with me?
20:35:58  * jmar777quit (Remote host closed the connection)
20:38:29  <trevnorris>othiym23: so you technically have a "stack" right? say an event emitted on req throws, it's going to bubble up to the domain around the connection listener.
20:40:35  <othiym23>yeah, unless req is explicitly added to a domain
20:41:26  <MI6>nodejs-v0.10: #1498 UNSTABLE linux-ia32 (2/600) smartos-x64 (2/600) osx-ia32 (1/600) osx-x64 (1/600) http://jenkins.nodejs.org/job/nodejs-v0.10/1498/
20:41:33  <trevnorris>othiym23: you mean to another domain? does that mean it'll then leave the original domain stack?
20:41:53  <trevnorris>wait. so the domain stack is controlled by the order in which they're created?
20:41:56  <MI6>joyent/libuv: tjfontaine created tag v0.10.17 - http://git.io/SJvYCg
20:41:58  <MI6>joyent/libuv: Timothy J Fontaine v0.10 * e9df7cb : Now working on v0.10.18 (+1 more commits) - http://git.io/NhnFEg
20:42:10  * mikealquit (Quit: Leaving.)
20:43:17  <othiym23>no, the order in which they're entered
20:43:49  * mikealjoined
20:44:01  <othiym23>trevnorris: there's only one domain stack across all requests
20:44:15  <othiym23>trying to think of a metaphor that doesn't disappear up the ass of abstract CS
20:44:29  <MI6>libuv-v0.10: #115 UNSTABLE windows (4/188) smartos (2/187) http://jenkins.nodejs.org/job/libuv-v0.10/115/
20:44:37  <trevnorris>yeah. I get that.
20:44:50  <trevnorris>so enter/exit can still maintain the stack.
20:45:32  <trevnorris>but issue is if we bypass calling enter before calling the callback from MakeCallback then we loose the stack.
20:46:39  <trevnorris>ok. think I need to go back and cleanup my 33 commits :P
20:46:49  <othiym23>being able to correctly maintain the active domain is the goal
20:46:54  <trevnorris>hopefully that'll refresh my head.
20:46:55  <othiym23>the stack is just an artifact in that process
20:47:19  <MI6>joyent/node: Timothy J Fontaine v0.10 * 55546f5 : uv: Upgrade to v0.10.17 - http://git.io/m15r3g
20:47:24  * bradleymeckquit (Quit: bradleymeck)
20:47:25  <trevnorris>and active domain is easy
20:47:55  * creationixquit (Ping timeout: 245 seconds)
20:47:58  * bradleymeckjoined
20:47:58  <trevnorris>seriously, just added EventEmitter.active = this.domain at the top of EE#emit; the set it to undefined afterwards.
20:50:00  <trevnorris>othiym23: and quick perf tests show that adds very little overhead.
20:50:13  <othiym23>cool
20:51:54  <trevnorris>bnoordhuis: so what we doing about v8?
20:52:38  <trevnorris>bnoordhuis: oh, nm. wasn't paying attention to MI6
20:53:04  <MI6>nodejs-v0.10-windows: #224 UNSTABLE windows-ia32 (9/600) windows-x64 (7/600) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/224/
20:53:31  * julianduquejoined
20:53:36  <tjfontaine>trevnorris: well we need to actually get a good test case and figure it out, instead of necessarily just letting the revert continue on
20:53:57  <trevnorris>tjfontaine: nod
20:54:08  <MI6>libuv-v0.10-gyp: #79 UNSTABLE smartos-x64 (2/187) linux-ia32 (1/187) smartos-ia32 (2/187) windows-x64 (5/188) windows-ia32 (4/188) http://jenkins.nodejs.org/job/libuv-v0.10-gyp/79/
20:57:53  <MI6>libuv-review: #75 UNSTABLE smartos-ia32 (3/187) smartos-x64 (2/187) windows-ia32 (5/188) windows-x64 (4/188) http://jenkins.nodejs.org/job/libuv-review/75/
20:59:13  * st_lukejoined
21:02:51  <tjfontaine>should I include the configure change int he v0.10.19 changelog, or leave it out?
21:04:34  <MI6>libuv-node-integration: #239 UNSTABLE smartos-ia32 (1/600) linux-ia32 (1/600) smartos-x64 (5/600) linux-x64 (1/600) http://jenkins.nodejs.org/job/libuv-node-integration/239/
21:04:52  <MI6>nodejs-v0.10: #1499 UNSTABLE linux-ia32 (2/600) linux-x64 (3/600) smartos-ia32 (3/600) smartos-x64 (3/600) http://jenkins.nodejs.org/job/nodejs-v0.10/1499/
21:04:54  <bnoordhuis>was there a configure change?
21:05:13  <bnoordhuis>the mips float-abi thing?
21:05:16  <tjfontaine>aye
21:05:21  <tjfontaine>seems not very interesting
21:05:52  <tjfontaine>but there wasn't a lot of churn, and we're on the outside of our normal release cycle
21:05:59  <trevnorris>bnoordhuis: in the case of SomeClass::SomeClass() : val(true), other_val(false) {.. do you always like the ":" on the next line, or just if it doesn't fit on the first?
21:06:36  <tjfontaine>https://gist.github.com/tjfontaine/6691223
21:06:45  <bnoordhuis>trevnorris: only when it doesn't fit
21:06:54  <trevnorris>coolio, thanks :)
21:07:52  <bnoordhuis>tjfontaine: lgtm
21:08:08  <trevnorris>does anyone else feel like v8 is taking longer to compile?
21:08:22  <trevnorris>maybe it's just me.
21:08:30  <tjfontaine>bnoordhuis: mk
21:08:42  <tjfontaine>trevnorris: have you met ccache? :)
21:08:52  <trevnorris>tjfontaine: eh?
21:08:54  * AvianFluquit (Remote host closed the connection)
21:09:10  <tjfontaine>trevnorris: http://ccache.samba.org/
21:09:47  <wolfeidau>TooTallNate: you around?
21:09:58  <TooTallNate>wolfeidau: jes
21:10:13  <trevnorris>tjfontaine: coolio, thanks. have it on my system but never used it.
21:10:19  <wolfeidau>TooTallNate: I got reliable compilation of nodejs going on ubuntu now
21:10:24  <tjfontaine>trevnorris: see also distcc :P
21:10:54  <MI6>joyent/node: tjfontaine created branch v0.10.19-release - http://git.io/2IIw_A
21:11:06  <wolfeidau>TooTallNate: https://gist.github.com/wolfeidau/6691220 this is the recipe i am using at the moment
21:11:13  <trevnorris>tjfontaine: whoa. cool. now that mozilla's given me key's to their aws servers I should go play around. :P
21:11:23  * bentkusquit (Ping timeout: 248 seconds)
21:11:26  <tjfontaine>heh
21:11:35  <TooTallNate>wolfeidau: nice :D
21:11:38  <wolfeidau>TooTallNate: Works on ubuntu as long as you install 32 bit libs
21:12:45  <wolfeidau>TooTallNate: Yeah learnt a ton about all the arm compile flags gotta test the binaries more thoroughly there are some flags there that seem to influence v8 compilation
21:13:24  <wolfeidau>Seems to be quite a balancing act getting it all to use the same group of flags relating to vfp and hard float
21:13:31  <TooTallNate>wolfeidau: your builds are probably more pi-optimized than mine at this point :D
21:13:45  <tjfontaine>maybe I'll use that to start producing them on a regular basis
21:13:55  <TooTallNate>tjfontaine: that'd be cool!
21:14:04  <tjfontaine>I say as I just kicked off the build for v0.10.19-release
21:14:11  <TooTallNate>though i don't really like singling it out to rpi's like that
21:14:18  * AvianFlujoined
21:14:22  <wolfeidau>TooTallNate: Yeah I want to prove that and understand some of those flags
21:14:26  <tjfontaine>I thought wolfeidau was doing beagles
21:14:39  <TooTallNate>wolfeidau: confirm or deny?
21:14:53  <wolfeidau>Yeah i am planning to do tar balls and deb packages for beagle and pi
21:15:07  <tjfontaine>that covers the majority of the market, right?
21:15:15  <TooTallNate>wolfeidau: but do they need different compile flags?
21:15:22  <wolfeidau>yes they do
21:15:27  <TooTallNate>separate tarballs/packages then
21:15:32  <TooTallNate>it sounds like?
21:15:32  <MI6>nodejs-v0.10-windows: #225 UNSTABLE windows-ia32 (8/600) windows-x64 (7/600) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/225/
21:15:54  <trevnorris>bnoordhuis: think it would be helpful/useful to make a way for sub classes to receive the ObjectTemplate so they can set the prototype methods, instead of always setting all the same methods in the parent classes?
21:16:10  <TooTallNate>what about AR.Drone!?!
21:16:12  * TooTallNatejk
21:16:16  <tjfontaine>trevnorris: that seems relatively sane
21:16:20  <wolfeidau>haha
21:16:33  <tjfontaine>trevnorris: at least if I am interpreting your meaning appropriately
21:16:57  <trevnorris>tjfontaine: probably are. just like, in TCPWrap we set a bunch of StreamWrap methods. why not just have those done in StreamWrap instead?
21:17:00  <bnoordhuis>trevnorris: maybe. i haven't really missed that so far
21:17:41  <wolfeidau>I will get some recipes for the beagle as well and see if i can get them all built from the one toolchain
21:17:52  <tjfontaine>I remember doing the rest of [un]ref and being depressed in how many places I needed to touch
21:18:00  <tjfontaine>wolfeidau: nod
21:18:19  <trevnorris>bnoordhuis: guess I just like the idea of a single point for change. i.e. was thinking of adding addAsyncListener to each handle so they could be controlled individually (like domains) but didn't want to go touch a bunch of files to add it.
21:18:49  * bradleymeckquit (Quit: bradleymeck)
21:22:10  <wolfeidau>tjfontaine TooTallNate would be great to have tar balls up for those two platforms especially if it is automated
21:23:31  * creationixjoined
21:24:48  * kazuponjoined
21:30:03  * kazuponquit (Ping timeout: 268 seconds)
21:31:10  <tjfontaine>wolfeidau: ya, I'd not like to get too far into alphabet soup for it
21:31:22  <tjfontaine>wolfeidau: but if we can do it reasonably, it'd be great
21:32:45  <trevnorris>indutny: three jslint errors on _tls_legacy
21:33:00  <indutny>trevnorris: oh really? may I ask you to fix them?
21:33:08  <indutny>trevnorris: I'm on a terrible internet right now
21:33:09  <trevnorris>indutny: sure. no problem :)
21:33:14  <indutny>trevnorris: thanks :)
21:35:07  <wolfeidau>tjfontaine: yeah agreed as long as you get the two main platforms and use an older glibc to cater for both debian and ubuntu it will work fine
21:35:55  <wolfeidau>tjfontaine: I will hassle you in a few days once i have tested out the recipes for both
21:36:05  <MI6>joyent/node: Trevor Norris master * 25dce6d : lint: fix a few semicolons - http://git.io/61mTIA
21:36:07  <tjfontaine>wolfeidau: please and thank you
21:38:42  * defunctzombiechanged nick to defunctzombie_zz
21:40:29  * julian_duquejoined
21:42:33  * julianduquequit (Disconnected by services)
21:42:37  * julian_duquechanged nick to julianduque
21:45:12  <MI6>nodejs-master: #576 UNSTABLE smartos-x64 (6/642) http://jenkins.nodejs.org/job/nodejs-master/576/
21:47:45  * st_lukequit (Remote host closed the connection)
21:53:18  * emilsedgh_part ("Konversation terminated!")
21:56:51  * jmar777joined
22:01:27  * rendarquit
22:02:00  * zotjoined
22:03:13  * jmar777quit (Remote host closed the connection)
22:04:31  * AvianFluquit (Read error: Connection reset by peer)
22:05:15  * AvianFlujoined
22:06:07  * julianduquequit (Ping timeout: 260 seconds)
22:08:59  <MI6>nodejs-master-windows: #369 UNSTABLE windows-x64 (23/642) windows-ia32 (21/642) http://jenkins.nodejs.org/job/nodejs-master-windows/369/
22:09:24  <MI6>joyent/node: tjfontaine created tag v0.10.19 - http://git.io/VgXFKw
22:10:26  <MI6>joyent/node: Timothy J Fontaine v0.10 * 9135c7f : blog: Post for v0.10.19 (+3 more commits) - http://git.io/MI8Wlg
22:11:16  <tjfontaine>fuckign shasums
22:11:45  <trevnorris>eh, they're overrated anyways. ;P
22:13:36  * c4miloquit (Remote host closed the connection)
22:14:10  * defunctzombie_zzchanged nick to defunctzombie
22:16:52  * bnoordhuisquit (Ping timeout: 260 seconds)
22:17:02  * zotquit (Quit: Leaving.)
22:17:11  <MI6>joyent/node: Timothy J Fontaine v0.10 * cfa03ad : blog: add missing shasums for v0.10.19 release - http://git.io/6WezaA
22:22:01  <MI6>nodejs-v0.10: #1500 UNSTABLE linux-ia32 (2/600) linux-x64 (1/600) smartos-x64 (3/600) osx-x64 (1/600) http://jenkins.nodejs.org/job/nodejs-v0.10/1500/
22:22:38  <othiym23>so uh dangerous times here
22:22:45  <tjfontaine>where there whenn why
22:22:56  <othiym23>does anybody know of a way to change the name of a function after it's been defined?
22:23:05  <trevnorris>othiym23: in js?
22:23:07  <tjfontaine>othiym23: TooTallNate has a module that I think can do it
22:23:10  <othiym23>trevnorris: y
22:23:16  <tjfontaine>oh
22:23:18  <tjfontaine>gwlat
22:23:21  <tjfontaine>*glwat
22:23:39  <othiym23>yeah, I'm gonna have to use eval, I think
22:23:41  <othiym23>barrrrf
22:23:57  <othiym23>fuckin' Express
22:24:17  <tjfontaine>Function(func.toString().replace())!
22:24:21  <trevnorris>othiym23: case?
22:24:52  <othiym23>tjfontaine: you're beautiful / that's horrible
22:25:07  <TooTallNate>othiym23: i think tjfontaine is talking about https://github.com/TooTallNate/node-function-name
22:25:18  <tjfontaine>othiym23: :)
22:25:19  <TooTallNate>tjfontaine: that won't work if the function relies on any closure variables
22:25:23  * kazuponjoined
22:25:37  <othiym23>trevnorris: got something looking for a middleware function by name, but I'm wrapping all the middlewares for pre-0.8 error tracing
22:25:44  <tjfontaine>TooTallNate: nod
22:25:47  <trevnorris>heh, fund.
22:25:49  <trevnorris>*fun.
22:26:27  <othiym23>TooTallNate: I think tjfontaine's approach will work just fine, because I'm doing the call within the scope of where the closure variables are
22:27:44  <othiym23>god, it sucks that I have to do this at all, I already have this nonsense to deal with: https://github.com/newrelic/node-newrelic/blob/master/lib/instrumentation/connect.js#L37-L66
22:28:08  <groundwater>line #37
22:28:36  <othiym23>yup
22:29:01  <othiym23>ES6 may allow you to set Function.length, but that's little comfort to me now
22:29:37  <MI6>nodejs-v0.10: #1501 UNSTABLE linux-x64 (1/600) smartos-x64 (2/600) http://jenkins.nodejs.org/job/nodejs-v0.10/1501/
22:30:06  * kazuponquit (Ping timeout: 264 seconds)
22:30:49  * julianduquejoined
22:30:53  * julianduquequit (Changing host)
22:30:53  * julianduquejoined
22:31:25  <othiym23>lib/instrumentation/connect.js|66 col 36 warning| The Function constructor is a form of eval.
22:31:31  <othiym23>YOU DON'T SAY JSHINT
22:31:32  <LOUDBOT>YES INDEED KITTY CAN USE MORE CATNIP
22:37:41  * jmar777joined
22:38:27  * dominictarrjoined
22:40:41  * hzquit
22:45:50  * Kakeraquit (Ping timeout: 264 seconds)
22:48:14  <MI6>node-review: #88 UNSTABLE windows-ia32 (7/600) osx-ia32 (1/600) centos-x64 (1/600) centos-ia32 (1/600) windows-x64 (8/600) osx-x64 (1/600) smartos-x64 (2/600) http://jenkins.nodejs.org/job/node-review/88/
22:49:24  <MI6>nodejs-v0.10-windows: #226 UNSTABLE windows-ia32 (9/600) windows-x64 (6/600) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/226/
22:59:38  <othiym23>https://gist.github.com/othiym23/6692494
23:01:04  <othiym23>new Function() just wraps the whole thing up in another closure
23:01:08  <othiym23>named 'anonymous'
23:01:10  <othiym23>thanks V8!
23:01:32  <tjfontaine>othiym23: I thought it had a parameter for name
23:01:47  <tjfontaine>hm so wrong.
23:02:31  <othiym23>it has parameters for parameters!
23:02:47  <tjfontaine>indeed
23:03:14  * jmar777quit (Remote host closed the connection)
23:09:51  <othiym23>actually like https://gist.github.com/othiym23/6692582 better
23:10:00  <othiym23>if I have to use eval anyway, might as well simplify that code
23:10:42  <tjfontaine>leading commas, they make the baby jesus cry :P
23:11:10  <tjfontaine>othiym23: that could be much worse I suppose
23:11:13  <othiym23>too bad, house style
23:11:20  <tjfontaine>hehe
23:12:15  * kellabytequit (Quit: Quit)
23:13:39  <tjfontaine>TooTallNate: 7196742852c9da5d29e2fb1e333bca9d21c8e205 "tls: don't push() incoming data when ondata is set" isn't valid on streams3 interface right? where there is no more ondata
23:14:00  <TooTallNate>tjfontaine: correct, it's a v0.10 only patch
23:14:11  <tjfontaine>ok, so the test doesn't need to follow along as well
23:14:17  <TooTallNate>tjfontaine: indeed
23:14:31  <TooTallNate>it would probably fail in fact
23:14:38  <tjfontaine>it timesout yes
23:15:20  * zotjoined
23:16:54  <MI6>nodejs-v0.10-windows: #227 UNSTABLE windows-ia32 (9/600) windows-x64 (10/600) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/227/
23:17:34  <TooTallNate>tjfontaine: lol, well yes, please delete when merging to master
23:17:36  * EhevuTovquit (Quit: This computer has gone to sleep)
23:17:39  <tjfontaine>TooTallNate: :)
23:18:26  * zotquit (Client Quit)
23:18:44  <tjfontaine>hmm, how is test-fs-watch failing for me
23:19:14  * mikealquit (Quit: Leaving.)
23:22:14  * mikealjoined
23:25:58  * kazuponjoined
23:27:33  * octetcloudquit (Ping timeout: 252 seconds)
23:33:37  * kazuponquit (Read error: Connection reset by peer)
23:33:59  * kazuponjoined
23:34:08  * kellabytejoined
23:38:56  * kazuponquit (Ping timeout: 268 seconds)
23:44:25  <tjfontaine>hmm stale files it seems
23:47:21  * hij1nx_onaboatquit (*.net *.split)
23:47:22  * MI6quit (*.net *.split)
23:47:22  * mburnsquit (*.net *.split)
23:47:23  * mikealquit (*.net *.split)
23:47:23  * dshaw_quit (*.net *.split)
23:47:23  * amartensquit (*.net *.split)
23:47:26  * rphillipsquit (*.net *.split)
23:47:26  * isaacsquit (*.net *.split)
23:47:28  * Raltquit (*.net *.split)
23:47:29  * julianduquequit (*.net *.split)
23:47:29  * brsonquit (*.net *.split)
23:47:31  * `3rdEdenquit (*.net *.split)
23:47:32  * philipsquit (*.net *.split)
23:47:35  * Benviequit (*.net *.split)
23:47:37  * toothrquit (*.net *.split)
23:47:37  * LOUDBOTquit (*.net *.split)
23:47:38  * defunctzombiequit (*.net *.split)
23:47:40  * skebcio_quit (*.net *.split)
23:47:42  * kuplatupsuquit (*.net *.split)
23:47:43  * sindresorhusquit (*.net *.split)
23:47:43  * dsantiagoquit (*.net *.split)
23:47:43  * chrisdickinsonquit (*.net *.split)
23:47:44  * pquernaquit (*.net *.split)
23:47:45  * blissdevquit (*.net *.split)
23:47:46  * LeftWingquit (*.net *.split)
23:47:50  * niskaquit (*.net *.split)
23:47:53  * isaacbwquit (*.net *.split)
23:47:53  * saghulquit (*.net *.split)
23:47:54  * Cheeryquit (*.net *.split)
23:47:55  * wolfeidauquit (*.net *.split)
23:47:56  * Domenic_quit (*.net *.split)
23:47:56  * dlmanningquit (*.net *.split)
23:47:57  * brucemquit (*.net *.split)
23:47:58  * AvianFluquit (*.net *.split)
23:47:58  * wwicksquit (*.net *.split)
23:47:59  * indutnyquit (*.net *.split)
23:48:00  * tjfontainequit (*.net *.split)
23:48:00  * nsmquit (*.net *.split)
23:48:00  * othiym23quit (*.net *.split)
23:48:01  * ralquit (*.net *.split)
23:48:02  * rjequit (*.net *.split)
23:48:03  * tellnesquit (*.net *.split)
23:48:05  * creationixquit (*.net *.split)
23:48:05  * inolenquit (*.net *.split)
23:48:05  * kenperkinsquit (*.net *.split)
23:48:05  * Damn3dquit (*.net *.split)
23:48:06  * KiNgMaRquit (*.net *.split)
23:48:07  * qmxquit (*.net *.split)
23:48:08  * kaesoquit (*.net *.split)
23:48:09  * MI6joined
23:48:09  * hij1nx_onaboatjoined
23:48:09  <trevnorris>wtf
23:48:09  <trevnorris>helloooo out there
23:49:27  <MI6>joyent/node: Timothy J Fontaine master * 42af62f : Merge remote-tracking branch 'upstream/v0.10' (+14 more commits) - http://git.io/TZ6qVA
23:50:32  <TooTallNate>trevnorris: helloooo?
23:50:50  <trevnorris>TooTallNate: wonder why so many people dropped.
23:51:12  <TooTallNate>trevnorris: i think freenode does some kind of sweep eventually
23:51:17  <TooTallNate>for stale sockets maybe
23:51:17  <trevnorris>ah, ok
23:51:22  * mikealjoined
23:51:22  * dshaw_joined
23:51:22  * amartensjoined
23:51:22  * rphillipsjoined
23:51:22  * isaacsjoined
23:51:31  <TooTallNate>i've seen it before
23:53:54  <trevnorris>TooTallNate: so here's a strange one. that async listeners patch. for some reason message/stdin_messages shows Pipe.tickCallback instead of process.tickCallback
23:54:14  * indexzerojoined
23:54:14  * mburnsjoined
23:54:14  * julianduquejoined
23:54:14  * AvianFlujoined
23:54:14  * creationixjoined
23:54:14  * inolenjoined
23:54:14  * kenperkinsjoined
23:54:14  * brsonjoined
23:54:14  * Damn3djoined
23:54:14  * sindresorhusjoined
23:54:14  * dsantiagojoined
23:54:14  * rjejoined
23:54:14  * toothrjoined
23:54:14  * philipsjoined
23:54:14  * chrisdickinsonjoined
23:54:14  * Benviejoined
23:54:14  * Raltjoined
23:54:14  * wwicksjoined
23:54:14  * LOUDBOTjoined
23:54:14  * `3rdEdenjoined
23:54:14  * wolfeidaujoined
23:54:14  * skebcio_joined
23:54:14  * LeftWingjoined
23:54:14  * Domenic_joined
23:54:14  * tellnesjoined
23:54:14  * pquernajoined
23:54:14  * dlmanningjoined
23:54:14  * kaesojoined
23:54:14  * blissdevjoined
23:54:14  * KiNgMaRjoined
23:54:14  * defunctzombiejoined
23:54:14  * qmxjoined
23:54:14  * niskajoined
23:54:14  * indutnyjoined
23:54:14  * tjfontainejoined
23:54:14  * isaacbwjoined
23:54:14  * kuplatupsujoined
23:54:14  * saghuljoined
23:54:14  * brucemjoined
23:54:14  * nsmjoined
23:54:14  * othiym23joined
23:54:14  * Cheeryjoined
23:54:14  * raljoined
23:55:18  * MI6changed nick to Guest16550
23:56:15  <TooTallNate>WOAH!
23:56:27  <trevnorris>yeah. and I can't figure out why
23:57:19  <tjfontaine>hm?
23:58:12  <tjfontaine>DON'T LEAVE ME HANGING IN THIS NETSPLIT GUYS
23:58:13  <tjfontaine>:P
23:58:13  <LOUDBOT>MIDGET PORN IS HARD TO COME BY HAHA GET IT?!
23:58:27  <tjfontaine>LOUDBOT: OMG URHILARIOUS
23:58:27  <LOUDBOT>tjfontaine: WIND UP THE SCOOTER WE'RE GOING TO CRACKER BARREL
23:58:50  <Guest16550>nodejs-master: #577 UNSTABLE smartos-ia32 (1/642) smartos-x64 (6/642) http://jenkins.nodejs.org/job/nodejs-master/577/
23:58:56  <tjfontaine>aww poor mi6
23:59:09  * Guest16550quit (Remote host closed the connection)