00:00:01  * ircretaryquit (Remote host closed the connection)
00:00:08  * ircretaryjoined
00:00:29  * dsantiagoquit (Quit: Computer has gone to sleep.)
00:06:52  * dsantiagojoined
00:18:05  * piscisaureus_joined
00:36:13  * qardjoined
00:37:53  * hzquit
00:40:22  * dsantiagoquit (Quit: Computer has gone to sleep.)
00:41:19  * dsantiagojoined
00:49:03  * qardquit (Quit: Leaving.)
00:57:46  * stagasjoined
01:01:38  * luigyquit (Ping timeout: 256 seconds)
01:07:21  * dominictarrquit (Quit: dominictarr)
01:20:31  * dominictarrjoined
01:48:47  * dominictarrquit (Ping timeout: 255 seconds)
02:41:34  * st_lukejoined
02:57:25  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
03:09:16  * kenperkinsquit (Quit: Computer has gone to sleep.)
03:25:00  * piscisaureus_joined
03:29:15  * brsonjoined
03:37:57  * st_lukequit (Remote host closed the connection)
04:01:48  * timoxleyjoined
04:21:17  * timoxleyquit (Quit: Computer has gone to sleep.)
04:24:26  * EhevuTovjoined
04:28:32  * EhevuTovquit (Client Quit)
04:49:27  * piscisaureus_quit (Ping timeout: 252 seconds)
04:51:06  * normanmjoined
05:06:58  * loladirojoined
05:22:32  * AvianFluquit (Remote host closed the connection)
05:49:44  * st_lukejoined
05:54:56  * loladiroquit (Quit: loladiro)
06:10:20  * stolsmajoined
06:12:10  * brsonquit (Quit: leaving)
06:29:15  * rendarjoined
06:30:03  * benoitcquit (Excess Flood)
06:38:56  * benoitcjoined
06:50:24  * `3rdEdenjoined
07:23:31  * EhevuTovjoined
07:34:10  * mraleph1joined
07:34:17  * mralephquit (Read error: Connection reset by peer)
07:36:59  * csaohjoined
08:20:26  * dominictarrjoined
09:01:58  * dominictarrquit (Quit: dominictarr)
09:15:23  * csaohquit (Quit: csaoh)
09:16:29  * kuebkjoined
09:16:58  <kuebk>hello
09:18:52  <kuebk>does anyone experienced something like this
09:18:57  <kuebk>make: Circular . <- Release/obj.target/libneon.stamp dependency dropped
09:19:05  <kuebk>while using node-gyp?
09:28:24  * kuebkquit (Read error: Connection reset by peer)
09:29:06  * kuebkjoined
09:39:41  * csaohjoined
09:46:23  * EhevuTovquit (Quit: This computer has gone to sleep)
09:46:41  * hzjoined
09:57:41  * stagasquit (Remote host closed the connection)
10:21:37  * inolenquit (Read error: Connection reset by peer)
10:21:55  * inolenjoined
10:35:56  * csaohquit (Quit: csaoh)
10:50:00  * csaohjoined
11:45:58  * rendar_joined
11:47:12  * rendarquit (Read error: Connection reset by peer)
12:38:57  * st_lukequit (Remote host closed the connection)
13:10:08  * AvianFlujoined
13:18:57  * kevinswiberjoined
13:30:47  * benoitcquit (Excess Flood)
13:31:29  * benoitcjoined
13:43:45  * kevinswiberquit (Remote host closed the connection)
13:50:26  * kevinswiberjoined
13:56:52  * c4milojoined
14:22:54  * loladirojoined
14:28:24  * kenperkinsjoined
14:36:49  * c4miloquit (Remote host closed the connection)
14:41:16  * c4milojoined
14:53:30  <isaacs>good morning heroes
15:19:25  * pachetjoined
15:29:06  * kuebkpart
15:35:05  * `3rdEdenquit (Remote host closed the connection)
15:43:06  * kevinswiberquit (Remote host closed the connection)
15:45:52  * kevinswiberjoined
15:46:27  * defunctzombie_zzchanged nick to defunctzombie
15:52:49  * kevinswiberquit (Remote host closed the connection)
15:53:38  <tjfontaine>all hail our monarch isaacs
15:56:09  <isaacs>hahah
15:56:31  <isaacs>the thing that sucks about monarchy is that there's usually only one way to quit the job.
15:57:13  <tjfontaine>violent overthrow, death, or the unlikely abdication
15:57:23  <tjfontaine>the first two usually are the same :)
15:59:40  <isaacs>yep
16:00:10  <isaacs>the trick is always to make sure the group you're pissing off is smaller than the group you're making happy.
16:00:34  <tjfontaine>how to win friends and influence people, as defined by isaacs
16:14:13  * dominictarrjoined
16:16:31  <isaacs>tjfontaine: well, if i was really a pro, i'd convince the people who i'm pissing off that actualy they want what the other group wants.
16:16:41  <isaacs>tjfontaine: and show them how i'm doing them a favor by not giving in to theier request.
16:16:50  <tjfontaine>true
16:20:01  <tjfontaine>isaacs: I keep meaning to get your opinion of this, https://github.com/tjfontaine/node/commit/bedd5d40daa75ce23b7f0cc44158140f0018301e
16:20:50  <isaacs>oh, i like that
16:21:08  <isaacs>after i make the pkg, i often try to `make install` and get screwed because it tries to put it in a weird place.
16:21:30  <isaacs>so, rubber-stamp lgtm. if it causes a problem when i next make a release, i'll bug you about it.
16:21:37  <tjfontaine>ok
16:21:41  <isaacs>go ahead and land
16:21:52  <isaacs>in stable, preferrably
16:22:50  <tjfontaine>ok, it's not particularly a problem in stable, but should be fine there as well
16:23:34  * dapjoined
16:34:53  * qardjoined
16:35:07  * hueniversejoined
16:41:25  * hzquit (Read error: Connection reset by peer)
16:41:43  * abraxasjoined
16:41:52  * hzjoined
16:41:56  * EhevuTovjoined
16:45:53  * abraxasquit (Ping timeout: 245 seconds)
16:48:54  * dapquit (Quit: Leaving.)
16:49:54  <trevnorris>morning.
16:50:02  <tjfontaine>hola
16:54:50  * EhevuTovquit (Quit: Leaving)
16:55:10  * loladiroquit (Quit: loladiro)
16:56:42  <indutny>morning
16:59:47  * kevinswiberjoined
17:06:35  <hueniverse>morning
17:09:03  * brsonjoined
17:10:09  * dominictarrquit (Quit: dominictarr)
17:11:34  * csaohquit (Quit: csaoh)
17:11:46  * inolenquit (Quit: Leaving.)
17:13:45  <MI6>joyent/node: Timothy J Fontaine v0.10 * 76cbd03 : build: only use DESTDIR instead of PREFIX for pkg - http://git.io/bZofww
17:14:40  <mmalecki>indutny: yo
17:14:49  <indutny>yo
17:14:56  <indutny>sup?
17:15:44  <mmalecki>indutny: Ken got a fix for your S3 bug
17:15:51  <indutny>I know
17:15:55  <indutny>he's talking to me right now :)
17:16:04  <mmalecki>indutny: oh, great!
17:16:18  <indutny>mmalecki: do you have dropbox account?
17:16:22  <indutny>I mean, are you using it? :)
17:16:32  <mmalecki>indutny: I have one, but not using it
17:16:35  <mmalecki>just the web ui
17:16:35  <indutny>ah ok
17:16:39  <mmalecki>what do you need tho?
17:16:46  <indutny>just in case if you'll ever want to backup it to S3 - I've a solution
17:16:57  * indutnydon't believes dropbox
17:17:01  <mmalecki>oh, nice
17:17:07  <mmalecki>I don't believe clouds at all
17:17:13  <indutny>yeah, me neither
17:17:28  <indutny>but its like a lot of data that I have no other place to store on
17:18:36  <mmalecki>yeah, I hear you. I don't really have data which isn't code :)
17:20:42  * dapjoined
17:20:51  <indutny>mmalecki: you just don't have wife yet :)
17:21:02  <indutny>there're like tons of family photos and videos
17:21:10  <indutny>which I don't want to loose
17:21:16  <mmalecki>hahahaha, true
17:21:42  <tjfontaine>what's the indutny solution for dropbox?
17:21:52  <mmalecki>not planning to have a wife just yet :)
17:22:03  * defunctzombiechanged nick to defunctzombie_zz
17:22:04  <indutny>tjfontaine: well, I just stream files from dropbox to s3
17:22:09  <tjfontaine>indutny: ah ok
17:22:12  <indutny>and then you can archive them to cheap glacier
17:22:21  <tjfontaine>was dropbox on s3 in the begining?
17:22:24  <tjfontaine>*wasn't
17:22:31  <indutny>well
17:22:34  <indutny>idk
17:22:34  <indutny>but
17:22:57  <indutny>its ok to have it in multiple places
17:23:05  <indutny>without uploading it twice to two different storages
17:23:11  <indutny>I mean, uploading twice from my laptop
17:23:17  <tjfontaine>I know a guy who works on the glacier team, I made a joke about "reel-to-reel" and he replied "no comment"
17:24:04  <indutny>:)
17:24:08  <indutny>well
17:24:13  <indutny>I don't think its secure there
17:24:15  <MI6>nodejs-v0.10: #174 UNSTABLE smartos-x64 (3/582) http://jenkins.nodejs.org/job/nodejs-v0.10/174/
17:24:20  <indutny>but I can upload it to a couple of different places
17:24:24  <indutny>like S3, rackspace storage
17:24:28  <indutny>dropbox
17:24:47  <indutny>chance of survival will be much bigger
17:26:21  <mmalecki>indutny: you're preparing for a nuclear fallout ;) ?
17:26:30  <indutny>:)
17:26:41  <indutny>basically, its this https://gist.github.com/indutny/252b1f9b91c8e02230f4
17:26:45  <indutny>if anyone is interested
17:26:51  <indutny>but it requires pkgcloud fixes
17:27:05  <indutny>also, code is pretty shitty
17:28:17  <MI6>joyent/node: Timothy J Fontaine master * 5037f9a : build: only use DESTDIR instead of PREFIX for pkg - http://git.io/EPtHug
17:29:42  * loladirojoined
17:32:39  * loladiroquit (Client Quit)
17:34:44  * othiym23quit (Quit: WeeChat 0.3.8)
17:37:21  * kevinswiberquit (Remote host closed the connection)
17:38:19  <MI6>nodejs-master: #195 UNSTABLE smartos-x64 (2/587) linux-ia32 (1/587) http://jenkins.nodejs.org/job/nodejs-master/195/
17:40:48  * loladirojoined
17:41:00  * indexzerojoined
17:45:27  * indexzeroquit (Client Quit)
17:48:25  * kevinswiberjoined
17:55:26  <trevnorris>how do we want to handle the cut over to v8 Typed Arrays?
17:56:32  <trevnorris>by 3.18.6 it'll have pretty much the entire implementation necessary.
17:57:42  * normanmquit (Quit: Computer has gone to sleep.)
17:58:06  * normanmjoined
17:58:07  * othiym23joined
18:00:19  <tjfontaine>trevnorris: will it not be as easy as removing what we have in our tree?
18:00:20  * c4miloquit (Remote host closed the connection)
18:00:56  <trevnorris>tjfontaine: might be. was wondering if we even care about any backwards compatibility support.
18:01:12  <tjfontaine>well, what won't be backwards compatible?
18:01:22  <trevnorris>the syntax.
18:01:38  <tjfontaine>but the v8 proper implementation is following the standard?
18:02:00  <trevnorris>there isn't a standard for the cc side.
18:02:13  <tjfontaine>oh well, ok that's what I was curious about
18:02:25  <tjfontaine>so you're worried about people who might be doing typed arrays from the native side
18:02:25  * amartensjoined
18:02:33  <tjfontaine>are you aware of any projects doing that?
18:02:37  <trevnorris>TooTallNate
18:02:46  <tjfontaine>which one?
18:03:25  <MI6>nodejs-v0.10-windows: #5 UNSTABLE windows-ia32 (7/582) windows-x64 (8/582) http://jenkins.nodejs.org/job/nodejs-v0.10-windows/5/
18:03:45  * inolenjoined
18:03:46  <tjfontaine>which module that is
18:04:05  <trevnorris>hm. maybe it was just the js side i'm remembering.
18:04:16  <tjfontaine>anyway, as far as that is concerned, it should be #ifdef'd around the node version
18:04:59  <tjfontaine>it is something that'd I'd be interested in seeing if someone is actually doing from the native side, I would expect most just make a buffer and then do the type work in js land
18:05:12  <trevnorris>tjfontaine: lol. nm. just checked v8_typed_array.h. no api is exposed anyways.
18:05:50  <trevnorris>there's an additional problem. right now you can create an array buffer from a buffer.
18:06:08  <trevnorris>i'm not sure that will be possible using v8 typed array. i'm reading through the implementation now
18:06:47  <tjfontaine>you can't just use an "iterable" for it?
18:06:54  <trevnorris>iterable?
18:07:20  <trevnorris>oh, and we can. they have an api that allows to pass existing data.
18:07:21  <tjfontaine>something it can for (s in src){}
18:07:52  <trevnorris>oooh. sure that's possible, but so so bad performance-wise.
18:08:21  <tjfontaine>well, yes well typedarrays don't instill much faith in me
18:09:02  <trevnorris>does this mean the mem won't be freed on gc?: "The memory block will not be reclaimed when ArrayBuffer is gc'd"
18:09:42  <tjfontaine>source?
18:10:18  <trevnorris>tjfontaine: http://git.io/pTSCuw
18:10:23  * indexzerojoined
18:11:39  <trevnorris>what i'm wondering is how to know when the memory could be free'd if we're not given a gc callback.
18:18:43  <MI6>nodejs-master-windows: #3 UNSTABLE windows-x64 (146/587) windows-ia32 (147/587) http://jenkins.nodejs.org/job/nodejs-master-windows/3/
18:22:03  * dominictarrjoined
18:23:31  * brsonquit (Quit: leaving)
18:23:48  * brsonjoined
18:26:13  * c4milojoined
18:44:28  * indexzeroquit (Quit: indexzero)
18:45:13  * stagasjoined
18:53:36  * isaacs_mobilejoined
19:02:11  * indexzerojoined
19:02:18  <trevnorris>tjfontaine: i swear. on of these days i'm going to get a green light from jenkins!
19:03:08  <tjfontaine>you were really close
19:03:40  * `3rdEdenjoined
19:03:54  <tjfontaine>trevnorris: anyway the situation is much worse if it were testing on windows :)
19:04:03  <tjfontaine>or if we included pummel
19:04:06  <trevnorris>lol
19:04:43  * philipsquit (Ping timeout: 245 seconds)
19:08:25  * isaacs_mobilequit (Remote host closed the connection)
19:09:30  <trevnorris>are any of those other "make test-*" even maintained?
19:10:06  <trevnorris>(other than bench)
19:13:08  * kevinswiberquit (Remote host closed the connection)
19:13:36  * philipsjoined
19:17:26  * indexzeroquit (Quit: indexzero)
19:23:48  * isaacs_mobilejoined
19:24:09  * kevinswiberjoined
19:34:46  * isaacs_mobilequit (Remote host closed the connection)
19:37:50  * `3rdEdenquit (Remote host closed the connection)
19:41:39  * isaacsfg
19:41:58  <isaacs>trevnorris: yeah, sort of
19:42:05  <isaacs>trevnorris: test-message gets run often
19:42:20  <isaacs>trevnorris: test-all is too slow, and has some known failures.
19:43:02  <trevnorris>test-valgrind has too much output.
19:43:27  * inolenquit (Ping timeout: 252 seconds)
19:48:08  * loladiroquit (Quit: loladiro)
19:57:27  * mikealjoined
19:57:44  <tjfontaine>simple, message, and pummel are run regularly on the CI
19:57:51  <tjfontaine>but "maintained" is another thing
20:02:00  * loladirojoined
20:07:16  * normanmquit (Quit: Computer has gone to sleep.)
20:10:22  * perezdjoined
20:37:14  * st_lukejoined
20:38:01  * saghul_joined
20:38:11  <tjfontaine>hmm no sblom or piscisaureus_
20:40:14  * stagasquit (Quit: ChatZilla 0.9.90-rdmsoft [XULRunner 1.9.0.17/2009122204])
20:44:00  * stolsmaquit (Ping timeout: 264 seconds)
20:44:41  * kevinswiberquit (Remote host closed the connection)
20:54:33  * kevinswiberjoined
20:55:58  * kevinswiberquit (Remote host closed the connection)
21:08:37  <MI6>joyent/node: tjfontaine created branch v0.7.8 - http://git.io/a_S5uQ
21:08:52  <tjfontaine>I did what?!
21:09:25  <tjfontaine>I'm going tos tart browsing github in incognito
21:10:38  <amartens>hah
21:11:12  * inolenjoined
21:23:22  <trevnorris>tjfontaine: wtf? how'd that happen?
21:24:15  <tjfontaine>trevnorris: tree/branch/tag drop down, you can also create a branch from that, I was in the filter section and just blindly typed what I was looking for and hit return
21:25:00  * mikealquit (Quit: Leaving.)
21:25:24  * saghul_quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
21:28:13  <trevnorris>lol
21:28:31  <trevnorris>they make it way too easy to mess w/ your repo online.
21:28:44  <tjfontaine>I would love to be able to turn it all off
21:28:52  * st_lukequit (Remote host closed the connection)
21:28:56  <tjfontaine>please to just be a pretty viewvc and issue tracker, kthx
21:29:54  * rendar_quit
21:30:53  <trevnorris>i created a greasemonkey script for that. :)
21:31:26  * timoxleyjoined
21:31:28  <trevnorris>no more merge button!
21:32:05  <trevnorris>but yeah. that "create a new branch" think is way too easy to screw up.
21:32:21  <trevnorris>and who seriously isn't creating a branch locally and pushing it?
21:32:39  <tjfontaine>people who click the edit file button
21:33:37  <trevnorris>i hate they made that possible.
21:34:27  <tjfontaine>join the club
21:34:35  <trevnorris>isaacs: can we do some remote branch cleanup? it makes me hurt inside. ;-)
21:34:49  <isaacs>trevnorris: sure.
21:34:54  <trevnorris>awesome.
21:35:00  <isaacs>git push ry-push :branchname
21:35:11  <isaacs>(assuming your ry-push remote is git@github.com:joyent/node
21:35:14  <tjfontaine>except for this case, where I created a branch that matches a tag name ...
21:35:21  <isaacs>tjfontaine: oh... weird.
21:35:31  <isaacs>how does one delete the branch in that case?
21:35:31  <tjfontaine>git push upstream-rw :v0.7.8
21:35:31  <tjfontaine>error: dst refspec v0.7.8 matches more than one.
21:35:31  <tjfontaine>error: failed to push some refs to 'git@github.com:joyent/node.git'
21:35:38  <tjfontaine>I haven't googled to find out yet
21:35:43  <isaacs>oh, you have to do something like refs/branches/blerg
21:35:50  <tjfontaine>probably
21:35:53  <trevnorris>isaacs: ok. since the release branches are tagged, they don't need to remain. cool if I remove all the old ones?
21:36:08  <isaacs>trevnorris: well.. people actually do occasionally refer to them
21:36:10  <isaacs>i'm not sure why
21:36:22  <isaacs>but i've gotten complaints about release branches not being there.
21:36:31  <trevnorris>hm. that's unfortunate. since the branch and tag just point to the same hash.
21:37:11  <trevnorris>isaacs: oh, it's probably because they still refer to it via *-release.
21:37:31  <isaacs>yeah
21:38:04  <trevnorris>oy vey
21:39:10  * c4miloquit (Remote host closed the connection)
21:47:19  <isaacs>indutny: did you get burned?
21:47:29  <indutny>huh? :)
21:47:31  <indutny>by what?
21:48:18  <trevnorris>(i'm guessing the sun ;-)
21:48:22  <indutny>oh
21:48:28  <indutny>well, there're rains in moscow
21:48:30  <indutny>and clouds
21:48:37  <indutny>so, I think, no
21:48:50  <indutny>in past years moscow became more like london
21:48:57  <indutny>in terms of weather
21:49:00  <trevnorris>eh, weren't you on the beach recently?
21:49:17  <indutny>yeah, I was
21:49:22  <indutny>and now I'm in moscow :)
21:50:49  <trevnorris>indutny: so my initial implementation of the allocator passed the "char*" as the make weak argument. and I didn't bother void* casting it.
21:50:54  <trevnorris>since I figured they were both pointers.
21:51:03  <indutny>and?
21:51:07  <trevnorris>but then I found v8 was running out of memory if I made a large buffer.
21:51:35  <trevnorris>it seemed like it was copying the entire buffer into the v8 heap.
21:51:43  <trevnorris>I had to explicitely cast the char* to a void*
21:52:55  <trevnorris>indutny: that doesn't seem like what should be happening. would you know why?
21:52:58  <isaacs>indutny: oh, i saw your tweet about 3rd degree bands.
21:53:04  <indutny>ah
21:53:11  <isaacs>indutny: figured it might be a typo if you were burned all up or something
21:53:16  <indutny>hahaha
21:53:17  <indutny>:D
21:53:18  <indutny>well
21:53:20  <isaacs>indutny: but it sounds like you'er fine :)
21:53:23  <indutny>no that's actualy band name
21:53:28  <indutny>3rdegree
21:53:37  <indutny>but I think they're too politicaly-social
21:53:42  <indutny>too much to listen them
21:53:45  <isaacs>indutny: i see
21:56:45  <tjfontaine>so
21:57:32  <tjfontaine>fs.writeSync is [generally] interruptable on unicies, but is definitely not on windows (at least in its current implementation)
22:04:18  <isaacs>tjfontaine: interruptable?
22:04:34  <tjfontaine>isaacs: consider test-fs-largefile, and hitting the timeout for the testsuite
22:04:40  <trevnorris>isaacs: ok to dump "json-api-docs"? just says in commit "still needs more cleanup" so want to double check
22:04:40  <isaacs>oic
22:04:52  <isaacs>trevnorris: sure.
22:04:57  <isaacs>trevnorris: that landed a long time ago
22:05:10  <tjfontaine>you cannot kill that process, while writeSync is happening, at least not from process explorer/taskill
22:05:18  <trevnorris>coolio
22:11:33  * AvianFluquit (Remote host closed the connection)
22:27:39  <trevnorris>isaacs: wow. streams2-stdin-wip is one mother of a branch.
22:27:53  <isaacs>trevnorris: hahah
22:27:54  <isaacs>yeah
22:27:59  <isaacs>trevnorris: kill at will
22:28:10  <isaacs>pretty much anything *-wip is probably garbage.
22:28:18  <isaacs>unless it's bene pushed to relatively recently
22:28:20  <isaacs>on joyent/node, anyway
22:28:27  * trevnorrissheds tears of joy
22:28:44  <trevnorris>coolio.
22:28:54  <trevnorris>first starting w/ anything that's been merged.
22:33:49  <trevnorris>isaacs: streams1?
22:34:11  <isaacs>trevnorris: i guess we can remove that now
22:34:23  <isaacs>it's established that we're not going to roll back streams2 :)
22:34:45  <trevnorris>heh
22:36:51  <trevnorris>ah. that feels much better. :)
22:39:53  * bnoordhuisjoined
22:43:05  * pachetquit (Quit: leaving)
22:51:32  <bnoordhuis>we need a faq. i'm getting tired of explaining to people that rss is a bad indicator for memory leaks
22:51:49  <trevnorris>lol. I was just looking at that.
22:52:06  <tjfontaine>even if we put that in the faq I doubt they'd read it
22:52:27  <trevnorris>but at least we could just paste it in the response.
22:53:02  * mikealjoined
22:55:25  <trevnorris>bnoordhuis: wanted to ask. you mentioned there's no need to void* cast a char*, but I found a strange case where passing the char* to MakeWeakCallback for buffers would cause v8 to run out of memory for very large buffers.
22:55:34  <trevnorris>bnoordhuis: it's like the entire buffer was being copied into the v8 heap.
22:56:31  <bnoordhuis>trevnorris: hard to say what that is without looking at the code
22:56:54  <bnoordhuis>which i don't have time for right now on account of going out for beer in a few minutes
22:57:01  <bnoordhuis>not to mention jetlag
22:57:24  <trevnorris>coolio
22:58:21  <tjfontaine>https://github.com/joyent/node/blob/master/test/pummel/test-net-many-clients.js#L40
22:58:35  <tjfontaine>should we still be emitting a 'connect' on the client after we're in a requestListener event?
22:59:03  <tjfontaine>I mean it's implicit that the client is connected by this time, right?
22:59:29  <tjfontaine>that being said, we aren't doing that, and it's broken
23:01:22  <bnoordhuis>tjfontaine: i removed that a while ago
23:01:30  <tjfontaine>so the test just needs updated?
23:01:36  <bnoordhuis>i thought i did
23:01:49  <tjfontaine>I mean, I see no real mention of the behavior
23:02:25  <bnoordhuis>sorry, gotta go
23:02:29  <tjfontaine>enjoy
23:02:34  <bnoordhuis>back in an hour or two (probably. or not)
23:06:55  * bnoordhuisquit (Ping timeout: 258 seconds)
23:08:31  <isaacs>ircretary: tell bnoordhuis Please write up a wiki page on this. Then at least you can paste the link, instead of getting annoyed each time.
23:08:31  <ircretary>isaacs: I'll be sure to tell bnoordhuis
23:10:17  <tjfontaine>(I say this knowing full well I'll have to write it) but we should have an anchor tag for the api wrt resume/pause/data-streams2
23:11:14  <tjfontaine>isaacs: lgty https://github.com/joyent/node/pull/5410 ?
23:11:47  * c4milojoined
23:12:57  <isaacs>tjfontaine: lgtm
23:13:01  <tjfontaine>k
23:14:07  <tjfontaine>I wonder who in jakarta is 36.72.84.210 and why the hell they insist on hammering our jenkins install
23:27:18  * mikealquit (Quit: Leaving.)
23:27:48  * mikealjoined
23:35:15  <MI6>joyent/node: Timothy J Fontaine master * a90dc41 : test: reap children when cluster-bind-twice fails - http://git.io/fNSg7g
23:49:09  <isaacs>hm. so, it seems that this writev problem is actually jsut a streams2 writable problem :)
23:49:39  <isaacs>indutny: w.write('hello', function() { should happen first }); w.on('finish', function() { should happen second }); w.end();
23:49:45  <isaacs>indutny: that order is not guaranteed! ^
23:50:05  <isaacs>indutny: i only noticed because i added a test for writev to addthis cork counter thingie
23:51:25  * hzquit
23:53:19  <isaacs>so, the first question shoudl be: do we care?
23:53:33  <isaacs>anyone care?
23:54:10  <tjfontaine>hm
23:54:56  <tjfontaine>what determines the order in which they happen now? when the buffer gets the actual write out?
23:55:31  <tjfontaine>or rather, how difficult would it be to delay the finish until all outstanding writecbs have been called?
23:55:53  <isaacs>tjfontaine: well, if the write is flushed out synchronously, then the write cb is deferred
23:56:11  <isaacs>tjfontaine: BUT, there's no flag to indicate that we're in the middle of deferring a write cb, so the end()->'finish' event is NOT deferred
23:56:16  <tjfontaine>hm right
23:57:10  <indutny>aha