00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:07  * ircretaryjoined
00:06:55  <isaacs>indutny: https://github.com/joyent/node/pull/5280
00:07:00  <isaacs>sorry for delay, computer is having some troubles
00:07:16  <indutny>wow
00:08:04  <isaacs>tjfontaine: ! it's showing "merge with caution" and stuff when tests fail now!
00:10:01  <tjfontaine>indeed
00:10:09  <tjfontaine>there's not a good set of statues to pick from
00:10:17  <tjfontaine>pending/success/error/failure
00:10:28  <tjfontaine>I picked error for unstable, and failure for actual build failures
00:10:31  <tjfontaine>(or aborts)
00:11:17  <tjfontaine>isaacs: the downside, is that Nodejs-Jenkins has to be a member of the joyent/node and joyent/libuv teams
00:12:25  <isaacs>meh. that's fine
00:12:36  <isaacs>i approve of your choices there. good work :)
00:12:52  <tjfontaine>heh good
00:13:07  <tjfontaine>I'm still unclear if we get a github push for forcepush's
00:16:31  <trevnorris>think I can throw up the new buffers tomorrow.
00:17:23  <trevnorris>indutny: might need your help though. while all tests pass, bench-tls dies.
00:19:53  * brsonjoined
00:29:00  * indexzerojoined
00:30:09  <trevnorris>isaacs: heh. this is funny. I created a SlowBuffer that ties directly to cc for backwards compat. now SlowBuffer's are ~15% faster than buffers.
00:31:04  * trevnorrisquit (Quit: Leaving)
00:43:32  * stagasquit (Quit: Bye)
00:43:45  <isaacs>trevnorris: interesting
00:46:37  <isaacs>ok, so... i'm gonna land this unshift() thing
00:46:50  <isaacs>it prevents a bunch of bugs.
00:47:04  <MI6>joyent/node: isaacs v0.10 * b0de1e4 : stream: Fix unshift() race conditions Fix #5272 The consumption of a re - http://git.io/D6qzLw
00:51:08  <isaacs>TooTallNate: repl pull req from niggler: https://github.com/joyent/node/pull/5274
00:51:27  <isaacs>TooTallNate: i'm reviewing as well, but assigned to you, since you are mr repl
00:56:12  <TooTallNate>isaacs: kewl, i'll take a look
00:56:41  <isaacs>TooTallNate: don't land it, i wanna clean up the commit message.
00:56:49  <isaacs>this is good stuff, though, i think
01:04:49  * bnoordhuisquit (Ping timeout: 258 seconds)
01:06:39  <MI6>nodejs-v0.10: #133 UNSTABLE osx-x64 (1/576) linux-ia32 (1/576) windows-x64 (5/576) windows-ia32 (6/576) http://jenkins.nodejs.org/job/nodejs-v0.10/133/
01:08:39  <isaacs>TooTallNate: that's funny, you had the exact same inline comments as i did
01:09:43  <isaacs>TooTallNate: ohh, i commented on the commit page, you commented on the .../files url, so it doesn't show them together. that's a bit weird.
01:10:20  <TooTallNate>isaacs: ya, wtf
01:10:22  <TooTallNate>github
01:11:28  * abraxasjoined
01:12:01  * inolenquit (Quit: Leaving.)
01:13:32  <isaacs>meh. it's cool :)
01:13:41  <isaacs>still the best online code review tool that isn't a huge pita to use.
01:14:14  <indutny>isaacs: sorry, just got back
01:14:17  <indutny>so... unshift, yeah?
01:16:23  * dapquit (Quit: Leaving.)
01:16:36  <isaacs>indutny: yeah.
01:16:46  <isaacs>indutny: if you strongly object, i can revert, but i figured you were gone :)
01:16:59  <indutny>lets take a look first :)
01:18:45  * c4miloquit (Remote host closed the connection)
01:54:20  * inolenjoined
01:56:42  * brsonquit (Quit: leaving)
02:26:04  * TooTallNatequit (Quit: ["Textual IRC Client: www.textualapp.com"])
02:33:01  <isaacs>indutny: thoughts? https://github.com/joyent/node/pull/5280 was the pr
02:33:10  <isaacs>indutny: or just http://git.io/D6qzLw
02:45:02  * c4milojoined
02:53:03  * c4miloquit (Remote host closed the connection)
03:10:48  * inolen1joined
03:11:43  * inolenquit (Ping timeout: 246 seconds)
03:40:21  <MI6>joyent/node: isaacs v0.10 * 8ee4300 : build: Typo in tools/msvs/msi/product.wxs - http://git.io/czPPSw
04:04:06  * mmaleckijoined
04:45:43  * abraxasquit (Remote host closed the connection)
04:52:57  * inolenjoined
04:53:31  * inolen1quit (Read error: No route to host)
05:00:27  * AvianFluquit (Read error: Connection reset by peer)
05:00:59  * AvianFlujoined
05:24:49  * brsonjoined
05:47:12  * benoitcquit (Excess Flood)
05:52:32  * benoitcjoined
06:02:29  * abraxasjoined
06:25:09  * AvianFluquit (Remote host closed the connection)
06:35:20  * `3rdEdenjoined
06:35:55  * rendarjoined
06:56:23  * brsonquit (Quit: leaving)
07:02:24  * stagasjoined
07:24:58  * bnoordhuisjoined
07:46:43  * hzjoined
07:48:10  * benoitcquit (Changing host)
07:48:11  * benoitcjoined
08:04:02  * bnoordhuisquit (Ping timeout: 252 seconds)
08:12:53  * defunctzombiechanged nick to defunctzombie_zz
08:15:47  * paddybyersquit (Ping timeout: 255 seconds)
08:26:44  * dominictarrjoined
08:30:03  * dominictarrquit (Client Quit)
08:44:43  * dominictarrjoined
08:57:57  * indexzeroquit (Quit: indexzero)
09:10:02  * bnoordhuisjoined
09:14:36  * bnoordhuisquit (Ping timeout: 245 seconds)
09:15:48  * paddybyersjoined
09:52:08  * `3rdEdenchanged nick to `3E|BRB
10:01:53  * loladiroquit (Quit: loladiro)
10:03:41  * abraxasquit (Remote host closed the connection)
10:15:28  * loladirojoined
10:22:08  * squirrel_mjoined
10:22:11  <squirrel_m>hi folks!
10:22:19  <squirrel_m>just wanted to ask a thing about pyuv ...
10:22:45  <squirrel_m>in pyuv.TCP object, there is callback for reading
10:23:01  <squirrel_m>where can I find coumetantion
10:23:11  <squirrel_m>what is returned in "error" and on what case
10:26:55  * loladiroquit (Quit: loladiro)
10:41:25  * `3E|BRBchanged nick to `3E
11:10:13  * piscisaureus_joined
11:13:30  <piscisaureus_>github :(
11:21:39  <MI6>joyent/libuv: Bert Belder master * aa23eea : Merge branch 'v0.10' (+4 more commits) - http://git.io/PtYY2w
11:22:08  * squirrel_mquit (Ping timeout: 245 seconds)
11:33:27  * piscisaureus_quit (Ping timeout: 258 seconds)
12:47:17  * bnoordhuisjoined
12:58:52  * paddybyers_joined
13:02:00  * paddybyersquit (Ping timeout: 258 seconds)
13:02:00  * paddybyers_changed nick to paddybyers
13:08:25  * piscisaureus_joined
13:20:24  * AvianFlujoined
13:23:45  * defunctzombie_zzchanged nick to defunctzombie
13:30:06  * piscisaureus_quit (Ping timeout: 264 seconds)
13:32:31  * piscisaureus_joined
13:57:44  * bradleymeckjoined
14:02:30  * piscisaureus_quit (Ping timeout: 264 seconds)
14:09:05  <saghul>libuv changelog, the proper way: http://starlogs.net/#joyent/libuv
14:18:59  * piscisaureus_joined
14:20:15  <piscisaureus_>bnoordhuis : hey
14:23:45  * stagas_joined
14:26:50  * stagasquit (Ping timeout: 252 seconds)
14:26:55  * stagas_changed nick to stagas
14:33:29  * bradleymeckquit (Quit: bradleymeck)
14:35:11  * bradleymeckjoined
14:38:06  * bradleymeckquit (Client Quit)
14:54:24  * loladirojoined
14:56:16  <indutny>hey bnoordhuis
14:56:18  <indutny>how are you doing?
15:00:31  <bnoordhuis>indutny: ho. i'm in a call atm
15:01:59  <piscisaureus_>ho! me too
15:03:51  <indutny>hoho
15:03:56  <indutny>hey over there
15:03:56  <indutny>o
15:03:57  <indutny>kk
15:04:27  <indutny>piscisaureus_: have you updated libuv in master/v0.10 ?
15:04:30  <indutny>I mean node.js
15:04:46  <piscisaureus_>indutny: no
15:04:52  <indutny>can I?
15:04:57  <indutny>don't you mind? :)
15:04:59  <piscisaureus_>isaacs: you need to push the libuv v0.10.4 tag
15:05:10  <piscisaureus_>indutny: you may upgrade to any released version
15:05:17  <indutny>ook :)
15:05:32  <piscisaureus_>indutny: it could be that someone else already upgraded libuv in node though
15:12:05  * bnoordhuisquit (Ping timeout: 248 seconds)
15:13:53  <indutny>piscisaureus_: https://github.com/joyent/node/pull/5285
15:13:57  <indutny>please help me reviewing
15:16:54  <piscisaureus_>indutny: still in a call
15:17:17  <indutny>ok, anyway that pull request is wrong
15:17:25  <indutny>fixing it
15:25:56  <indutny>http://starlogs.net/#joyent/node
15:32:07  * `3Equit (Remote host closed the connection)
15:35:35  <indutny>ok, I think I'll just push it if tests are passing
15:36:34  <MI6>joyent/node: Fedor Indutny master * 259839f : Merge branch 'v0.10' Conflicts: ChangeLog deps/uv/src/version.c src/n (+33 more commits) - http://git.io/PpuFcg
15:38:10  * bnoordhuisjoined
15:43:47  <bnoordhuis>and back!
15:43:48  <MI6>joyent/node: Fedor Indutny master * 061151c : uv: Upgrade to v0.11.1 - http://git.io/7foY6g
15:44:37  <indutny>good!
15:44:45  <indutny>bnoordhuis: I've merged v0.10 in master and updated libuv
15:44:48  <indutny>going to rebase writev
15:44:54  <indutny>fix some nits
15:45:02  <indutny>and push it if you've no other objections
15:45:05  <indutny>also
15:45:12  <indutny>ah, nvm
15:45:15  <bnoordhuis>well...
15:45:16  <indutny>I'll move that +15 thing
15:45:21  <indutny>and improve net.js
15:45:43  <bnoordhuis>i have an objection that's more philosophical than technical
15:45:55  <bnoordhuis>it's a lot of code, some things get faster, others get slower
15:46:04  <indutny>what got slower?
15:46:17  <bnoordhuis>you posted benchmark numbers, didn't you?
15:46:33  <indutny>yeah, but it was very little degradation on small amount of benchmarks
15:46:44  <indutny>and I believe it can improve lives of many people
15:46:50  <indutny>because chunked encoding is not something uncommon
15:46:57  <indutny>and it works very good with it
15:47:13  <indutny>also, about that += 15 thing
15:47:16  <bnoordhuis>well, hold off on landing it today
15:47:21  <indutny>np
15:47:27  <bnoordhuis>i think at least one more person besides me should review it
15:47:32  <indutny>sure
15:47:35  <indutny>so, about += 15
15:47:42  <indutny>I think its in the right place right now
15:47:49  <indutny>i.e. right before allocation
15:49:25  <indutny>bnoordhuis: can you please tell me what bothers you about it?
15:49:26  <bnoordhuis>indutny: i was thinking more of doing e.g. char* storage = new char[ROUND_UP(storage_size, 16)]
15:49:39  <indutny>that won't work
15:49:43  <bnoordhuis>i wouldn't say i'm bothered by it :)
15:49:56  <indutny>because each string write should happen at aligned address
15:50:08  <bnoordhuis>and?
15:50:14  <indutny>example
15:50:18  <indutny>storage_size = 200
15:50:25  <indutny>writing 5 one-byte strings
15:50:29  <indutny>first one goes at 0
15:50:32  <indutny>second one at 16
15:50:34  <indutny>third one at 32
15:50:41  <indutny>see, its not the same thing as ROUND_UP
15:50:47  <indutny>err, storage_size = 5
15:50:48  <bnoordhuis>oh wait, it's in a loop
15:50:49  <indutny>:)
15:50:52  <indutny>yeah
15:50:53  <bnoordhuis>hadn't spotted that
15:51:25  <bnoordhuis>but then it should be something like storage_size = ROUND_UP(storage_size, 16)
15:51:57  * dapjoined
15:53:54  * bradleymeckjoined
15:56:02  <indutny>bnoordhuis: and that won't work either
15:56:12  <indutny>we don't know actual sizes of strings before copying them :(
15:56:15  <indutny>that's pretty ugly thing
15:56:22  <indutny>ok, gtg
15:56:27  <indutny>time to buy apple stuff
15:56:29  * bradleymeckquit (Client Quit)
15:59:47  * bnoordhuisquit (Ping timeout: 252 seconds)
16:04:52  <MI6>joyent/libuv: isaacs created tag v0.10.4 - http://git.io/bzEOXg
16:05:14  * bradleymeckjoined
16:05:40  <piscisaureus_>isaacs: whoop whoop
16:06:04  <isaacs>indutny: you and bnoordhuis were talking about writev? i'll review it as well.
16:06:28  <tjfontaine>indutny got all surly with jankins
16:08:42  * bradleymeckquit (Client Quit)
16:12:08  * piscisaureus_quit (Ping timeout: 258 seconds)
16:30:33  * bnoordhuisjoined
16:43:46  * bnoordhuisquit (Ping timeout: 245 seconds)
16:48:51  * trevnorrisjoined
16:50:43  <trevnorris>tjfontaine: might be a little late on this, but the jenkins builds are now tied into github?
16:50:59  * inolenquit (Quit: Leaving.)
16:51:01  <tjfontaine>a little, they use the github statuses api
16:51:07  <trevnorris>cool
16:51:22  <tjfontaine>so folks who don't want the chrome plugin can see things
16:51:35  <trevnorris>yeah. just saw a red X next to one of the PR's w/ a link to the jenkins build.
16:51:42  * TooTallNatejoined
17:01:57  * dominictarrquit (Quit: dominictarr)
17:02:01  * brsonjoined
17:03:59  <MI6>nodejs-master: #147 FAILURE osx-x64 (1/575) linux-x64 (3/575) windows-ia32 (6/575) smartos-ia32 (3/575) smartos-x64 (2/575) linux-ia32 (1/575) http://jenkins.nodejs.org/job/nodejs-master/147/
17:06:05  * piscisaureus_joined
17:07:42  <MI6>libuv-master: #75 UNSTABLE linux (4/187) smartos (6/187) windows (7/188) osx (1/187) http://jenkins.nodejs.org/job/libuv-master/75/
17:08:57  * bnoordhuisjoined
17:09:44  * loladiroquit (Quit: loladiro)
17:14:44  <trevnorris>indutny: what about this scenario: pass a flag to SetIndexedPropertiesToExternalArrayData that tells the gc to automatically free the data backing the js object when it's cleaned up?
17:18:43  <trevnorris>hm. though then SetWrapperClassId would need to be extended to support any object that has attached array data, not just persistents.
17:19:37  <isaacs>bnoordhuis: re 5284, if we escape any of those chars, we have to escape all of them, i think. we should be like a browser, or like curl, but not some kind of hybrid.
17:21:23  * qmxchanged nick to qmx|lunch
17:25:51  <trevnorris>bnoordhuis: fyi latest uv in node during compile: "unix/core.c:69:1: warning: zero size arrays are an extension"
17:28:49  <isaacs>bnoordhuis: I failed at reading :) Yes, you are right, there's a bug in url.parse, we should be escaping those things
17:32:43  <bnoordhuis>trevnorris: yeah, that's my STATIC_ASSERT macro
17:32:45  * inolenjoined
17:33:07  <bnoordhuis>i'll rewrite it in the near future to use something else
17:33:49  <bnoordhuis>my initial implementation used typedefs but then gcc 4.8 complains they're unused :-/
17:33:59  <trevnorris>lol, freakin compilers.
17:40:32  <trevnorris>ok, this is annoying. getting a "corrupted double-linked list" from tls-connect.js bench. but doesn't appear in gdb or w/ debug.
17:42:00  <MI6>joyent/libuv: Ben Noordhuis v0.10 * ef85bda : unix: silence STATIC_ASSERT compiler warnings Newer versions of gcc comp - http://git.io/MsEV8w
17:42:09  <bnoordhuis>trevnorris: ^
17:42:12  <trevnorris>=)
17:42:45  <bnoordhuis>trevnorris: re that malloc bug, turn on core dumps and inspect the core file
17:42:51  <bnoordhuis>on os x, it's written to /cores/
17:43:02  <trevnorris>cool. thanks.
17:43:26  <tjfontaine>presuming that directory exists, which it may not
17:43:27  <bnoordhuis>btw, that's `ulimit -c unlimited`, then `gdb /path/to/node /cores/core.etc`
17:43:51  <bnoordhuis>tjfontaine: it exists on my system and i never created it
17:44:12  <bnoordhuis>might be xcode's work though
17:44:29  <tjfontaine>right, I'm not entirely sure what the scenarios are, but I have had encountered systems without it
17:49:48  <trevnorris>ah, found the problem. thanks.
17:50:06  * loladirojoined
17:50:14  * `3rdEdenjoined
18:02:27  * bnoordhu1sjoined
18:02:49  * rendar_joined
18:05:01  * bnoordhuisquit (Ping timeout: 245 seconds)
18:05:01  * rendarquit (Ping timeout: 245 seconds)
18:19:29  * piscisaureus_quit (Ping timeout: 252 seconds)
18:27:25  * bnoordhu1squit (Ping timeout: 256 seconds)
18:27:53  <trevnorris>isaacs: forget what I said about streams. crypto makes me hate my life.
18:31:59  * piscisaureus_joined
18:35:30  <trevnorris>indutny: you poking around?
18:36:30  * dominictarrjoined
18:37:43  * dominictarrquit (Client Quit)
18:43:58  <isaacs>review, somebody? https://github.com/isaacs/node/compare/GH-5284
18:45:50  * dominictarrjoined
18:48:17  <trevnorris>isaacs: code looks good.
18:53:40  * qmx|lunchchanged nick to qmx
19:00:43  <MI6>nodejs-v0.10: #134 ABORTED windows-x64 (4/576) windows-ia32 (8/576) http://jenkins.nodejs.org/job/nodejs-v0.10/134/
19:01:01  <MI6>nodejs-master: #148 ABORTED windows-x64 (9/579) osx-x64 (2/579) linux-x64 (1/579) windows-ia32 (6/579) smartos-ia32 (3/579) smartos-x64 (3/579) osx-ia32 (1/579) linux-ia32 (3/579) http://jenkins.nodejs.org/job/nodejs-master/148/
19:01:20  <MI6>libuv-node-integration: #17 ABORTED smartos-x64 (2/579) smartos-ia32 (1/579) windows-ia32 (8/579) osx-ia32 (2/579) osx-x64 (1/579) windows-x64 (7/579) http://jenkins.nodejs.org/job/libuv-node-integration/17/
19:03:18  <MI6>libuv-v0.10: #35 UNSTABLE smartos (4/186) linux (2/186) osx (1/186) windows (6/187) http://jenkins.nodejs.org/job/libuv-v0.10/35/
19:09:26  * inolenquit (Quit: Leaving.)
19:12:04  <piscisaureus_>tjfontaine: is there a more convenient jenkins dashboard somewhere?
19:13:53  * inolenjoined
19:14:10  <piscisaureus_>tjfontaine: I'd love something like this: http://build.chromium.org/p/client.v8/console
19:17:36  <trevnorris>anyone up to a few crypto questions? I'm getting a sort of race condition w/ freeing memory from SSL_CTX_free
19:17:47  <trevnorris>but i'm not seeing how that's related to my buffer changes.
19:18:47  <MI6>libuv-node-integration: #18 FAILURE linux-ia32 (1/576) windows-ia32 (8/579) osx-x64 (1/576) windows-x64 (6/576) http://jenkins.nodejs.org/job/libuv-node-integration/18/
19:19:43  <piscisaureus_>tjfontaine: half my kingdom if you can give me dashboard like that -> http://build.chromium.org/p/client.v8/console
19:40:22  * bradleymeckjoined
19:42:43  <trevnorris>ah ha! found it. no idea how to fix it, but at least I know where it is.
19:47:09  * brsonquit (Ping timeout: 256 seconds)
19:48:11  <tjfontaine>piscisaureus_: there are some plugins that do that sort of thing, lemme look into it, but unfortunately jenkins is pretty happy with their dashboard
19:49:10  <piscisaureus_>tjfontaine: there is no jenkins "dashboard", that's the problem.
19:49:30  <piscisaureus_>The great thing about the v8 dashboard is that it provides complete overview
19:49:59  <CoverSlide>or maybe you can work some jsp magic
19:50:20  <piscisaureus_>I'm too far away from the heat
19:50:26  <CoverSlide>i mean ... half a kindom ... i'd do it
19:50:32  <tjfontaine>CoverSlide: nah, I have a js framework that sits on top of jenkins that I could do it in
19:50:33  <piscisaureus_>Does it have an API? I could hack something client side
19:50:41  * CoverSlideshrugs
19:50:46  <tjfontaine>piscisaureus_: yup, basically any page append /api/json
19:50:52  <CoverSlide>likely
19:51:03  <piscisaureus_>tjfontaine: is it cors-enabled?
19:51:11  <tjfontaine>piscisaureus_: yes
19:51:17  <piscisaureus_>nice
19:51:18  <tjfontaine>well
19:51:23  <tjfontaine>I haven't had a problem doing it
19:51:24  <tjfontaine>but
19:51:30  <tjfontaine>anything you want to do I can stick in jankins
19:51:46  <tjfontaine>for instance, jenkins.nodejs.org/nightlies
19:52:28  <piscisaureus_>oh you mean I can publish a page there
19:52:57  <tjfontaine>ya, jankins is just a restify server that if it doesn't find the path proxies to jenkins
19:53:13  <tjfontaine>because, fuck writing jenkins java plugins.
19:53:34  * sgallaghquit (Remote host closed the connection)
19:54:05  <CoverSlide>oh wow, that's pretty cool
19:54:11  <piscisaureus_>tjfontaine: right... jankins is really a thing and not a typo ...
19:54:17  <tjfontaine>right
19:54:32  <piscisaureus_>tjfontaine: where is it located?
19:54:33  <tjfontaine>https://github.com/tjfontaine/jankins
19:54:44  <tjfontaine>not really the prettiest, but meh
19:54:48  <piscisaureus_>shrug
19:56:00  <piscisaureus_>tjfontaine: cool. thanks! I'll try to throw something together
19:56:31  <tjfontaine>piscisaureus_: np
20:02:00  * hzquit (Disconnected by services)
20:02:04  * hzjoined
20:09:09  <piscisaureus_>tjfontaine: http://jenkins.nodejs.org/job/libuv-v0.10/tapResults/nodata ?
20:09:51  <tjfontaine>piscisaureus_: when you use a multi-config job you have to ask each configuration for its results
20:10:00  <tjfontaine>http://jenkins.nodejs.org/job/libuv-v0.10/lastCompletedBuild/testReport/
20:10:17  <tjfontaine>http://jenkins.nodejs.org/job/libuv-v0.10/lastCompletedBuild/label=smartos/tapTestReport/
20:10:20  <piscisaureus_>tjfontaine: right.... thanks
20:10:41  <piscisaureus_>tjfontaine: I just realized it's not the chome status board I want. I want one that shows individual tests and their history.
20:11:40  <tjfontaine>piscisaureus_: if you tell me what you want, I will wire something up, and then I'll let you make it pretty :)
20:11:46  <piscisaureus_>kewl
20:13:34  * bradleymeckquit (Quit: bradleymeck)
20:15:37  * brsonjoined
20:19:05  * brsonquit (Client Quit)
20:19:24  * brsonjoined
20:24:32  * indexzerojoined
20:40:20  <trevnorris>anyone know the reason for using SlabBuffer's in lib/tls.js?
20:40:41  <trevnorris>the commit comments don't say anything.
20:48:28  * bnoordhuisjoined
21:00:26  <piscisaureus_>tjfontaine: hmm, the jenkins APIs are not cors enabled ...
21:01:11  <tjfontaine>hm I thought they were, lemme see if there's a plugin I can do to tweak it, but chrome and file:/// should be fine with it
21:02:23  <tjfontaine>I bet I can tweak that with restify
21:02:25  <piscisaureus_>tjfontaine: no, file:/// isn't
21:02:35  <piscisaureus_>tjfontaine: oh, that'd be cool!
21:02:55  <piscisaureus_>tjfontaine: it's true that it responds to OPTIONS but it doesn't set the proper headers
21:03:28  <tjfontaine>I should be able to just proxy the options, and then insert the cors header
21:06:18  <piscisaureus_>tjfontaine: yeah, that sounds not too bad
21:08:21  <tjfontaine>piscisaureus_: I just turned on cors for restify, can you see if that small change enables it for you?
21:09:26  <piscisaureus_>tjfontaine: I'm seeing no difference
21:11:12  <piscisaureus_>tjfontaine: you need to respond to OPTIONS with something that has 'access-control-allow-origin: *'
21:11:34  <trevnorris>isaacs: buffer.offset isn't used anymore. set it to 0 anyways?
21:11:44  <isaacs>trevnorris: sure
21:12:12  <trevnorris>ok. and how about "buffer.parent = buffer"?
21:14:24  <tjfontaine>piscisaureus_: ya, restify isn't catching it if it doesn't know the path, so I need to do it in the proxy
21:16:26  <trevnorris>eh, doesn't hurt perf. full backwards compatibility!
21:18:01  <tjfontaine>piscisaureus_: please try again
21:18:51  * loladiroquit (Quit: loladiro)
21:19:08  <piscisaureus_>tjfontaine: well... yes and no
21:19:19  <piscisaureus_>tjfontaine: it works insofar I see the access-control header
21:19:33  <tjfontaine>but jenkins doesn't do the right thing on OPTIONS
21:19:36  <piscisaureus_>tjfontaine: but my browser isn't digging it. Probably we need to set some more headers.
21:27:21  * loladirojoined
21:28:06  * rendar_quit
21:36:08  <tjfontaine>piscisaureus_: ok, it interrupts on OPTIONS and doesn't proxy, does that make your browser happier?
21:36:29  <tjfontaine>oh case probably matters doesn't it
21:41:16  * loladiroquit (Quit: loladiro)
21:43:41  * loladirojoined
21:59:26  <trevnorris>anyone feeling charitable to help me find a race condition w/ something to do w/ cleaning up memory?
22:00:03  <piscisaureus_>tjfontaine: here's what github sends back: https://gist.github.com/piscisaureus/5375496
22:01:30  * hzquit
22:02:09  <tjfontaine>is that on OPTIONS or GET?
22:02:52  * indexzeroquit (Quit: indexzero)
22:08:20  <piscisaureus_>tjfontaine: the response to OPTIONS
22:08:31  <piscisaureus_>tjfontaine: that's the only thing that matters, really
22:08:53  <tjfontaine>ok well I haven't seen you hit with OPTIONS since my last restart
22:08:59  <tjfontaine>there that time it did
22:09:11  <tjfontaine>CORS
22:09:11  <tjfontaine>{"name":"audit","hostname":"jenkins.nodejs.org","pid":6862,"audit":true,"level":30,"remoteAddress":"38.104.194.102","remotePort":36193,"req":{"method":"OPTIONS","url":"/","headers":{"user-agent":"curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8r ...
22:09:16  <tjfontaine>... zlib/1.2.5","host":"jenkins.nodejs.org","accept":"*/*","origin":"http://example.com","access-control-request-method":"POST","access-control-request-headers":"X-Requested-With"},"httpVersion":"1.1","trailers":{}},"res":{"statusCode":200,"headers":{"access-control-allow-origin":"*"},"trailer":false},"latency":2,"_audit":true,"msg":"handled: 200","time":"2013-04-12T22:08:28.916Z","v":0}
22:09:22  <tjfontaine>oh that was me I guess
22:14:03  * AvianFluquit (Read error: Connection reset by peer)
22:14:36  * AvianFlujoined
22:14:50  <tjfontaine>there you were
22:15:00  <piscisaureus_>tjfontaine: yup, works!
22:15:04  <piscisaureus_>tjfontaine: thanks
22:15:07  <tjfontaine>k
22:26:05  <trevnorris>bnoordhuis: need a tip. the bt from gdb vary, strace and --debug run too slowly to have the problem. I can't find the source of this memory issue.
22:26:34  <trevnorris>it seems there's a race condition between crypto trying to access the memory, and v8 gc cleaning it up.
22:26:43  <tjfontaine>trevnorris: try to force it by adding some gc()'s?
22:26:56  <trevnorris>tjfontaine: ah. good idea.
22:27:19  <tjfontaine>trevnorris: don't forget that as you find these it would be helpful to have a test for your code for it :)
22:27:46  <trevnorris>tjfontaine: i'll add a test as soon as I figure out the problem. =)
22:27:59  <trevnorris>right now it just pops up when running benchmark/tls/tls-connect.js
22:28:41  <tjfontaine>is today friday?
22:28:55  <trevnorris>yeah
22:29:20  <trevnorris>tjfontaine: ok. so that's not it. had it gc on every loop and doesn't happen.
22:29:58  <tjfontaine>try it on 100 or 1000 loops
22:30:57  <tjfontaine>piscisaureus_: you know you can get more data in one query by adding "?depth="
22:31:08  <piscisaureus_>tjfontaine: no, I didn't
22:31:25  <piscisaureus_>tjfontaine: actually, I'll probably add ?depth=999
22:31:41  <tjfontaine>piscisaureus_: there's depth and tree, where tree lets you filter on the data you actually want
22:34:17  <piscisaureus_>tjfontaine: depth=20 produces tons of data for me
22:34:32  <piscisaureus_>tjfontaine: chrome just crashted the tab after getting 100MB
22:34:36  <trevnorris>tjfontaine: that's working better. though I've analyzed 3 core dumps, and each backtrace was different...
22:34:39  <tjfontaine>piscisaureus_: ya, it's a lot of data
22:34:53  <trevnorris>this is driving me insane. i'm about to go kill a puppy.
22:34:54  <tjfontaine>piscisaureus_: which is why I mentioned tree parameter as well
22:35:11  <piscisaureus_>tjfontaine: sure, but I need to see what the data looks like first :)
22:35:44  <tjfontaine>hehe, indeed
22:36:07  <tjfontaine>if you're just inspecting in your browser as well, you can do ?pretty=true
22:36:36  <tjfontaine>the depth you're looking for though is probably only 3 or 4 levels deep
22:36:56  <tjfontaine>it's still a crap load of data
22:44:59  <MI6>joyent/node: isaacs master * 17a379e : url: Escape all unwise characters This makes node's http URL handling lo - http://git.io/e-GpKg
22:47:24  * `3rdEdenquit (Quit: Zzz)
23:13:43  <tjfontaine>bnoordhuis: are you around?
23:19:53  * loladiroquit (Quit: loladiro)
23:20:58  * paddybyersquit (Ping timeout: 246 seconds)
23:23:01  <piscisaureus_>tjfontaine: did I just crash your server?
23:23:52  <tjfontaine>everything seems kosher here
23:24:31  <piscisaureus_>http://jenkins.nodejs.org/job/libuv-v0.10/13/label=windows
23:25:00  * stagasquit (Read error: Connection reset by peer)
23:25:05  <tjfontaine>who knows what's up with that build
23:25:48  <piscisaureus_>seems that no detail pages are responding
23:26:06  <tjfontaine>I did see you hammer everything
23:26:40  <tjfontaine>so, did you know jenkins has no db? so nearly every request you do results in java hitting disk opening file interpreting it and sending it back to you
23:26:51  <piscisaureus_>ah
23:26:55  <piscisaureus_>that sucks
23:26:56  <piscisaureus_>:)
23:26:58  <tjfontaine>yes
23:27:09  <piscisaureus_>http://jenkins.nodejs.org/job/libuv-v0.10/27/label=osx/
23:27:18  <tjfontaine>ya I see everything is hanging
23:27:27  <tjfontaine>there we go
23:27:41  <piscisaureus_>tjfontaine: kewl
23:28:22  <bnoordhuis>tjfontaine: i'm here
23:28:51  <bnoordhuis>trevnorris: still having that issue?
23:28:53  <tjfontaine>bnoordhuis: so fs_event_close_in_callback hangs on smartos, and I'm a little confused as to why
23:29:26  <bnoordhuis>tjfontaine: hangs how exactly?
23:29:26  <tjfontaine>(on smartos)
23:29:39  <tjfontaine>port_getn is waiting for events, the cb never fires
23:30:20  <bnoordhuis>what does the dtruss output look like?
23:30:24  * loladirojoined
23:30:38  <bnoordhuis>you can run the test like this: sudo dtruss out/Debug/run-tests fs_event_close_in_callback fs_event_close_in_callback
23:30:58  <trevnorris>bnoordhuis: yeah
23:31:15  <bnoordhuis>trevnorris: is it reproducible?
23:31:27  <bnoordhuis>reproducible in the sense that it always happens at the same point, i mean
23:32:21  <tjfontaine>bnoordhuis: https://gist.github.com/tjfontaine/5376017
23:34:33  <trevnorris>bnoordhuis: sort of. right now it happens during SSL_CTX_free, in ~SecureContext
23:34:54  <trevnorris>bnoordhuis: but if I comment that out then it happens somewhere else.
23:36:40  <MI6>joyent/node: Ben Noordhuis master * 31d0d5a : test: add extra checks (+2 more commits) - http://git.io/i2lDfQ
23:38:56  <trevnorris>bnoordhuis: yeah. here's a bt w/ and w/o SSL_CTX_free: https://gist.github.com/trevnorris/5376042
23:39:23  <trevnorris>so they both go back to WeakCallback.
23:40:40  <tjfontaine>one thing I meant to look into was how you do slice
23:41:39  <trevnorris>tjfontaine: https://github.com/trevnorris/node/blob/buffer-buffet/lib/buffer.js#L311-L342
23:41:44  <bnoordhuis>tjfontaine: odd. the file watcher port is created. it's also added to the main port
23:42:05  <tjfontaine>trevnorris: ah ok
23:42:09  <bnoordhuis>has this test started failing recently or?
23:42:24  <tjfontaine>bnoordhuis: no I think this has been failing for a while, lemme verify
23:43:00  <tjfontaine>http://jenkins.nodejs.org/job/libuv-master/48/label=smartos/tapTestReport/ happened on march 12th
23:43:27  <tjfontaine>so a while, I'm fairly sure it's been happening since I've been testing
23:44:05  <tjfontaine>ya all the way back to my intially builds in jenkins
23:45:23  <bnoordhuis>tjfontaine: it's possible it's failing for the same reason it fails on kqueue-based platforms
23:45:28  <bnoordhuis>(until i disabled it, that is)
23:45:48  <tjfontaine>right, and if that's what needs to be done we can do that
23:50:47  * btraskjoined
23:51:44  <bnoordhuis>tjfontaine: hm, looking at it again, i'm not sure that's actually it
23:51:58  <bnoordhuis>the PORT_SOURCE_FILE gets created when you call uv_fs_event_init()
23:52:14  * loladiroquit (Quit: loladiro)
23:52:24  <bnoordhuis>if you open an issue, i'll look into it next week
23:52:36  <tjfontaine>ok
23:52:41  <tjfontaine>thanks
23:53:43  <bnoordhuis>don't say thanks just yet, i'll feel bad when i close it as WONTFIX :)
23:54:02  <tjfontaine>no, it won't be WONTFIX it will be DISABLEDTEST :)
23:54:18  <bnoordhuis>trevnorris: i wager you're doing an actual double free
23:54:45  <bnoordhuis>trevnorris: add some printfs before the code that calls free / delete[] that print the address
23:55:05  <bnoordhuis>i.e. printf("freeing %p\n", (void*) ptr);
23:55:49  <bnoordhuis>and maybe also after the call to malloc / new[] so you can correlate addresses
23:56:06  <bnoordhuis>and exclude addresses that have been reused after a previous call to free
23:57:29  <tjfontaine>joyent/libuv#776
23:57:31  <trevnorris>bnoordhuis: ah, ok. thanks.
23:57:47  * defunctzombiechanged nick to defunctzombie_zz