00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:08  * ircretaryjoined
00:00:10  * euoiaquit (Ping timeout: 244 seconds)
00:00:18  * dap_quit (Read error: Connection reset by peer)
00:00:24  * dap_1joined
00:00:38  * xaqjoined
00:00:50  * AvianFluquit (Ping timeout: 264 seconds)
00:03:37  * brsonjoined
00:09:31  * xaqquit (Remote host closed the connection)
00:14:47  * mikealquit (Quit: Leaving.)
00:15:30  * mikealjoined
00:20:50  * dap_1quit (Read error: Connection reset by peer)
00:20:52  * dap_joined
00:21:32  * thlorenzquit (Remote host closed the connection)
00:27:35  * mikealquit (Quit: Leaving.)
00:28:01  <tjfontaine>TooTallNate: so what if at the top of _http_client you add Agent.prototype.request = function () { new ClientRequest({ agent: this }) }; sort of thing?
00:28:32  <TooTallNate>tjfontaine: for backwards compat, or what?
00:28:37  <TooTallNate>that brings back the circular dependency
00:29:42  * mikealjoined
00:30:17  <TooTallNate>tjfontaine: IMO Agent shouldn't have any awareness of the ClientRequest... it should just expose an API
00:30:31  <TooTallNate>tjfontaine: and if Agent#request() *is* going to be a thing, we should document it
00:32:42  * zz_karupanerurachanged nick to karupanerura
00:33:00  <tjfontaine>TooTallNate: well it is offhandedly documented, taht certainly was a deficiency in the preivous landing of this feature, in general though I like the concept of being able to issue a request directly on the Agent -- not life or death though
00:33:47  <TooTallNate>tjfontaine: i don't think it's that valuable of a feature personally, and just introduces complexity in the code
00:35:24  * dap_quit (Read error: Connection reset by peer)
00:35:27  * dap_1joined
00:36:25  * hzquit
00:37:05  * xaqjoined
00:39:42  * euoiajoined
00:41:00  * rvaggquit (Quit: ta ta)
00:51:09  * m76quit (Read error: Connection reset by peer)
00:51:19  * wolfeidauquit
00:51:28  * wolfeidaujoined
00:55:47  * AvianFlujoined
01:00:09  * mikolalysenkoquit (Ping timeout: 272 seconds)
01:01:06  * kevinsimperjoined
01:01:29  * kazuponjoined
01:04:47  <TooTallNate>tjfontaine: this one should be less controversial :) https://github.com/joyent/node/pull/7190
01:04:59  <TooTallNate>of course jenkins disagrees: p
01:05:07  * dap_1quit (Read error: Connection reset by peer)
01:06:29  * kevinsimperquit (Ping timeout: 272 seconds)
01:09:00  * dap_joined
01:14:04  * dap_quit (Quit: Leaving.)
01:15:59  <swaj>is child_process.execSync() going to happen for sure? noticed it in the source, but no docs yet
01:16:49  <swaj>oh it's int he master docs, nvm, ignore me :P
01:17:57  * mikealquit (Quit: Leaving.)
01:18:52  * mikealjoined
01:19:36  <TooTallNate>swaj: ya, it's in v0.11.x finally
01:21:03  * Ralithquit (Ping timeout: 272 seconds)
01:23:56  <tjfontaine>swaj: :)
01:25:24  * xaqquit (Remote host closed the connection)
01:25:52  <tjfontaine>TooTallNate: man, talk about undocumented features, createConnection is no where to be seen in our http documentation
01:26:38  <TooTallNate>tjfontaine: oh geez. i didn't check. I assumed it was :p
01:26:45  <TooTallNate>tjfontaine: i can add some docs
01:26:55  <TooTallNate>it's been around for a while so probably not a bad idea
01:27:07  <tjfontaine>I went to see wtf the semantics of that even were :)
01:27:22  * mikolalysenkojoined
01:29:47  * bajtosjoined
01:31:06  * kazuponquit (Ping timeout: 244 seconds)
01:34:09  * mikealquit (Quit: Leaving.)
01:36:16  * kazuponjoined
01:36:29  * bajtosquit (Quit: bajtos)
01:38:53  <TooTallNate>tjfontaine: fwiw, that patch can probably be backported to v0.10 to fix the fact that you must pass `agent: false` in order for createConnection to be honored
01:39:03  <tjfontaine>haha
01:39:14  <tjfontaine>I'm still not even sure why it exists?
01:39:36  <TooTallNate>ya... i'd just use a custom Agent personally
01:39:47  <TooTallNate>or https://github.com/TooTallNate/node-agent-base
01:39:54  <tjfontaine>I think it is backwards compat for sure, the functionality has been subsumed by Agent
01:39:57  <tjfontaine>is my guess
01:40:08  <TooTallNate>ya, that's probably what happened
01:40:12  <TooTallNate>like back in the v0.3.x days
01:40:12  <TooTallNate>hahah
01:41:27  * abraxasjoined
01:41:32  <swaj>tjfontaine: is node-addon-layer baked in 0.12?
01:41:59  <tjfontaine>swaj: no, it didn't make the cut, too many things landed on my plate too quickly, but it will work with 0.12 or my name isn't Inigo Montoya
01:42:25  <swaj>tjfontaine: haha, okay, fair enough. I'm gonna make you fix it on RHEL then :P
01:42:27  <tjfontaine>TooTallNate: mikeals fault, https://github.com/joyent/node/commit/103990b64043b980cf98feefcae12019ae6e3108
01:42:36  <tjfontaine>swaj: I need to fix it for everyone :)
01:42:45  <swaj>it works on OSX FWIW
01:42:51  <tjfontaine>and sunos
01:42:55  <tjfontaine>it's an rtld difference
01:42:55  <TooTallNate>tjfontaine: lol :)
01:42:59  <swaj>but my prod servers are mad at it
01:43:16  <TooTallNate>tjfontaine: so yup, precedes http.Agent, haha
01:43:28  <TooTallNate>http2.js, haha
01:43:30  <TooTallNate>i forgot about that
01:43:37  <tjfontaine>yes, anyone with code running like this has been lucky to survive this long
01:43:38  <tjfontaine>(frankly)
01:43:52  <tjfontaine>which is why `{ agent: false }` is even a thing
01:44:33  <tjfontaine>so that being said, I now understand where this horrible thing came from
01:44:48  <TooTallNate>tjfontaine: so what's the call then... deprecate? or document+fix?
01:45:20  <tjfontaine>TooTallNate: so my main question is this, if I supply both agent and createConnection should creatConnection actually take precedence -- I am relatively against that
01:45:36  <tjfontaine>TooTallNate: mostly because agent was introduced to replace that feature
01:45:52  <TooTallNate>tjfontaine: ya good point... i think i'm on the same page
01:46:12  <TooTallNate>tjfontaine: however... agent: false, createConnection: fn *does* invoke the fn on v0.10
01:46:17  <TooTallNate>so we may want to maintain that
01:46:25  <tjfontaine>yes, that's fine, it's a very very explicit path
01:46:30  <TooTallNate>(only the falsey agent case I mean)
01:46:38  <TooTallNate>yup
01:47:08  <tjfontaine>well undefined is a special case, it uses the default agent
01:47:33  * `3rdEdenquit (Quit: Connection closed for inactivity)
01:49:41  * Ralithjoined
01:51:50  <swaj>tjfontaine: so I'm trying to get the right people to tell me yes, but there has been positive reception so far about us contributing to node's core crypto module.
01:52:28  <tjfontaine>swaj: that's excellent, I'm hoping that I will also be able to help convince, I think it's a fantastic story
01:52:37  <tjfontaine>it's the model i would like to follow for other features/subsystems
01:52:44  <swaj>they tend to appreciate being good patrons, and upstreaming stuff is good for future maintainability anyway.
01:53:07  <tjfontaine>companies with resources and interest, who are having particular pain points, should feel like their is a path forward for fixing and maintaining those subsystems
01:53:18  <tjfontaine>for instance the guy from yahoo who loves the debugger and its repl
01:53:40  <tjfontaine>I have no love for that, and if someone does love it, I'm happy to work with them
01:53:57  <swaj>yeah, crypto is painful, but enjoyable for geeks like me :P
01:54:25  <tjfontaine>yes, also your company is well suited to be someone who validates and verifies changes
01:55:01  <tjfontaine>it should be that everyone can feel like they're an owner and stake holder in Node.js -- it's one of our key attributes
01:55:45  <swaj>we actually have a pretty solid history of doing it for other projects too. You'd might be surprised at how many security features in JBoss that we either sponsored or contributed.
01:56:01  <tjfontaine>it makes total sense.
01:56:04  <swaj>so I'm trying to lean on that history :)
01:56:33  <tjfontaine>those who have a vested interest in the trustworthiness of the code should be those the most interested in making sure it meets their standards
01:57:53  <swaj>we also have a core contributor to the apache web server who we employ :P
01:58:23  <swaj>he's pretty damn awesome, too.
01:59:27  <tjfontaine>:)
02:03:37  * toothrotquit (Read error: Connection reset by peer)
02:04:49  * nickleefly___quit (Quit: Connection closed for inactivity)
02:07:08  * toothrjoined
02:16:50  <TooTallNate>tjfontaine: how many failing tests are "normal" on v0.10 branch?
02:16:51  * AlexisMochaquit (Read error: Connection reset by peer)
02:22:46  * thlorenzjoined
02:22:55  * paulfryzelquit (Remote host closed the connection)
02:28:18  <tjfontaine>TooTallNate: depends, generally 3
02:28:34  <TooTallNate>tjfontaine: yup... 3 :) so i think i didn't break anything :p
02:28:40  <tjfontaine>:)
02:28:51  <tjfontaine>are they these http://jenkins.nodejs.org/job/node-pullrequest/DESTCPU=ia32,label=osx/2053/tapTestReport/
02:29:11  <TooTallNate>tjfontaine: only the last one actually
02:29:24  <TooTallNate>the other two were different... but i cleared the screen already
02:29:32  <tjfontaine>anyway that build is for your PR
02:29:43  <TooTallNate>oh
02:29:59  <tjfontaine>http://jenkins.nodejs.org/job/node-pullrequest/2053/ is all yours
02:30:06  <tjfontaine>if you're in #node-ci you can see bot spew about it
02:30:33  <tjfontaine>maybe I'll add a parameter that includes the owner so if they're in there it highlights you
02:30:35  <TooTallNate>i mean none of them are http-related it seems like
02:30:43  <tjfontaine>not really no
02:31:12  <tjfontaine>net-listen-fd0 is just a broken test actually
02:31:20  <tjfontaine>I just need to backport the fix
02:31:31  <tjfontaine>the tls sessino one fedor looked at briefly
02:32:36  <TooTallNate>nod
02:32:38  * brsonquit (Ping timeout: 264 seconds)
02:32:57  <tjfontaine>our master tests are in far better shape and more trustworthy
02:33:05  <TooTallNate>:)
02:33:10  <tjfontaine>we only have on flakey one there, cluster-disconnect
02:33:23  <tjfontaine>I've even enabled gc and internet by default in the CI
02:33:28  * TooTallNatequit (Quit: Computer has gone to sleep.)
02:33:47  <swaj>I feel sorry for you... supporting centos-5.7
02:34:01  <tjfontaine>ya well, I need an ancient glibc
02:34:29  <tjfontaine>but I had to build a whole new toolchain for it because v8 uses c++ features that older linkers drop out
02:40:13  * euoiaquit (Ping timeout: 272 seconds)
02:43:19  * euoiajoined
02:50:45  * AWintermanquit (Remote host closed the connection)
02:54:05  * mikealjoined
03:00:23  * jmar777_quit (Remote host closed the connection)
03:01:31  * mikealquit (Quit: Leaving.)
03:02:05  * kevinsimperjoined
03:06:19  * kevinsimperquit (Ping timeout: 246 seconds)
03:21:23  * kazuponquit (Ping timeout: 272 seconds)
03:45:43  * TooTallNatejoined
03:56:51  * mikolalysenkoquit (Ping timeout: 272 seconds)
04:00:28  * mikolalysenkojoined
04:01:11  * AWintermanjoined
04:01:51  * TooTallNatequit (Quit: Computer has gone to sleep.)
04:02:27  * kazuponjoined
04:06:37  * AWintermanquit (Ping timeout: 244 seconds)
04:06:48  * kazuponquit (Ping timeout: 251 seconds)
04:32:25  * kazuponjoined
04:39:27  * abraxasquit (Remote host closed the connection)
04:40:50  * euoiaquit (Ping timeout: 240 seconds)
04:42:10  * kazupon_joined
04:42:34  * kazuponquit (Read error: Connection reset by peer)
04:53:26  * mikolalysenkoquit (Ping timeout: 246 seconds)
05:02:48  * kevinsimperjoined
05:06:09  * abraxasjoined
05:07:04  * kevinsimperquit (Ping timeout: 244 seconds)
05:07:29  * thlorenzquit (Remote host closed the connection)
05:08:04  * thlorenzjoined
05:08:57  * mikolalysenkojoined
05:09:05  * AvianFluquit
05:09:16  * AvianFlujoined
05:10:13  <andrewrk>is there a race condition free way to do a rename on posix systems? e.g. the 'wx' flags as in fopen, but for rename
05:11:21  * mikealjoined
05:12:11  <andrewrk>aha, I think what I want is to link and then unlink
05:12:20  * thlorenzquit (Ping timeout: 246 seconds)
05:15:55  * calvinfoquit (Quit: Leaving.)
05:53:24  * calvinfojoined
06:08:20  * TooTallNatejoined
06:22:32  * TooTallNatequit (Quit: ["Textual IRC Client: www.textualapp.com"])
06:26:27  * xaqjoined
06:32:13  * calvinfoquit (Quit: Leaving.)
06:32:59  * sinclair|workjoined
06:34:18  * calvinfojoined
06:42:39  * toothrquit (Ping timeout: 244 seconds)
06:57:46  * guilleiguaranquit (Quit: Connection closed for inactivity)
07:00:19  * xaqquit (Remote host closed the connection)
07:01:39  * kevinsimperjoined
07:06:13  * kevinsimperquit (Ping timeout: 272 seconds)
07:12:04  * seldo_quit (Quit: Leaving...)
07:16:26  * mikolalysenkoquit (Ping timeout: 264 seconds)
07:19:30  * brsonjoined
07:24:54  * paulfryzeljoined
07:29:39  * paulfryzelquit (Ping timeout: 272 seconds)
07:39:14  * brsonquit (Ping timeout: 264 seconds)
07:43:42  * kevinsimperjoined
07:53:35  * dshaw_joined
08:11:53  * mikolalysenkojoined
08:16:41  * mikolalysenkoquit (Ping timeout: 244 seconds)
08:23:32  <MI6>joyent/libuv: Saúl Ibarra Corretgé master * b055538 : unix, windows: clarify what uv_stream_set_blocking does - http://git.io/yJmavw
08:26:01  <MI6>joyent/libuv: Saúl Ibarra Corretgé master * b2341e7 : windows: fix typo - http://git.io/cSF5pg
08:26:07  <saghul_>FML :-/
08:26:08  <LOUDBOT_>COCK LIGHT COCK BRIGHT FIRST COCK I SEE TONIGHT
08:33:13  * m76joined
08:43:21  * calvinfoquit (Quit: Leaving.)
08:47:27  * abraxasquit (Remote host closed the connection)
08:51:38  * `3rdEdenjoined
08:52:02  * `3rdEdenchanged nick to Guest92445
08:56:20  * abraxasjoined
09:02:23  * Guest92445changed nick to V1
09:02:36  * V1changed nick to `3rdEden
09:12:37  * mikolalysenkojoined
09:13:53  * calvinfojoined
09:14:08  * calvinfoquit (Read error: Connection reset by peer)
09:14:21  * calvinfojoined
09:15:19  * calvinfo1joined
09:15:19  * calvinfoquit (Read error: Connection reset by peer)
09:16:08  * calvinfojoined
09:16:08  * calvinfo1quit (Read error: Connection reset by peer)
09:17:05  * calvinfo1joined
09:17:06  * calvinfoquit (Read error: Connection reset by peer)
09:17:38  * mikolalysenkoquit (Ping timeout: 264 seconds)
09:18:00  * calvinfojoined
09:18:00  * calvinfo1quit (Read error: Connection reset by peer)
09:18:05  * Kakerajoined
09:22:14  * calvinfoquit (Ping timeout: 246 seconds)
09:24:50  * daviddiasjoined
09:26:40  * paulfryzeljoined
09:28:02  * abraxasquit (Remote host closed the connection)
09:31:15  * paulfryzelquit (Ping timeout: 272 seconds)
09:31:41  * abraxasjoined
09:34:58  * abraxasquit (Remote host closed the connection)
09:41:55  * rendarjoined
09:42:17  * kazupon_quit (Remote host closed the connection)
09:42:43  * kazuponjoined
09:43:18  * janjongboomjoined
09:44:03  * calvinfojoined
09:45:48  * karupanerurachanged nick to zz_karupanerura
09:47:06  * kazuponquit (Ping timeout: 244 seconds)
09:48:40  * hzjoined
09:48:59  * calvinfoquit (Ping timeout: 272 seconds)
10:08:34  <MI6>joyent/libuv: indutny created branch fix/gh-1128 - http://git.io/ydjAYw
10:08:38  <indutny>oi
10:11:51  * mikolalysenkojoined
10:16:50  * mikolalysenkoquit (Ping timeout: 246 seconds)
10:27:24  * paulfryzeljoined
10:32:03  * paulfryzelquit (Ping timeout: 272 seconds)
10:44:28  * janjongboomquit (Ping timeout: 246 seconds)
10:44:47  * calvinfojoined
10:45:47  * calvinfo1joined
10:45:47  * calvinfoquit (Read error: Connection reset by peer)
10:46:40  * calvinfojoined
10:46:40  * calvinfo1quit (Read error: Connection reset by peer)
10:47:22  * calvinfoquit (Read error: Connection reset by peer)
10:47:38  * calvinfojoined
10:51:32  * kazuponjoined
10:51:50  * calvinfoquit (Ping timeout: 246 seconds)
10:59:47  * motleyjoined
11:01:10  * benviequit (Read error: Connection reset by peer)
11:01:29  * benviejoined
11:12:36  * mikolalysenkojoined
11:17:22  * mikolalysenkoquit (Ping timeout: 246 seconds)
11:28:03  * paulfryzeljoined
11:32:51  * paulfryzelquit (Ping timeout: 272 seconds)
11:32:58  <saghul_>oi!
11:33:09  <saghul_>indutny: does https://github.com/joyent/libuv/pull/1156 lgty?
11:35:53  * abraxasjoined
11:41:02  * abraxasquit (Ping timeout: 264 seconds)
11:43:02  * kazuponquit (Remote host closed the connection)
11:43:30  * kazuponjoined
11:48:14  * kazuponquit (Ping timeout: 264 seconds)
11:48:20  * calvinfojoined
11:49:52  * daviddiasquit (Read error: Connection reset by peer)
11:51:28  * lgierthjoined
11:52:44  * calvinfoquit (Ping timeout: 246 seconds)
12:01:39  * kazuponjoined
12:04:42  * euoiajoined
12:05:04  * dshaw_quit (Quit: Leaving.)
12:06:44  * stagasquit (Ping timeout: 245 seconds)
12:07:59  * kazuponquit (Remote host closed the connection)
12:08:26  * kazuponjoined
12:09:41  * daviddiasjoined
12:12:50  * kazuponquit (Ping timeout: 264 seconds)
12:14:26  * calvinfojoined
12:15:17  * calvinfoquit (Read error: Connection reset by peer)
12:15:24  * calvinfo1joined
12:16:19  * calvinfojoined
12:16:19  * calvinfo1quit (Read error: Connection reset by peer)
12:17:14  * calvinfo1joined
12:17:14  * calvinfoquit (Read error: Connection reset by peer)
12:18:08  * calvinfojoined
12:18:08  * calvinfo1quit (Read error: Connection reset by peer)
12:18:50  * sinclair|workquit (Remote host closed the connection)
12:18:50  * calvinfoquit (Read error: Connection reset by peer)
12:19:06  * calvinfojoined
12:20:02  * calvinfo1joined
12:20:02  * calvinfoquit (Read error: Connection reset by peer)
12:20:17  * daviddiasquit (Read error: Connection reset by peer)
12:20:57  * calvinfojoined
12:20:57  * calvinfo1quit (Read error: Connection reset by peer)
12:21:50  * calvinfo1joined
12:21:51  * calvinfoquit (Read error: Connection reset by peer)
12:22:06  * Kakeraquit (Ping timeout: 244 seconds)
12:22:26  * euoiaquit (Ping timeout: 264 seconds)
12:22:40  * calvinfo1quit (Read error: Connection reset by peer)
12:22:48  * hzquit
12:22:49  * calvinfojoined
12:23:42  * calvinfo1joined
12:23:43  * calvinfoquit (Read error: Connection reset by peer)
12:24:40  * calvinfojoined
12:24:41  * calvinfo1quit (Read error: Connection reset by peer)
12:25:26  * calvinfoquit (Read error: Connection reset by peer)
12:25:39  * calvinfojoined
12:30:03  * calvinfoquit (Ping timeout: 257 seconds)
12:31:29  * xaqjoined
12:35:40  * dshaw_joined
12:38:12  <indutny>yes
12:38:13  <indutny>I assume it builds now :P
12:38:16  <indutny>and tests are passing
12:44:48  * dshaw_quit (Ping timeout: 240 seconds)
12:50:13  * daviddiasjoined
12:55:22  * petka_joined
13:00:16  * kpdeckerjoined
13:03:02  * janjongboomjoined
13:17:29  <MI6>joyent/node: orangemocha created branch orangemocha-testing2 - http://git.io/PynX7g
13:18:52  * kazuponjoined
13:23:04  * kazuponquit (Ping timeout: 244 seconds)
13:26:20  * calvinfojoined
13:27:16  * calvinfo1joined
13:27:16  * calvinfoquit (Read error: Connection reset by peer)
13:28:14  * calvinfojoined
13:28:14  * calvinfo1quit (Read error: Connection reset by peer)
13:29:06  * calvinfo1joined
13:29:06  * calvinfoquit (Read error: Connection reset by peer)
13:33:24  <MI6>joyent/node: Maxwell Krohn master * a22a2d8 : tls: stop NodeBIO::Gets from reading off end of buffer - http://git.io/A8lt5g
13:33:32  * calvinfo1quit (Ping timeout: 245 seconds)
13:35:28  * kevinsimperquit (Remote host closed the connection)
13:36:42  * abraxasjoined
13:42:03  * abraxasquit (Ping timeout: 272 seconds)
13:42:13  * dshaw_joined
13:45:59  * daviddiasquit (Remote host closed the connection)
13:46:36  * dshaw_quit (Ping timeout: 245 seconds)
13:49:28  * kevinsimperjoined
13:53:56  * dshaw_joined
13:58:00  * dshaw_quit (Ping timeout: 240 seconds)
14:14:27  * AlexisMochajoined
14:14:41  <AlexisMocha>tjfontaine: https://github.com/joyent/node/pull/7196
14:14:55  * janjongboomquit (Ping timeout: 253 seconds)
14:14:58  <AlexisMocha>I tested it and it works like a charm
14:16:46  * guybrushquit (Excess Flood)
14:16:52  * guybrushjoined
14:16:54  * rosskjoined
14:19:58  * janjongboomjoined
14:20:58  * m76quit (Read error: Connection reset by peer)
14:25:00  * lgierthquit (Quit: Ex-Chat)
14:25:35  * euoiajoined
14:29:52  * calvinfojoined
14:30:19  * paulfryzeljoined
14:30:50  * calvinfo1joined
14:30:51  * calvinfoquit (Read error: Connection reset by peer)
14:31:45  * calvinfojoined
14:31:45  * calvinfo1quit (Read error: Connection reset by peer)
14:32:42  * calvinfo1joined
14:32:42  * calvinfoquit (Read error: Connection reset by peer)
14:32:46  * mikolalysenkojoined
14:32:59  * Kakerajoined
14:33:20  * calvinfo1quit (Read error: Connection reset by peer)
14:33:35  * calvinfojoined
14:34:32  * calvinfo1joined
14:34:32  * calvinfoquit (Read error: Connection reset by peer)
14:35:14  * jmar777joined
14:35:14  * calvinfo1quit (Read error: Connection reset by peer)
14:35:15  * paulfryzelquit (Ping timeout: 272 seconds)
14:35:23  * xaqquit (Remote host closed the connection)
14:35:25  * calvinfojoined
14:36:20  * calvinfo1joined
14:36:20  * calvinfoquit (Read error: Connection reset by peer)
14:37:21  * calvinfojoined
14:37:22  * calvinfo1quit (Read error: Connection reset by peer)
14:37:59  * calvinfoquit (Read error: Connection reset by peer)
14:38:12  * calvinfojoined
14:39:53  * indexzerojoined
14:39:53  * calvinfoquit (Read error: Connection reset by peer)
14:40:02  * calvinfojoined
14:40:34  <AlexisMocha>tjfontaine: regarding vm-infinite-recursion: between https://github.com/joyent/node/pull/7197 and your suggested fix (just checking a failure exit code), I am hoping we can land one :) The v8 reviewer is asking for a test and a couple nits but seemed generally ok with the fix.
14:44:11  * calvinfoquit (Ping timeout: 244 seconds)
14:44:36  * thlorenzjoined
14:45:19  * janjongboomquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:46:30  * rosskquit
14:48:54  * daviddiasjoined
14:51:17  * daviddiasquit (Read error: No route to host)
14:51:47  * daviddiasjoined
14:54:05  * daviddia_joined
14:54:40  * dshaw_joined
14:56:11  * daviddiasquit (Ping timeout: 245 seconds)
14:57:27  * m76joined
14:57:47  * rosskjoined
14:58:16  * daviddia_quit (Ping timeout: 245 seconds)
14:58:56  * dshaw_quit (Ping timeout: 246 seconds)
15:01:11  * daviddiasjoined
15:05:21  * daviddiasquit (Ping timeout: 245 seconds)
15:07:51  <MI6>joyent/node: orangemocha created branch orangemocha-testing3 - http://git.io/dJlDpA
15:10:47  * daviddiasjoined
15:12:58  * daviddia_joined
15:13:11  <MI6>joyent/node: Alexis Campailla orangemocha-testing3 * f9aa2fd : testing: testing - http://git.io/Z17-KQ
15:14:56  * daviddiasquit (Ping timeout: 245 seconds)
15:15:46  * kenperkinsjoined
15:15:47  * [m76]joined
15:16:29  * [m76]quit (Client Quit)
15:16:37  * m76quit (Write error: Connection reset by peer)
15:17:54  * m76joined
15:20:13  * AvianFluquit (Ping timeout: 272 seconds)
15:20:59  * kazuponjoined
15:26:22  * kazuponquit
15:27:25  <MI6>joyent/libuv: Saúl Ibarra Corretgé master * 515be33 : build: fix compilation warning if dtrace doesn't work - http://git.io/2-A3zQ
15:28:30  * paulfryzeljoined
15:30:25  * kenperkinsquit (Quit: Computer has gone to sleep.)
15:31:11  * kevinsimperquit (Remote host closed the connection)
15:33:02  * indexzeroquit (Quit: indexzero)
15:34:51  * kazuponjoined
15:41:45  * calvinfojoined
15:42:29  * calvinfoquit (Read error: Connection reset by peer)
15:42:41  * calvinfojoined
15:43:35  * calvinfo1joined
15:43:35  * calvinfoquit (Read error: Connection reset by peer)
15:44:07  * mikolalysenkoquit (Ping timeout: 244 seconds)
15:44:31  * calvinfojoined
15:44:31  * calvinfo1quit (Read error: Connection reset by peer)
15:48:10  <MI6>joyent/libuv: Brian White master * 269ff0b : fs: use preadv on Linux if available - http://git.io/-tIOQg
15:49:21  * calvinfoquit (Ping timeout: 272 seconds)
15:51:16  <MI6>joyent/libuv: Marc Schlaich master * 2f6d4b4 : build, windows: do not fail on Windows SDK Prompt - http://git.io/khr16g
15:53:23  * kazuponquit (Remote host closed the connection)
15:53:33  <saghul_>can any windows-er in the house review this, please? https://github.com/joyent/libuv/pull/1151 :-)
15:55:26  * dshaw_joined
15:59:36  * dshaw_quit (Ping timeout: 240 seconds)
15:59:50  * mikolalysenkojoined
16:01:21  * saghul_changed nick to saghul
16:01:33  * m76quit (Read error: Connection reset by peer)
16:07:20  * dap_joined
16:10:30  * AvianFlujoined
16:16:38  <tjfontaine>AlexisMocha: what are your thoughts for doing this in createWritableStream instead of in lib/net.js?
16:17:23  <tjfontaine>would require a slightly different interface
16:18:03  <tjfontaine>in pipe_wrap.cc we would have a setBlocking() and then in src/node.js createWritableStream we _handle.setBlocking(true)
16:20:43  <tjfontaine>or maybe actually in stream_wrap.cc
16:24:23  <tjfontaine>indutny: are you saying that the createConnection change interacts poorly with the agent: false change?
16:24:46  <indutny>tjfontaine: haven't dug much into it, but looks like so
16:24:59  <indutny>I'm sure Nate will figure it out
16:25:03  <indutny>but we shall not land it until then
16:25:16  <tjfontaine>right, I mean lgtm in the sense that I presume he ran the test suite ;)
16:25:31  <tjfontaine>if I were going to land it I would run them myself :P
16:25:47  <indutny>haha
16:25:49  <AlexisMocha>tjfontaine: that would also work, but it doesn't cover the case where the user creates a Socket passing a stdio fd. Maybe a corner case but it's included in the test. In any case, I see this as a temporary solution for 0.11
16:25:54  <indutny>yeah, but this test is introduced in another PR
16:26:32  <tjfontaine>AlexisMocha: it may be that we want to allow the scenario of someone explicitly creating a socket with an fd where they are expecting this behavior, this way we know that only when node is setting up stdio we're making it blocking
16:26:43  <tjfontaine>indutny: both are his, totally his fault :)
16:27:14  <tjfontaine>AlexisMocha: plus, if we ever want other streams to be blocking we have the API already in place :)
16:27:29  <indutny>tjfontaine: ahaha
16:28:22  * kazuponjoined
16:28:29  * guilleiguaranjoined
16:29:16  <AlexisMocha>tjfontaine: could do. Then another part of test-stream2-stderr-sync would be invalid
16:29:34  <tjfontaine>that test specifically watches for opening fd 1 or 2?
16:29:56  <AlexisMocha>opens a socket with {fd:2}
16:30:20  <tjfontaine>I swear, sometimes this software is so frustrating.
16:30:30  * xaqjoined
16:30:32  <AlexisMocha>not sure if we should give the user any control on this behavior
16:30:40  <AlexisMocha>*all* software? :)
16:30:53  <tjfontaine>especially this one, since it's the one we're responsible for ;)
16:31:51  * xaq_joined
16:32:13  <AlexisMocha>I mean, if users have the need to open sockets with {fd:2} to go around the sync behavior, we know we have screwed up.
16:32:41  <tjfontaine>I dunno, it's all fucked
16:32:42  <tjfontaine>:)
16:33:09  <AlexisMocha>I promise you a follow up after the release to review the sync vs async api for streams
16:33:31  <tjfontaine>anyway, from an api standpoint, lets leave this in the same place in js, but change the api from c++/js to be a setBlocking method inherited from stream_wrap.cc
16:34:01  <AlexisMocha>that's fine
16:34:13  <tjfontaine>that is to say, leave the logic in net.js but out of handle.open()
16:34:15  <tjfontaine>thansk
16:34:23  * mikealquit (Quit: Leaving.)
16:34:42  <AlexisMocha>but we call setBlocking in the same place where we open the handle, right?
16:34:45  * xaqquit (Ping timeout: 244 seconds)
16:35:00  <indutny>tjfontaine: does libusdt work properly on osx?
16:35:26  <indutny>tests seems to be failing
16:35:33  <indutny>`make test` in their repo
16:37:21  * AWintermanjoined
16:37:38  <tjfontaine>indutny: I haven't tried the test suite, no
16:37:51  <indutny>so it may not work on osx, right?
16:37:53  <indutny>oh
16:37:57  <indutny>better file the issue
16:39:20  * AWinterm_joined
16:39:25  <tjfontaine>the bindings I have work, the test suite may be broken yes
16:41:59  * AWintermanquit (Ping timeout: 240 seconds)
16:45:15  * calvinfojoined
16:46:11  * calvinfo1joined
16:46:11  * calvinfoquit (Read error: Connection reset by peer)
16:47:09  * calvinfojoined
16:47:10  * calvinfo1quit (Read error: Connection reset by peer)
16:49:05  <isaacs>indutny: ping
16:49:05  * calvinfoquit (Read error: Connection reset by peer)
16:49:09  <indutny>pong
16:49:18  <isaacs>indutny: i wasn't getting emailed by people asking why those modules exist.
16:49:24  <isaacs>their names make it clear what they DO with npm
16:49:32  <indutny>isaacs: man, seriously
16:49:37  <indutny>that's lame
16:49:42  * mikealjoined
16:50:09  <indutny>let's start from the fact that no one installs npm from npm
16:50:14  <indutny>people only do updates
16:50:18  <indutny>and it is quite rare
16:50:39  <isaacs>indutny: people were asking me what was up with that module.
16:50:55  <isaacs>it was causing confusion in the community of knowledgible parties.
16:51:00  <indutny>isaacs: I'm reading a lot of odd questions every day :)
16:51:05  <isaacs>that's the specific test of copyright infringement.
16:51:05  <indutny>in the mailing list
16:51:23  <tjfontaine>hmm?
16:51:46  <indutny>isaacs: I understand what you say, but npm is free and npmjs is free too
16:52:01  * daviddia_quit (Ping timeout: 245 seconds)
16:52:07  <indutny>and you do not own trademark
16:52:08  <indutny>anyway
16:52:21  <indutny>has owner of `npm` asked you to DCMA it?
16:52:38  <isaacs>indutny: i created the term, and modules that are hosted in my system are hosted there at my pleasure.
16:52:50  * benviequit (Ping timeout: 244 seconds)
16:52:52  * daviddiasjoined
16:52:53  <indutny>do you suggest that we should fork your system and use it?
16:53:08  <indutny>if we want to have unfiltered packages in it
16:54:00  <isaacs>indutny: if you want *literally unfiltered*, then yes.
16:54:17  <isaacs>indutny: if you want to upload disney movies, and child porn, and copyright infringement, then you'll have to do it on your own.
16:54:32  <indutny>isaacs: I see in what category you put `npmjs`
16:54:54  <indutny>isaacs: seriously, you was the last person from whom I expected such actions
16:54:57  <indutny>but whatever
16:55:01  <indutny>we all have different opinions
16:55:54  * mikolalysenkoquit (Quit: Lost terminal)
16:56:32  <isaacs>indutny: if it was not confusing, i have no problem with it
16:56:47  <isaacs>indutny: like i said, i was asked from several knowledgible people why arnout was making an official module for npm
16:57:04  <isaacs>indutny: and i immediately did what i am legally required to do in this case, and defended the mark from confusion in the community.
16:57:17  <isaacs>i've done it before, the only difference is that it's MY mark in this case.
16:57:24  <indutny>ok, I guess node forks would be next?
16:57:24  * benviejoined
16:57:38  <indutny>haha
16:57:41  <isaacs>indutny: everyone has a node fork.
16:58:04  <isaacs>indutny: i told arnout specifically that i have no problem with him writing such a module, he just has to name it so that it's not the same exact name as the website.
16:58:19  <isaacs>indutny: it's unclear what it is, and it seems like it's associated with the site, which it isn't.
16:58:57  * mikealquit (Quit: Leaving.)
16:59:15  <indutny>isaacs: the problem is how you did it. instead of devoting some time to find proper solution that works for everyone, you just removed it
16:59:45  * benvie_joined
17:00:10  <indutny>there could be a temporary redirect from npmjs to the new package name
17:00:20  <indutny>or something like that
17:00:26  <indutny>I understand that it takes time
17:00:30  <indutny>but that's the point
17:00:40  <indutny>being gentle to people who are using your service
17:00:50  <indutny>both publishers and the ones who are downloading stuff
17:01:11  <isaacs>indutny: i suggested several names that he could use instead.
17:01:30  <isaacs>indutny: if he hadn't responded, he would've gotten the customary 4 weeks to respond.
17:01:39  <isaacs>indutny: we DO actually have a policy, and i followed it
17:02:02  * benviequit (Ping timeout: 264 seconds)
17:02:16  <indutny>isaacs: AFAIK this policy is new
17:02:23  <indutny>and something that bother me too :)
17:02:35  * mikealjoined
17:02:38  <isaacs>indutny: it is not new. in fact, it's several years old. if you published something called "spiderman" and Marvel Entertainment told me to take it down, I woud.
17:02:44  <AvianFlu>I've been involved in several cases where modules got deleted on people
17:02:45  <isaacs>indutny: this has happened, several times.
17:03:01  <AvianFlu>marak camped a bunch of names, and his packages got deleted at nodejitsu's request, hilariously
17:03:04  <isaacs>indutny: first, the owner sends an email to me and the offender, saying, "Hey, please change this"
17:03:33  <isaacs>indutny: if the owner and publisher can't resolve it in 4 weeks, i intervene. If it's truly an eggregious case, I'll intervene earlier.
17:03:36  <indutny>AvianFlu: I'm quite sure those names was offending :P
17:03:53  <AvianFlu>well, they were the names of things that nodejitsu was going to publish, and the packages were empty
17:03:55  <isaacs>indutny: i don't pro-actively police for other peoples' IP, that would be madness.
17:03:55  <AvianFlu>so it's clearer
17:03:57  * benvie_quit (Client Quit)
17:04:00  <AvianFlu>but this stuff happens, is my point
17:04:07  <isaacs>AvianFlu: right, *empty* packages are deleted on sight.
17:04:22  <indutny>AvianFlu: I'm fine with deleting illegal or offending stuff
17:04:36  <indutny>but that thing sounds like a censoring stuff
17:04:53  <indutny>how much time passed since you guys emailed each other?
17:06:47  <isaacs>indutny: i emailed him friday. he responded right away, saying no, he won't remove it.
17:06:55  <isaacs>indutny: i emailed him again, saying, sorry, it's not a request. remove it or i will.
17:07:04  <indutny>ok, perhaps you are right
17:07:16  <isaacs>he replied again saying that he intends to keep it.
17:07:17  <indutny>but still it sucks that you haven't find solution that works for both of you
17:07:25  <isaacs>so i said, "Well, there it goes. Pick a new name"
17:10:35  <AlexisMocha>tjfontaine: https://github.com/joyent/node/pull/7196 I hope this is what you meant
17:10:48  <AlexisMocha>https://github.com/MSOpenTech/node/commit/5d1269a4879716b01285f2b3bf603e83e97098ed
17:11:35  <indutny>isaacs: ok
17:14:46  <tjfontaine>AlexisMocha: ok, yes this is precisely what I think we're looking for, except for two things :) ... 1: we need to check the return value and throw an errnoException, 2: we only actually want to do this when it's a pipe, right? -- man this is a frustrating api ...
17:16:54  <AlexisMocha>tjfontaine: 1: sure. 2: I didn't think we could create sockets on stdio other than on pipes. But how about I check if the setBlocking method is available?
17:19:50  <tjfontaine>AlexisMocha: right sorry, net.js can only do pipes, I was thinking of createWritableStdio which does work on files
17:20:38  * euoiaquit (Ping timeout: 264 seconds)
17:21:42  <tjfontaine>AlexisMocha: I am now just trying to think through the scenario of what happens if this is a tcp socket, they would have to go through great lengths to get here though
17:21:58  <AlexisMocha>i'll add the check anyway
17:22:23  <tjfontaine>AlexisMocha: I think we just want to add setBlocking to tcp_wrap.cc and tty_wrap.cc, even though there are no consumers
17:22:36  <tjfontaine>AlexisMocha: libuv will return an error properly anyway
17:23:46  <AlexisMocha>tjfontaine: but in that corner case in which this was a tcp socket, we shouldn't fail, just not call setBlocking
17:24:58  <tjfontaine>AlexisMocha: the scenario of them being able to do that involve someone explicitly do fs.closeSync(2) net.connect()/net.createServer().listen() and *then* doing new Socket({ fd: 2 })
17:25:19  <tjfontaine>they have to really work for it
17:26:10  <AlexisMocha>still, i'd rather not set it to blocking than fail. The less we expose of this the better, no?
17:27:52  <tjfontaine>yes
17:28:39  <tjfontaine>AlexisMocha: it's just that sniffing the method isn't the best way to do that, we actually want to do a type check on the handle
17:29:27  * benviejoined
17:36:07  * mikealquit (Quit: Leaving.)
17:36:31  * kazuponquit (Remote host closed the connection)
17:38:12  * mikealjoined
17:38:40  * abraxasjoined
17:40:01  * calvinfojoined
17:41:03  * euoiajoined
17:43:26  * abraxasquit (Ping timeout: 264 seconds)
17:46:20  * TooTallNatejoined
17:50:22  <AlexisMocha>tjfontaine: https://github.com/MSOpenTech/node/commit/1215c453d490e79eb85dbf8f20f5ac47768d5c94
17:50:42  <tjfontaine>looking
17:51:23  <AlexisMocha>couldn't find a prettier way of checking the handle type
17:51:38  <tjfontaine>I think we can probably make that nicer
17:52:12  <AlexisMocha>how?
17:52:50  <AlexisMocha>it's a temporary hack anyway
17:53:04  <tjfontaine>well it's somewhat expensive in that mechanism
17:53:21  <tjfontaine>since the handle is already created
17:53:22  <AlexisMocha>i liked sniffing the method better
17:53:38  <AlexisMocha>is there a cheap elegant way?
17:53:51  <tjfontaine>(handle instanceof Pipe)
17:55:46  <AlexisMocha>handle instanceof process.binding('pipe_wrap').Pipe
17:56:16  <tjfontaine>Pipe should already be cached at that point
17:56:23  <tjfontaine>https://github.com/MSOpenTech/node/blob/1215c453d490e79eb85dbf8f20f5ac47768d5c94/lib/net.js#L37
17:58:12  <AlexisMocha>it's not. that's a local var
17:58:19  <tjfontaine>well it could be hoisted
17:58:30  * dshaw_joined
17:58:36  <AlexisMocha>you want me to make it module global?
17:58:49  <tjfontaine>yes please
17:59:01  <tjfontaine>or
17:59:11  <tjfontaine>in createHandle stash type on handle
17:59:18  <tjfontaine>handle._type = type
17:59:43  * mikealquit (Quit: Leaving.)
18:00:20  <AlexisMocha>I think that it doesn't cover all code paths
18:01:22  * dshaw_1joined
18:01:26  * Ralithquit (Ping timeout: 264 seconds)
18:02:22  * mikealjoined
18:02:35  * dshaw_quit (Ping timeout: 244 seconds)
18:05:50  * mikealquit (Client Quit)
18:06:15  * mikealjoined
18:08:20  <AlexisMocha>tjfontaine: https://github.com/MSOpenTech/node/commit/b3ed2d2c6ddd2451ed83dd217d90bbb85786e5d5
18:10:08  * thlorenzquit (Remote host closed the connection)
18:10:44  * thlorenzjoined
18:15:22  <tjfontaine>AlexisMocha: yes this is it
18:15:30  * AvianFluquit (Ping timeout: 244 seconds)
18:15:48  * AvianFlujoined
18:16:23  <tjfontaine>AlexisMocha: after I integrate a new version of libuv we'll actually do it slightly differently, we will remove the check for isWindows and instead except out hte ENOSYS error
18:17:38  * mikealquit (Quit: Leaving.)
18:17:53  * mikolalysenkojoined
18:18:25  <AlexisMocha>tjfontaine: fine, as you prefer. It should be noted that unix is already blocking so it would make more sense for setBlocking to always succeed
18:19:03  <tjfontaine>AlexisMocha: right at least in the case of pipes
18:19:13  <tjfontaine>anyway thanks for your work on this
18:19:17  * mikealjoined
18:19:55  <tjfontaine>indutny: you have a fix for cluster disconnect flakeyness?
18:19:57  <AlexisMocha>tjfontaine: thank you. I am looking forward to seeing those 0 failures on windows
18:20:10  <tjfontaine>AlexisMocha: me too, me too. :)
18:27:43  * AvianFluquit (Remote host closed the connection)
18:28:12  * AvianFlujoined
18:38:59  <MI6>joyent/node: Trevor Norris master * 846f304 : test: fix async-listener-run-error-once (+2 more commits) - http://git.io/7SS_2Q
18:39:41  <tjfontaine>AlexisMocha: another one down :)
18:50:04  <tjfontaine>oh hell ya, I can feel a green build for unix coming on
18:51:43  * dshaw_1quit (Quit: Leaving.)
18:52:02  * dshaw_joined
18:52:23  <tjfontaine>bah so close
18:56:11  * petka_quit (Quit: Connection closed for inactivity)
18:57:11  <tjfontaine>AlexisMocha: http://jenkins.nodejs.org/job/nodejs-master-windows/lastCompletedBuild/DESTCPU=x64,label=windows/tapTestReport/simple.tap-540/
19:02:12  * dshaw_quit (Quit: Leaving.)
19:04:05  * daviddia_joined
19:04:06  * daviddiasquit (Ping timeout: 245 seconds)
19:18:24  <MI6>joyent/node: Nathan Rajlich v0.10 * 0a01a42 : http: invoke createConnection when no agent - http://git.io/zTRLog
19:19:22  * janjongboomjoined
19:19:53  * thlorenzquit (Remote host closed the connection)
19:20:35  * thlorenzjoined
19:20:41  * thlorenzquit (Remote host closed the connection)
19:20:55  * thlorenzjoined
19:35:12  * AvianFluquit
19:37:52  <indutny>tjfontaine: nope
19:37:55  <indutny>tjfontaine: I thought Alexis had
19:37:58  <indutny>TooTallNate: hey man
19:38:02  <indutny>about 0a01a42
19:38:11  <indutny>I think it breaks test from your another PR
19:38:16  <indutny>have you seen my comments?
19:38:19  <TooTallNate>indutny: ya, i noticed that
19:38:22  <TooTallNate>indutny: the test is faulty actually
19:38:25  <TooTallNate>i'm fixing it now
19:38:29  <indutny>oh
19:38:31  <TooTallNate>it doesn't consume the response
19:38:31  <indutny>it works with v0.10
19:38:44  <TooTallNate>ya, but it's a coincidence...
19:39:14  * mikolaly1enkojoined
19:39:35  * abraxasjoined
19:40:40  * AvianFlujoined
19:41:05  * xaqjoined
19:41:13  <TooTallNate>indutny: see https://github.com/TooTallNate/node/commit/47abdd9c43ae97ed11cb8cc4c770b43043718308
19:41:22  <TooTallNate>indutny: i it LGTY then i'll land on v0.10 branch
19:41:25  <TooTallNate>*if
19:42:10  * janjongb_joined
19:42:26  * daviddiasjoined
19:43:07  * dsantiag_joined
19:43:58  * abraxasquit (Ping timeout: 240 seconds)
19:44:18  * daviddia_quit (Ping timeout: 240 seconds)
19:47:36  * janjongboomquit (*.net *.split)
19:47:36  * mikolalysenkoquit (*.net *.split)
19:47:36  * xaq_quit (*.net *.split)
19:47:36  * guilleiguaranquit (*.net *.split)
19:47:38  * `3rdEdenquit (*.net *.split)
19:47:39  * dsantiagoquit (*.net *.split)
19:47:39  * rphillipsquit (*.net *.split)
19:47:40  * hueniversequit (*.net *.split)
19:48:55  * dshaw_joined
19:52:00  * mikolaly1enkoquit (Quit: Reconnecting)
19:52:06  * dsantiag_changed nick to dsantiago
19:52:16  * mikolalysenkojoined
19:52:59  <AlexisMocha>tjfontaine: why is that?
19:55:22  * dshaw_quit (Quit: Leaving.)
19:57:45  * AWintermanjoined
19:58:36  * guilleiguaranjoined
19:58:37  * `3rdEdenjoined
19:58:37  * rphillipsjoined
19:58:43  * hueniversejoined
19:59:07  * guilleiguaranquit (*.net *.split)
19:59:08  * `3rdEdenquit (*.net *.split)
19:59:08  * rphillipsquit (*.net *.split)
19:59:54  * wolfeida_joined
20:01:26  * Damn3dquit (Ping timeout: 272 seconds)
20:01:27  * AvianFluquit (Ping timeout: 272 seconds)
20:01:27  * wolfeidauquit (Ping timeout: 272 seconds)
20:01:28  * AWinterm_quit (Ping timeout: 272 seconds)
20:02:18  * AvianFlujoined
20:03:06  * Damn3djoined
20:04:32  * prettymuchbrycequit (Ping timeout: 244 seconds)
20:05:24  * guilleiguaranjoined
20:05:24  * `3rdEdenjoined
20:05:24  * rphillipsjoined
20:05:27  * dshaw_joined
20:05:28  * prettymuchbrycejoined
20:08:51  <AlexisMocha>tjfontaine: for that test i think you need a libuv update
20:09:24  * kpdecker1joined
20:09:29  * daviddiasquit (Remote host closed the connection)
20:09:33  * mikealquit (Quit: Leaving.)
20:09:34  * thlorenz_joined
20:09:59  <indutny>haha
20:10:07  <indutny>just finished adding `%dtrace()` to the v8
20:10:13  <indutny>now I need to hook it up with libusdt
20:10:23  <AlexisMocha>tjfontaine: for this: https://github.com/joyent/libuv/commit/0943d3cb33e2c28afb8252759e7a2313c9ec8ce4
20:12:57  * inolen1joined
20:12:58  * daviddiasjoined
20:14:09  * prettymuchbryce_joined
20:18:45  * prettymuchbrycequit (*.net *.split)
20:18:46  * AvianFluquit (*.net *.split)
20:18:46  * thlorenzquit (*.net *.split)
20:18:47  * kpdeckerquit (*.net *.split)
20:18:47  * jan____quit (*.net *.split)
20:18:47  * inolenquit (*.net *.split)
20:18:47  * prettymuchbryce_changed nick to prettymuchbryce
20:21:08  * AvianFlujoined
20:21:11  * Ralithjoined
20:25:34  * brsonjoined
20:27:13  <TooTallNate>tjfontaine: can i get an LGTY? https://github.com/TooTallNate/node/commit/47abdd9c43ae97ed11cb8cc4c770b43043718308
20:27:25  <TooTallNate>wanna land just the test on v0.10, so that we have test coverage there
20:31:09  * wolfeida_changed nick to wolfeidau
20:31:41  * dshaw_quit (Quit: Leaving.)
20:37:29  * kpdecker1quit (Read error: Connection reset by peer)
20:38:04  * kpdeckerjoined
20:41:33  * jan____joined
20:41:37  * jan____quit (Changing host)
20:41:37  * jan____joined
20:43:24  * dshaw_joined
20:48:37  * jmar777quit (Remote host closed the connection)
20:49:33  * kpdeckerquit (Quit: Leaving.)
20:49:50  * kpdeckerjoined
20:53:33  * dshaw_quit (Quit: Leaving.)
20:54:13  * jmar777joined
20:56:38  * calvinfoquit (Ping timeout: 240 seconds)
20:59:34  * kenperkinsjoined
21:02:10  * dshaw_joined
21:05:32  * kenperkinsquit (Remote host closed the connection)
21:05:41  * kevinsimperjoined
21:06:12  * kenperkinsjoined
21:08:09  * janjongb_quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:11:10  * mikolalysenkoquit (Ping timeout: 244 seconds)
21:18:39  <tjfontaine>TooTallNate: lgtm
21:18:44  <tjfontaine>AlexisMocha: ok I'll update libuv
21:20:13  <MI6>joyent/node: Nathan Rajlich master * f3189ac : http: remove the circular dependency (+3 more commits) - http://git.io/KJtH5g
21:20:25  <MI6>joyent/node: Nathan Rajlich v0.10 * 47abdd9 : test: add `agent: null` http client request test - http://git.io/JPpEUQ
21:20:36  * AvianFluquit (Ping timeout: 252 seconds)
21:20:48  * AvianFlujoined
21:21:27  * kenperkinsquit (Quit: Computer has gone to sleep.)
21:22:15  * kenperkinsjoined
21:22:51  * mikolalysenkojoined
21:24:07  * kevinsim_joined
21:24:34  * WalrusPony1joined
21:25:35  * jmar777_joined
21:26:27  * kenperkinsquit (Client Quit)
21:26:36  * c4milojoined
21:26:44  <tjfontaine>saghul, indutny: cutting a libuv release for 0.11, I don't think we're going to get the pipe thing reviewed before that, but we can always cut another if necessary
21:26:52  <indutny>sure
21:26:56  <indutny>I have no objections
21:27:31  * euoia_joined
21:28:30  * euoiaquit (Write error: Broken pipe)
21:28:34  * starkhalojoined
21:28:36  * jmar777quit (Write error: Connection reset by peer)
21:32:39  * WalrusPony2joined
21:33:04  * c4miloquit (Remote host closed the connection)
21:33:08  * starkhaloquit (Quit: WeeChat 0.4.3)
21:34:12  * daviddiasquit (Remote host closed the connection)
21:34:34  * thlorenzjoined
21:34:44  * kevinsimperquit (Ping timeout: 252 seconds)
21:35:05  * WalrusPonyquit (Ping timeout: 252 seconds)
21:35:25  * kenperkinsjoined
21:35:36  <MI6>joyent/libuv: tjfontaine created tag v0.11.20 - http://git.io/0yGp4A
21:35:38  <MI6>joyent/libuv: Timothy J Fontaine master * 6d463bf : Now working on v0.11.21 (+1 more commits) - http://git.io/XlD1iw
21:35:45  * WalrusPony1quit (Ping timeout: 252 seconds)
21:35:46  * thlorenz_quit (Ping timeout: 252 seconds)
21:36:24  * prettymuchbrycequit (Ping timeout: 240 seconds)
21:36:55  * rendarquit
21:39:55  * mikealjoined
21:40:28  * abraxasjoined
21:40:32  * starkhalojoined
21:43:01  * indexzerojoined
21:43:07  * prettymuchbrycejoined
21:45:18  * abraxasquit (Ping timeout: 265 seconds)
21:45:35  * jmar777_quit (Remote host closed the connection)
21:50:11  * mikealquit (Quit: Leaving.)
21:50:48  * mikealjoined
21:54:00  * indexzeroquit (Quit: indexzero)
21:56:32  * bradleymeckjoined
21:58:58  * indexzerojoined
22:00:57  * indexzeroquit (Client Quit)
22:02:16  * dshaw_quit (Quit: Leaving.)
22:04:10  <tjfontaine>saghul, indutny: doing the integration of the new libuv with vectored IO, so the alloc cb is called for each buf that is returned?
22:05:39  * kevinsim_quit
22:06:33  <indutny>tjfontaine: this is only for files
22:06:42  <indutny>i.e. only uv_fs_read/uv_fs_write changed
22:06:51  <tjfontaine>ya, I'm just looking at how to fix node_file.cc
22:06:52  <indutny>and they had no alloc_cb
22:07:00  <tjfontaine>the length is associated with each uv_buf_t?
22:07:15  <tjfontaine>and we pass the number of uv_buf_t's
22:07:46  * dshaw_joined
22:07:58  <indutny>you are not passing len explicitly
22:08:05  <indutny>it is calculated using length of buffers
22:08:07  <indutny>and yeah
22:08:15  <indutny>you need to pass count of buffers too
22:08:20  <indutny>you could just do 1
22:08:21  <indutny>for now
22:08:24  <tjfontaine>right
22:08:35  <tjfontaine>and the order has changed drastically
22:08:47  <indutny>not that much
22:08:49  <indutny>it was
22:08:52  <indutny>buf, lengths
22:08:55  <indutny>erm
22:08:56  <indutny>buf, length
22:08:57  <tjfontaine>I guess I need to not compare the node macro
22:08:57  <indutny>now it is
22:08:59  <indutny>bufs, count
22:09:09  <tjfontaine>the node macro handles it slightly differently
22:09:24  <tjfontaine>another question, can I use a stack allocated uv_buf_t?
22:09:48  <indutny>uv_buf_t - yes
22:09:55  <indutny>but their .data should be not allocated on stack
22:10:06  <tjfontaine>right, I was just checking to see how hard I had to change this :P
22:10:24  <indutny> ;;; <@72,#72> dtrace-probe
22:10:24  <indutny>0x34cefb490a3b 123 4887de REX.W xchgq rbx,rsi
22:10:25  <indutny>0x34cefb490a3e 126 0f1f440000 nop
22:10:25  <indutny>0x34cefb490a43 131 4887de REX.W xchgq rbx,rsi
22:10:26  <indutny>hahaha!
22:10:28  <indutny>eat this v8
22:10:42  <tjfontaine>you need to show bcantrill and dap_ :)
22:11:03  <indutny>not yet
22:11:05  <indutny>ok
22:11:05  <indutny>ttyl
22:11:08  <tjfontaine>k
22:17:18  * brsonquit (Ping timeout: 240 seconds)
22:17:32  * `3rdEdenquit (Quit: Connection closed for inactivity)
22:22:10  * kenperkinsquit (Quit: Computer has gone to sleep.)
22:22:57  <tjfontaine>well that almost worked the first time around
22:23:51  * AvianFluquit
22:24:02  * dshaw_quit (Quit: Leaving.)
22:25:20  * udpjoined
22:26:38  * paulfryz_joined
22:27:02  * paulfryzelquit (Read error: Connection reset by peer)
22:31:57  <tjfontaine>saghul: hey
22:32:11  <saghul>tjfontaine: hi!
22:32:30  <tjfontaine>can you do a quick sanity check with master's uv_fs_write for me
22:32:53  <tjfontaine>what's the req->result of a uv_fs_write on a uv_buf_init(NONNULL, 0)
22:33:09  * saghulchecks
22:33:33  <tjfontaine>offset of 0
22:33:51  <tjfontaine>I think I see the problem
22:34:45  <tjfontaine>saghul: https://github.com/joyent/libuv/blob/master/src/unix/fs.c#L601 should be -1 probably
22:35:30  <saghul>result will be the amount of data read /me thinks
22:35:51  <saghul>oh, you said write
22:35:55  <tjfontaine>yes
22:36:16  <tjfontaine>this is darwin, which I don't think has pwritev
22:36:41  <tjfontaine>so written = 0; we do the loop exactly once, but written is still 0, and therefore r is 1
22:36:55  <saghul>yeah, I see
22:36:58  <tjfontaine>so we call the callback with 1 bytes written, even though nothing happened
22:37:47  <tjfontaine>in this case, r == 0 is actually useful :)
22:37:58  * mikealquit (Quit: Leaving.)
22:40:22  <saghul>you are calling uv_fs_write with an emopty buffer, right?
22:40:24  * kpdeckerquit (Read error: Connection reset by peer)
22:40:34  <tjfontaine>that's write^Wright
22:40:41  <tjfontaine>fs.write(fd, '', 0, 'utf8', function(err, written) {
22:40:45  <tjfontaine>in the node logic
22:41:53  <saghul>can you try: https://gist.github.com/saghul/9240298
22:43:00  <saghul>tjfontaine: ^
22:43:01  <tjfontaine>saghul: works
22:43:11  <tjfontaine>lemme runt he whole test suite though ;)
22:43:16  <saghul>sure
22:43:24  * kpdeckerjoined
22:44:45  <saghul>also, I think Windows also needs a fix
22:45:13  <tjfontaine>ok, did I previously give you access to a scratch vm for testing against windows?
22:45:26  <saghul>ah, no, Windows is fine
22:45:43  <saghul>no, bug I have a windows laptop around here for testing
22:45:56  <saghul>*but
22:46:12  <tjfontaine>ok
22:46:41  <tjfontaine>it's not clear to me how, or if it's unrelated, but now our execsync tests are failing *sigh*
22:47:10  <saghul>do you want me to push the unix fix?
22:47:22  <tjfontaine>do you have time to test against windows?
22:47:46  <tjfontaine>I wouldn't push until you've also verified this case against windows, but the unix fix lgtm
22:47:53  <saghul>yes, let me write a uv test
22:49:20  <tjfontaine>it will be curious to see if reverting this change fixes the spawnsync stuff
22:49:56  <saghul>how does it fail?
22:50:23  <tjfontaine>they're timing out, unfortunately without anything useful in output
22:50:31  <tjfontaine>it could be completely unrelated of course
22:51:46  <tjfontaine>saghul: ya, ignore it, unrelated to this change
22:52:10  <saghul>ok
22:52:26  <saghul>since spawn sync opens a new loop it could be related to my uv_loop_init changes
22:52:34  <saghul>hopefully not though
22:52:56  <tjfontaine>well, I can't imagine what you did there to make that happen ;)
22:53:15  <saghul>in reality I just moved code around
22:53:23  <saghul>but who knows, software!
22:53:47  * mikealjoined
22:53:58  * paulfryz_quit (Remote host closed the connection)
22:54:01  <tjfontaine>you have tests that spin up multiple loops still with uv_loop_new?
22:54:39  <tjfontaine>oh I think we're just stuck waiting on the old loop to cleanup
22:54:47  <tjfontaine>saghul: is there an extra cleanup path?
22:55:00  * brsonjoined
22:55:18  <saghul>yes, there are tests which spin multiple loops
22:55:41  <saghul>the only difference is that now uv_loop_destroy will fail with an error if the loop is busy doing something
22:56:20  <tjfontaine>https://gist.github.com/tjfontaine/9240569
22:57:03  <saghul>hum, that doesn't really explain it
22:57:11  <saghul>one sec
22:57:35  <tjfontaine>that's jsut where I was when I attached
22:59:10  <tjfontaine>when I -ABRT'd it, it's just in io_polll
22:59:13  <saghul>tjfontaine: if I push to a feature branch jenkins do the magic, right?
22:59:19  <tjfontaine>yup
23:00:00  <MI6>joyent/libuv: saghul created branch fix_fs_write_emul - http://git.io/mWwwdg
23:00:26  * janjongboomjoined
23:02:58  <saghul>hum, looks like the windows slave is not actually running the tests? http://jenkins.nodejs.org/job/libuv-review-windows/label=windows/41/console
23:03:13  <tjfontaine>bah
23:03:14  <tjfontaine>moment
23:04:32  <saghul>osx tests passed, that's a Good Thing (TM)
23:04:45  <tjfontaine>seems like vcbuild.bat is different in master now
23:05:02  * mikolalysenkoquit (Ping timeout: 264 seconds)
23:05:34  <saghul>yep, but nothing major changed
23:05:45  <saghul>do you have the Windows SDK installed in those slaves?
23:06:19  <tjfontaine>yes
23:06:33  <tjfontaine>vcbuild.bat %DESTCPU% release
23:06:33  <tjfontaine>set UV_TAP_OUTPUT=1
23:06:34  <tjfontaine>Release/run-tests 2>&1 | tee test.tap
23:06:37  <tjfontaine>is what it's doign
23:06:55  <tjfontaine>only difference I guess is it's missing an .exe
23:07:13  <saghul>uh, maybe that Windows SDK thing is not liking something
23:07:26  <saghul>let me run tests locally, we can figure that out after
23:07:36  <tjfontaine>I just kicked it off
23:07:49  <tjfontaine>wtf
23:08:00  <tjfontaine>ohI know
23:08:03  <tjfontaine>silly batch
23:08:36  * toothrjoined
23:08:43  <saghul>maybe you can set WindowsSDKDir to ''
23:10:00  <tjfontaine>ok *now* it's working ;)
23:10:50  <saghul>test passes in my local machine, lets wait for Mr Jenkins
23:11:19  <tjfontaine>that is to say the normal ones still fail? :)
23:11:37  <saghul>yes!
23:13:16  <saghul>"ok 167 - fs_file_write_null_buffer"
23:13:22  <saghul>I think we are good
23:13:25  * bradleymeckquit (Quit: bradleymeck)
23:15:33  * toothrchanged nick to toothrot
23:17:17  <saghul>tjfontaine: fs test passes, other failures are unrelated, I'll push the fix then
23:17:48  <tjfontaine>k
23:18:11  <tjfontaine>I still need to track the spawnsync issue down, I can't do a libuv upgrade with that being brokne :/
23:18:12  <MI6>joyent/libuv: Saúl Ibarra Corretgé master * 3a8767e : unix: fix uv_fs_write when using an empty buffer - http://git.io/ocVodQ
23:18:16  * hueniversequit (Ping timeout: 244 seconds)
23:18:28  * hueniversejoined
23:19:30  <saghul>ok, let's check that one then
23:19:42  <saghul>apparently uv_run doesn't return, right?
23:22:27  <tjfontaine>saghul: without adding further debugging, that would seem to be hte case, yes
23:22:43  <saghul>tjfontaine: is your merge branch somewere? any way I can repro?
23:23:01  <tjfontaine>sure I push to my merge-review branch
23:23:15  <tjfontaine>it will be clean, that is to say, without your uv_fs_write fix
23:23:33  <saghul>ok
23:24:24  <MI6>joyent/node: Timothy J Fontaine merge-review * fab4b06 : src: update to vectored io in libuv (+9 more commits) - http://git.io/7cmWrA
23:24:33  <tjfontaine>saghul: pushed
23:24:59  <tjfontaine>erm that made jenkins unhappy :)
23:25:18  <saghul>oops
23:25:21  <tjfontaine>oh I missed adding my configure change
23:26:03  <saghul>gyp: Undefined variable uv_library in /home/saghul/src/node/deps/uv/uv.gyp while loading dependencies of /home/saghul/src/node/node.gyp while trying to load /home/saghul/src/node/node.gyp
23:26:03  <saghul>Error running GYP
23:26:03  <saghul>
23:26:06  <tjfontaine>ya
23:26:06  <MI6>joyent/node: Timothy J Fontaine merge-review * a4a7808 : src: update to vectored io in libuv - http://git.io/yFSEQA
23:26:18  <tjfontaine>repushed this time the commit includes configure
23:26:40  <saghul>ok, building
23:26:51  <saghul>how can I run the failing test?
23:27:04  <saghul>sorry, my node-fu is not strong
23:27:17  * brsonquit (Ping timeout: 265 seconds)
23:27:37  <tjfontaine>./node test/simple/test-child-process-execsync.js
23:28:09  <saghul>ok
23:28:20  <tjfontaine>that's the spirit debuggers, way to not expose the problem!
23:28:41  <mmalecki>tjfontaine: story of my life
23:28:54  <tjfontaine>get me some ctf and a core file
23:28:57  <tjfontaine>this shit would be easier :)
23:29:51  <mmalecki>I'm considering using something with DTrace as my laptop system for that exact reason
23:29:55  <mmalecki>archbsd is quite nice
23:30:00  <mmalecki>brb, heading home
23:31:59  <tjfontaine>ok I see uv__io_poll handle the uv__chld
23:32:03  <tjfontaine>signal 20
23:32:12  * thlorenzquit (Remote host closed the connection)
23:32:47  * thlorenzjoined
23:33:48  <saghul>neovim is using libuv, kewl!
23:34:31  <saghul>fuck, I forgot -j and this is taking ages :-/
23:35:54  * paulfryzeljoined
23:36:47  <tjfontaine>right kevent is stuck in infinite poll
23:36:48  * thlorenzquit (Ping timeout: 240 seconds)
23:38:06  <tjfontaine>saghul: Assertion failed: (r == sizeof msg || (r == -1 && (errno == EAGAIN || errno == EWOULDBLOCK))), function uv__signal_handler, file ../deps/uv/src/unix/signal.c, line 166.
23:38:15  * brsonjoined
23:38:19  <tjfontaine>which is not blowing on the release builds
23:38:24  * janjongboomquit (Ping timeout: 240 seconds)
23:39:00  <saghul>*** Error in `./out/Release/node': free(): invalid pointer: 0x000000000143b668 ***
23:39:19  <tjfontaine>la la la
23:39:23  <tjfontaine>are you on 10.6 still?
23:39:55  <saghul>this in on Debian
23:39:58  <tjfontaine>oh
23:40:11  <tjfontaine>fuck that's not good. :/
23:41:32  * abraxasjoined
23:41:57  <tjfontaine>so, we're in SIGCHLD handler, we are trying to write to the signal pipe, but it returns -1
23:44:06  * sblomquit (Read error: Connection reset by peer)
23:44:45  <saghul>the loop is not done yet, apparently
23:44:55  <saghul>uv_run returned but there are still handles there
23:45:11  <tjfontaine>uv_run returned for you?
23:45:22  <saghul>I called uv_loop_close and added an assert to see if it returned 0, but it doesn't
23:45:23  <saghul>yes
23:45:59  * abraxasquit (Ping timeout: 240 seconds)
23:46:20  <tjfontaine>if this is a difference between release and debug, Imma gone be pissed ;)
23:46:42  <saghul>apparently I built a release build
23:46:50  <tjfontaine>ya we default to it
23:47:43  * janjongboomjoined
23:48:28  <saghul>I'll make a debug build, the backtrace is useless now
23:52:48  <tjfontaine>saghul: https://gist.github.com/tjfontaine/9240569#file-gistfile2-txt
23:52:52  <tjfontaine>it's definitely racey as hell
23:53:06  <tjfontaine>but in this case, uv_run has *not* exited when I'm hung
23:53:34  <saghul>uh
23:54:07  <tjfontaine>https://gist.github.com/tjfontaine/9240569#file-gistfile3-txt
23:54:10  <tjfontaine>with the debug build
23:54:14  * euoia_quit (Ping timeout: 264 seconds)
23:54:19  <saghul>then that doesn't seem to be related to the changes to uv_loop_init and friends
23:54:35  <tjfontaine>no it's probably related to some pipe changes
23:54:43  * paulfryzelquit (Remote host closed the connection)
23:54:57  * Kakeraquit (Ping timeout: 244 seconds)
23:55:19  <saghul>I think the current uv_loop_delete implementation should have an assert though
23:55:38  <saghul>otherwise you'd be able to delete a loop which has hanles or requests in it
23:55:41  <saghul>as is the case
23:56:09  * euoiajoined
23:56:24  * kazuponjoined
23:59:10  * mikolalysenkojoined