00:00:01  * ircretaryquit (Remote host closed the connection)
00:00:08  * ircretaryjoined
00:04:26  * kazuponquit (Ping timeout: 268 seconds)
00:07:25  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
00:07:47  <kellabyte>godamnit, 2 days later and I still can't get this gyp to work, followed mailing list, libuv examples, I've made a mistake somewhere dunno where
00:09:58  <indutny>kellabyte: is it opensource?
00:10:02  <indutny>can you show me the code
00:10:14  <indutny>I used gyp in a couple of projects almost without problems
00:10:18  * loladirojoined
00:10:43  <kellabyte>indutny: yup! lemme get the link, and thanks :)
00:10:49  <indutny>np
00:10:53  <indutny>haven't helped yet
00:10:58  * kazuponjoined
00:11:28  <kellabyte>indutny: https://gist.github.com/kellabyte/5448923
00:11:41  <kellabyte>indutny: I don't care, anyone willing to spend the time to help me I'm thankful for :)
00:12:03  <indutny>ok, that mostly looks good to me
00:12:06  <indutny>what's happening?
00:12:13  <indutny>how are you generating makefile or whatever from it?
00:12:43  <indutny>is it like `<(library)` undefined, or something like this?
00:12:45  <indutny>python exception
00:12:52  <kellabyte>./bin/gyp/gyp -f make --depth=. --build=Debug -Dlibrary=static_library --generator-output=./build haywire.gyp
00:12:55  <kellabyte>it says..
00:13:11  <kellabyte>gyp: Invalid config specified via --build: Debug
00:15:57  <TooTallNate>kellabyte: what if you don't specify --build=Debug
00:16:09  <TooTallNate>kellabyte: you are specifying that as the default built type so maybe it's redundant
00:16:16  <TooTallNate>(though it shouldn't crash like that…)
00:17:51  <indutny>aha
00:17:55  <indutny>ok
00:18:01  <indutny>you need one really interesting thing
00:18:04  <indutny>-Lcommon.gypi
00:18:13  <indutny>and remove it from sources
00:18:19  <indutny>of haywire.gyp
00:18:38  <kellabyte>oh really? I saw a gypi in sources in uv.gyp
00:19:00  <indutny>well, you can leave it here
00:19:01  <indutny>ok
00:19:06  <indutny>just add -Lcommon.gypi
00:19:08  <indutny>to the build line
00:19:20  <TooTallNate>kellabyte: i've never seen a .gypi file in "sources" like that...
00:19:24  <TooTallNate>i'm not sure that does anything
00:19:34  <TooTallNate>in any case, indutny is the winnar
00:19:38  <TooTallNate>for a second time today :p
00:19:50  <indutny>well, it will add is a dependency
00:19:52  <kellabyte>-L is not a parameter gyp supports
00:19:56  <indutny>and will rebuild stuff it will change
00:20:03  <indutny>kellabyte: -Icommon.gypi?
00:20:04  <indutny>:)
00:20:16  <TooTallNate>kellabyte: it's -I
00:20:20  <TooTallNate>capitol "i"
00:20:24  <indutny>sorry
00:21:10  <indutny>does it works?
00:22:00  <indutny>s/works/work/
00:22:43  <kellabyte>no it just spewed a big error lol is gyp picky about order of arguments?
00:22:55  <indutny>haha
00:22:57  <indutny>well, lets see
00:22:59  <TooTallNate>what is the error?
00:23:22  <TooTallNate>if it's a different one than before than we're making progress :p
00:24:21  <indutny>gtg
00:24:22  <indutny>ttyl
00:24:24  <indutny>bar time
00:24:32  * loladiroquit (Quit: loladiro)
00:24:52  <kellabyte>https://gist.github.com/kellabyte/5448923
00:24:58  <kellabyte>I pasted the error at the bottom
00:25:03  <kellabyte>new error, so progress! lol
00:25:44  <kellabyte>if I specify configuration, target has to be specified perhaps? before I was just using defaults
00:26:35  <TooTallNate>kellabyte: i think you need to `cd` into "build" directory first
00:27:56  <kellabyte>TooTallNate: ahh, I think I misunderstood what that parameter was for!
00:28:51  <TooTallNate>kellabyte: ya, that's the dir when gyp will output the files it's generated
00:29:00  <kellabyte>TooTallNate: I was trying to tell it not to compile to "out"
00:29:25  <TooTallNate>kellabyte: i think you want -Goutput_dir=blah
00:29:58  <TooTallNate>kellabyte: for node-gyp, it does both: --generator-output=/path/to/build -Goutput_dir=.
00:31:01  <kellabyte>ah interesting, why -Goutput_dir=.?
00:31:09  <kellabyte>it'll create ./Release?
00:33:13  * trapitojoined
00:35:21  <TooTallNate>kellabyte: ya, cause otherwise it would build into ./build/out/Release :)
00:36:25  <kellabyte>TooTallNate: ah gotcha, thanks so much for the help! I think this line is what confused me, any idea what its there for? https://github.com/joyent/libuv/blob/master/uv.gyp#L61
00:37:02  <TooTallNate>kellabyte: not really… it could likely be removed
00:37:27  <kellabyte>ah
00:37:48  <kellabyte>omg so thankful for your help! I was ripping my hair out lol
00:39:00  * dominictarrquit (Quit: dominictarr)
00:41:24  * amartensquit (Quit: Leaving.)
00:42:05  * kazuponquit (Remote host closed the connection)
00:44:32  * nodeskippyjoined
00:44:47  * TooTallNatequit (Quit: Computer has gone to sleep.)
00:46:09  * nodeskippyquit (Client Quit)
00:47:58  <trapito>hi
00:59:35  * kazuponjoined
01:02:30  * kazuponquit (Remote host closed the connection)
01:03:00  * dapquit (Quit: Leaving.)
01:15:28  * inolenquit (Quit: Leaving.)
01:15:59  * nodeskippyjoined
01:16:30  * AvianFluquit (Remote host closed the connection)
01:33:06  * kazuponjoined
01:41:25  * kazuponquit (Ping timeout: 248 seconds)
01:43:06  * st_lukequit (Remote host closed the connection)
01:45:50  * nodeskippyquit (Quit: Leaving.)
01:51:08  * inolenjoined
02:03:01  * mikealjoined
02:14:29  * mikealquit (Quit: Leaving.)
02:14:36  * hzquit (Ping timeout: 264 seconds)
02:28:47  * bradleymeckjoined
02:30:28  * trapitoquit (Ping timeout: 245 seconds)
02:35:55  * c4milojoined
02:38:03  * TooTallNatejoined
02:44:49  * mikealjoined
02:58:33  * c4miloquit (Remote host closed the connection)
02:58:59  * c4milojoined
02:59:39  * kazuponjoined
03:03:25  * c4miloquit (Ping timeout: 256 seconds)
03:07:24  * pooyaquit (Quit: pooya)
03:08:05  * bradleymeckquit (Quit: bradleymeck)
03:08:27  * kazuponquit (Read error: Connection reset by peer)
03:09:00  * st_lukejoined
03:13:54  * AvianFlujoined
03:20:20  * loladirojoined
03:24:14  * brsonquit (Ping timeout: 256 seconds)
03:26:22  * kazuponjoined
03:35:02  * loladiroquit (Quit: loladiro)
03:40:50  * TooTallNatequit (Quit: ["Textual IRC Client: www.textualapp.com"])
03:47:35  * mikealquit (Quit: Leaving.)
04:01:43  * nodeskippyjoined
04:10:31  * mmaleckichanged nick to mmalecki[zzz]
04:11:32  * mmalecki[zzz]changed nick to mmalecki
04:13:42  * pooyajoined
04:28:29  * trevnorrisjoined
04:28:41  <trevnorris>hello
04:28:48  * brsonjoined
04:29:41  <tjfontaine>hola
04:30:32  * mikealjoined
04:30:49  <trevnorris>so, I've been test driving the crazy idea of passing back a non-persistent buffer.
04:30:58  <trevnorris>in terms of speed, it's ridiculously fast.
04:31:16  <trevnorris>not so safe, but oh so fast.
04:34:01  <tjfontaine>heh
04:35:52  <trevnorris>tjfontaine: what's interesting is that it clearly shows that malloc'ing is pretty much a noop vs time v8 needs to set everything up.
04:38:12  <tjfontaine>heh, well sure, lots of things get fast when you ignore the safety requirements :)
04:38:30  <trevnorris>lol, seriously.
04:39:15  <trevnorris>it's totally safe as long as no reference to the buffer is kept around, and no one slices the data. ;-)
04:44:22  * AvianFluquit (Remote host closed the connection)
04:54:13  * nodeskippyquit (Quit: Leaving.)
04:57:26  * st_lukequit (Remote host closed the connection)
05:00:21  * loladirojoined
05:18:54  * timoxleyjoined
05:22:41  * pooyaquit (Quit: pooya)
05:28:08  * trevnorrisquit (Quit: Leaving)
05:35:54  * pooyajoined
05:44:18  * loladiroquit (Quit: loladiro)
05:47:42  * indexzeroquit (Quit: indexzero)
05:50:34  * nodeskippyjoined
05:52:23  * nodeskippyquit (Client Quit)
06:00:01  * indexzerojoined
06:00:04  * indexzeroquit (Client Quit)
06:00:58  * indexzerojoined
06:01:12  * bajtosjoined
06:01:27  * indexzeroquit (Client Quit)
06:12:59  * mikealquit (Quit: Leaving.)
06:23:44  * rendarjoined
06:38:05  * `3rdEdenjoined
06:45:12  * nodeskippyjoined
06:51:19  * pooyaquit (Quit: pooya)
06:58:02  * nodeskippyquit (Quit: Leaving.)
07:08:37  * rvaggchanged nick to polyhackbot
07:10:57  * csaohjoined
07:12:02  * polyhackbotchanged nick to rvagg
07:12:10  * timoxleyquit (Quit: Computer has gone to sleep.)
07:15:15  * csaohquit (Client Quit)
07:17:44  * st_lukejoined
07:33:45  * csaohjoined
07:33:54  * benoitcquit (Quit: unexpected thing makes me to quit)
07:38:45  * benoitcjoined
07:40:11  * brsonquit (Ping timeout: 252 seconds)
07:41:01  * stolsmajoined
07:55:59  * `3rdEdenchanged nick to `3E|NOT|HERE|K
07:58:10  * bnoordhuisjoined
08:11:00  * st_lukequit (Remote host closed the connection)
08:19:42  * loladirojoined
08:20:26  * loladiroquit (Client Quit)
08:20:51  * `3E|NOT|HERE|Kquit (Remote host closed the connection)
08:37:19  * hzjoined
08:40:20  * kazupon_joined
08:41:40  * kazupon__joined
08:41:44  * kazuponquit (Read error: Connection reset by peer)
08:44:05  * bajtos_joined
08:44:16  * bajtos_quit (Client Quit)
08:44:32  * bajtos_joined
08:45:20  * kazupon_quit (Ping timeout: 258 seconds)
08:45:51  * bajtosquit (Ping timeout: 260 seconds)
08:49:21  * rendarquit (Read error: Connection reset by peer)
08:49:34  * Correnjoined
08:49:36  * rendarjoined
08:49:38  * rendarquit (Excess Flood)
08:49:54  * rendarjoined
08:49:55  * rendarquit (Excess Flood)
08:50:14  * rendarjoined
08:50:16  * rendarquit (Excess Flood)
08:50:33  * rendarjoined
08:50:34  * rendarquit (Excess Flood)
08:50:55  * rendarjoined
08:50:57  * rendarquit (Excess Flood)
08:51:25  * rendarjoined
08:51:26  * rendarquit (Excess Flood)
08:52:22  * rendarjoined
08:52:29  * bajtos_quit (Quit: bajtos_)
08:52:44  * kenperkinsquit (Ping timeout: 258 seconds)
08:52:44  * Correnchanged nick to kenperkins
08:52:46  * bajtosjoined
08:55:12  * stagas_joined
08:59:28  * kazupon__quit (Remote host closed the connection)
09:02:42  * st_lukejoined
09:03:44  * trevnorrisjoined
09:16:18  <bajtos>bnoordhuis: ping
09:22:38  * c4milojoined
09:35:58  <trevnorris>ircretary: tell isaacs did a quick implementation of those quick reference buffers I talked about. here are some of the results: https://gist.github.com/trevnorris/5458625
09:35:58  <ircretary>trevnorris: I'll be sure to tell isaacs
09:41:05  * dominictarrjoined
09:43:09  <bnoordhuis>bajtos: did you get my last message?
09:44:13  * Benviequit (Ping timeout: 240 seconds)
09:44:48  * st_lukequit (Remote host closed the connection)
09:45:11  * Benviejoined
09:49:56  * kazuponjoined
09:50:32  <bnoordhuis>bajtos: for posterity, it's okay to submit related changes in a single PR but they should go into separate commits
09:50:43  <bnoordhuis>the stress is on 'related' though
09:54:54  <trevnorris>bnoordhuis: so those quick cleanup buffers I mentioned are unstable and error prone, but man they're fast.
09:55:46  <bnoordhuis>trevnorris: shouldn't you be asleep by now?
09:56:19  <trevnorris>lol, probably. i usually try to sleep by 3am.
09:58:20  <trevnorris>well, just about there. off to bed. night.
10:03:22  * trevnorrisquit (Quit: Leaving)
10:07:27  * stagas_quit (Quit: Bye)
10:19:15  * kazupon_joined
10:20:21  * kazuponquit (Ping timeout: 256 seconds)
10:24:16  * kazuponjoined
10:26:45  * kazupon_quit (Ping timeout: 248 seconds)
10:29:06  * kazuponquit (Ping timeout: 276 seconds)
10:30:30  * kazuponjoined
10:32:09  * wolfeida_joined
10:33:12  * wolfeidauquit (Ping timeout: 264 seconds)
10:33:27  * csaohquit (Quit: csaoh)
10:55:25  * abraxasquit (Remote host closed the connection)
10:55:48  * bajtosquit (Quit: off for lunch)
10:55:58  * wolfeida_changed nick to wolfeidau
10:58:28  * kazuponquit (Remote host closed the connection)
11:00:38  * csaohjoined
11:00:44  * bnoordhuisquit (Ping timeout: 260 seconds)
11:04:25  * `3Ejoined
11:11:05  * piscisaureus_joined
11:20:53  * slurp1joined
11:20:53  * indutnytopic: And we're going to die at "liberal utopian vacation" day. ~ http://logs.libuv.org/libuv ~ http://groups.google.com/group/libuv
11:24:06  * stephank_joined
11:29:17  * mraleph1quit (*.net *.split)
11:29:17  * slurpquit (*.net *.split)
11:29:18  * stephankquit (*.net *.split)
11:29:23  * stephank_changed nick to stephank
11:31:51  * dominictarrquit (Quit: dominictarr)
11:32:03  * kazuponjoined
11:34:09  * bajtosjoined
11:36:28  * kazuponquit (Remote host closed the connection)
11:43:41  * bajtos_joined
11:46:47  * bajtosquit (Read error: Operation timed out)
11:50:35  * bajtosjoined
11:50:37  * bajtos_quit (Read error: Connection reset by peer)
11:52:22  * wolfeida_joined
11:52:22  * wolfeidauquit (Read error: Connection reset by peer)
11:52:33  * bajtosquit (Read error: Connection reset by peer)
11:52:34  * bajtos_joined
11:53:13  * bajtos_quit (Read error: Connection reset by peer)
11:53:29  * bajtosjoined
11:56:45  * wolfeida_changed nick to wolfeidau
11:58:03  * dominictarrjoined
12:01:17  * `3Echanged nick to `3E|BRBLOLZ
12:09:11  * bajtosquit (Read error: Connection reset by peer)
12:12:21  * piscisaureus_quit (Ping timeout: 248 seconds)
12:21:38  * timoxleyjoined
12:22:28  * sgallaghjoined
12:27:44  * bajtosjoined
12:29:37  * AvianFlujoined
12:32:08  * defunctzombie_zzchanged nick to defunctzombie
12:35:08  * sgallaghquit (Remote host closed the connection)
12:39:27  * sgallaghjoined
12:39:44  * sgallaghquit (Remote host closed the connection)
12:43:02  * piscisaureus_joined
12:46:56  * kazuponjoined
12:51:29  * kazuponquit (Ping timeout: 256 seconds)
12:56:44  * bnoordhuisjoined
13:01:11  * bnoordhuisquit (Ping timeout: 258 seconds)
13:04:48  * c4miloquit (Remote host closed the connection)
13:05:16  * c4milojoined
13:08:28  * piscisaureus_quit (Ping timeout: 258 seconds)
13:10:02  * c4miloquit (Ping timeout: 256 seconds)
13:13:45  * piscisaureus_joined
13:15:15  * bradleymeckjoined
13:18:21  * timoxleyquit (Quit: Computer has gone to sleep.)
13:29:49  * loladirojoined
13:34:24  * piscisaureus_quit (Ping timeout: 256 seconds)
13:40:20  * bnoordhuisjoined
13:42:20  * c4milojoined
13:44:40  * `3E|BRBLOLZchanged nick to `3rdEden
13:46:18  * bajtosquit (Quit: bajtos)
13:57:08  * piscisaureus_joined
14:01:45  * kevinswiberjoined
14:19:30  * dominictarrquit (Quit: dominictarr)
14:24:56  * sgallaghjoined
14:29:12  * loladiroquit (Quit: loladiro)
14:34:07  * defunctzombiechanged nick to defunctzombie_zz
14:36:57  * jmar777joined
14:45:26  <piscisaureus_>boring here
14:45:28  <piscisaureus_>:)
14:46:17  <bnoordhuis>hey, i know you
14:46:26  <bnoordhuis>didn't you use to do something with libuv and node.js?
14:47:09  <bnoordhuis>piscisaureus_: if you're bored, you can review miroslav's PR: https://github.com/joyent/node/pull/5365
14:47:28  <piscisaureus_>bnoordhuis: not really bored but ok :)
14:47:32  <piscisaureus_>after this. at 5pm
14:47:52  <bnoordhuis>what's 'this' in this context?
14:48:46  * bradleymeckquit (Quit: bradleymeck)
14:49:25  <piscisaureus_>test with thread pool scaling with a blocktime-based approach
14:54:51  * kazuponjoined
14:56:45  * bradleymeckjoined
15:06:12  * hzquit (Ping timeout: 264 seconds)
15:06:28  * hzjoined
15:17:30  * bnoordhuisquit (Ping timeout: 264 seconds)
15:18:36  * kazuponquit (Remote host closed the connection)
15:27:10  * timoxleyjoined
15:37:42  * dominictarrjoined
15:38:28  <CoverSlide>ooh thread pool scaling? sounds nifty
15:38:46  <CoverSlide>is there any reason behind the choice of 5 threads anyway?
15:44:33  * bradleymeckquit (Quit: bradleymeck)
15:46:16  * octetcloudjoined
15:46:44  * bradleymeckjoined
15:49:58  <isaacs>good morning
15:59:35  * dominictarrquit (Ping timeout: 252 seconds)
16:05:27  <piscisaureus_>CoverSlide: well the reason is it works okay in practice. Otherwise its pretty arbitrary
16:05:43  <piscisaureus_>CoverSlide: scaling is hard :)
16:05:55  <piscisaureus_>well - doing it sensibly
16:07:13  <tjfontaine>good day good sirs
16:11:12  <isaacs>piscisaureus_: can you check this out and provide some comments? https://github.com/joyent/node/issues/5358
16:11:48  <piscisaureus_>isaacs: Oh, great work.
16:12:03  <isaacs>piscisaureus_: i'd even be ok with: var stdout = execSync(cmd); var exitCode = child_process.lastExit
16:12:05  <piscisaureus_>isaacs: Yes, I have to step on my bike now but I can do that when I get home.
16:12:08  * TooTallNatejoined
16:12:11  <isaacs>or something similarly ugly
16:12:14  <isaacs>piscisaureus_: kewl
16:12:15  <piscisaureus_>ARGH
16:12:25  <piscisaureus_>isaacs please don't make me cry
16:12:26  <isaacs>piscisaureus_: mostly with execSync you only want the output
16:12:28  <isaacs>hahah
16:12:31  <tjfontaine>heh
16:12:32  <isaacs>there's precedent!
16:12:36  <isaacs>RegExp.$1
16:12:46  <isaacs>child_process['$?']
16:12:50  <isaacs>;P
16:12:53  <tjfontaine>brr
16:12:55  * dapjoined
16:13:06  <isaacs>better yet! global.$?
16:13:17  <isaacs>var output = execSync(cmd); exitCode = $?
16:13:29  <isaacs>except '$?' isn't a valid var name in js.
16:13:38  <isaacs>$QUESTIONMARK
16:14:08  * isaacsis dangerously close to exposing his bash fetish..
16:14:13  <tjfontaine>hehe
16:15:23  * jmar777quit (Remote host closed the connection)
16:15:35  <CoverSlide>but what to do with stderr or exit codes?
16:15:47  <tjfontaine>stderr goes to parent's stderr
16:15:52  <CoverSlide>maybe it could return an object?
16:16:01  * jmar777joined
16:16:21  <CoverSlide>{stdout:Buffer ..., stderr: Buffer ..., exitCode: 0}
16:16:40  <tjfontaine>CoverSlide: did you read the issue?
16:16:57  <CoverSlide>oic
16:17:00  <CoverSlide>sorry
16:18:12  * piscisaureus_quit (Ping timeout: 258 seconds)
16:19:17  * kazuponjoined
16:19:27  * dominictarrjoined
16:20:21  * jmar777quit (Ping timeout: 248 seconds)
16:20:50  * csaohquit (Read error: Connection reset by peer)
16:21:29  * mikealjoined
16:21:37  * csaohjoined
16:25:43  * nodeskippyjoined
16:28:02  <CoverSlide>personally I'm a fan of how python does subprocesses. subprocess.Popen() returns a Popen object with attributes stderr, stdout, stdin, returncode and pid. of course you have to wait() to get sync-like behavior so that would be implicit in this example, in which case stdin and pid are irrelevant
16:34:06  * hzquit
16:37:33  * mikealquit (Quit: Leaving.)
16:37:35  * indexzerojoined
16:38:10  * piscisaureus_joined
16:41:13  * rendar_joined
16:43:32  * rendarquit (Read error: Connection reset by peer)
16:46:36  * TooTallNatequit (Quit: Computer has gone to sleep.)
16:48:24  * csaohquit (Quit: csaoh)
16:49:48  * dominictarrquit (Quit: dominictarr)
16:50:41  * pooyajoined
16:56:29  <isaacs>CoverSlide: that's kind of where the "less simple" approaches have gone.
16:56:45  <isaacs>CoverSlide: i'm ok with those. but mostly, i think people just want their fucking output, and don't want to use a callback for it.
16:57:01  <isaacs>CoverSlide: the "fucking" is them talking to node, not me talking to you :)
16:57:26  <isaacs>CoverSlide: the hard part is splitting stdout and stderr
16:57:58  * timoxleyquit (Quit: Computer has gone to sleep.)
17:00:52  * indexzeroquit (Quit: indexzero)
17:02:16  * brsonjoined
17:02:16  * TooTallNatejoined
17:03:14  * kazuponquit (Remote host closed the connection)
17:04:41  * kazuponjoined
17:09:32  * loladirojoined
17:10:01  * inolenquit (Quit: Leaving.)
17:10:21  * bradleymeckquit (Quit: bradleymeck)
17:11:44  * mikealjoined
17:24:36  * dominictarrjoined
17:24:50  <piscisaureus_>isaacs: if I would take another shot at execSync with separate stderr/stdout would you mind?
17:26:21  <piscisaureus_>isaacs: I think being pragmatic here is good so we should go your route if I don't manage shortly (like, a week from now or so)
17:29:41  * dominictarrquit (Ping timeout: 268 seconds)
17:29:45  * dominictarr_joined
17:32:00  <TooTallNate>piscisaureus_: i don't think anyone in their right mind would mind you volunteering to write code :)
17:35:24  <piscisaureus_>hahaha
17:35:41  <piscisaureus_>well I mean I did things and never finished
17:35:48  <piscisaureus_>I'm trying to remember what the real problem was here
17:35:58  <piscisaureus_>writing a nested event loop is not that hard...
17:39:00  * loladiroquit (Quit: loladiro)
17:39:44  * kazuponquit (Remote host closed the connection)
17:40:55  * kazuponjoined
17:41:08  * bnoordhuisjoined
17:45:30  * loladirojoined
17:46:36  * jmar777joined
17:51:30  * inolenjoined
17:54:50  * loladiroquit (Quit: loladiro)
17:59:56  * dominictarr_quit (Quit: dominictarr_)
18:02:27  <isaacs>piscisaureus_: you have 7 days. Get it done ;)
18:02:47  <isaacs>piscisaureus_: i'm totally ok with doing it complicated, if it gets done soon. people wnat this, though, we should just deliver something reasonable soon.
18:03:27  <tjfontaine>isaacs: how painful would it be for the layout of /dist to change going forward?
18:04:21  * octetcloudquit (Ping timeout: 248 seconds)
18:04:31  <isaacs>tjfontaine: as long as it's bw compatible, then that might be fine
18:05:51  <tjfontaine>not sure who it would hurt, I basically want to stick the windows stuff in its own directory something like /{dist,nightlies}/<version>/win/{ia32,x64}/
18:06:51  * loladirojoined
18:07:09  <tjfontaine>maybe I could just put the .pdb/.exe stuff there, as that's the only problem portion
18:08:00  * TooTallNatequit (Ping timeout: 276 seconds)
18:10:12  * loladiroquit (Client Quit)
18:10:41  * octetcloudjoined
18:10:49  * TooTallNatejoined
18:14:33  * piscisaureus_quit (Ping timeout: 240 seconds)
18:29:57  * squeaky_pljoined
18:40:06  * bradleymeckjoined
18:42:36  * kazuponquit (Remote host closed the connection)
18:46:02  * st_lukejoined
18:51:06  * mralephquit (Quit: Leaving.)
18:56:14  * indexzerojoined
18:56:59  * toothrotquit (Ping timeout: 256 seconds)
18:59:34  * mralephjoined
19:03:40  * toothrjoined
19:04:33  * `3rdEdenquit (Remote host closed the connection)
19:13:14  * squeaky_plquit (Remote host closed the connection)
19:21:46  * mikealquit (Quit: Leaving.)
19:37:12  * trevnorrisjoined
19:39:23  <isaacs>tjfontaine: the thing that would suck is that version managers would all have to be updated.
19:39:36  <isaacs>tjfontaine: adding a *new* type of file/folder for nightlies would be fine.
19:39:51  <isaacs>tjfontaine: but even if we change how it uploads dist builds, we'll have to symlink or some crap.
19:40:22  <isaacs>tjfontaine: the last time we changed it (from a tarball-per-release to a folder-per-release) it was super annoying for those of us who maintain node version managers.
19:41:08  * timoxleyjoined
19:41:59  * brsonquit (Ping timeout: 246 seconds)
19:49:36  <trevnorris>good afternoon
19:52:10  * mikeal1joined
19:53:04  * `3rdEdenjoined
19:56:50  * brsonjoined
19:56:57  * jmar777quit (Remote host closed the connection)
19:58:54  <tjfontaine>isaacs: this would only effect the windows stuff, and is there a version manager that does windows binary work?
20:01:45  * TooTallNatequit (Quit: Computer has gone to sleep.)
20:02:24  * mikeal1quit (Ping timeout: 276 seconds)
20:13:15  * kazuponjoined
20:14:44  <trevnorris>bnoordhuis: don't usually stay up that late. actually I fell asleep early then kid woke up, couldn't get back to sleep.
20:14:49  <trevnorris>you have a kid right?
20:17:06  * TooTallNatejoined
20:18:07  * loladirojoined
20:21:39  * kazuponquit (Ping timeout: 260 seconds)
20:22:14  <trevnorris>TooTallNate: remind me what cloudup is?
20:22:32  <TooTallNate>trevnorris: my company's new product
20:22:33  <tjfontaine>who cares, I just filled it out and decided I'll find out later :)
20:22:44  <TooTallNate>well not "my" company, but who I work for :p
20:22:56  <tjfontaine>hipster tj just wants to say he was a beta luser
20:23:01  <trevnorris>lol. ok.
20:25:37  <TooTallNate>trevnorris: in a nutshell, i'd say it's kinda like dropbox, but with a focus on handling file types natively in the browser
20:26:00  <trevnorris>cool.
20:26:01  <TooTallNate>trevnorris: i.e. i give you a link to a, say, PSD file, and the website will decode that into an image to look at
20:26:08  <TooTallNate>so you don't *need* to have photoshop
20:26:24  <trevnorris>that's awesome.
20:26:30  <TooTallNate>that's just an example
20:26:48  <tjfontaine>presuming using gimp and not photoshop :)
20:26:49  <bnoordhuis>trevnorris: yep, i have a kid. second one is WIP actually
20:27:21  <bnoordhuis>due date sometime early september
20:27:28  <trevnorris>bnoordhuis: heh. so I'm sure you know what I mean. and congrats.
20:27:37  <bnoordhuis>i do and thanks :)
20:27:44  <tjfontaine>I vote for the 4th
20:28:06  <trevnorris>mine was born 2 days before PI day. I was disappointed.
20:30:35  <trevnorris>(can't tell if that came of like i'm an ass, or it was a joke)
20:31:05  <bnoordhuis>don't worry, i got it :)
20:31:22  * Kjerskijoined
20:31:40  * jmar777joined
20:31:43  <MI6>joyent/node: Ben Noordhuis master * 45ed546 : test: cluster: unhide child errors in bind-twice (+2 more commits) - http://git.io/aY6b9g
20:31:50  <trevnorris>bnoordhuis: what would you say to adding an optional third argument to "net.createServer" that would trigger quick buffer references?
20:32:38  <bnoordhuis>trevnorris: i would answer with a counterquestion
20:32:52  <bnoordhuis>namely, what is a quick buffer reference?
20:33:41  <tjfontaine>objects that aren't persisted
20:34:03  * loladiroquit (Quit: loladiro)
20:34:05  <trevnorris>bnoordhuis: ah, ok. basically you attach the char* to the buffer, pass the buffer, then clean it up immediately after the callback.
20:34:07  * dapquit (Read error: Connection reset by peer)
20:34:18  <bnoordhuis>ah, like that
20:34:18  * dapjoined
20:34:20  <trevnorris>so they just can't keep a reference to the buffer, or slice.
20:35:00  <bnoordhuis>right. it's an idea i've toyed with myself
20:35:00  <trevnorris>wrote up a test case last night. net was performing 20-30% faster.
20:35:46  <bnoordhuis>i don't know about exposing it in the public api right away
20:35:59  <bnoordhuis>but there's probably a few places where we can use it internally
20:36:09  <tjfontaine>should we update the windows bind twice tickets to indicate there's only one version now?
20:36:09  <trevnorris>good point.
20:36:38  <bnoordhuis>tjfontaine: they're still failing on windows?
20:37:11  <tjfontaine>bnoordhuis: before or after this commit?
20:37:13  * indexzeroquit (Quit: indexzero)
20:37:33  <bnoordhuis>tjfontaine: before. you're referring to this issue, i take it? https://github.com/joyent/node/issues/4966
20:37:48  <tjfontaine>yes
20:38:12  <bnoordhuis>right. i'll post a comment for bert to ignore
20:38:18  <tjfontaine>right
20:41:00  * wankdankerquit (Quit: Konversation terminated!)
20:42:13  * TooTallNatequit (Ping timeout: 248 seconds)
20:43:12  * indexzerojoined
20:43:13  * sgallaghquit (Remote host closed the connection)
20:43:38  <bnoordhuis>i still owe you that spec file btw
20:44:01  <bnoordhuis>didn't get around to it, the weather was just too great to stay indoor
20:44:32  <tjfontaine>hah, I understand, I changed the deb stuff slightly and added a little script to generate the debian changelog from ChangeLog
20:44:38  * TooTallNatejoined
20:44:58  * rendar_quit
20:45:31  <tjfontaine>I'm probably also going to make it do the reprepo stuff, but didn't get all the deps built for smartos, so I may punt and do it on linux and then just rsync it into place
20:47:06  <tjfontaine>I'm also going to have to put into my queue a jenkins plugin to spin up build drones on demand
20:47:55  <trevnorris>hm. interesting problem. how to deprecate a specific syntax of a method...
20:48:11  <tjfontaine>hmm, you mean a signature
20:48:53  <trevnorris>don't know. basically buffer .write() supports an old syntax that requires extra checks. just want to deprecate that so it can be removed.
20:49:24  <trevnorris>guess I should ask, what do you mean by "signature"?
20:49:30  <tjfontaine>a method signature
20:49:50  <tjfontaine>http://en.wikipedia.org/wiki/Type_signature
20:49:52  <bnoordhuis>function prototype for c/c++ programmers
20:50:10  <bnoordhuis>but that just confuses the hell out of js programmers
20:50:19  <MI6>nodejs-master: #173 UNSTABLE smartos-x64 (5/582) osx-ia32 (1/582) windows-ia32 (144/582) smartos-ia32 (2/582) windows-x64 (144/582) osx-x64 (1/582) linux-ia32 (2/582) http://jenkins.nodejs.org/job/nodejs-master/173/
20:50:32  <tjfontaine>oh god, that'll need rebuilt
20:50:39  <tjfontaine>I hate windows.
20:50:41  <tjfontaine>hate.
20:50:48  <bnoordhuis>hah
20:50:53  <bnoordhuis>i mean, i feel your pain
20:51:38  <tjfontaine>I'm really just going to pull windows into its own jobs out of the normal matrix builds, that way we can just ignore it (since everyone else does) and maybe get more people to watch the unicies test failures
20:52:38  <trevnorris>bnoordhuis: you're right. that's strange.
20:52:48  * mikealjoined
20:52:59  <trevnorris>(for a js developer)
20:53:36  <tjfontaine>syntax is generally a property of the language, at least when we're discussing them in polite conversation :)
20:55:09  * mikeal1joined
20:57:08  * mikealquit (Ping timeout: 252 seconds)
20:58:15  * timoxleyquit (Ping timeout: 256 seconds)
21:00:31  * stolsmaquit (Ping timeout: 256 seconds)
21:01:41  * octetcloudquit (Ping timeout: 245 seconds)
21:03:49  * qardjoined
21:05:39  * qardpart
21:07:18  * indexzeroquit (Quit: indexzero)
21:09:41  * st_lukequit (Remote host closed the connection)
21:10:00  <txdv>humanity has faught wars for less
21:11:23  <txdv>fought
21:11:33  <txdv>I'm a dead man
21:15:32  * hij1nxquit (*.net *.split)
21:16:12  <tjfontaine>htf
21:16:48  <tjfontaine>I hate windows.
21:17:06  <MI6>nodejs-master: #174 UNSTABLE smartos-x64 (2/582) osx-ia32 (1/582) windows-ia32 (144/582) smartos-ia32 (1/582) windows-x64 (144/582) linux-ia32 (1/582) http://jenkins.nodejs.org/job/nodejs-master/174/
21:17:39  <tjfontaine>I'm not sure if it's windows or jenkins fault that it can't reap all the left over processes
21:18:13  * kazuponjoined
21:23:11  * kazuponquit (Ping timeout: 256 seconds)
21:23:29  * hij1nxjoined
21:23:52  <tjfontaine>trevnorris: you certainly do forcepush a lot :0
21:23:55  <tjfontaine>er :)
21:25:14  <trevnorris>tjfontaine: heh, yeah.
21:25:21  <trevnorris>need me to tame it down?
21:25:53  <tjfontaine>nah I don't care
21:26:24  <tjfontaine>and for the most part the CI doesn't either
21:31:46  <trevnorris>=)
21:33:49  * `3rdEdenquit
21:38:17  * stagasquit (Read error: Connection reset by peer)
21:50:17  * bradleymeckquit (Quit: bradleymeck)
21:51:09  <MI6>nodejs-master: #175 UNSTABLE smartos-x64 (1/582) windows-ia32 (145/582) smartos-ia32 (1/582) windows-x64 (144/582) linux-ia32 (1/582) http://jenkins.nodejs.org/job/nodejs-master/175/
21:51:56  * loladirojoined
21:55:08  * jmar777quit (Remote host closed the connection)
22:01:59  * TooTallNatequit (Ping timeout: 255 seconds)
22:08:02  * loladiroquit (Quit: loladiro)
22:12:59  * octetcloudjoined
22:13:56  <bnoordhuis>morning sam
22:14:10  <bnoordhuis>or is it afternoon for you?
22:15:02  * dscape_quit (*.net *.split)
22:15:02  * joclekquit (*.net *.split)
22:15:26  * dscape_joined
22:15:27  * joclekjoined
22:18:56  * kazuponjoined
22:23:39  * TooTallNatejoined
22:25:01  * kazuponquit (Ping timeout: 245 seconds)
22:25:52  * dapquit (Quit: Leaving.)
22:29:22  * dapjoined
22:29:55  * c4miloquit (Remote host closed the connection)
22:30:22  * c4milojoined
22:34:53  * TooTallNatequit (Ping timeout: 240 seconds)
22:35:00  * c4miloquit (Ping timeout: 264 seconds)
22:37:31  * wolfeidauquit (Remote host closed the connection)
22:38:47  * TooTallNatejoined
22:43:33  * jmar777joined
22:45:53  * bradleymeckjoined
22:48:01  <tjfontaine>did I miss a change in the testsuite?
22:49:49  <bnoordhuis>tjfontaine: what time span are we talking here?
22:50:25  <tjfontaine>I dunno, I don't see a reason for this to be true, but the master-windows testsuite seems to be hating me, and throwing tons of addrinuse info
22:52:32  <tjfontaine>it's like NODE_COMMON_PORT isn't passing through the way I want it to
22:52:39  <MI6>nodejs-master: #176 UNSTABLE smartos-x64 (1/582) windows-ia32 (144/582) smartos-ia32 (1/582) windows-x64 (145/582) http://jenkins.nodejs.org/job/nodejs-master/176/
22:53:25  * TooTallNatequit (Ping timeout: 246 seconds)
22:54:23  <tjfontaine>I do wish eaddrinuse gave a bit more information
22:55:53  <tjfontaine>oh you know what, it could be bnoordhuis's fault
22:56:26  <tjfontaine>:)
22:57:35  <bnoordhuis>did i break something?
22:58:17  <tjfontaine>I believe your bind-twice test is leaving a hanging process
22:58:33  <tjfontaine>holding onto the common port
22:58:42  <bnoordhuis>oh? on what os?
22:58:56  <tjfontaine>windows of course
22:59:00  <bnoordhuis>ah
23:02:16  <bnoordhuis>okay, signing off. have a good night everyone
23:02:21  <tjfontaine>night
23:02:31  <trevnorris>night.
23:02:49  <tjfontaine>it's definitely leaving a hanging process
23:03:00  * TooTallNatejoined
23:03:07  <trevnorris>bnoordhuis: when you have a moment, mind taking a peek at: http://git.io/eThFJQ
23:03:23  <trevnorris>it's a more complex buffer pooling mechanism that's a hecka lot faster.
23:07:00  * bnoordhuisquit (Ping timeout: 268 seconds)
23:08:09  <trevnorris>isaacs: ^ wouldn't mind your feedback either. =)
23:09:09  <tjfontaine>heh, where did you come up with this mechanism
23:09:21  * russell_hquit (Changing host)
23:09:21  * russell_hjoined
23:13:01  <trevnorris>tjfontaine: just now. was testing for the "optimal" pool size, and noticed that almost no gain was achieved after a pool of 32x's the allocation size.
23:13:42  <trevnorris>so decided to break up the larger allocations (> 256 bytes) into chunks and segment each their own based on that.
23:14:21  <tjfontaine>sounds like an arena allocator :)
23:15:06  <trevnorris>heh, well it could be. didn't make it very far through my cs classes.
23:19:59  * mikeal1quit (Quit: Leaving.)
23:20:31  * mikealjoined
23:21:43  * kazuponjoined
23:21:56  <trevnorris>tjfontaine: so is this type of thing safe to use? (i'm reading up on them now ;-)
23:22:25  <tjfontaine>trevnorris: arena allocators? they're quite popular things
23:22:40  <trevnorris>oh, well. cool.
23:22:47  * bradleymeckquit (Quit: bradleymeck)
23:23:29  <tjfontaine>trevnorris: I'm sure they're probably covered to an extent in those papers isaacs linked you to yesterday
23:24:47  * c4milojoined
23:25:01  <trevnorris>tjfontaine: looked through those. think I got caught more up on the discussion of multi-threaded/processor allocations.
23:26:30  * kazuponquit (Ping timeout: 264 seconds)
23:26:59  <tjfontaine>the MT stuff is interesting, but there are some other parts that are interesting especially in the caching side
23:27:24  * loladirojoined
23:29:54  * nodeskippyquit (Quit: Leaving.)
23:31:57  <trevnorris>tjfontaine: what I'm curious is how those could help. check the perf output using that pooling: https://gist.github.com/trevnorris/5464076
23:33:23  <trevnorris>only ~8% of the time is spent in libc. and 3% in smalloc.
23:33:39  <trevnorris>the rest is in v8, which i'm not sure how much we could work around.
23:38:37  <tjfontaine>trevnorris: I'm not saying they're a panacea, nor that I think they're the only fix, just that when most of what you're fighting is against v8 you have to be willing to cut everywhere else you can
23:39:23  * mikealquit (Quit: Leaving.)
23:40:36  <trevnorris>tjfontaine: first, nice use of the word panacea. second, yeah. I understand.
23:41:00  <tjfontaine>heh :)
23:41:56  <trevnorris>then I could use your feedback. for example, the caching thing. docs say "preserve the invariant portion of an object's initial state"
23:42:24  <tjfontaine>right
23:42:30  <trevnorris>just not sure how that can be done w/ v8 since preserving any persistent handles is actually more expensive.
23:43:17  <tjfontaine>getting that to work with proper v8 objects probably isn't going to be ideal, at least if you can do that you'll be winning a lot more in the platform as a whole :)
23:46:04  <tjfontaine>I love doing a local change, and then looking at the breaking change and seeing that I just added in what ben took out
23:47:50  <trevnorris>lol
23:50:06  <trevnorris>tjfontaine: what i'd like is to pass a flag to SetIndexedPropertiesToExternalArrayData that says, cleanup data on gc.
23:50:19  <trevnorris>we'd have to remove "slice", but man that would be fast.
23:50:48  <trevnorris>been looking through the v8 code. easy to get lost in there.
23:50:55  <trevnorris>it's like a big bamboo forest.
23:51:14  <tjfontaine>it's certainly not for the faint at heart
23:52:33  <tjfontaine>I wish there were a half state for you, where you could reuse a persistent handle after the weak callback
23:53:07  <trevnorris>technically you can. since a persistent isn't destroyed until .Dispose() is called.
23:53:29  <trevnorris>actually wrote it that way at first, but the problem is after the make weak callback is called, the IS_WEAK flag is cleared.
23:53:38  <trevnorris>so you have to run MakeWeak on it again.
23:53:38  <tjfontaine>you might be able to get away with just allocating N persistent objects at start, and then as buffers are needed attach the char*, and then when you've lost M you reup to N
23:53:58  * jmar777quit (Remote host closed the connection)
23:54:10  <tjfontaine>dunno if its worth it
23:55:00  <trevnorris>it very much would be if didn't have to run MakeWeak every time we used the persistent.
23:55:18  <tjfontaine>on use, or on creation?
23:55:30  <tjfontaine>I mean, make 50, wait until you only have 25 left then make another 25
23:59:00  * mikealjoined