02:36:23  <piscisaureus_>aaaaaand slurp is back
02:36:26  * piscisaureus_quit (Client Quit)
02:38:42  * brsonjoined
02:47:48  * ericktquit (Quit: erickt)
02:48:29  * piscisaureus_joined
02:48:34  * piscisaureus_quit (Client Quit)
03:04:43  * xaqjoined
03:06:38  * xaqquit (Remote host closed the connection)
03:13:33  * mikealquit (Quit: Leaving.)
03:15:35  * xaqjoined
03:37:05  * xaqquit (Remote host closed the connection)
04:03:24  * avalanche123joined
04:07:53  * xaqjoined
04:09:28  * xaqquit (Remote host closed the connection)
04:10:39  * mikealjoined
04:17:32  * iraquit (Quit: Leaving...)
04:35:24  * bnoordhuisquit (Ping timeout: 252 seconds)
04:36:09  * loladiro_joined
04:38:39  * loladiroquit (Ping timeout: 260 seconds)
04:42:00  * ericktjoined
04:43:57  * xaqjoined
04:44:27  * Ariaquit (Remote host closed the connection)
04:54:36  * xaqquit (Remote host closed the connection)
05:02:04  * avalanche123quit (Quit: Computer has gone to sleep.)
05:08:29  * ericktquit (Quit: erickt)
05:13:23  * toothrquit (Ping timeout: 240 seconds)
05:15:21  * avalanche123joined
05:19:57  * toothrjoined
05:40:09  * isaacsjoined
05:58:39  <CIA-112>node: Charlie McConnell master * r4b021a3 / (5 files in 5 dirs): child_process: expose UV_PROCESS_DETACHED as options.detached - http://git.io/fra4IQ
06:00:22  <AvianFlu>lol wrong channel, there's docs in a second commit
06:00:46  <AvianFlu>https://github.com/AvianFlu/node/commit/6860e6ab10073772b7a1ab24d64dab7ea3797372
06:00:47  <isaacs>AvianFlu: squashed 'em into one
06:00:50  <AvianFlu>oh, great
06:00:52  <AvianFlu>nevermind then
06:00:57  <AvianFlu>I should read commit logs first lol
06:01:00  <isaacs>https://github.com/joyent/node/commit/4b021a3541770ae64a415760339bdd3223fdf602#diff-0
06:01:03  <isaacs>:)
06:01:11  <isaacs>a good concern :)
06:03:47  * isaacsquit (Remote host closed the connection)
06:10:14  * loladiro_quit (Ping timeout: 260 seconds)
06:21:54  * mikealquit (Quit: Leaving.)
06:26:16  * felixgejoined
06:26:16  * felixgequit (Changing host)
06:26:16  * felixgejoined
06:29:20  * AvianFluquit (Quit: This computer has gone to sleep)
06:30:00  * hij1nxjoined
06:30:07  * isaacsjoined
06:30:17  * mikealjoined
06:33:09  <CIA-112>node: Marc Harter master * r569acea / (3 files in 3 dirs): Fix #3379 prevent domain.intercept passing 1st arg to cb - http://git.io/BSpH0g
07:07:17  * avalanche123quit (Quit: Computer has gone to sleep.)
07:34:08  * rendarjoined
07:41:08  * TheJHjoined
07:49:58  * isaacsquit (Remote host closed the connection)
07:55:03  * hij1nxquit (Quit: hij1nx)
07:57:40  * dshaw_joined
08:13:48  * paddybyersjoined
08:18:16  * paddybyersquit (Client Quit)
08:21:14  * rendarquit
08:23:00  * rendarjoined
08:41:47  * AvianFlujoined
09:53:55  * AndreasMadsenjoined
10:00:32  * paddybyersjoined
10:17:51  * dshaw_quit (Quit: Leaving.)
11:20:33  * brsonquit (Remote host closed the connection)
11:47:22  * toothrquit (Ping timeout: 240 seconds)
11:49:56  * toothrjoined
11:55:41  * toothrquit (Read error: Connection reset by peer)
11:56:26  * toothrjoined
11:58:37  * AndreasMadsenquit (Remote host closed the connection)
12:14:27  * felixgequit (Quit: felixge)
12:34:01  * paddybyers_joined
12:37:11  * paddybyersquit (Ping timeout: 265 seconds)
12:37:11  * paddybyers_changed nick to paddybyers
12:39:58  * irajoined
12:44:55  * felixgejoined
12:44:55  * felixgequit (Changing host)
12:44:55  * felixgejoined
12:46:38  * felixgequit (Client Quit)
12:50:39  * mmaleckijoined
12:52:04  * mmaleckiquit (Client Quit)
12:53:16  * mmaleckijoined
12:53:52  * AndreasMadsenjoined
13:12:47  * hij1nxjoined
13:25:25  * iraquit (Quit: Computer has gone to sleep.)
13:59:52  * mmaleckiquit (Quit: Reconnecting)
14:00:03  * mmaleckijoined
14:35:14  * piscisaureus_joined
14:36:11  * ericktjoined
14:47:06  * isaacsjoined
14:57:18  * mmaleckiquit (Ping timeout: 244 seconds)
15:06:23  * mmaleckijoined
15:19:06  <isaacs>wow, v8 3.9 is WAY faster.
15:19:30  <isaacs>it's funny that virtually all of our performance in silly benchmarks is due to v8
15:20:14  <isaacs>while almost all of our performance in real-world stuff is due to real things like the number of syscalls we make and leaking memory and stuff.
15:23:35  <einaros>isaacs: did you ever get a chance to test the websocket bench stuff I pointed you to?
15:23:48  <isaacs>einaros: no, you got a link?
15:24:38  <einaros>https://github.com/einaros/websocket-benchmark
15:25:01  <piscisaureus_>isaacs:: so we're sticking to 3.9?
15:25:21  <einaros>what I observed was pretty massive perf drop for the benchmark of my module from 0.6.18 to 0.7.x
15:25:28  <piscisaureus_>isaacs: what benchmarks were the most significant btw?
15:26:05  <piscisaureus_>isaacs: maybe I can turn them into js-only benchmarks and have the v8 team take a look
15:26:34  <isaacs>piscisaureus_: the most significant (though it pains me) is http_simple.
15:26:41  <piscisaureus_>although.. hmm
15:26:50  <isaacs>it's also the least indicative of real-world performance.
15:26:57  <isaacs>but i've basically given up trying to convince people of this.
15:27:12  <isaacs>everyone just wants to see a big number from ab when they're returning a pre-rendered buffer.
15:27:31  <piscisaureus_>yeah, that's lame
15:28:05  <isaacs>everybody things that their boring non-dynamic express site is going to be slashdotted and hackernewsed and show up on the front page of yahoo and cpan.
15:28:14  <isaacs>but it's just pictures of their breakfast.
15:28:19  <isaacs>no one cares about your boring blog.
15:28:50  <piscisaureus_>isaacs: which benchmark were you running? bytes/123 ?
15:29:00  <isaacs>the performance impact that i care about is when i give it to voxer and jitsu, and they tell me its faster in real life.
15:29:26  <piscisaureus_>the latest v8, or node 0.7 ?
15:29:43  <isaacs>piscisaureus_: https://github.com/joyent/node/blob/master/benchmark/http.sh
15:29:56  <isaacs>i'm about to land rolling v8 back to 3.9
15:29:59  <isaacs>i'd previously done 3.10
15:31:28  <isaacs>piscisaureus_: defaults to /bytes/1024
15:33:18  <isaacs>piscisaureus_: the improvement is in isaacs/v8-upgrade
15:33:46  <isaacs>just making sure now that i didn't fuck the build on windows
15:42:17  * bnoordhuisjoined
15:45:55  <einaros>isaacs: benchmark output for 0.6.18 and 0.7.9 - https://gist.github.com/0aecce8dd50c78062e4d
15:46:31  <isaacs>piscisaureus_: if you get any time this weekend, wanna weigh in on https://github.com/joyent/node/pull/3393?
15:46:46  <isaacs>einaros: one sec...
15:47:24  <CIA-112>node: isaacs master * r940a686 / (254 files in 20 dirs): Roll V8 back to 3.9.24.31 - http://git.io/0JRdDA
15:47:24  <CIA-112>node: isaacs master * rc08357c / deps/v8/build/common.gypi : v8: Floating patches - http://git.io/pO64Vg
15:47:24  <CIA-112>node: Ben Noordhuis master * r26f754d / deps/v8/src/debug-agent.cc : v8: debug: fix error handling in SendConnectMessage() - http://git.io/TsckDA
15:47:31  <isaacs>einaros: can you try on joyent/master?
15:47:49  <einaros>alright
16:00:37  <einaros>isaacs: great - master is on par with 0.6.x
16:00:44  <isaacs>huzzah!
16:11:47  <CIA-112>libuv: Maciej Małecki master * rc92788e / test/runner.c : test: fix test runner progress bar - http://git.io/1zuH2w
16:12:25  <piscisaureus_>isaacs: re 3393: is this properly using windows resolution rules on windows ?
16:12:36  * madsonjoined
16:12:50  * AndreasMadsenquit (Read error: Connection reset by peer)
16:13:44  * travis-cijoined
16:13:44  <travis-ci>[travis-ci] joyent/libuv#396 (master - c92788e : Maciej Małecki): The build passed.
16:13:44  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/6a47e3a...c92788e
16:13:44  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1577513
16:13:44  * travis-cipart
16:16:11  <bnoordhuis>ah, back to 3.9?
16:18:12  <isaacs>bnoordhuis: yeah
16:18:18  <isaacs>performance regressions.
16:18:31  <isaacs>piscisaureus_: i'm not sure. that's why i asked you and igorzi_ to look at it :)
16:18:46  <bnoordhuis>yeah, i noticed last night when i benchmarked http_simple
16:18:49  <piscisaureus_>isaacs: ok
16:19:14  <isaacs>bnoordhuis: in my tests, master is now quite a bit faster than 0.6
16:19:15  <bnoordhuis>i was going to raise the subject today but it's no longer necessary :)
16:19:26  <bnoordhuis>oh, that's always good
16:19:35  <piscisaureus_>what were the major changes bettween 3.9 and 3.10 ?
16:19:49  <isaacs>piscisaureus_: not sure.
16:19:59  <isaacs>piscisaureus_: the chagne i care about is that 3.10 is slower.
16:20:04  <isaacs>piscisaureus_: but i'm sure that's just temporary :)
16:20:14  <piscisaureus_>isaacs: well incremental gc could be nice
16:20:19  <isaacs>in a few weeks it'll probably catch up and pass by, but we'll get that in 0.9
16:20:20  <piscisaureus_>but I am not sure when that landed
16:20:30  <isaacs>piscisaureus_: sure.
16:20:34  <isaacs>we'll pull it in eventually
16:22:00  <isaacs>piscisaureus_: about the windows symlink resolution rules...
16:22:15  <isaacs>piscisaureus_: it seems in my cursory testing like the fs.realpath matches what i was seeing in the shell
16:22:32  <isaacs>piscisaureus_: are there edge cases that you're aware of that i should check?
16:22:37  <piscisaureus_>isaacs: yes, this
16:22:54  <piscisaureus_>mkdir a\b\c
16:23:05  <piscisaureus_>mklink /d link a\b\c
16:23:17  <piscisaureus_>realpath('link/..')
16:23:37  <piscisaureus_>^-- that should point to '.' and not to './a/b'
16:23:44  <isaacs>hm.
16:23:59  <bnoordhuis>piscisaureus_: re v8 changes, there were quite a few but only one or two that look like they could have a performance impact
16:24:25  <bnoordhuis>and the missing write barrier fix probably shouldn't have that much impact
16:24:41  <bnoordhuis>the other one is improved support for harmony proxies, which could be expensive
16:26:53  <isaacs>piscisaureus_: https://gist.github.com/2901661
16:26:58  <isaacs>piscisaureus_: that's on unix
16:27:04  <isaacs>piscisaureus_: so, i think it's actually the same on both
16:28:06  * mmaleckiquit (Ping timeout: 246 seconds)
16:28:13  <piscisaureus_>isaacs: aha. so we just found a bug in unix :-)
16:28:27  <piscisaureus_>isaacs: you should try to open a file that does this
16:28:30  <CIA-112>node: Andreas Madsen master * r77cfbd9 / lib/domain.js : domain: dry decorate using util._extend - http://git.io/wHVi0w
16:28:31  <CIA-112>node: Andreas Madsen master * r535e109 / (lib/domain.js test/simple/test-domain.js): domain: run now return callback result - http://git.io/cnZrMA
16:28:31  <CIA-112>node: Andreas Madsen master * r60b45dc / (2 files in 2 dirs): domain: document and test dispose event - http://git.io/zKPezA
16:28:42  <piscisaureus_>isaacs: e.g.
16:28:52  <piscisaureus_>ln -s a/b/c link
16:28:56  <piscisaureus_>touch file
16:29:12  <piscisaureus_>fs.open('link/../file', 'r') <--
16:29:21  <piscisaureus_>isaacs: rhe open is going to fail
16:29:45  <piscisaureus_>isaacs: even if realpath will point you to 'link/file'
16:30:14  <isaacs>yeah
16:30:33  <isaacs>i'd written the js realpath impl to match what we had from the c implementation
16:30:47  <piscisaureus_>isaacs: so realpath does the right thing on windows , but not on unix
16:30:50  <isaacs>but realpath(2) is notoriously busted
16:31:05  <isaacs>piscisaureus_: clearly we need another test
16:31:29  <isaacs>or another case in test/simple/test-fs-realpath.js at least
16:31:54  <piscisaureus_>isaacs: we should verify that realpath(x) returns the path of the file that would be opened by open(x)
16:32:09  <isaacs>piscisaureus_: hm... yeah.
16:32:40  <piscisaureus_>because that is whay you want, eventually :-)
16:34:15  * piscisaureus_is now the path resolution nazi
16:35:56  <isaacs>piscisaureus_: yes, you are correct. https://gist.github.com/2901691
16:36:49  <isaacs>piscisaureus_: i'm content to let this lie for 0.9. meticulous x/../y handling is not super important. the pull req makes it work at all, which is important for npm ls to work properly in the face of symlinks.
16:37:09  <piscisaureus_>isaacs: well, on windows, this behaviour is actually correct
16:37:18  <piscisaureus_>isaacs: so yeah, for that part, +1
16:37:27  <piscisaureus_>all that's left now is a bug on unix :-)
16:38:16  <isaacs>yep
16:38:31  <isaacs>so we need to fix it without fixing it on windows ;)
16:38:34  <isaacs>since that'll break it
16:38:36  <piscisaureus_>yes
16:38:51  <isaacs>but i agree, realpath(x) should answer the question "what file will be opened by open(x)"
16:39:02  <isaacs>or more generally, "what file will be stat'd by stat(x)"
16:39:11  <piscisaureus_>yes
16:40:15  <piscisaureus_>we could also do
16:40:20  <isaacs>so, you're +1 on the patch for now then?
16:40:32  <isaacs>we've been living with this edge case unix bug forever, i think we'll be fine with it for another release cycle.
16:40:42  <isaacs>in fact, we may cause surprises by fixing it
16:40:50  <piscisaureus_>throw new YoureAnIdiotException("Wut? Resolving the parent of a symlink? Are you nuts?")
16:41:24  <piscisaureus_>isaacs: that's why it should be fixed in a major release. however, yes, +1 on your patch
16:42:49  <isaacs>piscisaureus_: as far as i'm concerned, 0.7 is basically stable already.
16:42:53  <isaacs>we need to think of it as such, anyway
16:43:42  <CIA-112>libuv: Ben Noordhuis master * r0ae026d / (test/benchmark-list.h uv.gyp test/benchmark-million-timers.c): bench: add lots-of-timers benchmark - http://git.io/6Yaphw
16:44:03  <piscisaureus_>isaacs: btw - I still have to review your commit really.
16:44:04  <bnoordhuis>spooky, that benchmark is faster in debug mode than it is in release mode...
16:44:11  <bnoordhuis>nearly twice as fast, even
16:44:28  <piscisaureus_>isaacs: this was my biggest concern. I might find other nits (in fact, I already did :-)
16:44:40  <CIA-112>node: isaacs master * r6332a4c / lib/fs.js : Expose posix realpath on windows as well - http://git.io/3FFt2w
16:44:40  <CIA-112>node: isaacs master * r424bca1 / lib/fs.js : Fix fs.realpath to work on Windows - http://git.io/eXGvRA
16:44:40  <CIA-112>node: isaacs master * r131a67e / (lib/fs.js test/simple/test-fs-realpath.js): Fix #3394 fs.realpath: Properly cache symlink targets - http://git.io/2ZfygA
16:44:41  <CIA-112>node: isaacs master * r65242ab / test/simple/test-fs-realpath.js : Fix fs.realpath tests so that they actually run - http://git.io/Kxlwag
16:44:55  <isaacs>bnoordhuis: that IS spooky.
16:45:35  <bnoordhuis>yeah, i guess i should look at the generated assembly and the profiler graph
16:45:45  * travis-cijoined
16:45:45  <travis-ci>[travis-ci] joyent/libuv#397 (master - 0ae026d : Ben Noordhuis): The build passed.
16:45:45  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/c92788e...0ae026d
16:45:45  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1577813
16:45:45  * travis-cipart
16:46:44  <piscisaureus_>isaacs: oops too late :-)
16:47:16  <isaacs>bnoordhuis: "Oh, you're in debug mode. You don't need to do those syscalls..."
16:47:21  <isaacs>piscisaureus_: whoops.
16:47:37  <isaacs>piscisaureus_: the only controversial aspect is that the test will fail if you're not an administrator.
16:47:38  <piscisaureus_>isaacs: nvm. I will review and maybe fix issues.
16:47:56  <isaacs>piscisaureus_: or if you don't have symlink ability, really
16:48:08  <isaacs>piscisaureus_: kewl.
16:48:11  <piscisaureus_>isaacs: I bet you my kingdom that I can find a bug in there :-p
16:48:23  <isaacs>piscisaureus_: you already did find a bug in there :)
16:48:35  <piscisaureus_>isaacs: another one, then :-)
16:48:47  <isaacs>kk, happy hunting.
16:48:54  <isaacs>add tests, please :)
16:49:09  <isaacs>i was really confused that those tests weren't running, btw. test-fs-realpath is 80% too clever.
16:50:46  <bnoordhuis>piscisaureus_: what numbers do you get on that benchmark?
16:52:13  <isaacs>ok, i'm heading out. leave a message if there are problems from that realpath stuff.
16:52:19  <isaacs>`npm link` in windows lands tonight :)
16:53:27  <piscisaureus_>isaacs: kewl
16:53:57  <piscisaureus_>isaacs: my preliminary tests show that realpath works as long as you don't do something really funky... so I wouldn't worry too much about it
16:54:06  <piscisaureus_>isaacs: eventually realpath should really move to libuv anyway
16:54:10  <bnoordhuis>curiously it's not faster when i run it through callgrind...
16:54:51  <bnoordhuis>gcc does a good job at inlining uv__timer_cmp() btw
16:56:23  <piscisaureus_>hurray for gcc
16:56:26  <piscisaureus_>and dinner time
16:56:49  <piscisaureus_>the only day I eat before 9pm is on saturday
16:56:54  <piscisaureus_>let's keep up the habit
16:57:00  * loladirojoined
16:57:21  <bnoordhuis>ah... the release makes 49 epoll_wait syscalls, whereas the debug build only makes 2
16:57:30  <piscisaureus_>whe lok
16:57:32  <piscisaureus_>lol
16:57:39  <piscisaureus_>you have an uv_idle in the test?
16:57:41  <piscisaureus_>or a bug?
16:57:46  <bnoordhuis>no, it's just timers
16:58:04  <bnoordhuis>i think it's an artifact of how i test the timers
16:58:18  <piscisaureus_>awa
16:58:20  <piscisaureus_>bbl
16:58:26  <bnoordhuis>wakka wakka
16:58:33  <piscisaureus_>ehe e
17:02:00  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
17:03:52  * xaqjoined
17:05:04  * hij1nxquit (Quit: hij1nx)
17:05:16  * mikealquit (Quit: Leaving.)
17:18:11  * isaacsquit (Remote host closed the connection)
17:31:13  * hij1nxjoined
17:31:23  * mikealjoined
17:53:34  * isaacsjoined
17:56:26  <bnoordhuis>isaacs: https://github.com/joyent/node/pull/3396 <- does that ring a bell?
17:58:48  <isaacs>bnoordhuis: yeah
17:59:06  <isaacs>he pinged me about that after writing the test. i told him to test on master, and it was working
17:59:18  * isaacshas learned to be much more careful about upgrading V8
17:59:45  <isaacs>it fails in 3.11.1, but passes in 3.11.7. but we are on 3.9 anyway, which is stable
18:00:56  <isaacs>bnoordhuis: we can probably just close the pull req. it's unlikely to regress unless i make the mistake of pulling in an unstable v8 again
18:01:28  * hij1nxquit (Quit: hij1nx)
18:04:06  <bnoordhuis>isaacs: would it hurt to pull it in though?
18:04:19  <bnoordhuis>it looks like a reasonable regression test to me
18:06:05  <isaacs>bnoordhuis: sure, should run fast.
18:14:16  * TooTallNatejoined
18:15:10  <tjfontaine>TooTallNate: ohai :)
18:15:25  <TooTallNate>tjfontaine: yo
18:15:33  <TooTallNate>i was just gonna look into your stuff
18:15:46  * TooTallNatesitting in a hotel in vegas… programming :D
18:16:08  <tjfontaine>heh you'll probably be luckier with code :)
18:16:44  <tjfontaine>TooTallNate: I'm really just confused as to what's going on with that reinterpret_cast it ends up out in the woods
18:20:01  <TooTallNate>tjfontaine: well i think something is wrong in your test case
18:20:24  <tjfontaine>TooTallNate: ok, do tell
18:20:41  <TooTallNate>tjfontaine: so after libvlc_media_get_tracks_info, "info" is a pointer to an array of these structs, yes?
18:21:22  * isaacsquit (Remote host closed the connection)
18:21:30  <tjfontaine>an array of pointers to the structs, but yes
18:21:56  <tjfontaine>well ok an array of structs
18:21:57  <TooTallNate>now i *think* what we have to do, instead of reinterpret(), is readPointer() there, at 0, and specifying the size of the returned array
18:22:15  <TooTallNate>so `TrackInfo.size * tracks`
18:22:32  <tjfontaine>hm
18:22:43  <TooTallNate>that will return a new appropriately sized buffer read from the address returned from the vlc call
18:23:27  <tjfontaine>if I don't reinterpret won't I have to allocate larger to start?
18:23:46  <TooTallNate>and then you can iterate through it, doing `ref.get(infoArray, i * TrackInfo.size, TrackInfo)`
18:24:07  <TooTallNate>tjfontaine: no, that's the point. readPointer() reads the memory address written to "info"
18:24:48  <TooTallNate>and returns a new buffer, and you can specify any size for it
18:25:19  * mikealquit (Quit: Leaving.)
18:25:26  <tjfontaine>ok, well I'm getting a different assertion this time then
18:25:39  <tjfontaine>oh I used get wrong
18:25:52  <TooTallNate>lemme try to make a gist
18:26:17  <tjfontaine>I got it, it works now
18:26:25  <TooTallNate>ya :D
18:26:34  <tjfontaine>we need to some serious documentation on this bad boy :)
18:26:40  <TooTallNate>this is very true
18:26:54  <TooTallNate>should be possible to whip something up with dox
18:27:00  <TooTallNate>there's a lot of inline comments
18:27:13  <TooTallNate>tjfontaine: so your pull is invalid I think
18:27:25  <tjfontaine>yes it is now
18:27:44  <TooTallNate>needed a test anyways ;)
18:27:53  <tjfontaine>ok before you leave and blow your inheritance on blackjack, is there an interface for free'ing function allocated memory?
18:28:35  <TooTallNate>well you called `new Buffer`, so node does that for you
18:28:50  <TooTallNate>once the buffer gets gc'd
18:29:14  <TooTallNate>otherwise, you can just make an ffi'd free() function. i have to do that for NodObjC
18:29:15  <tjfontaine>even if the caller is responsible for freeing the memory?
18:29:23  <tjfontaine>ok
18:29:40  <TooTallNate>when are you freeing in the C code sample?
18:29:56  <tjfontaine>I wasn't, but according to the libvlc docs I'm supposed to
18:30:30  <TooTallNate>what specifically, "info"?
18:30:38  <tjfontaine>ya, info
18:31:12  <CIA-112>node: Nuno Job master * r284816e / test/simple/test-bad-unicode.js : test: add test for bad unicode sequences - http://git.io/hscxEg
18:31:36  <TooTallNate>well try free()ing manually
18:31:44  <TooTallNate>and if there's a double free assertion error when the garbage collector runs you'll know
18:31:50  <tjfontaine>heh, indeed
18:32:30  <bnoordhuis>...
18:32:57  <bnoordhuis>TooTallNate: we need to talk about your approach to mentorship...
18:33:42  <TooTallNate>uh oh
18:34:41  * xaqquit (Remote host closed the connection)
18:34:55  <bnoordhuis>i half-kid. but double freeing as a way to catch errors is arguably not great advice
18:35:28  <TooTallNate>bnoordhuis: sure i agree, i'm just not familiar with the libvlc api
18:36:05  * bnoordhuisis off to dinner
18:36:36  <tjfontaine>heh, another way would try and watch the valgrind output I suppose :)
18:36:56  <tjfontaine>forcing gc()s doesn't seem to result in a double free, so seems relatively safe
18:37:23  <TooTallNate>tjfontaine: that seems proper to me
18:37:35  <tjfontaine>and I can make it double free, by doing it twice so this looks good
18:37:40  <TooTallNate>because once you call readPointer(), you're essentially reading into a buffer allocated by libvlc
18:38:00  <TooTallNate>and if their docs tell you to free it, then that makes total sense
18:38:22  * paddybyersquit (Quit: paddybyers)
18:38:28  <tjfontaine>indeed, I was just thinking maybe there should be a Buffer.noReallyIKnowIamSupposedToFreeThis()
18:38:34  <TooTallNate>tjfontaine: anyways, glad we got it figured out :)
18:38:41  * TooTallNateworking on docs meow
18:39:24  <TooTallNate>tjfontaine: you can do that with a "type" though
18:39:44  <TooTallNate>I'm gonna make a "readUtf8StringAndFree" type for NodObjC
18:39:58  <tjfontaine>ok
18:44:36  * ibcjoined
18:45:29  <TooTallNate>tjfontaine: actually a type wouldn't really work there. i think you just have to call free() manually
18:45:37  <ibc>hi, I init a async handle with uv_async_init(), and later, without calling uv_async_send() it occurs that uv_is_active() returns true for the async handle, is it correct?
18:45:50  <TooTallNate>tjfontaine: there is a possibility of adding a callback function to readPointer(), if you wanna lobby for it ;)
18:46:05  <TooTallNate>which would be invoked when the returned Buffer got gc'd
18:46:11  <TooTallNate>seems like a good fit in your case
18:47:00  <tjfontaine>ya, I was going to suggest either passing the cb back into js, or being explicit
18:47:08  <tjfontaine>I'll look at adding it
18:53:16  <ibc>I've reported it: https://github.com/joyent/libuv/issues/458
19:02:43  * TooTallNatequit (Ping timeout: 244 seconds)
19:04:01  * ericktquit (Quit: erickt)
19:05:32  * mikealjoined
19:06:08  * TooTallNatejoined
19:26:04  * AvianFlupart ("Leaving")
19:30:19  * ljacksonquit (Ping timeout: 252 seconds)
19:33:22  * ljacksonjoined
19:34:04  * mikealquit (Quit: Leaving.)
19:35:16  * paddybyersjoined
19:46:13  * AvianFlujoined
19:54:35  * ibcquit (Remote host closed the connection)
19:55:38  * brsonjoined
19:59:29  * brsonquit (Client Quit)
19:59:55  * brsonjoined
20:05:45  * piscisaureus_joined
20:06:43  * AvianFluquit (Ping timeout: 250 seconds)
20:06:44  <piscisaureus_>bnoordhuis: hey
20:06:51  <piscisaureus_>bnoordhuis: no, uv_async should keep the loop alive
20:06:55  <piscisaureus_>unless unrefed
20:07:24  <bnoordhuis>hey, i didn't make it up
20:08:09  <piscisaureus_>bnoordhuis: https://github.com/joyent/libuv/issues/347
20:08:38  <piscisaureus_>bnoordhuis: but, srsly, I don't know what that's supposed to mean but uv_async should definitely keep the loop alive
20:08:43  <piscisaureus_>and people can unref it if they don't care
20:09:06  <bnoordhuis>oh, it's not that i disagree
20:09:24  <bnoordhuis>i'm pointing out that this has been the behavior since forever
20:09:27  <bnoordhuis>and a day
20:09:39  <piscisaureus_>bnoordhuis: not in windows no :-p
20:09:59  <piscisaureus_>bnoordhuis: we should really be writing down this stuff, tbh. Especially the "exceptional cases"
20:10:53  <bnoordhuis>piscisaureus_: i agree. i expect the report on my desk at 4 pm
20:11:26  <piscisaureus_>bnoordhuis: the committee has deliberated and we decided to fire the guy that's responsible
20:11:34  <piscisaureus_>bye, ryah :-p
20:12:02  <bnoordhuis>he's been slacking off since like forever anyway
20:12:10  <piscisaureus_>yeah
20:12:58  <piscisaureus_>when I said it was okay to take some time off, I didn't mean he could just disappear for 6 months
20:13:34  * AvianFlujoined
20:13:55  <piscisaureus_>bnoordhuis: does uv-unix use uv_async internally somewhere?
20:14:41  <bnoordhuis>piscisaureus_: no
20:14:59  <piscisaureus_>bnoordhuis: ok, I was just about to say, nvm
20:15:03  <piscisaureus_>because it does :-p
20:15:36  <bnoordhuis>oh, for eio
20:15:49  <bnoordhuis>i thought you mean 'use' as in 'always'
20:16:04  * loladiroquit (Ping timeout: 260 seconds)
20:18:14  <tjfontaine>TooTallNate: is this not the right way to handle making char **argv https://github.com/tjfontaine/node-vlc/blob/master/vlc.js#L14
20:21:18  <CIA-112>libuv: Bert Belder reviewme * rfab7b7b / (src/unix/async.c src/unix/uv-eio.c): unix: uv_async handles should not be unref'ed automatically - http://git.io/ZyuUDQ
20:21:25  <piscisaureus_>bnoordhuis: ^-- lgty?
20:23:14  * travis-cijoined
20:23:15  <travis-ci>[travis-ci] joyent/libuv#398 (reviewme - fab7b7b : Bert Belder): The build was fixed.
20:23:15  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/commit/fab7b7b
20:23:15  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1578978
20:23:15  * travis-cipart
20:23:46  <tjfontaine>TooTallNate: never mind, that is indeed working
20:32:27  <CIA-112>libuv: George Yohng master * r87bf0a1 / src/win/util.c : windows: uv_interface_addresses() should not report disconnected adapters - http://git.io/uJQTfA
20:33:32  * avalanche123joined
20:34:25  * travis-cijoined
20:34:25  <travis-ci>[travis-ci] joyent/libuv#399 (master - 87bf0a1 : George Yohng): The build passed.
20:34:25  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/0ae026d...87bf0a1
20:34:25  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1579044
20:34:25  * travis-cipart
20:39:18  <CIA-112>libuv: George Yohng master * rfbe9974 / src/win/util.c : windows: uv_interface_addresses() should not report disconnected adapters - http://git.io/m7u55A
20:41:18  * travis-cijoined
20:41:19  <travis-ci>[travis-ci] joyent/libuv#400 (master - fbe9974 : George Yohng): The build passed.
20:41:19  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/0ae026d...fbe9974
20:41:19  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1579139
20:41:19  * travis-cipart
20:43:27  * brsonquit (Quit: Lost terminal)
20:44:47  * brsonjoined
20:49:21  * loladirojoined
21:10:54  * loladiroquit (Ping timeout: 260 seconds)
21:16:03  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
21:17:08  * madsonquit (Remote host closed the connection)
21:19:04  * TooTallNatequit (Ping timeout: 260 seconds)
21:24:41  * TooTallNatejoined
21:25:12  * c4milojoined
21:29:06  * TooTallNatequit (Ping timeout: 248 seconds)
21:50:00  * loladirojoined
21:59:30  * felixgejoined
21:59:31  * felixgequit (Changing host)
21:59:31  * felixgejoined
22:00:02  * rendarquit
22:03:51  * felixge_joined
22:03:51  * felixge_quit (Changing host)
22:03:51  * felixge_joined
22:08:04  * felixgequit (Ping timeout: 260 seconds)
22:08:06  * felixge_changed nick to felixge
22:10:22  * felixge_joined
22:10:22  * felixge_quit (Changing host)
22:10:22  * felixge_joined
22:12:20  * felixgequit (Ping timeout: 244 seconds)
22:12:20  * felixge_changed nick to felixge
22:14:09  * mikealjoined
22:15:57  * felixgequit (Quit: felixge)
22:28:09  * TooTallNatejoined
22:30:30  * theColejoined
22:47:09  * loladiroquit (Ping timeout: 260 seconds)
22:54:49  * paddybyersquit (Quit: paddybyers)
23:10:43  * loladirojoined
23:11:21  * hij1nxjoined
23:16:25  * theColequit (Quit: theCole)
23:42:34  * loladiroquit (Ping timeout: 260 seconds)
23:56:34  * isaacsjoined