00:02:54  <seebees1>If I may add something on the conversation of domains
00:04:07  <seebees1>In terms of managing callbacks, I think that this may be done by giving each "turn" of the event loop an number
00:04:32  <seebees1>and then exposing both this number and the turn from whence you came
00:04:35  <seebees1>(as it were)
00:04:46  <seebees1>It does not get you error handling
00:05:18  <seebees1>but it does allow you to unwind the events offline, as well as hand a bit of the problem off to userland.
00:05:53  <seebees1>I issued two pulls over the weekend, but I may not have made this argument clear.
00:12:11  * txdvquit (Ping timeout: 265 seconds)
00:14:01  * mmaleckichanged nick to mmalecki[zzz
00:14:04  * mmalecki[zzzchanged nick to mmalecki[zzz]
00:17:35  * paddybyersquit (Quit: paddybyers)
00:27:48  <isaacs>seebees1: i saw the pull reqs, but iddn't have a chance to dig into them
00:39:40  * bulatshakirzyanojoined
00:56:44  * Ariajoined
00:58:05  * dapquit (Quit: Leaving.)
00:58:36  * dapjoined
01:02:54  * CoverSlidequit (Ping timeout: 246 seconds)
01:06:29  * mikealquit (Quit: Leaving.)
01:07:39  * dapquit (Quit: Leaving.)
01:17:32  * pieternquit (Quit: pietern)
01:20:16  * bnoordhuisquit (Ping timeout: 260 seconds)
01:21:07  * mikealjoined
01:28:00  * txdvjoined
01:30:52  * xaqjoined
01:37:42  * mikealquit (Quit: Leaving.)
01:39:37  * mikealjoined
01:51:01  * mikealquit (Quit: Leaving.)
01:52:00  * mikealjoined
01:55:37  * abraxasjoined
01:58:20  * xaqquit (Remote host closed the connection)
02:02:11  * brsonquit (Quit: leaving)
02:11:24  * perezdquit (Quit: perezd)
02:25:30  * isaacsquit (Remote host closed the connection)
02:36:04  * dshaw_1joined
02:36:10  * dshaw_1quit (Client Quit)
02:40:53  * dshaw_quit (Ping timeout: 272 seconds)
02:50:54  * hij1nxjoined
03:05:47  * mikealquit (Quit: Leaving.)
03:13:01  * mikealjoined
03:19:05  * mikealquit (Quit: Leaving.)
03:20:14  * dshaw_joined
03:44:56  * mikealjoined
03:51:30  * Ariaquit (Read error: Connection reset by peer)
03:53:09  * Ariajoined
04:12:46  * Ariaquit (Remote host closed the connection)
04:44:44  * bulatshakirzyanoquit (Ping timeout: 245 seconds)
04:45:05  * pfox___joined
04:46:39  * bulatshakirzyanojoined
04:51:35  * isaacsjoined
04:53:12  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
04:55:34  * creationix_joined
04:55:41  * creationixquit (Excess Flood)
04:55:42  * creationix_changed nick to creationix
05:08:22  * isaacsquit (Remote host closed the connection)
05:19:53  * isaacsjoined
05:23:49  * xaqjoined
05:25:31  * isaacsquit (Remote host closed the connection)
05:26:28  * isaacsjoined
05:30:39  * isaacsquit (Ping timeout: 246 seconds)
05:32:39  * mralephjoined
05:49:50  * pfox___quit (Ping timeout: 264 seconds)
05:54:52  * dshaw_quit (Quit: Leaving.)
05:56:01  * mikealquit (Quit: Leaving.)
06:06:15  * orlandovftwjoined
06:07:24  * mikealjoined
06:10:30  * dshaw_joined
06:11:14  * mralephquit (Quit: Leaving.)
06:17:45  * txdvquit (Ping timeout: 245 seconds)
06:18:43  * txdvjoined
06:30:39  * txdvquit (Ping timeout: 260 seconds)
06:40:56  * paddybyersjoined
06:42:15  * txdvjoined
06:42:41  * hij1nxquit (Quit: hij1nx)
06:44:53  * stephankquit (Quit: *Poof!*)
06:57:41  * dshaw_quit (Quit: Leaving.)
07:10:15  * rendarjoined
07:12:07  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
07:18:47  * txdvquit (Read error: Connection reset by peer)
07:22:34  * txdvjoined
07:33:48  * dshaw_joined
07:39:10  * mikealquit (Quit: Leaving.)
07:53:24  * mikealjoined
08:01:26  * mikealquit (Quit: Leaving.)
09:30:27  * dshaw_quit (Quit: Leaving.)
09:32:02  * mmalecki[zzz]changed nick to mmalecki
09:34:28  * mmaleckichanged nick to mmalecki[brb]
10:06:06  * AndreasMadsenjoined
10:09:04  * dylukesjoined
10:27:15  * orlandovftwquit (Ping timeout: 265 seconds)
10:29:15  * dylukesquit (Quit: Computer has gone to sleep.)
10:33:41  * dylukesjoined
10:46:13  * dylukesquit (Quit: Computer has gone to sleep.)
10:46:33  * paddybyers_joined
10:49:49  * paddybyersquit (Ping timeout: 272 seconds)
10:49:49  * paddybyers_changed nick to paddybyers
10:55:07  * dylukesjoined
11:03:50  * dylukesquit (Quit: Computer has gone to sleep.)
11:15:45  * mmalecki[brb]changed nick to mmalecki
11:18:35  * abraxasquit (Remote host closed the connection)
12:06:36  * AndreasMadsenquit (Remote host closed the connection)
12:32:18  * piscisaureus_joined
13:05:50  * piscisaureus_quit (Read error: Connection reset by peer)
13:06:23  * piscisaureus_joined
13:23:27  * bnoordhuisjoined
13:57:54  * travis-cijoined
13:57:54  <travis-ci>[travis-ci] Skomski/node#8 (timeout-persistent-callback - 3df520b : Karl Skomski): The build failed.
13:57:54  <travis-ci>[travis-ci] Change view : https://github.com/Skomski/node/compare/f41901c^...3df520b
13:57:54  <travis-ci>[travis-ci] Build details : http://travis-ci.org/Skomski/node/builds/1057514
13:57:54  * travis-cipart
14:01:04  * skomskijoined
14:09:05  <piscisaureus_>bnoordhuis: hey
14:09:21  <piscisaureus_>bnoordhuis: do you have any idea how widespread fanotify-supporting kernels are?
14:12:30  * isaacsjoined
14:13:13  <tjfontaine>hmm 2.6.36 was the initial release of that
14:14:37  <tjfontaine>piscisaureus_: that means the 10.04 the previous LTS release of Ubuntu doesn't have support for it
14:15:04  <bnoordhuis>piscisaureus_: what tjfontaine said
14:15:04  * pfox___joined
14:15:16  <bnoordhuis>also, fanotify was present but not enabled in 2.6.36
14:15:22  <tjfontaine>right
14:15:25  <bnoordhuis>that wasn't until 2.6.37 or 2.6.38, i think
14:15:56  <bnoordhuis>and... *show stopper alert* fanotify requires CAP_SYS_ADMIN, i.e. root :(
14:16:09  <bnoordhuis>at least in the 3.x kernels
14:16:29  <tjfontaine>lame, CAP_SYS_ADMIN is a terrible CAP
14:16:33  <bnoordhuis>yes
14:18:17  <tjfontaine>the good news is I see patches working on removing that
14:19:03  <bnoordhuis>i'm not holding my breath, i think CAP_SYS_ADMIN is the most widely used capability
14:19:25  <tjfontaine>it is, there was an lwn article on it recently
14:19:40  <tjfontaine>http://lwn.net/Articles/486306/
14:19:55  * bradleymeckjoined
14:20:00  <tjfontaine>Which brings us to where we are today: of the 1167 uses of capabilities in C files in the Linux 3.2 source code, 451 of those uses are CAP_SYS_ADMIN. That's rather more than a third of all capability checks.
14:20:16  <bradleymeck>bnoordhuis you wanted to talk about distributing node binaries for addons?
14:20:29  <bnoordhuis>hah, right - that feels about right
14:20:37  <bnoordhuis>bradleymeck: who? me?
14:20:51  <mmalecki>bnoordhuis: yesterday, remember?
14:20:51  <bradleymeck>mmalecki said you were interested in something about it
14:21:17  <bnoordhuis>oh, that was something else - he told me you were compiling node a thousand times over
14:22:08  <mmalecki>yeah, we talked about distributing node itself
14:22:19  * c4milojoined
14:23:43  <bradleymeck>bnoordhuis: yea we have been out of some parnoia, a few months back were were just throwing binaries around and saw a few servers segfault from precompiled, we were crunched and compiling seemed to fix the issue so we just have been using that
14:24:34  <bradleymeck>didn't put too much research into it, but it was on Rackspace across different accounts ~700 servers at the time and around 10 failed
14:25:21  <bnoordhuis>bradleymeck: cosmic rays probably
14:25:27  <bradleymeck>bnoordhuis: heh
14:29:02  <mmalecki>I wouldn't mind fixing these 10 by hand btw ;)
14:30:00  * AndreasMadsenjoined
14:39:26  * AndreasMadsenquit (Remote host closed the connection)
14:59:47  * bradleymeck_joined
15:03:04  * bradleymeckquit (Ping timeout: 245 seconds)
15:03:04  * bradleymeck_changed nick to bradleymeck
15:06:31  * piscisaureus__joined
15:06:34  * piscisaureus_quit (Read error: Connection reset by peer)
15:10:10  * perezdjoined
15:17:36  * isaacsquit (Remote host closed the connection)
15:30:14  <indutny>bnoordhuis: hey ben
15:30:54  <bnoordhuis>indutny: ho fedor
15:39:04  * philipsquit (Excess Flood)
15:41:56  * philipsjoined
15:45:52  * pieternjoined
16:12:47  * dapjoined
16:17:18  * isaacsjoined
16:20:19  * orlandovftwjoined
16:21:27  * orlandovftwquit (Client Quit)
16:21:40  * orlandovftwjoined
16:22:59  <isaacs>good morning
16:31:09  <creationix>morning
16:38:55  * bulatshakirzyanojoined
16:40:07  * stephankjoined
16:44:10  * xaqquit (Remote host closed the connection)
16:49:54  * dshaw_joined
17:01:29  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
17:16:00  * orlandovftwquit (Ping timeout: 260 seconds)
17:20:45  * TooTallNatejoined
17:22:16  <pfox___>wow. heh.
17:22:36  <pfox___>i was pretty sure that Node.js vs Ruby on Rules email to the nodejs list was a joke/troll
17:22:48  <pfox___>surprised so many people replied with real, thoughtful answers
17:23:19  <pfox___>Ruby on Rails, even
17:36:33  * mralephjoined
17:50:00  * mikealjoined
17:54:52  <bnoordhuis>benchmarks...
18:00:49  * saghuljoined
18:04:23  * orlandovftwjoined
18:06:34  <piscisaureus__>yeah, speaking of benchmarks
18:06:44  <piscisaureus__>http_simple is really slow in my linux vm
18:07:00  <piscisaureus__>like, 2500 or so
18:07:57  <bnoordhuis>piscisaureus__: what numbers do you see with 0.4?
18:08:14  <piscisaureus__>bnoordhuis: let me try that out
18:08:27  <bnoordhuis>piscisaureus__: also, what http_simple benchmark are you running? buffer/1024, bytes/32, etc.?
18:08:50  <piscisaureus__>bnoordhuis: well buffer/x is just as slow
18:09:12  <piscisaureus__>I am going to check something else first
18:09:25  <piscisaureus__>Hopefully I did not accidentally install virtualbox 32 bit
18:09:35  <piscisaureus__>because that would be ehm, bad
18:13:50  <bnoordhuis>TooTallNate: ping
18:14:01  <TooTallNate>bnoordhuis: pong
18:14:08  <bnoordhuis>TooTallNate: hey
18:14:18  <bnoordhuis>do you have time to test a os x 10.5 patch today?
18:14:34  <TooTallNate>bnoordhuis: ya i should be able to
18:15:09  <piscisaureus__>no umm it's all x64
18:15:29  * bulatshakirzyanojoined
18:15:50  <bnoordhuis>TooTallNate: cool. i fixed that fs.watch ENOSYS error - at least, i think so
18:16:03  * skomskiquit (Quit: Leaving.)
18:16:08  <TooTallNate>bnoordhuis: is that functionality supposedly on 10.5?
18:16:57  <bnoordhuis>TooTallNate: maybe. apple docs are somewhat unclear on the subject
18:17:11  <TooTallNate>bnoordhuis: ok, well we'll just test it then :)
18:18:37  <bnoordhuis>TooTallNate: https://github.com/bnoordhuis/libuv/compare/issue3075
18:19:01  <bnoordhuis>TooTallNate: you can test it with -> for TEST in fs_event_ref fs_event_watch_dir fs_event_watch_file fs_event_watch_file_current_dir fs_event_no_callback_on_close fs_event_immediate_close fs_event_close_with_pending_event fs_event_close_in_callback; do echo "*** $TEST"; out/Debug/run-tests $TEST; done
18:19:20  <bnoordhuis>there's one test failing but that's because of something else
18:19:35  <TooTallNate>bnoordhuis: ok, on it
18:20:08  <bnoordhuis>thanks
18:20:35  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
18:21:46  * mralephquit (Quit: Leaving.)
18:21:52  * bulatshakirzyanojoined
18:22:02  * mralephjoined
18:26:02  * `3rdEdenjoined
18:27:35  <piscisaureus__>bnoordhuis:
18:27:35  <piscisaureus__>0.4 string/1024: 4248.11 [#/sec] (mean)
18:27:35  <piscisaureus__>0.4 buffer/1024: 3481.38 [#/sec] (mean)
18:27:35  <piscisaureus__>master string/1024: 3927.78 [#/sec] (mean)
18:27:35  <piscisaureus__>master buffer/1024: 2583.59 [#/sec] (mean)
18:27:35  <piscisaureus__>
18:27:47  <piscisaureus__>bnoordhuis: so, always slow, but also a significant regression
18:30:23  <bnoordhuis>piscisaureus__: string/x doesn't exist, you're probably testing the 404 page :)
18:30:35  <piscisaureus__>bnoordhuis: no I run "make bench"
18:30:49  <piscisaureus__>bnoordhuis: so it says stuff like ab-hello-world-buffer-1024
18:30:58  <piscisaureus__>but I just translated that to buffer/1024
18:31:05  <piscisaureus__>I know, it's bytes/xxx
18:31:20  <bnoordhuis>oh, right
18:31:58  <bnoordhuis>that bytes/x is a little slower is not terribly surprising
18:32:04  <bnoordhuis>buffer/x otoh...
18:32:24  <bnoordhuis>ab -g $data_fn -c 100 -t 10 http://$SERVER:$PORT/$uri > $summary_fn
18:32:30  <piscisaureus__>bnoordhuis: why are you not surprised that bytes/xxx is slower?
18:32:33  <bnoordhuis>-t 10 is too short
18:32:43  <bnoordhuis>piscisaureus__: isolates overhead for one
18:32:56  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
18:33:28  * `3rdEdenquit (Quit: Leaving...)
18:33:38  <bnoordhuis>piscisaureus__: can you rerun the benchmarks with -t 30?
18:35:02  <piscisaureus__>bnoordhuis: did you ever notice a perf degradation on your machines?
18:35:30  <bnoordhuis>piscisaureus__: between which releases?
18:35:47  <piscisaureus__>bnoordhuis: 0.4 .. master ?
18:35:55  <bnoordhuis>piscisaureus__: yes
18:36:16  <bnoordhuis>the introduction of isolates in v8 caused a pretty big drop
18:36:21  <piscisaureus__>bnoordhuis: right... so it is not surprising, then :-)
18:36:34  <piscisaureus__>bnoordhuis: then I will focus on perf issues with my VM in general
18:36:50  * bulatshakirzyanojoined
18:37:02  <bnoordhuis>piscisaureus__: well... the difference betwee v0.4 and master on the bytes/x benchmark is about what i suspected it would be
18:37:12  <bnoordhuis>the buffer/x benchmark is another story, that's a pretty steep drop
18:37:29  <bnoordhuis>also... now that i think of it, just after v0.6, we were faster than v0.4
18:40:11  <piscisaureus__>Yeah, I thought so too
18:40:19  <piscisaureus__>although maybe we weren't testing on x64
18:45:44  <bnoordhuis>upgrading the in-tree openssl is a royal PITA...
18:46:04  <TooTallNate>bnoordhuis: how come?
18:46:26  <mmalecki>#firstworldproblems
18:46:45  <bnoordhuis>mmalecki: there are no others
18:46:51  <bnoordhuis>TooTallNate: we have a couple of windows patches on top
18:46:59  <mmalecki>bnoordhuis: YOU have no others!
18:47:11  <bnoordhuis>also, every time you touch common.gypi, gyp rebuilds the whole tree...
18:54:30  * brsonjoined
18:54:40  <TooTallNate>bnoordhuis: here's the test results: https://gist.github.com/fce310d1a7a32487bd4a
18:55:44  <bnoordhuis>TooTallNate: thanks. that's better than i expected actually :)
18:56:10  <TooTallNate>bnoordhuis: haha, nice :)
18:56:53  <bnoordhuis>TooTallNate: if you're up to it, can you try this?
18:57:08  <bnoordhuis>gdb --args out/Debug/run-tests fs_event_watch_dir fs_event_watch_dir
18:57:08  <kohai>gdb has -1 beer
18:57:27  <bnoordhuis>`run`, then when the assert hits, `p events`
18:58:00  <TooTallNate>bnoordhuis: yup, one sec
19:00:21  <TooTallNate>bnoordhuis: hmm, doesn't seem to like that https://gist.github.com/5685cc1e620e11c1c625
19:01:01  <bnoordhuis>TooTallNate: oh, sorry. try `bt full`
19:02:01  <TooTallNate>bnoordhuis: here you are https://gist.github.com/a3209f2a8c36d91ab730
19:02:21  <bnoordhuis>events = <value temporarily unavailable, due to optimizations>
19:02:26  * bnoordhuisbangs head against desk
19:02:34  <bnoordhuis>in a debug build no less
19:02:47  <TooTallNate>hahah, indeed
19:03:04  <tjfontaine>I've started -O0'ing out of habit these days
19:03:06  <TooTallNate>should I recompile with -O0 or something/?
19:03:12  <bnoordhuis>TooTallNate: yes please
19:03:19  <TooTallNate>bnoordhuis: ok one sec
19:04:12  <TooTallNate>bnoordhuis: also, is there a backwards-compatible way to specify octals in Python? i'm using 2.5 on this machine and have to contantly reapply this patch for gyp :p
19:04:26  <TooTallNate>i should sent it upstream if there's a "preferred" way to do it
19:04:38  <bnoordhuis>TooTallNate: what do you mean with 'backwards-compatible'?
19:05:30  <tjfontaine>you mean octal literals?
19:06:56  <TooTallNate>tjfontaine
19:06:57  <TooTallNate>ya
19:07:12  <TooTallNate>gyp is using 0o666 vs. 0666
19:07:28  <TooTallNate>bnoordhuis: seems worse this time, haha https://gist.github.com/4c859a6fd7b37edb16ed
19:07:40  <bnoordhuis>TooTallNate: 0o123 is python3 syntax
19:07:55  <TooTallNate>bnoordhuis: ya but apparently it's 2.6 and up
19:08:11  <tjfontaine>2.6 in "preview" mode
19:08:17  <bnoordhuis>exactly
19:08:25  <TooTallNate>i think oct() is the backwards-compat way?
19:08:36  <bnoordhuis>just use decimal or hex :)
19:08:53  <tjfontaine>binary!
19:08:57  <bnoordhuis>oct() returns a string btw
19:09:11  <TooTallNate>oh, probably not what i want
19:09:13  <tjfontaine>I agree though, decimal or hex seems saner
19:14:53  * wankdankerjoined
19:19:17  <bnoordhuis>why is there no termio.h on os x? *sigh*
19:21:32  <tjfontaine>termios.h? :)
19:22:03  <tjfontaine>I honestly don't know why
19:22:44  <bnoordhuis>139 /* If unistd.h defines _POSIX_VERSION, we conclude that we
19:22:44  <bnoordhuis>140 * are on a POSIX system and have sigaction and termios. */
19:22:44  <bnoordhuis>141 # if defined(_POSIX_VERSION)
19:22:51  <bnoordhuis>from openssl :/
19:23:30  <bnoordhuis>apparently it doesn't export _POSIX_VERSION on os x...
19:23:44  * `3rdEdenjoined
19:28:33  <bnoordhuis>isaacs: https://github.com/joyent/node/issues/3081 <- have people never complained to joyent about that?
19:30:10  * mikealquit (Quit: Leaving.)
19:38:20  * piscisaureus__quit (Quit: ~ Trillian Astra - www.trillian.im ~)
19:42:03  * `3rdEdenquit (Quit: Leaving...)
19:46:03  * dylukes_joined
19:46:25  * dylukes_changed nick to dylukes
19:52:49  * `3rdEdenjoined
20:00:54  * benviequit
20:04:11  * mikealjoined
20:07:28  * piscisaureus_joined
20:11:10  <isaacs>bnoordhuis: not that i'm aware of.
20:11:11  <isaacs>i'll check
20:16:08  <bnoordhuis>wow, msbuild in multi-core mode means business, i have ~20 cl.exe processes running now
20:25:04  <TooTallNate>bnoordhuis: p.s. i found out what '\x1b[?1005h' does w.r.t. terminal mouse reporting
20:25:25  <TooTallNate>bnoordhuis: when you do that as well, then it's "utf8-mode"
20:25:40  <TooTallNate>i.e. x and y coords can go > 255
20:25:44  <bnoordhuis>TooTallNate: interesting
20:25:59  <bnoordhuis>and hacky, by the sound of it :)
20:26:04  * perezdquit (Quit: perezd)
20:27:15  <tjfontaine>welcome to terminal emulation
20:28:23  * perezdjoined
20:29:47  * piscisaureus_quit (Ping timeout: 246 seconds)
20:32:15  * avalanche123quit (Quit: Computer has gone to sleep.)
20:32:15  * bulatshakirzyanochanged nick to avalanche123
20:33:28  * avalanche123quit (Quit: Computer has gone to sleep.)
20:33:35  * avalanche123joined
20:35:28  * `3rdEdenquit (Quit: trololol)
20:35:33  <bnoordhuis>TooTallNate: how come node-gyp tries to download node-v0.7.5.tar.gz when node v0.7.6-pre is installed and on the path?
20:36:02  <bnoordhuis>i'm running `node-gyp configure` btw
20:36:14  <TooTallNate>bnoordhuis: closest match. isaacs added that. i've been meaning to add a switch to point to a source dir instead
20:37:01  <bnoordhuis>TooTallNate: there's no way to override it? will it try to build the source tree?
20:37:28  <TooTallNate>bnoordhuis: no and no. it does that and *only* extracts the .h files to build against
20:37:44  <bnoordhuis>okay, i suppose that's acceptable
20:37:46  <TooTallNate>bnoordhuis: they get extracted to ~/.node-gyp/x.y.z
20:38:13  <TooTallNate>bnoordhuis: it wouldn't really be needed except that the .msi and .pkg installers don't install header files
20:38:26  * piscisaureus_joined
20:38:53  <TooTallNate>i think it should stay this way though; it's nice sometimes to specify a different target version than the version of node that's running
20:39:53  <bnoordhuis>TooTallNate: is it expected behavior for `node-gyp install` to download the tarball again?
20:40:16  <TooTallNate>bnoordhuis: no, that sounds weird
20:40:29  <TooTallNate>bnoordhuis: you can `node-gyp ls` to see the dev headers that are installed
20:41:15  <bnoordhuis>yep, prints 0.7.5
20:41:34  <TooTallNate>bnoordhuis: really there should be no way that `install` downloads anything :p
20:42:00  <bnoordhuis>TooTallNate: https://gist.github.com/71892b8aa9b3eb68c2bd
20:43:18  <TooTallNate>bnoordhuis: oh, haha, i was thinking "build". ya "install" will re-install when specified. usually you never need to use "install"
20:43:40  <bnoordhuis>okay, clear
20:43:51  <bnoordhuis>does node-gyp install the module automagically?
20:44:16  <TooTallNate>bnoordhuis: no, it will be in build/Release/blah.node
20:44:37  <TooTallNate>bnoordhuis: err, what do you mean by "install"?
20:45:03  <bnoordhuis>TooTallNate: oh wait, nvm - i was thinking of `node-waf install`-like functionality
20:45:20  <bnoordhuis>but it's better not to have that
20:45:25  <TooTallNate>bnoordhuis: ahh right. ya node-gyp doesn't have anything like that
20:45:45  <TooTallNate>it's all up to npm after node-gyp builds it
20:55:42  * perezdquit (Quit: perezd)
20:57:38  * piscisaureus_quit (Ping timeout: 276 seconds)
21:01:58  * perezdjoined
21:03:11  <bnoordhuis>isaacs: http://nodejs.org/changelog.html <- kind of outdated :)
21:03:58  <isaacs>bnoordhuis: whoa..
21:04:05  <isaacs>yeah, that's kind of dumb.
21:04:16  * c4miloquit (Quit: c4milo)
21:04:31  <isaacs>bnoordhuis: i should do a make rule to generate that on each release.
21:05:52  <bnoordhuis>isaacs: want me to open an issue?
21:06:01  <isaacs>bnoordhuis: sure. assign to me.
21:06:10  <isaacs>i'll fix it for the next release, or before.
21:06:46  <mikeal>indutny: hey, when will candor have bindings to libuv?
21:06:59  <TooTallNate>mikeal: see candor.io
21:07:11  <bnoordhuis>isaacs: #3089
21:07:17  <isaacs>thanks :)
21:07:18  <TooTallNate>mikeal: err, creationix/candor.io
21:13:12  <TooTallNate>bnoordhuis: did you need me to run any more Leopard tests?
21:15:42  <bnoordhuis>TooTallNate: no. but did you get a backtrace from that one test?
21:16:23  <TooTallNate>bnoordhuis: i did: https://gist.github.com/4c859a6fd7b37edb16ed
21:16:26  <TooTallNate>that's with -O0
21:16:39  <TooTallNate>but it seems like that made things worse possibly :D
21:17:01  <bnoordhuis>TooTallNate: can you try `bt full`?
21:17:02  <tjfontaine>TooTallNate: should have given a bt full there as well :)
21:17:23  <TooTallNate>ahh, shoot, my bad :p
21:17:23  <TooTallNate>one sec
21:17:36  * mikealquit (Quit: Leaving.)
21:18:15  <TooTallNate>bnoordhuis: https://gist.github.com/01bcc546f27ec319d344
21:20:22  * mikealjoined
21:20:32  <bnoordhuis>TooTallNate: thanks. curious, looks like it's failing to open() the directory
21:20:50  <bnoordhuis>well, no matter - hardly anyone uses leopard anymore
21:21:10  <bnoordhuis>i'll just wait until the bug reports come in :)
21:22:00  <TooTallNate>bnoordhuis: i'm more interested in < /dev/tty personally :)
21:24:01  <tjfontaine>aren't you all happy that apple users are upgrade-early-upgrade-often users?
21:24:07  <CIA-99>libuv: Ben Noordhuis master * r3c41597 / (4 files in 3 dirs):
21:24:07  <CIA-99>libuv: unix: don't conditionally compile kqueue fs watcher
21:24:07  <CIA-99>libuv: Always compile in the kqueue-based fs event watcher and handle it at run-time
21:24:07  <CIA-99>libuv: if the kernel doesn't actually support it.
21:24:07  <CIA-99>libuv: Works around build issues when -mmacosx-version-min is not set properly.
21:24:08  <CIA-99>libuv: Fixes joyent/node#3075. - http://git.io/zqTZ6w
21:24:16  <bnoordhuis>yes. the one good thing about apple
21:24:38  * benviejoined
21:24:44  <tjfontaine>I say that, and yet I have 2 10.4 users I still have to support, and they unfortunately own the company
21:25:06  <tjfontaine>10.4 can still run classic you see, and they have applications that only run in classic ...
21:25:25  <mmalecki>enterprise.
21:26:06  <tjfontaine>they don't have to use the classic apps, they just refuse to accept the theme change of the new apps
21:26:20  <TooTallNate>tjfontaine: haha, that's the worts
21:26:22  <TooTallNate>worst
21:26:28  <TooTallNate>old people; set in their ways :p
21:26:51  <tjfontaine>TooTallNate: indeed
21:27:30  <tjfontaine>two years ago I finally got rid of my HP-UX 10.20 box, it's been replaced by a Power6 AIX5.3 box *sigh*
21:27:41  <TooTallNate>bnoordhuis: so for < /dev/tty are we gonna revert-revert the -mmin-macosx-version patch, or were you gonna look into it some more?
21:27:59  <bnoordhuis>TooTallNate: no, it can be reverted
21:28:04  <TooTallNate>ok
21:28:30  <bnoordhuis>upgrading libuv, recompiling the whole tree for the 10th time today, then i'll revert it
21:28:34  <bnoordhuis>or you can do it if you want
21:29:40  * pfox___quit (Read error: Operation timed out)
21:33:39  <CIA-99>node: Ben Noordhuis master * r0db4dc0 / (41 files in 9 dirs): deps: upgrade libuv to 3c41597 - http://git.io/om-Wsw
21:33:43  <bnoordhuis>TooTallNate: she's all yours
21:33:58  <mmalecki>that's what he said!
21:34:06  <TooTallNate>lulz
21:34:14  <mmalecki>when untying her.
21:39:37  <CIA-99>node: Nathan Rajlich master * r70a5b53 / common.gypi :
21:39:37  <CIA-99>node: Re-apply "build: target OSX 10.5 when building on darwin"
21:39:37  <CIA-99>node: This reverts commit 93eca95aece9804d22b012cc89487f4b0064a86d.
21:39:37  <CIA-99>node: Fixes #3072 (once again). - http://git.io/0hfBog
21:40:35  <CIA-99>node: Nathan Rajlich master * r7b71fd0 / common.gypi : build: add comment explaining MACOSX_DEPLOYMENT_TARGET. - http://git.io/41S54g
21:45:16  <bnoordhuis>Samba 3.0.x to 3.6.3 are affected by a vulnerability that allows remote code execution as the "root" user. <- awesome
21:45:36  <bnoordhuis>3.0.x. so that hole must've existed for close to 10 years...
21:45:43  <AvianFlu>hahahahha awesome
21:45:53  <mmalecki>well, that's what you get for using anything created for windows.
21:46:15  <tjfontaine>yes, it also doesn't requier authentication with samba to achieve it, quite troublesome
21:46:25  <bnoordhuis>yep. time to upgrade asap
21:46:59  <mmalecki>or rethink your software stack :)
21:47:25  <AvianFlu>I mean, last I checked, that came preinstalled on a lot of linux distros
21:47:56  <mmalecki>and change your distro, if it comes with samba preinstalled :D
21:48:51  <TooTallNate>but it's not like you're exposing your samba shared over the internet
21:48:58  * mikealquit (Quit: Leaving.)
21:48:59  <bnoordhuis>hah
21:49:18  <bnoordhuis>you're cute when you're naive, nathan
21:49:25  <tjfontaine>indeed
21:49:36  <tjfontaine>also consider if you run a public network, like an .edu
21:49:58  <TooTallNate>true. guess i've just never used SMB except for at home :D
21:51:53  * mikealjoined
21:52:01  <TooTallNate>bnoordhuis: i don't know what's different now, but even with -mmacosx-version-min=10.5 now < /dev/tty is hanging for me
21:52:28  <bnoordhuis>TooTallNate: on 10.5 or later?
21:52:36  <TooTallNate>bnoordhuis: Lion
21:54:05  <bnoordhuis>TooTallNate: what was the full line again?
21:54:19  <TooTallNate>bnoordhuis: `node < /dev/tty`
21:54:38  <TooTallNate>so weird; it was just working the other day...
21:54:49  <bnoordhuis>right
21:54:58  <bnoordhuis>echo "out/Release/node < /dev/tty" | sh prints >, then hangs
21:55:06  <TooTallNate>yup
22:03:52  <bnoordhuis>why doesn't os x have pidof or pkill? *rage*
22:04:37  <mmalecki>bnoordhuis: killall
22:08:02  <isaacs>man, javascript is super annoying sometimes.
22:08:34  <isaacs>i've managed to get almost all the necessary features of domains implemented, without performance falling off a cliff, except for the auto-adding of EventEmitters to the active domain.
22:08:44  <isaacs>that cuts http_simple bytes/1024 in half.
22:08:59  <isaacs>Thou Shalt Not Modify the EventEmitter Constructor.
22:15:14  * pfox___joined
22:32:23  * rendarquit
22:40:51  * mmaleckichanged nick to mmalecki[zzz]
22:42:08  * elijah-mbpquit (Remote host closed the connection)
22:42:19  * travis-cijoined
22:42:19  <travis-ci>[travis-ci] joyent/libuv#183 (master - 3c41597 : Ben Noordhuis): The build is still failing.
22:42:19  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/8e6f332...3c41597
22:42:19  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1061173
22:42:19  * travis-cipart
22:46:33  * piscisaureus_joined
22:51:10  * pfox___quit (Ping timeout: 252 seconds)
22:58:36  * dylukesquit (Quit: Pipes are broken. Sending packets via Fedex.)
22:59:13  * mikealquit (Quit: Leaving.)
23:03:34  * pijewskijoined
23:10:27  * paddybyersquit (Quit: paddybyers)
23:11:17  * mikealjoined
23:11:22  <bnoordhuis>piscisaureus_: ping
23:11:37  <piscisaureus_>bnoordhuis: bing
23:11:47  <bnoordhuis>piscisaureus_: drinking the ms koolaid?
23:12:02  <bnoordhuis>can you take a look at my openssl branch?
23:12:06  <bnoordhuis>i can't get it to compile on windows
23:12:19  <bnoordhuis>linux, sunos and os x finally work though
23:12:19  <piscisaureus_>bnoordhuis: http://www.youtube.com/watch?v=7GM4Lt5k24s
23:12:40  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/node/compare/openssl <- warning: big changeset
23:12:54  <bnoordhuis>1553 files changed, 176087 insertions(+), 59957 deletions(-)
23:13:02  <piscisaureus_>bnoordhuis: what changes?
23:13:19  <bnoordhuis>piscisaureus_: upgrade openssl to 1.0.0f
23:13:40  <bnoordhuis>plus assorted workarounds to get it actually compile on various platforms :/
23:13:47  <bnoordhuis>plus fixing some bad tests
23:13:49  <piscisaureus_>bnoordhuis: doing if for the shortstats? :-p
23:14:05  <bnoordhuis>plus fixing a bug in gyp
23:14:11  <bnoordhuis>piscisaureus_: yes :)
23:14:37  <piscisaureus_>ok
23:14:38  <piscisaureus_>cloning
23:15:08  <piscisaureus_>got bored of WSAEventSelect woes anyway
23:15:38  <piscisaureus_>and hacker news
23:15:50  <bnoordhuis>there's always good ol' slashdot
23:17:24  <pijewski>Hi everyone, I'm working on a HTTP file server and want to benchmark node vs. nginx. I've been talking with isaacs about how to get the file descriptor out of an HTTP request for use with fs.sendfile(). How would I go about that with v0.6/v0.7?
23:17:57  <bnoordhuis>pijewski: you don't, it's not exposed
23:18:09  <piscisaureus_>bnoordhuis: so what do we have sendfile for then? :-p
23:18:28  <pijewski>is there a way to do this in a non-portable, hacked-up way? :-) FWIW, I'm running on SunOS
23:18:30  <bnoordhuis>piscisaureus_: nothing really, it's pretty much useless now
23:19:11  <bnoordhuis>pijewski: yes. you could expose uv_stream_t->fd
23:20:11  <bnoordhuis>pijewski: it works like this: the net.Socket object has a .handle property which corresponds to a StreamWrap object
23:20:38  <bnoordhuis>you can find it in src/stream_wrap.cc
23:21:50  <pijewski>cool; I'll poke around in there
23:22:14  <bnoordhuis>pijewski: have fun :) let me know if you want some more pointers
23:22:18  <piscisaureus_>bnoordhuis: what a mess :-/
23:22:33  <bnoordhuis>piscisaureus_: ?
23:22:42  <piscisaureus_>bnoordhuis: compiling that thing on windows
23:22:46  <mjr_>Doesn't DTrace need that fd exposed also in order to work well?
23:22:50  <bnoordhuis>piscisaureus_: yeah. i couldn't get it to work
23:23:10  <bnoordhuis>mjr_: yes, but that takes place on a different level. it's not js land
23:23:22  <bnoordhuis>mjr_: you should see dap's latest patches
23:23:59  <mjr_>All of this good stuff is happening on 0.7, but I haven't been able to check out yet.
23:24:14  <mjr_>We are still back in the stone ages on whatever the latest 0.6 is.
23:24:39  * perezdquit (Quit: perezd)
23:27:19  <bnoordhuis>saghul: feel like beer this friday?
23:43:01  * trondnjoined
23:47:50  <isaacs>bnoordhuis: hey, so, mjr_ and friends are seeing some weird tls stuff.
23:47:57  <isaacs>bnoordhuis: bryan just showed me this: https://github.com/isaacs/node-lru-cache
23:48:01  <isaacs>er, wrong thing
23:48:05  <isaacs>bnoordhuis: https://gist.github.com/2355625
23:48:26  <mjr_>I did indeed do that.
23:49:05  <isaacs>bnoordhuis: so, this gist is a sample of the jsstack whenever the heap break point is pushed out.
23:49:28  <mjr_>Updated
23:49:29  <mjr_>https://gist.github.com/2355720
23:49:41  * mikealquit (Quit: Leaving.)
23:49:47  <trondn>If I where to start playing with libuv… which branch should I use?
23:49:58  <mjr_>That other one ran for longer, but it had other processes included instead of just the 8 production processes.
23:53:10  <mjr_>I wish there was a way to start a node program and tell V8, go ahead and compute line numbers on everything right now.
23:53:38  * mikealjoined
23:53:44  * perezdjoined
23:57:29  <isaacs>mjr_: what version of node are you using?
23:57:32  <isaacs>0.6.14?
23:58:21  <mjr_>0.6.15
23:58:57  <isaacs>ok, cool
23:59:51  <mjr_>So it's always when we are writing something back to an HTTPS client.