00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:03  <trevnorris>what do you mean "how they were wanting to achieve it"?
00:00:09  * ircretaryjoined
00:00:41  <tjfontaine>just in what their expected requirements for those mechanisms were and how template rendering was being delievered for their platforms
00:01:12  <othiym23>IRHydra is for different use cases than dtrace / perf
00:01:26  <othiym23>it has been very useful for me in understanding deopt behavior with monkeypatching, for example
00:01:35  <tjfontaine>also -- undisclosed enterprise -- has a new employee specifically just for performance
00:01:48  <othiym23>it's definitely more a developer tool than an ops tool, but that doesn't mean it's not a really valuable thing to have around in dealing with operational issues
00:01:59  <othiym23>also JADE SUCKS
00:02:03  <othiym23>that's right, I said it
00:02:19  <othiym23>no offence to Forbes, I'm sure the work he's done on the module is great
00:03:53  * jmar777joined
00:04:15  <trevnorris>ok. well, i'm still missing how this is an argument to not allow the irhydra2 patches into v0.12. for kicks I did an upgrade to 3.24.40 and it compiled and tested fine. not saying we bring that in wholesale, but that backporting the patch should be simple.
00:04:43  <othiym23>+1 on getting those patches in
00:05:00  <tjfontaine>the easiest way to get them in to node 0.12 is to have v8 backport htem
00:05:20  <tjfontaine>otherwise, the people who are equipped to understand the output are also the people equipped to recompile node with them though
00:05:59  <othiym23>I think the same rationale that applies to getting node-tick-processor into the default distro applies here
00:06:19  <othiym23>I also think making these tools more accessible will also help get them more widely used
00:06:26  <trevnorris>ok, i'm really confused about node-tick-processor. what the crap is it?
00:06:35  <tjfontaine>that's a different scenario, and is rather non controversial, though the applicability of applying the output is really suspect
00:06:49  <tjfontaine>the tick processor is already somethign we can build today
00:06:57  <othiym23>trevnorris: a hacked-up version of the tick-processor that comes with v8 to use with d8
00:07:07  <trevnorris>ah, ok
00:07:22  <othiym23>for using with the profiler
00:08:00  <othiym23>tjfontaine: I'm not advocating the maintenance of a large floating patch, but I do think small patches to support tooling should not be super controversial
00:08:21  <tjfontaine>certainly, the tick processor is something we can and probalby should ship, since we're in the best position to do so
00:08:38  <tjfontaine>we have precedence for including patches that help with dtrace, etw, and systemtap as well
00:08:51  <othiym23>if I'd had access to something like IRHydra when I was doing serious Java development, I FOR SURE would have wanted access to it
00:09:06  <tjfontaine>I'm not opposed to adding convenient things, but just like I'm not trying to get ahead of v8 for harmony, I'm not trying to get ahead of it for irhydra
00:09:07  <othiym23>I'm not really sure there's even anything today that will give you a similar level of insight into what HotSpot is up to
00:09:26  <othiym23>tjfontaine: I'm pretty sure Slava just made his changes on whatever happened to be HEAD at the time
00:09:33  <othiym23>but I guess at this point I'm just whistling into the wind
00:10:02  <tjfontaine>the point is that we're quickly going to be converging into a world where we have more opportunity outside v8
00:11:45  <othiym23>"opportunity outside v8"?
00:14:09  <trevnorris>and v8 is also giving us awesome abilities, like gdbjit
00:14:38  <trevnorris>freak I hate icu
00:14:58  * mikealquit (Quit: Leaving.)
00:17:08  * mikealjoined
00:17:23  * thlorenzjoined
00:17:38  * mcavagequit (Ping timeout: 240 seconds)
00:19:02  * tjkrusinskijoined
00:19:17  * mikealquit (Client Quit)
00:20:02  * mikealjoined
00:23:10  * bradleymeck_joined
00:27:25  * mikealquit (Quit: Leaving.)
00:36:20  * mikealjoined
00:37:49  * cphooverquit (Quit: Leaving.)
00:41:11  * mikealquit (Client Quit)
00:41:40  <isaacs>tjfontaine: (or anyone really) wanna lgtm the url parse bugfix?
00:42:03  <isaacs>https://gist.github.com/isaacs/27d1ca6fa4c53992da24
00:43:07  * paulfryzelquit (Remote host closed the connection)
00:44:19  * mikealjoined
00:45:12  <MI6>joyent/node: isaacs v0.10 * bd547d6 : npm: upgrade to 1.4.6 - http://git.io/coj3uw
00:45:20  <isaacs>tjfontaine: changelog'ed ^
00:46:41  * eugenewarejoined
00:46:58  * mikealquit (Client Quit)
00:47:25  * hzquit
00:49:00  * rmgquit (Remote host closed the connection)
00:49:14  * Kakeraquit (Ping timeout: 246 seconds)
00:49:38  * mikealjoined
00:52:05  * mikealquit (Client Quit)
00:54:31  * mikealjoined
00:59:22  * mikealquit (Client Quit)
00:59:24  * hzjoined
01:02:57  * rosskquit
01:10:14  * seldoquit (Remote host closed the connection)
01:11:11  * bradleymeckquit (Ping timeout: 252 seconds)
01:11:12  * bradleymeck_changed nick to bradleymeck
01:19:43  * rmgjoined
01:20:44  * bradleymeckquit (Quit: bradleymeck)
01:24:20  * tjkrusinskiquit (Ping timeout: 252 seconds)
01:26:11  * rmgquit (Ping timeout: 252 seconds)
01:30:54  * c4milojoined
01:31:02  * kevinsimperjoined
01:35:47  * kevinsimperquit (Ping timeout: 265 seconds)
01:35:53  * c4miloquit (Ping timeout: 268 seconds)
01:44:00  * paulfryzeljoined
01:44:36  * brunklequit (Quit: brunkle)
01:48:23  * paulfryzelquit (Ping timeout: 246 seconds)
01:53:17  * wolfeidauquit (Ping timeout: 246 seconds)
01:56:28  * wolfeidaujoined
01:58:26  * tjkrusinskijoined
02:02:24  * wolfeidauquit (Ping timeout: 268 seconds)
02:03:20  * tjkrusinskiquit (Ping timeout: 265 seconds)
02:04:57  * wolfeidaujoined
02:05:00  * m76quit (Read error: Connection reset by peer)
02:06:58  * mikolalysenkojoined
02:08:01  * eugenewarequit (Read error: Connection reset by peer)
02:10:16  * nickleeflyjoined
02:14:58  * wolfeidauquit (Ping timeout: 240 seconds)
02:19:51  * thlorenzquit (Remote host closed the connection)
02:20:26  * thlorenzjoined
02:21:47  * jmar777quit (Remote host closed the connection)
02:24:53  * thlorenzquit (Ping timeout: 252 seconds)
02:26:28  * perezdjoined
02:37:42  * wolfeidaujoined
02:43:26  * perezdquit (Quit: perezd)
02:44:28  * eugenewarejoined
02:45:01  * paulfryzeljoined
02:47:52  * paulfryz_joined
02:48:05  * tdmackeyjoined
02:49:17  * paulfryzelquit (Ping timeout: 246 seconds)
02:49:53  * eugenewarequit (Ping timeout: 268 seconds)
02:52:09  * perezdjoined
02:59:13  * tjkrusinskijoined
03:04:07  * tjkrusinskiquit (Ping timeout: 264 seconds)
03:07:05  * rmgjoined
03:17:22  * mikealjoined
03:19:57  * c4milojoined
03:20:41  * eugenewarejoined
03:24:07  <trevnorris>tjfontaine: not that it matters, but i'm getting a lot of fatal errors from v8 on the ExtensionsRegistration and MultithreadedParallelIsolates tests.
03:25:02  * c4miloquit (Ping timeout: 268 seconds)
03:26:35  * paulfryz_quit (Remote host closed the connection)
03:28:06  * rmgquit (Remote host closed the connection)
03:32:00  * kevinsimperjoined
03:36:37  * kevinsimperquit (Ping timeout: 265 seconds)
03:45:43  * paulfryzeljoined
03:46:10  * petka_quit (Quit: Connection closed for inactivity)
03:49:15  * mikealquit (Quit: Leaving.)
03:49:50  * paulfryzelquit (Ping timeout: 246 seconds)
04:00:00  * tjkrusinskijoined
04:01:08  <trevnorris>indutny: have an idea how I submit a backport to v8?
04:05:07  * tjkrusinskiquit (Ping timeout: 268 seconds)
04:05:25  * hzquit
04:06:22  * mikealjoined
04:07:02  * rmgjoined
04:14:54  * nickleeflyquit (Quit: Connection closed for inactivity)
04:14:59  * mikolalysenkoquit (Ping timeout: 268 seconds)
04:17:24  * rmgquit (Ping timeout: 252 seconds)
04:18:23  * mikealquit (Quit: Leaving.)
04:19:30  * mikealjoined
04:20:02  * mikealquit (Client Quit)
04:24:46  * jirwinquit (Ping timeout: 245 seconds)
04:26:17  <trevnorris>tjfontaine / indutny: https://codereview.chromium.org/211393007/
04:27:48  * jirwinjoined
04:44:53  * eugenewarequit (Remote host closed the connection)
04:45:26  * eugenewarejoined
04:49:04  * mikealjoined
04:52:43  * AvianFluquit (Remote host closed the connection)
04:53:13  * brsonquit (Ping timeout: 268 seconds)
04:53:13  * AvianFlujoined
04:58:09  * AvianFluquit (Ping timeout: 268 seconds)
05:00:20  <trevnorris>hueniverse: here's the PR i'm working on: https://github.com/joyent/node/pull/7145
05:09:02  * c4milojoined
05:13:46  * c4miloquit (Ping timeout: 265 seconds)
05:21:38  * brsonjoined
05:32:42  * kevinsimperjoined
05:37:17  * kevinsimperquit (Ping timeout: 246 seconds)
05:48:45  * perezdquit (Quit: perezd)
05:49:56  * eugenewa_joined
05:52:47  * eugenewarequit (Ping timeout: 252 seconds)
06:08:15  * eugenewa_quit (Read error: Connection reset by peer)
06:09:29  * wolfeidauquit (Remote host closed the connection)
06:13:02  * mrhoorayjoined
06:26:15  * eugenewarejoined
06:47:57  * paulfryzeljoined
06:52:52  * paulfryzelquit (Ping timeout: 265 seconds)
06:57:44  * c4milojoined
07:02:38  * c4miloquit (Ping timeout: 252 seconds)
07:07:35  * janjongboomquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
07:07:35  * eugenewarequit (Read error: Connection reset by peer)
07:10:06  * petka_joined
07:26:38  * brsonquit (Ping timeout: 240 seconds)
07:30:39  <indutny>morning
07:35:11  * m76joined
07:36:12  * c4milojoined
07:39:44  <MI6>joyent/node: Ben Noordhuis v0.10 * f68a116 : src: ensure that openssl's PRNG is fully seeded (+1 more commits) - http://git.io/mwl8uw
07:41:38  * m76quit (Read error: Connection reset by peer)
07:45:13  * m76joined
07:46:04  * janjongboomjoined
07:48:40  * paulfryzeljoined
07:52:13  <MI6>joyent/node: Fedor Indutny master * e781832 : crypto: fix lint (+5 more commits) - http://git.io/TuNzHg
07:53:05  * paulfryzelquit (Ping timeout: 246 seconds)
07:55:19  * kevinsimperjoined
08:02:27  * mrhoorayquit
08:07:12  <indutny>saghul: hey man
08:07:22  <indutny>how are you?
08:07:23  <saghul>indutny: ohai!
08:07:30  <saghul>how are you?
08:07:36  <indutny>good
08:08:20  <indutny>and you? :)
08:08:50  <saghul>good too!
08:09:27  <indutny>so
08:09:36  <indutny>what are you working on? ;)
08:09:38  <indutny>need any help?
08:09:53  <saghul>I think I'm going to land #1206 today
08:10:07  <saghul>and then move on to the other UDP/multicast issues
08:10:14  <indutny>ok, great
08:10:15  <saghul>you?
08:10:31  <saghul>oh, there is a thing you can help with, actually
08:10:34  <indutny>well, perhaps I'll add parallel move to uv_spawn()
08:10:37  <indutny>saghul: really? :)
08:10:38  <indutny>cool
08:10:41  <indutny>what is it?
08:10:51  <saghul>https://github.com/joyent/libuv/pull/1210
08:11:04  <indutny>you want me to test it?
08:11:07  <saghul>can you see it that fixes the pm2 issue? I couldn't reproduce it
08:11:11  <indutny>sure
08:11:13  <saghul>yes, please :-)
08:11:18  <saghul>kewl, thanks!
08:11:39  <indutny>np
08:38:12  <indutny>saghul: can't reproduce it either
08:38:17  <indutny>guess the author has fixed it
08:38:20  <indutny>somehow
08:38:45  <saghul>uh
08:49:27  * paulfryzeljoined
08:53:38  * paulfryzelquit (Ping timeout: 246 seconds)
09:20:08  * inolenjoined
09:22:05  * inolen1quit (Ping timeout: 268 seconds)
09:23:27  * c4miloquit (Remote host closed the connection)
09:27:45  * c4milojoined
09:39:52  * c4miloquit (Remote host closed the connection)
09:50:10  * paulfryzeljoined
09:54:32  * paulfryzelquit (Ping timeout: 246 seconds)
09:57:00  * hzjoined
10:14:35  * m76quit (Ping timeout: 252 seconds)
10:23:32  * m76joined
10:26:41  * rmgjoined
10:31:07  * rmgquit (Ping timeout: 264 seconds)
10:32:14  * AvianFlujoined
10:38:26  * mitsuhikoquit (Quit: ZNC - http://znc.sourceforge.net)
10:44:26  * mitsuhikojoined
10:47:14  * Kakerajoined
10:50:56  * paulfryzeljoined
10:55:30  * paulfryzelquit (Ping timeout: 265 seconds)
11:19:29  * mitsuhikoquit (Ping timeout: 252 seconds)
11:28:07  <roxlu>indutny: do you have time to explain some ssl stuff?
11:30:53  <indutny>hey man
11:30:56  * mitsuhikojoined
11:31:04  <indutny>sorry, I went afk yesterday
11:31:05  <indutny>shoot
11:31:19  <roxlu>no problem, thanks for taking time
11:31:44  <roxlu>is in here ok?
11:41:18  <roxlu>.. silence is YES right :)
11:42:12  <roxlu>I need to implement a DTLS server. I want to manage my own memory buffers and using a bio_method_st for this seems to be a good solution.
11:44:06  * wolfeidaujoined
11:45:17  * mitsuhikoquit (Ping timeout: 265 seconds)
11:45:40  <roxlu>Once I receive a ClientHello, I call SSL_read (after setting the input buffer that I use in my bio_read function). openssl will start it's handshake process and it calls the _bio_write() function of my bio_method_st. But this where I have an issue.. how am I supposed to provide a buffer for the output data that openssl creates?
11:51:36  * paulfryzeljoined
11:51:46  <indutny>щл
11:51:47  <indutny>ok
11:52:01  <indutny>have you seen SSL_set_bio() ?
11:52:02  <indutny>roxlu: ^
11:52:18  <roxlu>yes I'm using that
11:52:26  <indutny>then you have passed two bios to it
11:52:30  <indutny>one for output data
11:52:33  <indutny>and for input data
11:52:35  <roxlu>yes
11:53:27  * mitsuhikojoined
11:56:24  * paulfryzelquit (Ping timeout: 265 seconds)
11:59:21  * mitsuhikoquit (Ping timeout: 245 seconds)
12:05:57  * mitsuhikojoined
12:07:26  * janjongboomquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
12:07:29  * cphooverjoined
12:10:39  * eugenewarejoined
12:12:42  * m76quit (Read error: Connection reset by peer)
12:13:41  * m76joined
12:16:54  * cphooverquit (Quit: Leaving.)
12:20:16  * tjkrusinskijoined
12:28:59  * jmar777joined
12:32:47  * janjongboomjoined
12:35:57  * mitsuhikoquit (Quit: ZNC - http://znc.sourceforge.net)
12:37:57  * mitsuhikojoined
12:43:08  * m76quit (Ping timeout: 240 seconds)
12:49:54  * jmar777quit (Remote host closed the connection)
12:52:31  * paulfryzeljoined
12:57:18  * paulfryzelquit (Ping timeout: 265 seconds)
12:57:52  * janjongboomquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
13:32:50  * jmar777joined
13:36:33  * paulfryzeljoined
13:38:02  * thlorenzjoined
13:53:17  * paulfryz_joined
13:58:12  * paulfryz_quit (Ping timeout: 265 seconds)
14:02:36  * paulfryzelquit (Remote host closed the connection)
14:03:12  * paulfryzeljoined
14:07:43  * paulfryzelquit (Ping timeout: 264 seconds)
14:13:28  * m76joined
14:21:04  * eugenewarequit (Remote host closed the connection)
14:26:55  * tjkrusinskiquit (Ping timeout: 264 seconds)
14:31:32  * jmar777quit (Read error: Connection reset by peer)
14:32:00  * jmar777joined
14:33:52  * paulfryzeljoined
14:35:38  * mikolalysenkojoined
14:54:03  * paulfryz_joined
14:55:05  * paulfryzelquit (Read error: No route to host)
14:55:37  * paulfryzeljoined
14:58:37  * paulfryz_quit (Ping timeout: 265 seconds)
15:00:11  * paulfryzelquit (Ping timeout: 245 seconds)
15:01:59  * kevinsimperquit (Remote host closed the connection)
15:02:35  * kevinsimperjoined
15:05:23  * rmgjoined
15:06:38  * kevinsimperquit (Ping timeout: 240 seconds)
15:06:42  * paulfryzeljoined
15:08:39  * eugenewarejoined
15:13:35  * eugenewarequit (Ping timeout: 268 seconds)
15:22:18  * eugenewarejoined
15:24:09  * kkaeferjoined
15:24:15  <kkaefer>good afternoon
15:25:01  <kkaefer>I have two threads, A and B. Thread B runs a thread loop. Can thread A create a uv_timer and target it at thread B's loop?
15:25:13  * burlakquit (Ping timeout: 265 seconds)
15:27:11  <saghul>kkaefer: it's not thread safe to do that
15:27:19  * paulfryzelquit (Read error: Connection reset by peer)
15:27:28  <saghul>the only thread safe function is uv_async_send
15:27:44  <kkaefer>ok, so I'd have to create a uv_async and send it to thread B, create the timer in there?
15:27:46  * paulfryzeljoined
15:34:33  * hueniversequit (Ping timeout: 268 seconds)
15:36:11  * burlakjoined
15:40:41  * burlakquit (Ping timeout: 265 seconds)
15:42:11  * burlakjoined
15:42:17  <roxlu>indutny: my ssl server replies only with: ServerHello, Certificate, Certificat Request and ServerHello Done. No "Server Key Exchange" is done. Do you maybe have an idea what that might cause?
15:42:46  * bradleymeckjoined
15:43:03  * bradleymeckquit (Client Quit)
15:43:25  * bradleymeckjoined
15:43:55  * mikealquit (Quit: Leaving.)
15:45:03  * mikealjoined
15:46:14  * brunklejoined
15:54:45  * jmar777quit (Remote host closed the connection)
15:55:40  * eugenewarequit (Ping timeout: 265 seconds)
15:56:25  * paulfryz_joined
15:57:51  * mikealquit (Quit: Leaving.)
16:00:32  * m76quit (Ping timeout: 252 seconds)
16:01:05  * paulfryz_quit (Ping timeout: 252 seconds)
16:03:17  * hueniversejoined
16:08:59  * rmgquit (Remote host closed the connection)
16:12:07  * rosskjoined
16:20:10  * rmgjoined
16:37:04  * paulfryz_joined
16:39:27  * sblomquit (Read error: Connection reset by peer)
16:41:26  * paulfryz_quit (Ping timeout: 245 seconds)
16:42:44  <MI6>joyent/libuv: Tonis Tiigi v0.10 * fd77a5d : process: don't close stdio fds during spawn - http://git.io/yhbYYA
16:45:24  <MI6>joyent/libuv: Fedor Indutny master * c84796e : Merge branch 'v0.10' (+3 more commits) - http://git.io/gnRpBA
16:45:48  <indutny>roxlu: is it present in DTLS?
16:47:08  <bradleymeck>indutny: you did the zlib sync stuff in 0.11, is there a way to get that in 0.10 that I am unaware of (even if internal functions)
16:47:23  <indutny>that wasn't me :)
16:47:32  <tjfontaine>do you mean one off backport
16:47:35  <indutny>no way
16:47:39  <tjfontaine>or add it to an official release?
16:47:42  <indutny>this is against our principle
16:47:47  <indutny>tjfontaine: isn't it?
16:48:07  <tjfontaine>it is, which is why I'm trying to figure out what he wants :)
16:48:08  <bradleymeck>indutny: no, I was just wondering if it was hidden and not exposed somewhere
16:48:16  <indutny>nope
16:48:32  <bradleymeck>JS only solution works for those, just is slow
16:48:43  <roxlu>indutny: (on phone)
16:50:23  * jmar777joined
16:51:19  <Kakera>glad to see it's useful to someone :)
16:52:30  <bradleymeck>tjfontaine: working on moving this all out to _third_party_main.js to get a 0.10 single binary release available, new beta build probably tonight
16:52:56  <bradleymeck>not going to write up the archive loader and see what people complain is broken
16:53:00  <bradleymeck>about*
16:56:47  * calvinfojoined
16:58:14  * mikealjoined
17:02:55  * mikealquit (Ping timeout: 264 seconds)
17:03:52  <indutny>mmalecki: hey man
17:04:09  <indutny>how are you?
17:04:42  <indutny>trevnorris: hey man
17:04:50  <indutny>how are you? :)
17:04:51  <indutny>hehe
17:09:54  * bajtosjoined
17:22:38  * m76joined
17:26:08  <roxlu>indutny: just got off the phone ... is it okay if I ping you later today? (or tomorrow)
17:26:25  <indutny>sure
17:26:50  * mikealjoined
17:26:54  <roxlu>thanks
17:30:07  <indutny>np
17:34:38  * kpdeckerjoined
17:35:08  * seldojoined
17:37:28  <mmalecki>indutny: hey
17:37:32  <mmalecki>indutny: I'm good, you?
17:37:35  <indutny>good too
17:37:48  <indutny>reading about differential equations :P
17:37:54  * paulfryz_joined
17:38:00  <indutny>mmalecki: have you got some time for bud's tests?
17:38:09  <indutny>just curious
17:38:10  <indutny>:)
17:38:37  <indutny>I'm tempted to start working on this thing
17:38:43  <indutny>if you don't have time for it, that's fine :P
17:38:54  <indutny>its just that I am terribly procrastinating
17:39:01  <indutny>and want to pick up anything
17:39:09  <indutny>but on other hand it would be great to have you involved
17:40:40  <indutny>long monologue
17:42:29  * paulfryz_quit (Ping timeout: 265 seconds)
17:44:47  <tjfontaine>indutny: btw, do you have some bandwidth to add a buidlsystem feature?
17:44:55  <indutny>certainly
17:44:59  <indutny>what are we talking about?
17:45:13  <tjfontaine>indutny: ./configure --v8-options="--abort-on-uncaught-exception,--always-opt"
17:45:29  <tjfontaine>indutny: at build time define the SetFlags() that people would want
17:45:39  <tjfontaine>indutny: so if someone wants to enable harmony at compile time
17:45:47  <indutny>hm...
17:45:56  <indutny>sounds like a matter of adding one define
17:46:00  <othiym23>that seems fraught
17:46:17  <tjfontaine>othiym23: you mean people distributing their own versions of node?
17:47:08  <othiym23>yeah, with harmony enabled
17:47:14  <othiym23>I worry about fragmentation
17:47:38  <tjfontaine>well, we're basically already there -- the number of conversations I have to have about trademark ...
17:47:59  <tjfontaine>the reality is this is why chrome/chromium firefox/iceweasel also exist
17:48:36  <indutny>nodeharm
17:48:37  <indutny>:D
17:48:40  <tjfontaine>:P
17:48:44  <tjfontaine>I still love that name
17:48:48  <tjfontaine>or noduh
17:49:21  <indutny>ah, yeah
17:49:26  <indutny>anyway
17:49:28  <indutny>let's do it
17:50:14  <tjfontaine>I actually think there are some v8 flags we probably should be enabling for our release builds
17:51:34  <indutny>example?
17:53:22  * eugenewarejoined
17:54:05  * calvinfoquit (Read error: Connection reset by peer)
17:54:25  * calvinfojoined
17:54:45  <bradleymeck>othiym23: fragmentation will continue if we do get the bundler/single executable working, but even with private distros ala SL I am not sure it will cause many waves
17:55:36  <tjfontaine>indutny: potentially --always_opt (always try to optimize functions)
17:56:01  <indutny>no-no
17:56:02  <indutny>no
17:56:04  <indutny>nonono
17:56:07  <indutny>fuck it
17:56:09  <tjfontaine>I haven't looked at what it does
17:56:12  <indutny>it is a waste of time
17:56:22  <indutny>it is really bad for everything
17:56:47  <tjfontaine>well I don't know, I am just worried we may end up tweaking for us as we're a long running process
17:57:02  <tjfontaine>not that we can't do it in code, I would like to do it more generically
17:58:07  * eugenewarequit (Ping timeout: 264 seconds)
17:58:18  <trevnorris>indutny: sup dude?
17:58:20  <indutny>tjfontaine: I'm totally against it
17:58:29  <indutny>tjfontaine: we are both cli and long-running thing
17:58:36  <indutny>there is no common thing here
17:58:38  <indutny>trevnorris: hey man
17:58:45  <indutny>trevnorris: what is a status of AL?
17:58:51  <indutny>are you waiting for NR to test it?
17:59:11  <trevnorris>dead. i erased all my code. i'm threw with it. :P
17:59:28  <indutny>ooook
17:59:37  <indutny>I think last time tjfontaine has updated me on it
17:59:44  <indutny>we was waiting for you to squash something
17:59:50  <indutny>is there any blockers for you?
17:59:56  <indutny>anything I could help you with?
18:00:32  <trevnorris>yeah. could you just do a quick overview and see if I missed anything obvious.
18:00:44  <indutny>sure
18:00:45  <indutny>link?
18:01:04  <trevnorris>https://github.com/joyent/node/pull/7145
18:01:35  <trevnorris>my blocker is a failing NR test
18:02:03  * brunklequit (Quit: brunkle)
18:02:11  <trevnorris>and there're so many layers of code between the test and AL that i'm having a hard time pin it down.
18:02:14  <tjfontaine>that's not actually a blocker, we just need to get through review and squash and land
18:02:37  <trevnorris>tjfontaine: your mom's my blocker
18:02:45  <trevnorris>sorry, was up until 3 last night trying to figure this out
18:03:04  * mikealquit (Quit: Leaving.)
18:03:18  <trevnorris>tjfontaine: so, I opened a PR to v8 w/ the backport. it was refused. looks like 3.25 is coming to the end of its cycle and they'll be on 3.26 soon
18:03:27  <trevnorris>he said in about 2 weeks
18:03:40  <indutny>oh
18:03:43  <indutny>well
18:03:53  * mikealjoined
18:05:32  <trevnorris>tjfontaine: also, what do you think about the failing v8 tests?
18:05:58  * tfogalpart
18:07:02  <trevnorris>tjfontaine / indutny: also I think it could be useful to add the gyp flag v8_enable_disassembler=1 to the build
18:07:48  * mikealquit (Client Quit)
18:07:57  * mikealjoined
18:11:09  <tjfontaine>I dont' see that being very useful by default in our builds, but didn't I already change the build such that you can easily add arbitrary defines at compile time?
18:13:06  <tjfontaine>./configure -- -Dv8_enable_disassembler=1
18:13:50  <trevnorris>ah, really? I've been doing export GYP_DEFINES+=" v8_enable_disassembler=1"
18:13:57  <trevnorris>cool
18:14:37  <tjfontaine>it does pass to gyp_node so if that doesn't work for what you need let me know
18:16:44  <trevnorris>will do
18:18:37  <trevnorris>man, figuring out how to properly squash 30 commits is sort of a pain. but it is in progress.
18:19:19  <trevnorris>tjfontaine: I suggest we backport that patch to 3.24.35. I already have and have run every v8 release and debug test. everything works fine
18:19:28  * sblomjoined
18:19:29  <trevnorris>oh, and yeah. the -D thing works
18:21:30  <tjfontaine>link me again?
18:22:28  <tjfontaine>https://codereview.chromium.org/211393007/
18:22:28  <tjfontaine>?
18:23:08  * rosskquit (Ping timeout: 240 seconds)
18:23:30  * rosskjoined
18:24:47  <tjfontaine>if that's what you're referring to, it's non-trivial to float, it's fine if you want to keep it around for doing your own stuff
18:25:39  <tjfontaine>there's not even a good way to hide that behind a build option, it's invasive to hydrogen.cc
18:27:51  <roxlu>indutny: hey I'm back
18:28:24  <roxlu>this is the captured data I get with wireshark: https://gist.github.com/roxlu/763aaeaea848a8d125ff (which fails)
18:32:00  * sh1mmerquit (Quit: sh1mmer)
18:33:46  <trevnorris>tjfontaine: i'm confused about why you have a problem w/ the patch. it's now part of 3.25, and when we do eventually upgrade v8 again it will be there.
18:34:14  <trevnorris>and it is trivial to float. except for one minor exception it applies cleanly
18:36:07  <MI6>joyent/node: trevnorris created branch hgraph-backport - http://git.io/ozk6mA
18:36:11  <trevnorris>tjfontaine: ^
18:36:18  <tjfontaine>it is not trivial to float if v8 starts touching hydrogen for its own fixes
18:36:27  <tjfontaine>which means we can't do clean backports of their 3.24 patches
18:36:41  <tjfontaine>we always have to make sure we're also not breaking the irhydra2 patch
18:36:57  * mikealquit (Quit: Leaving.)
18:37:05  <tjfontaine>also this is a crucial part of how v8 actually works, and floating this patch is *not* a massive win for the majority of people running node today
18:37:12  <tjfontaine>so there's no critical need for us to float it
18:38:21  <trevnorris>well peachy.
18:38:38  * paulfryz_joined
18:38:42  * brunklejoined
18:38:48  * sh1mmerjoined
18:39:18  * sh1mmerquit (Client Quit)
18:39:48  <trevnorris>but, ya know, I could say the same about the majority of dtrace probes being place into core, if we're talking about the "majority"
18:40:04  * sh1mmerjoined
18:40:09  <tjfontaine>those probes are not in a dependency we don't control
18:40:19  <tjfontaine>they are in our source, and work with etw and systemtap as well
18:40:40  <tjfontaine>they are applicable generically for everyone running node in production
18:40:42  <trevnorris>you mean mdb_v8?
18:41:07  <tjfontaine>mdb_v8 has nothing to do with dtrace, and is built and included into the node binary but has no runtime effect on node
18:41:44  <tjfontaine>and is only included on sunos builds, not linux, darwin, win32, etc
18:41:56  <tjfontaine>because that's the only place libproc and mdb exist
18:42:29  <tjfontaine>if there were facilities that helped debugging of node in production for linux and darwin etc that needed build time support from node they would be considered equally for inclusion
18:43:05  * paulfryz_quit (Ping timeout: 252 seconds)
18:43:15  <tjfontaine>if the irhydra patch changed things in the postmortem generation script for v8, which is a python file that is run at build time, we could work on floating it there
18:43:26  <tjfontaine>floating irhydra2 patch otoh means touching codegen paths
18:43:51  <tjfontaine>which is a dangerous conversation, unless there are clear gains that everyone is going to feel, and that v8 has already included upstream
18:44:27  <trevnorris>um... v8 has already included upstream.
18:44:45  <trevnorris>next time we upgrade v8, past 3.24.35, that patch will be there
18:44:54  <tjfontaine>yes, and you'll be free to use it at that time
18:45:29  <tjfontaine>but backporting to 3.24 is non-trivial, because it touches codegen, and could get messy when v8 backports other fixes that node also needs to include
18:45:59  * yunongjoined
18:46:11  * petka_quit (Quit: Connection closed for inactivity)
18:46:11  <othiym23>when core says "floating patch", does it mean just copying stuff into deps/v8 and then committing the diff?
18:46:19  <tjfontaine>yes
18:46:26  * sblomquit (Read error: Connection reset by peer)
18:46:33  <tjfontaine>floating a patch, generally refers to touching code we do not primarily maintain
18:46:36  <othiym23>what about using e.g. quilt to float the patch in a patches directory?
18:46:38  * sblomjoined
18:47:09  <tjfontaine>debian/ubuntu/fedora and other down stream providers are free to do what they want -- though at the peril of core deciding not to support requests for modified versions of node
18:47:21  <othiym23>sure
18:47:49  <tjfontaine>I will not entertain people complaining node is falling over because of some miscompile because of a broken merge of a v8 change
18:47:56  <othiym23>I'm just saying that it's easier to deal with the kinds of conflicts you're talking about if things like the irhydra2 patch are applied at build time on a clean copy of upstream v8
18:48:36  <tjfontaine>to an extent yes, though we try and run with 0 patches, so we only get into a world of quilt if we decide to fork v8 entirely
18:48:37  <othiym23>whether that build is the blessed distro or an individual version is sort of immaterial at that point, if the build-process is quilt-aware
18:48:48  <tjfontaine>node will not be adopting quilt
18:49:09  <tjfontaine>if we are in a situation where we need to float significant patches to support upstream, we will fork that project and maintain it ourselves
18:49:14  * brsonjoined
18:49:45  <tjfontaine>otherwise floating patches should only be considered if it's clear upstream is going to accept our change (or a version of it) in the next release of their project such that we can include the blessed version
18:49:58  <othiym23>hmmm
18:50:00  <othiym23>that's clear, at least
18:50:26  <othiym23>just looking at ways to get past the current logjam so both you and trevnorris can get what you want
18:50:52  * brunklequit (Quit: brunkle)
18:52:06  <tjfontaine>I don't see it as a logjam
18:55:40  <othiym23>"tension", then
18:55:53  <othiym23>I too would like to at least see irhydra2 land in 0.12
18:56:23  <tjfontaine>if the patch were not invasive to codegen, or coudl easily be hidden by a build flag, that might be possible
18:56:33  <othiym23>and being able to see the disasm of JITted code is at the very least interesting to me in some circumstances, although there I agree that it's not so useful in prod / postmortem, and I can use my own builds for that
18:56:50  <tjfontaine>disasm of jit'd code is always possible regardless of the tool
18:57:04  <tjfontaine>what you mean is something that keeps more information about why it was jitd in that way
18:57:09  <othiym23>y
19:01:49  <trevnorris>it's more than that. it includes the IC location in the code so you can quickly identify the operation that caused a deopt.
19:02:02  <trevnorris>instead of having to read though a load of disassembly
19:02:33  <trevnorris>indutny: there's one thing you can help me with
19:05:53  <trevnorris>indutny: I need a generic way to do this: https://github.com/trevnorris/node/commit/319f19d2538d4e5ea430e71f903fe4c7c655d0c1
19:06:32  <trevnorris>it's ugly, and also needs to be done in a few other places. but I don't want to litter code w/ that crap.
19:07:06  <trevnorris>indutny: it wouldn't be much an issue except the only way to do, say, new TCPWrap() is using ->NewInstance()
19:07:16  <trevnorris>which limits my ability to pass an argument
19:07:21  * sh1mmerquit (Quit: sh1mmer)
19:08:15  <trevnorris>tjfontaine: ^ that's my last blocker. get that and squashing the rest should be pretty easy
19:09:22  * sh1mmerjoined
19:16:06  * brunklejoined
19:24:04  * sh1mmerquit (Quit: sh1mmer)
19:27:11  * bajtosquit (Quit: bajtos)
19:27:18  <indutny>back
19:27:40  <indutny>trevnorris: looking
19:28:00  <trevnorris>indutny: ok. think I might have a fix. have to go to the doc for a bit but i'll post it when I get back
19:28:03  * sh1mmerjoined
19:28:17  <indutny>should I check 319f19d253 ?
19:29:54  <indutny>trevnorris: ^
19:30:43  * c4milojoined
19:30:57  <trevnorris>indutny: yeah. basically a generic way of passing the *_wrap from, say, OnConnection to the new *Wrap() call for TCP/UDP/Pipe
19:31:28  <roxlu>indutny: I got it working
19:31:32  <indutny>roxlu: great
19:31:42  <roxlu>the communication needs a fingerprint .. I was using a wrong one :#
19:31:58  * jmar777_joined
19:32:02  <indutny>trevnorris: you decided to take a simpler path? :)
19:32:19  <indutny>instead of using class based scope variable assignment
19:32:44  <trevnorris>the way I did it was a hack just to see if the idea would work
19:32:53  <trevnorris>but it's definitely not something I want to publish
19:32:54  <indutny>ok
19:33:17  <indutny>seems to be legit
19:35:03  * jmar777quit (Ping timeout: 268 seconds)
19:35:03  <trevnorris>problem is that it needs to happen for TCP/UDP/Pipe, and also there's a case in StreamWrap::AcceptHandle
19:36:37  <indutny>you are struggling with simpler common solution?
19:37:21  * mikealjoined
19:38:46  <trevnorris>ah, wait. think I have it.
19:38:56  <trevnorris>ok. need to run but bbiab
19:40:53  * kpdeckerquit (Quit: Leaving.)
19:41:38  * mikealquit (Ping timeout: 252 seconds)
19:41:55  <tjfontaine>dynamic_casts really aren't necessary, we don't have rtti
19:44:51  <trevnorris>i won't be doing any dynamic_casts
19:45:03  <trevnorris>besides, the linter doesn't allow it :P
19:45:10  <indutny>they are true evil
19:45:29  <indutny>the only thing that it is worth
19:45:38  <indutny>is a generics + virtual functions simultaneously
19:51:16  * stagasjoined
19:51:20  <indutny>tjfontaine: almost finished ./configure --v8-flags
19:55:47  * kpdeckerjoined
19:56:22  <tjfontaine>indutny: neat
20:08:20  <bradleymeck>worst monkey patch ive done in a while, but works on 0.10 now
20:08:49  * kpdeckerquit (Read error: Connection reset by peer)
20:09:07  * kpdeckerjoined
20:14:16  * petka_joined
20:34:26  <indutny>tjfontaine: https://github.com/joyent/node/pull/7362
20:34:39  * paulfryzelquit (Remote host closed the connection)
20:37:36  * benvie_joined
20:37:43  <tjfontaine>indutny, mmalecki: https://gist.github.com/tjfontaine/9771851 now node's initialization time in my forkbomb test is being limited by cpu throttle and not kstat contention
20:37:50  * mikealjoined
20:38:01  <tjfontaine>by far most of the Init() time is 200 microseconds
20:38:18  <tjfontaine>where as today even in best conditions it's 4ms
20:38:22  <indutny>nice!
20:38:28  <indutny>finally
20:39:40  * jmar777_quit (Remote host closed the connection)
20:39:56  * benviequit (Ping timeout: 246 seconds)
20:40:07  * paulfryzeljoined
20:40:42  <tjfontaine>harder to identify the total gain for !sunos, but everyone else is avoiding at least one extra syscall at startup
20:42:08  * mikealquit (Ping timeout: 240 seconds)
20:42:09  * mikeal1joined
20:42:17  * mikeal1quit (Client Quit)
20:44:49  * paulfryzelquit (Ping timeout: 252 seconds)
20:48:45  <trevnorris>indutny: ping
20:50:08  <trevnorris>indutny: in StreamWrap::AcceptHandle() is the WrapType of the uv_stream_t* pipe always going to match the WrapType that will run Instantiate?
20:50:35  <trevnorris>indutny: for reference: https://github.com/joyent/node/blob/master/src/stream_wrap.cc#L123-L140
20:51:19  * brunklequit (Quit: brunkle)
20:52:38  * brunklejoined
20:53:17  * tjkrusinskijoined
20:54:11  * mikealjoined
20:54:48  * paulfryzeljoined
20:57:24  <tjfontaine>you mean like https://github.com/joyent/node/blob/master/src/stream_wrap.cc#L718
20:57:24  <tjfontaine>?
21:00:24  <trevnorris>tjfontaine: yeah. i'm wondering if, in that specific case, handle->data will also always be a TCPWrap*. i'm pretty sure it will be, but just want to double check.
21:00:32  <indutny>trevnorris: back
21:00:35  <indutny>sorry, was afk
21:01:30  <indutny>trevnorris: it should, isn't it?
21:01:59  <trevnorris>indutny: i'm just making that assumption.
21:02:09  <indutny>ok
21:02:17  <trevnorris>and since there's no way I know of to assert that it is, just wanted to double check w/ you
21:03:03  <indutny>you could assert it
21:03:08  <indutny>using handle->type
21:03:15  <indutny>and adding static consts to classes
21:04:20  <trevnorris>ah, ok.
21:04:46  <indutny>if you want to
21:04:50  <trevnorris>but isn't handle a uv_stream_t* ?
21:04:52  <indutny>not sure if you really need to :)
21:04:57  <indutny>it is
21:05:05  <indutny>but uv_stream_t could be a uv_tcp_t
21:05:13  <indutny>or
21:05:14  <indutny>uv_pipe_t
21:05:21  <indutny>or even
21:05:24  <indutny>uv_udp_t
21:05:24  <trevnorris>yeah. does uv_stream_t store that info?
21:05:29  <indutny>yes, it does
21:05:33  <indutny>stream->type
21:05:35  <trevnorris>ah, cool. didn't know that. ok
21:05:36  <indutny>or handle->type
21:05:37  <indutny>whatever
21:05:43  <trevnorris>thanks
21:06:33  <indutny>np
21:06:35  <indutny>you are welcome
21:17:35  * paulfryzelquit (Remote host closed the connection)
21:20:55  * mikolalysenkoquit (Ping timeout: 264 seconds)
21:27:50  <trevnorris>indutny: ok, here's my general solution: https://github.com/trevnorris/node/commit/108d3490bbb812e8789a25e9ed4dc73c36a6c8a6
21:27:58  <trevnorris>mainly the part in env.h and env-inl.h
21:28:41  <indutny>eh
21:28:47  <indutny>why not a on-stack class instance? :)
21:28:57  <indutny>but seems to be fine anyway
21:29:06  <trevnorris>... um, what?
21:29:49  * sh1mmer_joined
21:31:06  * sh1mmerquit (Ping timeout: 265 seconds)
21:31:12  <trevnorris>indutny: i'm cool to change the code around, but first i'll need to know what you're talking about.
21:31:19  <indutny>haha
21:31:24  <indutny>so you are doing this
21:31:26  <indutny>set_parent()
21:31:30  <indutny>code
21:31:32  <indutny>remove_parent()
21:31:34  <indutny>you could do
21:31:42  <indutny>{ AutoSetParent(env, parent);
21:31:42  <indutny>code
21:31:43  <indutny>}
21:31:52  <indutny>actually
21:31:57  <indutny>AutoSetParent auto(env, parent);
21:32:17  <indutny>where AutoSetParent will call remove_parent() in destructor automatically
21:32:22  <indutny>RAII
21:32:26  <indutny>http://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization
21:32:45  <trevnorris>okie dokie. i'm reading up. :)
21:33:02  <indutny>haha
21:33:03  <indutny>ok
21:36:54  * sh1mmer_quit (Ping timeout: 265 seconds)
21:40:57  * paulfryzeljoined
21:45:19  * paulfryzelquit (Ping timeout: 252 seconds)
21:46:55  * mikolalysenkojoined
21:48:54  * mikealquit (Quit: Leaving.)
21:51:59  * mikolalysenkoquit (Ping timeout: 252 seconds)
21:55:12  * mikealjoined
22:02:15  <tjfontaine>indutny: that pattern is so broken, it's not even funny -- people should really be explicit about their code and not so automagical
22:02:18  <tjfontaine>anyway
22:02:25  <tjfontaine>thanks C++ for that nonsense
22:02:32  * rmgquit (Remote host closed the connection)
22:03:55  * Qardjoined
22:04:17  * tjkrusinskiquit (Ping timeout: 268 seconds)
22:05:55  * eugenewarejoined
22:06:41  * rmgjoined
22:08:50  <Domenic_>If people know about UDP they can help guide the web UDP streams API here https://github.com/whatwg/streams/issues/64#issuecomment-38602115
22:09:36  <Domenic_>the spec author seems to think it should be stream.write({ port, address, data }) and var { port, address, data } = stream.read(), but Node's API seems to set port/address ahead of time and just use .write(data)/ data = read()
22:09:46  <Domenic_>so I pointed him at Node but don't really know enough to say what's better
22:11:02  <tjfontaine>well, there are multiple things
22:11:32  <tjfontaine>you can bind a port to a specific interface and address, but in general the tuple you get in on message is indeed port,address,data
22:11:48  <tjfontaine>outbound you always specify to whom and where you're sending this packet
22:11:57  <tjfontaine>the binding is just about how you appear to the remote end
22:13:20  * yunongquit (Remote host closed the connection)
22:16:00  * yunongjoined
22:19:54  <indutny>tjfontaine: I'd like to revisit --v8-flags tomorrow
22:19:56  <indutny>I think you are right
22:20:00  <indutny>there is no need in computation
22:20:20  <tjfontaine>indutny: ok, thanks
22:21:39  <indutny>ttyl
22:22:47  * bradleymeckquit (Ping timeout: 268 seconds)
22:24:57  <trevnorris>indutny: real quick
22:25:05  <trevnorris>indutny: you mean something more like this: https://github.com/trevnorris/node/commit/b64a3b4
22:25:29  * m76quit (Read error: Connection reset by peer)
22:25:41  <Domenic_>tjfontaine: hmm OK, that's quite interesting; thanks.
22:25:45  <trevnorris>tjfontaine: your thoughts on using that type of whatnot ^
22:27:17  <trevnorris>tjfontaine: or would you rather it be more explicit and call the methods before and after, instead of depending on the scope?
22:30:23  <trevnorris>also, what was that tweet about weekmaps being on by default on v0.12?
22:40:51  * eugenewarequit (Remote host closed the connection)
22:41:43  * paulfryzeljoined
22:46:26  * paulfryzelquit (Ping timeout: 252 seconds)
22:47:37  * AvianFluquit (Remote host closed the connection)
22:48:19  * thlorenzquit (Remote host closed the connection)
22:48:51  * thlorenzjoined
22:53:45  * thlorenzquit (Ping timeout: 265 seconds)
22:54:08  <trevnorris>um.. weakmaps
22:55:42  <othiym23>trevnorris: weak collections and ES6 promises are behind --es-staging and are going to be turned on by default in 3.25 / 3.26, according to Andreas Rossberg on v8-users
22:55:56  <tjfontaine>trevnorris: my perference are not to rely on scope to manage this
22:57:34  <trevnorris>othiym23: sure, but we're not going to 3.25 in v0.12 afaik
22:57:49  <othiym23>trevnorris: I know
22:57:52  <trevnorris>tjfontaine: sounds good. it does feel too implicit for my liking
22:57:56  <othiym23>trevnorris: but Domenic_ didn't, y esterday
22:58:02  <trevnorris>ah, ok
22:58:24  <Domenic_>:(
22:58:33  <trevnorris>I wish we were going to upgrade to 3.25. maybe if I can stall the release another 2 weeks we can :-P
22:59:20  <tjfontaine>there will be no more stalling :)
22:59:24  <Domenic_>what's new in 3.25 that you like trevnorris? Just the IRHydra fix?
22:59:24  <trevnorris>tjfontaine: in your personal opinion, is there ever a case to use that technique?
22:59:39  <tjfontaine>arm64
22:59:57  <mmalecki>V8 IS GETTING ARM64?!
22:59:57  <LOUDBOT>ALWAYS COKE, NEVER PEPSI. SAUCE ON THE SIDE.
23:00:11  <trevnorris>Domenic_: well, IRHydra is just taking advantage of new features. but the ability to track jit changes to every IC is freaking amazing
23:00:12  <mmalecki>sorry, I do get overly excited about ARMs.
23:00:13  <tjfontaine>trevnorris: I'm not entirely sure -- it's a common pattern in c++, but there are times when people don't really understand the implications and then it's impossible to split them up
23:00:35  <trevnorris>tjfontaine: yeah. and i'm always for explicit > implicit
23:00:41  <mmalecki>(that was to be expected, since chromeboooks. still, awesome.)
23:02:35  <trevnorris>tjfontaine: oh, apparently i'm doing a nodeup show tomorrow morning. so guess I won't be on the call.
23:02:43  <trevnorris>or, hey, maybe I can do both at the same time.
23:02:51  <tjfontaine>that would be awkward
23:02:53  <trevnorris>honestly have no idea what i'm supposed to talk about
23:03:04  <tjfontaine>rarely is it clear :)
23:03:06  <othiym23>AL, duh
23:03:07  <mmalecki>trevnorris: make shit up as you go, that's my experience
23:03:10  <trevnorris>hehe
23:03:14  <othiym23>since the show is about CLS and all
23:03:26  <tjfontaine>looks like it's CLS
23:03:44  <mmalecki>trevnorris: not sure if it's still about kangaroos, if it is, crack a few jokes about them :-D
23:03:50  <tjfontaine>just so long as everyone is clear on the disposition for v0.12
23:05:08  <trevnorris>ok. that's the last code change i wanted before finishing the squash. now, time to start making sense of all these commits.
23:05:11  * eugenewarejoined
23:06:00  <trevnorris>tjfontaine: so, the native code changes will land, correct? just so we can do tracing externally?
23:06:51  <tjfontaine>they will be in a process.binding('async_listener') thing, so yes all C++ changes will be there
23:07:24  <trevnorris>ok
23:07:50  <trevnorris>tjfontaine: we still can't use - in names can we?
23:07:56  <trevnorris>e.g. binding('async-wrap')
23:09:22  * mikealquit (Quit: Leaving.)
23:10:00  <tjfontaine>well can't is strong, we can -- we just don't because of how the macros work
23:10:12  <tjfontaine>if you wanted to you could just define the module without using the macro
23:11:57  * mikealjoined
23:12:01  <trevnorris>eh, whatevs
23:12:14  <mmalecki>oh trevnorris, you just gave me the best idea ever. JS disallows dashes (-) in identifiers but allows random UTF8 charactes. thus, EN DASH. http://www.fileformat.info/info/unicode/char/2013/index.htm My coworkers are going to hate me.
23:12:49  <trevnorris>mmalecki: hahaha. that is awesome.
23:13:40  <tjfontaine>what are we, go?
23:14:41  <othiym23>you could go for em dashes too -- I think that will actually look better in most monospaced fonts
23:14:50  <othiym23>–-—
23:15:01  <othiym23>I guess in consolas the hyphen is narrowest
23:15:12  <othiym23>for some reason in a bunch of monospaced fonts the en dash is narrower than the hyphen
23:16:03  * eugenewarequit (Remote host closed the connection)
23:16:11  * thlorenzjoined
23:21:47  <trevnorris>tjfontaine: ok, this squash is going to be pretty terse since a bunch is going to be ripped out anyways.
23:23:59  * jmar777joined
23:24:08  * thlorenzquit (Ping timeout: 240 seconds)
23:24:17  * mikolalysenkojoined
23:25:08  <tjfontaine>that's fine, we just need to preserve the stuff that's going to be going into the other repo
23:25:14  <tjfontaine>which you do have commit access to
23:29:19  * calvinfoquit (Quit: Leaving.)
23:29:39  * brunklequit (Quit: brunkle)
23:29:53  * mikealquit (Quit: Leaving.)
23:30:31  * brunklejoined
23:31:17  * wolfeidauquit (Ping timeout: 240 seconds)
23:38:58  * wolfeidaujoined
23:39:03  * hzquit (Ping timeout: 260 seconds)
23:40:32  * brunklequit (Quit: brunkle)
23:42:30  * paulfryzeljoined
23:46:51  * paulfryzelquit (Ping timeout: 245 seconds)
23:48:30  * wolfeidauquit (Ping timeout: 252 seconds)
23:51:17  * mikealjoined
23:53:51  * eugenewarejoined
23:57:26  <tjfontaine>saghul, indutny: http://jenkins.nodejs.org/job/libuv-master-gyp/DESTCPU=ia32,label=linux/563/tapTestReport/test.tap-156/
23:58:53  * hzjoined