00:04:43  * loladiro_joined
00:06:18  * loladiroquit (Ping timeout: 264 seconds)
00:06:19  * loladiro_changed nick to loladiro
00:36:55  <CIA-108>node: Ben Noordhuis v0.6 * reef57dd / (50 files): core: use proper #include directives - http://git.io/-uLb6A
00:37:23  <bnoordhuis>isaacs: ^
00:37:45  * piscisaureus_joined
00:37:45  * piscisaureus_quit (Remote host closed the connection)
00:41:12  * indexzerojoined
00:44:34  * loladiro_joined
00:44:49  * loladiroquit (Ping timeout: 245 seconds)
00:44:50  * loladiro_changed nick to loladiro
00:47:19  * indexzeroquit (Quit: indexzero)
00:54:44  <isaacs>bnoordhuis: thanks :)
00:55:00  <isaacs>bnoordhuis: btw, we're looking really good for 0.8 being done soon.
00:55:33  <isaacs>there are free hands to help with the ev_ fixes, awaiting your direction.
01:06:54  * loladiroquit (Ping timeout: 240 seconds)
01:20:46  <AvianFlu>isaacs, btw, I can now say with complete certainty that "just disabling the assertion" for that fs-watch failure on smartos is not a good idea :D
01:21:06  <isaacs>AvianFlu: orly? what happens?
01:21:14  <isaacs>i mean, i figured.
01:21:17  <isaacs>but now i'm curious
01:21:22  <AvianFlu>well, there are more problems than just that
01:21:32  <AvianFlu>like some events seem to double-fire
01:21:43  <AvianFlu>I started looking at the libuv tests, since they fail similarly
01:21:49  <AvianFlu>as I posted on the thing
01:22:09  <AvianFlu>a couple of them pass if you remove the assertion, but most still fail
01:35:37  * \toothrotjoined
01:36:54  * toothrquit (Ping timeout: 240 seconds)
01:42:00  * elijah-mbpjoined
01:43:17  * abraxasjoined
02:32:57  <CIA-108>node: Ben Noordhuis master * rff552dd / (lib/tls.js test/pummel/test-tls-ci-reneg-attack.js): tls: fix off-by-one error in renegotiation check - http://git.io/RROiRA
02:36:25  * brsonjoined
02:50:46  <AvianFlu>hahahaha
02:51:20  <AvianFlu>so, the file watching test on libuv that happens to make the most files in the watch_dir is timing out...
02:51:41  <AvianFlu>leaving a bunch of extra files that the other tests don't clean up, which is making 3 or 4 tests fail just from the rmdir calls at the end of the tests XD
02:52:02  <AvianFlu>ah, comedy
03:11:24  * sh1mmerjoined
04:22:49  * \toothrotchanged nick to toothr
05:22:11  * paddybyersjoined
05:31:16  * bnoordhuisquit (Ping timeout: 248 seconds)
05:50:08  * brsonquit (Ping timeout: 276 seconds)
06:11:07  * \toothrotjoined
06:13:10  * toothrquit (Ping timeout: 272 seconds)
06:51:28  * stephankquit (Quit: *Poof!*)
07:02:12  * hzjoined
07:04:22  * sh1mmerquit (Quit: sh1mmer)
07:04:59  * rendarjoined
07:19:47  * mmaleckijoined
07:40:17  * rendarquit
07:42:03  * rendarjoined
07:52:25  <hz>compiliamo per msvc10
07:52:27  <hz>vediamo...
07:54:15  <hz>e boost funzia
07:54:21  <hz>il mio project...
07:56:25  <hz>sbloccato il supporto a msvc10
07:56:39  <hz>vediamo se compila... rullo di tambri
07:57:10  <indutny>hz: sei sicuro che sei in una stanza corretta?
07:57:22  <indutny>google translate ^
07:57:35  <hz>nice translation
07:57:41  <hz>btw wrong channel :D
07:57:50  <hz>really sorry :P
07:57:53  <indutny>that's what I was trying to ask :)
07:58:18  <hz>ye
07:58:28  <hz>damn touchpad
08:10:32  * mmaleckiquit (Ping timeout: 276 seconds)
08:21:17  * loladirojoined
08:34:52  * mmaleckijoined
09:32:15  * loladiroquit (Quit: loladiro)
09:32:19  * mmaleckiquit (Ping timeout: 245 seconds)
09:47:36  * loladirojoined
09:49:53  * mmaleckijoined
09:56:37  * toothrjoined
09:58:24  * \toothrotquit (Ping timeout: 240 seconds)
10:01:34  * loladiroquit (Quit: loladiro)
10:03:47  * benviequit
10:36:08  * mmaleckiquit (Quit: Lost terminal)
10:36:30  * mmaleckijoined
11:22:35  * `3rdEdenjoined
11:23:07  * \toothrotjoined
11:24:24  * toothrquit (Ping timeout: 245 seconds)
12:09:20  * piscisaureus_joined
12:10:46  * abraxasquit (Remote host closed the connection)
12:15:31  <einaros>piscisaureus_: the patch mraleph and you discussed last night fixes it, as you commented
12:15:47  <piscisaureus_>einaros: yeah, I checked
12:16:05  <piscisaureus_>einaros: when this lands in bleeding edge I will land it in node as well
12:17:00  <einaros>running a few other perf checks on bleeding edge + that patch now
12:17:03  <einaros>all looks well so far
12:18:16  <piscisaureus_>cool
12:20:28  <einaros>the deopts and bailouts from that log makes me curious to know if there are any particular code constructs one should avoid for 3.11+, though
12:21:02  * rendar_joined
12:21:03  * rendar_quit (Excess Flood)
12:21:25  * rendar_joined
12:21:26  * rendar_quit (Excess Flood)
12:22:10  * rendar_joined
12:22:11  * rendar_quit (Excess Flood)
12:22:29  * rendar_joined
12:23:16  <hz>rendar?
12:23:18  <hz>:D
12:23:36  * rendarquit (Read error: Connection reset by peer)
12:25:44  <einaros>piscisaureus_: is the --code-comments flag that he mentioned something that should work?
12:26:12  <piscisaureus_>einaros: I never tried. But you could try...
12:26:19  <piscisaureus_>einaros: wasn't it --print-code-comments ?
12:26:53  <piscisaureus_>hmm, no, --code-comments
12:27:12  <einaros>piscisaureus_: ah, no, it was code-comments, but he apparently uses an irc client which replaces two dashes with that extra long dash
12:27:15  <einaros>so copy-paste broke it
12:27:21  <piscisaureus_>hehe
12:28:01  <piscisaureus_>hmm
12:28:04  <einaros>http://i.imgur.com/CbUnv.png
12:28:09  <piscisaureus_>race condition encountered in test-cluster-message
12:28:09  <einaros>sneaky
12:32:44  * kohaiquit (Remote host closed the connection)
12:33:35  <einaros>500 runs, no race
12:43:55  * bnoordhuisjoined
12:50:19  <piscisaureus_>aah
12:50:28  <piscisaureus_>insanity
12:52:08  * einarosstops running that test now
12:52:38  <piscisaureus_>einaros: the test sends a server handle over the ipc channel
12:52:42  <piscisaureus_>and then connects to it
12:52:55  <piscisaureus_>einaros: but what's happening for me... the master process is getting the connection
12:55:15  <einaros>piscisaureus_: the master is getting the net.connect(common.PORT connection attempt?
12:57:18  * mmaleckiquit (Ping timeout: 250 seconds)
13:04:25  * TheJHjoined
13:08:23  * piscisaureus_changed nick to piscisaureus
13:09:14  <piscisaureus>nope
13:10:19  <piscisaureus>well, hmm
13:10:20  <piscisaureus>it is
13:10:28  <piscisaureus>hmmmmmmmmm very weird
13:11:11  <einaros>seeing as the master can't possibly be executing server.listen, that sounds odd
13:13:40  <piscisaureus>worker.send('message from master');
13:13:48  <piscisaureus>^-- this is not always received by the worker
13:14:56  <piscisaureus>yep, race condition :-)
13:15:15  <piscisaureus>sometimes the message would arrive before the worker knew it had received a message
13:15:28  <piscisaureus>ER: *sometimes the message would arrive before the worker knew it had accepted a connection
13:15:46  <piscisaureus>not a node bug, there is not guaranteed ordering in callbacks
13:17:14  <CIA-108>node: Ben Noordhuis master * rf805139 / tools/installer.js : build: fix `make install DESTDIR=/path` - http://git.io/WvRl1A
13:22:32  <einaros>there's no reason why that process.on('message' should be within the connection block in either case
13:23:02  <einaros>ah, well it writes it back to the socket
13:23:04  <einaros>:P
13:23:15  <einaros>holy contrived example, batman
13:23:16  <piscisaureus>yep
13:23:17  <piscisaureus>fixing that
13:23:34  <piscisaureus>it relies on some idea that events will keep ordered across networks / processes
13:23:42  <piscisaureus>wrong assumption
13:27:42  <indutny>bnoordhuis: hi
13:27:44  <indutny>piscisaureus: hoy
13:27:51  <bnoordhuis>indutny: ho
13:27:54  <indutny>bnoordhuis: can I close that child_process issue?
13:28:00  <indutny>bnoordhuis: or is it reproducible for you?
13:28:05  <indutny>ah, you just answered
13:28:08  <bnoordhuis>yeah :)
13:28:18  <indutny>nice :)
13:28:20  <piscisaureus>indutny: 嘿
13:28:24  * mmaleckijoined
13:28:27  <indutny>piscisaureus: привет
13:28:29  <indutny>haha
13:28:55  <indutny>so we've only 3 issues
13:29:14  <indutny>one of them is not really an issue
13:29:35  <piscisaureus>indutny: I also still have to fix the resume-after-connect issue
13:30:03  <bnoordhuis>https://github.com/joyent/node/issues/3355 <- https://gist.github.com/f03485eb0abbb29af568 <- ev_io and ev_timer emulation
13:30:36  <bnoordhuis>should i shim ev_ref and ev_unref?
13:30:56  <indutny>bnoordhuis: they work on the event loop, ain't they?
13:31:14  <bnoordhuis>indutny: yes. libuv officially doesn't support that but i can make it work
13:31:23  <bnoordhuis>i don't really want to, though :)
13:31:26  <indutny>bnoordhuis: ah, well, more shims - better
13:31:35  <indutny>bnoordhuis: we'll remove that in 0.9 (I suppose)
13:31:45  <indutny>people should start migrating to new API
13:31:48  <bnoordhuis>yep
13:31:59  <bnoordhuis>i'm not going to maintain that shim indefinitely :)
13:32:04  <piscisaureus>We should postpone the release until http://codereview.chromium.org/10554011/ lands
13:32:14  <piscisaureus>atleast, until it lands in bleeding edge
13:32:45  <indutny>piscisaureus: you should create an issue with milestone for that
13:33:02  <indutny>so everyone would be able to track changes
13:34:16  <indutny>oh, nice
13:34:40  <indutny>MakeCallback's signature has changed
13:34:49  <indutny>is it internal API
13:34:51  <indutny>?
13:34:52  <indutny>bnoordhuis: ^
13:35:29  <piscisaureus>And we're back to 5 issue :-)
13:35:31  <indutny>that may broke some modules (I suppose)
13:35:42  <indutny>haha
13:35:46  <piscisaureus>hmm
13:35:48  <piscisaureus>that's not nice
13:35:55  <piscisaureus>bnoordhuis: can we add a shim for the old MakeCallback ?
13:36:03  <piscisaureus>it shouldn't be an internal API
13:36:28  <indutny>piscisaureus: I think we should
13:36:30  <indutny>I can do it
13:36:30  <bnoordhuis>piscisaureus: eh?
13:36:40  <piscisaureus>indutny: go for it
13:37:02  <bnoordhuis>there's more than one MakeCallback now, right?
13:37:11  <bnoordhuis>one still has the old signature
13:37:17  <piscisaureus>oh
13:37:27  <piscisaureus>yes, that's exactly what I wanted
13:37:29  <piscisaureus>indutny: ^\
13:37:57  <indutny>piscisaureus: ah, so we do have a shim?
13:38:06  <indutny>ok, just didn't know that
13:38:39  <indutny>oh, we have 3 different signatures for it
13:39:19  <piscisaureus>not enough
13:39:21  <piscisaureus>moar
13:39:25  <indutny>hehe
13:39:42  <piscisaureus>I was thinking
13:39:42  <indutny>we should add one more MakeCallback(obj, property, subproperty, argc, argv)
13:39:54  <indutny>in case if anyone will want to do this
13:39:55  <indutny>:D
13:40:08  <piscisaureus>indutny: I was thinking... what about making the debugger more useful
13:40:19  * indutnyis listening
13:40:42  <piscisaureus>indutny: when you start the debugger for a process that is hanging in epoll or whatever, you're sort of stuck
13:40:49  <piscisaureus>because you can't inspect js vars etc
13:40:54  <indutny>indeed
13:40:56  <piscisaureus>what about using some uv_async to wake up the event loop
13:41:01  <piscisaureus>and then call some empty js function
13:41:04  <piscisaureus>and break there
13:41:06  <indutny>good idea
13:41:15  <indutny>piscisaureus: ++
13:41:43  <piscisaureus>first, fixing tests :-)
13:41:51  <indutny>piscisaureus: haha
13:42:21  <indutny>about nextTick ordering issues: https://github.com/indutny/defer-tick/blob/master/test/load-test.js#L17
13:45:00  <piscisaureus>I have the impression that the gui who wrote the cluster tests doesn't really try on windows
13:45:11  <piscisaureus>I don't think even a single one of them passed out of the box
13:45:50  <indutny>piscisaureus: apparently it's hard to try anything on windows
13:45:58  <indutny>we should have a windows box for developers
13:46:04  <piscisaureus>yeah
13:46:15  <piscisaureus>maybe isaac can provision one :-)
13:46:18  <indutny>:)
13:46:30  <indutny>subtle proposal
13:46:34  <piscisaureus>I used to have a couple of from rackspace but I believe they have been rounded up
13:46:45  * c4milojoined
13:46:56  <piscisaureus>and if not, I forgot the passwords
13:47:38  * indutny5 open issues · Due in about 17 hours https://github.com/joyent/node/issues?milestone=10&state=open
13:47:41  <indutny>err
13:47:47  * indutnytopic: 5 open issues · Due in about 17 hours https://github.com/joyent/node/issues?milestone=10&state=open
13:47:47  <piscisaureus>so what's test-eio-limit supposed to test
13:47:52  <bnoordhuis>good question
13:47:56  <piscisaureus>and why should I care?
13:47:57  <indutny>piscisaureus: starvation
13:48:01  <indutny>piscisaureus: no :)
13:48:24  <indutny>there're no way to prevent starvation
13:48:39  <piscisaureus>indutny: it seems that actually, it wants starvation
13:48:45  <piscisaureus>if there is no starvation it fails
13:48:49  <indutny>piscisaureus: em...
13:49:10  <indutny>aaah, that's my test
13:49:14  <indutny>nice :)
13:49:21  <indutny>I thought you were talking about eio-race
13:49:38  <piscisaureus>indutny: ok, so what's eio-limit supposed to test?
13:50:49  <indutny>piscisaureus: well, it's sort of regression test. previously we had eio queue limit = 10
13:51:12  <indutny>and callbacks was added to queue and was never called
13:51:30  <piscisaureus>indutny: well, on windows there is no limit, and I have not seen a single bug that was caused by the absence of a limit
13:51:31  <indutny>I think this test is failing due to the bad heuristic
13:51:42  <piscisaureus>indutny: so I want to skip it on windows
13:51:49  <indutny>piscisaureus: yeah, np at all
13:52:01  * hzquit
13:52:03  <indutny>piscisaureus: why is it failing?
13:53:07  <indutny>this's an important question
13:53:08  <piscisaureus>indutny:
13:53:09  <piscisaureus>assert.js:104
13:53:09  <piscisaureus> throw new assert.AssertionError({
13:53:09  <piscisaureus> ^
13:53:09  <piscisaureus>AssertionError: false == true
13:53:09  <piscisaureus> at repeat (D:\node4\test\simple\test-eio-limit.js:30:12)
13:53:10  <piscisaureus> at D:\node4\test\simple\test-eio-limit.js:35:5
13:53:31  <indutny>well, that's a result, not the cause
13:53:57  <indutny>this means that a lot of callbacks wasn't fired
13:54:00  <indutny>100
13:55:08  <piscisaureus>indutny: well, not *yet* :-)
13:55:14  <indutny>heh
13:55:24  <indutny>you may replace > 1000 condition with > 100000
13:55:28  <indutny>and 100 with 1000
13:55:32  <indutny>and see if it'll fail
13:55:41  <piscisaureus>sure
13:55:47  <piscisaureus>***generating code***
13:55:52  <indutny>heh
13:57:13  <piscisaureus>indutny: yes, when I do that it passes
13:57:41  <piscisaureus>indutny: with 200 it passes too, btw
13:57:57  <indutny>piscisaureus: well, you may increase number then ;)
13:58:19  <indutny>before removing eio limit it was failing even on 200
13:59:44  <CIA-108>node: Bert Belder master * r1b7d23e / test/simple/test-cluster-message.js : Fix test-cluster-message so it passes on Windows - http://git.io/myTgnw
13:59:44  <CIA-108>node: Bert Belder master * r30ab1f5 / test/simple/test-eio-limit.js : test-eio-limit: bump the number of allowed background tasks to 200 - http://git.io/SAkMuw
14:02:36  * toothrjoined
14:03:30  * sh1mmerjoined
14:03:54  * \toothrotquit (Ping timeout: 240 seconds)
14:13:09  <piscisaureus>I do not understand what test-cluster-worker-disconnect is supposed to do, either
14:14:45  <indutny>I hope someone does
14:18:51  <piscisaureus>there was some npm package to start a file server quickly
14:18:54  <piscisaureus>what was the name again?
14:19:36  * \toothrotjoined
14:21:03  <indutny>piscisaureus: may be this https://github.com/balupton/simple-server ?
14:21:09  <indutny>honestly, I can't remember the name
14:21:28  * toothrquit (Ping timeout: 272 seconds)
14:21:46  <indutny>or this https://github.com/amasad/static-here
14:28:03  <piscisaureus>indutny: I refuse to use anything written in coffeescript
14:28:07  <indutny>haha
14:28:07  <piscisaureus>as a matter of principle
14:28:11  <indutny>last one isn't
14:30:07  * ericktjoined
14:30:30  <piscisaureus>indutny: I wonder if it is possible to control upnp from node
14:30:33  <piscisaureus>that would be awesome
14:30:44  <indutny>piscisaureus: I don't know much about it
14:30:45  <piscisaureus>static-here should set up port forwarding automatically
14:30:49  <indutny>piscisaureus: i.e. just know what it is
14:30:52  <indutny>aaah
14:30:56  <indutny>heh
14:31:02  <indutny>I think it's possible
14:31:10  <indutny>but its a matter of principle to me :)
14:31:56  <mmalecki>piscisaureus: http-server
14:32:10  <mmalecki>piscisaureus: `npm -g i http-server && http-server`
14:32:13  <mmalecki>THERE.
14:32:22  <piscisaureus>yeah
14:32:28  <piscisaureus>static-here doesn't work for me
14:32:31  <piscisaureus>gets me in a redirect loop
14:33:04  <piscisaureus>http has a strong nodejitsu shell to it
14:33:09  <piscisaureus>it installs a ton of dependencies
14:33:30  <piscisaureus>I suspect it also creates a nodejitsu account for me
14:33:40  <mmalecki>yeah, and takes over your computer
14:34:15  <piscisaureus>and it flatirons all my browser cookie through optimist into a hookio on Maraks cell phone
14:34:45  <piscisaureus>and then it doesn't work :-)
14:34:45  <piscisaureus>ode.js:384
14:34:45  <piscisaureus> var b = process.binding('signal_watcher');
14:34:53  <piscisaureus>wut
14:34:58  <mmalecki>how'd you know?!
14:35:08  <mmalecki>also, it doesn't depend on hookio :)
14:35:42  <piscisaureus>ok, no, that was not the module that once worked for me
14:36:20  <mmalecki>jeez, it really has lots of deps
14:36:41  <piscisaureus>it should not call process.binding('signal_watcher')
14:36:43  <piscisaureus>:-(
14:36:49  <mmalecki>wut
14:36:54  <mmalecki>it doesn't?
14:37:01  <mmalecki>it does process.on('sigint')
14:37:06  <mmalecki>oh, wait, I can fix it, actually
14:37:21  <mmalecki>just add a conditional check for windows
14:38:48  <mmalecki>piscisaureus: process.platform is 'win32' for windows, correct?
14:39:23  <piscisaureus>mmalecki: yep
14:40:59  <indutny>piscisaureus: so do you really need a server ? :)
14:41:03  <indutny>piscisaureus: you can do npm install -g bem
14:41:05  <indutny>ah
14:41:07  <indutny>no it won't work too
14:41:14  <indutny>I will write a book
14:41:24  <mmalecki>piscisaureus: https://github.com/nodeapps/http-server
14:41:24  <piscisaureus>indutny: not any more, I stuck to good ol' apache for now
14:41:24  <indutny>"30 simple steps to install a static server with a node.js"
14:41:33  <piscisaureus>hahaaa
14:42:03  <piscisaureus>mmalecki: cool, next time :-)
14:42:21  <mmalecki>piscisaureus: sure :)
14:42:48  <mmalecki>also, got rid of one unused dep, too!
14:43:35  <piscisaureus>\o
14:43:56  <isaacs>bnoordhuis: \o/
14:43:59  <indutny>mmalecki: haha
14:44:06  * toothrjoined
14:44:10  <indutny>mmalecki: we should chop down the rest too :P
14:44:22  <indutny>mmalecki: not to say, I love flatiron!
14:44:24  * \toothrotquit (Ping timeout: 240 seconds)
14:44:32  <mmalecki>indutny: yeah, that sounds good
14:44:41  <isaacs>bnoordhuis: we should also make the fakey ev_ functions warn at compile time, if possible.
14:45:05  <indutny>isaacs: just add if (rand() == 0) assert(0 && "Fakey function fails fast")
14:45:10  <mmalecki>indutny: I can remove flatiron pretty easily
14:45:19  <indutny>mmalecki: I don't really care, actually :P
14:45:23  <mmalecki>indutny: if you can review it later :)
14:45:26  <indutny>haha
14:45:27  <indutny>mmalecki: ok
14:46:08  <isaacs>indutny, bnoordhuis: well, i mean, if we'd warned about them in 0.6, we could remove then in 0.8. we need to warn in 0.8, so that we can remove them in 0.10
14:46:33  <indutny>isaacs: yes, that's understood. I was just kidding\
14:46:38  <piscisaureus>I still think the cluster is sort of lugubrish
14:46:40  <isaacs>ehe
14:46:44  <isaacs>sorry, still pre-coffee :)
14:46:49  <piscisaureus>all these deaths and suicides
14:47:09  <bnoordhuis>isaacs: sure, i'll add that
14:47:25  <piscisaureus>*lugubrious
14:47:26  <isaacs>bnoordhuis: as soon as we get that bit in, i wanna cut 0.7.12
14:47:32  <isaacs>and then get 0.8 wednesday maybe
14:47:47  <isaacs>rapid fire :)
14:47:55  <indutny>isaacs: hey
14:48:02  <indutny>isaacs: may be add stdin to forks then?
14:48:29  <isaacs>indutny: what are you working on at the moment? if you're free and wanna do it, sure. i can probably take a look at it, otherwise.
14:48:48  <isaacs>it's a pretty easy win, probably. i think it just wasn't quite trivial when you tried before, right?
14:48:57  <einaros>I hope the v8 fixes land before 0.8
14:49:20  <isaacs>bnoordhuis: didn't you find something also where fork({silent:true}) doesn't work as expecteD?
14:49:44  <indutny>isaacs: yes, it worked fine
14:49:58  <indutny>isaacs: I'm working on the SmartOS stuff
14:50:13  <isaacs>indutny: sweet
14:50:14  <indutny>isaacs: so if you don't have anything more important, I would be glad if you'll help me
14:50:15  <piscisaureus>isaacs: are we also getting a smartos drone?
14:50:22  <indutny>piscisaureus: we have ;)
14:50:25  <piscisaureus>er, a *windows* drone
14:50:32  <piscisaureus>isaacs: sorry, weird twist in my head
14:50:33  <isaacs>piscisaureus: yes, i'd emailed you the admin password
14:50:54  <piscisaureus>isaacs: /me got nothing
14:50:54  <mmalecki>indutny: review? https://github.com/nodeapps/http-server/pull/22
14:52:00  <indutny>mmalecki: I've no access to nodeapps team
14:52:19  <mmalecki>indutny: lame. put a 'lgtm' there, then :)
14:52:24  <indutny>mmalecki: already put
14:52:30  <mmalecki>kthx
14:53:24  <mmalecki>piscisaureus: there you go :)
14:53:41  <mmalecki>I like how it took 8 minutes
14:56:41  <indutny>piscisaureus: at least maintainer is very friendly ;)
14:56:51  * xaqjoined
14:57:06  <indutny>few days ago I've joined an #apache channel
14:57:22  <indutny>asked if they could give me some advices about OpenSSL performance
14:57:35  <indutny>basically, they answered me with "GTFO!"
14:57:42  <indutny>err/
14:57:48  <indutny>s/#apache/#httpd
14:58:25  <indutny>but one man sent me a private message, asking what I've interest in
14:58:31  <indutny>this was very friendly too
14:58:40  <indutny>moral: ask directly or GTFO
14:59:05  <bnoordhuis>isaacs: https://gist.github.com/f03485eb0abbb29af568
14:59:16  <bnoordhuis>isaacs: re silent:true, it worked again after a rebuild
14:59:28  <indutny>bnoordhuis: gyp bug, I suppose
14:59:51  <bnoordhuis>yeah, maybe. not sure
15:01:06  <piscisaureus>mmalecki: good job fixing http-server. As I said, next time :-)
15:02:06  <isaacs>bnoordhuis: sweet
15:02:13  * AndreasMadsenjoined
15:02:59  <mmalecki>piscisaureus: sure :)
15:03:14  <isaacs>bnoordhuis: an fputs will print at run-time, though, right?
15:03:23  <isaacs>bnoordhuis: i mean, i guess that's fine.
15:03:36  <bnoordhuis>isaacs: yes, at run-time
15:04:19  <indutny>isaacs: but only once
15:04:48  <bnoordhuis>well, once per file that does something ev-y
15:04:48  <isaacs>yeah, i guess that's fine
15:05:11  <isaacs>it should go to stderr, though, not stdout
15:08:01  <CIA-108>libuv: Ben Noordhuis master * rc1d8e01 / src/unix/linux/syscalls.c : linux: fix typo in syscall name - http://git.io/vGv39g
15:08:07  <bnoordhuis>isaacs: it does
15:08:22  <isaacs>oh, right.
15:08:31  * isaacsforgot that fputs is backwards
15:09:29  <isaacs>bnoordhuis: i guess all that's left is to install ev_emul.h instead of ev.h. then what, we use ev_real.h or something?
15:09:34  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
15:09:35  <einaros>piscisaureus: the patch was merged with bleeding edge
15:09:52  <piscisaureus>einaros: cool. I will backport soon
15:10:12  * travis-cijoined
15:10:12  <travis-ci>[travis-ci] joyent/libuv#423 (master - c1d8e01 : Ben Noordhuis): The build passed.
15:10:12  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/6d67cf1952d5...c1d8e011f4c1
15:10:12  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1647531
15:10:12  * travis-cipart
15:10:38  <bnoordhuis>isaacs: no, ev.h and ev-emul.h should both be installed.
15:10:42  <isaacs>indutny: i'm sure that there's a lot of HELP HOW COM WORDPRESS REWRITE BROKEN I DID REWRIETRULEZ AND IT BREADKS INDEXPHP WTF?! HELP PLEASE THANGS
15:10:47  <isaacs>indutny: in #httpd
15:11:17  <bnoordhuis>isaacs: node.h should include it if !defined(WANT_NODE_INTERNALS)
15:11:39  <bnoordhuis>somewhere at the bottom of node.h
15:11:46  <isaacs>i see.
15:12:06  <isaacs>so you have to ./configure --want-node-internals which will be a bit scary opt-in.
15:13:29  <bnoordhuis>eh?
15:13:40  <bnoordhuis>oh, no. WANT_NODE_INTERNALS is set when compiling node itself
15:13:43  <isaacs>i mena, in order to get the "real" ev.h
15:13:44  <isaacs>right
15:13:55  <isaacs>./configure in node, i mean
15:14:07  <bnoordhuis>yeah, but node's configure already does that
15:14:22  <indutny>isaacs: heh
15:14:25  <bnoordhuis>wait, i'll write the patch
15:14:32  <isaacs>yes, that would be helpful :)
15:14:51  <indutny>isaacs: HELP HOW COM OPENSSL HANDSHAKE PERFORMANCE FAILING MULTyTHREAD
15:15:06  <isaacs>indutny: did you try asking like that? it might've elicited more compassion.
15:15:15  <indutny>isaacs: yes, I just asked that
15:15:34  <isaacs>lol
15:15:57  * mmaleckiquit (Ping timeout: 250 seconds)
15:16:06  <indutny>isaacs: I think they're going to ban me
15:16:55  <indutny>they're much more gentle today
15:16:56  <indutny>"thumbs> indutny: did you have a question related to apache httpd?"
15:20:04  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/ev-emul
15:20:23  <bnoordhuis>indutny: mod_ssl has its own mailing list
15:20:48  <isaacs>that ^ is one of the reasons why i've always pushed for npm questions to be asked in #Node.js
15:21:14  <isaacs>express and socket.io have their own lists, but people ask in #node.js anyway, and it just makes it more tricky for new users.
15:21:20  <indutny>bnoordhuis: oh, nice
15:21:37  <indutny>bnoordhuis: btw, can I ask you for an C++ advice?
15:22:19  <isaacs>lol: /* Evil. Using this will void your warranty. */
15:23:24  * AvianFluquit (Ping timeout: 240 seconds)
15:23:50  * xaqquit (Remote host closed the connection)
15:23:54  <isaacs>bnoordhuis: testing now
15:24:43  <isaacs>piscisaureus: can you join #joyent?
15:26:11  <bnoordhuis>indutny: sure
15:26:45  <isaacs>bnoordhuis: tests pass on mac, no surprise there.
15:27:06  * \toothrotjoined
15:28:36  * toothrquit (Ping timeout: 248 seconds)
15:30:05  <isaacs>trying out the zookeeper thingie
15:32:14  <indutny>bnoordhuis: so I'm doing this thing in almost every class https://github.com/indutny/candor/compare/master...feature-ssa#L15R121
15:32:30  <indutny>bnoordhuis: is there any better pattern, or I should consider it as just a C++ verbosity
15:32:43  <indutny>i.e. better pattern w/o black magic
15:33:17  <isaacs>bnoordhuis: ../src/node-zk.cpp:253: error: ‘struct node::__ev_timer’ has no member named ‘data’
15:33:34  <bnoordhuis>indutny: you mean the linked list pointers?
15:33:54  <bnoordhuis>isaacs: oh, indeed it does not
15:33:58  <indutny>bnoordhuis: nono
15:34:01  <indutny>bnoordhuis: I mean accessors
15:34:20  <bnoordhuis>indutny: oh, like that. you can use macros for that if you want
15:34:40  <tjfontaine>you may consider that black magic though :)
15:34:48  <indutny>bnoordhuis: that's considered black magic :D
15:34:53  <bnoordhuis>yes. it's not very idiomatic c++
15:35:28  <bnoordhuis>but otherwise you'll just have to live with the verbosity
15:35:52  <indutny>ok
15:35:53  <indutny>:)
15:36:08  <isaacs>bnoordhuis: btw, this is the test i'm using: make install; npm install zookeeper
15:36:35  <bnoordhuis>isaacs: yeah, i simply forgot to add the data field :)
15:36:36  <isaacs>bnoordhuis: the next step would be: npm explore zookeeper -- node example.js
15:36:48  <isaacs>and it should not exit immediately
15:37:04  <isaacs>if we get that far, i'll hand it off to others for more extensive banging on it.
15:37:36  <bnoordhuis>isaacs: fixed
15:38:53  <isaacs>looks a bit better...
15:40:09  <isaacs>/usr/local/include/node/uv-private/ev.h:567: error: too few arguments to function ‘ev_tstamp ev_now(ev_loop*)’
15:40:12  <isaacs>../src/node-zk.cpp:226: error: at this point in file
15:43:19  <isaacs>hm. seems like that's fixed in the git version, but not in the one on npm.
15:43:36  * toothrjoined
15:44:34  <indutny>isaacs: btw, npm is quite old in 0.7.x, isn't it?
15:44:56  <isaacs>indutny: i just updated it yesterday, and i'll have nother update today :)
15:45:02  <indutny>isaacs: ok
15:45:07  <isaacs>it should be 1.1.27 or 1.1.26
15:45:10  <indutny>isaacs: because we should test it with 0.7.x before release
15:45:24  * \toothrotquit (Ping timeout: 240 seconds)
15:45:33  * loladirojoined
15:45:36  <isaacs>indutny: yes, i've been testing it repeatedly :)
15:45:53  <isaacs>indutny: there's a bug i'm runing into now, that i want to fix first.
15:46:24  <indutny>ok :) I'm just worried about such things
15:46:30  <indutny>bbiab
15:46:36  <isaacs>bnoordhuis: so, it builds, and i get the ev_io error. but it still exits right away.
15:47:24  <isaacs>bnoordhuis: test: npm install git://github.com/yfinkelstein/node-zookeeper --node-version 0.6.19 && npm explore zookeeper -- node example.js
15:47:59  <isaacs>of course, you also have to have zookeeper installed, so the test is pretty painful to set up the first time
15:50:44  * mark__joined
15:52:04  <isaacs>hmm... it seems like ti segfaults now, actually
15:52:07  <isaacs>progress :)
15:52:32  * AndreasMadsenquit (Remote host closed the connection)
15:54:44  * xaqjoined
15:58:58  <isaacs>bnoordhuis: https://gist.github.com/2949089
16:01:42  * stephankjoined
16:05:15  <bnoordhuis>isaacs: sorry, had a phone call. back now
16:05:32  <bnoordhuis>trying zookeeper now
16:05:33  * pieternjoined
16:06:43  <isaacs>bnoordhuis: #0 uv__io_stop (loop=0x0 <-- that looks suspicious
16:07:42  * ericktquit (Quit: erickt)
16:07:48  * mark__changed nick to mcavage
16:09:38  <bnoordhuis>isaacs: yeah. it's not picking up the shims
16:10:56  * dapjoined
16:14:32  * toothrquit (Read error: Connection reset by peer)
16:14:42  <isaacs>bnoordhuis: looks like it is...? #2 0x0000000102201b75 in __ev_io_start (w=0x100d18400) at ev-emul.h:128
16:14:50  <isaacs>that's coming from ev-emul
16:17:06  * toothrjoined
16:17:30  <bnoordhuis>node-zookeeper is pretty annoying to build...
16:17:36  <isaacs>bnoordhuis: but when it gets to uv__io_stop(uv_loop_t* loop, uv__io_t* handle), the loop==NULL
16:17:42  <isaacs>bnoordhuis: tell me about it
16:17:53  <isaacs>waf /o\
16:18:27  <isaacs>you don't have to actually fetch it fresh each time, thgouh. you can just run `npm install` in the root of it
16:19:17  <bnoordhuis>isaacs: yeah, i cloned the repo
16:19:29  <bnoordhuis>it still insists on rerunning the configure script each time though
16:20:31  <bnoordhuis>isaacs: the problem is this: node-zookeeper never inits the watcher, it calls ev_io_set, then ev_io_start
16:20:41  <bnoordhuis>technically wrong but i guess we'll have to support it for now
16:20:49  <isaacs>did ev support this?
16:21:14  <bnoordhuis>in a roundabout way
16:21:31  <isaacs>also, btw, i consider it technically wrong to even be doing this, but what are you gonna do? i'd prefer to just not support ev *at all* in some future version.
16:21:48  <tjfontaine>and by future I mean .9
16:21:54  <isaacs>yes.
16:22:03  <isaacs>like, stop exporting it, on master, next week.
16:22:43  <tjfontaine>what about leaving the shim as is, and having zookeper init the watcher :)
16:25:56  <bnoordhuis>well... i can fix so i guess i will
16:26:11  <isaacs>bnoordhuis: ++ :)
16:27:47  * Guest40501joined
16:32:03  <bnoordhuis>isaacs: but i'm afraid it'll have to wait for an hour or two or i'll have to deal with a very irate SO
16:32:07  <bnoordhuis>back in a bit
16:34:22  * Guest40501quit (Ping timeout: 250 seconds)
16:34:41  <isaacs>bnoordhuis: cheers, have fun, tell her we appreciate your time very much :)
16:36:32  * AvianFlu_joined
16:36:39  * pieternquit (Quit: pietern)
16:37:12  * ericktjoined
16:37:36  * pieternjoined
16:39:40  * piscisaureusquit (Ping timeout: 246 seconds)
16:40:38  * piscisaureus_joined
16:41:38  <piscisaureus_>damn, these httpd guys are aggressive
16:41:50  <piscisaureus_>I typed notes in that box accidentally
16:42:35  <piscisaureus_>I got banned within the second
16:42:49  * TooTallNatejoined
16:43:27  <isaacs>wow
16:48:10  * TooTallNatequit (Read error: Connection timed out)
16:48:17  * paddybyersquit (Quit: paddybyers)
16:48:54  * AvianFlu_quit (Ping timeout: 240 seconds)
16:48:56  <isaacs>AndreasMadsen: I'm not able to reprodce the test you describe in https://github.com/joyent/node/issues/3468
16:49:07  <isaacs>oh, he's nto here
16:52:40  * AvianFlu_joined
16:57:42  * paddybyersjoined
16:57:42  * paddybyersquit (Client Quit)
17:01:14  * TooTallNatejoined
17:01:53  <isaacs>indutny: hey
17:02:12  <isaacs>indutny: what smartos thing are you looking into? the fs.watch thing?
17:02:21  * ericktquit (Quit: erickt)
17:03:01  * ericktjoined
17:03:19  <piscisaureus_>I wish my IRC client had an ignore option
17:04:30  <isaacs>piscisaureus_: it doesn't?
17:04:57  <piscisaureus_>no
17:08:24  * txdvquit (Read error: Connection reset by peer)
17:08:36  * txdvjoined
17:21:07  * mmaleckijoined
17:22:47  <isaacs>piscisaureus_: you can't just do /ignore ... to ignore stuff? i thought that was like an irc command, not a client command
17:23:03  * isaacscould be wrong
17:23:14  <piscisaureus_>ah
17:23:16  <piscisaureus_>it works
17:23:35  <piscisaureus_>isaacs: thanks
17:23:44  <piscisaureus_>now to see if it also works for people PM-ing me
17:23:54  <tjfontaine>isaacs: it's not usually an irc command, but he's likely using trillian and that's special
17:24:01  <isaacs>oic
17:24:08  <isaacs>piscisaureus_: you getting trolled?
17:24:14  <isaacs>need some banhammers?
17:24:29  <piscisaureus_>isaacs: not trolled, just annoyed by some idiot
17:24:40  <piscisaureus_>isaacs: i have banhammers - will use them if really needed :-)
17:25:18  <isaacs>heh
17:25:20  <isaacs>ok
17:27:14  * brsonjoined
17:28:55  <isaacs>bnoordhuis, piscisaureus_: https://github.com/joyent/libuv/commit/reviewme
17:29:04  * pfox__joined
17:29:08  <CIA-108>libuv: isaacs reviewme * r72d334f / src/unix/sunos.c : Allow suppression of ifaddrs.h on SunOS (+422 more commits...) - http://git.io/soFcyQ
17:29:28  <isaacs>basically replaces that ifdef we took out a while back, but as an opt-out instead of an opt-in
17:29:47  <isaacs>then we can expose that as a ./configure flag, and S10 users can unfuck themselves.
17:29:57  <piscisaureus_>isaacs: is it not possible to just detect solaris 10, or smartos?
17:30:15  <isaacs>piscisaureus_: it's actually a little bit hairy, i guess
17:30:18  <piscisaureus_>isaacs: not that I care that much - but having to config that stuff is painful
17:30:26  <isaacs>granted, S10 people are fucked in many many ways anyway.
17:30:27  <piscisaureus_>although I could care more about solaris 10 users
17:30:44  <isaacs>and smartos is not the only illumos or sunos that supports ifaddrs
17:31:04  <isaacs>pretty much every sunos family since 2007 or something has had that.
17:31:20  * travis-cijoined
17:31:20  <travis-ci>[travis-ci] joyent/libuv#424 (reviewme - 72d334f : isaacs): The build passed.
17:31:20  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/f9fcaf5ccf60...72d334faa1d1
17:31:20  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1648545
17:31:20  * travis-cipart
17:31:24  <isaacs>but solaris 10 is older than that, as is openindiana 147, snv121, etc.
17:32:18  <isaacs>piscisaureus_: if it was a lot of first-class supported systems, i'd say go for feature detection, but it's just a few folks hanging on to outdated tech. they can add a ./configure switch.
17:40:10  * paddybyersjoined
17:53:54  * toothrquit (Ping timeout: 240 seconds)
17:54:06  * \toothrotjoined
17:56:12  <indutny>isaacs: yes
17:56:22  <isaacs>indutny: rawk
17:56:35  <indutny>isaacs: sup?
17:56:38  <isaacs>indutny: we can release 0.7.12 without that, but not 0.8
17:56:45  <isaacs>indutny: oh, just saying that you rock :)
17:56:56  <indutny>ah, I thought you was angry :P
17:57:10  <indutny>Like I did something very wrong
17:57:20  <isaacs>oh, no, you're doing the right things :)
17:57:25  <isaacs>very right, carry on :)
17:57:30  <indutny>haha :)
17:57:44  <indutny>brb
17:58:24  <indutny>isaacs: please don't forget to push fork's stdin
17:58:35  <indutny>it should definitely get into 0.7.12, or it'll be only in 0.9 :(
18:03:17  * hzjoined
18:03:52  <TooTallNate>bnoordhuis: isaacs: you guys think we should shim eio_custom() like we're doing for the ev stuff?
18:04:08  <TooTallNate>and we should add deprecation warnings to these shims if we aren't already
18:04:17  <bnoordhuis>TooTallNate: libeio isn't going away anytime soon
18:04:25  <bnoordhuis>libev however is
18:04:45  <bnoordhuis>still, getting people to migrate to uv_work would be nice
18:04:51  <TooTallNate>bnoordhuis: i'm more talking about this: https://github.com/joyent/node/issues/3445
18:05:01  <TooTallNate>but yes, i totally agree about them migrating to uv_work
18:05:31  <TooTallNate>idk how we feel about a handful of modules breaking… but then again that just pushes them to upgrade
18:05:43  <TooTallNate>i think the problem though was that we didn't add a deprecation warning sooner
18:05:45  <TooTallNate>like at v0.6
18:05:57  <bnoordhuis>ah, the channel bit
18:06:21  <bnoordhuis>oh, i'm required elsewhere again
18:06:23  <bnoordhuis>biab
18:13:35  <isaacs>bnoordhuis, TooTallNate: Yes, we should shim for the eio_custom signature change
18:14:12  <isaacs>and yes, agreed, the takeaway here is that we must deprecate loudly, so that it prints to the terminal when you use it.
18:14:28  <isaacs>we do this well for JS, but just underestimated the effect it would have in C-land
18:18:57  * mmaleckiquit (Ping timeout: 246 seconds)
18:21:54  * mmaleckijoined
18:22:24  * AvianFlu_changed nick to AvianFlu
18:24:34  * brsonquit (Ping timeout: 256 seconds)
18:29:15  * brsonjoined
18:31:58  * mmalecki_joined
18:34:47  * mmaleckiquit (Ping timeout: 256 seconds)
18:35:05  * mmalecki_changed nick to mmalecki
18:35:06  * sh1mmerquit (Quit: sh1mmer)
18:35:54  * mcavagequit (Ping timeout: 244 seconds)
18:36:16  * sh1mmerjoined
18:38:13  * mikealjoined
18:44:56  * sh1mmerquit (Ping timeout: 276 seconds)
18:46:20  * sh1mmerjoined
18:50:27  * mikealquit (Read error: Connection reset by peer)
18:50:32  * mikeal1joined
18:52:06  * toothrjoined
18:54:06  * toothr-joined
18:54:26  * \toothrotquit (Ping timeout: 272 seconds)
18:56:28  <indutny>isaacs: does smartos works in virtualbox?
18:56:54  * toothrquit (Ping timeout: 240 seconds)
18:57:14  <indutny>isaacs: internet is really borked in my place today
18:57:25  <mmalecki>indutny: it does, at least, did, when I tried last time
18:57:29  <indutny>mmalecki: ok
18:57:36  <indutny>mmalecki: last time I tried that - it didn't
18:57:39  <indutny>therefore I'm asking
18:58:36  * toothrjoined
18:59:30  * toothr-quit (Ping timeout: 272 seconds)
19:00:33  * mcavagejoined
19:02:15  <indutny>92 bytes from 192.168.10.1: Destination Net Unreachable
19:02:15  <indutny>Vr HL TOS Len ID Flg off TTL Pro cks Src Dst
19:02:15  <indutny> 4 5 00 5400 260f 0 0000 40 01 79df 192.168.10.3 8.8.8.8
19:02:15  <indutny>Request timeout for icmp_seq 705
19:02:34  <indutny>FAIL
19:03:45  <AvianFlu>lol :(
19:12:13  * paddybyersquit (Quit: paddybyers)
19:12:44  * mikeal1quit (Quit: Leaving.)
19:16:24  <isaacs>indutny: yeah, it works in virtualbox
19:16:34  <isaacs>er, actually, i dunno
19:16:39  <isaacs>i *know* it works in vmware
19:16:42  <isaacs>but i dont' use vbox
19:16:49  <isaacs>smartos: [07:28|% 100|+ 466|- 5]: Done
19:17:13  <isaacs>pummel/test-net-connect-econnrefused times out, the rest are all fs.watch
19:17:15  <isaacs>not bad!
19:20:25  * mikealjoined
19:21:15  * AndreasMadsenjoined
19:21:49  * mikealquit (Client Quit)
19:22:26  <AndreasMadsen>quick Q: why dose fs.read not invoke makeCallback and why is the 'Offset is out of bounds' error thrown and not passed to callback?
19:22:46  <isaacs>AndreasMadsen: that sounds like a bug
19:23:04  <isaacs>AndreasMadsen: (the makecallback i mean)
19:23:13  <isaacs>AndreasMadsen: the 'offset out of bounds' is becasue that's an argument error.
19:23:23  <isaacs>AndreasMadsen: you're passing in that number and the buffer, it's up to you to make sure it's right
19:24:19  <AndreasMadsen>isaacs: I know, but somehow I forgot this time and that made me wonder
19:24:40  <isaacs>but yeah, it should be using MakeCallback.
19:25:01  <isaacs>that should always be our way from C++ to JS, so that domains get handled properly
19:25:34  * stephankquit (Quit: *Poof!*)
19:26:03  <AndreasMadsen>isaacs: I don't see that makecallback invokes domain -> https://github.com/joyent/node/blob/master/lib/fs.js#L290
19:26:40  <AndreasMadsen>in fact domain isn't a part of fs.js
19:27:34  <AndreasMadsen>fs.write is miss the makecallback too
19:28:45  <TooTallNate>AndreasMadsen: i think he meant these MakeCallbacks https://github.com/joyent/node/blob/master/src/node.h#L259-275
19:29:42  <AndreasMadsen>TooTallNate: perhaps but the source code comment says it all -> https://github.com/joyent/node/blob/master/lib/fs.js#L287
19:30:50  * xaqquit (Read error: Connection reset by peer)
19:31:18  * xaqjoined
19:32:04  <TooTallNate>AndreasMadsen: well it seems superfluous for fs.read and fs.write since they both use a "wrapper" callback function instead
19:32:42  <AndreasMadsen>TooTallNate: ah, right
19:33:09  <isaacs>ohhh
19:33:14  <isaacs>yeah, sorry, diferent name, similar job
19:33:28  <isaacs>yes, makecallback in fs.js is only for this-binding
19:33:42  <isaacs>it's not relevant unless the underlying fs req wrap object woudl be leaked, which is not good
19:37:27  <AndreasMadsen>what is these functions "fs.sendfile, fs.sendfileSync, fs.fdatasyncSync, fs.fdatasync" they aren't documented or used in any lib/ file
19:37:43  <AndreasMadsen>s/is/are
19:40:07  <tjfontaine>probably because they're not generally portable functions
19:43:36  * \toothrotjoined
19:45:08  * toothrquit (Ping timeout: 244 seconds)
19:48:36  * toothrjoined
19:50:24  * \toothrotquit (Ping timeout: 240 seconds)
19:52:13  <indutny>bnoordhuis: https://github.com/joyent/libuv/blob/master/src/unix/sunos.c#L215
19:52:46  <indutny>bnoordhuis: shouldn't we do uv__handle_stop(...) here?
19:53:32  <indutny>piscisaureus_: ^^
19:53:58  <indutny>isaacs: ^^
19:53:58  <indutny>:D
19:54:42  <isaacs>indutny: i do not know. does that make the tests pass?
19:55:05  <AvianFlu>not by itself. (just tried it)
19:56:05  <indutny>isaacs: I can't check
19:56:16  <AvianFlu>indutny, send me patch files and I'll test them, lol
19:56:17  <indutny>isaacs: internet is always reconnecting...reconnecting...reconnecting...
19:56:22  <isaacs>ugh.
19:56:25  <isaacs>indutny: that's sad.
19:58:39  <indutny>AvianFlu: just add uv__handle_stop(handle) after that line
19:58:52  <AvianFlu>isaacs, I take that back - adding the uv__handle_stop() stops the assertion from firing
19:58:58  <AvianFlu>but there are still 3 failures
19:59:03  <indutny>hahaha
19:59:07  <indutny>+1 me
19:59:08  <AvianFlu>(I forgot I put the assertion in an #ifndef __sun)
19:59:27  <indutny>AvianFlu: what are other failures
19:59:36  <AvianFlu>https://gist.github.com/c775a417e424f536b490
19:59:52  <AvianFlu>one gets the wrong event, one times out, one fires twice
19:59:55  <AvianFlu>all the fun ones :D
20:00:27  <indutny>what if you'll manually run last one?
20:00:28  <AvianFlu>note that I'm running the test patch I showed you before, indutny
20:00:31  <AvianFlu>just a sec
20:00:39  <indutny>can you please remind me about that patch?
20:00:42  <isaacs>AvianFlu: wrong event name i don't care about very much
20:01:01  <indutny>isaacs: well, it may be a twice call problem too, actually
20:01:05  <AvianFlu>indutny, last one fails the same by itself
20:01:12  <isaacs>true
20:01:30  <AvianFlu>https://github.com/joyent/libuv/pull/469
20:01:38  <AvianFlu>indutny, ^^
20:01:42  <indutny>ah
20:01:45  <indutny>libuv stuff
20:01:46  * mikealjoined
20:01:55  <indutny>I'll open uv__handle_stop PR
20:02:27  <indutny>isaacs: smartos doesn't work in virtualbox for me
20:02:57  <isaacs>indutny: that's a shame
20:03:34  <AvianFlu>I'm trying that uv patch in node now, it'll be a minute
20:03:53  <isaacs>indutny: but yes, uv__handle_stop PR sounds lovely, thank you for making progress on that.
20:04:10  <indutny>haha
20:04:13  <indutny>I just looked at source
20:04:13  <isaacs>indutny: in the meantime, you can use me or AvianFlu as your SSH connection to run tests :)
20:04:17  <indutny>wasn't able to run anything
20:04:18  <indutny>haha
20:04:23  <indutny>remote debugging ftw!
20:04:26  <isaacs>haah
20:04:28  <isaacs>indeed!
20:05:03  * loladiroquit (Quit: loladiro)
20:06:13  <AvianFlu>hahahahahahhahahahahahah
20:06:22  <AvianFlu>so, after moving in the updated libuv
20:06:23  <AvianFlu>[02:20|% 100|+ 430|- 0]: Done
20:06:43  <AvianFlu>my test fixes + uv__handle_stop()
20:07:46  <indutny>yeeeeeeeah
20:07:56  <indutny>SmartOS is running in virtualbox
20:08:03  <indutny>I was trying to run it in 32bit VM
20:08:13  <isaacs>woot
20:08:29  <isaacs>you guys are getting so much cake.
20:08:46  <indutny>hahaha
20:08:47  <indutny>AvianFlu: ++
20:08:51  <indutny>for testing
20:11:46  * loladirojoined
20:12:58  <indutny>isaacs: what's equivalent of sudo on smartos
20:13:18  <indutny>?
20:13:36  <TooTallNate>indutny: pfexec
20:13:44  <indutny>TooTallNate: thanks
20:13:57  <indutny>err
20:14:01  <indutny>pfexec passwd
20:14:08  <indutny>Unexpected failure
20:14:14  <indutny>Password database unchanged
20:14:34  <isaacs>indutny: if you're running on vbox, why not just do stuff as root if you need root?
20:14:42  <isaacs>indutny: the default password is "root"
20:14:46  <indutny>isaacs: root's password is too long
20:14:47  <indutny>noo
20:14:51  <indutny>isaacs: https://download.joyent.com/pub/iso/SINGLE_USER_ROOT_PASSWORD.20120531T220306Z.txt
20:15:06  <TooTallNate>indutny: also, there's those smartos joyent boxes that isaacs set up that we can use too
20:15:08  <isaacs>oh, holy wtf
20:15:10  <isaacs>that's so big
20:15:26  <isaacs>indutny: just log in as root once, and do passwd and set it to something short
20:15:30  <isaacs>srsly, it's a vm, who cares?
20:15:36  <isaacs>it's not like you're putting this image on the internet
20:15:55  <indutny>I can't
20:16:08  <indutny>database error
20:16:14  <isaacs>oh, that's lame
20:16:23  <isaacs>indutny: go raise a fuss in #smartos
20:16:29  <isaacs>indutny: or if that doesn't work: #joyent
20:16:39  <isaacs>(though mostly the smae people, and they're probably mostly at lunch right now)
20:16:53  <indutny>already did
20:16:57  <isaacs>ok
20:17:11  <indutny>so where're we now
20:17:30  <indutny>two issues are fixed, but waiting for merging some commits
20:17:35  <indutny>i.e. shim and that smartos stuff
20:17:43  <indutny>one issue is waiting for v8 team to apply fix
20:17:57  <indutny>another is still a shim
20:18:02  * loladiroquit (Ping timeout: 272 seconds)
20:18:03  <indutny>so looks like only one real issue is left
20:18:21  <indutny>isaacs: https://github.com/joyent/node/issues/3118
20:18:38  <indutny>isaacs: what it is about?
20:19:23  <isaacs>indutny: if you do connect(),then pause(), then resume() right away on a net socket, it throws an assertion failure, because the fd is invalid.
20:19:34  <isaacs>indutny: i was just going to slap a flag on it in js-land
20:20:24  <isaacs>indutny: you'd said that the performance fix was landed in bleeding-edge?
20:20:33  <isaacs>indutny: for https://github.com/joyent/node/issues/3478
20:21:36  * isaacsoff to get some food, bbiab
20:21:52  * stephankjoined
20:22:23  <saghul>hi, what is test-gethostbyname supposed to test? I mean, it doesn't do real DNS queries, it tries to connect to a TCP echo server
20:22:39  <saghul>bnoordhuis piscisaureus_ ^
20:23:54  * mikealquit (Quit: Leaving.)
20:33:43  * mmaleckiquit (Ping timeout: 252 seconds)
20:36:11  <bnoordhuis>saghul: it's complicated
20:36:43  <bnoordhuis>it's essentially a test to check if libuv is driving c-ares right
20:36:50  <saghul>bnoordhuis I see
20:37:13  <saghul>I see it behaves different in Windows than Unix at least in my VM
20:37:33  <bnoordhuis>saghul: it's using uv_poll now, right?
20:37:36  <saghul>on unix it get ECONNREFUSED instantaneously, on Windows it waits 5s
20:37:37  <saghul>yes
20:38:12  <bnoordhuis>it's supposed to be identical in behavior now
20:38:14  <bnoordhuis>piscisaureus_: ^
20:38:36  <saghul>i checked master (c-ares 1.7.x) and I get the same timeout
20:39:17  <saghul>it would be nice if someone with windows could run that test (compiled with mingw32) :-)
20:39:35  <saghul>though my Windows VM is pretty naked
20:40:33  <indutny>bnoordhuis: can you please merge my libuv PR? https://github.com/joyent/libuv/pull/470
20:41:49  <piscisaureus_>bnoordhuis: is it true that with lxc/cgroups you cannot virtualize loopback devices?
20:42:00  <bnoordhuis>indutny: that's not right
20:42:08  <piscisaureus_>TooTallNate: what about http://yfrog.com/oc9p9jp for an icon?
20:42:35  <bnoordhuis>indutny: at least.... on sunos, it's loop->fs_event_watcher that's being polled, not the handle
20:42:43  <TooTallNate>piscisaureus_: woah, haha, pretty cool. does it look good small?
20:42:44  <bnoordhuis>indutny: is the handle started somewhere?
20:43:05  <piscisaureus_>TooTallNate: for the really small versions I want to do a flat one
20:43:26  <TooTallNate>for sure
20:43:29  <TooTallNate>piscisaureus_++
20:43:29  <bnoordhuis>indutny: oh, it is. okay, lgtm
20:44:15  <tjfontaine>piscisaureus_: the loopback stuff was true at one point in the openvz stuff, I haven't checked the status for the current container stuff
20:45:25  <CIA-108>libuv: Fedor Indutny master * r61a6e06 / src/unix/sunos.c : sunos: fix assertion errors in uv__finish_close - http://git.io/WJ0j3g
20:47:06  <indutny>bnoordhuis: thanks
20:47:40  <indutny>bnoordhuis: now we should not forget to merge it back into node :P
20:47:54  * brsonquit (Ping timeout: 240 seconds)
20:48:15  <bnoordhuis>indutny: you can upgrade libuv in node, right?
20:49:58  <indutny>bnoordhuis: right
20:50:31  <indutny>doing it atm
20:50:31  * brsonjoined
20:50:49  * brsonquit (Client Quit)
20:51:02  * brsonjoined
20:52:08  * brsonquit (Client Quit)
20:52:26  * brsonjoined
20:52:39  <isaacs>bnoordhuis: review? https://github.com/joyent/libuv/pull/471 and https://github.com/joyent/libuv/pull/469
20:52:50  <bnoordhuis>reviewing
20:53:09  * paddybyersjoined
20:53:27  <tjfontaine>piscisaureus_: I don't think it's the case anymore, I'm pretty sure they use some remapping to make loopback work but if you bind something ont he host only to lo it won't be accessible to the container
20:53:37  * isaacsheading to a coffeeshop, will be online soon
20:53:54  <bnoordhuis>isaacs: re #471, i'd be happier if we could sniff for some compiler define but lgtm failing that
20:54:09  <bnoordhuis>isaacs: but change the commit log to "sunos: allow suppression" etc. :)
20:56:39  * brsonquit (Client Quit)
20:56:50  <indutny>=== release test-http-dns-fail === on osx
20:56:54  * brsonjoined
20:59:26  * brsonquit (Client Quit)
20:59:52  * brsonjoined
21:00:35  * brsonquit (Client Quit)
21:00:41  <piscisaureus_>saghul: I don't know, you shouldn't get ECONNREFUSED at all
21:01:01  <indutny>die moscow ISP, die!
21:01:19  * brsonjoined
21:01:24  <saghul>piscisaureus_ does the test run in a split second for you?
21:01:58  <CIA-108>node: Fedor Indutny master * r39b7ab1 / (16 files in 8 dirs): uv: upgrade to 61a6e06e - http://git.io/J4jRkA
21:02:02  <piscisaureus_>saghul: yes
21:02:17  <saghul>doh
21:02:32  * brsonquit (Read error: Connection reset by peer)
21:03:10  <saghul>ok, I'll put the code on a branch, maybe my Windows decided not to be my friend :-S
21:03:16  * hzquit
21:03:34  * benviejoined
21:03:39  * benviequit (Client Quit)
21:03:39  <indutny>oh crap
21:03:43  <indutny>force push to save me
21:03:49  * benviejoined
21:04:08  <indutny>isaacs: any objections against force push?
21:04:22  <CIA-108>node: Fedor Indutny master * r057b806 / (18 files in 8 dirs): uv: upgrade to 61a6e06e - http://git.io/qHIyEw
21:04:26  <indutny>isaacs: some files were missing
21:04:38  * dapquit (Quit: Leaving.)
21:04:47  <indutny>piscisaureus_: sadly, but that happened with me too :(
21:05:08  <piscisaureus_>indutny: what happened?
21:05:13  <piscisaureus_>indutny: gethostbyname failure?
21:05:15  <indutny>piscisaureus_: I force pushed master
21:05:23  <indutny>:(
21:06:24  <tjfontaine>I blame githubs pullreqs for making force push too popular
21:06:48  * brsonjoined
21:09:01  <CIA-108>libuv: isaacs master * r382f2a2 / src/unix/sunos.c : sunos: allow suppression of ifaddrs.h - http://git.io/rSzRxQ
21:09:34  <creationix>github pull request caused howtonode to go down for a few minutes today
21:09:45  <creationix>followed by some unholy forced pushed to get it back up
21:11:00  * travis-cijoined
21:11:00  <travis-ci>[travis-ci] joyent/libuv#426 (master - 382f2a2 : isaacs): The build passed.
21:11:00  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/61a6e06ec2e2...382f2a262bc0
21:11:00  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1650441
21:11:00  * travis-cipart
21:12:36  * \toothrotjoined
21:13:08  * toothrquit (Ping timeout: 248 seconds)
21:13:12  <saghul>piscisaureus_ do we still need to float this for skipping bogus resolvers on Windows? https://github.com/joyent/libuv/blob/master/src/ares/ares_init.c#L701
21:13:28  * indutnytopic: 4 open issues https://github.com/joyent/node/issues?milestone=10&state=open
21:15:20  <piscisaureus_>saghul: sorry, I am busy doing other stuff atm
21:15:24  * piscisaureus_part
21:15:30  * piscisaureus_part
21:16:54  <indutny>isaacs: about that .pause() .resume() issue
21:17:01  <indutny>isaacs: we've already seen similar thing
21:17:14  <indutny>isaacs: I proposed removing that assertion
21:17:21  <indutny> /cc bnoordhuis ^
21:17:42  <indutny>but I can't remind how that ended :)
21:17:49  <indutny>s/remind/remember
21:19:43  <isaacs>\o/
21:20:01  <indutny>isaacs: isaacs: you can't fix that in js land, because .setNoDelay and other stuff that sets up flags
21:20:12  <indutny>before dns lookup will be finished
21:20:17  <isaacs>indutny: when i said flag, i meant a js flag
21:20:29  <indutny>isaacs: and what do you want to do with that flag?
21:20:30  <isaacs>indutny: like, on connect, it'll actually .pause() if .paused is set
21:20:38  <indutny>aah
21:20:41  <indutny>that sort of shim
21:20:42  <isaacs>indutny: not actually call the underlying pause/resume until it's connected
21:20:43  <isaacs>yeah
21:20:44  <indutny>ok, that'll work
21:20:55  <indutny>isaacs: what about stdio stuff?
21:20:57  <isaacs>easier than a real fix, and just as good :)
21:21:01  <indutny>I'm not really on anything
21:21:10  <indutny>I think I may implement it
21:21:16  <indutny>fork's stdin
21:21:22  <isaacs>indutny: awesome, go for it
21:21:56  <isaacs>i'm gonna fix this S10 thing
21:23:32  * TooTallNatequit (Quit: Computer has gone to sleep.)
21:24:01  <indutny>cool
21:26:07  * AndreasMadsenquit (Remote host closed the connection)
21:30:55  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/ev-emul <- seems to work now
21:31:11  <isaacs>bnoordhuis: !! :D
21:31:13  <isaacs>awesome
21:32:21  <indutny>oh, interesting
21:33:18  * dapjoined
21:36:26  <indutny>isaacs: review please https://gist.github.com/fffde7f21393ba3c48f1
21:36:35  <indutny>isaacs: also there're an issue with it https://github.com/joyent/node/issues/3484
21:36:39  <indutny>though not really notable
21:36:42  <indutny>but quite interesting...
21:36:46  <indutny>bnoordhuis: https://github.com/joyent/node/issues/3484
21:37:08  <indutny>somehow stdin in child process is automatically resumed when write to stdout has occured
21:37:38  <isaacs>indutny: why is stdin a 'pipe' when it's not silent?
21:38:01  <indutny>isaacs: well stdin is always silent, anyway :P
21:38:12  <indutny>isaacs: I didn't want to break existing things
21:38:29  <indutny>ah, nvm
21:38:32  <indutny>it won't break them
21:38:40  <indutny>isaacs: do you think it should be 0? why?
21:39:15  <isaacs>indutny: hm...
21:39:24  <isaacs>indutny: maybe we should just let the user pass in a stdio array?
21:39:33  <isaacs>indutny: and then push a ipc on it?
21:41:19  <indutny>isaacs: em... that's another option
21:41:24  * TheJHquit (Ping timeout: 240 seconds)
21:41:30  <indutny>isaacs: does anyone need it?
21:41:42  <indutny>isaacs: they can use .spawn() and IPC methods will be added automatically
21:41:57  <isaacs>indutny: well, then we could leave the defaults as you have them, and if anyone barks, then they can just set it differently
21:41:59  <indutny>so the only difference is automatic push of 'ipc' to stdio options
21:42:08  <isaacs>indutny: yeah :)
21:42:17  <isaacs>indutny: so child_process.fork() is entirely implementable in userland.
21:42:23  <indutny>isaacs: indeed
21:42:25  <isaacs>i think that's somewhat winful :)
21:42:33  <indutny>isaacs: that was your idea, actually :)
21:42:44  <indutny>isaacs: so does my patch looks good to you?
21:44:31  <isaacs>indutny: i'd prefer for stdin to default to 0 if silent:true
21:44:47  <isaacs>indutny: and only do that if there's no options.stdio already.
21:45:03  <isaacs>er, if silent:false
21:46:54  <indutny>ok
21:47:19  <indutny>pushing
21:47:31  <CIA-108>node: Fedor Indutny master * r0a89e8b / (lib/child_process.js test/simple/test-child-process-fork3.js): child_process: add .stdin stream to forks - http://git.io/vmV9Qg
21:48:20  <indutny>I need a tea
21:49:37  <bnoordhuis>thoughts on https://github.com/joyent/node/pull/3483 ?
21:50:25  * rendar_quit
21:50:51  <indutny>no, I need to sleep
21:51:20  <tjfontaine>bnoordhuis: shouldn't installPrefix go through the deprecation procedure?
21:51:29  <isaacs>omg, the internet is so slow here and you guys are so fast!
21:51:40  <isaacs>tjfontaine: installPrefix has been empty for all of 0.6, i though
21:51:55  <tjfontaine>it very well maybe, as I notice the commit doesn't actually remove it
21:52:10  <bnoordhuis>tjfontaine: yes, it doesn't exist anymore
21:52:25  <bnoordhuis>$ out/Release/node -pe process.installPrefix
21:52:25  <bnoordhuis>undefined
21:52:37  <CIA-108>node: isaacs master * rd0111aa / deps/uv/src/unix/sunos.c : uv: Upgrade to 382f2a26 - http://git.io/9auE_g
21:52:39  <isaacs>bnoordhuis: land at will
21:52:40  <tjfontaine>process->Set(String::NewSymbol("installPrefix"), String::New(NODE_PREFIX))
21:52:50  <tjfontaine>apparently the ifdef wasn't there anymore :)
21:53:29  * c4miloquit (Remote host closed the connection)
21:53:57  <CIA-108>node: isaacs reviewme * r0ef6cfd / configure : Fix #3465 Add ./configure --no-ifaddrs flag - http://git.io/xuLvVA
21:54:00  <isaacs>can i get a +1?
21:55:27  <bnoordhuis>isaacs: --without-ifaddrs?
21:55:38  <isaacs>sure.
21:55:49  <bnoordhuis>oh, we have a --no-ssl2 flag, it seems
21:55:52  <bnoordhuis>how consistent of us
21:55:59  <isaacs>yeah
21:56:20  <indutny>ttyl guys
21:56:28  <bnoordhuis>sleep tight, fedor
21:56:29  <isaacs>bnoordhuis: i thought that --without is for "don't build/install xyz" whereas --no is "disable feature xyz"
21:56:34  <isaacs>g'nite indutny, thanks so much!!
21:56:54  <bnoordhuis>isaacs: yeah, you're right
21:56:57  <bnoordhuis>+1 then
21:57:00  <isaacs>kewl
21:57:18  <CIA-108>node: isaacs master * r0ef6cfd / configure : Fix #3465 Add ./configure --no-ifaddrs flag - http://git.io/xuLvVA
21:57:46  <CIA-108>node: Trent Mick master * r196207c / (5 files in 2 dirs): node: remove unused flags and files - http://git.io/wZF6Gg
21:57:51  * isaacsreviewing ev-emul
21:58:40  * mikealjoined
22:03:05  * alejandromgjoined
22:03:05  <isaacs>bnoordhuis: hm. it still exits immediately when running that example
22:05:54  * mikealquit (Quit: Leaving.)
22:07:42  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/master <- how do you feel about that v8 commit?
22:07:54  <bnoordhuis>seems kind of nice
22:08:25  <bnoordhuis>yeah, node-zookeeper is crashing for me again. it used to work before the last compile
22:08:35  <bnoordhuis>fickle code...
22:09:02  <isaacs>bnoordhuis: (re eval line nums) it is cute, but i think we can wait for it to land in a regular v8 release
22:09:18  <isaacs>bnoordhuis: we get this with vm.* anyway
22:19:04  <CIA-108>node: Trent Mick reviewme * r196207c / (5 files in 2 dirs): node: remove unused flags and files - http://git.io/wZF6Gg
22:19:04  <CIA-108>node: isaacs reviewme * rde65ba7 / (2 files in 2 dirs): net.Socket: Delay pause/resume until after connect - http://git.io/iCaLZw
22:19:11  <isaacs>bnoordhuis: ^
22:19:30  <isaacs>not sure why trent's thing is showing up there...
22:21:40  * brsonquit (Quit: leaving)
22:21:57  * brsonjoined
22:22:20  <bnoordhuis>isaacs: why is that necessary?
22:22:50  * mcavage_joined
22:22:57  <bnoordhuis>oh right, that's in the works
22:23:09  <bnoordhuis>smarter uv_read_start(), that is
22:23:20  <isaacs>sure
22:23:32  <isaacs>but in the meantime, this prevents the assertion from being blown
22:24:08  * brsonquit (Client Quit)
22:24:44  <bnoordhuis>this should be fixed in libuv, really
22:24:54  <bnoordhuis>but it's okay as a stopgap measure
22:24:54  * mcavagequit (Ping timeout: 240 seconds)
22:25:01  <isaacs>agreed
22:25:05  <isaacs>ok, i'm stopping the gap.
22:25:08  * brsonjoined
22:25:26  * brsonquit (Client Quit)
22:25:58  * brsonjoined
22:26:01  <CIA-108>node: isaacs master * rde65ba7 / (2 files in 2 dirs): net.Socket: Delay pause/resume until after connect - http://git.io/iCaLZw
22:26:51  * mcavage_quit (Ping timeout: 244 seconds)
22:27:50  * mcavagejoined
22:28:06  * brsonquit (Client Quit)
22:28:22  * brsonjoined
22:29:22  * mikealjoined
22:30:15  <isaacs>three! https://github.com/joyent/node/issues?milestone=10&sort=updated&state=open
22:30:25  * isaacstopic: 3 open issues https://github.com/joyent/node/issues?milestone=10&sort=updated&state=open
22:31:11  <isaacs>ok, i'm gonna finish fixing up this next npm release..
22:34:27  * dapquit (Quit: Leaving.)
22:35:08  * mikealquit (Read error: Connection reset by peer)
22:35:22  * mikealjoined
22:35:44  * alejandromgquit (Quit: leaving)
22:39:15  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/ev-emul
22:39:31  <bnoordhuis>i'm getting decidedly less enthusiastic about that wrapper though
22:39:42  <bnoordhuis>the bug this time was in node-zookeeper
22:40:05  <bnoordhuis>no doubt lots of other modules have similar-but-not-quite bugs
22:40:12  * mikealquit (Read error: Connection reset by peer)
22:40:21  * mikealjoined
22:40:57  <isaacs>bnoordhuis: i think if we fix node-zookeeper, which is clearly only minimally maintained to keep working, then that's a good test case.
22:42:21  * mikeal1joined
22:43:21  <isaacs>bnoordhuis: if we make a good solid try, at least the majority can start to move onto 0.8
22:43:24  * mikealquit (Read error: Connection reset by peer)
22:43:50  <isaacs>anyone who is still burned, will at least have a clear message that we'll try to accommodate them, or a warning that it's deprecated and they should use a different thing
22:45:25  <isaacs>hm, got a segfault. doing a clean rebuild
22:46:57  * dapjoined
22:47:09  * paddybyersquit (Quit: paddybyers)
22:48:31  <isaacs>bnoordhuis: whoa, weird, `make isntall` isn't putting a node in the path..
22:48:43  <isaacs>it's just doing mkdir -p /usr/local/include/node/
22:49:09  <bnoordhuis>oh?
22:49:47  <bnoordhuis>isaacs: what happens if you undo f805139?
22:50:03  <isaacs>yeah, i was just looking at that :)
22:55:30  <isaacs>hm.. still no go
22:56:04  <bnoordhuis>strange. wfm
22:56:14  <bnoordhuis>both with and without DESTDIR
22:57:31  <isaacs>bisectin
22:57:54  <isaacs>bnoordhuis: i wonder if it only fails after a distclean
22:57:57  <isaacs>which would be super painful.
22:59:27  <isaacs>whoa... apparently de65ba7aba0e48a846e8cad86278f51372d3de23 broke this??
22:59:58  <bnoordhuis>fatal: bad object de65ba7aba0e48a846e8cad86278f51372d3de23
23:00:04  <isaacs>(that's the current master)
23:00:11  <isaacs>the pause/resume stuff
23:00:22  <isaacs>i wonder if that behavior somehow confuses the installer?
23:01:05  * \toothrotchanged nick to toothr
23:01:19  * dapquit (Quit: Leaving.)
23:01:45  * dapjoined
23:02:07  <bnoordhuis>isaacs: right, now it's broken for me too. thanks a lot, isaac :)
23:02:15  <isaacs>:)
23:02:39  <isaacs>yeah, this breaks child_process.exec somehow..
23:02:56  <isaacs>see, this is why you eat the dog food :)
23:03:35  * toothrquit (Quit: here we are)
23:03:48  <isaacs>ohhhhh... i found it.
23:03:52  <isaacs>streams are hard.
23:04:35  <isaacs>also, undefined is falsey.
23:05:16  <bnoordhuis>_paused is never set?
23:05:55  <bnoordhuis>seems like it
23:06:01  <isaacs>https://gist.github.com/2951331
23:06:06  <isaacs>lgty?
23:06:11  * mikeal1quit (Read error: Connection reset by peer)
23:06:12  <isaacs>fixes it for me
23:06:22  <isaacs>we actually only ever care about _paused being *true* anyway
23:06:22  * mikealjoined
23:06:35  <isaacs>so the if (this._paused) guard is excessive
23:06:36  * toothrjoined
23:06:52  <isaacs>since you might want to call .resume() without ever having called pause()
23:07:31  * mikealquit (Read error: Connection reset by peer)
23:07:36  * mikeal1joined
23:08:52  <isaacs>bnoordhuis: ^
23:09:34  <bnoordhuis>isaacs: right, makes sense. lgtm
23:09:49  <CIA-108>node: isaacs master * r1df222f / lib/net.js : Fix breakage introduced in de65ba7 - http://git.io/7xzsBg
23:10:36  <isaacs>ok, lemme try ev-emul now that it's actually installing the files that zookeeper is pulling in :)
23:13:07  <isaacs>bnoordhuis: ev-emul looks pretty tight
23:13:19  <isaacs>bnoordhuis: just rebase on master, and let's ship it
23:13:28  <isaacs>bnoordhuis: the eio_custom signature change would be nice to get in there, too
23:14:03  <bnoordhuis>isaacs: do you know of a module that uses it?
23:18:58  <isaacs>zsock does
23:19:07  <isaacs>oh, you know what? my old example module probably does
23:19:13  <isaacs>i haven't touched it since the 0.2 days
23:19:16  <isaacs>one sec
23:20:29  <isaacs>sweet
23:20:35  <isaacs>npm install git://github.com/isaacs/node-async-simple
23:20:38  <isaacs>bnoordhuis: ^
23:20:55  <isaacs>bnoordhuis: someone had sent me a pull req to make it work with 0.7, and i never bothered to merge it in
23:21:00  <isaacs>now i'm glad i was so lazy :)
23:21:05  <bnoordhuis>:)
23:26:45  * mikeal1quit (Read error: Connection reset by peer)
23:59:11  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/ev-emul
23:59:19  <bnoordhuis>you will need some patches to nas though: https://gist.github.com/9a86a9d77ce48f28c353
23:59:38  * ryahjoined
23:59:45  <isaacs>bnoordhuis: !!
23:59:49  <isaacs>bnoordhuis: eio for windows?!?!?
23:59:51  <bnoordhuis>the 'return int' thing is something that's been gone from libeio since the pre-0.4 days
23:59:52  <isaacs>zomg
23:59:58  <bnoordhuis>isaacs: #ifndef :)