00:00:15  <isaacs>piscisaureus_: it could be thinking that its prefix is 'config' or something?
00:00:17  <isaacs>i dunno
00:02:05  <piscisaureus_>isaacs: https://github.com/isaacs/npm/blob/master/node_modules/mkdirp/index.js#L16-40
00:03:41  <isaacs>piscisaureus_: so it makes the child, gets enoent, makes the parent, gets eexist, repeat?
00:03:59  <piscisaureus_>isaacs: well, I think it should be getting EPERM
00:04:11  <isaacs>right
00:04:18  <isaacs>only EEXIST and ENOENT would be relevant here.
00:04:23  <piscisaureus_>but something is funny
00:04:35  <piscisaureus_>when I try to reproduce i get UNKNOWN
00:04:49  <isaacs>piscisaureus_: right, so why doens't it just abort immediately at that point?
00:05:04  <piscisaureus_>yeah, weird
00:05:11  * mikealjoined
00:06:18  <piscisaureus_>unless... the cwd if borked
00:06:30  <piscisaureus_>igorzi: yt?
00:08:52  <CIA-99>node: Nathan Rajlich master * re60b18b / Makefile : Make a fat binary for the OS X `make pkg`. - http://git.io/qP6KBQ
00:08:52  <CIA-99>node: Colton Baker master * rac9fa2b / lib/_debugger.js : Removed `rl.on('close', ...)` callback - http://git.io/KV30hQ
00:08:54  <CIA-99>node: Colton Baker master * rfd61bfc / (doc/api/readline.markdown lib/readline.js lib/repl.js): (log message trimmed)
00:08:54  <CIA-99>node: readline: ^Z (SIGSTP) handling
00:08:54  <CIA-99>node: Bugfix and update.
00:08:54  <CIA-99>node: - Fixed bug where Node's REPL wouldn't continue when returning from ^Z
00:08:54  <CIA-99>node: (SIGTSTP)
00:08:54  <CIA-99>node: - Removed old readline callback
00:08:55  <CIA-99>node: Readline API update with docs.
00:12:11  <CIA-99>libuv: Ben Noordhuis master * r8f716a9 / .gitignore : .gitignore gyp build artifacts - http://git.io/6HaO6Q
00:14:05  * travis-cijoined
00:14:05  <travis-ci>[travis-ci] joyent/libuv#100 (master - 8f716a9 : Ben Noordhuis): The build is still failing.
00:14:05  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/e504719...8f716a9
00:14:05  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/724527
00:14:05  * travis-cipart
00:18:50  * mikealquit (Quit: Leaving.)
00:22:46  * philipsquit (*.net *.split)
00:22:47  * saghulquit (*.net *.split)
00:22:47  * elijah-mbpquit (*.net *.split)
00:22:47  * mrb_bkquit (*.net *.split)
00:23:14  * travis-cijoined
00:23:14  <travis-ci>[travis-ci] joyent/node#482 (master - fd61bfc : Colton Baker): The build is still failing.
00:23:14  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/7f94b5c...fd61bfc
00:23:14  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/724503
00:23:14  * travis-cipart
00:23:37  * mikealjoined
00:23:44  <isaacs>lol bnoordhuis: https://github.com/joyent/node/commit/b72277183f9cb79274d6f8915e02b895c56b2288
00:23:55  <isaacs>bnoordhuis: do you know how many times i pushed ry for that?
00:24:36  <piscisaureus_>isaacs: did ry not want it?
00:24:53  <isaacs>no, he did not want the repl to be polluted ever.
00:25:00  <isaacs>he was wrong. this is awesome.
00:25:19  <isaacs>for a long time i was floating something somewhat like this, but eventually it got annoying when i was doing more node contributions.
00:25:29  <isaacs>ie, before i had commit access.
00:25:36  <isaacs>holy crap it's so nice.
00:25:42  <mmalecki>why does he list all the libs there?
00:25:51  <isaacs>mmalecki: so that we don't pull in deprecated modules by accident.
00:25:59  <mmalecki>Object.getOwnPropetyNames(process.binding('natives'))
00:26:03  <mmalecki>ah, right
00:26:05  <isaacs>it'd be better to just be Object.keys(process.binding('natives')) probably
00:26:09  <benvie>might be better to filter that list instead of the other way around
00:26:27  <isaacs>but... meh.
00:26:31  <isaacs>yeah
00:26:37  <mmalecki>yeah, it's not like it changes every day
00:26:37  <isaacs>whatever, that list will probably never change anyway
00:26:43  <isaacs>no, it will probably not ever change again.
00:26:50  <mmalecki>it'll!
00:26:56  <isaacs>what for?
00:26:57  <benvie>what about a clowns module
00:26:59  <mmalecki>sys.js is going away :)
00:27:11  <isaacs>mmalecki: sys isn't on the list ;)
00:27:12  <benvie>platform
00:27:39  <isaacs>i'm going to drag my feet on require('posix') or require('platform') or whatever.
00:27:40  <benvie>but yeah there's also a bunch of random stuff in natives
00:27:48  <isaacs>some _ stuff
00:28:34  * philipsjoined
00:28:34  * saghuljoined
00:28:34  * elijah-mbpjoined
00:28:34  * mrb_bkjoined
00:28:42  <piscisaureus_>I feel like checking out.
00:28:43  <benvie>even some non _ stuff you wouldn't ever include like freelist and buffer_ieee754
00:28:59  <benvie>and even buffer itself doesn't work well if you try to include it and run it again
00:29:09  <piscisaureus_>igorzi: I'd like to know what process.cwd() reports in that context where npm fails.
00:29:25  <piscisaureus_>igorzi: I've put a comment requesting `npm config list` output.
00:29:29  <isaacs>anyway, i'll go back to making the build. just wanted to come here and cheer
00:29:36  <piscisaureus_>bnoordhuis: what time are you coming tomorrow?
00:29:39  <isaacs>piscisaureus_: thanks for taking a look at that.
00:33:06  <bnoordhuis>piscisaureus_: around noon, i think
00:33:16  <piscisaureus_>bnoordhuis: right.
00:35:26  * mikealquit (Quit: Leaving.)
00:35:41  * pfox___joined
00:36:13  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
00:39:12  <bnoordhuis>you know, simple/test-dgram-send-error is not very nice - it sends udp packets to random ip addresses...
00:44:07  * piscisaureus_joined
00:58:34  * mikealjoined
01:00:51  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/compare/udp-fixes <- review?
01:02:05  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
01:03:07  * mikealquit (Client Quit)
01:05:13  <isaacs>bnoordhuis: lgtm
01:07:00  * isaacsquit (Remote host closed the connection)
01:07:26  <bnoordhuis>cool
01:08:26  <CIA-99>node: Ben Noordhuis master * rdefa637 / (3 files in 3 dirs): dgram: fix out-of-bound memory read - http://git.io/3m0Beg
01:08:26  <CIA-99>node: Ben Noordhuis master * r3502e45 / lib/dgram.js : dgram: defer send error to next tick - http://git.io/sF1rrg
01:15:37  * pfox___quit (Ping timeout: 244 seconds)
01:17:23  * isaacsjoined
01:19:33  * orlandovftwquit (Ping timeout: 252 seconds)
01:23:12  * travis-cijoined
01:23:12  <travis-ci>[travis-ci] joyent/node#483 (master - 3502e45 : Ben Noordhuis): The build is still failing.
01:23:12  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/fd61bfc...3502e45
01:23:12  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/724713
01:23:12  * travis-cipart
01:26:12  <isaacs>review please: https://github.com/isaacs/node/commit/7bd7f524c1c598885eac06e5df4190ee3f5cf93d
01:26:29  <isaacs>better link: https://github.com/isaacs/node/commit/v0.7.5-release
01:31:57  <TooTallNate>nice, 2 for me this time :)
01:40:13  * mikealjoined
01:45:45  * mikealquit (Quit: Leaving.)
01:50:16  * mikealjoined
02:11:57  * mikealquit (Quit: Leaving.)
02:13:16  <isaacs>ok, i've got a smoking gun that there's a race condition in chmod.
02:13:27  <isaacs>at least, fchmod.
02:13:58  <isaacs>looking at this test/simple/test-fs-chmod.js that fails randomly, if you run it in a loop, eventually it'll crash. dtruss shows that the stat is coming *before* the chmod.
02:18:59  * mikealjoined
02:23:49  * mikealquit (Client Quit)
02:24:47  * dapquit (Quit: Leaving.)
02:26:49  <isaacs>or at least, this test. simpler test seems to beok.
02:27:22  * dapjoined
02:33:03  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
02:33:42  * mralephjoined
02:38:03  * dapquit (Quit: Leaving.)
02:45:44  * mmaleckichanged nick to mmalecki[zzz]
02:49:45  * brsonquit (Ping timeout: 252 seconds)
02:51:41  <isaacs>omg, i'm dumb.
02:51:47  <isaacs>this race is so dumb.
02:51:49  <isaacs>holy shit.
02:52:03  * isaacsasync programming is too hard, i guess.
02:52:46  <isaacs>oh, wait, no, nvm.
02:52:50  <isaacs>they're different files. heh.
03:02:10  * orlandovftwjoined
03:14:58  * bnoordhuisquit (Ping timeout: 244 seconds)
03:17:57  * mralephquit (Ping timeout: 244 seconds)
03:22:32  * orlandovftwquit (Ping timeout: 240 seconds)
03:30:18  * pieternquit (Quit: pietern)
03:32:53  <isaacs>no, they're not different files! hah. symlinks.
03:32:55  <isaacs>foiled again.
03:33:08  * xaqjoined
03:38:00  * piscisaureus_joined
03:38:21  * piscisaureus_quit (Client Quit)
04:08:31  * orlandovftwjoined
04:10:37  * pfox___joined
04:55:51  * pfox___quit (Remote host closed the connection)
04:58:39  <CIA-99>node: isaacs master * r92cb684 / (test/fixtures/a1.js test/fixtures/a1.js):
04:58:40  <CIA-99>node: Change test fixture from symlink to regular file
04:58:40  <CIA-99>node: The only test using this is test/simple/test-fs-chmod.js, and it was
04:58:40  <CIA-99>node: treating a.js and a1.js as two separate files, resulting in a race
04:58:40  <CIA-99>node: condition. (Interestingly enough, it was *not* using the symlink file to
04:58:40  <CIA-99>node: test lchmod, which uses a different temp file.) - http://git.io/4YfQww
04:58:41  <CIA-99>node: isaacs master * r9a6f936 / test/simple/test-fs-watch.js : test-fs-watch: Add faster failure, and link to #2813 - http://git.io/9kqUSQ
04:59:06  * mikealjoined
05:13:20  * travis-cijoined
05:13:20  <travis-ci>[travis-ci] joyent/node#484 (master - 9a6f936 : isaacs): The build is still failing.
05:13:20  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/3502e45...9a6f936
05:13:20  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/725587
05:13:20  * travis-cipart
05:13:24  <indutny_sleeping>isaacs: morning :)
05:13:26  * indutny_sleepingchanged nick to indutny
05:13:31  <isaacs>morning :)
05:15:34  <indutny>that pull request lgtm
05:15:41  <indutny>going to build and test it
05:15:50  <indutny>isaacs: %
05:15:52  <indutny>^
05:17:35  * mikealquit (Quit: Leaving.)
05:17:51  <isaacs>indutny: oh, the cluster thingie?
05:18:10  <indutny>yes
05:18:14  <isaacs>indutny: awesome, thanks.
05:18:35  <isaacs>indutny: if you can find time at some point to figure out why the debugger tests seem kind of flaky, that'd be good, too
05:18:43  <isaacs>indutny: they seem like they pass sometimes, but fail other times. maybe a race somewhere.
05:18:56  <indutny>ok, of course
05:19:05  <isaacs>indutny: thanks. the sooner we can get to 0 failures, the better.
05:19:24  <isaacs>we're not that far away, really :)
05:19:52  <indutny>:)
05:28:37  <indutny>isaacs: that PR works fine to me
05:28:50  <isaacs>kewl
05:28:57  <isaacs>thanks
05:29:04  * mikealjoined
05:29:04  <indutny>ok, I'll let Andreas fix one small issue with it
05:29:10  <indutny>and then merge it in master
05:29:16  <isaacs>k
05:29:23  <isaacs>it'd be nice to squash it down to a few logical changes.
05:29:31  <isaacs>there's a bit of back-and-forth there that's kind of confusing
05:29:49  <indutny>haha
05:29:55  <indutny>of course
05:30:03  <indutny>it's too raw now
05:30:11  <isaacs>if that's too much of a pain, a `git merge --no-ff` is good sometimes, too
05:30:25  <isaacs>then you can at least have one "thing" that corresponds to the change
05:31:05  <indutny>isaacs: no, better squash such things
05:31:26  <indutny>this PR contains various small-fix-for-my-previous-commit stuff
05:31:30  * isaacsquit (Remote host closed the connection)
05:32:57  * orlandovftwquit (Read error: Connection reset by peer)
05:34:56  * orlandovftwjoined
05:37:50  <indutny>ttyl
06:01:23  * mikealquit (Quit: Leaving.)
06:11:30  * mikealjoined
06:13:04  * mikealquit (Client Quit)
06:16:07  * pfox___joined
06:30:24  * paddybyersquit (Quit: paddybyers)
06:35:06  * hij1nxjoined
06:39:33  * paddybyersjoined
07:19:50  * isaacsjoined
07:37:38  * sh1mmerquit (Read error: Connection reset by peer)
07:38:05  * sh1mmerjoined
07:39:05  * sh1mmerquit (Client Quit)
07:59:59  * mmalecki[zzz]changed nick to mmalecki
08:06:19  * mralephjoined
08:19:10  <CIA-99>node: isaacs master * r1b5048b / doc/api/globals.markdown : doc: refer to modules doc rather than src/node.js - http://git.io/sZ1c5g
08:33:27  * travis-cijoined
08:33:27  <travis-ci>[travis-ci] joyent/node#485 (master - 1b5048b : isaacs): The build is still failing.
08:33:27  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/9a6f936...1b5048b
08:33:27  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/726580
08:33:27  * travis-cipart
08:44:15  * mjr_quit (Quit: mjr_)
08:49:46  * xaqquit (Remote host closed the connection)
08:55:21  * isaacsquit (Remote host closed the connection)
09:16:08  * mralephquit (Quit: Leaving.)
09:16:38  * hij1nxquit (Quit: hij1nx)
09:23:50  * AvianFluquit (Ping timeout: 240 seconds)
10:08:59  * dshaw_quit (Quit: Leaving.)
10:14:21  * mmaleckichanged nick to mmalecki[out]
10:18:51  * orlandovftwquit (Ping timeout: 245 seconds)
10:46:58  * AndreasMadsenjoined
11:16:54  <benvie>haha wtf "Okay, so I managed to fork() a console process with NtCreateProcess and RtlCreateUserThread, and actually print something to the console from the forked clone."
11:19:45  <benvie>http://processhacker.sourceforge.net/forums/viewtopic.php?f=8&t=147 NtCreateProcess, RtlCloneUserProcess, NtCreateUserProces
11:21:30  * bnoordhuisjoined
11:21:40  <benvie>it's a real live fork for windows. Totally hack but it's the real thing
11:28:13  <benvie>http://processhacker.sourceforge.net/forums/download/file.php?id=49 is an amazing thing, wow
12:09:33  * bnoordhuisquit (Quit: leaving)
12:10:07  * bnoordhuisjoined
12:42:42  <CIA-99>node: Ben Noordhuis master * rb92a919 / (src/node_buffer.cc src/node_script.cc src/v8_typed_array.cc): Remove unused variables. - http://git.io/hSVtCw
12:57:19  * travis-cijoined
12:57:19  <travis-ci>[travis-ci] joyent/node#486 (master - b92a919 : Ben Noordhuis): The build is still failing.
12:57:19  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/1b5048b...b92a919
12:57:19  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/727672
12:57:19  * travis-cipart
13:00:43  * piscisaureus_joined
13:12:39  <CIA-99>node: Shigeki Ohtsu master * r680d75a / (common.gypi configure tools/gyp_addon): build: make default_configuration consistent with BUILDTYPE - http://git.io/xQZxyA
13:12:39  <CIA-99>node: Shigeki Ohtsu master * r58e5d69 / (common.gypi tools/gyp_addon): build: fix the case where config.gypi doesn't exist - http://git.io/WH-Z6Q
13:12:39  <CIA-99>node: Shigeki Ohtsu master * r491ec17 / common.gypi : build: change default BUILDTYPE of gyp to Release - http://git.io/OKqjoA
13:12:40  <CIA-99>node: Shigeki Ohtsu master * rc1fcb1d / vcbuild.bat :
13:12:40  <CIA-99>node: build: change default BUILDTYPE of vcbuild.bat
13:12:40  <CIA-99>node: From Debug to Release. - http://git.io/KtTQOg
13:17:41  * mmalecki[out]changed nick to mmalecki
13:27:23  * travis-cijoined
13:27:23  <travis-ci>[travis-ci] joyent/node#487 (master - c1fcb1d : Shigeki Ohtsu): The build is still failing.
13:27:23  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/b92a919...c1fcb1d
13:27:23  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/727886
13:27:23  * travis-cipart
13:35:48  <mmalecki>oh? I segfaulted node?
13:36:01  <mmalecki>interesting...
13:42:50  <mmalecki>bnoordhuis: you around to help debug it? me and gdb don't really get together too well
13:43:10  <indutny>mmalecki: I can help you probably
13:43:27  <indutny>mmalecki: I'm using it a lot in last month
13:43:36  <indutny>(or two)
13:43:39  <mmalecki>indutny: oh, awesome! so, how do I get started XD ?
13:43:55  <indutny>gdb /path/to/your/app
13:43:57  <indutny>oh
13:43:57  <indutny>wait
13:43:59  <indutny>man gdb
13:44:01  <indutny>^^
13:44:11  * wankdankerjoined
13:44:15  <indutny>and google it
13:44:15  <CIA-99>libuv: Ben Noordhuis master * ra5082e8 / test/run-tests.c : test: add proper type casts - http://git.io/PGCgnA
13:44:16  <CIA-99>libuv: Ben Noordhuis master * r4240f0d / test/test-fs.c : test: check uv_fs_*() return values - http://git.io/jfP_2w
13:44:35  <bnoordhuis>mmalecki: what's up?
13:44:52  <mmalecki>indutny: the problem is, it's npm segfaulting node
13:44:57  <mmalecki>not node itself
13:45:08  <mmalecki>and I'm not sure how to give it params
13:45:09  <bnoordhuis>mmalecki: gdb --args path/to/node npm [args]
13:45:12  <bnoordhuis>then r for run
13:45:13  <indutny>^^
13:45:14  <indutny>yes
13:45:25  <bnoordhuis>make sure you use a debug build of node
13:45:26  <indutny>bnoordhuis: you're fast :)
13:45:48  <bnoordhuis>that's what my girlfriend always says too
13:45:48  <mmalecki>it's release, actually. sec, I'll rebuild it
13:45:59  <mmalecki>sec == 5 minutes on this piece of shit
13:46:01  <indutny>bnoordhuis: oh, btw
13:46:06  * travis-cijoined
13:46:06  <travis-ci>[travis-ci] joyent/libuv#101 (master - 4240f0d : Ben Noordhuis): The build is still failing.
13:46:06  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/8f716a9...4240f0d
13:46:06  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/728024
13:46:06  * travis-cipart
13:46:18  <indutny>bnoordhuis: I think I will develop twss.js clone but for another phras
13:46:25  <indutny>bnoordhuis: "That's what Ben said"
13:46:29  <mmalecki>lol
13:46:30  <bnoordhuis>oh dear...
13:46:32  <indutny>haha
13:46:41  <mmalecki>also, good guy Ben https://github.com/joyent/node/pull/2571#issuecomment-4136113
13:46:54  <indutny>like "you're smart kid, but that's to dumb"
13:46:58  <indutny>s/to/too
13:47:10  <indutny>ok, brb
13:47:19  * indutnyis going to drink tea
13:47:24  <mmalecki>bnoordhuis: ./configure --debug?
13:47:48  <mmalecki>indutny: tea? don't you guys drink only vodka there?
13:47:49  <indutny>mmalecki: make BUILDTYPE=Debug
13:47:58  <indutny>mmalecki: are you from Poland?
13:48:03  <mmalecki>indutny: yeah
13:48:18  <indutny>so you should know it better :)
13:48:25  <mmalecki>haha, yeah, right
13:48:44  <indutny>people are drinking beer here
13:49:01  <indutny>vodka is an old-schoold
13:49:20  <indutny>brb
13:49:27  * indutnychanged nick to indutny_tea_vodk
13:49:31  <indutny_tea_vodk>f*ck
13:49:35  <mmalecki>lulz
13:50:23  <mmalecki>oh, debug build is faster, sweet
13:51:38  <AndreasMadsen>indutny_tea_vodk: did you figure out if net._setSimultaneousAccepts(sendHandle); need to be called for sockets?
13:53:01  <mmalecki>argh, I need to find myself some legit linux laptop
13:53:55  <AndreasMadsen>mmalecki: I'm reinstalled my windows machine, not done yet :)
13:54:46  <mmalecki>AndreasMadsen: I decided to avoid windows when possible :)
13:55:04  <indutny_tea_vodk>AndreasMadsen: I think it doesn't
13:55:20  <indutny_tea_vodk>but wait
13:55:21  <mmalecki>don't trust him, he's drunk already
13:55:21  <AndreasMadsen>mmalecki: I need to test immortal on it, and also to check if net._setSimultaneousAccepts(sendHandle); can be avoided
13:55:21  <indutny_tea_vodk>hm...
13:55:28  <indutny_tea_vodk>AndreasMadsen: can I accept on socket?
13:55:47  <indutny_tea_vodk>AndreasMadsen: can you please test it?
13:55:56  * indutny_tea_vodkchanged nick to i_drunk_anyway
13:56:03  <i_drunk_anyway>mmalecki: what did ya say?
13:56:15  <AndreasMadsen>indutny_tea_vodk: I don't have a working windows machine yet.
13:56:22  <mmalecki>i_drunk_anyway: nothing, nothing :D
13:56:34  <AndreasMadsen>i_drunk_anyway: ^
13:56:41  <i_drunk_anyway>AndreasMadsen: sorry, I was talking about setSimultaneousAccepts
13:56:52  <i_drunk_anyway>AndreasMadsen: can you test if that's needed for Sockets
13:57:01  <i_drunk_anyway>AndreasMadsen: from short look - no
13:57:15  <AndreasMadsen>i_drunk_anyway: I know but it is a windows thing so I can't test it.
13:57:19  <AndreasMadsen>yet
13:57:22  <i_drunk_anyway>aaah
13:57:23  <i_drunk_anyway>I see now
13:57:29  <AndreasMadsen>man he is drunk
13:57:34  <i_drunk_anyway>who?
13:57:41  <AndreasMadsen>you bitch
13:57:52  * i_drunk_anywaychanged nick to indutny
13:57:55  <indutny>that wasn't funny
13:58:23  <mmalecki>AndreasMadsen: behave
13:58:55  <AndreasMadsen>indutny: but since you are drunk you will properly not remember it
13:58:56  <mmalecki>uhm... I have a weird feeling that this build got into loop
13:59:08  <indutny>AndreasMadsen: great
13:59:27  <mmalecki>oh man, these russians, they always remember things like that!
13:59:40  * AvianFlujoined
13:59:40  <mmalecki>watch out for black vans
13:59:44  <AndreasMadsen>oh shit
13:59:49  * AndreasMadsenchanged nick to madson
13:59:59  <mmalecki>yo AvianFlu
14:00:01  <madson>going undercover
14:00:06  <AvianFlu>sup
14:00:30  <mmalecki>AvianFlu: just casually sigfaulting node :)
14:00:56  <AvianFlu>segfaults! love a segfault
14:00:59  <AvianFlu>what kind today
14:01:24  <mmalecki>not sure yet! building node in debug mode to find out
14:01:41  <mmalecki>AvianFlu: you're at rs already, btw?
14:01:59  <AvianFlu>mmalecki, it's 6am
14:02:06  <mmalecki>AvianFlu: oh lol
14:02:09  <AvianFlu>worldtimebuddy.com ftw
14:02:18  <mmalecki>AvianFlu: we need it in kohai
14:02:49  <mmalecki>also, why is this building v8 the second time
14:02:58  <mmalecki>s/the//
14:06:33  <madson>indutny: I have squached https://github.com/joyent/node/pull/2772 now, will fix the .connections issue and add documentation now.
14:06:48  <indutny>madson: ok
14:06:50  <mmalecki>indutny, bnoordhuis: what should I do if it doesn't segfault with debug build?
14:07:26  <madson>mmalecki: get drunk
14:07:46  <mmalecki>madson: no no, not now
14:09:30  <madson>oh, hate writing documentation to child_process, the Structure is so bad.
14:11:40  <indutny>mmalecki: you can try debugging release build
14:11:46  <indutny>mmalecki: :)
14:12:56  <mmalecki>indutny: I guess this is going to be very painful?
14:13:28  <indutny>mmalecki: no, but it may work fine too
14:13:38  <indutny>mmalecki: that means that something is racey
14:13:44  <mmalecki>actually, it segfaulted!
14:13:47  <indutny>oh, cool
14:13:51  <indutny>bt then
14:13:56  <indutny>run `bt`
14:14:05  <mmalecki>got it
14:14:24  <indutny>show it!
14:14:28  <mmalecki>lemme put it somewhere
14:16:24  <mmalecki>https://gist.github.com/1893028
14:16:38  <indutny>ok
14:16:48  <mmalecki>I probably should install debug packages for things, sec
14:17:03  <indutny>yeah
14:17:48  <indutny>send it to mraleph anyway
14:17:51  <indutny>s/send/sent
14:19:52  <indutny>mmalecki: is that latest node?
14:21:08  <mmalecki>indutny: yeah, master
14:21:09  * mralephjoined
14:21:10  <mmalecki>sec, food
14:21:17  <indutny>mmalecki: mraleph is here :)
14:21:39  <indutny>mraleph: 4:20 < indutny> mmalecki: is that latest node?
14:21:40  <indutny>14:21 < mmalecki> indutny: yeah, master
14:22:16  <mraleph>mmalecki: does node_g crash on you?
14:23:09  <indutny>mraleph: I think he's out for a food
14:23:39  <mraleph>ok
14:23:40  <mmalecki>mraleph: no
14:23:47  <mraleph>mmalecki: sad
14:23:49  <mmalecki>mraleph: only release build
14:24:02  <mraleph>mmalecki: can you build release build with -g to get better debugging information?
14:24:46  <mmalecki>mraleph: actually, mind if I grab some food first? 20 minutes?
14:27:17  <mmalecki>ok, brb
14:27:18  <mraleph>mmalecki: how can I :-)
14:27:23  <mraleph>foooooood
14:28:07  <piscisaureus_>and theeeeen?
14:28:16  <AvianFlu>BRAAAAINSSS!!
14:28:18  <AvianFlu>...oh wait, nevermind
14:33:11  <mraleph>piscisaureus_: and then some zubrowka
14:33:45  <indutny>mraleph: hahahaha
14:33:57  <indutny>omg
14:34:06  <indutny>pertzovka
14:40:37  <mmalecki>oh man, I love having KFC near
14:40:47  <mmalecki>zubrowka! <3
14:41:11  <mmalecki>so, how do I build with -g?
14:42:22  <mraleph>I don't know :-)
14:42:30  <mraleph>CXXFLAGS=-g CFLAGS=-g
14:42:33  <tjfontaine>CFLAGS=-g CXXFLAGS=-g make
14:42:33  <mraleph>something like that.
14:42:40  <mraleph>\o/
14:42:42  <tjfontaine>hypothetically
14:42:45  <tjfontaine>at least :P
14:43:00  <mraleph>probably will conflict with some other flags like -fdata-sections though.
14:43:10  <mraleph>I usually just edit gyp file
14:43:19  <mraleph>or whatever is used these days.
14:43:42  <mmalecki>build systems change fast, don't they :) ?
14:45:57  <mmalecki>ok, modified config.gypi, looks like it works
14:46:14  <mmalecki>btw, I see to be getting different backtraces every time
14:47:36  <indutny>mmalecki: that's bad
14:47:44  <indutny>mmalecki: probably stack is corrupt
14:47:47  <mmalecki>indutny: I'm aware :)
14:47:49  <indutny>and it's segfaulting on return
14:47:54  <mmalecki>indutny: yeah, gdb says so
14:47:55  <indutny>actually, after return
14:48:06  <indutny>gdb is smart
14:48:47  <mmalecki>(btw, if anyone wants to ssh in, ping me)
14:49:55  <mraleph>so what's the back trace with better debug info?
14:50:02  <mmalecki>mraleph: still building
14:50:06  <mraleph>ah ok
14:50:43  <mmalecki>mraleph: it's an old tablet I wanted to use for running seeeecret devops script for monitoring and stuff
14:51:01  <mmalecki>so things take time here :/
14:51:53  <mraleph>I see
14:52:09  <mraleph>I am accustomed to building V8 with -j18 :-)
14:52:22  <mmalecki>mraleph: oh? 18 cores?
14:52:45  <indutny>xeon
14:52:47  <indutny>probably
14:52:49  <mmalecki>oh man
14:53:02  <indutny>some people have more :D
14:53:23  <mmalecki>I should try building node on cluster with distcc
14:53:34  * mmaleckichecks his account balance
14:53:46  <madson>I don't understant, it don't follow 2, 4, 8, 16, 32 :/
14:54:19  <madson>*understand
14:57:04  <mmalecki>ok, I guess I can't afford it yet
14:57:12  <mmalecki>anyway, build's over
14:58:07  <madson>I've got so many kicks for not using if statements, and now I see this -> https://github.com/joyent/node/blob/master/lib/net.js#L713
14:59:33  <mmalecki>mraleph: https://gist.github.com/1893191
15:00:09  * sh1mmerjoined
15:12:21  * philipsquit (Excess Flood)
15:13:14  * philipsjoined
15:14:50  <mmalecki>mraleph: let me know if you want to ssh, btw
15:17:56  <AvianFlu>madson, conditional assignment is appropriate sometimes, but too dense to read in others
15:18:03  <AvianFlu>that port function is clean on one line with it
15:18:26  <AvianFlu>but it gets confusing quick, especially if you nest them like I do to troll people sometimes
15:19:22  <madson>AvianFlu: I just didn't think node had any conditional assignment.
15:19:41  <AvianFlu>madson, some if statements are just too short I guess :D
15:20:10  <madson>AvianFlu: :)
15:20:10  <mmalecki>are we talking about `val || (val = blah)`?
15:20:13  <AvianFlu>this is probably a case where you just need a good reason to make it less readable
15:20:17  <AvianFlu>mmalecki, no, ternary operator
15:20:21  <mmalecki>AvianFlu: ah
15:20:25  <madson>mmalecki: see -> https://github.com/joyent/node/blob/master/lib/net.js#L713
15:20:42  <mmalecki>nice one
15:20:48  <mraleph>a lot of generated code, sad.
15:20:55  <AvianFlu>personally, I like that, cause it makes that one line
15:21:01  <mraleph>does it repro on newest V8?
15:21:06  <indutny>mmalecki: heya
15:21:19  <indutny>can you run disas at segfault
15:21:27  <mmalecki>mraleph: it's on node's master, is it newest?
15:21:31  <indutny>would be interesting to see what causes it
15:21:32  <mmalecki>probably not
15:21:36  <mmalecki>indutny: ok, sec
15:22:14  <mmalecki>indutny: "No function contains program counter for selected frame."
15:22:16  <madson>x = Number(x); if (x >= 0) return x; else return false;
15:22:27  <madson>^ that isn't very much longer
15:22:55  <madson>15 chars longer
15:23:00  <indutny>mmalecki: use offset
15:23:10  <indutny>mmalecki: like `disas 0x....,+100`
15:23:33  <mmalecki>indutny: sorry for being a noob, but what should I put in place of dots?
15:23:34  <indutny>mmalecki: x=+x
15:23:42  <indutny>mmalecki: top address from bt
15:24:03  <indutny>madson: x=+x
15:24:10  <mraleph>mmalecki: no it is not newest
15:25:21  <madson>indutny: you need to check for >= 0, but yes :D
15:25:46  <indutny>madson: I mean that instead of x = Number(x) you can use x = + x
15:25:56  <indutny>ooh
15:26:03  <madson>indutny: so why don't we :/
15:26:05  <indutny>unary plus and minus, that's what I forgot in dotlang
15:26:11  <indutny>crap
15:26:32  <mmalecki>indutny: updated https://gist.github.com/1893191
15:26:51  <indutny>mmalecki: ooh
15:27:04  <indutny>mmalecki: can you decrease address in 0x... a little bit
15:27:18  <mmalecki>indutny: how much is a little bit?
15:27:42  <indutny>mmalecki: 0x20
15:28:12  <mraleph>what's in ebx? info reg
15:28:25  <indutny>mmalecki: yeah, that's what I'm trying to figure out
15:28:34  <indutny>err
15:28:38  <indutny>mraleph: ^
15:28:41  <mraleph>looks like an instance type check. probably ebx does not contain a valid object for whatever reason.
15:29:02  <mraleph>I meant ecx
15:29:16  * hij1nxjoined
15:29:18  <mmalecki>indutny, mraleph https://gist.github.com/1893191
15:29:52  <mmalecki>uhm... 0xbeeddead?
15:30:01  <indutny>yeah
15:30:08  <mraleph>yeah that looks nasty
15:30:08  <indutny>so it checks if ecx is not smi
15:30:16  <mraleph>it's uint32_t zap value.
15:30:40  <mraleph>can you run node with --trace-deopt?
15:31:04  <mraleph>also --trace-opt --trace-osr
15:31:04  <kohai>also has -1 beer
15:31:12  <indutny>zap value!!!
15:31:12  <mmalecki>also++
15:31:12  <kohai>also has 0 beer
15:31:14  <mmalecki>sure
15:31:24  <indutny>I like it :)
15:31:29  <AvianFlu>what's a zap value? what it gets set to if things blow up?
15:31:31  <mmalecki>can I close this session or should I keep it?
15:32:08  <indutny>AvianFlu: I think slot is getting filled with zap value on GC or on init
15:32:18  <indutny>it's needed to perform checks
15:32:22  <indutny>mmalecki: keep it :)
15:32:36  <AvianFlu>gotcha
15:32:38  <mraleph>if it reproable then you don't
15:33:00  <mraleph>this particular zap value is used during stack translation between optimized and non-optimized frames.
15:33:08  <indutny>mraleph: gotcha
15:35:38  <mmalecki>mraleph: https://gist.github.com/1893330
15:36:47  <indutny>mmalecki: Programowanie
15:36:51  <indutny>Poland ^ :)
15:36:53  <indutny>like it
15:37:02  <indutny>it's like funny russian :D
15:37:25  <mmalecki>indutny: haha :). yeah, it's last place where it's called "Programowanie"
15:37:37  <mmalecki>I've "dev" all over the place now
15:38:02  <mmalecki>because Downloads is "Pobrane", so I can't tab complete "Programowanie" with one key
15:39:01  <indutny>hahaha
15:39:05  <indutny>Pobrane
15:39:18  <indutny>I loooooooove polish
15:39:32  <mmalecki>haha, I can learn you some later :)
15:39:44  <mmalecki>er, teach
15:40:01  <AvianFlu>it's more fun pronouncing "Programowanie" in english, though, I gotta say
15:40:15  <AvianFlu>especially like a texan
15:40:43  * mmaleckitries
15:40:54  <mmalecki>oh lol
15:41:36  * AvianFluquit (Quit: Leaving)
15:46:55  * AvianFlujoined
15:57:22  * madsonchanged nick to AndreasMadsen
15:59:57  <benvie>windows fork --> https://gist.github.com/1893415
16:06:25  * AndreasMadsenquit (Remote host closed the connection)
16:06:28  <DrPizza>benvie: idgi
16:06:34  <DrPizza>benvie: after all that code
16:06:41  <DrPizza>benvie: it just calls the OS CreateProcess function
16:07:07  * piscisaureus_changed nick to piscisaureus
16:07:31  * xaqjoined
16:19:07  <benvie>well it also copies it no?
16:19:11  <benvie>before doing that
16:19:19  <DrPizza>well I don't know, does it?
16:19:25  <DrPizza>it would be more convincing if it called it!
16:21:27  <benvie>I mean there's exported functions in it that do things to that effect, but it's a lib
16:24:49  * AndreasMadsenjoined
16:27:44  <benvie>the magic is in CreateProcessInternalW
16:29:08  * isaacsjoined
16:29:43  <isaacs>good morning
16:30:18  <AndreasMadsen>morning
16:30:24  <indutny>morn
16:33:28  * sh1mmerquit (Quit: sh1mmer)
16:42:36  * hij1nxquit (Quit: hij1nx)
16:54:05  * hij1nxjoined
16:54:27  <isaacs>test, please: http://nodejs.org/dist/v0.7.5/node-v0.7.5-RC0.tar.gz
16:54:51  * stephankjoined
17:13:02  * xaqquit (Remote host closed the connection)
17:20:47  * pieternjoined
17:28:34  * orlandovftwjoined
17:29:45  * dapjoined
17:40:02  * AvianFluquit
17:41:27  <bnoordhuis>piscisaureus: https://github.com/bnoordhuis/libuv/commit/d3efefb0435e950251350744f84e9bbb8d783871 <- review!
17:42:34  * mikealjoined
17:42:56  <pfox___>so, im working still on landing the libuv bindings for rust, and the windows/mingw32 build is failing
17:43:24  <pfox___>any of the tests that have handlers registered are hanging.. gdb isn't showing anything besides the hangs, either
17:43:32  <pfox___>curious if there are any known landmines on the libuv side, here?
17:43:52  <pfox___>ive repro'd on a 2008r2 64bit vm i have
17:44:05  * AvianFlujoined
17:45:24  <pfox___>could very be caused by a rust messaging hiccup, but just curious if there's anything i should know. im about to begin with the debug printfs
17:50:02  * AndreasMadsenquit (Remote host closed the connection)
17:52:46  * sh1mmerjoined
17:53:13  <bnoordhuis>pfox___: the win32 side of libuv isn't entirely thread-safe
17:53:32  <bnoordhuis>(yet anyway - bert's going to fix that if he knows what's good for him)
17:53:53  <pfox___>what do you mean by that?
17:54:10  <pfox___>i pretty much do everything on the libuv thread anyways, via uv_aync_send..
17:54:20  <bnoordhuis>in that case you should be fine
17:54:31  <pfox___>that's kind of the cornerstone of the rust binding.. we isolate libuv in a single thread and do operations against via a uv_async_t
17:54:53  <bnoordhuis>in that case you can ignore my previous remark :)
17:55:01  <pfox___>but running multiple loops in multiple threads should be ok?
17:55:10  <bnoordhuis>no
17:55:21  <bnoordhuis>timers, for example, are stored in a single, static binary tree
17:55:42  <bnoordhuis>piscisaureus: ^
17:55:48  <bnoordhuis>no putting it off anymore
17:56:20  <pfox___>is this just a windows limitation, or in general?
17:56:55  * hij1nxquit (Quit: hij1nx)
17:56:56  <pfox___>because i built the library around the assumption that i could spawn multiple uv_loop_t's in different threads, and as long as we don't, ahem, cross-the-stream we should be ok?
17:57:13  <bnoordhuis>pfox___: it's a bug in the windows implementation that we'll fix
17:57:21  <pfox___>gotcha
17:57:24  <bnoordhuis>and when i say 'we' i mean 'bert' :)
18:00:26  <bnoordhuis>pfox___: https://github.com/joyent/libuv/issues/320
18:02:35  * xaqjoined
18:05:23  <pfox___>:)
18:16:40  <CIA-99>libuv: Ben Noordhuis master * r09c722e / (src/uv-common.c src/uv-common.h):
18:16:40  <CIA-99>libuv: common: make uv__set_*_error() return -1
18:16:40  <CIA-99>libuv: So you can do `if (errno) return uv__set_sys_error(loop, errno);`. - http://git.io/Vs5qmQ
18:16:40  <CIA-99>libuv: Ben Noordhuis master * rd3efefb / (9 files in 5 dirs):
18:16:40  <CIA-99>libuv: linux: share inotify fd across event loop
18:16:41  <CIA-99>libuv: Previously, a new inotify fd was created for each watcher, making it quite easy
18:16:42  <CIA-99>libuv: to run into the system-wide fs.inotify.max_user_instances limit (usually 128).
18:16:42  <CIA-99>libuv: Fixes #300. - http://git.io/Esxi7Q
18:17:37  * AndreasMadsenjoined
18:18:34  * travis-cijoined
18:18:34  <travis-ci>[travis-ci] joyent/libuv#102 (master - d3efefb : Ben Noordhuis): The build is still failing.
18:18:34  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/4240f0d...d3efefb
18:18:34  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/729375
18:18:34  * travis-cipart
18:24:22  * TooTallNatejoined
18:28:36  * bnoordhuisquit (Quit: Leaving)
18:38:56  * piscisaureusquit (Ping timeout: 265 seconds)
18:39:30  * sh1mmer_joined
18:39:43  * sh1mmer_quit (Remote host closed the connection)
18:40:25  * sh1mmer_joined
18:43:13  * sh1mmerquit (Ping timeout: 260 seconds)
18:43:14  * sh1mmer_changed nick to sh1mmer
18:45:13  * orlandovftwquit (Ping timeout: 265 seconds)
18:52:41  * `3rdEdenjoined
18:52:54  * igorziquit (Ping timeout: 245 seconds)
18:53:27  * sh1mmer_joined
18:53:36  * sh1mmer_quit (Remote host closed the connection)
18:54:14  * sh1mmer_joined
18:56:46  * sh1mmerquit (Ping timeout: 245 seconds)
18:56:47  * sh1mmer_changed nick to sh1mmer
19:01:29  <AndreasMadsen>indutny: think I finally got server.on('close') to work -> https://github.com/AndreasMadsen/node-1/commit/f2ebcc8307673996bc540aefbfc7ba437d37742d
19:01:44  <AndreasMadsen>But another testcase would be in order.
19:04:18  <AndreasMadsen>^ I'm just surprised about how complicated this is, because we can't relay the destroy call live (the speed issue).
19:06:04  <AndreasMadsen>I'm open if there are easier ways
19:06:18  <AndreasMadsen>isaacs: ^ you too are assigned to this, right?
19:11:19  * brsonjoined
19:17:47  * xaqquit (Remote host closed the connection)
19:17:57  * mikealquit (Quit: Leaving.)
19:21:29  <isaacs>AndreasMadsen: fluch?
19:21:33  <isaacs>AndreasMadsen: i think you mean flush?
19:21:41  <AndreasMadsen>isaacs: oh, hehe
19:21:43  <AndreasMadsen>will fix
19:22:03  <isaacs>also, should handleConvertion be handleConversion?
19:22:26  <isaacs>(i know, english was a bad choice for humans to use for programming ;)
19:22:55  <AndreasMadsen>oh, it miss a "s" will also fix that
19:25:10  <AndreasMadsen>isaacs: but besides the spelling ...
19:25:15  <AndreasMadsen>?
19:28:46  <isaacs>s/_useingSlaves/_usingSlaves/g and a few lint failures introduced. (sorry, still reading...)
19:29:09  <isaacs>how important is it to make 'connections' private and a getter?
19:29:47  <isaacs>can we just add the getter when usingSlaves is set, and not in the "normal" case?
19:30:03  <isaacs>or just document that it's not relevant when using slaves, and leave it as-is?
19:32:18  * xaqjoined
19:32:51  <pfox___>question: calls to uv_async_send() before calling uv_run() .. they will cause the handle's callback to processed (probably once) after the loop starts
19:33:18  <pfox___>ive observed this on linux.. does this work on windows, too?
19:33:24  <pfox___>i mean, it should obv
19:34:11  <isaacs>AndreasMadsen: this.slave.on('internalMessage', function removeMe?
19:34:44  <isaacs>AndreasMadsen: L37, that one doesn't get removed
19:34:53  <isaacs>er, L137
19:35:26  <AndreasMadsen>isaacs: oh it did at a port, but the design was changed so It shouldn't be removed :)
19:35:48  <AndreasMadsen>I will make a jslint commit when done
19:35:55  <isaacs>yeah, that's minor stuff
19:35:59  <isaacs>(lint i mean)
19:36:04  <isaacs>but stuff that's confusing is more important
19:36:19  <AndreasMadsen>isaacs: how is connections private?
19:36:32  <isaacs>AndreasMadsen: i mean, changing connections -> _connections
19:36:36  * `3rdEdenquit (Quit: Leaving...)
19:36:42  <isaacs>AndreasMadsen: is that really necessary?
19:36:49  <isaacs>a getter seems like more magic than necessary here.
19:37:25  <AndreasMadsen>isaacs: not sure, I can change it back. Hopefully userland will read the documentation carefully :)
19:38:19  <AndreasMadsen>isaacs: but it shouldn't be a speed issue, if thats the problem.
19:38:29  <isaacs>AndreasMadsen: no, it's more a confusion issue :)
19:39:20  * orlandovftwjoined
19:39:23  * orlandovftwquit (Client Quit)
19:39:41  * orlandovftwjoined
19:39:53  <AndreasMadsen>isaacs: If it reverted there should be some other public property indicating .connections dosn't work.
19:40:24  <isaacs>AndreasMadsen: you can just set it to -Infinity
19:40:29  <isaacs>then it'll always be -Infinity
19:40:44  <isaacs>and still the same type, and connections>maxConnection will always be false
19:41:01  <AndreasMadsen>isaacs: but it is needed as a counter
19:41:07  <isaacs>even in the using slaves case?
19:41:22  <AndreasMadsen>or ._connections is, isaacs yes the close event depend on it.
19:41:28  <isaacs>oh, ok
19:41:55  <isaacs>oh, right if (self._handle || self._connections) return;
19:42:17  <isaacs>what is self._connections when using slaves?
19:42:42  <AndreasMadsen>what self.connections was before
19:43:14  <AndreasMadsen>self.connections was replaced with self._connections, and self.connections become a getter
19:44:23  <isaacs>right.
19:44:32  <AndreasMadsen>self.connections will return self._connections - unless slaves was used in that case it return false,
19:44:33  <isaacs>when using slaves, though, what does the _connections counter count?
19:45:04  <isaacs>what is _connections counting in the "usingSlaves" case?
19:45:07  <AndreasMadsen>isaacs: connections, but in a unreliable way.
19:45:22  <isaacs>so... why are we waiting on it to emit('close') then?
19:45:26  <AndreasMadsen>It is first reliable when it becomes zero
19:45:32  <isaacs>ok.
19:46:09  <isaacs>why is it not reliable?
19:46:51  <AndreasMadsen>isaacs: because socket.destroy() isn't called in the parent after it has been moved to the slave.
19:47:22  <AndreasMadsen>isaacs: socket.destroy() is called in the salve, not the parent where the server live.
19:47:48  <AndreasMadsen>and socket.destroy decrease .connections
19:48:36  <AndreasMadsen>so when calling server.close() the server tell all saves to send a callback when all there connections are destroyed.
19:49:05  <AndreasMadsen>then the parent do a manual socket.destroy() so .connections is decreased.
19:49:14  <AndreasMadsen>and the close event will eventually emit.
19:50:40  <AndreasMadsen>server.connections is the least magical thing here. it just false or a number :D
19:51:25  <AndreasMadsen>But don't get me wrong, this stuff will be totally ridiculous to implement in userland.
19:57:53  * igorzijoined
20:00:23  <igorzi>using cluster on unix.. when the master process goes down - does it bring all slave processes down with it?
20:10:46  <isaacs>yeah, ok
20:11:13  <isaacs>so, you'll have a number of connections that artificially higher, because it's counting sockets that have been closed on the child.
20:11:14  <AndreasMadsen>igorzi: no, not yet
20:11:27  <AndreasMadsen>igorzi: next patch will fix that :)
20:11:57  <igorzi>AndreasMadsen: are you doing it in libuv? or node?
20:12:05  * mraleph1joined
20:12:27  <AndreasMadsen>igorzi: I have not made the patch, but I will property just do it in node.
20:12:52  <isaacs>AndreasMadsen: oh, also, indutny gave that other one a +1, and it looks ok to me. just squash it into one or a few commits with clear messages
20:13:01  <isaacs>AndreasMadsen: should that come before this one, or after?
20:13:10  <isaacs>or are they just different?
20:13:23  <mraleph1>mmalecki: yeah that definetely looks like a bug in OSR.
20:13:30  <mraleph1>mmalecki: nasty
20:13:46  <AndreasMadsen>isaacs: the cluster patch there kill workers should come after.
20:13:49  <igorzi>AndreasMadsen: how is it going to work? are going to send heart-beat messages between master/slaves?
20:14:20  <AndreasMadsen>igorzi: I would keep it simple and listen on process.on('exit')
20:14:28  <mmalecki>can't child do process.stdin.on('end')?
20:14:38  <AndreasMadsen>mmalecki: no
20:14:43  <mmalecki>child's stdin should be closed after parent dies
20:15:05  <AndreasMadsen>mmalecki: but in node 0.7 you can use on('disconnect')
20:15:15  <mmalecki>so use it :)
20:15:17  <AndreasMadsen>mmalecki: I tried in the immortal module
20:15:30  <AndreasMadsen>^ with process.stdin.on('close')
20:15:34  <CoverSlide>'disconnect' ?
20:15:38  <CoverSlide>for streams?
20:15:51  <AndreasMadsen>CoverSlide: when using fork IPC
20:15:54  * stephankquit (Quit: *Poof!*)
20:16:01  <CoverSlide>oh
20:16:16  <CoverSlide>it's only for fork? not for spawn / etc ?
20:16:21  <igorzi>AndreasMadsen: k thanks
20:16:22  <AndreasMadsen>http://nodejs.org/docs/v0.7.4/api/child_processes.html#event_disconnect_
20:16:30  <AndreasMadsen>CoverSlide: only fork
20:16:44  <mraleph1>mmalecki: I suppose you are using stock master branch node? what is deps/npm/cli.js?
20:17:29  <mmalecki>mraleph1: yeah, it's master. it's npm - node package manager
20:17:34  <mmalecki>it's cli interface
20:17:38  <CoverSlide>.disconnect? is that like disown?
20:18:18  <AndreasMadsen>CoverSlide: yes
20:18:23  <CoverSlide>awesome
20:18:39  <AvianFlu>isaacs, I got detached child processes working on unix and windows, btw :D
20:18:42  <isaacs>mraleph1, mmalecki: what bug are you tracking down?
20:18:48  <isaacs>AvianFlu: rad!
20:18:58  <AndreasMadsen>AvianFlu: ++
20:18:58  <kohai>AvianFlu has 77 internet points
20:18:59  <mmalecki>isaacs: node segfaulting
20:19:00  <mraleph1>mmalecki: yeah I know what npm is :-) I was just curious which version
20:19:14  <isaacs>mraleph1: it should be version 1.1.1, i believe
20:19:22  <isaacs>mraleph1: check deps/npm/package.json
20:19:33  <mmalecki>AvianFlu: oh! testing on os x, sec
20:19:34  <AvianFlu>the libuv side: https://github.com/AvianFlu/libuv/compare/spawn-detached
20:19:43  <AvianFlu>and node: https://github.com/AvianFlu/node/compare/spawn-detached
20:19:59  <AvianFlu>after I get a few general sanity checks and maciej tells me if it runs on a mac, I'll PR
20:20:00  <isaacs>mraleph1: or ./node deps/npm -v
20:20:25  <mmalecki>isaacs: given it doesn't segfault :D
20:21:07  <mraleph1>ok I will take a look tomorrow.
20:21:17  <mmalecki>`git fetch AvianFlu`, lol
20:21:23  <mmalecki>mraleph1: want ssh access?
20:22:27  <mraleph1>nuh it should be reproducible
20:22:38  <mraleph1>if not I will ask for an access
20:22:44  <mmalecki>sounds good :)
20:27:42  <mmalecki>cd
20:27:52  <mmalecki>er, derp
20:29:25  * mikealjoined
20:30:22  <mmalecki>AvianFlu: I'm getting some build errors, looking
20:30:33  <igorzi>AndreasMadsen: when do you think your patch (to kill worker processes if master dies) will land?
20:31:04  <AndreasMadsen>igorzi: I will make it right after issacs has landed the current patch.
20:31:18  <AndreasMadsen>igorzi: so you better ask him :)
20:31:48  <igorzi>isaacs: will that make it into v0.7.5?
20:32:05  <AndreasMadsen>^ https://github.com/joyent/node/pull/2740
20:32:20  <mmalecki>AvianFlu: ahahahaha, what a fail
20:32:28  <mmalecki>AvianFlu: it's called MAP_ANON on OS X
20:32:37  <mmalecki>AvianFlu: http://code.google.com/p/mupen64plus/issues/detail?id=446 :)
20:33:19  <isaacs>igorzi: no, i think it'll wait for 0.7.6
20:33:30  <isaacs>0.7.5 was forked yesterday.
20:33:34  <igorzi>isaacs: k thanks
20:33:38  <AvianFlu>mmalecki, LULZ
20:33:38  <isaacs>i'm about to push it live, actually.
20:34:16  <AvianFlu>mmalecki, easy fix at least, incoming
20:36:58  <AvianFlu>mmalecki, try it now
20:37:56  <mmalecki>building
20:38:07  * mikealquit (Quit: Leaving.)
20:39:17  <TooTallNate>WERE DOING IT LIVE!!!
20:40:54  <mmalecki>AvianFlu: WOOOOT WOOOT
20:41:04  <mmalecki>TooTallNate: doing devops :D ?
20:41:19  <AvianFlu>mmalecki, win
20:41:28  <mmalecki>AvianFlu: hell yeah win
20:41:35  * mmaleckidances around
20:42:26  * mikealjoined
20:42:51  <mmalecki>TooTallNate: you know what they say, btw? you're not in devops unless you say "we're doing it live" 3 times a day
20:43:33  <AndreasMadsen>isaacs: another assert patch -> https://github.com/joyent/node/pull/672
20:44:41  <isaacs>AndreasMadsen: thanks
20:45:08  <AndreasMadsen>isaacs: and this one -> https://github.com/joyent/node/pull/706
20:45:38  <AndreasMadsen>isaacs: but perhaps it should land if it's a problem. But somehow I don't see why it is.
20:45:40  * xaqquit (Ping timeout: 255 seconds)
20:49:08  <AndreasMadsen>since we gyp now, this could also be closed -> https://github.com/joyent/node/pull/709
20:50:12  <mmalecki>oh man, ^ is ancient
20:50:18  <mmalecki>AvianFlu: so, pull request?
20:50:41  * hij1nxjoined
20:50:45  <AvianFlu>mmalecki, I'm going to clean up the commit history some
20:50:55  <mmalecki>AvianFlu: all right :)
20:50:56  <AvianFlu>add that ifdef stuff in the first commit
20:52:11  <mmalecki>AvianFlu: I think I'll start working on forever later tomorrow :)
20:52:41  <mmalecki>it'd be super cool to have it work with node master without binary addons
20:53:15  <AvianFlu>you're getting ahead of yourself
20:53:23  <AvianFlu>but yes, ditching daemon.node will be good for the whole world
20:53:42  <mmalecki>AvianFlu: no, I'm not, I just want to actually release it as an alpha :)
20:53:50  <AndreasMadsen>mmalecki: you don't need addons to deattach processors, you can just use a JS hack.
20:54:29  <AvianFlu>AndreasMadsen, not on windows you can't
20:54:47  <AvianFlu>and it's easier this way, trust me :D
20:55:02  <AndreasMadsen>AvianFlu: okay. Did not try.
20:55:10  <mmalecki>yeah, windows is really important. we're moving our cloud to windows NT soon, so
20:55:23  <AndreasMadsen>AvianFlu: way easier, and no hacks.
20:55:39  <AndreasMadsen>mmalecki: oh, that would be so cool
20:55:40  <AvianFlu>I mean, it's this api, at present: https://gist.github.com/ce20858d1bab6d553aec
20:55:44  <mmalecki>or wait, we're going to use windows me, I think
20:56:04  <TooTallNate>mmalecki: haha, i don't think node runs there
20:56:26  <mmalecki>TooTallNate: I want to try it now :D
20:56:37  <TooTallNate>i think XP is the baseline
20:56:56  <TooTallNate>was XP based off of NT?
20:57:05  <mmalecki>TooTallNate: oh well, then we're going to virtualize XP on top of Me, I guess
20:57:54  <AndreasMadsen>AvianFlu: Immortal.start('script.js', {stategy: 'deattached'}) <- so API is not a problem :)
20:58:07  <TooTallNate>me was the start of m$ thinking their users are retarded
20:58:14  * TooTallNateclick on My Computer
20:58:28  <TooTallNate>on the left, "oh did you mean "My Documents"?"
20:58:33  <mmalecki>yeah. I remember that windows 2000 wasn't that bad!
20:58:39  <TooTallNate>no, i meant "My Computer", that why I clicked on it !
20:58:55  <TooTallNate>windows 2000 was really nice actually
20:58:57  <AndreasMadsen>night
20:59:07  <mmalecki>see you AndreasMadsen
20:59:29  <AndreasMadsen>mmalecki: yeah, I will follow you about that forever thing. Just so you know :D
20:59:49  * AndreasMadsenquit (Remote host closed the connection)
20:59:52  <mmalecki>well, ok, I guess?
21:05:07  * xaqjoined
21:06:43  <pfox___>any docs on using libuv in mingw32?
21:06:51  <pfox___>i got it to build, but linking it is proving troubling
21:07:39  <pfox___>after pointing gcc to libuv via -L and -I, it's exploding in the linker with a bunch of undefined references to what looks like windows platform calls
21:08:32  <pfox___>https://gist.github.com/1895065
21:10:38  * mikealquit (Quit: Leaving.)
21:11:57  * hij1nxquit (Ping timeout: 252 seconds)
21:25:27  * AvianFluquit (Ping timeout: 240 seconds)
21:29:47  * felixgequit (Quit: felixge)
21:30:55  * felixgejoined
21:30:55  * felixgequit (Changing host)
21:30:55  * felixgejoined
21:32:27  * sh1mmerquit (Quit: sh1mmer)
21:32:46  * AvianFlujoined
21:33:34  * felixgequit (Client Quit)
21:37:33  * mikealjoined
21:46:34  * xaqquit (Remote host closed the connection)
21:49:33  * mikealquit (Quit: Leaving.)
21:58:41  * perezdjoined
22:02:20  <isaacs>igorzi: hey, you around?
22:02:26  <isaacs>igorzi: i'm having trouble building master on windows
22:02:51  <isaacs>it worked on v0.7.4
22:04:07  <isaacs>oh, wait... i was saying "all", i guess it errors out now rather than ignoring it?
22:05:11  * xaqjoined
22:06:41  <isaacs>oh, no, my mistake. i just don't have my git shell prompts to help me :) yeah, it definitely works on 0.7.4, but not master or 0.7.5
22:21:40  <isaacs>bisect ftw!
22:23:02  * perezd_joined
22:25:16  * perezdquit (Ping timeout: 248 seconds)
22:25:16  * perezd_changed nick to perezd
22:27:18  <TooTallNate>isaacs: let me make a patch
22:27:25  <isaacs>TooTallNate: ok
22:27:46  <isaacs>i'd rather just revert for 0.7.5, but if you want to work on https://github.com/joyent/node/issues/2818 then be my guest :)
22:37:49  * dshaw_joined
22:42:44  <CIA-99>node: isaacs v0.7.5-release * r73806b5 / AUTHORS : Add new authors - http://git.io/D1tWYg
22:42:44  <CIA-99>node: isaacs v0.7.5-release * rd3b8372 / lib/querystring.js : lint, missed in QS speed merge - http://git.io/IUq6qQ
22:42:46  <CIA-99>node: isaacs v0.7.5-release * ra913022 / (configure node.gyp):
22:42:46  <CIA-99>node: Revert "build: support shared V8 properly"
22:42:46  <CIA-99>node: This reverts commit 3d1b67064085ef486a69c2fc69d195dc35f4f50c.
22:42:46  <CIA-99>node: Breaks build on windows.
22:42:46  <CIA-99>node: https://gist.github.com/1895279 - http://git.io/cG_5dw
22:42:46  <CIA-99>node: isaacs v0.7.5-release * rd384b8b / (7 files in 6 dirs): (log message trimmed)
22:42:46  <CIA-99>node: 2012.02.23, Version 0.7.5 (unstable)
22:42:47  <CIA-99>node: * startup speed improvements (Maciej Małecki)
22:42:47  <CIA-99>node: * crypto: add function getDiffieHellman() (Tomasz Buchert)
22:43:42  <CIA-99>node: * buffer: support decoding of URL-safe base64 (Ben Noordhuis)
22:43:42  <CIA-99>node: * Make QueryString.parse() even faster (Brian White)
22:43:42  <CIA-99>node: * url: decode url entities in auth section (Ben Noordhuis)
22:43:42  <CIA-99>node: isaacs master * ra913022 / (configure node.gyp):
22:43:42  <CIA-99>node: Revert "build: support shared V8 properly"
22:43:43  <CIA-99>node: This reverts commit 3d1b67064085ef486a69c2fc69d195dc35f4f50c.
22:43:43  <CIA-99>node: Breaks build on windows.
22:43:44  <CIA-99>node: https://gist.github.com/1895279 - http://git.io/cG_5dw
22:43:44  <CIA-99>node: isaacs master * rd384b8b / (7 files in 6 dirs): (log message trimmed)
22:43:44  <CIA-99>node: 2012.02.23, Version 0.7.5 (unstable)
22:43:45  <CIA-99>node: * startup speed improvements (Maciej Małecki)
22:43:46  <CIA-99>node: * crypto: add function getDiffieHellman() (Tomasz Buchert)
22:43:46  <CIA-99>node: * buffer: support decoding of URL-safe base64 (Ben Noordhuis)
22:43:48  <CIA-99>node: * Make QueryString.parse() even faster (Brian White)
22:43:48  <CIA-99>node: * url: decode url entities in auth section (Ben Noordhuis)
22:43:48  <CIA-99>node: isaacs master * rc5cbc3d / (11 files in 7 dirs): Merge branch 'v0.7.5-release' - http://git.io/mF4iMA
22:44:21  <TooTallNate>isaacs: https://github.com/joyent/node/pull/2819
22:44:28  <TooTallNate>isaacs: simple enough :p
22:45:11  <isaacs>sweet, thanks
22:45:22  <TooTallNate>np
22:45:27  <isaacs>i'll merge in one sec
22:53:21  <CIA-99>node: isaacs master * r59ecf2c / (configure node.gyp):
22:53:21  <CIA-99>node: Revert "Revert "build: support shared V8 properly""
22:53:21  <CIA-99>node: This reverts commit a9130222bd36ac2c5dd7ae560f02bb99569749a7. - http://git.io/1DeF5A
22:53:21  <CIA-99>node: Nathan Rajlich master * ra84adad / node.gyp :
22:53:22  <CIA-99>node: Default node_shared_v8 to false.
22:53:22  <CIA-99>node: Fixes #2818. - http://git.io/v94QYQ
22:53:34  <isaacs>TooTallNate: thanks!
22:54:48  * sh1mmerjoined
22:57:35  * travis-cijoined
22:57:35  <travis-ci>[travis-ci] joyent/node#488 (v0.7.5-release - d384b8b : isaacs): The build failed.
22:57:35  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/73806b5^...d384b8b
22:57:35  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/731005
22:57:35  * travis-cipart
22:58:27  * mmaleckiupdates node version on travis ci
22:58:30  * travis-cijoined
22:58:30  <travis-ci>[travis-ci] joyent/node#489 (master - f492baa : isaacs): The build is still failing.
22:58:30  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/c1fcb1d...f492baa
22:58:30  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/731008
22:58:30  * travis-cipart
22:58:36  <rphillips>does libuv export the libev headers on unix systems?
23:00:14  * igorzi_joined
23:00:23  <mmalecki>rphillips: yeah, we use ev_default_fork in daemon.node
23:01:06  <rphillips>i'm working on a project that needs libev but I'm working with gyp
23:01:20  <rphillips>may as well reuse libuv
23:01:23  * diogosnowsjoined
23:01:33  <mmalecki>rphillips: luvit?
23:01:51  <rphillips>making a monolithic build of stud
23:02:43  <mmalecki>rphillips: https://github.com/joyent/libuv/blob/master/include/uv-private/uv-unix.h#L27
23:02:59  <mmalecki>rphillips: and it's included by uv.h on unix systems, so, yeah, it does
23:03:16  <rphillips>awesome
23:03:58  <rphillips>would be cool do have a hack-day project to port stud over to pure libuv
23:04:10  <rphillips>s/do/to
23:05:06  <mmalecki>rphillips: 29.02? or is it just node?
23:06:19  <rphillips>29.02?
23:06:23  <AvianFlu>googling 'stud' was a bad idea. what is it?
23:06:41  <rphillips>https://github.com/bumptech/stud
23:06:45  <AvianFlu>cool, thanks
23:06:47  <rphillips>it's an SSL endpoint
23:06:58  <AvianFlu>oh, I've heard of this
23:06:58  <AvianFlu>right
23:07:54  <mmalecki>rphillips: isn't 29.02 some kind of a hack day?
23:08:12  <mmalecki>rphillips: (at rackspace, I mean?)
23:08:54  <rphillips>sure is
23:09:02  * travis-cijoined
23:09:02  <travis-ci>[travis-ci] joyent/node#490 (master - a84adad : Nathan Rajlich): The build is still failing.
23:09:02  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/f492baa...a84adad
23:09:02  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/731106
23:09:02  * travis-cipart
23:09:24  <rphillips>i forgot about that
23:12:32  * AvianFluquit (Quit: Leaving)
23:13:44  * brsonquit (Quit: leaving)
23:17:31  * sh1mmerquit (Quit: sh1mmer)
23:21:55  * elijah-mbpquit (Remote host closed the connection)
23:27:36  <igorzi_>isaacs: no x64 binaries for 0.7.5?
23:28:06  * sh1mmerjoined
23:28:15  * sh1mmerquit (Remote host closed the connection)
23:28:28  <isaacs>igorzi_: i didn't make them this time, no.
23:28:29  * sh1mmerjoined
23:28:35  <TooTallNate>igorzi_: did you guys figure out why it was unstable before?
23:28:39  <isaacs>igorzi_: wasn't there some v8 issue that needed to get fixed?
23:28:45  <isaacs>or is that resolved now?
23:28:51  <igorzi_>isaacs: that was in v0.6
23:28:57  <isaacs>ohh!
23:29:03  <isaacs>ok, hold on, lemme add it
23:29:11  <igorzi_>isaacs: afaik master didn't have that problem
23:29:21  <igorzi_>isaacs: let me test it on master, to make sure
23:29:43  <isaacs>ok
23:35:41  * perezd_joined
23:35:52  * mjr_joined
23:36:19  * perezdquit (Ping timeout: 252 seconds)
23:36:19  * perezd_changed nick to perezd
23:41:55  * perezdquit (Ping timeout: 252 seconds)
23:42:10  * perezdjoined
23:44:10  * perezd_joined
23:46:30  * perezdquit (Ping timeout: 240 seconds)
23:47:26  * perezdjoined
23:47:38  * dshaw_quit (Quit: Leaving.)
23:49:17  * perezd_quit (Ping timeout: 244 seconds)
23:50:32  <isaacs>arrrggg! forgot to do the build in the x64 shell.
23:51:02  <TooTallNate>does that just build 32-bit again?
23:53:08  * piscisaureus_joined
23:59:29  <CIA-99>libuv: Brandon Philips v0.6 * rfca18c3 / src/win/fs.c :
23:59:29  <CIA-99>libuv: win: fs: handle EOF in read
23:59:29  <CIA-99>libuv: in luvit after upgrade libuv from 243cfc to d3efef readSync started
23:59:29  <CIA-99>libuv: failing. It seems that the code cleanup stopped handling EOF
23:59:29  <CIA-99>libuv: Trivially reproduced with this
23:59:30  <CIA-99>libuv: local fs = require('fs')
23:59:30  <CIA-99>libuv: print(fs.readFileSync('foo.luvit')) - http://git.io/9aDiDQ