00:02:02  * AvianFlujoined
00:09:52  * AvianFluquit (Ping timeout: 260 seconds)
00:13:06  * AvianFlujoined
00:13:11  * AvianFluquit (Read error: Connection reset by peer)
00:14:14  * mralephjoined
00:15:27  <igorzi>piscisaureus_: hey yt?
00:15:43  <piscisaureus_>yup
00:15:48  <piscisaureus_>I'm doing named shared memory
00:16:16  <piscisaureus_>igorzi: ^
00:16:26  <igorzi>piscisaureus_: cool
00:16:58  <igorzi>piscisaureus_: i'm trying to get v0.4.12 built with cygwin, and for the life of me i can't get passed ./configure
00:16:58  <piscisaureus_>igorzi: but what's up?
00:17:06  <piscisaureus_>heh
00:17:18  <igorzi>i keep getting "error: could not configure a cxx compiler!"
00:17:24  <igorzi>and i already did rebaseall
00:17:25  <piscisaureus_>igorzi: what does it say?
00:17:30  <piscisaureus_>hmm
00:17:35  <piscisaureus_>did you install the cxx compiler?
00:18:23  <piscisaureus_>igorzi: open cygwin's setup and check mark g++ for installation
00:18:57  <igorzi>yeah, i can do "gcc --help" or "g++ --help"
00:19:08  <igorzi>(in cygwin cmd prompt)
00:19:18  * AvianFlujoined
00:19:47  <piscisaureus_>igorzi then try `CC=gcc` and `CXX=g++` before running configure
00:19:58  <piscisaureus_>igorzi: otherwise we could screen share
00:20:41  <piscisaureus_>I hope cygwin still compiles
00:21:38  * isaacsquit (Read error: Connection reset by peer)
00:21:56  * isaacsjoined
00:22:07  <igorzi>piscisaureus_: i tried it.. doesn't work. it's actually on ndjs-win4, you could RDP directly to it.
00:26:40  * AvianFluquit (Ping timeout: 260 seconds)
00:27:09  <piscisaureus_>igorzi: I think we needed gcc4
00:27:15  <piscisaureus_>trying that now
00:27:30  <igorzi>piscisaureus_: thx! sorry to randomize you with this
00:28:17  <piscisaureus_>meh, no luck
00:29:01  <igorzi>ryah: the cluster.js docs currently give an example to use signals to detect when a child process dies and spin up a new one. is that how v0.6 will release?
00:32:28  * ericktjoined
00:40:04  <ryah>igorzi: nope - need to fix
00:40:11  * AvianFlujoined
00:41:01  <ryah>export CXX=`which g++`
00:41:09  <piscisaureus_>igorzi: http://boxysystems.com/index.php/step-by-step-instructions-to-install-nodejs-on-windows/comment-page-1/#comment-8682
00:41:12  <piscisaureus_>OMG
00:41:35  <piscisaureus_>people want node so badly, they actually try to figure this shit out
00:41:50  <igorzi>damn!
00:42:17  <igorzi>ryah: will the cluster api also allow killing worker processes?
00:42:40  <ryah>igorzi: yes- there's going to be some last minute massive changes there :)
00:43:05  <igorzi>ryah: ok thanks
00:43:07  <ryah>workers should be worker.kill()
00:43:12  <ryah>er
00:43:18  <ryah>killing should already work, is what i meant
00:43:46  <ryah>but we need an event when they die
00:45:42  <igorzi>ryah: cluster.startMaster doesn't return workers now, right?
00:48:42  <igorzi>ryah: or you could just do startMaster with workers=0, and then use spawnWorker to start workers.. is that the idea?
00:51:52  <ryah>igorzi: maybe i'll make the "workers" array public
00:55:27  <igorzi>ryah: that would be cool
00:55:46  <igorzi>piscisaureus_: are you still trying to get cygwin to work on ndjs-win4?
00:55:51  <piscisaureus_>yeah
01:00:45  * bnoordhuisjoined
01:02:05  * mralephquit (Quit: Leaving.)
01:02:12  <bnoordhuis>yay, internet again
01:02:35  * dapquit (Quit: Leaving.)
01:02:56  <bnoordhuis>igorzi: i think you're about to head home?
01:04:25  <ryah>bnoordhuis: piscisaureus_ is working on compiling cygwin
01:04:34  <ryah>bnoordhuis: apparently it's non-trivial
01:05:01  <piscisaureus_>yes it fucks up
01:05:05  <bnoordhuis>i don't doubt that for a second
01:06:36  <igorzi>bnoordhuis: yeah, in about 40 min
01:06:38  <ryah>do we have a non-cygwin build for v0.4.12?
01:06:56  <piscisaureus_>wut? a mingw build?
01:07:01  <ryah>yes
01:07:11  <piscisaureus_>that's much less likely to work
01:07:16  <ryah>select + sockets?
01:07:18  <ryah>huh.
01:07:28  <piscisaureus_>well I can try
01:07:37  <ryah>well im glad the work we did for v0.6 is apparent :)
01:07:46  <bnoordhuis>igorzi: no worries, i wanted to do the benchmarks but i'll wait until piscisaureus_ is finished
01:07:51  <ryah>v0.4 = doesn't work. v0.6 = works
01:08:07  <piscisaureus_>heh
01:08:24  <piscisaureus_>cygwin is wonderful
01:08:44  <ryah>tell me more
01:09:00  <piscisaureus_>There's no software with more failure modes
01:09:14  <bnoordhuis>[16:21|% 100|+ 613|- 15]: Done <- sunos
01:09:17  <ryah>oo, sounds enticing
01:09:40  <ryah>i like my software with many failure modes
01:14:44  <piscisaureus_>ANGER &^*$%^$*%(
01:15:24  <igorzi>piscisaureus_: do you have any machines with cygwin already installed, that worked with node at some point?
01:15:32  <piscisaureus_>yeah my home machine
01:15:35  <piscisaureus_>but I am not at home
01:15:37  <piscisaureus_>:-(
01:15:54  <piscisaureus_>the nodejs buildbot
01:16:06  <piscisaureus_>I wonder what that looks like these days
01:16:25  <ryah>i think it got reaped
01:16:38  <piscisaureus_>oh nooos
01:16:53  <bnoordhuis>ryah: test/simple/test-net-pipe-connect-errors.js passes for you now?
01:17:14  <bnoordhuis>because it doesn't on linux, not with master and not with 0.4
01:17:24  * brsonquit (Quit: leaving)
01:17:46  <bnoordhuis>AssertionError: "ECONNREFUSED" == "ENOTSOCK"
01:19:56  <ryah>bnoordhuis: yep
01:20:10  <ryah>i tested both
01:20:27  <bnoordhuis>different operating systems are a wonderful thing
01:20:31  <ryah>i guess linux acts different than osx
01:20:32  <bnoordhuis>i'll check it out
01:20:55  <ryah>it would be lovely to get a test in libuv for it
01:20:56  <piscisaureus_>any python experts here?
01:21:08  <ryah>piscisaureus_: not expert... but i can generally use it
01:21:31  <bnoordhuis>piscisaureus_: i call guido my home boy so speak to me
01:21:39  <piscisaureus_>what is the python equivalent of
01:21:40  <piscisaureus_>node -e "require('child_process).exec('fubar')"
01:21:55  <bnoordhuis>os.spawn()
01:22:26  <bnoordhuis>though the proper thing is to use the subprocess module
01:22:41  <piscisaureus_>bnoordhuis: I just want to run a process and see the output
01:22:52  <piscisaureus_>preferrably from the repl
01:23:14  * isaacsquit (Quit: isaacs)
01:24:22  <bnoordhuis>piscisaureus_: import subprocess; subprocess.call('ls')
01:25:13  <bnoordhuis>piscisaureus_: import subprocess; subprocess.call(['ls', '-hl']) # if you want to pass args
01:27:17  * AvianFlu_joined
01:29:31  <piscisaureus_>aaah
01:29:35  <piscisaureus_>cygmpfr-4.dll <-- not found
01:29:39  * AvianFluquit (Ping timeout: 260 seconds)
01:31:49  <piscisaureus_>igorzi: it's rolling
01:32:08  <igorzi>piscisaureus_: ahh thanks!
01:32:28  <igorzi>piscisaureus_: i can take over
01:32:47  <piscisaureus_>igorzi: apparently the current cygwin distribution has an error, the wrong version of libmpfr (or something) gets installed
01:34:05  * avian__joined
01:34:46  <igorzi>piscisaureus_: can i bump you off that machine?
01:34:55  <piscisaureus_>sure, go ahead
01:35:02  <piscisaureus_>it's still building though
01:35:19  <igorzi>thanks again!
01:35:28  <piscisaureus_>nvm
01:36:49  * AvianFlu_quit (Ping timeout: 252 seconds)
01:38:04  * isaacsjoined
01:40:29  * isaacsquit (Client Quit)
01:40:51  * AvianFlu_joined
01:43:11  * avian__quit (Ping timeout: 260 seconds)
01:48:30  * AvianFlu_quit (Ping timeout: 260 seconds)
01:54:25  * ericktquit (Ping timeout: 252 seconds)
01:58:41  <bnoordhuis>[14:49|% 100|+ 623|- 7]: Done <- linux
01:58:56  <bnoordhuis>Command: out/Release/node /home/bnoordhuis/src/nodejs/node/test/simple/test-tcp-wrap-listen.js
01:58:56  <bnoordhuis>--- CRASHED ---
01:59:27  <bnoordhuis>spurious crash though, can't reproduce it
02:00:44  <bnoordhuis>i take that back
02:00:48  <bnoordhuis>$ out/Release/node /home/bnoordhuis/src/nodejs/node/test/simple/test-tcp-wrap-listen.js > /dev/null
02:00:48  <bnoordhuis>Segmentation fault (core dumped)
02:09:06  * isaacsjoined
02:14:50  <piscisaureus_>Allright-y, named shared memory works
02:14:53  <piscisaureus_>now off to bed
02:16:48  * isaacsquit (Quit: isaacs)
02:34:23  * piscisaureus_quit (Ping timeout: 248 seconds)
02:38:35  * isaacsjoined
02:57:40  * piscisaureus_joined
02:59:49  * isaacsquit (Quit: isaacs)
03:14:01  <bnoordhuis>https://github.com/joyent/node/blob/cf78d04/src/stream_wrap.cc#L155
03:14:24  <bnoordhuis>before the call to global->SetHiddenValue(), b->handle_.IsEmpty() is false
03:14:33  <bnoordhuis>afterwards it's true
03:14:43  <bnoordhuis>and only in release builds
03:14:46  <bnoordhuis>and only sometimes...
03:14:49  <bnoordhuis>another v8 bug?
03:15:13  <bnoordhuis>ryah: ^ suggestions
03:27:51  <bnoordhuis>nm, found it - fucking build system
03:31:52  * isaacsjoined
03:51:59  * isaacsquit (Quit: isaacs)
03:55:59  <bnoordhuis>[18:38|% 100|+ 621|- 9]: Done <- linux, ffs
03:56:57  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
04:11:25  * mikealjoined
04:16:37  <indutny>piscisaureus_: morning
04:16:58  <bnoordhuis>indutny: he's probably asleep - he should be!
04:17:13  <indutny>bnoordhuis: ok!
04:17:15  <indutny>thank you
04:32:57  <CIA-95>node: Ben Noordhuis master * r8974ba3 / test/simple/test-net-pipe-connect-errors.js :
04:32:57  <CIA-95>node: test: fix simple/net-pipe-connect-errors
04:32:57  <CIA-95>node: When trying to connect to something that is not a UNIX socket, Linux returns
04:32:57  <CIA-95>node: ECONNREFUSED, not ENOTSOCK.
04:32:57  <CIA-95>node: We cannot atomically determine if the other end is a) a stale socket, or b) not
04:32:57  <CIA-95>node: a socket at all, so let's accept both error codes. - http://git.io/CBOhqA
04:39:37  <CIA-95>libuv: saghul master * r677bb70 / src/uv-common.c : Added missing error codes to uv_err_name - https://github.com/joyent/libuv/commit/677bb70bfb6378a68da8438c62b1061b2bea57ab
04:55:08  * igorzi_joined
05:13:41  * bnoordhuisquit (Ping timeout: 240 seconds)
05:23:21  * ericktjoined
05:55:44  * felixgejoined
05:55:44  * felixgequit (Changing host)
05:55:44  * felixgejoined
06:00:52  * felixgequit (Quit: felixge)
06:13:46  * ericktquit (Ping timeout: 260 seconds)
06:26:50  * ericktjoined
06:29:29  * igorzi_quit (Quit: Page closed)
07:13:06  * ericktquit (Ping timeout: 260 seconds)
07:22:07  * AvianFlujoined
07:29:49  * ryahtopic: v0.6.0 TODAY
07:45:31  * felixgejoined
07:45:33  * felixgequit (Changing host)
07:45:33  * felixgejoined
08:11:48  <indutny>wow
08:12:00  <indutny>wooot
10:03:08  * mafintoshjoined
10:19:58  * bnoordhuisjoined
10:20:11  <bnoordhuis>piscisaureus: yo homez, u there?
10:25:21  <indutny>:)
11:02:17  * felixgequit (Quit: felixge)
11:36:50  * mafintoshquit (Quit: Linkinus - http://linkinus.com)
11:38:45  * felixgejoined
11:38:45  * felixgequit (Changing host)
11:38:45  * felixgejoined
11:41:06  * felixgequit (Client Quit)
11:41:20  * felixgejoined
11:44:01  * felixgequit (Client Quit)
11:44:24  * felixgejoined
11:44:24  * felixgequit (Changing host)
11:44:24  * felixgejoined
11:46:21  * felixgequit (Client Quit)
11:46:28  * felixgejoined
11:46:28  * felixgequit (Changing host)
11:46:28  * felixgejoined
11:47:03  * felixgequit (Client Quit)
12:10:52  <CIA-95>node: Ben Noordhuis master * r5fd012e / lib/fs.js : fs: don't use octal values, not allowed in strict mode - http://git.io/E-4F7A
12:27:37  * piscisaureus_joined
12:37:14  <CIA-95>node: Jordan Sissel master * r358f0ce / (lib/url.js test/simple/test-url.js):
12:37:14  <CIA-95>node: url: add '.' '+' and '-' in url protocol
12:37:14  <CIA-95>node: - Based on BNF from RFC 1738 section 5.
12:37:14  <CIA-95>node: - Added tests to cover svn+ssh and some other examples - http://git.io/q8iPxQ
12:59:00  <piscisaureus_>bnoordhuis: momentje
12:59:20  <bnoordhuis>piscisaureus_: vooruit dan
13:00:08  <bnoordhuis>leuk dat je me belt om te vertellen dat je me zo terug belt :)
13:00:23  <piscisaureus_>bnoordhuis: om een of andere manier pakt de macbook mijn headphones niet
13:00:35  <piscisaureus_>anders dan wordt iedereen op kantoor boos
13:00:40  <bnoordhuis>ah ja, mac he? it just works(tm)
13:19:58  <indutny>piscisaureus_: yt?
13:20:04  <piscisaureus_>indutny: yep
13:20:15  <indutny>piscisaureus_: debugger is continuing because of some bugs in v8
13:20:19  <piscisaureus_>indutny: I had 2 questions yesterday
13:20:27  <indutny>piscisaureus_: and I don't understand second question :)
13:20:42  <indutny>piscisaureus_: ah
13:20:43  <indutny>got it
13:20:53  <indutny>do you mean node debugger's command?
13:22:13  <indutny>piscisaureus_: I think we may add it, but v8's debugger protocol doesn't support this
13:22:39  <indutny>piscisaureus_: so I can only send evaluate request for "debugger" command
13:22:46  <indutny>quite hacky, but should work
13:22:50  <indutny>piscisaureus_: are you ok with that?
13:23:14  <piscisaureus_>indutny: somehow the node debugger issues a continue statement right after attaching the debugger (with `node debug -p pid`)
13:23:26  <piscisaureus_>indutny: why is that. I'd be happy if it just breaks
13:23:30  <piscisaureus_>indutny: the other thing
13:23:36  <piscisaureus_>indutny: there is no command to pause the debuggee
13:23:47  <indutny>piscisaureus_: yep, that's what I was talking about
13:23:47  <piscisaureus_>indutny: while that would be easy, just send SIGUSR1 there
13:23:52  <indutny>piscisaureus_: eval "debugger"
13:24:04  <indutny>I'll add `pause` method
13:24:09  <piscisaureus_>yeah
13:24:25  <indutny>piscisaureus_: about continue - how is that possible?
13:24:30  <piscisaureus_>indutny: I don't know
13:24:36  <indutny>piscisaureus_: are you sure that it's node debugger's behaviour, not v8 bug?
13:24:49  <piscisaureus_>indutny: I am not 100% sure
13:25:07  <piscisaureus_>but how does an actual `continue` command get queued up otherwise?
13:25:32  <indutny>piscisaureus_: ah, you debugged v8, right?
13:25:37  <indutny>ok, I'll investigate that
13:25:44  <indutny>but that's odd
13:25:46  <piscisaureus_>cool
13:26:03  <indutny>aaah
13:26:06  <indutny>I think I got it
13:26:17  <indutny>we're calling continue
13:26:18  <piscisaureus_>indutny: there will be another binding for starting a remote debugger btw
13:26:20  <indutny>right after connection
13:26:24  <piscisaureus_>indutny: exactly. why?
13:26:36  <indutny>piscisaureus_: investigating...
13:26:39  <piscisaureus_>indutny: the new binding will be process._debugProcess(pid)
13:27:04  <indutny>ok
13:27:04  <piscisaureus_>indutny: because we're stepping off of using console control handlers
13:27:20  <indutny>I think we can remove that continue line
13:27:23  <indutny>let me test it
13:28:43  <indutny>brb, going to take something in store
13:30:08  <indutny>piscisaureus_: looks like it works fine w/o it, btw
13:30:18  <indutny>but I'll need to fix tests and check everything
13:30:56  <piscisaureus_>indutny: kewl
13:52:27  * igorziquit (Ping timeout: 265 seconds)
14:15:48  * felixgejoined
14:15:49  * felixgequit (Changing host)
14:15:49  * felixgejoined
14:45:09  <wankdanker>bnoordhuis: regarding our conversation yesterday about some of the dgram stuff, I have posted a patch that enables setBroadcast (https://gist.github.com/1339461)
14:45:43  <wankdanker>bnoordhuis: i don't plan on forking node and libuv so, i won't be issuing a pull request, but if you want to use this, help yourself.
14:49:43  <piscisaureus_>indutny: hey
15:09:08  <piscisaureus_>indutny: which line do i have to remove to disable that continue statement?
15:13:33  * isaacsjoined
15:29:18  <CIA-95>node: Bert Belder reviewme * r189dd8f / (10 files in 2 dirs): Fix line endings and trailing whitespace - http://git.io/kBbdNg
15:29:18  <CIA-95>node: Bert Belder reviewme * r6ee73a2 / (src/node.cc src/node.h): Windows: ErrnoException shouldn't mix up crt and win32 errors - http://git.io/43Upmw
15:29:18  <CIA-95>node: Bert Belder reviewme * r35f4182 / src/node.cc : Make error reporting from node::DLOpen more consistent - http://git.io/bwtT-g
15:29:19  <CIA-95>node: Bert Belder reviewme * rdd5f4fc / (3 files in 3 dirs): process._debugProcess - http://git.io/Ucn94g
15:29:29  <piscisaureus_>bnoordhuis: igorzi: please review : ^
15:33:33  * ericktjoined
15:43:29  <bnoordhuis>piscisaureus_: reviewed
15:48:25  * dapjoined
15:55:48  * ericktquit (Ping timeout: 245 seconds)
15:57:05  * bnoordhuisis back in an hour or two
16:00:26  <indutny>piscisaureus_: sorry, I was afk
16:00:38  <indutny>piscisaureus_: do you still need it?
16:01:17  * bnoordhuisquit (Ping timeout: 240 seconds)
16:14:41  <ryah>piscisaureus_++
16:21:42  <indutny>ryah++
16:22:47  <indutny>ryah: https://github.com/joyent/node/pull/2011
16:23:05  <indutny>as bert pointed out, debugger was requesting `continue` right after connection initialization
16:23:13  <indutny>fixed that and updated test
16:24:12  <indutny>piscisaureus_++ !
16:24:15  <indutny>looks good
16:24:39  <piscisaureus_>indutny: does the debugger still work on unix?
16:24:45  <indutny>piscisaureus_: em.. yep
16:24:56  <indutny>piscisaureus_: please test it on windows, I'm on mac right now
16:24:57  <piscisaureus_>with my patch?
16:25:01  <indutny>piscisaureus_: ah
16:25:03  <indutny>piscisaureus_: dunno
16:25:04  <indutny>:D
16:25:34  <indutny>piscisaureus_: running make
16:26:01  <indutny>piscisaureus_: https://gist.github.com/c247f00ee0282a64f9fe
16:26:13  <indutny>not really
16:30:03  <piscisaureus_>indutny++
16:30:13  <piscisaureus_>with yout patch it works beautifully
16:31:11  <piscisaureus_>indutny: what if you change the signature of RegisterSignalHandler to return void?
16:32:17  <indutny>../src/node.cc: In function ‘void node::RegisterSignalHandler(int, void (*)(int))’:
16:32:17  <indutny>../src/node.cc:2362: error: return-statement with a value, in function returning 'void'
16:32:27  <indutny>ok, it works
16:32:28  <piscisaureus_>indutny: yeah I know
16:32:31  <indutny>if I'll remove return
16:33:46  <indutny>https://gist.github.com/24a322f56bb2ebccc621
16:33:48  <indutny>piscisaureus_: ^
16:34:12  <indutny>oh
16:34:15  <indutny>sorry, my fault
16:35:54  <indutny>piscisaureus_: wait a minute
16:39:28  <indutny>piscisaureus_: https://github.com/indutny/node/commit/df83b471c4e986b9cc8648b913e92e03f79b0558
16:48:25  <indutny>so is debugger working properly, except that stupid failing step out?
16:49:41  <ryah>indutny: so pull #2011 fixes the failing test?
16:49:48  <indutny>ryah: nope
16:50:05  <piscisaureus_>ryah: which test?
16:50:11  <indutny>piscisaureus_: repl test
16:50:23  <indutny>piscisaureus_: it sometimes steps out too far
16:50:35  <piscisaureus_>it's prolly a v8 bug rtight?
16:50:41  <indutny>probably
16:52:38  <indutny>aaah
16:52:43  <indutny>looks like v8 is fine
16:53:17  <ryah>test/simple/test-debugger-repl.js
16:53:25  <ryah>times out occasionally
16:53:37  <ryah>(very often actually)
16:54:24  <ryah>was any progress made with the benchmarks?
16:54:26  <piscisaureus_>indutny: I squashed your second fix
16:55:01  <CIA-95>node: Bert Belder master * r189dd8f / (10 files in 2 dirs): Fix line endings and trailing whitespace - https://github.com/joyent/node/commit/189dd8f8034213d476e79845ad43563eb934fcf9
16:55:02  <CIA-95>node: Bert Belder master * r6ee73a2 / (src/node.cc src/node.h): Windows: ErrnoException shouldn't mix up crt and win32 errors - https://github.com/joyent/node/commit/6ee73a2fee78eeda8eb5d171dbf0b9d13ce54119
16:55:02  <CIA-95>node: Bert Belder master * r35f4182 / src/node.cc : Make error reporting from node::DLOpen more consistent - https://github.com/joyent/node/commit/35f4182cee079e2c481dc436f5ea97966957def8
16:55:02  <CIA-95>node: Bert Belder master * r829735e / (3 files in 3 dirs): Implement process._debugProcess - http://git.io/hNZPfw
16:55:07  <CIA-95>node: Fedor Indutny master * rb5d32d4 / (3 files in 3 dirs):
16:55:07  <CIA-95>node: debugger: do not request `continue` on connection
16:55:07  <CIA-95>node: * Updated test
16:55:07  <CIA-95>node: * Use `node debug file`, not `node debug -e "..."` in test - http://git.io/t4Ayrw
16:55:22  <indutny>piscisaureus_: cool
16:57:37  <indutny>ryah: https://github.com/joyent/node/pull/2012
16:57:41  <indutny>should work fine now
16:57:56  <indutny>I think, sometimes input was merged
16:58:48  <indutny>small delay should be enough to ensure that everything is received separately
16:58:52  <ryah>indutny: have piscisaureus_ mrge it
16:58:56  <indutny>ok
16:58:57  <indutny>piscisaureus_: https://github.com/joyent/node/pull/2012
16:58:58  <indutny>:)
17:00:47  <CIA-95>node: Fedor Indutny master * r8ef2c13 / test/simple/test-debugger-repl.js : fix debugger-repl test - http://git.io/fnyTKQ
17:00:51  <indutny>great
17:00:59  <indutny>ok, time to sleep
17:01:04  <indutny>thank you all, guys
17:01:54  <ryah>indutny: night
17:02:23  <piscisaureus_>ryah: I really want this in for .6 : https://github.com/joyent/libuv/issues/234
17:02:31  <piscisaureus_>but I am distracted now ...
17:02:55  <ryah>piscisaureus_: ok
17:03:11  <ryah>we'll get it in
17:03:32  <ryah>im doing a last minute rewrite of the cluster api
17:05:33  <piscisaureus_>omg
17:05:35  <piscisaureus_>why not
17:06:00  <wankdanker>let me know if you guys need any help testing anything to get 0.6 out.
17:08:01  <ryah>wankdanker: sure if you can do "./configure --debug && make test-all" and gist the results for various platforms
17:10:47  <wankdanker>ryah: k
17:11:19  * AvianFluquit (Quit: Leaving)
17:15:49  <piscisaureus_>[02:59|% 100|+ 300|- 15]: Done
17:17:50  <ryah>piscisaureus_: osx?
17:17:58  <ryah>or win?
17:18:02  <piscisaureus_>win
17:18:06  <ryah>nice
17:18:10  <ryah>-15 isnt so bad
17:18:17  <ryah>but i thought igor fixed a few tests yesterday
17:18:49  <piscisaureus_>tls-peer-certificate still fails for me
17:18:57  <piscisaureus_>ryah: did you revert ben's patch for that?
17:18:57  <ryah>yeah, i thought we fixed that one
17:19:04  <ryah>no..
17:19:24  <ryah>maybe you can bisect it?
17:19:58  <piscisaureus_>I can't really
17:20:16  <ryah>im pretty sure it worked in v0.5.10
17:20:20  <piscisaureus_>yeh
17:21:17  <piscisaureus_>the v8 upgrades fuck up bisecting
17:21:24  <piscisaureus_>so slow
17:24:40  <isaacs>anyone available to review this? https://github.com/isaacs/node/commit/b79d6af731abea45b593076eb8695ad0f093a028
17:24:56  <isaacs>it lgtm, but i touched the code, so another pair of eyes would be a plus.
17:25:32  <ryah>indutny: sure - one sec
17:25:54  <ryah>er isaacs
17:25:58  <isaacs>:)
17:26:09  <ryah>isaacs: did this guy sign the cla?
17:26:13  <isaacs>yeah
17:26:15  <isaacs>just checked
17:26:18  * igorzijoined
17:26:47  <ryah>[16:07|% 100|+ 625|- 5]: Done
17:26:50  <ryah>^-- test-all, osx
17:27:00  <ryah>16minutes -_-
17:27:43  <ryah>isaacs: it'd be nice if the first line of the commit was 50 chars.
17:28:14  <ryah>isaacs: i dont think anyone cares to know if it's a squash - rather what the patch is about
17:28:20  <ryah>isaacs: otherwise LGTM
17:28:34  <isaacs>ryah: ?
17:28:35  <isaacs>> "Close #1930 Convenience methods for zlib".length
17:28:36  <isaacs>40
17:28:45  <isaacs>that's the first line
17:29:03  <isaacs>ryah: oh, you're getting confused by the github redesign ;P
17:29:16  <isaacs>that threw me a bunch of times, too
17:29:39  <isaacs>k, i'll land it.
17:30:15  <ryah>oh
17:30:36  <ryah>isaacs: made two comments about code style too
17:30:42  <isaacs>ok
17:31:08  * ericktjoined
17:31:24  <isaacs>ah, missed that, good catch. i'll fix.
17:33:47  <CIA-95>node: Matt Robenolt master * r5213c39 / (3 files in 3 dirs): (log message trimmed)
17:33:47  <CIA-95>node: Close #1930 Convenience methods for zlib
17:33:47  <CIA-95>node: This is a combination of 20 commits. Their commit messages are preserved
17:33:47  <CIA-95>node: below for the benefit of future generations.
17:33:47  <CIA-95>node: * Adding a shortcut to easily compress/decompress a string of text.
17:33:47  <CIA-95>node: * Making the API consistent. unzip should accept a Buffer for input as well.
17:33:48  <CIA-95>node: * Adding docs.
17:35:21  <ryah>cool
17:35:26  <igorzi>piscisaureus: are you working on this? https://github.com/joyent/libuv/pull/229
17:38:35  * bnoordhuisjoined
17:39:36  <ryah>bnoordhuis: did you get my email about ENOTSOCK?
17:39:47  <bnoordhuis>ryah: yes
17:41:33  <ryah>igorzi, bnoordhuis: which of you will run the linux numbers on the cluster?
17:41:43  <bnoordhuis>ryah: i will, i just need access
17:42:44  * brsonjoined
17:44:00  <igorzi>bnoordhuis: rebooting into linux now.. i'll let you know once it's ready
17:44:41  <piscisaureus_>ryah: 0e8a55d2a22b88dc3b9b0165f344602b0fa8c977 breaks tls-peer-certificate
17:44:52  <bnoordhuis>igorzi: cool, thanks
17:45:25  <igorzi>bnoordhuis: btw, you're not planning to install your private linux kernel on that machine, right?
17:45:34  <bnoordhuis>igorzi: heh, no
17:45:44  <ryah>https://github.com/joyent/node/commit/0e8a55d2a22b88dc3b9b0165f344602b0fa8c977
17:46:28  <ryah>sneaky dutch.
17:46:42  <ryah>once i gave ben access to my machine and he put his kernel on it too
17:47:18  <igorzi>sneaky indeed :)
17:48:42  <bnoordhuis>do we have the call in 10? because i think my SO is fixing up dinner
17:49:19  <ryah>bnoordhuis: let's do it now quickly
17:49:41  <ryah>igorzi, piscisaureus_: is that okay for you?
17:49:45  <igorzi>yep
17:49:50  <bnoordhuis>cool
17:50:14  <igorzi>bnoordhuis: the server is booted into linux now
17:50:20  <bnoordhuis>igorzi: sweet, thanks
17:51:11  <bnoordhuis>piscisaureus_: wth are you?
17:51:31  <ryah>piscisaureus_: can you go into an office or something?
17:51:39  <ryah>piscisaureus_: we could hear you but it was very noisy
17:53:17  <ryah>piscisaureus_: ...
17:53:26  <piscisaureus_>ryah: add me again
18:05:07  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
18:18:49  * piscisaureus_joined
18:18:59  * AvianFlujoined
18:22:16  <ryah>ryan@mac1234:~/projects/node% ~/attic/v0.2.6/bin/node benchmark/startup.js
18:22:16  <ryah>Started node 100 times in 4558 ms. 45.58 ms / start.
18:22:16  <ryah>ryan@mac1234:~/projects/node% ~/attic/v0.4.12/bin/node benchmark/startup.js
18:22:16  <ryah>Started node 100 times in 6925 ms. 69.25 ms / start.
18:22:16  <ryah>ryan@mac1234:~/projects/node% ~/attic/v0.5.10/bin/node benchmark/startup.js
18:22:18  <ryah>Started node 100 times in 4868 ms. 48.68 ms / start.
18:26:40  <isaacs>so, we're almost back to 0.2.6 startup time?
18:26:54  <ryah>it seems that way
18:28:06  <isaacs>nice.
19:03:38  <bnoordhuis>i don't understand this obsession with startup times...
19:04:01  <bnoordhuis>igorzi: are you done?
19:06:53  <igorzi>bnoordhuis: almost.. another 5 min
19:07:13  <bnoordhuis>okay, no rush
19:11:10  <CIA-95>node: Jacob H.C. Kragh master * r44a5452 / doc/api/fs.markdown : docs: fix typo in the documented fchown signature - http://git.io/rTF0Dw
19:19:35  * mikealquit (Quit: Leaving.)
19:21:37  <igorzi>bnoordhuis: it's booted into linux now
19:22:02  * isaacsquit (Quit: isaacs)
19:22:10  <bnoordhuis>igorzi: how do i get to the other machine? they have the same external address
19:22:27  <igorzi>bnoordhuis: i'll send you an email
19:22:44  <bnoordhuis>igorzi: okay, thanks
19:34:11  * AvianFlu_joined
19:34:52  * AvianFluquit (Ping timeout: 255 seconds)
19:35:41  * piscisaureus_quit (Ping timeout: 240 seconds)
19:42:12  <igorzi>piscisaureus: yt? can you do a quick review?
19:43:05  <igorzi>piscisaureus: https://gist.github.com/1340298
19:45:05  * AvianFlu_quit (Ping timeout: 252 seconds)
19:45:56  * AvianFlu_joined
19:50:29  <bnoordhuis>on a side note: https://github.com/bnoordhuis/linux/compare/epoll_ctlv
19:50:49  <bnoordhuis>>100% performance increase on a 'many active fds' micro benchmark!
19:51:44  <pquerna>wow, didn't know about ctlv
19:51:58  <bnoordhuis>pquerna: that's because i just wrote it :)
19:52:29  <pquerna>hah
19:52:39  <pquerna>just noticed this is a fork of linux :)
19:52:59  <bnoordhuis>expect it in your distro's kernel in four or five years
19:57:21  * AvianFlu_changed nick to AvianFlu
20:01:35  * AvianFlu_joined
20:03:02  * AvianFluquit (Disconnected by services)
20:03:04  * AvianFlu_changed nick to AvianFlu
20:07:21  * isaacsjoined
20:40:04  <indutny>bnoordhuis: yt?
20:40:32  <bnoordhuis>indutny: yep
20:42:07  * AvianFluquit (Read error: Connection reset by peer)
20:42:37  <indutny>ah
20:42:43  <indutny>just woke up
20:43:01  <indutny>wasn't able to figure out why, after rebase, #2011 was equal to master :P
20:43:01  <bnoordhuis>you or me?
20:43:10  <indutny>https://github.com/joyent/node/pull/2011
20:43:15  <indutny>me
20:43:24  <indutny>rebased it
20:43:40  <indutny>bnoordhuis: ok, time to get back to bed
20:43:42  <indutny>night
20:43:46  <indutny>or morning
20:43:47  <indutny>:)
20:43:48  <bnoordhuis>heh, sleep tight
20:44:49  * AvianFlujoined
20:47:04  * mikealjoined
21:10:47  <ryah>http://code.google.com/p/dart/source/browse/branches/bleeding_edge/dart/runtime/bin/eventhandler.h
21:12:26  <ryah>https://gist.github.com/1339957
21:13:23  <ryah>so cygwin -> iocp = 2x
21:13:28  <ryah>somewhat disappointing
21:25:37  * piscisaureus_joined
21:27:49  * AvianFluquit (Ping timeout: 255 seconds)
21:28:11  * AvianFlu_joined
21:30:24  * AvianFlu_changed nick to AvianFlu
21:33:23  <piscisaureus_>somewhat disappointing...
21:33:27  <piscisaureus_>yeah
21:33:48  <piscisaureus_>although... 2x is not so bad
21:34:05  <piscisaureus_>but 100x would've been better sure
21:40:31  <igorzi>piscisaureus: https://gist.github.com/1340298
21:40:57  <igorzi>bnoordhuis: you still using the cluster? (i want to run some more benchmarks on windows)
21:41:17  <igorzi>piscisaureus_: https://gist.github.com/1340298
21:41:32  <piscisaureus_>igorzi: I see it
21:41:41  <piscisaureus_>igorzi: is there a test for this
21:42:02  <igorzi>piscisaureus_: yeah, i wrote one (should be in the patch)
21:42:04  <piscisaureus_>(just curious; I can imagine you have no time to write one)
21:43:55  <piscisaureus_>igorzi: just curious, what's the difference between FILE_GENERIC_READ and GENERIC_READ ?
21:44:07  <piscisaureus_>igorzi: +1 btw, it looks good
21:45:48  <igorzi>piscisaureus_: the OS will figure out what GENERIC_READ means for a particular kernel object.. here we know that we're dealing with files, so we can just set it directly to FILE_GENERIC_READ. http://msdn.microsoft.com/en-us/library/windows/desktop/aa446632(v=vs.85).aspx
21:46:19  <piscisaureus_>aah, ok cool
21:47:35  <CIA-95>libuv: Igor Zinkovsky master * rfe97c4d / (src/win/fs.c test/test-fs.c test/test-list.h): windows: honor O_APPEND in uv_fs_open - http://git.io/wsRSjQ
22:09:08  <CIA-95>libuv: Igor Zinkovsky master * r82cf0b3 / src/win/process.c : windows: remove quotes from PATH in uv_spawn - http://git.io/-kDMhg
22:10:45  * ryahquit (Quit: leaving)
22:10:54  * ryahjoined
22:14:04  <igorzi>anyone object against upgrading libuv?
22:14:06  <CIA-95>node: Ryan Dahl master * r8652848 / (4 files in 3 dirs): new cluster api - http://git.io/K8e61w
22:14:59  <CIA-95>node: Ryan Dahl master * rd42006c / (doc/api/cluster.markdown lib/cluster.js):
22:15:00  <CIA-95>node: cluster: Remove eachWorker, workerCount
22:15:00  <CIA-95>node: unnecessary - http://git.io/wlUy7Q
22:16:01  * mralephjoined
22:16:13  <ryah>igorzi: go for it
22:16:24  * mralephquit (Client Quit)
22:17:16  * mralephjoined
22:17:47  <ryah>https://github.com/joyent/node/blob/d42006c80a9c827e4c147e22f56bb3d789f578ce/doc/api/cluster.markdown
22:18:24  <wankdanker>Added Windows XP test results https://gist.github.com/1339957
22:18:35  <ryah>wankdanker: nice thanks
22:18:54  <wankdanker>ryah: you are welcome.
22:21:18  <CIA-95>node: Igor Zinkovsky master * r5496e67 / (5 files in 3 dirs):
22:21:19  <CIA-95>node: upgrade libuv to 82cf0b38c0
22:21:19  <CIA-95>node: fixes #2004 - http://git.io/wDSWKQ
22:22:39  * piscisaureus_quit (Ping timeout: 248 seconds)
22:30:36  * felixgequit (Quit: http://www.debuggable.com/)
22:34:33  <CIA-95>node: Ryan Dahl master * r1e3e6b7 / doc/api/cluster.markdown : cluster: add example for message passing - http://git.io/CuyelQ
22:35:10  <ryah>https://github.com/joyent/node/blob/1e3e6b7c15f39500d30d91846ad3621fb72aa91b/doc/api/cluster.markdown
22:37:55  <bnoordhuis>ryah: why did you remove cluster.startMaster()? :(
22:38:15  <ryah>bnoordhuis: less typing
22:38:52  <bnoordhuis>ryah: more typing - now i need to do -> for (i = 0; i < ncpus; i++) cluster.fork()
22:39:12  <igorzi>bnoordhuis: still using the cluster?
22:39:26  <ryah>bnoordhuis: oh right - but more explicit
22:39:40  <bnoordhuis>igorzi: yes, for a little longer
22:39:55  <bnoordhuis>i wanted to bench the cluster api too and would've been done by now
22:40:10  <igorzi>bnoordhuis: k, np
22:40:28  <bnoordhuis>if a certain someone didn't modify the api on me >:(
22:41:10  <ryah>sorry
22:41:19  <ryah>im working with TJ to get this shit in place at the last minute
22:41:33  * piscisaureus_joined
22:44:01  * AvianFluquit (Ping timeout: 260 seconds)
22:45:14  <igorzi>ryah: i'm fixing pipe_connect_bad_name on windows.. is the idea that for a bad pipe name uv_pipe_connect should return 0, and the failure is notified with connect callback?
22:45:37  <ryah>igorzi: correct
22:46:22  <ryah>we could probably even make uv_pipe_connect return void
22:46:56  <ryah>(just to make it explicit)
22:47:22  <igorzi>i can do that.. want me to go ahead (for v0.6)?
22:47:51  <bnoordhuis>maxing out 8 cores is a pretty awesome sight in htop :)
22:48:04  * AvianFlujoined
22:49:29  * piscisaureus_quit (Read error: Connection reset by peer)
22:50:35  <ryah>igorzi: yeah
22:50:56  <ryah>im going to start preparing the changelog
22:51:09  * piscisaureus_joined
22:53:52  <CIA-95>node: Ben Noordhuis master * r89014c9 / benchmark/http_simple_cluster.js : bench: add http_simple cluster edition benchmark - http://git.io/2YPagA
22:53:53  <CIA-95>node: Jacob H.C. Kragh master * rc1472ac / doc/api/fs.markdown :
22:53:53  <CIA-95>node: docs: fix some minor typos in the fs documentation
22:53:53  <CIA-95>node: * fchmodSync: replace 'path' with 'fd'
22:53:53  <CIA-95>node: * lchmod: replace 'fd' with 'path'
22:53:54  <CIA-95>node: * utimes, futimes, fsync: mark 'callback' as optional - http://git.io/wFEShw
22:56:44  <ryah>bnoordhuis: nice
22:57:19  <CIA-95>node: Ryan Dahl master * r5c3954a / doc/api/child_processes.markdown :
22:57:19  <CIA-95>node: docs: make std*Stream spawn opts explicitly internal
22:57:19  <CIA-95>node: Fixes #1884. - http://git.io/VbRu9Q
22:57:19  <bnoordhuis>i'm seeing good numbers too
22:57:26  <bnoordhuis>not quite a linear speed-up but close
22:57:49  * AvianFluquit (Ping timeout: 255 seconds)
22:58:05  <ryah>SIGINT on windows?
22:58:10  <ryah>not happening, right?
22:58:18  <ryah>process.on('SIGINT') ?
22:58:39  <ryah>i dont think so
22:58:52  * AvianFlujoined
22:58:56  <ryah>we can condier this an API flaw and will make it in v0.6.1
22:59:20  <piscisaureus_>ryah: currently no signals happenb on windows at all
23:01:58  <ryah>piscisaureus_: what about this :https://github.com/joyent/node/issues/2000
23:02:04  <ryah>it's our last issue for v0.6 milestone
23:02:29  <piscisaureus_>ryah: look into libuv issues as well
23:02:32  <piscisaureus_>ryah: but that one is done
23:02:38  <ryah>piscisaureus_: can you close?
23:02:43  <piscisaureus_>aye
23:03:16  * AvianFluquit (Read error: Connection reset by peer)
23:04:08  <ryah>so let's get the void return on uv_pipe_close and also https://github.com/joyent/libuv/issues/234
23:04:16  <ryah>piscisaureus_: do you have time to add these flags?
23:04:33  <ryah>uv_pipe_connect that is
23:04:34  * AvianFlujoined
23:04:45  <piscisaureus_>ryah: I have a few minutes left
23:04:49  <piscisaureus_>so, yeah
23:05:06  <ryah>good
23:07:06  <CIA-95>libuv: Igor Zinkovsky master * rfaca140 / (7 files in 4 dirs): make uv_pipe_connect return void - http://git.io/tUHqfQ
23:07:29  <igorzi>ryah bnoordhuis: can you verify the unix change --^
23:07:49  <bnoordhuis>igorzi: i'm done with the cluster
23:08:07  <igorzi>bnoordhuis: cool, i'll reboot into windows now
23:08:14  <igorzi>bnoordhuis: how do the numbers look?
23:08:21  <bnoordhuis>igorzi: better than windows :)
23:08:46  <bnoordhuis>igorzi: https://gist.github.com/1340726
23:09:33  <ryah>only slightly better - but that's okay
23:10:34  <ryah>bnoordhuis: did you do the startup time and io benchmarks?
23:10:34  <igorzi>bnoordhuis: i noticed that you collected /bytes/10240, but not /bytes/102400
23:10:44  <igorzi>or is that a typo?
23:10:55  <ryah>igorzi: good catch!
23:11:02  <bnoordhuis>oh damn
23:11:34  <bnoordhuis>i suppose i should redo those :(
23:11:44  * bnoordhuislogs in again
23:11:46  <igorzi>bnoordhuis: i'd also redo the /bytes/1024 benchmark
23:11:55  <bnoordhuis>igorzi: why?
23:12:07  <igorzi>that 1st result looks like an outlier
23:12:21  <igorzi>8179.1, 6307.2, 6286.1
23:12:36  <bnoordhuis>right, i ran that one two times
23:12:52  <bnoordhuis>but it was pretty persistent - a first fast run, followed by two slower ones
23:13:13  <bnoordhuis>maybe a gc artifact
23:14:09  <bnoordhuis>but i'll re-run it just in case
23:14:16  <igorzi>bnoordhuis: wcat has an option to drop lowest+highest results.. that's how i usually ran windows benchmarks.. i think i took that out last night when i was playing with cygwin benchmarks
23:14:52  <igorzi>bnoordhuis: can i log into ndjs-win3 for a minute? i'll put that back into the script
23:15:04  <bnoordhuis>igorzi: sure
23:15:12  <bnoordhuis>do i need to log out first?
23:15:24  <igorzi>nope.. it'll just bump you off :)
23:15:30  <bnoordhuis>:)
23:15:59  <ryah>yeah we should be careful to be collecting that stuff in the exact same way
23:16:17  <ryah>usually i restart the server between runs
23:17:05  <igorzi>bnoordhuis: ok done.. it'll drop highest/lowest results
23:17:22  <piscisaureus_>ryah: bnoordhuis: igorzi: https://gist.github.com/1340747
23:17:28  <piscisaureus_>I need some input on flags names
23:17:36  <igorzi>bnoordhuis: you can just log back into ndjs-win3, and it'll bump me off
23:18:04  <ryah>piscisaureus_: maybe make the comment that they're not implemented yet in the uv.h
23:18:45  <ryah>piscisaureus_: instead of UV_FS_EVENT_POLL use UV_FS_EVENT_STAT ?
23:19:48  <igorzi>ryah: btw, any plans to do anything about the slab allocator for v0.8?
23:23:23  <ryah>igorzi: not yet - but we should discuss
23:23:46  <ryah>im leaning on allowing libuv to alloc the memory
23:25:00  <igorzi>ryah: cause there's a possible optimization (no 0-reads up to n connections) that we're not using on windows because of that
23:26:07  * CoverSlidequit (Quit: leaving)
23:27:04  * piscisaureus_quit (Ping timeout: 255 seconds)
23:29:07  <ryah>igorzi: yeah
23:30:16  * piscisaureus_joined
23:31:40  * AvianFluquit (Ping timeout: 260 seconds)
23:36:44  <bnoordhuis>node: src/uv-common.c:98: uv_err_name: Assertion `0' failed. <- *sigh*
23:38:56  <ryah>meh
23:38:59  <ryah>how did you get that?
23:39:00  * AvianFlujoined
23:40:42  * piscisaureus_quit (Ping timeout: 260 seconds)
23:42:40  <bnoordhuis>ryah: /bytes/102400 on master
23:43:01  <bnoordhuis>23.5 req/s, 35 errors...
23:43:32  <ryah>god
23:44:05  * mralephquit (Quit: Leaving.)
23:45:04  * mralephjoined
23:48:50  <ryah>alright got the release commit set
23:50:03  * mralephquit (Read error: Connection reset by peer)
23:50:10  <ryah>RC1 coming up
23:55:53  <ryah>http://nodejs.org/dist/v0.6.0/node-v0.6.0rc1.tar.gz