00:01:49  * ericktquit (Quit: erickt)
00:04:50  * isaacsjoined
00:25:23  <CIA-53>libuv: Bert Belder wintty * rd43240a / (6 files in 2 dirs): win: TTY work - WIP - http://git.io/wudrKA
00:38:05  <piscisaureus>ryah: what happens if you change the tty from line-buffered into raw mode, but the user has already typed some stuff?
01:06:33  * ericktjoined
01:07:16  <erickt>bah, laptop just got stolen with all of my libuv patches
01:09:46  <indutny>omg
01:09:58  <indutny>steven wilson's album in lossless on what.cd
01:12:55  * isaacsquit (Quit: isaacs)
01:13:03  <indutny>any progress with 0.5.8 milestone?
01:13:59  <indutny>ryah: are we really going to fix this before 0.5.8 - https://github.com/joyent/node/issues?sort=created&direction=desc&state=open&page=1&milestone=5 ?
01:16:36  * bnoordhuisjoined
01:18:47  * ericktquit (Quit: erickt)
01:27:24  <piscisaureus>erickt: wtf
01:27:55  <bnoordhuis>hah, i know what this is about
01:28:02  <piscisaureus>bnoordhuis: yt!!
01:28:07  <bnoordhuis>i am
01:28:17  * mikealjoined
01:28:20  <piscisaureus>bnoordhuis: you what *what* is a about?
01:28:35  <indutny>haha
01:28:54  <bnoordhuis>https://github.com/joyent/libuv/pull/200 <- amirite or amirite?
01:29:29  <piscisaureus>well, luckily he sent a pull request :-)
01:29:32  <piscisaureus>not everything is lost
01:29:54  <indutny>god, I love C++ (struct sockaddr*)&addr
01:30:09  <piscisaureus>indutny: it looks like C though :-)
01:30:10  <bnoordhuis>indutny: not just c++
01:30:11  <indutny>variable is unreferenced referenced type
01:30:16  <indutny>bnoordhuis: doesn't matter
01:30:19  <indutny>:)
01:30:50  <bnoordhuis>i kinda agree that the uv_ip?_addr interface kind of sucks from an error reporting POV
01:31:15  <bnoordhuis>we should probably document that it's a convenience function, nothing more
01:40:32  * brsonquit (Ping timeout: 276 seconds)
01:47:57  <piscisaureus>omg i gtg to bed
01:50:55  <bnoordhuis>good idea :)
01:50:58  <bnoordhuis>sleep tight, piscisaureus
01:52:29  <piscisaureus>bnoordhuis: can I ask you one more quiestion?
01:52:37  <piscisaureus>*question
01:52:40  <bnoordhuis>sure
01:53:03  <piscisaureus>bnoordhuis: I noticed that when I call `man` or `less` on linux the terminal switches to another buffer
01:53:10  * ericktjoined
01:53:17  <piscisaureus>and changes the height of the tty scrollback buffer etc
01:53:18  <bnoordhuis>buffer?
01:53:20  <piscisaureus>how does it do that?
01:53:21  <bnoordhuis>oh, right
01:53:34  <piscisaureus>and it switches back when you exit the application
01:53:50  <bnoordhuis>not sure, i'
01:53:59  <bnoordhuis>*i'd have to look it up
01:54:08  <piscisaureus>where would you look?
01:54:10  <piscisaureus>the source heh
01:54:14  <piscisaureus>but where can I found that
01:54:15  <bnoordhuis>the ncurses source
01:54:17  <piscisaureus>*find
01:54:34  <piscisaureus>on github? :-)
01:54:47  <bnoordhuis>easier even: apt-get source libncurses5
01:55:08  <piscisaureus>on windows?
01:55:12  <piscisaureus>:-p
01:55:27  <bnoordhuis>http://www.gnu.org/s/ncurses/
01:55:45  <bnoordhuis>i should've linked you go lmfgtfy.com :)
01:55:48  <bnoordhuis>*to
01:56:13  <piscisaureus>I know google works - but I find it difficult to find browseable sources for these things
01:56:22  <piscisaureus>usually you would just know where to look
01:56:55  <bnoordhuis>i suspect that in this case you'll have to download a tarball
01:57:02  <bnoordhuis>ncurses is positively ancient
02:20:23  <piscisaureus>Okay, I know enough. Thanx
02:20:25  <piscisaureus>bye
02:48:01  * dmkbotquit (Remote host closed the connection)
02:48:10  * dmkbotjoined
03:14:20  * bnoordhuisquit (Read error: Operation timed out)
03:38:55  * piscisaureusquit (Ping timeout: 248 seconds)
03:49:38  * isaacsjoined
04:38:08  <isaacs>nots
04:38:09  <isaacs>notes
04:38:17  <isaacs>erm... wrong room, sorry :)
04:50:40  <pquerna>http://omniti.com/surge/2011/sessions/training
04:50:43  <pquerna>node.js hack day eh
06:07:37  * mralephjoined
06:53:29  * piscisaureusjoined
06:54:10  <piscisaureus>mraleph: what's up with libuv performing poorly on mac?
06:56:42  <mraleph>piscisaureus: nuh, apparently it's a bad ab that measures complete crap
06:57:03  <piscisaureus>mraleph: oh, ok
06:58:04  <piscisaureus>mraleph: I would be really surprised if it was bad - the only difference is that is uses kqueue and not epoll. But miracles do happen ...
07:00:56  * mralephquit (Quit: Leaving.)
07:21:54  <piscisaureus>bnoordhuis: ryah: https://gist.github.com/1241230 <-- master still worse than 0.4 (on mac)
07:23:29  * piscisaureus_joined
07:38:29  * isaacsquit (Quit: isaacs)
08:25:47  <indutny>hey guys
08:25:55  <indutny>is anyone around?
08:26:10  <indutny>oh, nevermind
08:26:11  <indutny>:)
08:32:09  * piscisaureus_hmm. on my mac v0.4.13-pre is a little slower than master. But I may be missing limit issues, I don't know what else to do other than ulimit -n 10000
08:32:41  * piscisaureus_And I still wonder why colloquy makes me talk in /me statements
08:33:19  * piscisaureus_Does this fix it?
08:36:03  <indutny>haha :)
08:36:07  * indutnydon't know
08:36:18  * piscisaureus_this really sucks
08:36:40  * piscisaureus_macs are just so painful
08:36:57  <indutny>:)
08:37:11  * piscisaureus_mac is not so bad but there just is no good software for mac.
08:37:48  <indutny>UX is very good, though
08:39:38  * piscisaureus_UX very good, sure :-/
08:39:51  * piscisaureus_Now how do I disable this bloody /me mode
08:40:02  * piscisaureus_quit (Quit: piscisaureus_)
08:40:18  * piscisaureus_joined
08:40:18  * piscisaureus_quit (Client Quit)
08:40:36  * piscisaureus_joined
08:41:26  * piscisaureus_more #mac fun: I typed my password in #node.js once because osx switched focus while I was typing it.
08:41:33  * piscisaureus_Now that's UX I tell you
08:47:26  <indutny>haha
08:47:53  <indutny>piscisaureus_: I can't see password in logs :P
08:48:31  <piscisaureus>indutny: doesn't matter - i changed it anyway
08:48:36  <indutny>heh :)
08:48:45  <piscisaureus>indutny: ryah did notice it btw
08:48:53  <indutny>piscisaureus: hehe :)
08:49:00  <indutny>irccloud filtered that
08:49:24  * piscisaureus_hopes so
08:52:38  * piscisaureus_1joined
08:53:26  * piscisaureus_1quit (Client Quit)
08:53:43  * piscisaureus_quit (Quit: piscisaureus_)
08:54:03  * piscisaureus_joined
08:54:35  <piscisaureus_>maybe adium > colloquy
08:54:41  <piscisaureus_>:-)
08:54:49  <indutny>:)
08:54:57  <indutny>friend of my is using adium
08:55:04  <indutny>mraleph should be using it too
08:55:06  <indutny>AFAIK
08:55:45  <piscisaureus_>it looks a bit icky
08:55:54  <piscisaureus_>but at least it behaves reasonable
08:58:22  * piscisaureus_quit (Client Quit)
08:58:38  * piscisaureus_joined
08:59:59  * piscisaureus_quit (Client Quit)
09:38:18  * mikealquit (Quit: Leaving.)
14:05:16  * bnoordhuisjoined
14:05:30  <CIA-53>libuv: Bert Belder wintty * r47ded49 / (7 files in 2 dirs): win: TTY work - WIP - http://git.io/PyVKMg
14:07:43  <piscisaureus>erickt: so you got your laptop back?
14:22:03  <piscisaureus>[% 100|+ 65|- 7]: Done.
14:22:03  <piscisaureus>^-- libuv :-(
14:23:07  <bnoordhuis>piscisaureus: can you test the c-ares branch?
14:23:42  <piscisaureus>I guess- just a sec
14:25:06  <bnoordhuis>no rush
14:25:31  <bnoordhuis>i should've said: "can you test the c-ares branch some time today?"
14:25:39  <piscisaureus>probably :-)
14:26:01  <piscisaureus>uv_tty_get_winsize with fd == 0 --> no luck on windows
14:27:10  <piscisaureus>ryah: can't we just say that people have to call uv_tty_getwinsize on stdout/stderr?
14:29:06  <CIA-53>libuv: Bert Belder wintty * r00d50cc / (7 files in 2 dirs): win: TTY work - WIP - http://git.io/Ng2iEA
14:30:23  <piscisaureus>bnoordhuis: [% 100|+ 65|- 7]: Done.
14:30:27  <piscisaureus>same as master
14:31:25  <bnoordhuis>good, thanks
14:31:42  <bnoordhuis>ryah: if you have a sec, can you test the c-ares branch on darwin?
14:32:22  * piscisaureus_joined
14:35:20  <piscisaureus_>bnoordhuis: [% 100|+ 64|- 5]: Done. <— lion
14:35:46  <bnoordhuis>is that the same as master too?
14:35:49  <bnoordhuis>piscisaureus: ^
14:35:55  <bnoordhuis>err... piscisaureus_
14:35:58  <bnoordhuis>confusing!
14:36:35  <piscisaureus_>bnoordhuis: I am not that autistic, no worries :-)
14:36:48  <piscisaureus_>bnoordhuis: yes, same as master.
14:37:01  <piscisaureus_>bnoordhuis: btw,
14:37:05  <piscisaureus_>Output from process `fs_symlink`:
14:37:05  <piscisaureus_>run-tests(56984,0x7fff79902960) malloc: *** error for object 0x7fe183800000: pointer being freed was not allocated
14:37:15  <bnoordhuis>hmm
14:37:19  <bnoordhuis>will check
14:39:02  <piscisaureus_>bnoordhuis: btw, stack trace:
14:39:03  <piscisaureus_>Program received signal SIGABRT, Aborted.
14:39:03  <piscisaureus_>0x00007fff8ea3ece2 in __pthread_kill ()
14:39:03  <piscisaureus_>(gdb) bt
14:39:03  <piscisaureus_>#0 0x00007fff8ea3ece2 in __pthread_kill ()
14:39:03  <piscisaureus_>#1 0x00007fff92d8a7d2 in pthread_kill ()
14:39:04  <piscisaureus_>#2 0x00007fff92d7ba7a in abort ()
14:39:04  <piscisaureus_>#3 0x00007fff92dda84c in free ()
14:39:05  <piscisaureus_>#4 0x000000010003d5ef in eio_destroy (req=0x1003000c0) at eio.c:857
14:39:05  <piscisaureus_>#5 0x000000010003d73f in eio_finish (req=0x1003000c0) at eio.c:885
14:39:06  <piscisaureus_>#6 0x000000010003d13b in etp_poll () at eio.c:703
14:39:06  <piscisaureus_>#7 0x000000010003d92d in eio_poll () at eio.c:971
14:39:07  <piscisaureus_>#8 0x0000000100033322 in uv_eio_want_poll_notifier_cb (watcher=0x10008e968, status=0) at uv-eio.c:55
14:39:07  <piscisaureus_>#9 0x0000000100026215 in uv__async (loop=0x10008efc0, w=0x10008e9b0, revents=524288) at core.c:488
14:39:08  <piscisaureus_>#10 0x0000000100037b59 in ev_invoke_pending (loop=0x10008efc0) at ev.c:2143
14:39:08  <piscisaureus_>#11 0x00000001000384c4 in ev_run (loop=0x10008efc0, flags=0) at ev.c:2519
14:39:30  <bnoordhuis>weird that valgrind doesn't complain here
14:40:02  <erickt>piscisaureus: no, unfortunately. I'm on another computer. Fortunately-ish it was my work computer that I just installed an online backup, so I don't think I lost too much
14:40:03  <piscisaureus_>irc://irc.freenode.net:6667/#13
14:41:04  <piscisaureus_>erickt: hmm- fortunately-ish sounds just right
14:41:39  <piscisaureus_>I am kinda surprised though - you did a pull request and the next moment your laptop is gone
14:42:03  <erickt>I know! he just grabbed it from right under my fingertips
14:42:28  <piscisaureus_>erickt: huh you got mugged?
14:42:46  <erickt>no, i was at a starbucks coding away and he just grabbed it and ran
14:43:40  <piscisaureus_>oh hmm that sucks!
14:43:52  <bnoordhuis>that's why i have a 5 kg laptop
14:43:58  <bnoordhuis>let him try and run away with that!
14:44:06  <piscisaureus_>lol
14:44:07  <erickt>yeah, the bastard wouldn't even let me push my changes first
14:44:16  * ericktis jealous
14:45:44  <piscisaureus_>now I'll never dare to code at the starbucks.
14:46:35  <piscisaureus_>Or I should chain my laptop to myself
14:48:37  <bnoordhuis>put a sticker on it: "warning, don't touch - owner has a contagious skin disease"
14:50:51  <piscisaureus_>bnoordhuis: you got the email from kasper too right?
14:51:01  <bnoordhuis>piscisaureus_: yes
14:51:07  <indutny>bnoordhuis: I suggest you using IBM 350 as hard drive
14:51:19  <indutny>bnoordhuis: it's quite hard to steal you data when you using such disk
14:51:44  <piscisaureus_>just make your computer very ugly
14:52:01  <piscisaureus_>print out nodejs.org and glue it to the cover
14:52:15  <indutny>piscisaureus: a web page?
14:53:20  <piscisaureus_>indutny: ?
14:53:29  <indutny>piscisaureus_: nevermind
14:53:33  <indutny>:)
14:53:39  <indutny>piscisaureus_: i've couchbase stickers
14:53:45  <indutny>and I don't know where to glue them
15:04:30  <erickt>I was thinking about lugging my old desktop and 27in crt monitor next time I want to code in public
15:13:59  <indutny>erickt: 27in crt monitor?
15:14:15  <indutny>quite big
15:14:17  <erickt>yeah, it was way awesome back in 2002!
15:16:02  <indutny>I'm working on 17'' notebook :P
15:16:18  <indutny>going to buy external monitor this year
15:16:27  <indutny>s/this/in this
15:24:06  * piscisaureus_quit (Quit: Leaving.)
16:03:45  * isaacsjoined
16:10:55  * bentkusjoined
16:21:58  * rmustaccjoined
16:33:19  <ryah>bnoordhuis: sure.. one sec
16:44:44  <ryah>bnoordhuis: no failing c-ares tests
16:45:53  <ryah>er- sorry using master
16:45:59  <ryah>one sec
16:46:29  <ryah>origin/c-ares also okay
16:51:59  <ryah>http://arlolra.no.de/bench/ab-hello-world-string-1024
16:52:08  <ryah>^-- seems we got a jump recently
16:59:16  <ryah>bnoordhuis: f9fec3a <--
16:59:26  <ryah>bnoordhuis: we should not be changing API
16:59:31  <ryah>https://github.com/joyent/node/commit/f9fec3a
16:59:53  <ryah>v0.4.12 is bug fixes only - this is not a bug
17:10:26  <piscisaureus>ryah: a jusp - yeah, but we've lost it again on the last 2 bars
17:10:28  <piscisaureus>*jump
17:15:14  <bnoordhuis>ryah: re f9fec3a: you consider it an api change?
17:16:12  <bnoordhuis>i can get it to overflow the stack if you try write in a loop
17:17:28  <bnoordhuis>provided the writes always succeed immediately
17:18:58  <ryah>bnoordhuis: tell them to use v0.5
17:19:02  <ryah>(in the future)
17:19:13  <bnoordhuis>hah, okay
17:19:54  <ryah>i think we should be only making critical fixes to v0.4 now
17:21:22  <piscisaureus>at what time are we having the call? 40 mins or 1:40?
17:24:10  <indutny>looks like my commit decreased performance..
17:24:25  <indutny>https://github.com/joyent/node/commit/82d0ac7ac783e3532fa3f0a833be4721907e0ada
17:24:28  <indutny>hehe :)
17:24:51  <indutny>results seems to be inaccurate
17:24:54  <ryah>guys can we do the call at noon today?
17:25:05  <ryah>indutny: i think there is a lot of varience
17:25:12  <igorzi>ryah: yep.. that would actually be better for me too
17:25:17  <bnoordhuis>sure
17:25:19  <ryah>*variance
17:25:24  <piscisaureus>noon - now = ?
17:25:28  <ryah>piscisaureus: 1:40
17:25:31  <igorzi>bnoordhuis: hey... i'm trying to fix tcp_write_error on windows. the write doesn't fail. what is it in this test that makes the write fail? large buffer size?
17:25:39  <piscisaureus>ok
17:26:20  <bnoordhuis>igorzi: the test already passes? good
17:26:33  <piscisaureus>it fails for me too
17:26:59  <igorzi>no, it fails for me too, i'm trying to make it pass.. in order to make it pass the write needs to fail
17:27:08  <igorzi>the write succeeds on windows
17:27:09  <bnoordhuis>oh, right
17:27:19  <bnoordhuis>you mean it accepts the 10 MB write buffer?
17:27:23  <ryah>bnoordhuis: can you check that ./node test/simple/test-child-process-ipc.js is working on linux
17:27:43  <bnoordhuis>ryah: wfm
17:27:45  <igorzi>i'm not getting status == -1 in the callback
17:27:46  <ryah>thanks
17:28:36  * dmkbotquit (Remote host closed the connection)
17:28:42  * dmkbotjoined
17:29:07  <bnoordhuis>igorzi: the test in a nutshell asserts that after a failed write the write callback is still run
17:29:09  <ryah>igorzi: can you test this in windows: ./node test/simple/test-regress-GH-1726.js
17:29:20  <ryah>igorzi: if you're on HEAD
17:29:37  <ryah>igorzi: (it's probably broken)
17:29:41  <bnoordhuis>to make it fail it tries to do a stupendously large write (10 MB)
17:30:06  <bnoordhuis>fail as in "no way that's being sent out over the wire in one chunk"
17:30:42  <bnoordhuis>then close the connection
17:31:45  <bnoordhuis>so the close cleanup code should invoke the write callback
17:31:54  <ryah>bnoordhuis: we need to do the secret weapon http-parser binding trick this week
17:31:57  <CIA-53>libuv: Bert Belder wintty * r622eb99 / (7 files in 2 dirs):
17:31:57  <CIA-53>libuv: win: implement tty
17:31:57  <CIA-53>libuv: loose end: line-buffered input reads ascii, not unicode - http://git.io/n8Zyzw
17:32:00  <ryah>bnoordhuis: which will make us faster than v0.4
17:32:16  <ryah>bnoordhuis: (aka don't go to js for each header)
17:32:25  <bnoordhuis>ryah: right
17:32:40  <bnoordhuis>any ideas on how to go about that?
17:32:44  <piscisaureus>^-- if anybody feels like reviewing :-)
17:33:46  <igorzi>bnoordhuis: windows is quiet happy with 10MB buffer.. it even bypasses IOCP when it does the write
17:34:07  <piscisaureus>rly?
17:34:09  <piscisaureus>with 10mb buffer?
17:34:10  <ryah>bnoordhuis: yeah. add a bunch of pointer, length values to the HTTPParser class - say 50
17:34:25  <ryah>bnoordhuis: try to fill up those with the assumption that 99% of the time we get the headers in the first packet
17:34:53  <ryah>bnoordhuis: so we have a slow case of headers > 50 lines and header do not fit in one packet
17:35:27  <bnoordhuis>okay, makes sense
17:35:32  <ryah>bnoordhuis: in the on_headers_complete walk through the (pointer, length) pairs and create v8::String objects
17:36:56  <igorzi>piscisaureus: yeah
17:37:38  <bnoordhuis>igorzi: is there a good way to have a write abort half way through? or make it fail in another way?
17:39:08  <igorzi>bnoordhuis: maybe we can close the connection on the server when it accepts (instead of doing a read)?
17:40:37  <bnoordhuis>igorzi: no, that means the write will fail immediately (because the connection is closed, that was the first iteration of tcp-write-error)
17:40:58  <bnoordhuis>it needs to fail while it's in progress
17:41:34  <bnoordhuis>i understand windows accepts arbitrarily sized write buffers?
17:43:22  <igorzi>piscisaureus: you got any ideas on how to make the write fail while it's in progress?
17:44:53  <igorzi>ryah: test-regress-GH-1726.js passes for me
17:45:13  <piscisaureus>igorzi: yeah - queue a few writes, then close the handle
17:45:52  <piscisaureus>igorzi: or close the receiving
17:46:25  <ryah>igorzi: interesting! thanks
17:47:09  <ryah>sounds like #1726 is a bug in libuv perhaps...
17:47:48  <igorzi>bnoordhuis: i'll play with it. i'll update the test once i find a reliable way to make it fail while in-progress
17:47:59  <bnoordhuis>igorzi: cool
17:52:48  <indutny>ryah: https://github.com/joyent/node/pull/1768
17:54:52  * brsonjoined
17:58:00  <ryah>indutny: thanks!
17:58:08  <CIA-53>node: Fedor Indutny master * rc26cf84 / (3 files in 3 dirs):
17:58:09  <CIA-53>node: debugger: fix backtrace with no frames
17:58:09  <CIA-53>node: Fixes #1768 - http://git.io/pX4FTQ
17:58:53  * dmkbotquit (Remote host closed the connection)
17:58:57  * dmkbotjoined
17:59:37  <indutny>np
17:59:41  <indutny>brb
18:03:21  * bentkusquit (Remote host closed the connection)
18:19:34  <ryah>i guess we need to be closing the stdio pipes after the process exits in uv_spawn
18:25:40  <indutny>hm...
18:25:57  <indutny>piscisaureus: are you sure that GenerateConsoleCtrlEvent works for foreign process?
18:27:56  <indutny>piscisaureus: I'm definitely sending events to console, but I see no reaction from it
18:29:22  <indutny>guess we'll need to have mailslots in libuv first
18:29:31  <indutny>only then remote debugging by process id
18:35:28  <piscisaureus>indutny: link me to your code
18:42:56  <indutny>piscisaureus: one second
18:45:25  * igorzi_joined
18:45:38  <indutny>piscisaureus: https://github.com/indutny/node/tree/fsb-test2
18:46:01  <indutny>piscisaureus: try running benchmark/http_simple.js and then `node debug -p pid`
18:46:12  <indutny>or `node` and process.kill(pid, 'SIGUSR1')
18:46:19  <indutny>(signal is actually ignored)
18:46:40  <indutny>piscisaureus: and that's code https://github.com/indutny/node/blob/fsb-test2/src/node.cc#L1615
18:47:07  <isaacs>ryah: when is 0.5.8 going to be released?
18:47:14  * igorziquit (Ping timeout: 252 seconds)
18:52:01  <isaacs>bnoordhuis: review? https://github.com/isaacs/node/commit/c828ded0c2b6905c0f3ce0301e3c76923e9f5dec
18:52:14  <isaacs>(literally one char fix)
18:52:26  <bnoordhuis>hah
18:52:30  <bnoordhuis>sure, lgtm
18:53:09  <CIA-53>node: isaacs master * rc828ded / lib/zlib.js : zlib: Typo. s/opt/opts/ - http://git.io/1-vBoA
18:53:40  <indutny>piscisaureus: any ideas?
19:00:36  <bnoordhuis>call time?
19:01:38  <piscisaureus>yes, call me
19:02:33  <piscisaureus>ryah: ^
19:04:49  <ryah>yes
19:04:56  <ryah>one sec
19:04:57  <ryah>sorry guys
19:05:26  <isaacs>i'm available. we doing skype?
19:05:42  <piscisaureus>indutny: I think the child process that you want to signal needs to be created with the CREATE_NEW_PROCESS_GROUP flag
19:06:12  <piscisaureus>indutny: normally you send the signal to the "master" of the process group
19:16:53  <indutny>piscisaureus: heh
19:16:57  <indutny>piscisaureus: this is remote process
19:17:10  <indutny>piscisaureus: that's what the problem is in
19:17:32  <indutny>ok, thank you, anyway!
19:17:46  <indutny>it's 2am in my place and I'm going to sleep!
19:17:46  <indutny>ttyl, everyone!
19:28:25  <bnoordhuis>sleep tight, indutny
19:29:10  <piscisaureus>yes sleep well indutny
19:30:21  <ryah>piscisaureus: i think you ought to just land the tty stuff and let's see what breaks
19:36:48  <CIA-53>libuv: Bert Belder master * r622eb99 / (7 files in 2 dirs):
19:36:48  <CIA-53>libuv: win: implement tty
19:36:48  <CIA-53>libuv: loose end: line-buffered input reads ascii, not unicode - http://git.io/n8Zyzw
19:37:32  <ryah>http://arlolra.no.de/bench/startup_time <-- updated
19:45:06  <ryah>bnoordhuis: https://gist.github.com/325b90295f17c751ed06
19:45:09  <ryah>bnoordhuis: remember this?
19:45:22  <bnoordhuis>ryah: yes
19:45:28  <ryah>bnoordhuis: did that ever get relanded?
19:45:40  <bnoordhuis>piscisaureus: ^ ?
19:45:41  <ryah>it seems not
19:45:44  <piscisaureus>no
19:45:50  <ryah>okay, im putting it back int
19:46:09  <ryah>should i take bnoordhuis's fixes or piscisaureus's original?
19:46:15  <piscisaureus>we need pnoordhuis' patch
19:46:21  <bnoordhuis>to ask the question is to answer it
19:46:26  <piscisaureus>bnoordhuis's fixes should be okay
19:46:31  <bnoordhuis>okay...
19:46:37  <ryah>(is there a way to see the diff with gist?)
19:46:49  <piscisaureus>no :(
19:46:55  <piscisaureus>unless you clone to gist locally
19:47:10  <piscisaureus>but bnoordhuis' fix should avoid stray child processes being left
19:47:16  <piscisaureus>nothing otherwise
19:47:53  <piscisaureus>maybe on -> once is sane also
19:51:02  <bnoordhuis>piscisaureus: implicit declaration of function �UnregisterWait� <- are you missing a `#include <windows.h>` somewhere?
19:51:19  <CIA-53>node: Ryan Dahl master * r040cf02 / test/simple/test-regress-GH-1697.js :
19:51:19  <CIA-53>node: Re-land Bert's test for #1697
19:51:19  <CIA-53>node: Was reverted out in cde81a6675caa5d64d2215b75105cf353ed98dc6 - http://git.io/YPdkOg
19:51:21  <piscisaureus>hmm... probably not
19:51:31  <piscisaureus>bnoordhuis: which file?
19:51:54  <bnoordhuis>piscisaureus: tty.c
19:53:16  <piscisaureus>bnoordhuis: shouldn't happen - it includes uv.h, which includes uv-win.h, which includes windows.h
19:53:26  <piscisaureus>bnoordhuis: also, you would be missing a zillion other stuff
19:53:33  <piscisaureus>bnoordhuis: what toolchain are you using
19:53:37  <bnoordhuis>mingw-gcc
19:53:58  <bnoordhuis>it's also complaining about RegisterWaitForSingleObject and QueueUserWorkItem
19:54:38  <piscisaureus>bnoordhuis: no issues here
19:55:26  * igorzijoined
19:55:48  * igorzi_quit (Ping timeout: 252 seconds)
19:57:44  <bnoordhuis>piscisaureus: apparently it depends on _WIN32_WINNT >= 0x0500
19:57:53  <piscisaureus>bnoordhuis: that should be added by the makefile
19:58:06  <piscisaureus>bnoordhuis: are you compiling with a patched c-ares?
19:58:21  <bnoordhuis>piscisaureus: no
19:58:45  <bnoordhuis>but nm, i wasn't using the makefile
19:59:04  <piscisaureus>you should add -D_WIN32_WINNT=0x0501
20:01:44  <piscisaureus>bnoordhuis: I think I should split the utf8/16 de/encoders out
20:01:55  <piscisaureus>somwhow
20:11:43  <igorzi>bnoordhuis: can you pls try this on linux? https://gist.github.com/1243268
20:12:49  <bnoordhuis>igorzi: Assertion failed in ../test/test-tcp-write-error.c on line 119: tcp_client.write_queue_size == 0
20:13:25  <igorzi>bnoordhuis: hmm, does that indicate a bug? :)
20:13:29  <igorzi>(in uv-unix?)
20:14:43  <bnoordhuis>igorzi: hmm, maybe - i'll check
20:16:53  <bnoordhuis>igorzi: not really a bug, libuv is still flushing the write queue at that point in time
20:18:53  <igorzi>bnoordhuis: i think i'm missing the point of the test.. isn't it to make sure that write_queue_size is 0 when write cb fails?
20:19:27  <bnoordhuis>igorzi: yes, but the assumption was that there's only write taking place
20:20:05  <bnoordhuis>so by the time the write callback runs, the write request has been popped off the queue and the write_queue_size been decremented
20:20:54  <bnoordhuis>but your approach should be easy to fix, just check that tcp_client.write_queue_size == 0 after uv_run returns
20:21:32  <igorzi>bnoordhuis: ok will do
20:21:50  <bnoordhuis>igorzi: i already did :)
20:22:07  <bnoordhuis>it's still failing here though so maybe there *is* a bug in uv-unix :/
20:23:09  <igorzi>bnoordhuis: it's probably not because of multiple writes, it's probably because the handle is closed before write cb
20:24:45  <igorzi>before my change the test was asserting that req->handle->write_queue_size > 0 after the write, and now i'm breaking out of the loop when req->handle->write_queue_size > 0 is true. so it must still be doing just one write.
20:27:32  <igorzi>i'll be back in 20 min
20:29:35  <bnoordhuis>igorzi: https://gist.github.com/7b8565374d5958df05e0 <- wfm
20:31:37  * erickt_joined
20:42:00  <CIA-53>libuv: Ben Noordhuis master * r083c97e / (45 files in 10 dirs): c-ares: upgrade to 1.7.5 - http://git.io/Dker5g
20:42:01  <CIA-53>libuv: Ben Noordhuis master * rb288afd / src/uv-common.c :
20:42:01  <CIA-53>libuv: common: fix includes
20:42:01  <CIA-53>libuv: Don't rely on c-ares to pull in the headers for inet_addr, net_pton and
20:42:01  <CIA-53>libuv: inet_ntop. c-ares 1.7.4 did but 1.7.5 does not. - http://git.io/PKYgjw
20:44:13  <piscisaureus>bnoordhuis: I should go to bed, have to get up early tomorrow. Are you done reviewing?
20:44:32  <bnoordhuis>piscisaureus: i think so
20:45:22  <piscisaureus>bnoordhuis: so you actually understand what's going on
20:45:26  <piscisaureus>:-[
20:45:31  <bnoordhuis>mostly :)
20:45:45  <bnoordhuis>so sorry, no job security there
20:45:54  <piscisaureus>shit
20:48:10  <piscisaureus>bnoordhuis: but why didn't you catch https://github.com/joyent/libuv/commit/622eb99113b28222a8ca149447350d0762f540ef#commitcomment-612379
20:48:10  <piscisaureus>bnoordhuis: and - the refrigerator question - what does https://github.com/joyent/libuv/commit/622eb99113b28222a8ca149447350d0762f540ef#L6R944 do?
20:49:40  <bnoordhuis>rotate the screen?
20:53:42  <piscisaureus>bnoordhuis: correct
20:53:49  <piscisaureus>bnoordhuis: fridge coming your way
20:54:56  <bnoordhuis>i already have the fridge, can i have the pedaalemmer instead?
20:55:16  <piscisaureus>a used one, yeah
21:00:33  <bnoordhuis>piscisaureus: btw, is your mac on?
21:00:52  <piscisaureus>no but it can be on
21:01:16  <bnoordhuis>can you do to your mac what i did to your girlfriend last night?
21:01:50  <piscisaureus>bnoordhuis: hitting on it?
21:02:03  <piscisaureus>better not, it's not mine
21:02:08  <bnoordhuis>turning it on
21:02:37  <piscisaureus>bnoordhuis: oh
21:02:37  <piscisaureus>ok
21:02:54  <piscisaureus>yes, it's on now
21:03:03  * piscisaureus_joined
21:03:51  <bnoordhuis>piscisaureus: https://gist.github.com/e58f684dca20886c3120 <- my key
21:04:05  <piscisaureus>bnoordhuis: you gotta guide me - how do I create an account?
21:04:16  <bnoordhuis>piscisaureus: `adduser bnoordhuis`
21:05:02  <piscisaureus>adduser: command not found
21:05:15  <bnoordhuis>ffs, mac/bsd
21:05:33  <bnoordhuis>piscisaureus: `useradd bnoordhuis`?
21:05:45  <piscisaureus>nope
21:06:04  <bnoordhuis>are you logged in as root btw?
21:06:18  <piscisaureus>I can sudo yes
21:06:33  <bnoordhuis>okay, you need to run these commands as root
21:07:55  <piscisaureus>bnoordhuis: no difference
21:08:17  <bnoordhuis>macs make me sad :(
21:08:47  <bnoordhuis>piscisaureus: http://serverfault.com/questions/20702/how-do-i-create-user-accounts-from-the-terminal-in-mac-os-x-10-5
21:17:41  * piscisaureus__joined
21:21:23  * piscisaureusquit (Ping timeout: 256 seconds)
21:23:08  <piscisaureus_>bnoordhuis: 145.94.57.158
21:23:13  <piscisaureus_>^— try that
21:23:27  <bnoordhuis>yay!
21:23:34  <bnoordhuis>asks me for a password though
21:23:43  <piscisaureus_>I have added your key
21:23:47  <bnoordhuis>copy the key to ~bnoordhuis/.ssh/authorized_keys
21:24:02  <bnoordhuis>chmod 700 ~bnoordhuis/.ssh && chmod 600 ~bnoordhuis/.ssh/authorized_keys
21:24:11  <piscisaureus_>oh I copied it to id_rsa.pub
21:26:18  <piscisaureus_>bnoordhuis: try again
21:26:33  <bnoordhuis>piscisaureus_: still asks me for a pw
21:26:44  <piscisaureus_>passwd
21:27:06  <bnoordhuis>nothing
21:27:08  <bnoordhuis>what username?
21:27:12  <bnoordhuis>maybe pm that :)
21:28:20  <bnoordhuis>piscisaureus_: can you `chown -R bnoordhuis:bnoordhuis /Users/bnoordhuis`?
21:29:06  <piscisaureus_>invalid argument
21:29:12  <bnoordhuis>oh god, macs
21:29:22  <piscisaureus_>chown -R bnoordhuis /Users/bnoordhuis <— works
21:31:43  <igorzi>bnoordhuis: https://gist.github.com/7b8565374d5958df05e0 works.. thanks
21:59:38  <ryah>https://gist.github.com/1243529 <--
21:59:48  <ryah>not very good behavior for either...
22:06:10  <igorzi>bnoordhuis: ryah: i'm fixing the windows build after c-ares upgrade.. where would be a good place in uv-unix to move these? https://github.com/joyent/libuv/blob/master/src/uv-common.c#L29-31
22:07:48  <bnoordhuis>igorzi: internal.h probably
22:08:13  <bnoordhuis>igorzi: it compiled for bert though...
22:10:23  <igorzi>with mingw or vc?
22:10:53  <igorzi>it's also missing 2 files from c-ares, which are needed for windows
22:11:44  <bnoordhuis>piscisaureus_: ^
22:11:56  <bnoordhuis>igorzi: what files are missing?
22:12:34  <bnoordhuis>are you using the makefile? i might not have updated that one :/
22:12:57  <igorzi>i'm building with gyp
22:13:00  <igorzi>c/ares/ares_platform.c
22:13:05  <igorzi>src/ares/ares_iphlpapi.h
22:15:53  <CIA-53>libuv: Igor Zinkovsky master * rf6a365e / test/test-tcp-write-error.c : fix tcp_write_error for windows - http://git.io/yZKADw
22:16:25  <bnoordhuis>igorzi: right, they're full of win32 specific stuff - i'll add them
22:17:45  <igorzi>bnoordhuis: ok.. and please move the includes to uv-unix (https://github.com/joyent/libuv/blob/master/src/uv-common.c#L29-31)
22:18:32  <ryah>http://arlolra.no.de/bench/ab-hello-world-buffer-1024 <-- what caused this bump?
22:19:02  <ryah>it's too bad our other backends are off
22:19:34  <ryah>maybe that box just got faster...
22:20:39  <ryah>yeah nevermind
22:23:17  <bnoordhuis>igorzi: can you test my master branch?
22:24:20  <piscisaureus__>I'm off to bed
22:24:20  <igorzi>bnoordhuis: yep
22:24:23  <piscisaureus__>goodbye all
22:24:32  <bnoordhuis>piscisaureus__: sleep tight
22:24:37  <ryah>piscisaureus__: night
22:24:59  * piscisaureus__quit (Quit: ~ Trillian Astra - www.trillian.im ~)
22:25:02  * mralephjoined
22:27:56  <igorzi>ares_library_init.h(26): fatal error C1083: Cannot open include file: 'ares_iphlpapi.h': No such file or directory
22:28:55  <igorzi>src\ares is not on the include path. #include <ares_iphlpapi.h> --> #include "ares_iphlpapi.h"
22:35:10  <igorzi>damn - after fixing that, now gethostbyname test is failing.. i'll debug
22:44:38  * piscisaureus_part
22:45:04  * piscisaureus_joined
23:03:04  <CIA-53>libuv: Ben Noordhuis master * r9673abe / src/unix/fs.c :
23:03:04  <CIA-53>libuv: unix: fix pointer ownership bug
23:03:04  <CIA-53>libuv: libuv realloc'd a pointer that belonged to and was later freed by libev. - http://git.io/nmUXsA
23:09:01  * isaacsquit (Quit: isaacs)
23:27:08  * isaacsjoined
23:32:49  * isaacsquit (Quit: isaacs)
23:34:12  * isaacsjoined
23:35:17  * isaacsquit (Client Quit)
23:37:00  * erickt_quit (Quit: erickt_)
23:46:58  * mralephquit (Quit: Leaving.)
23:51:32  * isaacsjoined