00:25:00  * dmkbotquit (Remote host closed the connection)
00:25:04  * dmkbotjoined
00:26:16  * bnoordhuisjoined
00:54:07  * dmkbotquit (Remote host closed the connection)
00:54:11  * dmkbot1joined
00:54:30  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
01:35:07  * ericktquit (Quit: erickt)
01:56:39  <igorzi>ryah: do you want the new fs.watch api to completely replace the old fs.watchFile? or do you want to have both?
02:00:07  <ryah>igorzi: can we call it watchFile2 for now?
02:00:36  <ryah>but yeah, i want to replace it eventually
02:01:01  <igorzi>ryah: ok, watchFile2 it is
02:01:49  <ryah>bnoordhuis: rmustacc says that he can't do the solaris binding for watchFile
02:01:56  * brsonquit (Ping timeout: 252 seconds)
02:02:06  <bnoordhuis>ryah: okay, i'll have a go at it
02:02:08  <igorzi>ryah: though since it watches directories also, watchFile2 might be confusing?
02:02:11  <bnoordhuis>once i tame kqueue
02:02:21  <ryah>bnoordhuis: i also, kind of have the desire to take a swing at it
02:02:25  <ryah>i never did event port stuff
02:02:31  <bnoordhuis>sure, go ahead
02:03:23  <ryah>igorzi: should we change the name? i guess we're not going to be providing these stat objects anymore
02:03:34  <ryah>fs.watch() ?
02:04:20  <igorzi>ryah: yeah, I think fs.watch is better
02:04:54  <igorzi>ryah: also do you want to bubble status up to js-land?
02:05:45  <igorzi>ryah: when for example the buffer overflows.. we still fire the event cb with status = -1
02:11:29  <bnoordhuis>igorzi: how does the following work on windows? monitor file foo, unlink foo, recreate foo
02:11:39  <bnoordhuis>do you still get events for foo afterwards?
02:21:20  <igorzi>bnoordhuis: yeah, we'll still get events for foo
02:21:55  <bnoordhuis>igorzi: i suspected as much, thanks
02:22:21  <bnoordhuis>not sure how to resolve that with kqueue or event ports since they're file descriptor based
02:23:05  <ryah>igorzi: yeah we need to bubble up the overflows
02:31:26  * bnoordhuisquit (Read error: Operation timed out)
02:45:23  * bnoordhuisjoined
03:04:46  * isaacsquit (Quit: isaacs)
03:08:02  <bnoordhuis>*sigh* kqueue
03:08:08  <bnoordhuis>ryah: think with me please
03:08:43  <bnoordhuis>kqueue fds aren't pollable and i don't want to chop uv libev too bad
03:09:19  <bnoordhuis>so the only thing i can think of to make kqueue work in libuv, is to do it in a thread
03:09:47  <bnoordhuis>or alternatively poll kevent() regularly...
03:34:13  * dmkbot1quit (Remote host closed the connection)
03:34:17  * dmkbotjoined
03:40:35  * dmkbotquit (Remote host closed the connection)
03:40:39  * dmkbotjoined
04:19:25  <CIA-53>node: Ben Noordhuis v0.4 * r9ef962f / src/node_crypto.cc :
04:19:25  <CIA-53>node: crypto: fix error message buffer overrun
04:19:25  <CIA-53>node: ERR_error_string() expects a buffer of at least 256 bytes, the input buffer
04:19:25  <CIA-53>node: was not even half that size. Use ERR_error_string_n() instead. - http://git.io/-Z750w
04:19:26  <CIA-53>node: Ben Noordhuis v0.4 * r3a9b08f / src/node_crypto.cc : crypto: look up SSL errors with ERR_print_errors() - http://git.io/r93ywA
04:19:26  <CIA-53>node: Ben Noordhuis v0.4 * rfdb0eb5 / src/node_crypto.cc : crypto: fix read of potentially uninitialized variable - http://git.io/ADtIEw
04:19:27  <CIA-53>node: Ben Noordhuis v0.4 * r93aad55 / src/node_crypto.cc : crypto: fix delete of potentially uninitialized pointer - http://git.io/mw-lEA
04:54:00  * indutnyquit (Remote host closed the connection)
05:19:56  * dmkbotquit (Remote host closed the connection)
05:20:00  * dmkbotjoined
06:03:48  * piscisaureusjoined
06:39:38  * bnoordhuisquit (Ping timeout: 260 seconds)
06:48:17  * dmkbotquit (Remote host closed the connection)
06:48:23  * dmkbotjoined
06:56:13  * dmkbotquit (Remote host closed the connection)
06:56:33  * dmkbotjoined
06:56:45  * dmkbotquit (Remote host closed the connection)
06:56:55  * mralephjoined
06:59:14  * ericktjoined
07:05:30  * dmkbotjoined
07:05:42  * dmkbotquit (Remote host closed the connection)
07:10:53  * dmkbotjoined
07:11:05  * dmkbotquit (Remote host closed the connection)
07:12:45  * ericktquit (Quit: erickt)
07:23:02  * dmkbotjoined
07:23:29  * dmkbotquit (Remote host closed the connection)
07:24:09  * dmkbotjoined
07:33:05  * mralephquit (Quit: Leaving.)
07:58:15  * piscisaureus_joined
08:00:09  * piscisaureus_quit (Client Quit)
08:03:05  * piscisaureus_joined
08:06:37  * indutnyjoined
08:06:47  <indutny>hy everyone!
08:28:18  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
08:29:09  * piscisaureus_joined
08:29:27  * piscisaureus_quit (Client Quit)
08:31:02  * piscisaureus_joined
08:36:46  <CIA-53>libuv: Bert Belder prefork * r84e969e / (prefork.c src/win/tcp.c): prefork III - http://git.io/T0Rm1A
08:38:15  <piscisaureus_>hi indutny
08:38:50  * piscisaureus_changed nick to piscisaureus2
08:59:06  <dmkbot>joyent/node: jbergstroem: running test GH-814 sometimes leaves trailing processes - https://github.com/joyent/node/issues/914
09:38:09  <piscisaureus2>bnoordhuis: yt?
13:41:37  * ericktjoined
13:56:55  <indutny>ryah: yt?
13:57:00  <indutny>good morning
13:59:52  * bnoordhuisjoined
14:14:42  <piscisaureus2>ah, bnoordhuis just woke up :-)
14:15:01  <bnoordhuis>rise and shine!
14:47:08  <bnoordhuis>kqueue is turning out to be a real pain
14:47:15  <piscisaureus2>bnoordhuis: why?
14:47:24  <piscisaureus2>bnoordhuis: just chop up libev
14:47:43  <bnoordhuis>piscisaureus: i don't want to do that, that's going to suck when upgrading
14:48:03  <piscisaureus2>bnoordhuis: meh, we'll never upgrade again
14:48:17  <piscisaureus2>bnoordhuis: libev will be replaced by march 2012
14:48:25  <bnoordhuis>hmm, you think so?
14:48:35  <piscisaureus2>ok - it's probably better to talk to ryah
14:48:48  <indutny>2012 - end of libev
14:57:55  * mikealjoined
14:57:59  <mikeal>dammit ryah
14:58:05  <mikeal>now that you've put the idea of domains in my head
14:58:11  <mikeal>i can't stop seeing all these places that I need it :)
15:04:23  * ericktquit (Quit: erickt)
15:19:15  * piscisaureus2quit (Read error: Connection reset by peer)
15:19:23  * piscisaureus_joined
15:23:09  * ericktjoined
15:26:09  * ericktquit (Client Quit)
15:27:12  * dmkbotquit (Remote host closed the connection)
15:27:17  * dmkbotjoined
15:32:43  * piscisaureus_quit (Quit: piscisaureus_)
15:35:31  * piscisaureus_joined
15:36:53  * ericktjoined
15:38:19  <CIA-53>libuv: Marek Jelen master * rd9fb84c / config-unix.mk : build: absorb CFLAGS from environment - http://git.io/db0Ocw
15:48:34  * ericktquit (Quit: erickt)
15:51:20  * mikealquit (Quit: Leaving.)
15:59:49  * dmkbotquit (Remote host closed the connection)
15:59:54  * dmkbotjoined
16:02:29  * dmkbotquit (Remote host closed the connection)
16:03:47  <ryah>indutny: good morning
16:06:02  * isaacsjoined
16:10:48  <piscisaureus_>isaacs: hey
16:10:50  <indutny>ryah: https://github.com/indutny/node/compare/master...feature-debugger-round-5#L3L-1
16:10:59  <isaacs>hi
16:11:07  <indutny>ryah: work is in-progress, but can I ask you to review this
16:11:08  <indutny>?
16:11:25  <piscisaureus_>isaacs: I have 2 questions.
16:11:25  <piscisaureus_>first, when a package is unpublished on npm, is it really gone?
16:12:12  <piscisaureus_>isaacs: also, a user was complaining about all npm user passwords being exposed (http://img593.imageshack.us/img593/8964/npmw.png)
16:13:12  <isaacs>piscisaureus: i'm talking to TheJH now
16:13:31  <isaacs>piscisaureus: the passwords aren't exposed. teh salted hashes are. don't use asingle dictionary word as your password
16:14:17  <isaacs>piscisaureus: and yes-ish. you *can* go through the old revisions of the document after it's been deleted, but if the db gets compacted, it's actually really truly gone
16:14:38  <piscisaureus_>isaacs: can I call you on skype?
16:15:06  <isaacs>sure
16:15:16  <isaacs>isaac.schlueter
16:22:44  <ryah>bnoordhuis: no extra thread
16:23:00  <bnoordhuis>ryah: so now what?
16:23:19  <ryah>bnoordhuis: hack libev
16:23:35  <bnoordhuis>we're going to drop libev eventually?
16:23:49  <ryah>there's no plan for that
16:24:14  * dmkbotjoined
16:24:38  <bnoordhuis>okay, when needs must and all that
16:24:42  <bnoordhuis>but first... dinner!
16:26:34  <piscisaureus_>http://twitter.com/#!/nodenpm/status/116800617971843072
16:26:39  <piscisaureus_>^-- isaacs
16:31:47  * dmkbotquit (Remote host closed the connection)
16:37:55  <ryah>https://gist.github.com/1233593 <-- are we happy with this?
16:38:01  <ryah>i think i am
16:38:25  <ryah>indutny: let me get back to you in a few hours, i have some meetings
16:38:32  <ryah>indutny: glad you're working on this though!
16:38:45  <indutny>ryah: nice, I think I'll be up for 1-2 hour
16:38:56  <indutny>ryah: so if I'll be afk - feel free to reach me via email
16:39:35  <ryah>indutny: can you squash the commit and submit a pull request?
16:39:43  <ryah>indutny: it's easier for me to review
16:39:55  <indutny>ryah: squash all commits?
16:40:13  <ryah>indutny: the test commits
16:40:21  <indutny>ryah: got it
16:43:44  * dmkbotjoined
16:48:17  * dmkbotquit (Remote host closed the connection)
16:59:59  <indutny>ryah: https://github.com/joyent/node/pull/1754
17:00:12  <indutny>1 commit, and only important things
17:00:26  <indutny>ryah: all other fixes left in feature-debugger-round-5
17:00:42  <indutny>it's rebased too
17:09:22  * ericktjoined
17:09:26  * piscisaureus_taking the train. see you guys later
17:10:31  <bnoordhuis>fyi, mees brushed her tooth (singular) for the first time today!
17:10:42  <bnoordhuis>kids <3
17:13:21  <indutny>:)
17:13:24  <indutny>nice
17:14:53  * piscisaureus_quit (Ping timeout: 260 seconds)
17:20:27  <indutny>rebasing is just awesome!
17:27:35  * brsonjoined
17:37:12  <CIA-53>libuv: Igor Zinkovsky master * r019e6ed / src/win/fs-event.c : windows: NextEntryOffset is relative, not absolute - http://git.io/Tq6xNQ
17:52:59  <bnoordhuis>ryah piscisaureus igorzi: scrum call in 10, right?
17:54:15  <ryah>bnoordhuis: yes
17:57:24  <indutny>ryah: does that pull request looks good, or should I rebase anything else?
18:01:01  <bnoordhuis>piscisaureus: pick up your phone
18:02:55  * Bert0963joined
18:03:20  <Bert0963>sorry - i am on a train
18:07:04  <bnoordhuis>better that you're on a train than the other way aroun
18:10:23  * Bert0963quit (Ping timeout: 248 seconds)
18:15:41  * piscisaureuzjoined
18:31:05  <igorzi>bnoordhuis: ndjs-ubt5 is up
18:31:14  <bnoordhuis>igorzi: cool, thanks
18:31:43  <piscisaureuz>Moar Unix benchmarks?
18:32:06  <piscisaureuz>Trying to best Windows on prefork?
18:50:15  <ryah>"Currently the server handles between 10.000 and 15.000 concurrent WebSocket connections, about 350.000 clients per day, without any problems using 100-300 MB RAM."
18:50:19  <ryah>Currently the server handles between 10.000 and 15.000 concurrent WebSocket connections, about 350.000 clients per day, without any problems using 100-300 MB RAM.
18:50:25  <ryah>http://siriux.net/2011/09/running-a-large-websocket-server/
18:50:30  <ryah>\o/
18:57:42  <ryah>we need to do a blocking stdout
19:25:42  * mralephjoined
19:28:27  <piscisaureus>sorry for not attending the call - I forgot that we moved it
19:28:43  <piscisaureus>anything particular came up that I should know about?
19:43:48  <igorzi>ryah bnoordhuis piscisaureus: to indicate overflow for fs watcher, do you want to add UV_OVERFLOW to uv_fs_event enum? or return status = -1, and set uv_last_error to UV_OVERFLOW (or something like that)?
19:44:10  <bnoordhuis>igorzi: status=-1
19:46:41  <piscisaureus>bnoordhuis: and errno == ?
19:46:49  <piscisaureus>status == -1 implies that errno is set
19:47:01  <bnoordhuis>nuh uh
19:47:27  <bnoordhuis>you windows guys don't even have a useful errno, do you?
19:47:45  <bnoordhuis>anyway, consumers of libuv should assume nothing about errno
19:54:44  <ryah>i kind of like having a uv_fs_event for it
19:54:52  <ryah>makes it seem more normal
20:01:54  <ryah>mraleph: what happened to --print_code ?
20:02:24  <mraleph>ryah: nothing what do you mean?
20:02:39  <ryah>mraleph: % node --code_comments --print_code x.js
20:02:40  <ryah>Error: unrecognized flag --print_code
20:02:40  <ryah>Try --help for options
20:02:40  <ryah>-2
20:03:06  <mraleph>if it's release build you have to build with objectprint=on
20:03:14  <ryah>oh, ok
20:03:52  <igorzi>ryah bnoordhuis piscisaureus: i think i like having a special uv_fs_event for it too.. this way OSes that just report that something has changed could use that same uv_fs_event
20:05:11  <ryah>igorzi: yeah it would be weird for kqueue to only report "errors"
20:05:23  <ryah>maybe we should call it UV_FS_EVENT_UNKNOWN
20:05:24  <ryah>or something
20:05:34  <ryah>something happend but we don't know what
20:06:07  <ryah>UV_UNKNOWN_EVENT
20:06:40  <ryah>UV_CHANGE ?
20:10:23  <igorzi>we already have UV_CHANGE.. we could just use that, but without the filename?
20:13:46  * brsonquit (Ping timeout: 255 seconds)
20:18:55  * brsonjoined
20:20:08  <piscisaureus>bnoordhuis: I meant uv_last_err, not libc's errno
20:20:39  <bnoordhuis>piscisaureus: that should be set to a meaningful error
20:20:55  <piscisaureus>bnoordhuis: yes, whenever -1 is returned
20:27:28  <ryah>igorzi: yeah, i like that
20:44:36  * piscisaureuzquit (Quit: Bye)
21:10:21  <bnoordhuis>http://www.hp.com/hpinfo/newsroom/press/2011/110922xb.html <- is webos coming back? stay tuned!
21:16:35  <erickt>hi #libuv. Right now, uv_getaddrinfo automatically cleans up the addrinfo linked list. This means that we're forced to copy the list again if we want to use the addrinfos outside of the callback function (such passing the list to another thread for processing). What do you think about exposing freeaddrinfo to the libuv users?
21:18:52  <ryah>erickt: yes, please
21:18:57  <ryah>i also don't like the behavior we have now
21:19:19  <erickt>how so?
21:20:19  <ryah>erickt: by adding uv_freeaddrinfo ?
21:23:46  <erickt>oh. I thought you meant you didn't like the uv_getaddrinfo api
21:30:33  <erickt>Also, libuv isn't compiling on macs anymore. Looks like the fs monitoring api hasn't been implemented for darwin yet
21:32:14  <ryah>erickt: bnoordhuis is working on it - it should be stubbed out though.
21:33:11  <bnoordhuis>stubs, i'll add them
21:40:46  <bnoordhuis>erickt: feel like being a guinea pig?
21:41:39  <erickt>sure
21:44:18  * mralephquit (Quit: Leaving.)
21:44:31  <bnoordhuis>erickt: https://github.com/bnoordhuis/libuv/commit/e3767aa
21:46:54  <erickt>bnoordhuis: errored out with: ../include/uv.h:1026: error: expected specifier-qualifier-list before ‘UV_FS_EVENT_PRIVATE_FIELDS’
21:47:05  <bnoordhuis>gah, right
21:47:06  <bnoordhuis>1 sec
21:49:36  <bnoordhuis>erickt: take two: https://github.com/bnoordhuis/libuv/commit/71cf1aa
21:51:40  <erickt>that worked, and the tests failed with sensible failures: https://gist.github.com/1236152
21:52:32  <erickt>the test-pipe-bind-error.c failure has been happening for some time now, so that's not new
21:53:39  * mralephjoined
21:53:56  <bnoordhuis>test-pipe-bind-error is the bane of my existence :(
21:54:44  <bnoordhuis>thanks for testing btw, erickt
21:55:20  <erickt>glad to help
21:56:28  <CIA-53>libuv: Ben Noordhuis master * r3368d6c / (6 files in 2 dirs):
21:56:28  <CIA-53>libuv: unix: stub file watcher implementation
21:56:28  <CIA-53>libuv: The file watcher API has not been implemented on all Unices yet.
21:56:28  <CIA-53>libuv: Provide stubs on those platforms so libuv at least compiles. - http://git.io/S1HdLA
22:09:20  * mralephquit (Quit: Leaving.)
22:09:33  <CIA-53>libuv: Igor Zinkovsky master * rf9ad0a7 / src/win/fs-event.c : windows: fire UV_CHANGE in case of overflow - http://git.io/aFLEeA
22:14:39  <erickt>ryah: got the uv_freeaddrinfo implemented here: https://github.com/joyent/libuv/pull/196
22:38:34  * isaacsquit (Quit: isaacs)
23:06:59  <piscisaureus>meh
23:07:12  <piscisaureus>translating windows keys into vt100 is so painful
23:19:43  <piscisaureus>ryah: does v8 properly handle non-bmp utf8 characters, e.g. those that should be encoded as surrogate pairs?
23:20:05  * isaacsjoined
23:40:48  <bnoordhuis>piscisaureus: no
23:41:04  <piscisaureus>why is that?
23:41:17  <bnoordhuis>because they filter out everything that's outside the bmp
23:41:31  <bnoordhuis>and replace it with 0xfffe, if i'm not mistaken
23:41:34  <piscisaureus>that's fucked up
23:41:56  <bnoordhuis>piscisaureus: http://code.google.com/p/v8/issues/detail?id=761
23:42:00  <igorzi>ryah: https://github.com/igorzi/node/commit/e86b443cd27ff986b1edf031476c7b790becb8ca
23:42:13  <igorzi>^---- file watcher binding
23:45:59  <piscisaureus>bnoordhuis: it's a bad excuse - I don't care what v8's parser does, but it would be nice if String->Utf8Write would correctly generate non-bmp characters from surrogate pairs
23:46:36  <bnoordhuis>piscisaureus: world peace would be nice too but i don't think it's going to happen anytime soon :/
23:46:37  <piscisaureus>and vice versa
23:47:17  <piscisaureus>bnoordhuis: I have magic karma: http://code.google.com/p/v8/issues/detail?id=847 <-- got resolved too :-)
23:48:43  <bnoordhuis>piscisaureus: you can always try, won't hurt
23:49:07  <bnoordhuis>i think they're resistant to change because it's pretty much tied in with their unicode library, unibrow
23:49:21  <bnoordhuis>that's a really bad nerd joke btw, unibrow...
23:49:31  <piscisaureus>har har
23:49:52  <bnoordhuis>but hey, i didn't make it up
23:51:05  <piscisaureus>bnoordhuis: string encoding needs some love anyway -> http://code.google.com/p/v8/issues/detail?id=1665&q=cons%20string&colspec=ID%20Type%20Status%20Priority%20Owner%20Summary%20HW%20OS%20Area%20Stars
23:51:14  <piscisaureus>er, http://code.google.com/p/v8/issues/detail?id=1665
23:51:45  <bnoordhuis>yes
23:51:47  <bnoordhuis>we should send a patch
23:51:59  <piscisaureus>omg when I wrote that I must have been very tired
23:52:03  <piscisaureus>look at the language :-(
23:52:07  <piscisaureus>I am making a fool of myself
23:52:21  <bnoordhuis>no worse than usual, don't worry
23:53:21  <piscisaureus>node development ~ slavery
23:54:02  <bnoordhuis>that reminds me... http://beta.uitzendinggemist.nl/programmas/4147-de-slavernij <- worth watching
23:54:09  <bnoordhuis>there's also a junior edition :)
23:54:54  <bnoordhuis>fun fact, daphne bunskoek is the best friend of a good friend of mine
23:55:01  <bnoordhuis>not that i've ever met her
23:55:06  <bnoordhuis>but she's from almelo
23:55:33  <piscisaureus>I am not sure if I like her
23:56:00  <piscisaureus>she looks kinda cute but then I remember she's in rtl boulevard
23:56:13  <bnoordhuis>she is?
23:56:13  <piscisaureus>omg I want to watch rbl broznovo again
23:56:19  <bnoordhuis>it's over daphne!
23:58:11  <piscisaureus>bnoordhuis: http://www.youtube.com/watch?v=6YZWKvB0Abw&feature=related
23:59:03  <piscisaureus>nee doe mij dan toch maar leika vanlexmondski
23:59:14  <bnoordhuis>yolanthe *melts*
23:59:39  <bnoordhuis>or rather, yolanthski
23:59:55  <piscisaureus>Snolski Troski