01:39:55  * brsonquit (Read error: Operation timed out)
02:06:23  <dmkbot>joyent/node: tellnes: fix dns.resolveTxt - https://github.com/joyent/node/issues/1817
02:09:02  <ryah>dmkbot is back \o/
02:09:25  <ryah>pquerna: but.. why?
02:09:37  <ryah>also virgo? what's that
02:58:59  <isaacs>hey windows people! try this out: Don't use Cygwin. Use Windows native. http://npmjs.org/doc/README.html#Installing-on-Windows-Experimental
03:17:41  <ryah>isaacs++
03:17:46  <ryah>isaacs: now we need to make that one step
03:17:52  <isaacs>yeah, srsly
03:18:29  <isaacs>also, there's still some polishing to be done.
03:18:37  <isaacs>the whole "bin" prefix thing, it's just a bit odd.
03:18:45  <isaacs>i mean, mkes sense on posix, but odd for windows
03:20:20  <ryah>isaacs: what's the state of tar.js?
03:20:40  <ryah>would be good to use that now.
03:20:50  <isaacs>it's not good.
03:20:57  <isaacs>but i'm going to attack that very soon.
03:21:20  <isaacs>bundling a tar exe is highly suboptimal
03:21:22  <ryah>is there some bsd-licensed tar.c that we can drop into node?
03:21:30  <isaacs>yeah, there's bsdtar.
03:21:36  <isaacs>but it's silly. you shouldn't do that in node.
03:21:41  <ryah>(we would not make this a public interface)
03:21:47  <isaacs>tar is not a core internet protocol
03:22:04  <isaacs>and bsdtar is not as good as a tar can be.
03:22:15  <ryah>agreed, but the question is how we can get npm off the tar.exe as quickly as possible
03:22:45  <isaacs>the headaches involved in bundling bsd's tar.c in node are, i'm guessing, higher than those in just writing tar in javascript.
03:22:55  <ryah>okay
03:23:02  <isaacs>it's not a particularly difficult protocol, once you catalog all the edge cases (which i've spent most of the last 2 years doing)
03:23:22  <ryah>just wanted to put that on the table
03:23:31  <isaacs>i appreciate your support :)
03:32:38  <pquerna>libarchive
03:32:58  <pquerna>http://code.google.com/p/libarchive/
03:33:20  <pquerna>(but its still big, anyways)
03:33:47  <pquerna>ryah: virgo = c+lua agent for collecting monitoring things (incomplete, just startedish)
03:34:11  <pquerna>(well, and a port of an.. older thing... to use uv)
04:16:52  * isaacsquit (Quit: isaacs)
04:24:14  * isaacsjoined
04:29:18  * bradleymeckjoined
05:08:09  * ericktjoined
05:15:10  * isaacsquit (Quit: isaacs)
05:21:37  * ericktquit (Quit: erickt)
05:34:31  * ericktjoined
05:44:21  * ericktquit (Quit: erickt)
05:48:30  * bradleymeckquit (Quit: Leaving)
06:33:06  * isaacsjoined
08:18:37  * isaacsquit (Quit: isaacs)
08:56:50  * felixgejoined
08:56:50  * felixgequit (Changing host)
08:56:51  * felixgejoined
11:19:23  * bnoordhuisjoined
11:41:21  <CIA-53>node: Ben Noordhuis master * rf1ca485 / src/node_zlib.cc : zlib: fix assert message - http://git.io/5_Zypg
11:56:23  <dmkbot>joyent/node: tellnes: fix dns.resolveTxt - https://github.com/joyent/node/issues/1817
14:00:48  * piscisaureusjoined
14:10:55  * piscisaureusquit (Ping timeout: 248 seconds)
14:29:16  * ericktjoined
14:50:47  <CIA-53>node: Christian Tellnes master * rd9c67ae / src/cares_wrap.cc : dns: implement resolveTxt() - http://git.io/ReDp3g
14:50:47  <CIA-53>node: Ben Noordhuis master * r976c6b0 / test/internet/test-dns.js : test: add TXT DNS record lookup test - http://git.io/cxrrQQ
14:58:26  * bradleymeckjoined
15:00:45  <indutny>woot
15:01:02  <bradleymeck>?
15:08:12  <indutny>bradleymeck: resolveTxt
15:12:13  <bradleymeck>so we can do domain name verifications using that reliably now?
15:12:45  <indutny>looks like so
15:28:07  * luxigoquit (Read error: Connection reset by peer)
15:59:50  * Casanjoined
15:59:57  <Casan>#freebsd
16:14:58  * isaacsjoined
16:26:32  * piscisaureusjoined
16:30:31  * isaacsquit (Quit: isaacs)
16:33:03  * pieternjoined
16:33:11  <pietern>piscisaureus: hi bert
16:33:26  <piscisaureus>pietern: hey pieter
16:33:33  <indutny>pquerna: big kudos to you
16:33:36  <indutny>:)
16:33:38  <pietern>piscisaureus: just dropped in real quick to let you know the cla stuff is under review now
16:33:41  <indutny>love your work, really
16:34:02  <indutny>piscisaureus: hello bert! :)
16:34:11  <pietern>piscisaureus: and it's looking good from what I can see
16:34:13  <piscisaureus>pietern: oh good - we're needing it :)
16:34:26  <pietern>piscisaureus: I hope it isn't stalling other patches that may depend on it or something
16:34:53  <piscisaureus>pietern: not really - but we're just leaving a bad bug unfixed
16:35:14  <pietern>piscisaureus: so you found out it definitely did fix that socket.io node bug?
16:35:38  <pietern>or it was simply not reproducible with the patch applied.. which is effectively the same of course.. ;)
16:35:40  <piscisaureus>pietern: well... I've not seen so many issues with socketio at all honestly
16:36:14  <piscisaureus>pietern: but I am pretty sure that it fixes the failing assert in net.js
16:36:22  <piscisaureus>like, 99% :-)
16:36:32  <pietern>ah, ok, I think that's the one I'm thinking of then
16:36:36  <piscisaureus>pietern: we actually landed a test for the bug that your patch fixes in node
16:36:53  <pietern>piscisaureus: oh great, consistently reproducible?
16:37:48  <pietern>that's pretty hard with these scheduling-sensitive things
16:37:58  <pietern>or at least, usually hard..
16:38:04  <piscisaureus>pnoordhuis: yes - https://github.com/joyent/node/blob/master/test/simple/test-regress-GH-1697.js
16:40:04  <piscisaureus>pietern: so you are moving to palo alto?
16:40:09  <piscisaureus>or is that just temporary?
16:40:13  <pietern>piscisaureus: yes,
16:40:31  <pietern>piscisaureus: no it's pretty permanent
16:40:54  <pietern>as long as it's fun ;)
16:41:57  <piscisaureus>pietern: so you finished to finish school? Or you just decided to leave it?
16:42:11  <piscisaureus>indutny: hey fedor
16:42:30  <piscisaureus>indutny: so are you going to work on libuv all the time?
16:43:14  <pietern>piscisaureus: almost done, the degree itself will probably get issued when I'm away
16:43:32  <piscisaureus>ah, good
16:43:40  <piscisaureus>heh
16:43:43  <pietern>piscisaureus: definitely looking forward to leaving univ. behind me ;)
16:43:50  <pietern>I think we share that
16:44:05  <piscisaureus>:-) very much
16:44:08  <pietern>piscisaureus: hehe
16:44:19  <pietern>piscisaureus: anyway, I'll keep you posted on any progress
16:44:42  <pietern>if it is going to take too long, we'll find a solution
16:44:47  <pietern>talk to you later!
16:44:57  <piscisaureus>pietern: yeah - ttyl
16:45:00  * pieternquit (Quit: pietern)
16:56:15  * bradleymeckquit (Ping timeout: 248 seconds)
16:57:32  * ericktquit (Quit: erickt)
16:57:37  <indutny>piscisaureus: on anything related to node
16:58:02  <indutny>piscisaureus: I'll try to do my best with that open pull requests
16:58:22  <indutny>and then I'll be up for anything else
17:01:47  <piscisaureus>indutny: cool
17:02:09  <piscisaureus>indutny: I didn't really realize that you were really funded to do (almost) wheter you want
17:02:18  <piscisaureus>until I was talking to marak
17:02:26  <indutny>heh
17:02:50  <indutny>contribution to libuv and nodejs is cool anyway
17:02:59  <indutny>even without funding
17:03:18  <indutny>but thanks to nodejitsu I've more time to work on this things
17:03:54  <piscisaureus>indutny: cool, sure, but you wouldn't be able to actually work on it without being funded :-)
17:04:07  <indutny>yes! exactly :)
17:04:42  <indutny>love you guys
17:10:43  * ericktjoined
17:19:25  * bradleymeckjoined
17:22:20  <bradleymeck>piscisaureus i got windows modules mostly working, except for one thing, cannot figure out how to work around "C:\Users\bradley\Documents\node\src\node_object_wrap.h(56): warning C4251: 'node::ObjectWrap::handle_' : class 'v8::Persistent<T>' needs to have dll-interface to be used by clients of class 'node::ObjectWrap' [C:\Users\bradley\Documents\node\node.vcxproj]" , do you know how to make templated classes have accessible m
17:22:20  <bradleymeck>embers?
17:26:39  * piscisaureusquit (Ping timeout: 248 seconds)
17:35:40  <ryah>good morning
17:36:51  * isaacsjoined
17:37:12  <igorzi>morning
17:37:25  * pquernaquit (Ping timeout: 240 seconds)
17:38:31  * pquernajoined
17:39:37  * mralephjoined
17:40:15  <igorzi>ryah: bnoordhuis: could we do the call 1h later today?
17:40:21  <bnoordhuis>igorzi: sure
17:41:03  * bradleymeckquit (Ping timeout: 248 seconds)
17:41:51  <ryah>igorzi: yes
17:49:06  * Casanquit (Ping timeout: 256 seconds)
17:52:05  * felixgequit (Quit: felixge)
17:56:41  * pquernaquit (Changing host)
17:56:41  * pquernajoined
18:19:18  * aaaaaandjbdgjbfjjoined
18:21:47  * aaaaaandjbdgjbfjquit (Client Quit)
18:31:00  * piscisaureusjoined
18:31:29  * brsonjoined
18:43:04  * bradleymeckjoined
18:56:37  * piscisaureusquit (Read error: Connection reset by peer)
18:58:02  <ryah>mraleph: what's the status on newgc?
18:58:14  <ryah>mraleph: i've noticed your twice weekly releases have stopped
18:58:27  <mraleph>ryah: we are stabilizing it.
18:58:46  <mraleph>ryah: should be green now
18:58:59  <mraleph>ryah: soon will push to trunk
19:01:34  <ryah>mraleph: this week?
19:01:52  * piscisaureusjoined
19:02:14  <piscisaureus>omg guys - stop doing fibonacci please
19:02:32  <bradleymeck>mraleph is there a way to force v8_export to be set, even when building static for node.exe in windows
19:02:56  <mraleph>hmm. not sure
19:04:45  <ryah>piscisaureus, bnoordhuis: call?
19:04:54  <piscisaureus>bradleymeck: I haven't been able to check out the dll/template stuff yet
19:05:03  <piscisaureus>ryah: I am at the airport
19:05:11  <ryah>piscisaureus: oh okay
19:05:18  <bradleymeck>np piscisaureus ive got that on hold during work hours
19:05:31  <piscisaureus>you can try but the germans are so annoying
19:05:56  <piscisaureus>literally every 5 seconds there's an announcement and I can't escape it
19:06:54  <bnoordhuis>achtung bitte!
19:07:30  <piscisaureus>it's like they forgot to do announcements today
19:07:36  <piscisaureus>and they're now processing backlogs
19:23:25  * piscisaureusquit (Ping timeout: 258 seconds)
19:29:05  * bradleymeckquit (Quit: Leaving)
19:32:39  * Casanjoined
19:37:28  * erickt_joined
19:51:45  <CIA-53>libuv: Igor Zinkovsky ipc * r2d826fe / (test/test-ipc.c test/test-spawn.c uv.gyp): make test-ipc accept the pending tcp server - http://git.io/TXuz2Q
20:02:47  <ryah>igorzi: thanks
20:05:01  <ryah>bnoordhuis: we should do a blog post for v0.5.9 about how it's faster than v0.4 :)
20:05:26  <bnoordhuis>ryah: sure
20:05:35  <bnoordhuis>we'll need some hard numbers to back it up
20:06:13  <bnoordhuis>but i can type up a nice "putting node on steroids" piece of infotainment
20:06:20  <ryah>i think we should also do a blog post for v0.5.10 about fs.watch()
20:06:37  <ryah>we can start doing some work to convince people to port thier libs
20:06:57  <bnoordhuis>ryah: btw, are you still going to do the sunos port of fs.watch()?
20:07:17  <ryah>bnoordhuis: yes, once i finish this ipc stuff
20:12:14  <igorzi>ryah: i could finish the ipc test if you want.. lmk
20:15:16  <ryah>igorzi: you're quite welcome to - im still stuck on some spinning bug when i send a handle
20:17:53  <indutny>:)
20:18:43  <indutny>I'll write - I'll translate
20:18:48  <indutny>and post on russian blogs
20:19:08  <igorzi>ryah: ok.. we need a way to get both processes to accept a connection. i'm thinking of spinning up a client and keep making connections until both servers accept a connection.
20:19:11  <indutny>s/I'll write/You'll write/
20:20:48  <ryah>igorzi: after each accept, the servers can close their incoming connections
20:21:32  <ryah>might be less painful to just send a bunch though..
20:25:46  <igorzi>ryah: yep, i'll play with it. just sending a bunch of connections might make the test fail sometimes (if all connnections get accepted by only one server)
20:36:23  <dmkbot>joyent/node: egorFiNE: reproducible crash with 0.5.7 in fs.c on OS X 10.7.2 - https://github.com/joyent/node/issues/1739
20:39:14  <ryah>someone needs to get OSX 10.7 so we can fix these bugs--^
20:43:59  <bnoordhuis>ryah: why don't you upgrade?
20:47:11  <ryah>my boot loader is fucked up
20:47:31  <ryah>slap on the wrist from apple for having installed linux
20:48:13  <bnoordhuis>can't have that, of course, frivolously installing competing operating systems
20:50:29  <rmustacc>ryah: You can probably just get a second laptop and upgrade that...
20:51:41  * erickt_quit (Read error: Operation timed out)
20:53:07  * erickt_joined
20:53:18  * raggi_quit (Read error: Operation timed out)
20:53:55  * raggijoined
20:55:05  <bnoordhuis>i'm going to pick up a macbook this week, i'll make sure it comes with 10.7
21:15:01  <ryah>rmustacc: do we have extras?
21:15:23  <ryah>nah, i think i'll just reintall this computer from scratch
21:15:30  <ryah>there's a lot of shit on here anyway
21:16:32  * felixgejoined
21:16:32  * felixgequit (Changing host)
21:16:32  * felixgejoined
21:17:48  * mralephquit (Quit: Leaving.)
21:19:48  <rmustacc>ryah: Yes
21:42:57  <pquerna>"End of Sale for Windows 2003 OS" <- yaaaaaay
21:46:57  <bnoordhuis>ryah: https://github.com/bnoordhuis/libuv/compare/kqueue <- review?
21:49:35  <CIA-53>libuv: Ryan Dahl ipc * r7de7ff3 / (src/unix/stream.c test/test-ipc.c): unix: handle passing kind of working - http://git.io/MR--Kg
21:51:56  <bnoordhuis>yay, kernel.org is back up
21:52:22  <ryah>not for me :/
21:52:54  <bnoordhuis>ryah:
21:52:56  * indexzerojoined
21:57:53  <bnoordhuis>igorzi: https://github.com/bnoordhuis/libuv/compare/kqueue <- can you review the changes to the fs watcher test?
21:58:53  <bnoordhuis>fs_event_watch_file_current_dir modified the watched file before the event loop was started so i put that behind a timer
22:00:33  * pquernaquit (Ping timeout: 258 seconds)
22:01:25  <igorzi>bnoordhuis: lgtm
22:01:32  <bnoordhuis>igorzi: thanks
23:03:10  * ericktquit (Quit: erickt)
23:03:11  * erickt_changed nick to erickt
23:10:22  <ryah>http://dbus.freedesktop.org/doc/api/html/group__DBusConnection.html#gab3cbc68eec427e9ce1783b25d44fe93c
23:10:25  <ryah>^-- sloppy
23:12:11  * erickt_joined
23:13:24  * felixgequit (Quit: http://www.debuggable.com/)
23:19:18  * pquernajoined
23:31:18  * piscisaureusjoined
23:31:54  <piscisaureus>I'm back
23:31:58  <piscisaureus>finally
23:32:07  <piscisaureus>my travel scheduling can use some tweaking
23:32:38  <piscisaureus>Tomorrow I will be on the call again :-)
23:33:35  <piscisaureus>I have 2 questions -
23:33:35  <piscisaureus>How is IPC working out?
23:33:35  <piscisaureus>What is the directory watcher api in node going to look like?
23:33:48  <piscisaureus>(because I have some fears about that)
23:35:16  <piscisaureus>ryah igorzi: ^
23:40:11  <igorzi>piscisaureus: ipc is going...
23:40:17  <igorzi>piscisaureus: https://github.com/joyent/node/blob/master/lib/fs.js#L632
23:41:00  <piscisaureus>hey, indexzero has hijacked i<tab>
23:41:39  <indexzero>piscisaureus: My ridiculously circuitous plan is nearly one-quarter complete
23:42:15  <piscisaureus>indexzero: oh, you're going to hyjack everything?
23:42:43  <piscisaureus>igorzi: the problem with using fs.watch for both directories and files is that the trick people use now to watch entire trees is going to collapse
23:43:02  <piscisaureus>because they would now iterate over all the "stuff" in a directory and watch everything
23:43:33  <piscisaureus>but now, when one of this "stuff" is suddenly a directory you are watching the contents of that and not the entry itself
23:44:57  <igorzi>piscisaureus: i don't think the intention was to make fs.watchFile directly mappable to fs.watch. it's a different api, so people would have to rewrite their apps.
23:45:07  <piscisaureus>ah - ok
23:46:30  <piscisaureus>So we're also retaining fs.watchFile?
23:46:34  <piscisaureus>that seems odd
23:46:39  <piscisaureus>I'd like to see it go
23:46:59  <piscisaureus>But maybe we should get some feedback on the new api first
23:49:33  <igorzi>piscisaureus: i think at some point you mentioned that you always have pAcceptEx == pAcceptEx6, correct?
23:49:49  <piscisaureus>igorzi: yes, stimmt
23:50:13  <piscisaureus>igorzi: you said the same right?
23:50:46  <igorzi>piscisaureus: actually, on my dev box - they're different
23:51:17  <piscisaureus>that's just weird
23:51:24  <igorzi>piscisaureus: and here's the real kicker.. they're also different between duplicated socket and the original socket
23:51:30  <igorzi>(on my dev box)
23:51:38  <piscisaureus>hmm
23:51:39  <piscisaureus>wtf
23:52:07  <piscisaureus>But I don't like to fetch these pointers every time we create a socket
23:52:19  <igorzi>basically, if i use pAcceptEx to accept a duplicated socket - I get WSAENOTSOCK
23:52:31  <piscisaureus>hah
23:52:51  <piscisaureus>igorzi: what flavor of windows are you running on your dev box?
23:53:04  <igorzi>no, i'm not fetching them every time. i'm duplicating a dummy socket uv_winsock_init, and then stashing pAcceptExDup
23:53:11  <igorzi>2k8 r2
23:53:17  <piscisaureus>igorzi: I kind of assume you have some extra LSP stacked on top winsock
23:53:25  <igorzi>maybe
23:53:29  <piscisaureus>igorzi: maybe some VPS layer, or a virus :-)
23:55:17  <igorzi>anyway, i'm glad i ran into this, because this would have been a bitch to track down if someone ran into this :)
23:55:27  <igorzi>(after the release)
23:55:31  <piscisaureus>yeah
23:55:39  <piscisaureus>it kind of sucks though
23:59:06  <piscisaureus>igorzi: but this basically means that *any* socket can have a different pointer?
23:59:19  <piscisaureus>I mean, there may be LSPs that report different pointers every time
23:59:39  <igorzi>piscisaureus: potentially, i guess
23:59:54  <piscisaureus>igorzi: also, I kind of think that in case of a duplicated socket, the pointer may be dependent on which process owns the original socket