00:08:57  <creationix>piscisaureus: how do I get the struct ev_loop* from uv?
00:11:09  <creationix>ahh, it's a private field `ev` on uv loops
00:19:17  * creationixpart
00:49:54  <pquerna>piscisaureus: hey
00:51:14  * isaacsquit (*.net *.split)
00:51:15  * ircretaryquit (*.net *.split)
00:51:15  * rmustaccquit (*.net *.split)
00:51:15  * ryahquit (*.net *.split)
00:51:16  * CoverSlidequit (*.net *.split)
00:51:17  * indutnyquit (*.net *.split)
00:55:19  * isaacsjoined
00:55:19  * CoverSlidejoined
00:55:19  * ryahjoined
00:55:19  * ircretaryjoined
00:55:19  * rmustaccjoined
00:55:19  * indutnyjoined
00:55:20  * DrPizzaquit (Excess Flood)
00:55:39  * DrPizzajoined
01:05:30  * ericktquit (Quit: erickt)
01:09:48  * piscisaureus_joined
01:10:18  <piscisaureus_>pquerna: hey
01:10:52  <piscisaureus_>pquerna: sorry, my battery died
01:13:33  <piscisaureus_>ryah: hey, what time are you at the office tomorrow
01:13:45  <piscisaureus_>ryah: and, do we need to sign up or something or do we just get in?
01:14:45  <rmustacc>piscisaureus_: If you walk like you belong, then it's fine.
01:14:55  <rmustacc>Just come into the 20th floor.
01:15:40  <piscisaureus_>rmustacc: ok, we'll try
01:15:53  <piscisaureus_>rmustacc: if it doesn't work, who should be called?
01:16:01  <rmustacc>piscisaureus_: Depends on what time you're coming in.
01:16:14  <rmustacc>I'll be there by 9:15am or so, so I could give you a number.
01:16:21  <piscisaureus_>ah, ok
01:16:27  <piscisaureus_>rmustacc: do that. just in case
01:20:08  * piscisaureus_part
01:20:19  * piscisaureus_joined
01:30:06  <isaacs>can i get a hell yeah? https://github.com/joyent/node/pull/1904
01:30:09  <isaacs>(that is, a review)
01:32:54  <piscisaureus_>isaacs: I'm too tired to concentrate, will do it tomorrow
01:32:59  <piscisaureus_>(or ben or ryah)
01:33:47  <isaacs>piscisaureus: you in ca
01:33:47  <isaacs>?
01:34:03  <piscisaureus_>yes
01:34:10  <piscisaureus_>and ben is too
01:40:29  * brsonquit (Ping timeout: 244 seconds)
01:41:52  <isaacs>piscisaureus: awesome
01:42:00  <isaacs>where you guys staying? sf?
01:42:10  <piscisaureus_>we're in the orchard hostel
01:42:15  <piscisaureus_>in sf yes
01:42:18  <isaacs>nice
01:42:44  <piscisaureus_>well, thnx joyent :-)
01:44:43  <isaacs>awesome.
01:44:55  <isaacs>i'm doing some contracting thingie on wednesday, but i'll see if i can make it to office hours
01:45:02  <isaacs>t and th, i have class in berkeley
01:49:38  <rmustacc>Off for a bit. I'll see you tomorrow.
01:51:28  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
01:54:03  * piscisaureusjoined
02:07:27  * AvianFluquit (Ping timeout: 255 seconds)
02:14:40  * isaacsquit (Quit: isaacs)
02:27:30  * AvianFlujoined
02:29:47  * ericktjoined
02:42:20  * AvianFlu_joined
02:45:30  * AvianFluquit (Ping timeout: 260 seconds)
03:01:16  * CoverSlidequit (Read error: Connection reset by peer)
03:01:16  * CoverSlidejoined
03:15:25  * piscisaureus_quit (Ping timeout: 240 seconds)
03:27:57  * AvianFlu_quit (Ping timeout: 260 seconds)
03:31:25  * AvianFlu_joined
03:43:11  * isaacsjoined
03:48:51  * AvianFlu_quit (Read error: Connection reset by peer)
04:10:38  * AvianFlujoined
04:22:27  * isaacsquit (Quit: isaacs)
05:03:50  * ericktquit (Quit: erickt)
05:58:34  * mralephjoined
07:30:36  * AvianFluquit (Ping timeout: 260 seconds)
07:47:06  * AvianFlujoined
08:13:28  * mralephquit (Quit: Leaving.)
08:15:17  * mralephjoined
08:19:43  * mralephquit (Client Quit)
08:19:57  * AvianFluquit (Ping timeout: 252 seconds)
08:32:07  * AvianFlujoined
08:36:27  * mralephjoined
08:38:05  * mralephquit (Client Quit)
09:04:05  * AvianFluquit (Ping timeout: 260 seconds)
09:15:55  * AvianFlujoined
12:28:27  * daurnimatorjoined
12:28:31  <daurnimator>hi
12:29:07  <daurnimator>I checked out libuv and am trying to compile with mingw/msys
12:29:34  <daurnimator>ran make; and it only builds uv.a; how do i make it a dll?
12:50:59  <daurnimator>and if I run the vcbuild.bat I get: File "gyp_uv", line 20 print 'Error running GYP'; SyntaxError: invalid syntax Failed to create vc project files.
14:04:05  * ericktjoined
14:30:41  * piscisaureus_joined
15:22:24  * isaacsjoined
15:39:48  * ericktquit (Quit: erickt)
15:45:57  * piscisaureus__joined
15:49:20  * piscisaureus_quit (Ping timeout: 276 seconds)
15:52:28  * piscisaureus__quit (Ping timeout: 245 seconds)
15:58:08  * brsonjoined
16:08:33  * brsonquit (Ping timeout: 252 seconds)
16:24:47  <ryah>daurnimator: we don't make DLLs right now...
16:25:14  <ryah>daurnimator: but you can edit the uv.gyp where it says "library": "dynamic" to be "library": "static"
16:43:02  * dapjoined
16:48:08  * brsonjoined
17:07:23  <igorzi>daurnimator: which version of python do you have? (i think i've seen that GYP error at some point)
17:18:07  * ericktjoined
17:20:23  * mralephjoined
17:32:53  * dwwwjoined
17:33:03  <dwww>Hi!
17:33:14  <dwww>What could be the cause of such error
17:33:15  <dwww>luvit: src/unix/core.c:213: uv__finish_close: Assertion `!(handle->flags & UV_CLOSED)' failed.
17:33:36  <dwww>a double-close attempt?
17:55:07  <igorzi>ryah: call?
17:55:17  <ryah>igorzi: yep
17:55:32  <ryah>igorzi: is it earlier today? :/
17:55:40  <ryah>it says 10:45 in my calendar
17:56:13  * piscisaureus___joined
17:56:20  <igorzi>ryah: hmm, it says 11 for me... i think Claudio is travelling today, so skype
17:56:48  <ryah>ok
17:56:58  <ryah>igorzi: skype? okay
17:57:19  <igorzi>ryah: yep
17:58:14  <ryah>igorzi: let's use my conf line +1-415-400-0604 code 0615
17:58:26  <ryah>because bert and ben are in my office
17:58:53  <igorzi>ryah: k
18:16:14  * creationix|workjoined
18:38:11  * AvianFlu_joined
18:39:42  <indutny>yeeah
18:39:47  <indutny>finally I'm online :D
18:39:55  <indutny>how is it going?
18:41:37  * AvianFluquit (Ping timeout: 260 seconds)
18:46:54  * AvianFlu_quit (Ping timeout: 255 seconds)
18:47:00  * creationix|workchanged nick to creationix
18:49:23  * AvianFlu_joined
18:55:29  <ryah>indutny: hi
18:55:54  * avian__joined
18:57:02  <indutny>where is ben?
18:58:43  * dwwwpart
18:58:53  * AvianFlu_quit (Ping timeout: 260 seconds)
19:03:53  * avian__changed nick to AvianFlu
19:09:54  * bradleymeckjoined
19:35:11  * piscisaureus___quit (Ping timeout: 248 seconds)
19:53:20  * piscisaureus___joined
19:59:47  <ryah>indutny: here
20:04:44  <indutny>ryah: bnoordhuis?
20:04:45  <indutny>can't see him
20:05:16  <ryah>indutny: he's in my office
20:05:27  <indutny>ah
20:05:32  <ryah>indutny: he's coming on irc
20:05:44  <indutny>ryah: great! :)
20:05:51  <indutny>thanks
20:07:27  * bnoordhuisjoined
20:08:13  * bradleymeckquit (Ping timeout: 240 seconds)
20:13:30  <CIA-53>node: Ryan Dahl * rd77ce4b / (4 files in 3 dirs):
20:13:30  <CIA-53>node: Fixes #1860. Remove process.writeError
20:13:30  <CIA-53>node: Breaks a few tests in "make test-message" - http://git.io/DcE35A
20:14:01  * bnoordhuispart
20:15:11  * bnoordhuisjoined
20:16:05  <bnoordhuis>indutny: looking for me?
20:17:00  * creationixpart
20:17:53  <indutny>bnoordhuis: yeah
20:18:05  <indutny>bnoordhuis: what is status of my pull reqs?
20:18:36  <bnoordhuis>indutny: the status is that i haven't done anything the last three or four days :)
20:22:37  <indutny>:)
20:22:37  <indutny>me too
20:22:38  <indutny>just asking :D
20:26:15  * piscisaureus___quit (Quit: ~ Trillian Astra - www.trillian.im ~)
20:27:41  * piscisaureus_joined
20:36:15  <ryah>indutny: they'll get merged soon
20:36:18  <ryah>we're just catching up
21:04:13  * isaacsquit (Quit: isaacs)
21:11:44  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
21:12:05  <CIA-53>libuv: Frank Denis * re8a418e / include/uv.h : Fix pasto: uv_udp_recv_start() receives a UDP datagram, it doesn't send data. - http://git.io/yROX-w
21:21:34  <bnoordhuis>ryah piscisaureus_: https://gist.github.com/cf200af99b4ae26eb5a1
21:32:23  <CIA-53>node: Ryan Dahl * r5783a52 / (12 files in 3 dirs):
21:32:24  <CIA-53>node: Remove process.memoryUsage().vsize
21:32:24  <CIA-53>node: Not meaningful cross platform; unused. - http://git.io/knOhww
21:45:26  * mralephquit (Quit: Leaving.)
21:50:31  <bnoordhuis>piscisaureus_: http://code.google.com/p/gyp/issues/detail?id=216
22:01:34  <igorzi>ryah: https://github.com/joyent/node/blob/master/lib/net.js#L695
22:01:50  <igorzi>^---- this breaks fork().send({}, server._handle);
22:02:14  <igorzi>(after server.listen())
22:02:24  <igorzi>(because server._handle is not set)
22:02:37  <CIA-53>node: Ryan Dahl * rdbfc819 / (11 files in 3 dirs): Upgrade GYP to r1078 - http://git.io/BvTaeA
22:03:15  * bradleymeckjoined
22:05:39  <ryah>igorzi: hm - ok
22:06:02  <igorzi>ryah: does listen2 have to run on nexttick?
22:06:10  <ryah>igorzi: i think the nextTick can be removed
22:06:20  <ryah>that was something i was just playing around with
22:06:50  <ryah>https://github.com/joyent/node/blob/dbfc8198a6879a0764314db701b610ca2f0c038b/lib/net.js#L682-684 <-- this is already in nextTick
22:13:37  <CIA-53>node: Ryan Dahl * r6cc0c9e / lib/net.js :
22:13:37  <CIA-53>node: Remove superfluous nextTick during server binding
22:13:37  <CIA-53>node: This breaks fork().send({}, server._handle) after server.listen() because
22:13:37  <CIA-53>node: server._handle is not set. - http://git.io/JOg9gQ
22:13:58  <ryah>igorzi--^
22:15:54  <igorzi>ryah: cool, thx
22:23:21  * isaacsjoined
22:32:51  <isaacs>ryah: about to remove the contexts from the repl
22:33:03  <isaacs>any strong feelings about this? i str you were for it. bnoordhuis lgtm'ed it
22:35:16  <bradleymeck>isaacs removing all contexts?
22:35:35  <isaacs>bradleymeck: just from the repl
22:35:43  <bradleymeck>as in you cant set sandbox vars?
22:35:45  <isaacs>require("repl").context === global, plus a deprecation warning
22:36:04  <isaacs>not in the repl, that is correct.
22:36:26  <isaacs>it'll make modules and the repl not have different copies of Array, Object, RegExp, etc.
22:36:32  <isaacs>instanceof will work properly
22:37:06  <bradleymeck>but people using multiple repls might get pollution, though they could just spawn up children i guess
22:37:18  <bradleymeck>thats a pretty darn rare need
22:48:18  <igorzi>piscisaureus_ ryah: https://gist.github.com/1277039
22:48:35  <igorzi>^---- the auto-scale experiment
22:49:02  <igorzi>only marginally better, but not as good as just doing 1 pending accept
22:49:25  * bradleymeckquit (Ping timeout: 244 seconds)
22:50:45  <igorzi>as we discussed, libuv should probably expose a knob, which defaults to no pending accepts in case the socket is shared
22:51:01  <ryah>isaacs: why?
22:51:21  <ryah>isaacs: can we review that before you land?
22:51:31  <isaacs>ryah: because we get new user bugs very frequently being confused by this.
22:51:50  <isaacs>ryah: yeah: https://github.com/joyent/node/pull/1904
22:52:31  <isaacs>for the cli repl, i think any objections are pretty misplaced. however, if you have a web server with a repl, then you might not want to have them run in the same context for some reason.
23:09:43  <piscisaureus_>igorzi: where's the code for that auto-scaling?
23:12:33  <CIA-53>node: Ryan Dahl * rd9bc845 / test/simple/test-c-ares.js : Fix test-c-ares.js on Windows by disabling PTR test - http://git.io/2V1raA
23:20:52  <igorzi>piscisaureus_: https://github.com/igorzi/libuv/commit/e27006e534de3684db077952c6efc0097fa2ba1a
23:21:01  <igorzi>there are some hacky shortcuts there to make the experiment work.. so, please be gentle when assessing the code :)
23:24:30  <igorzi>ryah: are you planning to make "node cluster" to fork again if a child process dies?
23:27:03  <ryah>igorzi: we're discussing ways to make it plugable
23:27:07  <ryah>so people can define their own logic
23:27:38  <ryah>igorzi: digressing... Debug\node.exe test\fixtures\print-chars.js 50 | cat
23:27:59  <ryah>igorzi: i.e. when stdio is opened as blocking pipes
23:28:09  <ryah>igorzi: it needs to be done on the thread pool
23:29:38  <igorzi>ryah: is stdio being opened as blocking pipes already implemented? (on windows)
23:30:12  <piscisaureus_>igorzi: no, it's not, that's the problem
23:30:43  <piscisaureus_>igorzi: we need to have some sort of detection in uv_pipe_open_fd and branch in uv_pipe_write / uv_pipe_queue_read
23:31:05  <igorzi>ok, if nobody is working on that yet, i can work on that
23:31:23  <piscisaureus_>igorzi: do you have anything else you want to work on now
23:31:42  <piscisaureus_>I mean, finish your current work first before you start something else seems better
23:31:49  <piscisaureus_>but otherwise you can work on it
23:32:25  <igorzi>piscisaureus_: i'm fixing tests atm, and not having stdio in forked node processes is breaking a bunch of tests.. so this kind of falls in line with fixing tests
23:32:50  <piscisaureus_>igorzi: well, sounds good. Go for it :-)
23:35:36  <igorzi>piscisaureus_: we're not going to do the detection whether the pipe is blocking or non-blocking (beyond ipc=1 in uv_pipe_init), correct? basically, uv_pipe_open will assume a blocking pipe unless pipe->ipc is set?
23:36:09  <piscisaureus_>igorzi: I don't think so
23:36:26  <igorzi>piscisaureus_: you want to do the detection?
23:36:40  <piscisaureus_>igorzi: oh hmm I think you're right
23:36:52  <piscisaureus_>igorzi: yeah we can do without
23:37:00  <igorzi>piscisaureus_: k
23:37:01  <piscisaureus_>igorzi: although it feels a little brittle to me
23:37:47  <igorzi>piscisaureus_: as we discussed before though, in most cases (when node is started by some other process) it'll be a blocking pipe
23:38:24  <piscisaureus_>igorzi: yeah ... so we assume blocking when opening an fd and ipc is zero
23:39:04  <piscisaureus_>igorzi: although I could imagine that at some point people may want to spawn another libuv process and use overlapped pipes, but not use the ipc protocol
23:41:54  <igorzi>piscisaureus_: hmm.. we could include the detection. i remember you were doing some experiments with it.. do you have that detection code?
23:44:20  * AvianFluquit (Ping timeout: 260 seconds)
23:49:27  <piscisaureus_>igorzi: let me just look it up
23:49:49  <piscisaureus_>igorzi: I don't think i have the code still but it's gonna be a simple call and a flag test
23:51:11  <piscisaureus_>igorzi: you need NtQueryInformationFile with FileModeInformation
23:51:59  <igorzi>piscisaureus_: ok, thx.
23:52:42  * AvianFlujoined
23:52:50  <igorzi>piscisaureus_: if the pipe is non-blocking, then we'll try with iocp, and if it's already associated with iocp - then we'll do UV_HANDLE_EMULATE_IOCP
23:54:28  <piscisaureus_>igorzi: no
23:55:46  <igorzi>piscisaureus_: why not?
23:56:09  <piscisaureus_>igorzi: it's a waste of effort - when do people give us an *associated* iocp pipe?
23:56:17  <piscisaureus_>igorzi: we can have that some day
23:56:18  <piscisaureus_>maybe
23:56:50  <piscisaureus_>igorzi: but is seems to me that it would be more useful to support that for socket read/writes than for pipe stuff
23:57:04  <ryah>igorzi: this issue is demostrated with the broken test-child-process-stdout-flush.js
23:57:13  <piscisaureus_>igorzi: but if you really want to do it, go for it
23:57:35  <ryah>i think this would be a good thing for igor to take, or?
23:57:42  <piscisaureus_>yes
23:58:28  <piscisaureus_>igorzi: something else
23:58:38  * ryahtopic: v0.5.10 on Friday
23:58:50  <piscisaureus_>igorzi: is it possible to figure out if SetFileNotificationModes was used on a handle?
23:58:53  <piscisaureus_>igorzi: I can't find it
23:59:05  <piscisaureus_>(since I don't have the source code)
23:59:26  <igorzi>piscisaureus_: the code to handle iocp pipes is already there, so might as well try to use it..
23:59:36  <igorzi>piscisaureus_: i'll check
23:59:39  <piscisaureus_>igorzi: oh, cool, do that