00:00:29  <piscisaureus_>bnoordhuis: the file is probably still open when the readFile callback is made. So unlink() doesn't really delete the directory entry, it's just marked for deletion.
00:00:46  <piscisaureus_>but as soon as the user tries to delete the dir, there's still a file in there.
00:01:03  <bnoordhuis>piscisaureus_: how are you going to fix it?
00:01:09  <piscisaureus_>bnoordhuis: am i?
00:01:11  <piscisaureus_>:-p
00:01:44  <bnoordhuis>piscisaureus_: i expect a fix by 4 on my desk
00:01:51  * elijah-mbpjoined
00:02:05  <piscisaureus_>bnoordhuis: you also expect to get paid by the 20th, don't you :-p
00:02:16  <piscisaureus_>bnoordhuis: so maybe no fix is possible.
00:02:33  <piscisaureus_>bnoordhuis: maybe we can change fs.ReadStream to actually close the file *before* the end event
00:02:41  <bnoordhuis>piscisaureus_: okay. in that case it ought to be documented
00:02:54  <bnoordhuis>piscisaureus_: fs.ReadStream?
00:02:55  <piscisaureus_>bnoordhuis: why?
00:03:01  <piscisaureus_>bnoordhuis: I don't think anybody will notice :_)
00:03:18  <piscisaureus_>bnoordhuis: or fs.ReadFile (which uses readStream under the hood) could wait for the "close" event and not the "end" event.
00:03:42  <bnoordhuis>piscisaureus_: are you mixing issues?
00:03:57  <piscisaureus_>bnoordhuis: no
00:04:24  <bnoordhuis>piscisaureus_: in that case you need to explain to me how this is related to fs.unlinkSync and fs.rmdirSync :)
00:04:41  <piscisaureus_>bnoordhuis: On Windows, fs.unlinkSync does not work if called in the callback of fs.readFile
00:04:50  <piscisaureus_>bnoordhuis: the key is in "the callback of fs.Readfile"
00:05:35  <bnoordhuis>piscisaureus_: oh? that actually makes a difference?
00:05:55  <piscisaureus_>bnoordhuis: as I just explained, you cannot delete a file that is open (it just gets marked for deletion) and you cannot remove a directory with files in it (even those pending deletion) - it fails with ENOTEMPTY
00:06:05  <bnoordhuis>your windows file semantics are strange and confusing
00:06:08  <piscisaureus_>bnoordhuis: fs.readFile uses fs.ReadStream underneath
00:06:24  <piscisaureus_>bnoordhuis: fs.readFIle makes its callback when ReadStream emits "end"
00:06:32  <piscisaureus_>by that time the file is still open
00:06:41  <bnoordhuis>piscisaureus_: right. so essentially this could be fixed by deferring the callback with process.nextTick?
00:06:48  <piscisaureus_>so if you try to unlink it and then delete it's parent directory, it fails
00:06:49  <bnoordhuis>assuming that the file is closed on the current tick
00:07:02  <piscisaureus_>bnoordhuis: well, the close happens on the thread pool
00:07:12  <piscisaureus_>bnoordhuis: so you would have to wait until after "close" returns, somehow
00:07:12  <bnoordhuis>ah
00:07:43  <bnoordhuis>it's starting to make sense
00:07:58  <piscisaureus_>bnoordhuis: so here were my options: (a) actually close first, then emit "end"
00:08:05  <piscisaureus_>bnoordhuis: or make fs.ReadFile wait for "close"
00:08:10  <CIA-99>node: Igor Zinkovsky v0.6 * r052aaa4 / (src/node.h src/node_file.cc): windows: use 64bit offsets for uv_fs apis - http://git.io/Dkrt4w
00:08:27  <piscisaureus_>bnoordhuis, it would slightly increase latency so maybe we should do it only on windows
00:08:48  <bnoordhuis>piscisaureus_: i don't think people use node for its high-performance file I/O
00:08:49  <piscisaureus_>bnoordhuis: (if you care)
00:08:55  <piscisaureus_>which is a shame
00:09:19  <piscisaureus_>we should add unbuffered ReadStreams
00:09:20  <piscisaureus_>that would help
00:09:38  <bnoordhuis>i don't know
00:09:57  <piscisaureus_>bnoordhuis: no? where do you spend your time on unix?
00:10:07  <bnoordhuis>profiling suggests that it's malloc and the overhead of threads that is expensive
00:10:09  <piscisaureus_>bnoordhuis: with unbuffer ReadStream we can use kernel aio
00:10:15  <piscisaureus_>so no threads
00:10:18  <bnoordhuis>oh, you mean O_DIRECT
00:10:25  <piscisaureus_>yes
00:10:49  <bnoordhuis>yes, well... O_DIRECT and/or AIO doesn't work with all filesystems
00:11:06  <piscisaureus_>bnoordhuis: is it possible to detect failure and fall back?
00:11:14  <piscisaureus_>I suppose that is what windows would need to
00:11:25  <bnoordhuis>piscisaureus_: not reliably, for sure
00:11:31  <CIA-99>node: Igor Zinkovsky v0.6 * rf178f2a / (13 files in 5 dirs): upgrade libuv to d68b3d960b6d95bfc16027cecca2f3fa48bcc36f - http://git.io/PXUG9A
00:11:39  <bnoordhuis>it's a sad state of affairs, AIO is :/
00:12:01  <isaacs>I'd like to do a v0.6 release tomorrow or monday. anything pending to get into it?
00:12:08  <piscisaureus_>isaacs: license stuff.
00:12:27  <isaacs>piscisaureus_: right, but that's done in master, right? looks easy enough to port, i can handle it.
00:12:40  <piscisaureus_>isaacs: yeah, it's just... there are tasks left :-)
00:12:46  <isaacs>piscisaureus_: kewl :)
00:13:02  <isaacs>piscisaureus_: anything that you're aware of that needs to be finished or figured out?
00:13:41  <piscisaureus_>isaacs: I think the license file just needs a review. And maybe you want to install the license file itself as well (as opposed to just showing it in the installer)
00:13:51  <piscisaureus_>isaacs: https://github.com/joyent/node/issues/3056 <-- tracking bug
00:14:05  <isaacs>yeah, that's a big part of the motivation, actually
00:14:27  <piscisaureus_>isaacs: it would be good to get a test for the large file issue that igorzi just fixed in
00:14:32  <isaacs>mark was pretty cool about the whole thing, but i do feel kind of bad that we raised his ire even a little.
00:14:35  <piscisaureus_>isaacs: (in node)
00:15:35  <piscisaureus_>isaacs: so he just emailed ryan?
00:15:49  <piscisaureus_>isaacs: or was it legal department stuff?
00:16:18  * travis-cijoined
00:16:18  <travis-ci>[travis-ci] joyent/node#696 (v0.6 - 052aaa4 : Igor Zinkovsky): The build passed.
00:16:18  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/ffee873...052aaa4
00:16:18  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1028661
00:16:18  * travis-cipart
00:16:19  * orlandovftwquit (Ping timeout: 272 seconds)
00:16:24  <piscisaureus_>isaacs: I guess you should just tell him sorry, it was an oversight, we are just getting used to binary distributions
00:17:08  <igorzi>piscisaureus_ isaacs: i'll create a test for the large file
00:17:17  <piscisaureus_>igorzi: kewl
00:18:30  <seebees1>hey, can I ask a question about 0.7 -> 0.8?
00:19:28  <isaacs>piscisaureus_: yeah, that's exactly what ryan told him, and then i also apologized and told him that it'd be fixed in the next release.
00:19:47  <isaacs>piscisaureus_: like i said, he was pretty understanding.
00:19:54  <piscisaureus_>isaacs: ok, cool
00:20:00  * travis-cijoined
00:20:01  <travis-ci>[travis-ci] joyent/node#697 (v0.6 - f178f2a : Igor Zinkovsky): The build passed.
00:20:01  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/052aaa4...f178f2a
00:20:01  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1028692
00:20:01  * travis-cipart
00:20:18  <igorzi>https://github.com/joyent/node/blob/master/test/disabled/test-fs-largefile.js
00:20:27  <igorzi>piscisaureus_: do you know why it's disabled? ^
00:22:21  <bnoordhuis>seebees1: sure
00:22:46  <seebees1>What is the time frame for winding down 0.7?
00:23:02  <seebees1>Because I don't want to start on something and only get it half doneā€¦
00:23:15  <isaacs>seebees1: very soon.
00:23:20  <seebees1>hum
00:23:32  <seebees1>so if I bust out something this weekend?
00:23:40  <isaacs>seebees1: i'd like to finish the remaining features by May 1.
00:23:45  <isaacs>seebees1: and release 0.8 by June 1
00:23:52  <bnoordhuis>seebees1: depends on what it is
00:23:59  <isaacs>so we have a month to bang on it and can brag about how awesome it is at NodeConf on July 2
00:24:02  <bnoordhuis>that long stack traces thing?
00:24:08  <seebees1>yup
00:24:18  <seebees1>I think I know how I can make it work
00:24:21  <seebees1>but I need to find time
00:24:32  <seebees1>to work on it. and "real" work is being anoying...
00:24:50  <seebees1>but if I can get something real this weekend and make a pull then people can look at it.
00:24:55  <seebees1>so it might work...
00:24:58  <bnoordhuis>yes
00:25:13  <seebees1>ok. good to know. Monday at the latest then. thanks.
00:25:37  <bnoordhuis>seebees1: no rush. worst case, it doesn't make the cut and lands in 0.9
00:25:57  <seebees1>bnoordhuis: too true. just my pride ;)
00:26:35  <seebees1>my guess is if it get's in it would probably .9 in any event. But I can hope...
00:29:34  <piscisaureus_>isaacs: I am also doing a patch for #3051 quickly.
00:31:06  * avalanche123quit (Quit: Computer has gone to sleep.)
00:31:30  <isaacs>piscisaureus_: kewl
00:37:11  <toothr>i'm still wondering about the subtree watching for windows... will this ever be configurable, or is this passed over due to not being avail in other platforms? https://github.com/joyent/libuv/blob/master/src/win/fs-event.c#L240
00:43:47  <toothr>i guess i can just make my own version of uv_fs_event_init without requiring any changes to libuv.. right
00:47:17  <piscisaureus_>toothr: I am not agains it... but we have no story for other platforms except (maybe) recent versions of OS X
00:47:32  <piscisaureus_>toothr: but solaris and bsd and linux just won't do it
00:48:38  <toothr>aha, perhaps I will open an issue if that's the right thing to do?
00:56:18  <piscisaureus_>toothr: well, you could, but I doubt it'd help much
00:56:34  <piscisaureus_>toothr: it would be better to come up with a nice solution :-)
01:04:24  <CIA-99>node: Bert Belder reviewme * r48eb490 / (lib/fs.js test/simple/test-fs-readfile-unlink.js):
01:04:25  <CIA-99>node: fs.readFile: don't make the callback before the fd is closed
01:04:25  <CIA-99>node: On Windows it is not possible to unlink() the read file in the callback.
01:04:25  <CIA-99>node: This fixes #3051. A test is included. - http://git.io/TT5g5Q
01:04:25  <CIA-99>node: Bert Belder reviewme * rbb82907 / src/v8_typed_array.cc : Please the compiler - http://git.io/2bWI1Q
01:04:27  <piscisaureus_>^-- bnoordhuis igorzi - review?
01:12:30  * travis-cijoined
01:12:30  <travis-ci>[travis-ci] joyent/node#698 (reviewme - bb82907 : Bert Belder): The build is still failing.
01:12:30  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/56a8101...bb82907
01:12:30  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1028841
01:12:30  * travis-cipart
01:13:25  <igorzi>piscisaureus_: i'll enable this test: https://github.com/joyent/node/blob/v0.6/test/disabled/test-fs-largefile.js
01:13:51  <piscisaureus_>igorzi: kk
01:14:59  <piscisaureus_>igorzi: no more comments besides the callback(error) ?
01:15:21  <igorzi>piscisaureus_: nope.. both patches lgtm
01:15:30  <CIA-99>node: Bert Belder v0.6 * r40b7302 / (lib/fs.js test/simple/test-fs-readfile-unlink.js):
01:15:30  <CIA-99>node: fs.readFile: don't make the callback before the fd is closed
01:15:30  <CIA-99>node: On Windows it is not possible to unlink() the read file in the callback.
01:15:30  <CIA-99>node: This fixes #3051. A test is included. - http://git.io/l2-akg
01:15:30  <CIA-99>node: Bert Belder v0.6 * r1042a8d / src/v8_typed_array.cc : Please the compiler - http://git.io/g91hmA
01:19:41  <CIA-99>node: Igor Zinkovsky v0.6 * r0b57fee / (2 files in 2 dirs): enable test-fs-largefile.js test - http://git.io/t05Ycg
01:23:26  * travis-cijoined
01:23:27  <travis-ci>[travis-ci] joyent/node#699 (v0.6 - 1042a8d : Bert Belder): The build was broken.
01:23:27  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/f178f2a...1042a8d
01:23:27  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1028871
01:23:27  * travis-cipart
01:26:49  <piscisaureus_>igorzi: isaacs: on windows, would you want to install LICENSE or license.rtf?
01:27:13  <isaacs>piscisaureus_: i'm not sure we really need to *install* the license.
01:27:22  <isaacs>piscisaureus_: just need to display and have the user click-through
01:27:33  <isaacs>piscisaureus_: that being said, meh. LICENSE (text file) probably.
01:27:41  * travis-cijoined
01:27:42  <travis-ci>[travis-ci] joyent/node#700 (v0.6 - 0b57fee : Igor Zinkovsky): The build is still failing.
01:27:42  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/1042a8d...0b57fee
01:27:42  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1028881
01:27:42  * travis-cipart
01:27:57  <piscisaureus_>isaacs: ah, right. I recall that you said you wanted to install the thing, but I must have misunderstood.
01:29:45  <piscisaureus_>bnoordhuis: personally, I don't care
01:29:54  <bnoordhuis>piscisaureus_: ?
01:30:23  <piscisaureus_>bnoordhuis: re "?" - ?
01:30:41  <bnoordhuis>piscisaureus_: "bnoordhuis: personally, I don't care"
01:30:48  <piscisaureus_>oh, sorry
01:30:49  <bnoordhuis>don't care about what?
01:30:52  <piscisaureus_>confused you with isaacs
01:30:58  <piscisaureus_>getting tired
01:31:06  <bnoordhuis>i'm the tall guy, remember?
01:31:32  * TooTallNatequit (Ping timeout: 260 seconds)
01:31:35  <bnoordhuis>okay, off to bed - sleep tight, guys
01:32:13  <piscisaureus_>yeah, me too
01:32:15  <piscisaureus_>later guys
01:32:16  * abraxasjoined
01:32:19  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
01:37:06  * bnoordhuisquit (Ping timeout: 276 seconds)
01:51:23  * pieternquit (Ping timeout: 265 seconds)
02:04:26  * brsonquit (Ping timeout: 245 seconds)
02:19:02  * abraxasquit (Read error: Connection reset by peer)
02:19:28  * abraxasjoined
02:46:08  * TooTallNatejoined
02:51:48  * mikealquit (Quit: Leaving.)
02:53:14  * abraxas_joined
02:54:12  * abraxasquit (Read error: Connection reset by peer)
03:07:56  * mikealjoined
03:12:40  * mikealquit (Ping timeout: 240 seconds)
03:16:06  * mikealjoined
03:24:48  * dylukesquit (Quit: Pipes are broken. Sending packets via Fedex.)
03:24:58  * mikeal1joined
03:25:58  * mikealquit (Read error: Connection reset by peer)
03:33:07  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
03:48:40  * perezdquit (Quit: perezd)
04:14:23  * mikeal1quit (Read error: Connection reset by peer)
04:25:06  * mikealjoined
04:26:41  * mikealquit (Read error: Connection reset by peer)
04:26:47  * mikealjoined
04:36:03  * abraxas_quit (Read error: No route to host)
04:36:13  * abraxasjoined
04:40:41  * abraxasquit (Ping timeout: 245 seconds)
04:41:03  * abraxasjoined
04:45:11  * abraxas_joined
04:45:41  * abraxasquit (Ping timeout: 245 seconds)
04:49:42  * pieternjoined
04:50:53  * pieternquit (Client Quit)
04:54:02  * ryahquit (Ping timeout: 264 seconds)
05:13:32  * ryahjoined
05:13:58  * AvianFluquit (Ping timeout: 250 seconds)
05:14:37  * isaacsquit (Remote host closed the connection)
05:44:37  * AvianFlujoined
05:56:52  * AvianFluquit (Ping timeout: 250 seconds)
06:11:34  * AvianFlujoined
06:11:56  * mmalecki[away]changed nick to mmalecki
06:15:22  * mmaleckichanged nick to mmalecki[zzz]
07:00:10  * paddybyersquit (Quit: paddybyers)
07:00:52  * paddybyersjoined
07:23:36  * txdv_joined
07:27:21  * txdvquit (Ping timeout: 245 seconds)
07:38:49  * `3rdEdenjoined
08:22:32  * `3rdEdenquit (Quit: Leaving...)
08:51:56  * rendarjoined
09:12:22  * `3rdEdenjoined
09:34:27  * `3rdEdenquit (Quit: Leaving...)
09:43:41  * `3rdEdenjoined
09:52:42  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
10:16:54  * abraxas_quit
10:49:26  <txdv_>how is the windows async io api called?
10:49:51  <rendar>there are APC queues or io completion ports..
11:31:56  * piscisaureus_joined
11:36:20  * piscisaureus_quit (Client Quit)
11:42:46  * piscisaureus_joined
11:49:42  * bnoordhuis_joined
12:43:29  * `3rdEdenjoined
12:56:17  * bnoordhuisjoined
12:56:20  * bnoordhuis_quit (Read error: Connection reset by peer)
13:04:15  * dylukesjoined
14:02:13  * dylukesquit (Quit: Computer has gone to sleep.)
14:11:45  * skomskijoined
14:17:53  * pfox___joined
14:20:52  * bnoordhuisquit (Read error: Connection reset by peer)
14:21:07  * bnoordhuisjoined
14:22:05  * mmalecki[zzz]changed nick to mmalecki
14:22:16  * skomskiquit (Quit: skomski)
14:26:08  * dylukesjoined
14:26:26  * piscisaureus_quit (Ping timeout: 264 seconds)
14:28:44  * mikealquit (Quit: Leaving.)
14:32:21  * mikealjoined
14:37:14  * mikealquit (Ping timeout: 264 seconds)
14:39:25  * mikealjoined
14:40:20  * mikealquit (Read error: Connection reset by peer)
14:40:22  * mikeal1joined
14:55:58  * skomskijoined
14:59:13  * dylukesquit (Quit: Computer has gone to sleep.)
15:00:09  * dylukesjoined
15:06:06  * skomskiquit (Quit: skomski)
15:12:49  * piscisaureus_joined
15:26:00  * isaacsjoined
15:26:08  * TooTallNatejoined
15:31:04  * mikeal1quit (Quit: Leaving.)
15:31:33  * mikealjoined
15:36:02  * mikealquit (Ping timeout: 264 seconds)
15:54:03  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
15:57:46  * philipsquit (Excess Flood)
15:59:08  * philipsjoined
16:06:11  * dapjoined
16:16:10  * piscisaureus_quit (Ping timeout: 240 seconds)
16:30:07  * piscisaureus_joined
16:32:22  * bnoordhuis_joined
16:32:31  * bnoordhuisquit (Read error: Connection reset by peer)
16:45:48  * bnoordhuis_quit (Quit: Leaving)
17:08:47  * mmaleckichanged nick to mmalecki[brb]
17:15:33  * dylukesquit (Quit: Computer has gone to sleep.)
17:33:59  <TooTallNate>we should merge this https://github.com/joyent/node/pull/3059
17:43:42  <isaacs>TooTallNate: do you have access to the CLA document?
17:43:47  <isaacs>TooTallNate: the list of who's signed it, i mean
17:43:58  <TooTallNate>ya
17:44:00  <isaacs>kewl
17:45:07  <isaacs>so, i think you probably know the procedure. needs a test, and docs, and no added lint failures. squash to probably one commit (unless it makes sense to be multiple things, but like "X" and "Test for X" should be together) and push.
17:45:26  <isaacs>TooTallNate: the repl/readline stuff is your area of recent expertise, so feel free to land it if it meets with your approval.
17:46:18  <TooTallNate>ok cool, once he signs the cla, will do
17:58:22  * brsonjoined
18:05:00  * mikealjoined
18:13:23  * perezdjoined
18:25:12  * orlandovftwjoined
18:28:50  * dshaw_joined
18:37:41  * dapquit (Ping timeout: 265 seconds)
18:41:11  * piscisaureus_quit (Read error: Connection reset by peer)
18:41:26  * piscisaureus_joined
18:43:30  * dapjoined
18:44:33  * dylukesjoined
19:01:13  * dapquit (Quit: Leaving.)
19:01:34  * dapjoined
19:10:52  * piscisaureus_quit (Ping timeout: 260 seconds)
19:18:05  * pieternjoined
19:24:21  <TooTallNate>so test-repl.js had a flaw, which was exposed by #3059
19:24:25  <TooTallNate>so needs review: https://github.com/TooTallNate/node/compare/fix-repl-test
19:25:51  * mralephjoined
19:32:56  * CoverSlidejoined
19:35:50  * orlandovftwquit (Ping timeout: 252 seconds)
19:36:27  * Skomskijoined
20:16:19  * pietern_joined
20:18:10  * pieternquit (Ping timeout: 240 seconds)
20:18:10  * pietern_changed nick to pietern
20:28:12  <creationix>is there a way to open a file with O_EXCL in node?
20:28:27  <creationix>I want to only create the file, never truncate an existing one like the "w" string does
20:28:48  * dshaw_quit (Quit: Leaving.)
20:29:21  <tjfontaine>creationix: http://wankernews.com/item?id=3532299
20:29:52  <creationix>right, I remember seeing that
20:30:01  <creationix>but I can't require 0.8 yet
20:30:16  <creationix>maybe there is a way to do it the old way and pass in the number directly
20:30:39  <creationix>(I was the one that added the "w", "r" shortcuts originally)
20:34:04  <creationix>yep, stringToFlags passes through the number as-is. Cool
20:36:12  <creationix>hmm, I wonder why bnoordhuis did O_TRUNC with O_EXCL and O_CREAT
20:36:21  <creationix>if it's exclusive there would never be a file to truncate
20:36:30  <creationix>or maybe I don't understand what these flags do
20:42:12  * mmalecki[brb]changed nick to mmalecki
20:54:07  <isaacs>can i get a lgtm? https://gist.github.com/2321917
20:54:25  <isaacs>TooTallNate, igorzi, bnoordhuis ^
20:54:54  <TooTallNate>isaacs: sure why not :p
20:55:05  <TooTallNate>was that throwing for someone?
20:55:48  <isaacs>TooTallNate: well, no one's reported it, but i realized that you could close the channel in the child, and then writing toit would throw, and it's kind of brutal to have to check for its openness or wrap child.send() in a try/catch every time.
20:56:08  <TooTallNate>isaacs: ya agreed
20:56:10  <TooTallNate>+1
20:56:21  <TooTallNate>isaacs: igorzi: i need a review too: https://github.com/TooTallNate/node/compare/fix-repl-test
20:56:23  <isaacs>it IS an error condition, most likely, but an error event is the proper approach. (also, sketching out domains stuff, which will be a bit easier with error events than throws)
20:57:26  <isaacs>TooTallNate: + this._line_buffer += b.toString();
20:57:35  <isaacs>TooTallNate: consider using a StringDecoder?
20:57:50  <TooTallNate>shoot, you're probably right
20:57:54  <isaacs>if the intent is to decode utf8, then naively appending each string is not so good
20:58:02  <TooTallNate>right
20:58:07  <TooTallNate>haha, this guy's gonna kill me
20:59:20  <isaacs>which guy?
20:59:35  <TooTallNate>from the pull request
20:59:36  <TooTallNate>https://github.com/joyent/node/pull/3059
20:59:47  <TooTallNate>he's just had to do lots of little revisions already :p
20:59:49  <TooTallNate>mostly lint
20:59:57  <isaacs>oh, ok
21:00:03  <isaacs>you can do that as a separate commit on top of it.
21:06:30  * pieternquit (Quit: pietern)
21:07:07  <TooTallNate>isaacs: hmmm, so does that mean we need to add 'setEncoding' to readline (and an option in repl)?
21:07:40  <isaacs>TooTallNate: maybe we can just have it use utf8, and change it if anyone complains
21:07:42  * pieternjoined
21:07:43  * Skomskiquit (Quit: Nettalk6 - www.ntalk.de)
21:07:45  <isaacs>TooTallNate: since that's what it effectively does now
21:07:53  <TooTallNate>isaacs: ok, that sounds good
21:17:04  <CIA-99>node: isaacs master * rb9bfb1b / lib/child_process.js : child_process: Emit err rather than throw for IPC write failure - http://git.io/NBZmmQ
21:32:24  * piscisaureus_joined
21:32:29  * travis-cijoined
21:32:29  <travis-ci>[travis-ci] joyent/node#701 (master - b9bfb1b : isaacs): The build is still failing.
21:32:29  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/d73b257...b9bfb1b
21:32:29  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1034666
21:32:29  * travis-cipart
21:35:58  <TooTallNate>isaacs: StringDecoder-ified: https://github.com/TooTallNate/node/compare/fix-repl-test
21:43:08  <CIA-99>node: isaacs v0.6 * r37ac5e5 / (259 files in 29 dirs): Update npm to 1.1.16 - http://git.io/mG8Uqw
21:47:41  <isaacs>TooTallNate: too hungry. will review after lunch.
21:48:52  <CoverSlide>lunch at 2:45PM?
21:48:55  <CoverSlide>dayum
21:51:36  * travis-cijoined
21:51:36  <travis-ci>[travis-ci] joyent/node#702 (v0.6 - 37ac5e5 : isaacs): The build is still failing.
21:51:36  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/0b57fee...37ac5e5
21:51:36  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1034905
21:51:36  * travis-cipart
21:54:35  * mikealquit (Quit: Leaving.)
21:58:08  * mikealjoined
22:01:17  <piscisaureus_>\o/ uv_poll works
22:03:06  <CoverSlide>wtg!
22:05:53  <pfox___>piscisaureus_: is that a new bit of api?
22:06:33  <piscisaureus_>yeah
22:07:26  <piscisaureus_>if you want to integrate some external lib (cares, libssh, ...) you need to poll sockets for readability and writability
22:09:35  <pfox___>interesting
22:14:43  * dshaw_joined
22:14:44  <TooTallNate>piscisaureus_: nice!
22:15:03  <piscisaureus_>now I need to get the fallback mode properly working
22:16:10  * AvianFluquit (Ping timeout: 240 seconds)
22:27:56  * AvianFlujoined
22:29:35  * rendarquit
22:35:55  * dylukesquit (Quit: Computer has gone to sleep.)
22:36:23  * bnoordhuisjoined
22:37:22  <piscisaureus_>and then there was this guy who does COM ports... on windows
22:37:25  * wankdankerquit (Read error: Connection reset by peer)
22:37:35  <piscisaureus_>the 80s just rang
22:37:45  <TooTallNate>piscisaureus_: did you look at that guys branch?
22:37:50  <piscisaureus_>no
22:37:52  <TooTallNate>looks like he did a lot of work on it already
22:38:08  <piscisaureus_>TooTallNate: are we talking about the same
22:38:10  <piscisaureus_>guy?
22:38:13  <TooTallNate>piscisaureus_: https://github.com/jaredhanson/node-serialport/compare/windows-support
22:38:15  <TooTallNate>idk
22:38:36  <piscisaureus_>TooTallNate: I was talking about http://9292.nl/en/journeyadvice/amsterdam_bloemstraat-48f/station-rotterdam-centraal/arrival/2012-04-07T1115
22:38:37  <TooTallNate>which guy are you talking about? :p
22:38:39  <piscisaureus_>er
22:38:40  <TooTallNate>ah
22:38:43  <piscisaureus_>wrong link
22:38:44  <piscisaureus_>:-)
22:38:45  <TooTallNate>oh
22:39:04  <piscisaureus_>TooTallNate: that -> http://9292.nl/en/journeyadvice/amsterdam_bloemstraat-48f/station-rotterdam-centraal/arrival/2012-04-07T1115
22:39:33  <TooTallNate>piscisaureus_: same link?
22:39:51  <piscisaureus_>huh
22:39:53  <piscisaureus_>fuck
22:40:10  <piscisaureus_>my clipboard is borked
22:40:30  <piscisaureus_>I cannot copy it's silently ignored
22:40:40  <TooTallNate>highlight and drag?
22:40:43  <TooTallNate>haha, that's weird
22:40:44  <piscisaureus_>http://groups.google.com/group/nodejs/browse_thread/thread/c36d513f06a84d62
22:40:59  <piscisaureus_>TooTallNate: I ran ,mingw's gdb on windows
22:41:05  <piscisaureus_>it breaks the clipboard sometimes
22:41:07  <piscisaureus_>very weird
22:41:57  <piscisaureus_>I kind of forgot (when I was doing the mingw port a year ago it used to happen all the time. I wrote a tool to repair the clipboard even.)
22:43:25  <benvie>oh yeah I had some program that used to break it too
22:43:28  <TooTallNate>haha, very streange
22:43:29  <benvie>how does that even happen
22:43:30  <TooTallNate>strange
22:43:52  <benvie>that windows somehow lets you break something global like that
22:44:10  <benvie>something about now releasing the what's it called
22:44:30  <benvie>GlobalLock and GlobalUnlock
22:44:38  <piscisaureus_>benvie: no it's OpenClipboard
22:45:01  <piscisaureus_>benvie: if your app calls that and forgets to call CloseClipboard the clipboard remains locked
22:45:07  <benvie>oh yeah
22:45:23  <piscisaureus_>(even after it exits)
22:45:28  <piscisaureus_>that's the broken part
22:45:36  <benvie>yeah and now I remember
22:45:56  <benvie>I had to make the clipboard module basically have a ref tracking part to
22:46:22  <benvie>so when you want to use the clipboard you have to hit the ref and unref, and that keeps track of whether you opened it or not and opens/closes it
22:47:01  <benvie>makes it kind of annoying to do a cross-platform clipboard thing
22:47:55  * mikealquit (Quit: Leaving.)
22:48:38  <piscisaureus_>TooTallNate: the windows-support branch doesn't look like it should work now
22:48:50  <piscisaureus_>TooTallNate: the guy should come and talk to me :-)
22:49:19  <piscisaureus_>TooTallNate: he is working in the right direction but apparently he does not really understand the windows overlapped io concepts
22:49:35  <TooTallNate>piscisaureus_: i'll point him in your direction
22:49:45  * dylukesjoined
22:49:46  <piscisaureus_>TooTallNate: you know the guy?
22:50:04  <TooTallNate>no, i just found that in a "windows support" Issue on node-serialport
22:50:41  <benvie>reminds me, I wanted to add the ability to listen for clipboard changes to this but then I got all wrapped up in the binary data stuff
22:52:07  <piscisaureus_>benvie: I don't think you can do that in a nice way. Or... maybe, if you create a thread and a window and a message pump and an uv_async thing
22:52:20  <benvie>oh is it blocking?
22:52:27  <benvie>AddClipboardFormatListener
22:52:43  <benvie>oh wait
22:52:48  <benvie>I see, you need it on a hwnd
22:53:05  <benvie>nah you don't need a message pump to run an hwnd
22:53:19  <benvie>I was able to do it using setTimeout and PeakMessage
22:53:39  <piscisaureus_>yeah, that also works
22:53:47  <piscisaureus_>ugly tho :-)
22:53:52  <benvie>haha it's funny though
22:54:27  <benvie>https://gist.github.com/1938424
22:54:41  <benvie>making a window and running the loop entirely using ffi
22:54:45  <benvie>pretty ridiculous
22:55:35  <piscisaureus_>yeah well I don't mind
22:56:10  <piscisaureus_>it's pretty ridiculous that there is no fckng way to integrate a messate pump with io completion ports
22:56:32  <piscisaureus_>it would be nice to be able to do WaitMessage on another thread and then wake up the main thread
22:56:33  <benvie>part of why I've been trying to build out a better binary data interface is so ffi bindings for the whole windows api are actually usable
22:56:35  <piscisaureus_>no luck to
22:56:42  <piscisaureus_>*tho
22:57:17  <benvie>yeah I haven't actually tried the setInterval approach on anything besides that little thing. I imagine it would be pretty terrible to actually use it
22:57:29  <benvie>though I'm planning to at least try
22:58:16  <benvie>it'd be workable for something like a clipboard watcher and hilariously awesome to run all of a browser frame on it
22:59:09  <piscisaureus_>right
22:59:11  <piscisaureus_>I have to go now
22:59:26  <benvie>adios
23:07:47  * mikealjoined
23:09:38  * dylukesquit (Quit: Computer has gone to sleep.)
23:12:51  <isaacs>TooTallNate: rubber-stamp lgtm.
23:13:21  * dylukesjoined
23:13:30  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
23:13:49  <TooTallNate>isaacs: thanks
23:15:19  <CIA-99>node: Nathan Rajlich master * r78eb174 / (lib/readline.js test/simple/test-readline-interface.js):
23:15:19  <CIA-99>node: readline: use StringDecoder for decoding "normal" data
23:15:19  <CIA-99>node: The fix from #3059 was not handling multi-byte utf8 data properly. (+5 more commits...) - http://git.io/fBUi8g
23:16:56  <isaacs>TooTallNate: does b6d6a54f8057d0adad13d4bcc11eb3f443079ddc belong on v0.6?
23:17:10  <isaacs>"build: target OSX 10.5 when building on darwin"
23:17:33  <TooTallNate>isaacs: don't we use waf on 0.6 still?
23:17:39  <isaacs>oh, right
23:17:39  <isaacs>nvm
23:17:42  <isaacs>dumb question :)
23:17:50  <TooTallNate>:p
23:18:00  <isaacs>is it a problem in 0.6, though? should we port to waf-style?
23:18:27  <TooTallNate>hmmm, good question
23:18:45  <isaacs>i guess we can just try not doing it.
23:18:48  <isaacs>see how that goes.
23:18:52  <isaacs>:)
23:18:58  <TooTallNate>isaacs: i'll try out the 0.6.14 .pkg on a Leopard machine on monday and see if it's needed
23:19:15  <isaacs>or you can try out the 0.6.15 pkg on leopard!
23:19:39  <TooTallNate>ya for sure
23:19:54  <TooTallNate>isaacs: are we doing a release today?
23:20:43  <isaacs>TooTallNate: looking like a monday morning deal
23:20:54  <TooTallNate>cool
23:20:54  <isaacs>we can review over the weekend
23:21:24  <isaacs>testing out the port of the license stuff right now
23:21:29  <isaacs>making sure i got it all over properly
23:24:39  * mralephquit (Quit: Leaving.)
23:30:06  * travis-cijoined
23:30:06  <travis-ci>[travis-ci] joyent/node#703 (master - 78eb174 : Nathan Rajlich): The build is still failing.
23:30:06  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/b9bfb1b...78eb174
23:30:06  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/1035593
23:30:06  * travis-cipart
23:49:40  * bnoordhuisquit (Ping timeout: 240 seconds)
23:51:08  * mikealquit (Quit: Leaving.)
23:55:11  * mikealjoined
23:58:06  * dylukesquit (Quit: Pipes are broken. Sending packets via Fedex.)
23:59:49  * pietern_joined