00:00:51  * brsonquit (Ping timeout: 252 seconds)
00:03:30  * brsonjoined
01:57:11  * bnoordhuisjoined
02:06:22  * isaacsjoined
02:15:29  <dmkbot>joyent/node: BillBarnhill: Fix memory allocation in uv__fs_after - https://github.com/joyent/node/issues/1686
02:16:57  * brsonquit (Ping timeout: 258 seconds)
02:20:29  <dmkbot>joyent/node: Bonuspunkt: fs.watchFile not working on windows - https://github.com/joyent/node/issues/1358
03:14:53  * bnoordhuisquit (Ping timeout: 260 seconds)
04:34:42  * isaacsquit (Quit: isaacs)
04:42:10  * isaacsjoined
04:54:01  * isaacsquit (Quit: isaacs)
05:55:30  <dmkbot>joyent/node: tshinnic: add positioned writing feature to fs.WriteStream - https://github.com/joyent/node/issues/1645
06:07:32  * mralephjoined
06:36:51  <CIA-52>node: Thomas Shinnick master * re4ebeb6 / lib/fs.js :
06:36:51  <CIA-52>node: fs: minor corrections from examining stream read positioning
06:36:51  <CIA-52>node: Fix minor typos, one small refactor, and change emit() in a constructor
06:36:51  <CIA-52>node: to a throw - http://git.io/sQdZ7Q
06:36:51  <CIA-52>node: Thomas Shinnick master * re58c036 / (4 files in 3 dirs):
06:36:52  <CIA-52>node: fs: add positioned file writing feature to fs.WriteStream
06:36:52  <CIA-52>node: Patterned on same feature in ReadStream; a small bit of new code added
06:36:52  <CIA-52>node: plus two refactorings of previous code; added two test files.
06:36:53  <CIA-52>node: Fixes #1645. - http://git.io/iLX_vQ
06:44:39  * mralephquit (Quit: Leaving.)
06:47:22  * felixgejoined
06:47:22  * felixgequit (Changing host)
06:47:22  * felixgejoined
06:56:23  * piscisaureusjoined
06:58:59  <piscisaureus>ryah: hey
06:59:29  <piscisaureus>ryah: so are you going to remove uv_init :?
07:00:22  * ericktquit (Quit: erickt)
07:01:28  <CIA-52>node: Thomas Shinnick master * r7dc2c49 / (lib/fs.js test/simple/test-fs-watch-file.js):
07:01:28  <CIA-52>node: fs: unguarded fs.watchFile cache statWatchers checking fixed
07:01:28  <CIA-52>node: Use hasOwnProperty to check filepath cache; previous code could fail if
07:01:28  <CIA-52>node: a filepath duplicated a chained property name.
07:01:28  <CIA-52>node: Fixes #1637. - http://git.io/tOkTZw
07:03:10  <ryah>piscisaureus: i havne't got to it yet
07:03:18  <ryah>piscisaureus: tomorrow
07:08:06  * felixgequit (Quit: felixge)
07:09:32  <piscisaureus>ryah: ok :-)
07:32:47  <indutny>ryah: good evening
09:30:29  <dmkbot>joyent/libuv: piscisaureus: Specify error handling for uv_write and uv_connect - https://github.com/joyent/libuv/issues/185
09:36:53  <piscisaureus>igorzi: I am (re)starting with console input reading. Ping me before you start doing the same
09:41:54  <indutny>does anyone here have a experience of submitting patch to v8?
09:42:00  <indutny>I've a problems with finding reviewer
10:03:37  <piscisaureus>indutny: yes
10:03:57  <piscisaureus>indutny: what's up? can you link me to the rietveld issue?
10:04:12  <indutny>http://code.google.com/p/v8/issues/detail?id=1639
10:04:42  <indutny>I can't create it without reviewer
10:04:45  <piscisaureus>indutny: !
10:04:58  <piscisaureus>indutny: nice. you actually understand that?
10:05:26  <indutny>yep
10:05:28  <indutny>some parts
10:06:08  <piscisaureus>indutny: just assign [email protected] and [email protected]
10:06:22  <piscisaureus>you could add vyacheslav as well because he knows you
10:07:29  <indutny>R=nickname1,nick2,nick3?
10:07:34  <indutny>how to assign multiple?
10:09:24  <indutny>wow it worked
10:09:26  <indutny>interested
10:09:28  <indutny>:)
10:09:34  <indutny>s/interested/interesting
10:09:50  <indutny>haha :D thanks
10:10:10  <piscisaureus>indutny: so can you now link me to the rietveld issue?
10:10:17  <indutny>one second
10:10:31  <indutny>http://codereview.chromium.org/7780033/
10:11:18  <piscisaureus>indutny: getting patches landed in v8 is not easy. but this is a good start
10:11:32  <indutny>mraleph guided me
10:11:35  <indutny>with code
10:15:28  <indutny>piscisaureus: hm...
10:15:39  <piscisaureus>wut?
10:15:43  <indutny>piscisaureus: is http://codereview.chromium.org/ correct for v8?
10:16:01  <piscisaureus>indutny: yes
10:16:03  <piscisaureus>that's okay
10:16:04  <indutny>ok
10:16:11  <indutny>because issue link is incorrect
10:16:15  <indutny>as it refers to chromium
10:16:17  <indutny>not v8
10:16:17  <piscisaureus>indutny: although, if you reference a bug you should prefix with v8:
10:16:22  <piscisaureus>e.g. bug=v8:123445
10:16:24  <indutny>fck
10:16:33  <piscisaureus>indutny: don't worry
10:16:45  <piscisaureus>indutny: the committer will fix up the commit message
10:16:58  <indutny>oh
10:16:59  <indutny>thanks
10:20:38  <indutny>is it possible to become node.js committer?
10:21:12  <indutny>I don't think I can make harm with our current release system
10:21:25  <indutny>and I'll work only on a debugger
10:25:36  <piscisaureus>indutny: it is not impossible but is not going to happen immediately
10:26:10  <indutny>ok
10:26:20  <indutny>should I ask ryah about that?
10:26:26  <indutny>or it's enough to ask you
10:26:44  <piscisaureus>indutny: we appreciate the work you are doing on the debugger very much
10:26:54  <piscisaureus>that's not it
10:27:09  <piscisaureus>indutny: you can ask ryah but he'll give you the same answer
10:27:13  <indutny>ok
10:27:19  <indutny>thank you!
10:50:55  <indutny>great Yang started reviewing it
10:51:06  <indutny>s/great/great!/
10:51:07  <indutny>haha
10:51:11  <indutny>not Great Yang
10:51:13  <indutny>lol :)
12:43:32  * bnoordhuisjoined
13:10:29  <dmkbot>joyent/libuv: piscisaureus: Specify error handling for uv_write and uv_connect - https://github.com/joyent/libuv/issues/185
13:11:26  <bnoordhuis>piscisaureus: I am not sure about errors that are deferred from a failed bind (EADDRNOTVAIL). <- is bind deferred on windows?
13:11:47  <piscisaureus>bnoordhuis: yes
13:11:53  <piscisaureus>errors are deferred
13:11:59  <piscisaureus>maybe not to write
13:12:02  <piscisaureus>to connect probably
13:12:40  <bnoordhuis>we discussed this last week, didn't we?
13:12:49  <bnoordhuis>i have this tremendous feeling of deja-vu :)
13:12:52  <piscisaureus>bnoordhuis: you mean with pieter?
13:12:55  <piscisaureus>bnoordhuis: yes
13:13:04  <indutny>piscisaureus: are you from cloudkick?
13:13:08  <bnoordhuis>what was the conclusion if any?
13:13:24  <bnoordhuis>defer everything but programmer error and system failure like ENOMEM?
13:13:26  <piscisaureus>bnoordhuis: the conclusion was that we maybe should defer all errors
13:13:27  <piscisaureus>indutny: yes
13:13:31  <indutny>cool
13:13:41  <bnoordhuis>rackspace!
13:13:47  <bnoordhuis>there is no cloudkick any more
13:14:01  <piscisaureus>bnoordhuis when do you want to go to sf?
13:14:14  <bnoordhuis>piscisaureus: a whole week away from mees :`(
13:14:30  <bnoordhuis>i don't know, i have to renew my passport first
13:14:46  <bnoordhuis>any preferences?
13:15:43  <piscisaureus>bnoordhuis: not really
13:15:56  <piscisaureus>it shouldn't interfere with jsconf.eu
13:16:03  <bnoordhuis>when is that?
13:16:06  <piscisaureus>also I think we need a few weeks to get clearance
13:16:16  <piscisaureus>bnoordhuis: 1+2 october
13:16:22  <bnoordhuis>oh, you're on the no-fly list?
13:16:32  <piscisaureus>bnoordhuis: ghe
13:16:49  <piscisaureus>bnoordhuis: no I believe you have to either get a visa or this visa waiver clearance
13:16:54  <piscisaureus>ot takes some time
13:17:22  <bnoordhuis>visa for a one week business trip?
13:17:26  * bnoordhuisgoogles
13:51:55  * ericktjoined
13:55:51  * dmkbot1joined
13:59:02  * dmkbotquit (Ping timeout: 252 seconds)
14:05:48  <dmkbot1>joyent/node: tmoeller: os.cpus() and os.getNetworkInterfaces() empty on windows - https://github.com/joyent/node/issues/1687
14:22:48  * dmkbot1quit (Remote host closed the connection)
14:23:07  * dmkbotjoined
14:52:04  <bnoordhuis>piscisaureus: what are you working on?
14:57:01  * ericktquit (Quit: erickt)
15:19:00  * isaacsjoined
15:26:50  * felixgejoined
15:32:07  <piscisaureus>bnoordhuis: sorry, I was working on node-ftp
15:32:15  <piscisaureus>but now I'm back on file streams
15:32:17  <piscisaureus>er
15:32:21  <piscisaureus>on tty streams
15:33:07  * felixgequit (Quit: http://www.debuggable.com/)
15:33:20  <piscisaureus>bnoordhuis: what are you working on?
15:33:35  <bnoordhuis>piscisaureus: fixing bugs
15:33:40  <bnoordhuis>but i want to get started on stdio
15:33:50  <bnoordhuis>btw, what are you doing with node-ftp
15:34:08  <piscisaureus>bnoordhuis: just helping out a guy here
15:34:57  <bnoordhuis>piscisaureus: okay
15:35:05  <piscisaureus>can I call you
15:35:11  <bnoordhuis>sure, skype?
15:35:23  <piscisaureus>just a sec, globbing a headphone
15:35:46  <bnoordhuis>okay, let me turn off destiny's child
15:36:22  <indutny>:)
15:36:23  <indutny>lol
15:38:04  <dmkbot>joyent/node: khoomeister: Any chance of a synchronous child_process.exec/spawn in the future? - https://github.com/joyent/node/issues/1167
15:39:26  * brsonjoined
15:56:55  * isaacsquit (Quit: isaacs)
15:59:53  * isaacsjoined
16:13:04  * dmkbotquit (Remote host closed the connection)
16:13:20  * dmkbotjoined
16:23:03  * brsonquit (Ping timeout: 276 seconds)
16:23:39  * brsonjoined
16:28:18  <dmkbot>joyent/libuv: piscisaureus: Specify error handling for uv_write and uv_connect - https://github.com/joyent/libuv/issues/185
16:36:26  <ryah>piscisaureus: https://github.com/joyent/libuv/tree/prefork what's the status on this
16:36:35  <ryah>piscisaureus: i want to give it to igor for testing
16:37:14  <piscisaureus>ryah: it needs rebasing
16:37:27  <piscisaureus>ryah: igor should just send the login details of his cluster
16:37:30  <indutny>ryah: hi!
16:37:43  <indutny>seen you in v8 AUTHORS list
16:38:43  <ryah>igorzi: you there?
16:38:50  <ryah>indutny: hi
16:40:26  <CIA-52>libuv: Ben Noordhuis master * rb450d87 / (src/unix/eio/eio.c src/unix/fs.c): eio: fix memory leak in eio__scandir() - http://git.io/MH2ItA
16:43:34  <CIA-52>node: Ben Noordhuis master * r7e7e983 / (3 files in 3 dirs): uv: upgrade to b450d87 - http://git.io/2Kxq3g
16:44:27  * brsonquit (Ping timeout: 250 seconds)
16:45:56  <ryah>isaacs: better to discuss here
16:46:07  <isaacs>oh, yeah, i guess so
16:46:08  <isaacs>:)
16:46:25  <ryah>isaacs: i think node_zlib.h isn't necessary
16:46:29  <isaacs>k
16:46:30  <ryah>isaacs: it's all private to that module
16:46:45  <isaacs>k
16:46:49  <piscisaureus>isaacs: do you have a list with compiled addons
16:46:55  <piscisaureus>that are on npm?
16:47:06  <piscisaureus>or is there a way to find those
16:47:23  <ryah>isaacs: instead of defines for the mode, use enums (so that we can see the symbols in the debugger)
16:47:39  <isaacs>piscisaureus: this is mostly compiled stuff: http://registry.npmjs.org/-/needbuild
16:47:46  <isaacs>ryah: ok
16:47:50  <ryah>enum node_zlib_modes { DEFLATE = 1, ... };
16:47:54  <isaacs>sure
16:48:57  <piscisaureus>isaacs: thnx - that was what I was looking fotr
16:48:58  <ryah>isaacs: you going to be in today?
16:49:08  <isaacs>ryah: yeah
16:49:29  <ryah>isaacs: come by when you're here - let's discuss node_zlib.cc - looks good but i have some comments
16:49:43  <isaacs>will do
16:50:11  <isaacs>it's working for short strings, but i'm running into odd data corruption when i throw more data at it, so something's definitely broken.
16:51:26  <ryah>isaacs: basically i think more should be done in js - like the callback queue
16:51:40  <ryah>but let's discuss
16:51:43  <isaacs>sure
16:57:13  * piscisaureusquit (Ping timeout: 260 seconds)
17:04:30  * igorziquit (Ping timeout: 252 seconds)
17:05:39  * bnoordhuisis off to dinner
17:08:15  <indutny>isaacs: do you need any help with zlib?
17:08:20  <indutny>I've some experience
17:08:32  <isaacs>indutny: it's not too tricky.
17:08:38  <indutny>yeah, but anyway
17:08:42  <indutny>feel free to ask me anything
17:08:46  <isaacs>sure
17:08:47  <isaacs>thanks :)
17:08:50  <indutny>np
17:09:01  * igorzijoined
17:10:17  * dmkbotquit (Remote host closed the connection)
17:10:36  * dmkbotjoined
17:12:17  <isaacs>indutny: actually, if you can figure out why i'm getting garbage data coming out of deflate/inflate, that'd be helpful. but i think i'm just going to have to trace through it
17:13:06  <indutny>zlib branch, right?
17:13:33  <indutny>or not
17:13:55  <indutny>isaacs: where can I find code to look at?
17:14:04  <isaacs>my zlib branch
17:14:13  <indutny>ah
17:14:14  <indutny>yours
17:14:15  <indutny>ok
17:14:15  <isaacs>i have a rudimentary test at test/simple/test-zlib.js
17:14:20  <isaacs>yeha, isaacs/node
17:16:24  <isaacs>indutny: i'm pretty sure i'll find it today, though. if you see anything obvious, then that'd be helpful, but i don't want to take you away from working on the debugger stuff.
17:16:36  <indutny>k
17:17:03  <indutny>right now I'm on nodejitsu stuff
17:17:08  <isaacs>oh, neat
17:17:08  <indutny>but I'll take a break soon
17:20:40  <indutny>wow, a lot of LoC
17:20:45  <indutny>oh, copy-paste, actually
17:20:46  <indutny>:D
17:20:53  <indutny>at the end
17:39:54  <indutny>submitting patches to v8 is really hard
17:40:14  <indutny>but that's definitely good for ensuring that patches are good enough
17:40:33  * dmkbotquit (Remote host closed the connection)
17:40:44  * dmkbotjoined
17:40:46  <ryah>they control their code base - which is nice.
17:40:55  <ryah>i try to control mine - but do a much worse job
17:41:13  <ryah>would love to have their code review system for node
17:41:31  <isaacs>ryah: they = ?
17:41:34  <isaacs>v8?
17:42:25  <indutny>ryah: yeah, but have you seen AUTHORS list? :)
17:42:30  <indutny>not much people here
17:42:31  <indutny>:)
17:42:34  <indutny>compare to node's
17:42:40  <indutny>;)
17:42:57  <indutny>I think there should be a balance between hardness and openness
17:46:00  * mralephjoined
17:47:24  <indutny>mraleph: http://codereview.chromium.org/7780033/
17:47:27  <indutny>mraleph: hello!
17:51:10  <mraleph>indutny: hi
17:51:15  <mraleph>indutny: yep I saw that
17:51:36  <mraleph>indutny: I recomend running aaaaall tests :-)
17:51:39  <indutny>mraleph: yeah
17:51:43  <indutny>that's a problem
17:51:45  <indutny>:)
17:51:49  <indutny>that takes toooo long
17:51:59  <mraleph>how many cores your box has?
17:52:06  <indutny>mraleph: are all tests passing on bleeding_edge
17:52:12  <indutny>mraleph: i5 - 2 cores 4 threads
17:52:14  <mraleph>yes
17:52:16  <mraleph>ok
17:52:23  <mraleph>then run tests with -j4
17:52:32  <mraleph>or -j3
17:52:54  <indutny>mraleph: oh
17:52:59  <indutny>mraleph: that's much better
17:53:00  <indutny> :D
17:53:46  <indutny>at least it builds in 2 min
17:55:55  <mraleph>imagine how cool it is with -j12 ;-)
17:56:24  <indutny>mraleph: :)
17:56:41  <indutny>mraleph: is that chromium's buildserver?
17:56:51  <indutny>mraleph: or your PC?
17:57:09  <mraleph>:-)
17:57:29  <mraleph>that's my pocket alarm clock ;-)
17:57:46  <indutny>nice :)
18:00:48  <indutny>mraleph: aarrgh, stupid tests :)
18:00:51  <indutny>getting me crazy
18:00:59  <indutny>s/getting/making
18:06:12  * dmkbotquit (Remote host closed the connection)
18:06:26  * dmkbotjoined
18:20:14  <isaacs>indutny: congrats, you're the mvp of node now, apparently.
18:20:25  <isaacs>i think that's kinda like being the mayor of a coffee shop in 4square.
18:20:26  <indutny>isaacs: hahaha
18:20:39  <indutny>isaacs: yeah
18:20:58  <indutny>how can I delete myself from this stuff
18:21:05  <indutny>I hate receiving notifications from them
18:22:27  <isaacs>indutny: i just usually set up a filter in gmail to delete any messages i don't like
18:23:11  * isaacsquit (Quit: isaacs)
18:34:56  <CIA-52>libuv: Bert Belder master * rf790b68 / (5 files in 3 dirs): Remove uv_init from uv.h and windows backend - http://git.io/ou4eMQ
18:34:59  <CIA-52>libuv: Ryan Dahl master * r828b6eb / src/unix/core.c : unix: remove uv_init - http://git.io/1Ku4Tw
18:35:00  <CIA-52>libuv: Bert Belder master * rbb15c0e / src/win/timer.c : win: make uv_hrtime() thread-safe - http://git.io/I2E1vA
18:35:02  <CIA-52>libuv: Bert Belder master * r0dc564a / (33 files): Remove uv_init calls from tests and benchmarks - http://git.io/PzeJQg
18:39:15  * graydonjoined
18:48:57  * piscisaureuzjoined
18:49:03  <piscisaureuz>hey
18:49:33  <ryah>piscisaureuz: hi
18:49:36  <piscisaureuz>ill not be home until 21.15-31.30
18:50:14  <ryah>piscisaureuz: im not sure what that means, but okay
18:50:19  <piscisaureuz>so feel free to do the scrum without me if necessary
18:50:45  <ryah>piscisaureuz: ok
18:50:51  <ryah>piscisaureuz: i landed your noinit
18:50:58  <piscisaureuz>nice
18:51:16  <ryah>piscisaureuz, igorzi: void uv_pipe_open(uv_pipe_t*, uv_file file);
18:51:27  <ryah>piscisaureuz, igorzi: are either of you against this?
18:53:08  <piscisaureuz>let's discuss in 30 minutes
18:53:13  * piscisaureuzquit (Client Quit)
18:56:26  <ryah>bnoordhuis: is your modifications for gyp/sunos up somewhere?
18:56:50  <bnoordhuis>ryah: https://github.com/bnoordhuis/node/compare/gyp
18:57:04  <ryah>merci
18:57:13  <igorzi>ryah: is this for stdio? or something else?
18:57:18  <ryah>igorzi: yes
18:57:47  <ryah>igorzi: but i think we're going to need something more general than just for studio
18:58:02  <ryah>i'd like to have a way to open arbitrary handles
18:58:37  <ryah>bnoordhuis: what of this can we land?
18:59:07  <bnoordhuis>ryah: nothing, linking doesn't work reliable yet
18:59:30  <bnoordhuis>let me quickly check if the last gyp upgrade fixed that
18:59:39  <igorzi>ryah: it should probably return uv_stream_t*? (similar to this https://github.com/joyent/libuv/blob/master/include/uv.h#L370)
19:04:23  <igorzi>ryah: are we doing a call now? or later?
19:09:37  <bnoordhuis>ryah: https://github.com/bnoordhuis/libuv/compare/stdio-ideetje
19:13:13  <ryah>https://github.com/joyent/node/blob/master/src/node.js#L261-268
19:16:25  <indutny>ryah: shouldn't it be defined as static value?
19:16:29  <indutny>ah
19:16:34  <indutny>see now an `if`
19:18:44  * piscisaureusjoined
19:21:35  * piscisaureus_joined
19:21:47  <piscisaureus_>hey
19:21:53  <piscisaureus_>I'm home
19:21:58  <piscisaureus_>you guys already had the call?
19:22:08  <piscisaureus_>bnoordhuis ryah igorzi: ^ ?
19:22:08  <igorzi>damn it.. skype crashed
19:22:22  <bnoordhuis>piscisaureus: calling right now
19:22:25  <bnoordhuis>join the fun
19:22:31  <piscisaureus_>add ne
19:22:33  <piscisaureus_>add em
19:22:35  <piscisaureus_>add me
19:24:59  * ryahtopic: v0.5.7 https://github.com/joyent/node/issues?state=open&milestone=3
19:25:05  * piscisaureusquit (Ping timeout: 260 seconds)
19:26:00  * DrPizzajoined
19:56:58  * brsonjoined
20:04:39  * indutnyjoined
20:19:16  * isaacsjoined
20:23:17  <piscisaureus_>I am installing windows on my macbook pro
20:23:22  <piscisaureus_>that's so backwards :-(
20:25:49  <DrPizza>bootcamped or virtualized?
20:26:45  <piscisaureus_>bootcamped
20:27:30  <DrPizza>good.
20:27:37  <DrPizza>step 2: delete mac os x
20:27:44  <DrPizza>(if you don't, mac os x might try to delete your windows)
20:28:09  <piscisaureus_>c'mon drpizza
20:28:17  <piscisaureus_>I held you in higher regard
20:28:22  <DrPizza>lion has a bug that it wipes out the first 500 MB of the disk or something
20:28:26  <DrPizza>if you upgrade to lion when bootcamped
20:28:30  <DrPizza>it just trashes the Windows partition
20:28:35  <piscisaureus_>oh i am on lion already
20:28:42  <piscisaureus_>besides lion is the first partition
20:30:25  <DrPizza>ah, probably safe then
20:38:18  * brsonquit (Ping timeout: 260 seconds)
20:56:12  <igorzi>piscisaureus_: DrPizza: do you know if people that run node on windows have a need to do registry access? and if yes, what do they use?
21:00:08  * dmkbotquit (Remote host closed the connection)
21:00:21  * dmkbotjoined
21:07:03  <bnoordhuis>hah, i can't get x64 v8 to link on the no.de machines
21:07:21  <bnoordhuis>runs out of memory even with 2GB of memory
21:07:25  <bnoordhuis>the ia32 build works though
21:15:42  <CIA-52>libuv: Ben Noordhuis master * r2f2223f / src/unix/pipe.c : sunos: fix compiler errors, sun is a #define - http://git.io/IGltpA
21:18:04  <DrPizza>igorzi: I'm not sure, I think a good argument can be made that having registry access makes it more "windows native", if you see what I mean
21:18:24  <DrPizza>igorzi: otherwise, I think they'd probably have to resort to doing something like spawning reg.exe and praasing its (horrible) output
21:18:30  <DrPizza>parsing*
21:21:09  <piscisaureus_>igorzi: what kind of registry access do you think people need
21:21:17  <piscisaureus_>that is a more resonable question
21:21:27  <piscisaureus_>maybe if they are writing shell scripts
21:21:36  <piscisaureus_>but I can't really see it for server
21:21:44  <CIA-52>libuv: Ben Noordhuis master * r9bd8bd7 / uv.gyp : gyp: add sunos config section - http://git.io/L_5itg
21:22:32  <DrPizza>piscisaureus_: yeah Ican see it more for admin type tsaks
21:22:33  <DrPizza>tasks
21:22:58  <piscisaureus_>right
21:23:05  <piscisaureus_>but I'ts hardly a priority
21:23:38  * rmustaccjoined
21:28:23  <ryah>bnoordhuis: im going to manage this release - it's going to be released on thrusday
21:28:58  <ryah>bnoordhuis: you do v0.5.8
21:29:08  <bnoordhuis>er, sure
21:29:31  <ryah>we need a new release sooner rather than later because NPM on unix is broken
21:30:05  <isaacs>ryah: yes, your offhand mention about needing to keep a reference to the buffer was the reason i was getting bizarre corrupted data in teh zlib binding, btw.
21:30:16  <ryah>isaacs: k, coool
21:30:18  <dmkbot>joyent/node: aaronj1335: REPL sources file from NODE_STARTUP env var - https://github.com/joyent/node/issues/1688
21:30:21  <ryah>s/o//
21:32:29  <CIA-52>libuv: Ryan Dahl master * r4484d61 / (9 files in 5 dirs):
21:32:30  <CIA-52>libuv: Add interfaces for uv_pipe_open, uv_tty_init, uv_tty_set_mode
21:32:30  <CIA-52>libuv: Nothing works - no tests. This is just to coordinate efforts between Bert
21:32:30  <CIA-52>libuv: and I. - http://git.io/cLWe7g
21:48:11  <CIA-52>node: Ben Noordhuis master * r0d80040 / (41 files in 6 dirs): uv: upgrade to 9bd8bd7 - http://git.io/8Fet0g
21:49:21  <bnoordhuis>argh, v8 rebuilds *again*
21:49:29  * bnoordhuisbangs head against table
21:51:36  * ryahplaces a pillow on bnoordhuis's bed
21:51:38  <ryah>er
21:51:39  <ryah>table
21:51:55  <bnoordhuis>heh, thanks
21:52:12  <piscisaureus_>ryah is so nice today
21:55:18  <dmkbot>joyent/node: tshinnic: Continuing a thought: JS-level util.errnoException - https://github.com/joyent/node/issues/1689
21:55:19  <dmkbot>joyent/node: tshinnic: Continuing a thought: JS-level util.errnoException - https://github.com/joyent/node/issues/1689
21:57:07  <piscisaureus_>I don't quite dig the external connectors on this mac pro. esata->no_luck. vga->na-ah. hdmi->nope firewire->keine. usb->only2
21:58:01  <DrPizza>thunderbolt
21:58:49  <piscisaureus_>yes but can you actually buy thunderbolt drives?
21:59:03  <piscisaureus_>sure, from apple. It's only gonna cost me an arm and a leg
21:59:08  <piscisaureus_>and a kidney prolly
21:59:17  <igorzi>piscisaureus_: DrPizza: when I see this: https://github.com/joyent/node/issues/1167.. i think surely people that need this will need to do registry access
21:59:54  <igorzi>(when using node as a scripting tool)
22:00:22  <DrPizza>piscisaureus_: yes, you can get a raid enclosure for like 4999 + drives
22:00:23  <DrPizza>or something
22:00:27  <DrPizza>er
22:00:28  <DrPizza>$999
22:01:19  <ryah>igorzi: why do you need reg access for sync spawn?
22:01:39  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/libuv/compare/tcp-close-again <- review please?
22:01:57  <piscisaureus_>ok. that's the last think I do tonight :-)
22:02:16  <bnoordhuis>piscisaureus_: fun fact, that crash in test/pummel/test-keep-alive.js went away a couple of node commits ago
22:02:34  <igorzi>ryah: no, i don't mean i need reg access for sync spawn. i mean i see that people use node as a scripting tool (which is why sync spawn was requested), so it seems that same people will need reg access on windows
22:02:39  <bnoordhuis>so it wasn't (well, not entirely) that patch
22:04:33  <ryah>igorzi: what's requried for registry access? it's just a file isn't it?
22:05:08  * mraleph1joined
22:05:17  <DrPizza>no
22:05:20  <DrPizza>win32 api
22:05:34  * mralephquit (Read error: Connection reset by peer)
22:06:13  <ryah>sounds like a good external module
22:06:26  <ryah>*addon
22:07:25  <CIA-52>libuv: Ryan Dahl master * re7eeacb / src/unix/tty.c : unix: implement uv_tty_set_mode - http://git.io/5zcmIQ
22:07:44  <CIA-52>node: Ben Noordhuis master * r77e4abb / deps/v8/tools/gyp/v8.gyp : v8: add platform-solaris.cc to gyp build (+7 more commits...) - http://git.io/MQcg-Q
22:09:07  <bnoordhuis>ryah: two things broken with gyp right now: --debug doesn't work and neither does `make install`
22:09:20  <igorzi>ryah: right
22:09:39  <piscisaureus_>bnoordhuis: lgtm - but do you have any idea why the crash went away?
22:09:48  <piscisaureus_>or what was/is the root cause of it?
22:10:00  <bnoordhuis>piscisaureus_: something with the handle wrap / stream wrap
22:10:21  <bnoordhuis>it was corrupting memory somehow and now it no longer does
22:10:28  <piscisaureus_>bnoordhuis: we actually fixed it or it just accidentally went away?
22:10:30  <bnoordhuis>i have half a mind to git bisect it
22:10:41  <bnoordhuis>the latter, i think, i don't see a conscious commit
22:10:42  <CIA-52>node: Ryan Dahl master * r51f2e84 / (10 files in 5 dirs): Upgrade libuv to e7eeacb - http://git.io/0l4d0Q
22:10:42  <CIA-52>node: Ryan Dahl master * rcaaa59c / (lib/net_uv.js src/pipe_wrap.cc src/pipe_wrap.h):
22:10:42  <CIA-52>node: Wrap uv_pipe_open, implement net.Stream(fd);
22:10:42  <CIA-52>node: Fixes simple/test-child-process-ipc on unix. - http://git.io/qydYRA
22:11:21  * ericktjoined
22:12:05  <ryah>bnoordhuis: can we push https://github.com/joyent/node/commit/77e4abbc3e66505af89c57cd7bff555890a33f3f upstream?
22:12:14  <bnoordhuis>ryah: yes
22:12:18  <bnoordhuis>can and will
22:12:36  <ryah>they probably need this too https://github.com/joyent/node/commit/d104e5b91cfa3ef3ef846d5a0ab07c0336263a92
22:12:53  <ryah>not sure if it should be in v8-node.gyp or just v8-base.gyp
22:12:57  <piscisaureus_>it would be nice to have mingw support in gyp
22:13:00  <ryah>probably v8-base
22:13:06  <ryah>piscisaureus_: yes
22:13:07  <piscisaureus_>shouldn't be so hard - mingw is just a gcc/make platform
22:13:10  <ryah>piscisaureus_: it's doable
22:13:35  <piscisaureus_>I haven't looked at gyp's internals at all so far
22:13:43  <ryah>don't waste your time though
22:13:44  <piscisaureus_>and are they open to patches or notsomuch?
22:13:55  <ryah>they are open to patches
22:13:57  * mraleph1quit (Quit: Leaving.)
22:14:15  <bnoordhuis>ryah: v8-base == v8.gyp?
22:15:56  <ryah>bnoordhuis: yeah, sorry
22:17:53  <ryah>[01:11|% 100|+ 250|- 10]: Done
22:17:57  <ryah>^-- make test on OSX
22:21:40  <bnoordhuis>http://code.google.com/p/gyp/issues/detail?id=216
22:21:45  <bnoordhuis>hey, good work
22:21:50  <bnoordhuis>progress!
22:26:58  <isaacs>ryah: so.. if i create a ReqWrap, is there any reason to keep it around after the uv_work is done?
22:39:08  <isaacs>ryah: npm fail. passing in the CA cert doesn't work.
22:53:49  <isaacs>also, looks like we're getting the wrong errno from failed fs operations, and that breaks npm's mkdir function
22:53:57  <CIA-52>node: Ben Noordhuis master * r7180327 / (deps/v8/build/common.gypi deps/v8/tools/gyp/v8-node.gyp): v8: move __C99FEATURES__=1 into common.gypi - http://git.io/FM1D2g
22:54:23  <bnoordhuis>isaacs: wrong errno?
22:54:45  <isaacs>bnoordhuis: yeah, I'm getting -1 Unknown Error when it's clearly an EEXIST
22:55:01  <igorzi>isaacs: is this on unix or windows?
22:55:02  <bnoordhuis>isaacs: in c++ or js land?
22:55:39  <isaacs>unix, js land, but it's coming out of ErrnoException(req_wrap.req.errorno)
22:56:32  <isaacs>> fs.mkdirSync(process.cwd(), 0755)
22:56:33  <isaacs>Error: UNKNOWN, Unknown error: 0 '/Users/isaacs/dev-src/js/node'
22:56:57  <bnoordhuis>we probably don't have a mapping for it yet
22:57:02  <isaacs>> try { fs.mkdirSync(process.cwd(), 0755) } catch (er) { console.error(er) }
22:57:02  <isaacs>{ [Error: UNKNOWN, Unknown error: 0 '/Users/isaacs/dev-src/js/node'] errno: -1, code: 'UNKNOWN', path: '/Users/isaacs/dev-src/js/node', syscall: 'mkdir' }
22:57:12  <bnoordhuis>what errno do you see with strace / truss?
22:57:46  <isaacs>hm... i don't have either of those, it seems like..
22:58:06  <isaacs>that's odd, i thought strace was included on the mac by default.
22:58:16  <rmustacc>isaacs: You could probably use dtruss.
23:02:52  <isaacs>mkdir("/Users/isaacs/dev-src/js/node\0", 0x1ED, 0x100811490) = -1 Err#17
23:03:16  <isaacs>sure enough:
23:03:17  <isaacs>> require("constants").EEXIST
23:03:17  <isaacs>17
23:03:39  <isaacs>bnoordhuis: where are these mappings you speak of?
23:04:07  <bnoordhuis>isaacs: they're in uv_err_name() and uv_strerror()
23:04:20  <bnoordhuis>isaacs: the conversion is done in node_file.cc in function FSError()
23:04:53  <bnoordhuis>conversion, that is, from c string to js exception
23:05:00  <isaacs>bnoordhuis: i don't see any FSError() in node_file.cc
23:05:09  <bnoordhuis>what commit are you at?
23:05:32  <isaacs>* caaa59c Ryan Dahl (HEAD, ry/master, master) Wrap uv_pipe_open, implement net.Stream(fd); (56 minutes ago)
23:06:00  <isaacs>just pulled your last one
23:06:24  <bnoordhuis>er... line 240 in node_file.cc?
23:06:32  <isaacs>oh, nvm, vim. i'm dumb.
23:10:19  <isaacs>ah, yeah, definitely missing in deps/uv/src/uv-common.c
23:11:16  <isaacs>and deps/uv/include/uv.h
23:11:43  <bnoordhuis>isaacs: go ahead and add it
23:11:49  <isaacs>k
23:13:57  <isaacs>what's the thing to add to deps/uv/src/win/error.c?
23:16:51  <isaacs>it looks like either ERROR_ALREADY_EXISTS, ERROR_FILE_EXISTS, or ERROR_DUP_NAME all are basically the same as EEXIST, but i could be misreading this doc
23:17:07  <isaacs>oh, i guess ERROR_DUP_NAME is a networking thing...
23:17:25  <bnoordhuis>igorzi: ^
23:17:28  <isaacs>or maybe ERROR_CANNOT_MAKE?
23:18:00  <ryah>isaacs: the ReqWrap should be closed after the callback comes back from the thread pool
23:18:29  <isaacs>ryah: when the write() cb gets called?
23:19:32  <ryah>isaacs: yes
23:19:37  <isaacs>ryah: k
23:21:01  <ryah>piscisaureus_: when you do this tty stuff - it'd be good to stub out where you expect the vt100 parsing to go
23:21:24  <isaacs>ryah: it looks like ~ReqWrap does the Dispose() call...?
23:21:28  <ryah>piscisaureus_: that way we can progressively add functionality
23:22:31  <isaacs>ryah: you were talking about some need to manually close or clean up the ReqWrap, or was i misinterpreting you?
23:23:55  <ryah>isaacs: yeah.. mm. one sec
23:25:10  <ryah>dang - it seems like there might be a memory leak
23:25:16  <ryah>are we never calling ~ReqWrap ?
23:25:47  <ryah>oh i see - it's done in c++
23:26:59  <ryah>github is making it harder and harder to have links
23:27:33  <ryah>isaacs: take socket.shutdown() as an example
23:27:34  <ryah>https://github.com/joyent/node/blob/718032777d486b9dcedcfeabd9a35e8ab5d5e15b/src/stream_wrap.cc#L258-284
23:27:50  <ryah>isaacs: after the MakeCallback it calls "delete req_wrap"
23:29:31  <ryah>oh the "y" hotkey is very useful
23:29:42  <ryah>on github
23:30:16  <isaacs>i see.
23:30:40  <isaacs>so, i'd do that in my Flate::UVProcessAfter function, after calling the cb.
23:31:06  <ryah>yeah
23:31:19  <ryah>seems like you could shorten that name to Flate::After
23:31:40  <isaacs>sure
23:31:58  <isaacs>once the queue processing is moved out to js, the "Process" bit doesn't make much sense :)
23:32:16  * brsonjoined
23:34:12  <isaacs>anyone wanna review? https://gist.github.com/1212769
23:40:15  <bnoordhuis>sure, give me a sec
23:41:08  <bnoordhuis>isaacs: lgtm
23:42:10  <isaacs>bnoordhuis: what about the windows ERROR_FILE_EXISTS mapping? just go with it for now and let igorzi complain if it's wrong? ;)
23:42:32  <isaacs>or do you have reason to believe that my guess is accurate?
23:42:49  <bnoordhuis>isaacs: your guess is as good as mine, so...
23:42:58  <bnoordhuis>igorzi piscisaureus_: ^ don't say we didn't warn you
23:44:10  <isaacs>heh
23:44:19  <isaacs>hm... it's still failing the test in node, though...
23:45:51  <isaacs>oh, i see, i'm getting uv's errno, not require("constants").EEXIST
23:45:53  <isaacs>that's weird.
23:46:31  <isaacs>npm can live with it, though
23:46:39  <bnoordhuis>isaacs: uv errno's don't align with the system's errno's
23:46:45  <ryah>sy errnos aren't exposed
23:46:46  <isaacs>right
23:46:47  <ryah>sys
23:46:59  <isaacs>k. guess they'd be crazy different across platforms anyway.
23:47:10  <isaacs>can we get uv's errno's on require('constants') at some point?
23:47:23  <bnoordhuis>yes, i wrote a patch for that actually
23:47:47  <ryah>require('constants').EEXIST should be 'EEXIST'
23:47:52  <ryah>not an integer
23:47:54  <CIA-52>libuv: isaacs master * r3c00d87 / (4 files in 4 dirs):
23:47:54  <CIA-52>libuv: Add EEXIST handling
23:47:54  <CIA-52>libuv: Additionally, map ERROR_ALREADY_EXISTS to EEXIST on Windows. I'm a bit
23:47:54  <CIA-52>libuv: unsure about this mapping. Could a windows person confirm? - http://git.io/UrtYXw
23:48:06  <ryah>we want to use strings instead of integer error codes in the future
23:48:39  <isaacs>ryah: that's bizarre.
23:48:55  <isaacs>ryah: but bizarre things can be nice :)
23:49:07  <bnoordhuis>oh, i wrote this: https://github.com/bnoordhuis/node/compare/remap-uv-errors
23:49:33  <bnoordhuis>er, the one that counts: https://github.com/bnoordhuis/node/compare/remap-uv-errors#diff-3
23:49:58  <ryah>integers are not nice to use
23:50:14  <bnoordhuis>i like integers
23:50:18  <bnoordhuis>my best friends are integers
23:51:08  <ryah>ideally v8 is going to optimize our string cmps like crazy
23:51:28  <bnoordhuis>i confess i don't like strings and i know it's... irrational
23:51:50  * ericktquit (Ping timeout: 276 seconds)
23:52:13  <bnoordhuis>are there good benchmarks for that?
23:52:20  <bnoordhuis>v8's optimizer is quite moody
23:52:24  <ryah>bnoordhuis: no
23:52:25  <ryah>:)
23:54:21  <isaacs>what's the process for pulling uv changes into node/deps/uv?
23:56:32  <bnoordhuis>isaacs: (cd deps/uv && rsync -av --exclude=.git ../../../libuv/ . && git clean -dfx) && git commit . -m "uv: upgrade to <commit>"
23:56:44  <bnoordhuis>that's how i do it anyway :)
23:56:49  <isaacs>k, great.
23:57:45  <bnoordhuis>there should be `git add .` in there if you add new files
23:57:49  <ryah>hm
23:57:57  <ryah>use git archive
23:58:21  <bnoordhuis>if you give me a good reason why
23:58:23  <ryah>rm -rf ~/projects/node/deps/uv
23:58:28  <ryah>git archive --prefix=uv/ HEAD | (cd ~/projects/node/deps; tar -xf - )
23:58:37  <ryah>git add ~/projects/node/deps/uv
23:58:54  <bnoordhuis>bad reason, more typing