00:00:22  <isaacs_away>piscisaureus_: nono, you gain reputability by owning it.
00:00:30  <isaacs_away>trust me. people love me because i've fixed so many npm bugs.
00:00:34  <isaacs_away>no one remembers that i caused all of them.
00:00:36  <isaacs_away>:)
00:02:15  <piscisaureus_>isaacs_away: the issue was https://github.com/joyent/node/issues/2635
00:02:57  <isaacs_away>right, thanks
00:02:59  <isaacs_away>closed
00:06:28  <isaacs_away>test, please: http://nodejs.org/dist/v0.6.10/node-v0.6.10-RC0.tar.gz
00:09:08  * AvianFluquit (Quit: Leaving)
00:30:56  <isaacs_away>bnoordhuis: https://gist.github.com/1726689 <-- this make any sense to you?
00:31:41  <bnoordhuis>isaacs_away: yeah - he's probably got libv8 installed system-wide
00:32:00  <bnoordhuis>or she
00:32:27  <bnoordhuis>the beard seems to suggest it's a he
00:32:38  <isaacs_away>lol
00:32:41  <isaacs_away>you never know, man.
00:33:01  <isaacs_away>(i'm assuming from your voice and the fact that i've seen you in person and you're way taller than me that you'er a man.)
00:34:14  <bnoordhuis>oh, that's from all the hormonally enhanced meat i ate in my youth
00:34:17  <isaacs_away>SmartOS: [00:56|% 100|+ 327|- 2]: Done
00:34:30  <isaacs_away>lol
00:36:31  <isaacs_away>looking good so far.
00:36:55  <isaacs_away>windows compiling
00:46:09  <isaacs_away>d'oh!!
00:46:13  <isaacs_away>almost a perfect swish
00:48:37  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
00:51:21  <mjr_>bnoordhuis: did ircretary tell you about my issue?
00:51:36  <bnoordhuis>mjr_: no
00:51:43  <mjr_>oh, let me find it
00:51:54  <mjr_>Uncaught and I think uncatchable exception in net.js
00:52:29  <mjr_>Uncaught exception: Error: accept Unknown system errno 103 at errnoException (net.js:640:11) at TCP.onconnection (net.js:825:24)
00:52:37  <bnoordhuis>oh, that shouldn't happen
00:52:42  <bnoordhuis>103 is ECONNREFUSED
00:52:43  <mjr_>My thoughts exactly
00:52:51  <isaacs_away>bnoordhuis: you should add "/msg ircretary notes" to your irc signon command list
00:52:54  <bnoordhuis>err, assuming you run on linux
00:52:59  <mjr_>ECONNABORTED 103 /* Software caused connection abort */
00:53:06  <bnoordhuis>oh, i was close
00:53:07  <mjr_>That's my Linux errno.h header.
00:53:23  <mjr_>WTF even is that in this case?
00:54:02  <bnoordhuis>mjr_: is this you making a http request?
00:54:57  <mjr_>I don't know. That's the entire stack.
00:55:21  <bnoordhuis>looked it up in net.js, it's the server receiving a network error
00:55:39  <bnoordhuis>it's possible that the other end sends a RST
00:56:01  <bnoordhuis>do you happen to know if it's http or https?
00:56:18  <mjr_>no idea. I do know that this is not https client.
00:56:29  <mjr_>But this process handles both http and https server, and it does a ton of http client.
00:56:48  <mjr_>Looks like it only happens a few times a day.
00:57:19  <mjr_>From the logs, I don't see any obvious common context.
00:57:30  <mjr_>But these are all very busy processes.
00:58:53  <mjr_>Yet another exciting artifact from our high traffic node environment, I'm sure.
00:59:05  <bnoordhuis>no doubt :)
00:59:21  <bnoordhuis>it's kind of hard to say what it is exactly
00:59:38  <bnoordhuis>but it could, for example, be a client that resets the connection halfway through the tcp handshake
00:59:52  <mjr_>That's what I was thinking, but I'm not sure what Linux does in that case.
00:59:53  <bnoordhuis>we need to map the errno obviously but it's not a really unexpected error
01:00:57  <TooTallNate>isaacs_away: re: node-tar, when do "ignoredEntry", "longPath", and "longLinkpath" get fired?
01:01:00  <mjr_>OK, well let me know how I can help.
01:01:06  <mjr_>Do you want me to file a github issue on this one?
01:01:25  <isaacs_away>TooTallNate: so, in tarballs, you can have "extended header" files.
01:01:27  <bnoordhuis>mjr_: https://github.com/torvalds/linux/blob/dcd6c92/net/ipv4/af_inet.c#L663 <- that's what it does :)
01:01:30  <bnoordhuis>and yes, please
01:01:43  <isaacs_away>TooTallNate: which aren't files that should be unpacked, but sort of like header information or the next file.
01:01:49  <mjr_>oh man, linux is on github. I forgot about that.
01:01:56  <mjr_>badass
01:02:38  <TooTallNate>isaacs_away: ok so i would assume the "extendedHeader" handles that case?
01:02:49  <isaacs_away>TooTallNate: so, for instance, if you have a file with a 1024 character filename, you put the first 155 chars in the "path" field, and then have a "longPath" where the body of the longPath file is the long filename.
01:02:59  <isaacs_away>TooTallNate: yes, node-tar never *writes* longPath or longLinkpath files.
01:03:04  <isaacs_away>TooTallNate: that's a gnutar-ism
01:03:13  <isaacs_away>TooTallNate: welcome to the tar pits ;)
01:03:22  <TooTallNate>hahaha
01:03:24  <mjr_>bnoordhuis: so by that af_inet.c example, it sound alike this must be client, not server.
01:03:28  <mjr_>good one
01:03:29  <TooTallNate>well im just extracting node tar.gzs
01:03:36  <isaacs_away>TooTallNate: yeah.
01:03:44  <isaacs_away>TooTallNate: for the most part, you only care about regular entries
01:03:57  <TooTallNate>isaacs_away: but it would be nice to be able to pass Extractor some kind of filter function
01:04:02  <TooTallNate>but not too big a deal
01:04:17  <bnoordhuis>mjr_: https://github.com/torvalds/linux/blob/dcd6c92/net/socket.c#L155 <- that's the server counterpart
01:04:17  <isaacs_away>TooTallNate: you can
01:04:32  <isaacs_away>TooTallNate: { filter: function () { this === entry, return true or false } }
01:04:43  <TooTallNate>isaacs_away: perfect :)
01:04:44  <isaacs_away>TooTallNate: at least, i think that's how it works
01:04:49  <bnoordhuis>mjr_: sorry, copy/paste error: https://github.com/torvalds/linux/blob/dcd6c92/net/socket.c#L1552
01:04:52  <mjr_>bnoordhuis: oh, sure enough. Something about my search failed.
01:04:52  <isaacs_away>TooTallNate: whatcha doin?
01:05:10  <TooTallNate>isaacs_away: gotta extract the header and gyp-related files
01:05:15  <TooTallNate>for a given node version
01:05:29  <isaacs_away>TooTallNate: neat.
01:05:37  <TooTallNate>isaacs_away: i don't see "filter" in extract.js
01:05:39  <isaacs_away>TooTallNate: from the dist tarball?
01:05:48  <TooTallNate>isaacs_away: yes, dist tarball
01:05:50  <isaacs_away>TooTallNate: filter is implemented in the fstream parent classes.
01:06:05  <isaacs_away>TooTallNate: i think i do this in npm somewhere.
01:06:44  <isaacs_away>i hadn't thought of pulling that stuff out of the dist tarballs directly. that's clever.
01:07:02  <dap>isaacs_away: I've finished the fix for #2667 and got it wrapped up in a commit on my github fork. What do I do now? Just submit a pull request?
01:07:13  <isaacs_away>dap: yes, please.
01:07:20  <TooTallNate>isaacs_away: ya, that way we can store them in like ~/.node-gyp or something, and not have to re-download the dev files
01:07:22  * bnoordhuissigns off for today
01:07:26  <isaacs_away>dap: put "Fix #2667" in the commit message somewhere.
01:07:28  <isaacs_away>g'nite bnoordhuis
01:07:41  <bnoordhuis>thanks, good luck with release
01:07:47  <bnoordhuis>*the release
01:07:54  <TooTallNate>isaacs_away: one stumbling block so far is needed a precompiled node.lib file for windows somewhere on the web so node-gyp can download it
01:08:03  <TooTallNate>s/needed/needing
01:08:14  <isaacs_away>hm.
01:08:23  <isaacs_away>we upload a pdb file, is that useful?
01:08:40  <TooTallNate>no i don't think so, could be wrong though
01:08:47  <isaacs_away>k
01:08:50  <TooTallNate>but on windows, the modules link to node.lib, so it needs that
01:09:15  <TooTallNate>and that means someone has to host the .libs and compile them for releases, haha
01:09:28  <TooTallNate>but ideally, it would only be minor point releases that matter
01:09:33  <TooTallNate>0.6, 0.8...
01:09:35  * bnoordhuisquit (Read error: Operation timed out)
01:10:05  <TooTallNate>the patch version shouldn't change the ABI (if we're doing it right), so minor versions should be all that's needed
01:10:13  <dap>isaacs_away: thanks, done.
01:10:17  <isaacs_away>dap: thanks :)
01:10:36  <isaacs_away>dap: this is for master, yes?
01:11:24  <mjr_>isaacs_away: bnoordhuis left, but I opened this issue: https://github.com/joyent/node/issues/2683
01:11:28  <dap>isaacs_away: Yes. In fact, it has no impact on the node build, so there would be no point in backporting it. It's a file people have to install into /usr/lib/dtrace in order to use DTrace with node. The new file works with all previous node versions.
01:11:32  <mjr_>Not sure if you guys assign issues or whatever.
01:13:08  <isaacs_away>mjr_: we do.
01:13:10  <isaacs_away>mjr_: thanks
01:13:46  <mjr_>Also, issue #2683! Not exactly a small project at this point.
01:13:55  <isaacs_away>inorite!?
01:15:24  <isaacs_away>[node@ryannode ~]$ cd web/nodejs.org/dist/[TAB]
01:15:24  <isaacs_away>Display all 118 possibilities? (y or n)
01:15:26  <isaacs_away>mjr_: ^
01:15:29  <isaacs_away>118!!!
01:15:35  <mjr_>jezus
01:15:50  <mjr_>That's a large number.
01:17:27  <isaacs_away>srsy
01:19:56  * orlandovftwquit (Ping timeout: 240 seconds)
01:20:25  <CIA-114>node: isaacs v0.6 * r465f57c / (3 files in 2 dirs): Upgrade v8 to 3.6.6.20 - http://git.io/6ET1kg
01:20:25  <CIA-114>node: isaacs v0.6 * r71527e3 / (deps/v8/SConstruct deps/v8/tools/gyp/v8.gyp): Patches floating on v8 - http://git.io/subsvw
01:20:26  <CIA-114>node: isaacs v0.6 * r5937d02 / tools/msvs/msi/product.wxs : msi: npm bash shim fixes - http://git.io/9zseVA
01:20:28  <CIA-114>node: isaacs v0.6 * r051908e / (7 files in 6 dirs): (log message trimmed)
01:20:28  <CIA-114>node: 2012.02.02, Version 0.6.10 (stable)
01:20:28  <CIA-114>node: * Update V8 to 3.6.6.20
01:20:28  <CIA-114>node: * Add npm msysgit bash shim to msi installer (isaacs)
01:20:28  <CIA-114>node: * buffers: fix intermittent out of bounds error (Ben Noordhuis)
01:20:29  <CIA-114>node: * buffers: honor length argument in base64 decoder (Ben Noordhuis)
01:20:29  <CIA-114>node: * windows: Fix path.exists regression (Bert Belder)
01:20:29  <CIA-114>node: isaacs v0.6 * r74a2528 / src/node_version.h : Now working on v0.6.11 - http://git.io/c4K_Sw
01:25:03  * sj26quit (Excess Flood)
01:26:11  * sj26joined
01:26:15  <piscisaureus_>isaacs_away++
01:26:16  <kohai>isaacs_away has 1 beer
01:26:23  * isaacs_awaychanged nick to isaacs
01:26:27  <isaacs>piscisaureus_: thanks
01:27:56  * AvianFlujoined
01:28:26  * travis-cijoined
01:28:26  <travis-ci>[travis-ci] joyent/node#361 (v0.6 - 74a2528 : isaacs): The build was broken.
01:28:26  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/30fd2f7...74a2528
01:28:26  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/618256
01:28:26  * travis-cipart
01:29:36  * isaacsneeds an irl beer
01:31:01  <tjfontaine>sorry, but ctcp won't solve that
01:31:18  <tjfontaine>oakland, hmm perhaps I can order one online and have it delivered?
01:31:44  <piscisaureus_>I'm going. bye
01:32:02  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:04:46  <TooTallNate>isaacs: https://github.com/isaacs/node-which/pull/1
02:04:57  <TooTallNate>im retarded, the first one was a pull request to my own fork
02:04:59  <TooTallNate>Doh!
02:05:48  <isaacs>TooTallNate: np
02:06:04  <isaacs>TooTallNate: does it respect the PATHEXT env if you're not using the sync version?
02:07:16  <TooTallNate>isaacs: ya you already had that in there
02:07:18  <isaacs>TooTallNate: d'oh, also a path.charAt(0) === "/" in there. that should be a more fancy-pants isAbsolute function.
02:07:21  <isaacs>oh, nice.
02:07:26  <isaacs>thanks.
02:07:33  <TooTallNate>there's isAbsolute()
02:07:39  <TooTallNate>hah, never heard of it
02:07:45  <isaacs>i don't know if it's exposed.
02:07:49  <isaacs>should probably be.
02:09:08  <TooTallNate>isaacs: i don't see it, only a could isAbsolute variables that get set via the same str.charAt(0) === '/' code
02:09:19  <TooTallNate>s/could/couple
02:09:34  <isaacs>TooTallNate: check out node's lib/path.js
02:09:36  <isaacs>it's in there somewhere
02:09:50  <isaacs>we should definitely expose that in master, like asap
02:10:40  <TooTallNate>haha, i'd send a quick patch but i cant find it
02:12:57  <isaacs>TooTallNate: oh, i think there's some kind of bigass regexp that does that.
02:13:03  <isaacs>i forget now where i used that last...
02:13:05  <isaacs>glob, maybe?
02:13:10  <isaacs>somehwere in npm?
02:13:24  <isaacs>time to go :) g'nite
02:13:34  <TooTallNate>isaacs: nite. have a good one
02:14:08  * isaacsquit (Remote host closed the connection)
02:56:37  * pieternquit (Quit: pietern)
03:09:18  * brsonquit (Quit: leaving)
03:25:25  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
03:37:30  * dapquit (Quit: Leaving.)
03:40:12  * dapjoined
03:41:39  * dapquit (Client Quit)
03:43:11  * dshaw_quit (Ping timeout: 255 seconds)
05:34:29  * dshaw_joined
05:38:10  * dshaw_quit (Client Quit)
05:42:55  * dshaw_joined
05:48:16  * orlandovftwjoined
06:30:18  * paddybyersjoined
07:04:55  * mralephjoined
07:42:20  * mikealjoined
07:44:50  * mikealquit (Client Quit)
07:51:42  * mralephquit (Quit: Leaving.)
07:58:11  * mralephjoined
07:58:39  * igorziquit (Ping timeout: 245 seconds)
08:31:36  * mralephquit (Quit: Leaving.)
08:48:49  * mikealjoined
09:58:35  * dshaw_quit (Quit: Leaving.)
10:01:21  * orlandovftwquit (Ping timeout: 272 seconds)
10:57:33  * bnoordhuisjoined
12:01:16  * mralephjoined
13:18:37  * piscisaureus_joined
14:09:13  <CIA-114>node: Dave Pacheco v0.6 * r9fb088e / src/node.d :
14:09:13  <CIA-114>node: dtrace: add missing translator
14:09:13  <CIA-114>node: Add missing translator for node_dtrace_http_*_request_t types.
14:09:13  <CIA-114>node: Fixes #2667. - http://git.io/0_67cw
14:18:46  * travis-cijoined
14:18:46  <travis-ci>[travis-ci] joyent/node#362 (v0.6 - 9fb088e : Dave Pacheco): The build was fixed.
14:18:46  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/74a2528...9fb088e
14:18:46  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/620607
14:18:46  * travis-cipart
14:22:56  * AndreasMadsenjoined
14:36:21  <piscisaureus_>bnoordhuis: when did we rename all occurences of process to process_1 in node.cc?
14:36:27  <piscisaureus_>(also in 0.6)
14:36:51  <bnoordhuis>iirc when we started getting serious about isolates
14:38:39  <CIA-114>node: Bert Belder v0.6 * r68db206 / src/node.cc : Rename some occurrences of `process_1` back to `process` - http://git.io/1kZKrA
14:46:41  * travis-cijoined
14:46:41  <travis-ci>[travis-ci] joyent/node#363 (v0.6 - 68db206 : Bert Belder): The build passed.
14:46:41  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/9fb088e...68db206
14:46:41  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/620740
14:46:41  * travis-cipart
14:52:55  <AndreasMadsen>I want global.errno to be a number and not a string. I do https://github.com/AndreasMadsen/node-1/compare/base...errno-abstract#L6R983 but not all errnos are right - why?
15:09:07  <piscisaureus_>AndreasMadsen: errnos are from libuv
15:09:17  <piscisaureus_>AndreasMadsen: we don't want to change this btw
15:11:03  <AndreasMadsen>piscisaureus_: You should it is so frustrating too get only an errno code and not a readable message
15:18:36  <AndreasMadsen>any news on the "isolate double exit" bug, it is even more frustrating than errnos -> https://github.com/joyent/node/issues/2665
15:21:24  <CIA-114>libuv: Ben Noordhuis v0.6 * r2f182f4 / (src/unix/error.c src/unix/stream.c): unix: map ECONNABORTED - http://git.io/gS4rmA
15:22:28  <bnoordhuis>AndreasMadsen: isolates are out
15:22:41  <AndreasMadsen>bnoordhuis: how so?
15:22:49  <AndreasMadsen>bnoordhuis: we won't use it
15:22:54  <AndreasMadsen>?
15:23:07  <bnoordhuis>AndreasMadsen: yes, and remove from the code base
15:23:11  * travis-cijoined
15:23:11  <travis-ci>[travis-ci] joyent/libuv#76 (v0.6 - 2f182f4 : Ben Noordhuis): The build is still failing.
15:23:11  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/267e75d...2f182f4
15:23:11  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/620985
15:23:11  * travis-cipart
15:23:29  <AndreasMadsen>bnoordhuis: forever - or just for the moment?
15:23:46  <bnoordhuis>AndreasMadsen: forever and a day
15:24:05  <bnoordhuis>it was a nice experiment but it didn't work out
15:24:10  <AndreasMadsen>bnoordhuis: it solved a lot of cluster issues :(
15:24:25  <bnoordhuis>it introduced a lot more
15:24:59  <AndreasMadsen>bnoordhuis: I noticed that - thought it was only temporally bugs
15:26:23  <AndreasMadsen>need to see if https://github.com/joyent/node/pull/2544 still is relevant
15:26:56  <bnoordhuis>AndreasMadsen: isolates support cluttered the code base for only marginal returns
15:26:58  <AndreasMadsen>bnoordhuis: not really removed -> https://github.com/joyent/node/blob/master/lib/child_process.js#L543
15:27:13  <bnoordhuis>threads are not even that much faster than processes on linux
15:27:51  <bnoordhuis>in-memory ipc is nice but we can use shared memory for that
15:28:05  <bnoordhuis>and nobody ever complained about ipc being so slow anyway
15:28:59  <AndreasMadsen>bnoordhuis: it was not about speed, but when the process (master) died threads (workers) would die too.
15:29:06  <mmalecki>can someone take a look at https://github.com/joyent/libuv/pull/298?
15:29:42  <indutny>wweeehoo
15:29:51  <indutny>:D
15:30:01  <AndreasMadsen>bnoordhuis: but please remove it from child_process.js too. Confuses me -> https://github.com/joyent/node/blob/master/lib/child_process.js#L543
15:30:06  <AndreasMadsen>See you later
15:30:07  * AndreasMadsenquit (Remote host closed the connection)
15:30:43  <indutny>bnoordhuis: have you wiped out isolates?
15:30:57  <bnoordhuis>indutny: not yet but i'm working on it
15:31:11  <indutny>bnoordhuis: k
15:31:15  <indutny>just curious
15:41:19  <bnoordhuis>indutny: was 44e7033 an isolates-only bug fix?
15:44:38  <indutny>bnoordhuis: huh?
15:44:39  <indutny>one sec
15:44:54  <indutny>not only
15:45:01  <indutny>actually no
15:45:04  <indutny>bnoordhuis: ^
15:45:08  <bnoordhuis>okay, i'll replay that one later on
15:45:10  <indutny>bnoordhuis: that's very strange thing
15:45:25  * bnoordhuisis reverting the isolates commits in reverse order
15:45:33  <indutny>bnoordhuis: heh
15:47:36  <bnoordhuis>there are people ice skating in front of my house...
15:47:41  <bnoordhuis>better them than me
15:48:14  <bnoordhuis>it hasn't been freezing that long, i kind of doubt it's safe
15:50:07  * bnoordhuislikes how there is nearly no merge conflict you can't solve with -C0
15:51:19  <mmalecki>-C0?
15:52:52  <indutny>C3P0
15:53:07  <indutny>mmalecki: a shortcut
15:58:36  <mmalecki>indutny: ++, lol
15:58:36  <kohai>indutny has 9 beers
16:18:58  * AndreasMadsenjoined
16:22:37  <AndreasMadsen>who will review this https://github.com/joyent/node/pull/2661 - would like to make the cluster disconnect patch this weekend.
16:26:38  * felixgejoined
16:26:38  * felixgequit (Changing host)
16:26:38  * felixgejoined
16:39:56  * perezdquit (Ping timeout: 240 seconds)
16:41:43  * pieternjoined
16:45:07  * perezdjoined
16:46:03  * perezd_joined
16:49:27  * perezdquit (Ping timeout: 240 seconds)
16:49:27  * perezd_changed nick to perezd
16:50:11  <bnoordhuis>mmalecki: -C0 == reduce patch context to 0 lines
16:51:31  <bnoordhuis>mmalecki: ping me tonight about that ELOOP patch and i'll merge it
16:51:49  <mmalecki>bnoordhuis: define tonight
16:52:11  <bnoordhuis>mmalecki: 18.00-00.00 CEST
16:52:25  <bnoordhuis>make that 20.00-00.00
16:53:09  <mmalecki>bnoordhuis: ok, making `at` play some music then
16:54:51  * perezd_joined
16:56:39  * perezdquit (Ping timeout: 248 seconds)
16:56:40  * perezd_changed nick to perezd
16:58:48  * perezd_joined
16:59:17  * perezd_quit (Client Quit)
17:00:51  * isaacsjoined
17:01:32  * perezdquit (Read error: Connection reset by peer)
17:18:45  * stephankjoined
17:20:24  * orlandovftwjoined
17:20:24  * orlandovftwquit (Client Quit)
17:20:32  * piscisaureus_quit (Ping timeout: 244 seconds)
17:20:40  * orlandovftwjoined
17:30:09  * perezdjoined
17:42:28  * sh1mmerquit (Quit: sh1mmer)
17:43:53  * felixgequit (Quit: felixge)
17:45:09  <isaacs>good morning
17:45:16  <indutny>morning
17:45:21  * mikealquit (Quit: Leaving.)
17:45:22  <mmalecki>morning!
17:51:15  * igorzijoined
18:01:57  * perezdquit (Read error: Connection reset by peer)
18:03:37  * perezdjoined
18:03:44  * perezdquit (Client Quit)
18:05:02  * TooTallNatejoined
18:06:15  * mikealjoined
18:13:56  <bnoordhuis>https://github.com/bnoordhuis/node/commit/fa42dc8 <- review? reverts isolates for great good
18:16:34  <bnoordhuis>indutny: what could make simple/test-debugger-client fail with AssertionError: "3.8.9" == "3.6.6.15"?
18:18:15  <mmalecki>v8 version assert?
18:18:21  <mmalecki>*somewhere*?
18:18:37  <indutny>bnoordhuis: huh?
18:18:40  <mmalecki>just a wild guess, tho
18:18:58  <indutny>bnoordhuis: that's odd, what node version?
18:19:10  * perezdjoined
18:19:38  <indutny>bnoordhuis: ah I know
18:19:43  <bnoordhuis>indutny: see above, that's after reverting isolates
18:20:01  <indutny>bnoordhuis: process.versions.v8 is newer than one reported by v8 debugger server
18:20:17  <indutny>bnoordhuis: left is process.versions.v8
18:20:21  <indutny>bnoordhuis: right is actual v8
18:20:27  <bnoordhuis>ah, i have stray node processes...
18:20:33  <indutny>yeah
18:20:44  <indutny>bnoordhuis: I really love debugger in isolate ;)
18:20:50  <indutny>sad that it happens
18:21:07  <bnoordhuis>killed the node processes, works now, thanks
18:21:10  <indutny>offtopic, amazon is trying to bill 1$ from me for about 1 year
18:21:17  <indutny>sending email each two weeks or so
18:21:27  <indutny>kinda tired of it
18:22:05  <bnoordhuis>why is amazon billing you?
18:23:33  <indutny>bnoordhuis: can't remember exactly, but I think I turned on one virtual server in 2010
18:23:48  <indutny>but I really don't know
18:24:08  <bnoordhuis>oh well, you're in russia - they'll never get you
18:24:33  <indutny>bnoordhuis: haha
18:24:43  <indutny>bnoordhuis: even if they get me, I'll give them 1 $
18:24:46  <indutny>in cash
18:44:55  <mmalecki>benefit of living in Russia ;)
18:50:16  * brsonjoined
18:52:25  <isaacs>bnoordhuis: i've noticed that sometimes when the version you're testing is not the version in the PATH
18:52:38  <isaacs>or that
18:53:09  <isaacs>stray node process, yeah
19:00:25  <mmalecki>bnoordhuis: hey! there's some guy in my computer telling me to ping you about eloop.
19:15:11  <AndreasMadsen>bnoordhuis: https://github.com/bnoordhuis/node/commit/fa42dc8#L2R89 should not be reverted
19:27:44  * mikealquit (Quit: Leaving.)
19:31:33  * dshaw_joined
19:46:56  * dapjoined
20:01:18  * AndreasMadsenquit (Remote host closed the connection)
20:09:08  * mikealjoined
20:24:13  <isaacs>bnoordhuis: hey, so, on that fs.rename argument error thing... i'm wondering, what if we just add the "arguments" member to every error that gets generated from the fs methods?
20:24:52  * einarosquit (Quit: Lost terminal)
20:25:13  * pieternquit (Quit: pietern)
20:27:20  * `3rdEdenjoined
20:40:52  <bnoordhuis>isaacs: what arguments member do you mean?
20:41:04  <bnoordhuis>mmalecki: yes, reviewing
20:41:34  <isaacs>bnoordhuis: internal functions in javascript, when they throw, often tell you the arguments to the function on the error object
20:41:54  <isaacs>we could just put that on all fs.* methods when they pass an error to the cb
20:42:07  <bnoordhuis>isaacs: oh, like that. how expensive is that?
20:42:29  <isaacs>shouldn't be very expensive. it'd be better to make it an array of the strings, though, rather than the actual arguments object.
20:42:36  <isaacs>if you sneeze on the arguments object, v8 hits you.
20:43:06  <bnoordhuis>who or what carries the arguments around when the thread pool is doing its thing?
20:46:09  * mikealquit (Quit: Leaving.)
20:47:51  <isaacs>bnoordhuis: i guess we could just do it at the JS level, for async stuff anyway
20:48:17  <isaacs>bnoordhuis: but i think the cheapest way (though some manual labor) is to just tack on an array when we generate the error object in src/node_file.cc
20:48:26  <bnoordhuis>isaacs: the bindings don't return anything that js land can attach to though
20:48:37  * mikealjoined
20:48:43  <bnoordhuis>nearly everything in node_file.cc does return Undefined()
20:48:50  <isaacs>right
20:49:36  <isaacs>but you could have something like: fs.open = function(path, type, cb) { binding.open(path, type, function(er, fd) { if (er) ...; cb(er, fd) }) }
20:49:51  <isaacs>... => er.arguments = [path, type]
20:50:52  <bnoordhuis>yes. but it's probably a good idea to test the performance and memory overhead of that
20:51:15  <bnoordhuis>still, it's well worth a try
20:51:31  <isaacs>oyeah
20:51:42  <isaacs>or, it could potentially be done in the ASYNC_CALL macro
20:56:46  <bnoordhuis>yeah, that was my initial suggestion to diego
20:56:59  <bnoordhuis>it kind of saddens me that libeio also makes a copy of the new path
20:57:09  <bnoordhuis>but i guess i should just learn to live with that
20:57:42  <isaacs>more debug tracking is almost always good, but it should not require uv changes, i don't think, and *definitely* not require eio changes.
21:00:35  <isaacs>ok, i'm going to fork v0.7.3-release
21:07:05  <isaacs>bnoordhuis: actually, you know how we talked about freezing the release 1-2 days ahead of time?
21:07:13  <bnoordhuis>yep
21:07:29  <isaacs>$ glg -n5
21:07:29  <isaacs>* a9723df Ben Noordhuis (HEAD, ry/master, ry-push/master, master) Revert "Process symlinked shared library as .node" (2 days ago)
21:07:29  <isaacs>* ec79acb isaacs (v0.7.2-release) working on 0.7.3 now (2 days ago)
21:07:30  <isaacs>* a3efcd2 isaacs (v0.7.2) 2012.02.01, Version 0.7.2 (unstable) (2 days ago)
21:07:40  <isaacs>do we really want to do a release for that one commit? :\
21:07:44  <bnoordhuis>hah no
21:08:06  <bnoordhuis>let's wait for when isolations are backed out
21:08:09  <isaacs>it'd be an easy changelog to write :)
21:08:40  <isaacs>also: i wonder if it'd be better to do both releases on the same day
21:08:45  <isaacs>or if that'd be more confusing.
21:08:55  <isaacs>it'd certainly be nice to only destroy one day of my week, instead of 2
21:09:00  * pieternjoined
21:09:13  <bnoordhuis>yeah... it's probably more confusing but not insurmountable
21:09:15  <mmalecki>bnoordhuis: ELOOP?
21:10:15  <bnoordhuis>maybe add *DANGER DANGER DANGER* to the 0.7.x email
21:10:41  <bnoordhuis>mmalecki: looks good. did bert sign off the windows change?
21:11:13  <bnoordhuis>ERROR_CANT_RESOLVE_FILENAME seems like the proper error code to me but you never know with windows
21:11:32  <mmalecki>bnoordhuis: he told me to use it
21:11:58  <bnoordhuis>then i suppose i have no choice but to land it, eh?
21:12:02  <mmalecki>but, like, please run this test on windows, I didn't even test it
21:12:27  <mmalecki>bnoordhuis: I guess so?
21:12:54  <bnoordhuis>hmm, your patch doesn't apply to v0.6
21:13:14  <bnoordhuis>also, i really don't want to have to start up my mac first, then the windows 7 vm, compile libuv, etc...
21:13:15  <mmalecki>bnoordhuis: sec, I'll rebase
21:14:13  <mmalecki>bnoordhuis: then let's assume that Bert was right
21:14:41  <tjfontaine>test in production like all the rest of us
21:15:54  <mmalecki>lol, production and windows
21:16:05  <tjfontaine>heh
21:16:27  <bnoordhuis>don't laugh - my bank's ATMs run on windows
21:16:40  <bnoordhuis>i've seen the BSOD more than once on an ATM's terminal...
21:17:10  <mmalecki>that's just sad
21:17:13  <tjfontaine>I try not to think about such things, just like when I'm riding in roller coasters I try and not think humans maintain it
21:18:02  <mmalecki>bnoordhuis: rebased
21:18:47  <bnoordhuis>mmalecki: still doesn't apply
21:20:17  <mmalecki>bnoordhuis: oh, sorry, fail, pushed wrong branch
21:20:22  <mmalecki>bnoordhuis: try now
21:22:31  <CIA-114>libuv: Maciej Małecki v0.6 * re2565c3 / src/win/error.c : win: map `ERROR_CANT_RESOLVE_FILENAME` to `UV_ENAMETOOLONG` - http://git.io/_bHlfA
21:22:31  <CIA-114>libuv: Maciej Małecki v0.6 * r3ff3626 / (include/uv.h src/unix/error.c): unix: map `ELOOP` to `UV_ELOOP` - http://git.io/e1yO1g
21:22:32  <CIA-114>libuv: Maciej Małecki v0.6 * r0596c59 / (test/test-fs.c test/test-list.h): test: test if `UV_ELOOP` mapping works - http://git.io/cgTvjg
21:22:48  <mmalecki>bnoordhuis: thanks!
21:24:12  * travis-cijoined
21:24:12  <travis-ci>[travis-ci] joyent/libuv#77 (v0.6 - e2565c3 : Maciej Małecki): The build is still failing.
21:24:12  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/2f182f4...e2565c3
21:24:12  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/622946
21:24:12  * travis-cipart
21:25:45  * elijah-mbpquit (Read error: Connection reset by peer)
21:26:16  * elijah-mbpjoined
21:28:26  * paddybyersquit (Quit: paddybyers)
21:47:43  * paddybyersjoined
22:08:10  * `3rdEdenquit (Quit: Leaving...)
22:28:04  * felixgejoined
22:28:04  * felixgequit (Changing host)
22:28:04  * felixgejoined
22:56:23  <isaacs>Inline JavaDoc style comments: pros/cons?
22:56:45  <isaacs>bnoordhuis, piscisaureus, igorzi, indutny, anyone else ^
22:57:53  <bnoordhuis>isaacs: no strong opinion
22:58:10  <bnoordhuis>the main drawback is that people always forget to update the comment
22:58:32  <isaacs>true.. but they also forget to update the separate markdown file, and we manage to keep that pretty much in check.
22:58:41  <isaacs>imo, the main drawback is that those comments are annoying.
22:58:44  <isaacs>and i hate them.
22:58:59  <mmalecki>I kinda like docco style
22:59:28  <mmalecki>looks like https://github.com/nodejitsu/haibu/blob/master/lib/haibu/core/spawner.js#L24-30
22:59:58  <bnoordhuis>not my cup of tea
23:00:18  <mmalecki>I even wrote docstrings implementation in js
23:00:20  <bnoordhuis>i prefer earl grey
23:00:32  <bnoordhuis>speaking of...
23:00:41  * bnoordhuisgo gets a cup
23:00:44  <mmalecki>I'd release it, but I have no idea how to call it
23:01:06  <bnoordhuis>truth be told, i don't like comments
23:01:19  <isaacs>standalone markdown files are so nice.
23:01:25  <bnoordhuis>code should be so simple that you don't need them
23:01:39  <isaacs>bnoordhuis: yes, that's it, that's what bugs me. thank you.
23:01:52  <isaacs>javadoc comments are there for the benefit of people *not* looking at the code.
23:02:02  <isaacs>comments should be for the benefit of the person reading the code.
23:02:08  <isaacs>not pulled out as documentation.
23:02:13  <bnoordhuis>exactly
23:04:20  <isaacs>however, for API reference, there are some definite advantages.
23:04:28  <isaacs>despite the fact that they're ugly adn annoying and i hate them.
23:05:11  <isaacs>i really like the way that v8's doxygen docs can all link you directly to where in the source the function lives.
23:08:16  <igorzi>isaacs: no strong opinion.. i guess the advantage that wasn't mentioned is that this would probably minimize cases when docs go out of sync with the API
23:08:49  <bnoordhuis>mmalecki: how many frozen homeless people have you found on your doorstep so far?
23:09:06  * piscisaureus_joined
23:09:40  <piscisaureus_>isaacs: I am not strongly agains jsdoc in general. But I don't think we can put *all* documentation in there.
23:09:43  <mmalecki>bnoordhuis: dunno, I try to get rid of them as soon as I see them
23:09:49  <piscisaureus_>isaacs: I mean, we need examples etc
23:09:52  <mmalecki>bnoordhuis: 7 or 8 maybe?
23:10:02  <piscisaureus_>isaacs: you definitely don't want tot stuff those in comments
23:10:07  <mmalecki>bnoordhuis: and you?
23:10:21  <bnoordhuis>mmalecki: none so far
23:10:22  <isaacs>piscisaureus_: yeah, the yui project did that. giant massive markdown/html monsters in the comments.
23:10:28  <piscisaureus_>isaacs: the v8 way is nice but the v8 docs are fairly minimal compared to what we should have.
23:10:48  <piscisaureus_>I still very much prefer to have documentation in markdown
23:10:57  <bnoordhuis>but i'll probably dredge up some ice skaters from the canal come spring
23:11:33  <bnoordhuis>the ice was pretty dangerous today, lots of holes
23:12:01  <mmalecki>bnoordhuis: good plan. winter without finding frozen people is not a winter in my book
23:12:16  <bnoordhuis>wholeheartedly agreed
23:13:49  <isaacs>you guys are crazy.
23:13:56  <isaacs>it's sunny and beautiful here today
23:14:06  * mikealquit (Quit: Leaving.)
23:14:23  <isaacs>bit chilly, actually. 15˚C
23:14:29  <isaacs>need long sleeves
23:15:26  <piscisaureus_>isaacs: is jsdoc stuff really an option?
23:15:39  <piscisaureus_>isaacs: iow, did garen / msft suggest it?
23:15:39  <isaacs>piscisaureus_: i'm exploring ideas.
23:16:01  <bnoordhuis>isaacs: it's -15 here, i gather it's around -30 where mmalecki is
23:16:03  <isaacs>piscisaureus_: just trying to find a way to get all this info into our docs, in a way that isn't painful to maintain
23:16:11  <mmalecki>lol, long sleeves at 15 C.
23:16:22  <piscisaureus_>pussy :-p
23:16:36  <isaacs>my apt gets a lot of sun, though, so the windows are open.
23:16:40  <mmalecki>bnoordhuis: -15 as well
23:16:40  <isaacs>it's pretty warm in here.
23:16:51  <piscisaureus_>ok. I'm not supposed to be here, so goodbye again :-)
23:17:00  <igorzi>piscisaureus_: wait!!
23:17:07  <piscisaureus_>igorzi: ok, what's up
23:17:08  <bnoordhuis>mmalecki: oh, the -30 must've been farther east
23:17:37  <igorzi>piscisaureus_: we still didn't find the root cause of that socket hang issue, but there's probably a very high chance that it's this: http://support.microsoft.com/kb/2568167
23:17:48  <mmalecki>I'm pretty sure it's -30 where indutny lives
23:18:02  <igorzi>FILE_SKIP_COMPLETION_PORT_ON_SUCCESS for non-IFS LSPs
23:18:23  <piscisaureus_>igorzi: seems unlikely to me
23:18:31  <igorzi>piscisaureus_: why?
23:18:36  <bnoordhuis>isaacs: when are you releasing 0.6.10?
23:18:47  <isaacs>bnoordhuis: last night
23:18:58  <piscisaureus_>igorzi: but I'm not surprised, since SetFileCompletionNotificationModes doesn't even work correctly *with* it :-)
23:18:58  <bnoordhuis>oh... did you forget to tag it?
23:18:59  <isaacs>http://nodejs.org/dist/latest/
23:19:08  * `3rdEdenjoined
23:19:22  <piscisaureus_>igorzi: but, I asked many people for their 'netsh winsock show catalog' output
23:19:35  <isaacs>no, it's tagged
23:19:36  <isaacs>https://github.com/joyent/node/tree/v0.6.10
23:19:58  <piscisaureus_>igorzi: and noone seemed to have any funky LSPs installed
23:19:59  <bnoordhuis>sorry, my bad - my shell's git autocomplete was acting up
23:20:15  <igorzi>piscisaureus_: we should at the very least do what the docs suggest (which is to not use FILE_SKIP_COMPLETION_PORT_ON_SUCCESS with non-IFS LSPs), which should be easy to do
23:20:21  <piscisaureus_>igorzi: except for one guy who had some vmware stuff in there but I checked and it should not do anything for tcp
23:20:30  <piscisaureus_>igorzi: I agree sure
23:21:01  <piscisaureus_>igorzi: we could just the same we do for UDP (is check protocol chain length) for tcp too.
23:22:06  <igorzi>piscisaureus_: well, FILE_SKIP_COMPLETION_PORT_ON_SUCCESS should be ok with LSPs in general, we just need to detect non-IFS LSPs
23:22:18  <igorzi>XP1_IFS_HANDLES bit
23:22:28  <piscisaureus_>igorzi: right.
23:22:33  <piscisaureus_>igorzi: yes that seems fine.
23:23:06  * mikealjoined
23:24:04  <piscisaureus_>igorzi: https://github.com/isaacs/npm/issues/1894#issuecomment-3679754 <-- for some 'show catalog' output
23:24:09  <piscisaureus_>all seems pretty normal to me
23:25:52  <igorzi>piscisaureus_: %SystemRoot%\system32\vsocklib.dll.. this is vmware one?
23:26:08  * xycjoined
23:26:09  * elijah-mbpquit (Remote host closed the connection)
23:26:16  <igorzi>piscisaureus_: Marcas de servicio: 0x9
23:26:41  <xyc>i wonder what kind of research ryah will pursue
23:26:59  * elijah-mbpjoined
23:27:12  <igorzi>piscisaureus_: it doesn't have XP1_IFS_HANDLES bit set
23:27:47  <piscisaureus_>igorzi: vmci, yes that's the only unusual one I'v eseen
23:29:21  * felixgequit (Quit: felixge)
23:32:14  <piscisaureus_>igorzi: I am more afraid it could be that one other application is squatting the port
23:32:34  <piscisaureus_>igorzi: hmm that wouldn't explain why npm doesn't work
23:33:48  <piscisaureus_>ok now I really have to go. but as you might have noticed, I check the irc channel from time to time :-)
23:36:49  <igorzi>piscisaureus_: ttyl
23:36:58  * mikealquit (Quit: Leaving.)
23:37:26  <piscisaureus_>igorzi: ttyl. I really hope we find someone that can reproduce the issue once.
23:37:58  <igorzi>piscisaureus_: yep
23:42:39  <CIA-114>node: Stefan Rusu v0.6 * r4671e54 / lib/net.js :
23:42:39  <CIA-114>node: net: destroy socket on DNS error
23:42:39  <CIA-114>node: The socket was never destroyed on DNS errors. This broke some clients, including
23:42:39  <CIA-114>node: lib/https.js. - http://git.io/tb3d2Q
23:42:39  <CIA-114>node: Stefan Rusu v0.6 * r07a983a / (2 files):
23:42:40  <CIA-114>node: test: add tcp and https DNS error tests
23:42:40  <CIA-114>node: net-dns-error: specifc test for the net DNS issue.
23:42:40  <CIA-114>node: http-dns-error: now it works for HTTPS as well. - http://git.io/Gh2z5w
23:44:06  * piscisaureus_quit (Ping timeout: 240 seconds)
23:50:39  * travis-cijoined
23:50:39  <travis-ci>[travis-ci] joyent/node#364 (v0.6 - 07a983a : Stefan Rusu): The build was broken.
23:50:39  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/68db206...07a983a
23:50:39  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/623662
23:50:39  * travis-cipart
23:54:54  * xycquit (Quit: leaving)
23:57:46  * `3rdEdenquit (Quit: Zzzzz)
23:58:39  * paddybyersquit (Quit: paddybyers)
23:58:53  <isaacs>oh, also, inline javadoc comments are stupid for events.
23:59:00  <isaacs>unless you comment on your .emit() calls