00:03:46  * piscisaureus_quit (Read error: Connection reset by peer)
00:08:07  * piscisaureus_joined
00:10:28  * piscisaureus_quit (Read error: Connection reset by peer)
00:14:48  * piscisaureus_joined
00:21:13  * mikealquit (Quit: Leaving.)
00:25:41  * c4milojoined
00:26:19  * ericktquit (Ping timeout: 244 seconds)
00:26:58  <isaacs>cc: error: ../deps/uv/src/unix/cares.c: No such file or directory
00:26:58  <isaacs>cc: fatal error: no input files
00:26:58  <isaacs>compilation terminated.
00:27:00  <isaacs>make[1]: *** [/root/node/out/Release/obj.target/uv/deps/uv/src/unix/cares.o] Error 1
00:27:02  <isaacs>make[1]: *** Waiting for unfinished jobs....
00:27:03  <isaacs>cc: error: ../deps/uv/src/unix/check.c: No such file or directory
00:27:06  <isaacs>cc: fatal error: no input files
00:27:08  <isaacs>compilation terminated.
00:27:38  * perezdquit (Quit: perezd)
00:28:41  * bnoordhuisquit (Ping timeout: 252 seconds)
00:35:53  <piscisaureus_>yes
00:36:12  <piscisaureus_>igorzi__: the last time you upgraded libuv in node you forgot to add src/cares.c
00:36:39  <piscisaureus_>(which is a new file in libuv)
00:38:15  <piscisaureus_>hmm
00:38:24  <piscisaureus_>it seems that ben already fixe it
00:38:57  <piscisaureus_>isaacs: I think make clean; ./configure will make your problems go away
00:39:44  * ryahjoined
00:40:43  <isaacs>ok...
00:45:22  <isaacs>piscisaureus_: hm, still fails, but differently
00:45:24  <isaacs>[[email protected] ~/node]# make
00:45:24  <isaacs>make -C out BUILDTYPE=Release
00:45:25  <isaacs>make[1]: Entering directory `/root/node/out'
00:45:26  <isaacs> CC(target) /root/node/out/Release/obj.target/uv/deps/uv/src/unix/core.o
00:45:28  <isaacs>../deps/uv/src/unix/core.c: In function 'uv_getaddrinfo_done':
00:45:30  <isaacs>../deps/uv/src/unix/core.c:321:27: error: 'handle' undeclared (first use in this function)
00:45:32  <isaacs>../deps/uv/src/unix/core.c:321:27: note: each undeclared identifier is reported only once for each function it appears in
00:45:34  <isaacs>make[1]: *** [/root/node/out/Release/obj.target/uv/deps/uv/src/unix/core.o] Error 1
00:45:36  <isaacs>make[1]: Leaving directory `/root/node/out'
00:45:38  <isaacs>make: *** [node] Error 2
00:45:40  <isaacs>[[email protected] ~/node]#
00:45:52  <piscisaureus_>hmm well it seems libuv is busted, then
00:46:10  <isaacs>seems that way
00:46:56  * avalanche123joined
00:47:11  <piscisaureus_>isaacs: do you mind if I not fix it right away?
00:47:14  <piscisaureus_>I was about to go sleep
00:47:53  <isaacs>piscisaureus_: yeah, that's fine
00:47:55  <isaacs>i'm sick anywa
00:48:06  <piscisaureus_>oh hmm that's good
00:48:13  <piscisaureus_>erm, not good
00:48:14  <isaacs>good news is that i figured out how to deploy the node source using fleet, but it seems like fleet is designed more for websites than CI
00:48:17  <piscisaureus_>is it serious
00:48:19  <isaacs>nah
00:48:32  <isaacs>just a cold from going to a conf and staying up all night and then not-sleeping on a plane on the way back
00:48:47  <isaacs>i'm almost done with it
00:48:51  <piscisaureus_>ah
00:49:09  <piscisaureus_>is fleet like buildbot?
00:49:18  <isaacs>piscisaureus_: no, it's more like a deploy tool
00:49:24  * ryahquit (Quit: leaving)
00:49:27  <isaacs>piscisaureus_: but it's like buildbot in the sense that you send code to it and it does stuff
00:49:41  <piscisaureus_>ah right
00:49:41  <isaacs>it's all git-based and runs with node and dnode
00:50:09  <isaacs>ok, i'm out. get some sleep :)
00:50:12  <piscisaureus_>so we're supposed to send it "git pull; git clean -fdx; ./configure; make test"
00:50:14  <piscisaureus_>yeah
00:50:15  <piscisaureus_>later
00:50:20  <isaacs>piscisaureus_: nono
00:50:26  <isaacs>you just run `fleet deploy` from in the git repo
00:50:36  <isaacs>and since it's node, it'll run in windows, too
00:50:45  <piscisaureus_>hehe
00:50:49  <piscisaureus_>that's the theory
00:50:52  <isaacs>right
00:50:55  <piscisaureus_>nobody really knows :-p
00:51:07  <piscisaureus_>getting some rest, tty tomorrow
00:51:18  <isaacs>g'nite
00:51:23  * isaacsquit (Remote host closed the connection)
00:55:46  * piscisaureus_quit (Ping timeout: 256 seconds)
01:01:15  * ericktjoined
01:04:08  * benviejoined
01:06:52  * mrb_bkquit (Ping timeout: 246 seconds)
01:07:09  * mrb_bkjoined
01:13:31  * ericktquit (Quit: erickt)
01:14:41  * indexzeroquit (Quit: indexzero)
01:20:00  * pieternquit (Quit: pietern)
01:28:03  * abraxasjoined
01:28:15  * abraxasquit (Read error: Connection reset by peer)
01:28:47  * abraxasjoined
02:00:06  * ryahjoined
02:42:30  * iraquit (Quit: Leaving...)
02:53:53  * brsonquit (Quit: leaving)
03:12:50  * indexzerojoined
03:17:05  * ericktjoined
03:39:50  * mikealjoined
04:07:08  * c4miloquit (Remote host closed the connection)
04:10:08  * loladiroquit (Quit: loladiro)
04:12:18  * loladirojoined
04:18:08  * loladiroquit (Quit: loladiro)
04:26:38  * TheJHjoined
04:30:38  * loladirojoined
05:16:20  * TheJHquit (Quit: goodbye)
05:17:03  * ericktquit (Quit: erickt)
05:26:17  * mikealquit (Quit: Leaving.)
05:31:31  * mikealjoined
05:35:26  * paddybyersjoined
05:37:43  * loladiroquit (Quit: loladiro)
06:10:57  * mjr_joined
06:13:46  * leifquit (Ping timeout: 260 seconds)
06:16:28  * mikealquit (Quit: Leaving.)
06:17:43  * mmaleckijoined
06:20:42  * mikealjoined
06:21:01  * mikealquit (Client Quit)
06:26:17  * mikealjoined
06:34:01  * indexzeroquit (Ping timeout: 249 seconds)
06:51:17  * mikealquit (Quit: Leaving.)
07:01:33  * rendarjoined
07:15:37  * mikealjoined
07:31:38  * paddybyers_joined
07:33:53  * paddybyersquit (Read error: Operation timed out)
07:33:54  * paddybyers_changed nick to paddybyers
07:38:39  * ljacksonjoined
08:08:55  * avalanche123quit (Quit: Computer has gone to sleep.)
08:15:43  * rphillipsquit (Ping timeout: 245 seconds)
08:17:59  * rphillipsjoined
08:21:43  <CIA-155>node: Igor Zinkovsky master * r39e2552 / (4 files in 4 dirs): windows: enable creating directory junctions with fs.symlink - http://git.io/Wb00GA
09:18:25  * paddybyers_joined
09:22:04  * paddybyersquit (Ping timeout: 245 seconds)
09:22:05  * paddybyers_changed nick to paddybyers
10:00:45  * irajoined
10:05:44  * piscisaureus_joined
10:41:08  * iraquit (Quit: Leaving...)
10:58:59  * c4milojoined
11:52:33  * irajoined
11:57:27  * theColejoined
11:59:39  * bnoordhuisjoined
12:08:10  <CIA-155>node: Ben Noordhuis master * r078763a / test/addons/.gitignore : test: update addons .gitignore - http://git.io/oliOqg
12:09:23  * trondnjoined
12:12:17  <trondn>what's the latest "stable" version? I'm going to build a prototype to evaluate if I'm going to reimplement a component in C on top of libuv, or go to a totally different language ;)
12:18:06  <bnoordhuis>trondn: current master is pretty stable
12:18:19  <bnoordhuis>more stable than the v0.6 branch
12:18:30  <trondn>bnoordhuis: great. THen I'll start on top of that :)
12:22:55  <trondn>bnoordhuis: how do I "install" it?? (which header files do I need etc?)
12:23:16  <trondn>make: *** No rule to make target `install'. Stop.
12:33:38  <CIA-155>libuv: Ben Noordhuis master * r752ac30 / (4 files): unix: don't pass sockaddr to accept() - http://git.io/ReAXNQ
12:34:59  <bnoordhuis>trondn: you don't install it, it builds a static lib
12:35:25  * travis-cijoined
12:35:25  <travis-ci>[travis-ci] joyent/libuv#316 (master - 752ac30 : Ben Noordhuis): The build was broken.
12:35:25  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/cff2221...752ac30
12:35:25  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1421837
12:35:25  * travis-cipart
12:35:30  <bnoordhuis>trondn: it's meant to be easy to bundle with other projects, like node or luvit
12:36:58  <trondn>oh… sorry to hear :( I like dynamic libraries I can just upgrade without having to rebuild my entire app :S
12:44:59  <mmalecki>did we get anywhere with that req.connection.remoteAddress being null yesterday?
12:57:20  * c4miloquit (Remote host closed the connection)
13:33:11  * c4milojoined
13:52:03  <mmalecki>bnoordhuis: hey, mind if I use your Unix knowledge a bit?
14:05:33  <CIA-155>libuv: Ben Noordhuis master * r793a52f / (include/uv-private/uv-unix.h include/uv.h): unix: pack uv structs more - http://git.io/5klr7A
14:05:36  <bnoordhuis>mmalecki: shoot
14:06:04  <bnoordhuis>pfox__: take note of 793a52f ^
14:06:40  <mmalecki>bnoordhuis: basically, when I child_process.fork() and process.exit() right after that, child process is detached from the terminal. can I reattach it to parent's terminal? I tried `setsid()`, but it didn't work
14:07:40  <bnoordhuis>mmalecki: well... you can open /dev/tty
14:07:46  <bnoordhuis>but it's not quite the same as a real tty
14:07:54  <bnoordhuis>why do you want to reattach?
14:09:46  <mmalecki>bnoordhuis: I wrote a simple zero-downtime restart, which involves forking + passing a server handle, but I can either have a detached child or have a new, unused node process every time I do a restart
14:10:58  <mmalecki>bnoordhuis: https://gist.github.com/2781304 is the restart thing, kinda silly, but works fine for one of our usecases
14:14:58  * travis-cijoined
14:14:59  <travis-ci>[travis-ci] joyent/libuv#317 (master - 793a52f : Ben Noordhuis): The build is still failing.
14:14:59  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/752ac30...793a52f
14:14:59  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1422571
14:14:59  * travis-cipart
14:21:59  * ericktjoined
14:22:34  * avalanche123joined
14:29:53  <bnoordhuis>that eio_overflow test is starting to piss me off...
14:30:06  <bnoordhuis>indutny: ping
14:32:33  <bnoordhuis>piscisaureus_: https://github.com/joyent/node/issues/3314#issuecomment-5899616
14:34:22  <piscisaureus_>bnoordhuis: commented
14:39:31  * avalanche123quit (Quit: Computer has gone to sleep.)
14:44:42  * avalanche123joined
14:45:30  * isaacsjoined
14:46:10  <isaacs>good morning.
14:46:13  <isaacs>how are you gentlmen
14:46:20  <bnoordhuis>sup isaac?
14:46:43  <isaacs>good! i have completed my conference cold successfully, it seems.
14:47:04  <isaacs>how are things in gouda?
14:49:42  <bnoordhuis>i don't know, i'm in amsterdam
14:50:13  <bnoordhuis>but extrapolating from this morning's experiences i wager it's sunny
14:51:04  <isaacs>oh, cool.
14:52:18  <mmalecki>isaacs: did you get a chance to look at this req.connection.remoteAddress issue?
14:52:23  <mmalecki>isaacs: also, hi :)
14:52:27  <isaacs>mmalecki: no
14:52:30  <isaacs>mmalecki: link?
14:53:10  <mmalecki>isaacs: I didn't create one yet, I don't have too many details :(. I will put some logging into this thing and gather some info
14:53:19  <isaacs>oh, ok
14:53:21  <isaacs>kewl
14:53:26  <mmalecki>but the gist is, req.connection.remoteAddress is null sometimes
14:53:40  <deoxxa>it's a ghost!
14:54:11  <mmalecki>a production ghost!
14:54:22  <deoxxa>that's the scariest kind
14:54:39  <mmalecki>no, there is one, even more scary
14:54:44  <mmalecki>a cloud ghost!
14:55:10  <mmalecki>involves your VMs randomly disappearing and such
14:57:01  <deoxxa>randomly disappearing i can live with - it's randomly coming back that really ruins your day
14:57:09  * trondnpart
14:57:28  <isaacs>mmalecki: if you can reproduce it, that'd be great
14:57:33  <mmalecki>it's even worse when they come back as windows machines :)
14:57:55  <deoxxa>haha
14:58:27  <mmalecki>isaacs: I was able to reproduce few times with a websocket connection over http, but nothing testable
15:01:55  <isaacs>mmalecki: a test that's sporadically reproducible is better than a description in english. if you can get that, and it fails 10% of the time or whatever, even, please post an issue.
15:01:57  <isaacs>doesn't have to be 100%
15:02:15  <isaacs>of course, 100% failing is better :)
15:08:44  * rendarquit
15:15:05  <avalanche123>hello folks, just added shared library build target to gyp - https://github.com/joyent/libuv/pull/428
15:15:58  * pieternjoined
15:17:50  <avalanche123>bnoordhuis ping
15:18:32  <pfox__>bnoordhuis: ping
15:20:38  * pieternquit (Client Quit)
15:23:11  <bnoordhuis>avalanche123, pfox__: pong
15:23:39  <avalanche123>bnoordhuis I made the build patch in gyp, it now generates shared and static library products
15:23:56  <bnoordhuis>avalanche123: i'll look at it this week
15:24:00  <avalanche123>thanks
15:24:41  <pfox__>bnoordhuis: asking about that commit you ref'd for me.. is that struct layout change?
15:24:47  <bnoordhuis>pfox__: yes
15:24:50  * pfox__nods
15:25:00  <pfox__>i expect a fair amount of that once we update our submodule
15:25:09  <pfox__>we have tests for all of the structs we map from rust
15:25:17  <bnoordhuis>okay, cool
15:25:19  <pfox__>so it'll fail on the first run, and i'll know the size of the structs in C
15:25:33  <pfox__>so i just pack in void pointers, and we're good to go
15:26:15  <pfox__>i also have 8bit ints if the math is off. happened with, like, sockaddr_in6 i believe
15:26:33  <pfox__>anywho. i'd love a more elegant solution :)
15:27:15  <pfox__>someday maybe we'll have the ability to declare opaque records in rust that correspond to a struct in a linked library. it's entirely possible with libclang, which rust has taken a dep on
15:27:26  <pfox__>someday :)
15:27:29  <bnoordhuis>heh
15:35:02  * loladirojoined
15:45:31  * pieternjoined
15:49:09  <isaacs>bnoordhuis, piscisaureus_, igorzi__, TooTallNate, indutny: skype in 12 mins
15:49:26  <isaacs>TooTallNate might be jetlagged...
15:49:38  * rendarjoined
16:03:56  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
16:05:31  * piscisaureus_joined
16:07:05  * perezdjoined
16:21:28  <isaacs>igorzi__: do you have a ssh key?
16:25:57  * TheJHjoined
16:28:26  <isaacs>bnoordhuis: oh, i forgot to ask, did you guys get anywhere with the cluster perf issue?
16:29:35  * piscisaureus_quit (Ping timeout: 265 seconds)
16:39:49  * mmaleckichanged nick to mmalecki[brb]
16:45:34  * piscisaureus_joined
16:54:34  * isaacsquit (Remote host closed the connection)
16:55:23  * isaacsjoined
16:56:15  <CIA-155>node: Jeroen Janssen master * r66e12db / doc/api/process.markdown : doc: update possible values for process.platform - http://git.io/QZSamw
16:57:25  * isaacsquit (Remote host closed the connection)
17:02:04  * brsonjoined
17:04:29  <avalanche123>bnoordhuis looks like gyp is building wrong architecture for mac
17:04:38  <avalanche123>any ideas how I should hint it?
17:07:46  * avalanche123quit (Quit: Computer has gone to sleep.)
17:09:42  * brsonquit (Ping timeout: 265 seconds)
17:10:09  * mikealquit (Quit: Leaving.)
17:10:12  * brsonjoined
17:11:23  * mikealjoined
17:21:25  * piscisaureus_quit (Ping timeout: 252 seconds)
17:21:47  * bnoordhuisquit (Ping timeout: 256 seconds)
17:48:08  <creationix>what in libuv would emit "Couldn't get a file descriptor referring to the console" Best I can tell that's an openvt error
17:48:18  <creationix>This is on Linux
17:49:52  * AndreasMadsenjoined
17:51:07  * igorzi__quit (Ping timeout: 245 seconds)
17:52:55  * isaacsjoined
17:53:05  * AndreasMadsenquit (Remote host closed the connection)
17:53:12  * AndreasMadsenjoined
18:03:04  * ericktquit (Quit: erickt)
18:07:08  * Raynosquit (Ping timeout: 244 seconds)
18:12:58  * Raynosjoined
18:15:37  * paddybyers_joined
18:18:30  * paddybyersquit (Ping timeout: 244 seconds)
18:18:30  * paddybyers_changed nick to paddybyers
18:25:07  * perezdquit (Read error: Connection reset by peer)
18:25:28  * perezdjoined
18:25:41  * AvianFlujoined
18:26:28  * brsonquit (Read error: Operation timed out)
18:28:19  * rphillipsquit (Ping timeout: 244 seconds)
18:29:02  * mmalecki[brb]changed nick to mmalecki
18:31:54  * mikealquit (Quit: Leaving.)
18:32:46  <indutny>ircretary: tell isaacs that I'm sorry that I missed call. Was in a subway and thought call was at Wednesday
18:32:47  <ircretary>indutny: I'll be sure to tell isaacs
18:33:46  * avalanche123joined
18:34:02  * rphillipsjoined
18:34:53  * brsonjoined
18:43:50  <isaacs>indutny: hey, no worries
18:43:58  <indutny>isaacs: ok, cool
18:44:11  <indutny>sorry, I was confused by previous week's call
18:45:00  <isaacs>indutny: it's no big deal. we do things async in node :)
18:45:05  <indutny>hahahahaha
18:45:06  <indutny>right
18:45:07  <isaacs>indutny: how's stdio coming?
18:45:40  <isaacs>indutny: i saw that you were working on that, but i had spotty internet, and then was sick, so didnt' follow too closely.
18:45:48  <indutny>isaacs: well, it's going well. I think I'll push more updates to my pull request tomorrow
18:45:57  <indutny>isaacs: yeah, I was on very spotty internet too
18:46:09  <indutny>isaacs: connected to the open wi-fi of my neighboors :D
18:46:13  <isaacs>hahah
18:46:14  <isaacs>that's the best
18:46:20  <isaacs>make sure to change their router password.
18:46:30  <indutny>isaacs: hahaha
18:46:35  <isaacs>and then set up mac address filtering, so that only they and you can use it.
18:46:55  <indutny>nice
18:47:04  <indutny>you're very experienced in such things, I got it? :D
18:47:10  <indutny>arrgh
18:47:13  <isaacs>well, i mean, i care for my neighbors.
18:47:13  <indutny>I take it
18:47:21  <indutny>(how this should be written)
18:47:22  <isaacs>don't want just any old random person using their internet.
18:47:26  <indutny>hahaha
18:47:29  <indutny>isaacs: you saved them
18:48:05  <indutny>isaacs: but considering it once again, I can never be sure that those two MAC addressess are only mine and their
18:48:17  <isaacs>most people in oakland have very secure wifi, unfortunately. i live on the 7th floor of a building with lots of windows, in a neighborhood with very few tall buildings, so it's basically like being on an antenna.
18:48:30  <indutny>heh
18:48:32  <isaacs>maybe 50-100 wifi networks visible, but all of them are locked.
18:48:38  <indutny>isaacs: well, I'm living in the center of Moscow atm
18:48:54  <indutny>isaacs: many secure wi-fi networks here too
18:49:04  <indutny>isaacs: almost everyone has wi-fi router
18:49:11  <indutny>LTE is still not there
18:50:35  <mjr_>I will refrain from my wireless density rant.
18:50:59  <CIA-155>node: Igor Zinkovsky master * rb4ed3c1 / src/node_file.cc : remove NODE_USE_64BIT_UV_FS_API - http://git.io/AtXfjQ
18:53:51  <indutny>mjr_: heh, well it's not so dense in Omsk, that's why I stated it
18:54:21  <indutny>actually I seen only 3-4 networks in Omsk
18:54:26  <indutny>while about 10-12 in Moscow
18:54:31  <indutny>(though, that's expected)
18:56:04  <indutny>isaacs: how does 64bit FS works in node now?
18:57:06  <indutny>isaacs: fs.read is still accepting 31bit integers
18:57:18  <indutny>I mean javascript side, of course
18:59:46  * indexzerojoined
19:00:08  <isaacs>indutny: i'm not sure.
19:00:34  <isaacs>indutny: haven't looked too closely at it. igorzi would know.
19:00:55  <indutny>isaacs: I think we should allow users to push buffers instead of int offset
19:01:16  <indutny>or array of two ints
19:01:20  <isaacs>hm.
19:01:37  <indutny>I'm just talking about my use-case - btree implementation
19:02:05  <isaacs>right, so, like a JS bigint impl
19:02:44  <isaacs>or a int64 as a buffer[4]
19:04:49  * brsonquit (Ping timeout: 252 seconds)
19:05:28  * brsonjoined
19:05:38  <indutny>yep
19:05:43  <indutny>sort of
19:06:34  * iraquit (Quit: Textual IRC Client: http://www.textualapp.com/)
19:09:32  * AvianFluquit (Ping timeout: 244 seconds)
19:11:22  * TheJHquit (Read error: Operation timed out)
19:19:02  * igorzijoined
19:22:03  * brsonquit (Ping timeout: 244 seconds)
19:22:29  * mikealjoined
19:24:48  * AvianFlujoined
19:26:31  * piscisaureus_joined
19:31:43  * mikealquit (Quit: Leaving.)
19:33:14  * mikealjoined
19:33:40  * piscisaureus_quit (Ping timeout: 252 seconds)
19:34:20  * ericktjoined
19:34:48  <isaacs>igorzi, piscisaureus, bnoordhuis: review? https://gist.github.com/2783736
19:35:03  <isaacs>ircretary tell bnoordhuis review? https://gist.github.com/2783736
19:35:04  <ircretary>isaacs: I'll be sure to tell bnoordhuis
19:46:01  * AndreasMadsenquit (Remote host closed the connection)
19:51:29  <avalanche123>isaacs ping
20:36:26  * TheJHjoined
20:36:35  * brsonjoined
20:39:41  <isaacs>hola
20:39:43  <isaacs>avalanche123: pong
20:40:06  <avalanche123>isaacs I noticed we're forcing 32 bit arch in gyp builds
20:40:10  <avalanche123>do you know anything about it?
20:40:25  <isaacs>avalanche123: seems suboptimal
20:40:30  <avalanche123>yeah
20:40:35  <isaacs>avalanche123: we do both for the mac pkg, i know
20:40:50  <avalanche123>isaacs awesome, I added support to pass arch on the command line
20:40:53  <TooTallNate>isaacs: no it's a universal binary
20:41:22  <isaacs>TooTallNate: a true universal? i thought we just built both and then lipo'd them together
20:41:29  <TooTallNate>well, ya
20:41:40  <avalanche123>what is true universal?
20:41:57  <TooTallNate>avalanche123: -arch x64 -arch i386 at the same time
20:42:00  <tjfontaine>when you use clang and pass multiple -arch's
20:42:03  <TooTallNate>but gyp doesn't support it set
20:42:03  <avalanche123>ah
20:42:08  <avalanche123>yeah
20:42:10  <TooTallNate>afaik
20:42:18  <avalanche123>would be cool to use pkg approach in gyp too
20:43:13  <avalanche123>maybe I'll add the lipo step to it
20:43:35  <TooTallNate>avalanche123: are you talking about libuv build or node?
20:43:41  <avalanche123>libuv
20:43:46  <TooTallNate>ahh
20:43:56  <TooTallNate>cause in node, the ./configure script takes care of that
20:44:04  <TooTallNate>by creating a config.gypi file
20:44:13  <avalanche123>maybe I should look at that
20:44:32  <avalanche123>TooTallNate think it would be straightforward to port it to libuv?
20:44:43  <TooTallNate>ya i think so
20:45:24  <avalanche123>ok, I'll look into it
20:46:21  * piscisaureus_joined
20:56:03  * piscisaureus_quit (Read error: Connection reset by peer)
20:59:16  * piscisaureus_joined
21:00:51  * avalanche123quit (Quit: Computer has gone to sleep.)
21:02:20  <CIA-155>libuv: isaacs master * r604802a / src/unix/core.c : Fix sunos build: undeclared variable 'handle' - http://git.io/E0_gRw
21:02:48  <isaacs>piscisaureus_: thanks :)
21:02:57  <piscisaureus_>isaacs: np :-)
21:03:04  <isaacs>ok, now time to do 0.7.9 build.
21:03:08  <isaacs>got anything you want in? now's the time
21:03:28  <piscisaureus_>nope
21:03:41  * travis-cijoined
21:03:41  <travis-ci>[travis-ci] joyent/libuv#318 (master - 604802a : isaacs): The build is still failing.
21:03:41  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/793a52f...604802a
21:03:41  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1426757
21:03:41  * travis-cipart
21:03:42  <piscisaureus_>nothing that's close enough to being finished that can make it anyway
21:03:48  <isaacs>kewl
21:07:01  * avalanche123joined
21:10:35  * avalanche123quit (Client Quit)
21:14:46  * irajoined
21:19:05  * brsonquit (Ping timeout: 260 seconds)
21:22:59  * brsonjoined
21:23:40  * AvianFluquit (Remote host closed the connection)
21:36:34  <igorzi>piscisaureus_: hey.. yt?
21:42:31  <piscisaureus_>igorzi: yeo
21:42:33  <piscisaureus_>*yep
21:42:40  <piscisaureus_>igorzi: sup?
21:43:10  <isaacs>piscisaureus_: final answer: remove test-next-tick-ordering2.js?
21:43:24  <piscisaureus_>isaacs: no
21:43:36  <piscisaureus_>isaacs: unless we decide to actually change the nexttick semantics
21:43:39  <isaacs>that's what i mean
21:43:52  <piscisaureus_>isaacs: but right now it just catches an unix bug -
21:44:06  <igorzi>piscisaureus_: for stdio what uv_stream_t do we accept? just pipe and tty?
21:44:08  <isaacs>k.
21:44:19  <isaacs>i'm just so used to only making builds when all the tests pass...
21:44:21  <isaacs>feels so wrong..
21:44:21  <piscisaureus_>igorzi: yes that seems enough for now
21:44:29  <igorzi>piscisaureus_: ok
21:44:30  <piscisaureus_>isaacs: there *is* a bug
21:45:40  <igorzi>piscisaureus_: in case of uv_tty_t we just use the parent process's stdio handles, right?
21:45:46  <isaacs>piscisaureus_: ok, fair enough.
21:45:54  <piscisaureus_>igorzi: no
21:46:25  <piscisaureus_>igorzi: well, I don't think it matters but you can also just rip the handle out of the uv_tty_t
21:46:46  <piscisaureus_>that seems cleaner to me
21:46:50  <igorzi>piscisaureus_: ok
21:47:47  <igorzi>piscisaureus_: i guess we need to rework indutny's api a bit.. we either need to add UV_TTY flag or remove UV_CREATE_PIPE (and just switch on uv_stream_t.type)
21:49:02  <piscisaureus_>igorzi: no it works differently... if UV_CREATE_PIPE is specified uv_spawn *creates* a pipe for you and populates the uv_stream_t* field
21:49:30  <piscisaureus_>er, it initializes the handle that was passed
21:49:45  <igorzi>piscisaureus_: i thought we decided that uv_stream_t* must already be initialized, no?
21:49:59  <piscisaureus_>igorzi - ok, sure, but then it connects it
21:50:20  <igorzi>piscisaureus_: if it's initialized - then uv_stream_t.type is already set.. why do we need UV_CREATE_PIPE?
21:50:31  <piscisaureus_>igorzi: well there's two different modes
21:50:42  <piscisaureus_>with UV_CREATE_PIPE the passed stream is the parent end of a pipe
21:51:16  <piscisaureus_>igorzi: *without* that the passed pipe is set as the stdio for the child
21:51:52  <igorzi>piscisaureus_: oh
21:51:59  <igorzi>piscisaureus_: i kind of missed that
21:52:11  <piscisaureus_>yeah it's kind of non obvious
21:52:12  <piscisaureus_>:-)
21:52:48  <piscisaureus_>like, with CREATE_PIPE an actual pipe (with two ends) is created
21:52:52  <piscisaureus_>what node does now
21:53:08  <piscisaureus_>but the idea is that sometimes you actually have an fd or a thing that you want to set as the child end
21:53:46  <igorzi>piscisaureus_: ok.. makes sense.. so really UV_CREATE_PIPE is added convenience
21:53:54  <piscisaureus_>igorzi: in a way, yes
21:54:14  <piscisaureus_>igorzi: but additionally, libuv needs to decide whether it opens the pipe in overlapped mode or not
21:54:50  <piscisaureus_>igorzi: normally it is better to have the node end in overlapped mode, but for safety purposes we make the child end blocking
21:55:09  <piscisaureus_>igorzi: but in order to decide that libuv needs to know what end will be the parent and which end should be the child's
21:55:45  <igorzi>piscisaureus_: so without UV_CREATE_PIPE - it should always be non-overlapped, right?
21:55:59  <piscisaureus_>igorzi: well, I think we should ignore this problem
21:56:03  <igorzi>piscisaureus_: with UV_CREATE_PIPE - uv_pipe_t.ipc decides that
21:56:15  <piscisaureus_>the later is definitely true
21:56:24  <piscisaureus_>igorzi: obviously there's a problem here
21:56:52  <piscisaureus_>since if we do uv_pipe_connect() it will also create an overlapped pipe
21:57:14  <CIA-155>node: isaacs master * r24de89b / (20 files in 6 dirs): Upgrade uv to 604802a - http://git.io/lIRIMA
21:57:20  <piscisaureus_>hmm
21:57:40  <piscisaureus_>igorzi: yeah maybe we can't even support pipes that are created by libuv
21:58:05  <piscisaureus_>shit
21:59:21  <isaacs>jesus, this changelog is looooonnnnngggg
21:59:30  <isaacs>waited way too long to do 0.7.9
22:00:32  <igorzi>piscisaureus_: wait.. so what's the problem exactly?
22:00:32  * bnoordhuisjoined
22:01:01  <DrPizza>hi node people
22:01:06  <DrPizza>I mentioned you in a post I wrote today. http://arstechnica.com/information-technology/2012/05/no-cost-desktop-software-development-is-dead-on-windows-8/
22:01:06  <piscisaureus_>igorzi: so you create a pipe with uv_pipe_connect
22:02:01  <DrPizza>igorzi: can you refresh my memory, I think you said once you were working on some other project to help bring it to windows, but that you had to maintain it as a fork because the upstresam devs didn't care
22:02:04  <piscisaureus_>igorzi: then you try to pass that end to a child
22:02:13  <piscisaureus_>DrPizza: I think that was redis
22:02:18  <DrPizza>ah yeah
22:02:34  <igorzi>yeah redis
22:03:03  <DrPizza>is it that tehy just don't give a damn about windows, or are the changes too significant or something?
22:03:33  <igorzi>DrPizza: the former
22:04:05  <igorzi>piscisaureus_: we have some code in pipe.c that detects if the pipe is overlapped or not
22:04:55  <piscisaureus_>igorzi: true, so it works as long as the child process is node
22:05:15  <DrPizza>igorzi: thanks.
22:05:17  <piscisaureus_>igorzi: but I suppose we could live with the fact that it'll break if the child is not node
22:05:32  <piscisaureus_>igorzi: so, windows people, just don't do that
22:06:42  <igorzi>piscisaureus_: yep, if they want a pipe they should just use UV_CREATE_PIPE and let uv_spawn do the right thing
22:08:04  <piscisaureus_>igorzi: exactly
22:08:16  <piscisaureus_>igorzi: so, that's the reason that we have UV_CREATE_PIPE in the first place :-)
22:08:38  <igorzi>piscisaureus_: thanks.. that was enlightening :)
22:11:10  * loladiroquit (Quit: loladiro)
22:20:10  * paddybyersquit (Quit: paddybyers)
22:22:45  <isaacs>igorzi: lstat should work properly in windows now, yes? so fs.lstatSync("some-symlink").isSymlink() == true?
22:23:00  <isaacs>(haven't tried it yet, just writing changelog)
22:23:03  <igorzi>isaacs: yes
22:23:06  <isaacs>kewl
22:23:10  <isaacs>igorzi++
22:23:19  <isaacs>that's awesome.
22:29:29  <isaacs>piscisaureus_: got a couple-word summary of the uv poll refactor?
22:30:39  <piscisaureus_>isaacs: add uv_poll interface for joining external event loops through file descriptor state notifications
22:30:43  <piscisaureus_>isaacs: something like that
22:30:53  <piscisaureus_>isaacs: it's a long story in reality :-)
22:31:24  <isaacs>yeah, it's one of those things... the only thing i could come up with was "Bert said it was important and would make stuff good... so... he did it... um..."
22:32:02  <isaacs>piscisaureus_: we really need to start getting libuv having releases and changelogs and stuff.
22:32:15  <isaacs>piscisaureus_: then i can just do "Upgrade UV to x.y.z" and be done with it
22:32:39  <CIA-155>libuv: Ben Noordhuis v0.6 * raeb580d / src/unix/ev/ev.c : unix: remove abort() in ev_unref() - http://git.io/ClojIA
22:33:45  <CIA-155>libuv: Ben Noordhuis v0.6 * rcd69ea9 / test/runner.c : test: silence compiler warning - http://git.io/2JUzag
22:33:48  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/libuv/compare/v0.6...accept-backoff
22:34:13  * travis-cijoined
22:34:14  <travis-ci>[travis-ci] joyent/libuv#319 (v0.6 - cd69ea9 : Ben Noordhuis): The build is still failing.
22:34:14  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/bc4126b...cd69ea9
22:34:14  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1427818
22:34:14  * travis-cipart
22:36:58  <CIA-155>node: Ben Noordhuis v0.6 * r7535e39 / benchmark/http_simple_auto.js : bench: add http_simple_auto benchmark - http://git.io/k73F_Q
22:36:59  <CIA-155>node: Ben Noordhuis v0.6 * r5ff2ae8 / benchmark/http_simple_cluster.js : bench: start a worker for each CPU - http://git.io/k7jZmQ
22:38:26  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/node/compare/v0.6...accept-backoff <- in node
22:40:39  <piscisaureus_>bnoordhuis: ok so now I am going to re-land the patch that removes the env var :-0
22:40:49  <bnoordhuis>piscisaureus_: go for it :)
22:40:57  <isaacs>igorzi: this fs 64 bit offset thing... is this actually exposed in node somehow? it seems like we're still using just plain old js numbers.
22:41:15  <bnoordhuis>isaacs: i think we use NumberValue(), which is as good as it gets
22:41:38  <isaacs>hm... ok
22:41:51  * rendarquit
22:41:51  <isaacs>but that's still just 51 bits or whatever?
22:41:54  <isaacs>53? i forget.
22:41:56  * mmalecki_joined
22:41:57  <piscisaureus_>bnoordhuis: ah so probably you patched it another time after I pushed it to ajaxorg/node ;-)
22:42:18  <bnoordhuis>piscisaureus_: i never pushed anything to ajaxorg/node
22:42:27  <piscisaureus_>bnoordhuis: I know... I did
22:42:42  <piscisaureus_>because I wanted to put the remove-env-var patch somewhere
22:43:31  * mmalecki_quit (Client Quit)
22:43:43  * mmaleckiquit (Ping timeout: 252 seconds)
22:43:48  * mmalecki_joined
22:43:51  <bnoordhuis>isaacs: 52 :)
22:43:58  <bnoordhuis>actually, slightly less
22:44:40  <bnoordhuis>err, slightly less than 53 bits
22:44:49  <bnoordhuis>man, it's getting late
22:45:18  * mmalecki_quit (Client Quit)
22:45:36  * mmaleckijoined
22:46:31  <isaacs>right, i'ts like 2^53-1, right?
22:46:34  <isaacs>or something like that
22:46:56  <isaacs>i would do a lot of things for a proper int64 type in JavaScript
22:47:15  <isaacs>especially if it was like, no actual change in the language, just that math starts working right.
22:49:47  <bnoordhuis>DrPizza: interesting article
22:49:58  <bnoordhuis>isaacs: yes, agreed. one can dream...
22:50:09  <DrPizza>obviously for projects like node it doesn't make a difference immediately
22:50:14  <piscisaureus_>DrPizza: yes. And the vs express plans suck indeed
22:50:28  <DrPizza>but a few years down the line?
22:50:30  <piscisaureus_>I think that will just pin us to vs10 forever
22:50:33  <DrPizza>yes
22:50:46  <bnoordhuis>unless ms keeps handing out msdn subscriptions :)
22:50:47  <DrPizza>but what happens if google wants to switch chrome/v8 to VS11 or VS12 or whatever it'll be by then
22:50:52  <piscisaureus_>although gyp makes the pain more bearable :-)
22:51:11  <piscisaureus_>yes, that would suck
22:51:24  <DrPizza>obviously google doesn't care about $500/head
22:51:24  <piscisaureus_>handing out these subscriptions doesn't really help
22:51:25  * TheJHquit (Read error: Connection reset by peer)
22:51:28  <DrPizza>no
22:51:36  <DrPizza>it's like people saying "oh just use dreamspark/bizspark"
22:51:42  <DrPizza>that's not an effective long-term solution
22:51:50  <DrPizza>and if these subscriptions are so worthless to microsoft
22:51:53  <DrPizza>why not just make VS free
22:52:06  <piscisaureus_>That's not the reason
22:52:15  <piscisaureus_>they actually want people to stop writing desktop apps
22:52:29  <piscisaureus_>because metro apps need to be sold via marketplace
22:52:42  <piscisaureus_>so more metro apps equals more $$
22:52:44  <DrPizza>yes
22:52:45  <DrPizza>I know
22:52:51  <DrPizza>it's not som uch the commission, I don't think
22:53:03  <DrPizza>rather, it's that MS needs lots of metro apps if win8 is to have any chance of competing with iPad
22:53:10  <isaacs>anything missing? anything you think should be omitted? http://nodejs.org/dist/v0.7.9/blog.html
22:53:21  <DrPizza>but what I find so implausible is the implication that a hobbyist developer will sit down with vS11
22:53:26  <DrPizza>discover that he can't write a desktop app with express
22:53:32  <DrPizza>and then say "oh never mind, I'll write a metro app instead"
22:53:39  <piscisaureus_>yeah, that's not gonna work
22:54:09  <igorzi>isaacs: for "windows: enable creating directory junctions with fs.symlink" - pls put Bert's name as well
22:54:16  <isaacs>k
22:54:19  <DrPizza>it hurts the desktop, and It hink it's *supposed* to help Metro, I just can't see how it will
22:55:21  <isaacs>it's a good thing stdio didn't land in this release.
22:55:30  <isaacs>wouldn't've had room to put it in the changelog ;)
22:56:15  <bnoordhuis>isaacs: fb7348a crypto: add PKCS12/PFX support
22:56:30  <isaacs>bnoordhuis: i think that was in a different release already, wasn't it?
22:56:40  <tjfontaine>actually google is moving to clang, so long as the ms stuff continues to progress there they will have less reason to depend on vs
22:56:47  <isaacs>bnoordhuis: * crypto: add PKCS12/PFX support (Sambasiva Suda), v0.6.12
22:56:55  <isaacs>bnoordhuis: *v0.6.18
22:56:57  * indexzeroquit (Quit: indexzero)
22:57:25  <bnoordhuis>isaacs: oh, you're right. this is the first v0.7.x release it's in
22:57:48  <bnoordhuis>did bert's string optimization work land in v0.6 too?
22:59:22  <isaacs>i think there's a line for that...
22:59:37  <bnoordhuis>found it: Optimize writing strings with Socket.write (Bert Belder)
22:59:45  <isaacs>oh, no, i think i might've deleted it
22:59:59  <isaacs>ah, yeah, there it is
23:00:19  <bnoordhuis>lgtm
23:00:23  <isaacs>also, it's not going out tonight ;)
23:00:43  <isaacs>but i'll get it monday
23:03:42  * c4miloquit (Remote host closed the connection)
23:07:02  <bnoordhuis>piscisaureus_: upside to the new accept patch is that the performance difference is negligible
23:07:45  <piscisaureus_>bnoordhuis: really? So it did matter?
23:07:48  <bnoordhuis>i see the same numbers on tcp4_pound_100 before and after
23:07:54  <bnoordhuis>it would seem so :)
23:11:09  <piscisaureus_>bnoordhuis: so we can just put it into master and auto-enable for cluster :-)
23:11:09  <piscisaureus_>nice
23:11:37  <bnoordhuis>seems like it
23:12:01  <bnoordhuis>on tcp4_pound_1000 it scores even better than the non-relaxed version
23:12:13  <bnoordhuis>because it's not exhausting the available port range as fast :)
23:13:43  * perezdquit (Quit: perezd)
23:14:35  * perezdjoined
23:16:24  <bnoordhuis>https://github.com/mattgodbolt/gcc-explorer <- nice
23:16:47  * perezdquit (Client Quit)
23:18:05  <mmalecki>isaacs: does this test runner need to have all these silly features like few kinds of output modes and progress bars?
23:18:10  * mmaleckifinally got to writing it
23:18:25  <isaacs>mmalecki: that's not super important.
23:18:30  <bnoordhuis>i want a knight rider-style progress bar!
23:18:39  <isaacs>mmalecki: the important thing is that it prints out the results in a nice way at the end.
23:18:49  <isaacs>mmalecki: how many tests passed, how many failed, and the names of the failures.
23:18:58  <isaacs>mmalecki: also, don't bother printing stdout/stderr unless it fails.
23:19:13  <mmalecki>isaacs: ack'd
23:19:16  <mmalecki>bnoordhuis: :)
23:19:16  <isaacs>mmalecki: i personally really like tap-style output, but that's just me.
23:19:35  <mmalecki>bnoordhuis: I watched this series when I was a kid!
23:20:01  <bnoordhuis>mmalecki: was it still being broadcasted in 2004?
23:20:47  <mmalecki>bnoordhuis: haha. yeah, in Poland :)
23:25:33  <CIA-155>libuv: Ben Noordhuis master * rce28e13 / src/unix/internal.h : unix: remove unused __attribute__((unused)) - http://git.io/nvjSvQ
23:27:24  * travis-cijoined
23:27:24  <travis-ci>[travis-ci] joyent/libuv#320 (master - ce28e13 : Ben Noordhuis): The build is still failing.
23:27:24  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/604802a...ce28e13
23:27:24  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1428358
23:27:24  * travis-cipart
23:34:52  <CIA-155>libuv: Ben Noordhuis master * r28766dc / (test/test-list.h uv.gyp test/test-eio-overflow.c): test: remove test-eio-overflow.c - http://git.io/mfhekA
23:39:01  * isaacsquit (Remote host closed the connection)
23:42:31  * travis-cijoined
23:42:31  <travis-ci>[travis-ci] joyent/libuv#321 (master - 28766dc : Ben Noordhuis): The build was fixed.
23:42:31  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/ce28e13...28766dc
23:42:31  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1428454
23:42:31  * travis-cipart
23:47:54  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
23:50:36  * piscisaureus_joined