00:01:25  <paddybyers>isaacs: https://github.com/paddybyers/node/issues/19
00:01:51  <isaacs>paddybyers: you probably want to post that on https://github.com/joyent/node/issues
00:01:55  <isaacs>not on your fork
00:02:13  <paddybyers>I guess
00:02:33  <isaacs>paddybyers: or better yet, make it a pull request. sign the cla first: http://nodejs.org/cla.html
00:07:46  <paddybyers>https://github.com/joyent/node/pull/2125
00:30:44  * mikealjoined
00:41:58  * paddybyersquit (Quit: paddybyers)
01:09:25  * bnoordhuisquit (Ping timeout: 240 seconds)
01:18:47  * mikealquit (Quit: Leaving.)
01:19:23  * isaacsquit (Quit: isaacs)
01:23:12  * mralephquit (Quit: Leaving.)
01:30:24  * dapquit (Quit: Leaving.)
01:42:45  * AvianFlu_joined
01:42:53  * AvianFluquit (Disconnected by services)
01:42:55  * AvianFlu_changed nick to AvianFlu
01:52:17  * pieternquit (Quit: pietern)
02:15:19  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:22:17  * isaacsjoined
02:32:53  * isaacsquit (Quit: isaacs)
02:38:06  * AvianFluquit (Quit: Leaving)
02:43:53  * mikealjoined
02:47:56  * brsonquit (Quit: leaving)
03:49:46  * mikealquit (Quit: Leaving.)
04:15:57  * mikealjoined
05:02:49  * isaacsjoined
05:03:15  * isaacsquit (Client Quit)
05:58:19  * mikealquit (Quit: Leaving.)
06:26:53  * mikealjoined
06:44:20  * paddybyersjoined
07:09:44  * mikeal1joined
07:12:16  * mikealquit (Ping timeout: 240 seconds)
08:28:54  * mikeal1quit (Quit: Leaving.)
08:30:07  * kuebkjoined
08:36:11  * isaacsjoined
08:45:04  * isaacsquit (Read error: Connection reset by peer)
08:45:13  * isaacsjoined
09:07:20  * mikealjoined
09:12:59  * mralephjoined
09:56:01  * mralephquit (Quit: Leaving.)
09:58:19  * mralephjoined
09:58:57  * mralephquit (Client Quit)
10:13:50  * isaacsquit (Quit: isaacs)
10:27:19  * piscisaureus_joined
11:29:37  * bnoordhuisjoined
11:59:12  * piscisaureus_quit (Ping timeout: 245 seconds)
12:28:11  * creationixjoined
12:32:08  * piscisaureus_joined
13:32:23  <bnoordhuis>node + threads, default joinable or detached?
13:33:08  <bnoordhuis>piscisaureus_: does windows make that distinction?
13:38:07  <bnoordhuis>i have to say i like that almost everything is waitable on in windows - mutexes, threads, semaphores, sockets, etc.
13:44:15  * indutnyquit (*.net *.split)
13:48:23  * indutnyjoined
13:48:32  * DrPizza_quit (Excess Flood)
13:48:39  * DrPizzajoined
13:49:04  <paddybyers>bnoordhuis: are you planning to create a thread abstraction in uv to include things like uv_thread_join?
13:49:12  <bnoordhuis>paddybyers: yep
13:49:33  <bnoordhuis>though i'm not quite set on uv_thread_join yet
14:23:44  * creationixquit (Ping timeout: 248 seconds)
14:33:40  * indutnyquit (*.net *.split)
14:50:26  * indutnyjoined
14:50:58  <indutny>sorry, stupid irccloud
14:51:00  <indutny>just connected
14:51:16  <bnoordhuis>indutny: use a real irc client :)
14:54:38  <indutny>haha :)
14:54:57  <indutny>I like offline messages
14:55:08  <bnoordhuis>you have a server right?
14:55:28  <bnoordhuis>just start irssi in a screen session and reconnect whenever you turn on your machine
14:57:27  <indutny>hm...
14:57:42  <indutny>nice tip
14:57:49  <indutny>thanks
14:57:53  * indutnyquit (Remote host closed the connection)
15:08:22  * creationixjoined
15:11:33  * indutnyjoined
15:11:38  * piscisaureus_quit (Ping timeout: 252 seconds)
15:16:26  <bnoordhuis>ryah: we should fix https://github.com/joyent/libuv/issues/206 ...
15:19:33  * Marakjoined
15:24:37  * indutnyquit
15:24:38  * piscisaureus_joined
15:26:03  * indutnyjoined
15:26:07  <indutny>heya
15:26:12  <indutny>can you see me?
15:26:25  <indutny>bnoordhuis: ping
15:26:52  <bnoordhuis>indutny: pong
15:27:02  <indutny>cool
15:27:14  <indutny>going to restart this thing under non-root :D
15:27:17  <indutny>brb
15:27:19  * indutnyquit (Client Quit)
15:30:21  * indutnyjoined
15:30:27  <indutny>ok, finally back
15:33:00  <indutny>wow, window switching is interesting
15:33:01  <indutny>kinda odd
15:33:49  <bnoordhuis>you mean ctrl-a?
15:34:08  <indutny>what is ctrl-a ?
15:34:20  <indutny>ctrl-p, ctrl-n
15:34:36  <bnoordhuis>are we talking about screen or irssi?
15:34:51  <indutny>bnoordhuis: irssi, of course
15:35:20  <bnoordhuis>right, i thought it was about screen
15:37:38  <piscisaureus_>bnoordhuis: we are debugging node 0.4 dns issues
15:37:51  <piscisaureus_>bnoordhuis: we see recvfrom with fd=-1
15:37:54  <piscisaureus_>is that a bug?
15:38:07  <bnoordhuis>piscisaureus_: depends
15:38:18  <piscisaureus_>are there normal cases where you would do that?\
15:38:18  <bnoordhuis>i've observed c-ares doing that for no good reason
15:38:32  <bnoordhuis>no, it will always fail with EBADF
15:38:50  <piscisaureus_>but why does it do that in the first place
15:38:55  <piscisaureus_>it looks fucked up
15:39:03  <bnoordhuis>never looked much into it
15:39:12  <bnoordhuis>but i can speculate
15:39:41  <bnoordhuis>well... no, let's not
15:40:42  <bnoordhuis>piscisaureus_: can you describe the issue?
15:42:15  <piscisaureus_>bnoordhuis: c-ares randomly fails to look in the hosts file
15:42:19  <piscisaureus_>and returns ENOTFOUND
15:42:29  <piscisaureus_>but in happens intermittently
15:42:36  <piscisaureus_>so sometimes it works (without restarting the process)
15:42:59  <piscisaureus_>but we can't (yet?) reproduce it standalone
15:43:12  <bnoordhuis>piscisaureus_: what os?
15:43:15  <piscisaureus_>solaris
15:43:30  * bnoordhuisthinks
15:44:03  * indutnyquit (Quit: leaving)
15:44:50  <bnoordhuis>hmm, don't know - no immediate, plausible cause comes to mind
15:46:55  * indutnyjoined
15:49:36  <bnoordhuis>piscisaureus_: pro tip: compile gdb 6.x and put a breakpoint in the code that returns ENOTFOUND
15:49:53  <bnoordhuis>6.x because 7.x doesn't compile on sunos
16:06:13  <indutny>bnoordhuis: can process.stdin have write watcher? https://github.com/joyent/node/blob/v0.4/lib/tty_posix.js#L59
16:06:19  <indutny>as it was having in 0.4.12
16:06:37  <bnoordhuis>indutny: write watchers in node have gone the way of the dodo
16:06:43  <indutny>I'm not very deep in new libuv architecture
16:06:47  <bnoordhuis>i.e. tasty but extinct
16:06:48  <indutny>ah
16:06:54  <indutny>bnoordhuis: ok
16:18:20  <indutny>bnoordhuis: basically, what can hold libuv's event loop from exiting?
16:18:26  <indutny>some non-zero conters?
16:18:29  <indutny>s/conters/counters
16:19:12  * CoverSlidejoined
16:19:50  <bnoordhuis>indutny: no, active handles that are not uv_unref'd
16:25:52  <piscisaureus_>I walso wonder why cares calls ioctl tcgetattr on the hosts file fd
16:28:29  <piscisaureus_>^./1: open64("/etc/inet/hosts", O_RDONLY) = 6
16:28:30  <piscisaureus_>^./1: fstat64(6, 0x080470B0) = 0
16:28:30  <piscisaureus_>^./1: fstat64(6, 0x08046FC0) = 0
16:28:30  <piscisaureus_>^./1: ioctl(6, TCGETA, 0x08047060) Err#25 ENOTTY
16:28:30  <piscisaureus_>^./1: read(6, " # i d e n t\t " @ ( #".., 512) = 271
16:28:30  <piscisaureus_>^./1: read(6, 0x08543C9C, 512) = 0
16:28:30  <piscisaureus_>^./1: llseek(6, 0, SEEK_CUR) = 271
16:28:31  <piscisaureus_>^./1: close(6) = 0
16:31:34  * luxigojoined
16:32:59  * AvianFlujoined
16:33:23  <indutny>bnoordhuis: ok, I think I fixed that
16:33:24  <indutny>:D
16:33:29  <indutny>not sure if I broke anything
16:33:32  <indutny>hehe
16:33:59  <indutny>https://github.com/joyent/node/blob/master/src/tty_wrap.cc#L171
16:34:12  <indutny>just insert uv_unref(uv_default_loop()) right after this line
16:34:23  <indutny>is it correct?
16:36:38  <indutny>bnoordhuis: ?
16:36:52  <indutny>or should I do this somewhere on libuv level?
16:41:29  * kuebkpart
16:49:04  <indutny>bnoordhuis: btw, is that bug reproducable on windows?
17:03:50  <indutny>bnoordhuis: anyway, https://github.com/joyent/libuv/pull/251
17:11:07  <luxigo>how could I enable verbose build (with gcc commands echoed)... ?
17:13:08  <luxigo>my node internal module build fails now with "error: exception handling disabled, use -fexceptions to enable" ...
17:13:41  <luxigo>maybe adding -fexceptions could work but cannot try since build commands are not echoed to terminal :-D
17:14:24  <luxigo> I dont know where to add it in wscript or other makefiles..
17:16:43  <luxigo>on the other hand i can still revert the folder to a previous git state but if you can help while i try to fix it... :-)
17:19:42  * dapjoined
17:23:48  * isaacsjoined
17:26:08  <luxigo>as I workaround i will just comment out the "throw"... but if you can tell me how to do it the nodejs way ...
17:26:54  <luxigo>(in c++ code)
17:30:18  <luxigo>maybe http://groups.google.com/group/nodejs/browse_thread/thread/d444acb2c231ad45 answers
17:31:23  * brsonjoined
17:35:30  * kuebk^joined
17:43:46  * AvianFluquit (Quit: Leaving)
18:03:34  * igorzi_quit (Quit: Page closed)
18:03:50  * igorzijoined
18:10:36  <igorzi>https://github.com/joyent/node/issues/2128
18:10:59  <igorzi>piscisaureus_: ^--- what do you think?
18:15:55  <piscisaureus_>nah I don't know
18:16:04  <piscisaureus_>fixing it in the console functions seems to be fine
18:16:13  <piscisaureus_>but I am afraid of the slippery slope
18:21:50  <igorzi>piscisaureus_: seems like it would also need to go into util.js
18:43:38  * AvianFlujoined
18:49:02  <isaacs>piscisaureus_, igorzi: if you're adding \n, then it ought to add \r\n on windows, i think
18:49:13  <isaacs>console.* is already a convenience thing
18:49:41  <isaacs>what we must not do is magically translate \n to \r\n whenever writing to stdout
18:50:41  <piscisaureus_>I agree I think
18:50:53  <igorzi>isaacs: yep agreed
18:51:53  <igorzi>what about functions in util.js?
18:52:30  <isaacs>like util.inspect you mean?
18:53:13  <igorzi>util.puts, util.debug
18:53:21  <isaacs>ah, yeah
18:53:27  <isaacs>yeah, they should do the same thing
18:53:38  <isaacs>just use some "eol" var instead of \n
18:53:42  <isaacs>and set that to \r\n on windows
18:53:55  <isaacs>i'm already doing this on npm, so that the debug log is readable
18:55:00  <igorzi>it'd probably also be useful to export the "eol" var
18:57:44  <igorzi>call today?
18:58:00  <igorzi>ryah piscisaureus_ bnoordhuis: --^
18:59:30  <piscisaureus_>yeah why not
19:02:13  <ryah>yes
19:03:09  <piscisaureus_>wait 5 minutes if you want to make me happy
19:04:08  <ryah>bnoordhuis: call?
19:05:17  <isaacs>skype or phone?
19:19:24  <luxigo>what happened with node_native modules ?... adding modules in node/lib and using _third_party_main.js is not working anymore..
19:19:58  * mralephjoined
19:27:23  <ryah>igorzi, piscisaureus_
19:27:26  <ryah>ryan@mac1234:~/projects/node% time node benchmark/fast_buffer2_creation.js
19:27:26  <ryah>node benchmark/fast_buffer2_creation.js 0.29s user 0.02s system 96% cpu 0.320 total
19:27:29  <ryah>ryan@mac1234:~/projects/node% time node benchmark/fast_buffer_creation.js
19:27:32  <ryah>node benchmark/fast_buffer_creation.js 0.46s user 0.02s system 81% cpu 0.592 total
19:27:34  <ryah>ryan@mac1234:~/projects/node% time node benchmark/buffer_creation.js
19:27:37  <ryah>node benchmark/buffer_creation.js 2.11s user 0.21s system 94% cpu 2.455 total
19:52:13  <piscisaureus_>igorzi: you're obviously free to ignore that last request from me :-)
19:56:27  <indutny>what's due for 0.6.2?
20:08:00  <ryah>indutny: bug fixes
20:08:08  <piscisaureus_>fix bugs
20:16:30  * mikealquit (Quit: Leaving.)
20:19:40  * isaacsquit (Quit: isaacs)
20:20:17  <bnoordhuis>back
20:20:39  <bnoordhuis>indutny: does #251 pass all tests in libuv and node?
20:22:51  <ryah>igorzi, piscisaureus_ : https://gist.github.com/1371250
20:23:00  <ryah>no difference
20:25:17  <bnoordhuis>piscisaureus_: i doubt it's c-ares doing the TCGETA ioctl
20:28:25  <igorzi>ryah piscisaureus_: i'll play with splitting the 1mb buffer into n sub-buffers (as we discussed)
20:31:34  <CIA-95>node: Łukasz Walukiewicz master * r5d0b5a0 / (lib/buffer.js test/simple/test-writeint.js): buffer: fix minimum values for writeInt*() functions - http://git.io/ScE90Q
20:31:58  <CIA-95>node: Łukasz Walukiewicz v0.6 * r3b852d7 / (lib/buffer.js test/simple/test-writeint.js): buffer: fix minimum values for writeInt*() functions - http://git.io/hIskbg
20:32:34  * isaacsjoined
20:33:17  <bnoordhuis>piscisaureus_: https://github.com/joyent/node/issues/2036
20:38:44  * piscisaureus_quit (Ping timeout: 244 seconds)
20:41:35  * paddybyersquit (Quit: paddybyers)
20:45:37  * isaacsquit (Quit: isaacs)
20:53:30  <ryah>bnoordhuis: no need to land in both branches - just land in v0.6
20:53:33  <ryah>i'll merge back
20:53:50  <bnoordhuis>ryah: oh right, of course
20:54:04  <ryah>bnoordhuis: im close on https://github.com/joyent/libuv/issues/206
20:54:35  <bnoordhuis>ryah: what's causing it?
20:54:36  <ryah>https://github.com/joyent/libuv/blob/4ae40b6a3ffcfcfe8b800d0cad216b04cc4d4b6c/src/unix/stream.c#L493
20:55:00  <ryah>https://github.com/joyent/libuv/blob/4ae40b6a3ffcfcfe8b800d0cad216b04cc4d4b6c/src/unix/stream.c#L411
20:55:17  <bnoordhuis>ah, lemme guess... req->error is actually an errno?
20:55:21  <ryah>yeah
20:56:58  <bnoordhuis>good catch
21:02:50  <CIA-95>libuv: Ryan Dahl v0.6 * r2007eb8 / src/unix/stream.c :
21:02:50  <CIA-95>libuv: Set uv_write_t errno correctly
21:02:50  <CIA-95>libuv: Fixes #206. - http://git.io/NXix2A
21:06:34  <CIA-95>node: Ryan Dahl v0.6 * r44314cc / (10 files in 5 dirs): Upgrade libuv to 2007eb8 - http://git.io/GNmo3Q
21:08:02  <ryah>it would be nice to have a test for this...
21:08:05  <ryah>but .. hm
21:21:37  * indexzerojoined
21:25:56  * creationixquit (Remote host closed the connection)
21:32:45  * indexzeroquit (Quit: indexzero)
21:34:23  * mikealjoined
21:35:06  * paddybyersjoined
21:36:53  * isaacsjoined
21:38:30  <CIA-95>node: Łukasz Walukiewicz master * r3b852d7 / (lib/buffer.js test/simple/test-writeint.js): buffer: fix minimum values for writeInt*() functions - http://git.io/hIskbg
21:38:30  <CIA-95>node: Ryan Dahl master * r44314cc / (10 files in 5 dirs): Upgrade libuv to 2007eb8 - http://git.io/GNmo3Q
21:38:31  <CIA-95>node: Ryan Dahl master * r93dc713 / (14 files in 8 dirs): Merge remote branch 'origin/v0.6' - http://git.io/MJsmug
21:38:32  <CIA-95>node: Ryan Dahl master * r40cb398 / src/node_version.h : Working on v0.7.0 - http://git.io/RQHdTA
21:39:03  <ryah>https://groups.google.com/d/topic/nodejs/eVBOYiI_O_A/discussion
21:39:08  <ryah>^-- v0.8 roadmap
21:41:07  <mraleph>isolates, domains, no waf \o/
21:41:25  <bnoordhuis>ryah: i'm not seeing the libuv errno bug fix in node master or 0.6
21:44:19  <ryah>bnoordhuis: uh - hm
21:45:16  <ryah>bnoordhuis: it's there
21:45:16  <ryah>https://github.com/joyent/node/blob/40cb398867b1aae47aa4add04eb5cf7e8bfc131e/deps/uv/src/unix/stream.c#L148
21:45:37  * benviejoined
21:52:47  <bnoordhuis>ah sorry, i missed it with all that dragonflybsd stuff
21:54:04  <ryah>bnoordhuis: did the dragonflyBSD patch look okay to you?
21:54:13  <ryah>seems pretty minimal to me
21:54:18  <bnoordhuis>ryah: yes
21:54:34  <bnoordhuis>dragonfly is a fork of freebsd, they're pretty similar
22:08:03  * piscisaureus_joined
22:09:22  * AvianFluquit (Ping timeout: 255 seconds)
22:10:59  * AvianFlujoined
22:21:28  <piscisaureus_>igorzi: does `vcbuild nobuild` give you an error too
22:21:29  <piscisaureus_>?
22:21:41  <piscisaureus_>(on v0.6)
22:25:07  <igorzi>piscisaureus_: no.. what error do you get?
22:25:49  <piscisaureus_>igorzi:
22:25:50  <piscisaureus_>D:\node3>vcbuild.bat nobuild
22:25:51  <piscisaureus_>The system cannot find the batch label specified - arg-ok
22:25:58  <piscisaureus_>igorzi: the label is there though
22:26:07  <piscisaureus_>It works if I insert two blank lines before it
22:26:12  <piscisaureus_>otherwise, no luck
22:27:05  * creationixjoined
22:29:20  * indexzerojoined
22:29:56  <igorzi>piscisaureus_: hmm.. the only recent change was made by ryah to add upload
22:33:15  <piscisaureus_>igorzi: wut. If I remove that line it also works again
22:33:22  <piscisaureus_>I put it back, it breaks :-/
22:35:07  * indexzeroquit (Quit: indexzero)
22:35:21  <igorzi>piscisaureus_: wtf? :) maybe it has \n instead of \r\n as the eol?
22:35:54  <piscisaureus_>it has \n all over the place
22:36:16  <piscisaureus_>but no \r anywhere
22:36:26  <piscisaureus_>also `vcbuild test` just works
22:46:45  <igorzi>piscisaureus_: is your git setup to replace \n with \r\n?
22:47:38  <piscisaureus_>nope
22:47:41  <piscisaureus_>should it be?
22:47:49  <piscisaureus_>Me no gusto
22:48:27  <igorzi>piscisaureus_: mine is (that was the default).. maybe that's why i don't have this issue
22:49:13  <piscisaureus_>I think I will just land the 2 empty lines and be done with it
22:49:48  <igorzi>piscisaureus_: wait.. now i'm able to reproduce if i just download the raw copy of vcbuild.bat (without git)
22:54:49  * indexzerojoined
22:56:09  <ryah>question - on the MSI XML file do we need to list all files that will be installed?
22:57:50  <ryah>did i fuck up the vcbuild.bat ?
22:57:58  <piscisaureus_>ryah: yes and no
22:58:06  <piscisaureus_>ryah: you did not do anything wrong
23:00:10  <ryah>ok
23:00:44  <ryah>so isaacs, ideally we'd add the npm SHA1 to vcbuild.bat msi
23:00:45  * Marakquit (Disconnected by services)
23:01:01  <isaacs>ryah: the sha1 of what exactly?
23:01:11  <ryah>isaacs: the commit you want released
23:01:17  <isaacs>ryah: what about a tag? would that work?
23:01:23  <isaacs>are you just passing it as an argument to git checkout?
23:01:29  <ryah>yeah
23:01:35  <isaacs>k
23:01:37  * creationix_joined
23:01:47  <isaacs>the problem there is that we'll also have to build the docs, which means installing ronnjs
23:02:13  <CIA-95>node: Ben Noordhuis v0.6 * r5e3b009 / (lib/tls.js test/simple/test-tls-set-ciphers.js):
23:02:14  <CIA-95>node: tls: make cipher list configurable
23:02:14  <CIA-95>node: options.ciphers existed but didn't work, the cipher list was effectively
23:02:14  <CIA-95>node: hard-coded to RC4-SHA:AES128-SHA:AES256-SHA.
23:02:14  <CIA-95>node: Fixes #2066. - http://git.io/u8AhZw
23:02:16  <ryah>hmm - yeah that's kind of annoying
23:02:29  <ryah>you were suggesting downloading the tarball from the website?
23:04:21  <igorzi>piscisaureus_: pls go ahead with 2 empty lines
23:04:47  <piscisaureus_>igorzi: random weirdness, eh? :-)
23:05:20  <igorzi>piscisaureus_: i don't see it because git converted \n to \r\n for me
23:05:44  <piscisaureus_>igorzi: yeah but generally \n is n
23:05:50  <piscisaureus_>ot a problem for cmd.exe
23:06:29  <igorzi>piscisaureus_: right.. i guess we found a case when it is :)
23:06:41  <CIA-95>node: unknown v0.6 * r7acb92d / vcbuild.bat : Fix strange vcbuild "label not found" error - http://git.io/9kyDAQ
23:06:49  * piscisaureus_sighs
23:07:00  <igorzi>ryah: yes all files that you want included with the install need to be listed in wix
23:07:14  <ryah>eh - someone needs to set their git email ...
23:07:22  <ryah>https://github.com/joyent/node/commit/7acb92dee1584ab2017177b2a6ea3ab461cf3308 <--
23:07:52  <ryah>igorzi: is it possible ot just give it a directory?
23:08:01  * kuebk^quit
23:08:14  <ryah>igorzi: and tell it to install all sub-files?
23:08:42  <piscisaureus_>ok, force pushing
23:08:46  <piscisaureus_>scheisse
23:08:47  <ryah>igorzi: the problem is that npm has a huge number of files
23:08:53  <piscisaureus_>forgot to do it after git reinstall
23:09:23  <ryah>piscisaureus_: please only force push to v0.6 and make sure 5e3b009 is the parent of your commit
23:09:26  <isaacs>ryah: yeah, i was suggesting downloading the tarball, or i can even change up the npm build to put a zip somewhere.
23:09:53  <ryah>isaacs: tar is fine
23:09:58  <isaacs>k
23:10:20  <ryah>what else is built in the tarball besides docs?
23:10:25  <isaacs>the install.sh script reads the tarball location from https://registry.npmjs.org/npm/latest
23:10:48  <isaacs>also all the dependencies are bundled.
23:10:59  <CIA-95>node: Bert Belder v0.6 * r9c70b09 / vcbuild.bat : Fix strange vcbuild "label not found" error - http://git.io/18FnQA
23:10:59  <ryah>ah, ok
23:11:02  <isaacs>so you don't have to do git submodule update --init --recursive
23:11:33  <ryah>isaacs: and we're going to want to run the tests - is that going to be possible with the tarball?
23:11:44  <isaacs>ryah: sure.
23:11:51  <ryah>ok
23:11:51  <isaacs>ryah: um... on windows?
23:11:55  <ryah>yeah
23:11:56  <isaacs>no
23:12:01  <isaacs>the test runner is a bash script
23:12:06  <ryah>oh.. hm
23:12:08  <ryah>we need to fix that
23:12:13  <isaacs>yeah. on the todo list :)
23:12:35  <igorzi>ryah: i'll find out if you can point it to a directory
23:12:35  <piscisaureus_>many tests still rely on msys tools I think
23:12:42  <piscisaureus_>like curl etc
23:12:52  <ryah>yes
23:13:10  <ryah>isaacs: does the bash runner run on windows though?
23:13:15  * mralephquit (Quit: Leaving.)
23:13:16  <ryah>isaacs: windows with bash, that is
23:13:41  <ryah>that would be acceptable
23:13:44  <ryah>i have bash on my windows
23:13:56  <isaacs>ryah: i'm not sure. there's some oddness with npm using git/msys bash, because the .cmd files don't run
23:14:08  <ryah>hm yeah
23:14:36  <isaacs>afaict, there's no clear way to write a program that works with cmd.exe and also works with msys bash without compiling it.
23:15:13  <ryah>what's the plan for the test runner?
23:17:45  <ryah>isaacs: have you seen felix's new test runner?
23:17:50  <ryah>isaacs: it looks pretty nice
23:17:53  <isaacs>yeah, it's ok
23:17:58  <isaacs>i'll probably use tap.
23:18:06  <isaacs>since i don't have to get anyone's permission to change it when i hit bugs.
23:18:21  <ryah>yeah
23:18:27  <isaacs>test runners and flow control libs should not be black boxes.
23:18:39  <isaacs>i'm remarkably surprised at how many people use slide and tap. they're not even that great.
23:18:48  <igorzi>ryah: yes, it's possible to include directories
23:18:56  <ryah>igorzi: great
23:19:38  <ryah>so isaacs, this is getting installed at C:\Program Files\nodejs\node.exe and C:\Program Files\nodejs\npm.cmd ?
23:19:48  <isaacs>sounds right
23:19:54  <ryah>all the other npm files should be tucked away in C:\Program Files\nodejs\node_modules or something ?
23:19:59  <isaacs>bingo.
23:20:04  <ryah>cool
23:20:11  <isaacs>you guessed correctly, if that was a guess :)
23:20:48  <ryah>so we're going to download the tarball, unpack it into the node project root directory, then run the WiX scripts
23:20:54  <piscisaureus_>I wonder if it will end up in c:\program files (x86)
23:21:10  <igorzi>yeah, on x64 windows it goes into program files (x86)
23:21:11  <ryah>which will take the npm directory and put it in the MSI to be installed into C:\Program Files\nodejs\node_modules\npm
23:21:27  <isaacs>WiX scripts?
23:21:45  <ryah>the msi creating stuff
23:21:49  <isaacs>oh, ok
23:21:59  <isaacs>i can provide the npm.cmd file pre-built so that you don't have to generate that.
23:22:10  <ryah>isaacs: https://github.com/joyent/node/blob/3222a04d5e8d1fef90ac6247c6febeebe02cf659/vcbuild.bat#L78-85
23:22:22  <ryah>how big is the npm.cmd?
23:22:58  <isaacs>213 bytes.
23:23:03  <isaacs>oh, there's also npm-g.cmd and npm_g.cmd
23:23:04  <ryah>we could just include it in node's tree if you want
23:23:17  <isaacs>639 bytes total ;)
23:23:35  <ryah>in tools/msi/npm.cmd
23:23:39  <ryah>or whatever
23:24:02  <ryah>or - i guess it's better if it's just in the npm direcotry
23:24:27  <isaacs>yeah, i'll just put it there, the msi script can copy it out
23:24:48  <ryah>ok
23:25:00  <ryah>so we won't do this for v0.6.2 but we'll try for v0.6.3
23:25:12  <ryah>hopefully bsdtar.exe will be removed by then
23:25:39  <ryah>and then windows users will be happy
23:26:08  * creationixquit (Quit: Leaving.)
23:26:08  * creationix_changed nick to creationix
23:26:38  <ryah>isaacs: i want to specify exactly the version of npm we will install in the node tree - not the latest
23:26:46  <isaacs>ryah: sure.
23:27:27  <ryah>how big is the npm tarball?
23:27:32  <isaacs>ryah: https://github.com/isaacs/npm/commit/80c7f1eba41e05825f4c941f4361192a0619822e
23:27:33  <ryah>we could just drop it in deps
23:27:37  <igorzi>ryah isaacs: so program files\nodejs\node_modules will be the global node_modules folder?
23:27:59  <isaacs>Content-Length: 1187840
23:28:02  <isaacs>1M
23:28:22  <isaacs>igorzi: yes.
23:28:31  <isaacs>and program files\nodejs\ will be the global bin folder.
23:29:06  <ryah>yeah maybe we'll just drop the tarball into node
23:29:06  <igorzi>isaacs: this means that modules will be downloaded to program files\nodejs\node_modules when --global switch is used, correct?
23:29:18  <ryah>i dont like the idea of having to go over the internet
23:30:24  * creationixquit (Quit: creationix)
23:34:28  <igorzi>ryah isaacs: this means that users will need to run elevated to do "npm install --global"
23:34:53  <ryah>if we can include a directory from WiX then it seems quite easy to include npm
23:35:01  <ryah>igorzi: that seems reasonable
23:35:08  <ryah>igorzi: does it prompt you for a password?
23:35:18  <ryah>or how do you run elevated commands?
23:35:24  <igorzi>ryah isaacs: yep. that's actually a good thing
23:35:47  <igorzi>ryah: you'll get access denied when trying to write to program files
23:36:11  <isaacs>igorzi: that's pretty similar to the normal unix install mechanism, actually.
23:36:18  <igorzi>ryah: you can start cmd.exe elevated with right-click, run as admin
23:36:29  <isaacs>how hard would it be to have the msi let you choose an install location?
23:36:44  <ryah>isaacs: i think bert is adding that
23:36:54  <igorzi>ryah isaacs: what about when people uninstall node.. is it reasonable to wipe the global node_modules?
23:37:05  <ryah>igorzi: yes, i think so
23:37:44  <igorzi>what if someone just wants to upgrade from 0.6.2 to 0.6.3.. all their global modules will be gone
23:38:06  <igorzi>i guess they could just install those modules non-global
23:38:44  <ryah>igorzi: yeah that's a little annoying
23:38:51  <ryah>what's the process for upgrading an MSI?
23:39:06  <ryah>uninstall / install?
23:39:17  <ryah>or is there special upgrade process that can be defined?
23:39:28  <igorzi>yes, if side-by-side installs are not supported (which is how it is now)
23:40:24  <igorzi>i believe there's also a way to do upgrades
23:42:50  <piscisaureus_>ryah: should ObjectWrap be exported for use by modules
23:42:51  <piscisaureus_>?
23:43:01  <piscisaureus_>so modules can define their own
23:43:37  * paddybyersquit (Quit: paddybyers)
23:46:18  * AvianFlu_joined
23:47:46  * AvianFluquit (Disconnected by services)
23:47:49  * AvianFlu_changed nick to AvianFlu
23:51:07  <ryah>piscisaureus_: yes
23:51:17  <ryah>many people are using it
23:53:18  * bnoordhuisincluded
23:56:40  <CIA-95>node: Bert Belder reviewme * r3e696d1 / (src/node_buffer.h src/node_object_wrap.h):
23:56:40  <CIA-95>node: Windows: make Buffer and ObjectWrap available to compiled extensions
23:56:40  <CIA-95>node: Closes GH-2036 - http://git.io/9LO2AQ
23:56:46  <piscisaureus_>igorzi: ^- signoff?