00:01:32  * isaacsjoined
00:03:06  * creationixjoined
00:23:10  * mikealjoined
00:24:08  * mikealquit (Client Quit)
00:26:03  * mikealjoined
00:30:28  * isaacs_joined
00:30:28  * isaacsquit (Read error: Connection reset by peer)
00:30:28  * isaacs_changed nick to isaacs
00:40:39  <CIA-95>libuv: Ben Noordhuis master * r1e0aab0 / (src/unix/kqueue.c test/test-fs-event.c test/test-list.h):
00:40:40  <CIA-95>libuv: unix: properly disarm kqueue fs watcher
00:40:40  <CIA-95>libuv: Fixes "Assertion failed: (revents == EV_LIBUV_KQUEUE_HACK),
00:40:40  <CIA-95>libuv: function uv__fs_event, file ../src/unix/kqueue.c, line 58." - http://git.io/g-drKA
00:40:42  <CIA-95>libuv: Ben Noordhuis v0.6 * r2bd181a / (src/unix/kqueue.c test/test-fs-event.c test/test-list.h):
00:40:42  <CIA-95>libuv: unix: properly disarm kqueue fs watcher
00:40:42  <CIA-95>libuv: Fixes "Assertion failed: (revents == EV_LIBUV_KQUEUE_HACK),
00:40:42  <CIA-95>libuv: function uv__fs_event, file ../src/unix/kqueue.c, line 58." - http://git.io/yxqC3g
00:51:13  <isaacs>bnoordhuis, ryah: you around?
00:51:28  <isaacs>it seems like ./configure --debug exits 0 after printing a trace and saying it's got a ValueError
00:51:31  <isaacs>does that sound right?
00:52:00  <isaacs>https://gist.github.com/1375150
00:56:24  <bnoordhuis>isaacs: works for me
00:56:35  <isaacs>bnoordhuis: that's weird.
00:56:51  <bnoordhuis>oh, you're using the xcode generator
00:56:54  <bnoordhuis>let me try that
00:57:04  <isaacs>i don't know about xcode
00:57:13  <isaacs>i have xcode on my system, but i mean, i didn't choose that explicitly or anythihg
00:58:16  <isaacs>it looks from the error like it's finding multiple build files for libz?
00:59:33  <bnoordhuis>xcode generator wfm
00:59:46  <bnoordhuis>isaacs: try running gyp with -fmake
01:00:07  <isaacs>bnoordhuis: $ make
01:00:07  <isaacs>tools/gyp_node -f make
01:00:07  <isaacs>['-f', 'make', '/Users/isaacs/dev-src/js/node/node.gyp', '-I', '/Users/isaacs/dev-src/js/node/common.gypi', '-I', '/Users/isaacs/dev-src/js/node/options.gypi', '--depth=.', '--generator-output', '/Users/isaacs/dev-src/js/node/out', '-Goutput_dir=/Users/isaacs/dev-src/js/node/out', '-Dtarget_arch=ia32', '-Dcomponent=static_library', '-Dlibrary=static_library']
01:00:07  <isaacs>make -C out BUILDTYPE=Release
01:00:08  <isaacs> CC(target) /Users/isaacs/dev-src/js/node/out/Release/obj.target/http_parser/deps/http_parser/http_parser.o
01:00:08  <isaacs> LIBTOOL-STATIC /Users/isaacs/dev-src/js/node/out/Release/libhttp_parser.a
01:00:08  <isaacs> CC(target) /Users/isaacs/dev-src/js/node/out/Release/obj.target/uv/deps/uv/src/uv-common.o
01:00:09  <isaacs>that's running now
01:00:38  <bnoordhuis>okay, so the make generator works?
01:00:41  <isaacs>it's in v8_base
01:00:55  <isaacs>well... i don't know
01:01:04  <isaacs>i mean, configure said it failed,but exited with 0 status
01:01:18  <isaacs>and "make" is doing stuff
01:01:28  <isaacs>whether it worked or is hideously broken, i don't know
01:01:29  * piscisaureus__joined
01:01:34  <bnoordhuis>we spawn another process that runs tools/gyp_node
01:01:36  <isaacs>will find out soon.
01:01:47  <bnoordhuis>so configure always exits with status 0
01:02:24  <isaacs>yeah
01:02:25  <bnoordhuis>try checking out 4f8d6d0^ and running configure again
01:02:31  <isaacs>also, there's in "install" make target.
01:02:41  <isaacs>that is deeply problematic.
01:02:49  <bnoordhuis>we'll fix it :)
01:02:53  <isaacs>i know :)
01:03:08  <isaacs>still building v8
01:03:21  <isaacs>can't wait for my new faster laptop, coming this week :)
01:03:32  <bnoordhuis>heh :)
01:03:39  <bnoordhuis>okay, off to bed - sleep tight, isaacs
01:03:46  <isaacs>have fun
01:04:16  * piscisaureus_quit (Ping timeout: 248 seconds)
01:05:17  * mralephquit (Quit: Leaving.)
01:06:01  <ryah>isaacs: have a look in the configure file and see if you can debug it
01:06:14  <ryah>gyp generates makefiles, msvs files, or xcode files
01:06:19  <isaacs>k
01:06:25  <isaacs>debug the missing install thing?
01:06:25  <ryah>but we've set it up that when you type "configure; make"
01:06:30  <ryah>it only generates make file
01:06:59  <ryah>no the python stack trace thing you reported
01:07:17  <isaacs>ok
01:07:45  <ryah>it seems your thing is generating xcode for some reason
01:08:01  <isaacs>ryah: how would it know that "make" is the next command...?
01:08:09  <isaacs>ryah: i did ./configure && make
01:08:18  <isaacs>or you could type `./configure \n make`
01:08:42  <ryah>oh i see - on the last line of the configure script
01:08:48  <ryah>change it from
01:08:49  <ryah>subprocess.call('tools/gyp_node')
01:08:50  <ryah>to
01:08:54  <ryah>subprocess.call('tools/gyp_node -f make')
01:09:16  <ryah>and possibly remove the calls to "tools/gype_node" from the Makefile
01:12:08  <isaacs>so, make worked
01:12:18  <isaacs>it looks like it's just throwing an error, swallowing it, and continuing on.
01:13:34  <ryah>isaacs: https://gist.github.com/1375212
01:13:43  <ryah>isaacs: try 'make distclean' with this
01:14:18  <ryah>er - rebuild it with this patch
01:14:23  <ryah>i think it should fix your initial error
01:15:16  <isaacs>no stack trace there
01:16:33  <ryah>ok
01:16:38  <isaacs>make is making
01:16:47  <isaacs>but doesn't that mean that it won't work for non-make systems?
01:16:56  * mikealquit (Quit: Leaving.)
01:16:59  <isaacs>windows/xcode/etc?
01:17:02  <ryah>there are no non-make systems
01:17:08  <isaacs>ha
01:17:12  <isaacs>there's windows
01:17:17  <ryah>they use vcbuild.bat
01:17:22  <isaacs>oh, right
01:17:23  <isaacs>k
01:24:00  * mikealjoined
01:25:44  * mikealquit (Client Quit)
01:27:44  <isaacs>ryah: you wanna land https://raw.github.com/gist/1375212/91d0802ef5a1114b96451f3635676cc3090f01ca/gyp.diff then?
01:27:46  <isaacs>lgtm
01:33:20  * bnoordhuisquit (Ping timeout: 260 seconds)
01:36:28  <ryah>igorzi: where do i get signcode.exe from?
01:36:43  <CIA-95>node: Ryan Dahl master * r48d21dd / (Makefile configure): Force makefile generation in gyp - http://git.io/pi3xHA
01:37:03  <ryah>oh nevermind - i guess it's SignTool.exe
01:37:49  <igorzi>ryah: did you find where to get signtool.exe?
01:38:26  * dapquit (Quit: Leaving.)
01:38:47  * dapjoined
01:38:53  <ryah>igorzi: yep
01:39:16  <ryah>i got the joyent certs finally
01:40:13  <igorzi>nice
01:41:23  <ryah>igorzi: does the .exe and .msi need to be signed or just .msi?
01:42:14  <igorzi>ryah: both
01:42:21  <ryah>k
01:52:35  <ryah>igorzi: can you test this http://nodejs.org/dist/v0.6.2/node-v0.6.2.msi
01:53:00  <igorzi>one sec
01:53:33  <ryah>igorzi: it should be signed
01:55:48  <igorzi>ryah: installed fine.. i can see that msi is signed
01:57:31  <ryah>great
01:57:49  <isaacs>bam. i can install stuff using the js tar.
01:57:55  <isaacs>(not locally, though, that breaks.
01:58:15  <isaacs>but installs from the registry are working.
01:58:42  * isaacsquit (Quit: isaacs)
02:14:49  * creationixquit (Quit: creationix)
02:15:38  * isaacsjoined
02:26:56  * isaacsquit (Quit: isaacs)
02:30:28  * isaacsjoined
02:39:01  * igorziquit (Ping timeout: 265 seconds)
02:40:22  * indexzerojoined
02:56:00  * indexzeroquit (Quit: indexzero)
02:59:24  * brsonquit (Quit: leaving)
03:03:02  * isaacsquit (Quit: isaacs)
03:35:45  * indexzerojoined
03:42:20  * isaacsjoined
03:45:58  * AvianFluquit (Ping timeout: 255 seconds)
03:53:29  * AvianFlujoined
03:55:02  * isaacsquit (Quit: isaacs)
04:53:51  * isaacsjoined
04:54:35  * isaacsquit (Client Quit)
05:15:48  * igorzijoined
05:24:25  * mikealjoined
05:43:04  * indexzeroquit (Quit: indexzero)
05:43:33  * indexzerojoined
06:02:49  * dapquit (Quit: Leaving.)
06:04:46  * isaacsjoined
06:21:25  * mikealquit (Quit: Leaving.)
06:41:15  * mikealjoined
06:58:42  * paddybyersjoined
07:03:22  * mralephjoined
07:29:04  * paddybyersquit (Quit: paddybyers)
08:19:37  * kuebkjoined
08:23:18  <indutny>godd evening
08:25:45  * paddybyersjoined
08:25:57  * isaacsquit (Quit: isaacs)
08:29:21  * piscisaureus____joined
08:29:26  * piscisaureus____changed nick to piscisaureus_
08:31:25  * paddybyers_joined
08:32:42  * piscisaureus__quit (Ping timeout: 252 seconds)
08:33:52  * paddybyersquit (Ping timeout: 240 seconds)
08:33:52  * paddybyers_changed nick to paddybyers
08:51:17  * piscisaureus_quit (Ping timeout: 245 seconds)
09:03:01  * piscisaureus_joined
09:47:37  * paddybyersquit (Quit: paddybyers)
09:52:04  * indexzeroquit (Quit: indexzero)
09:52:43  * mralephquit (Quit: Leaving.)
09:54:39  * indexzerojoined
09:57:01  * indexzeroquit (Client Quit)
10:02:55  * bnoordhuisjoined
10:15:25  <CIA-95>node: Ben Noordhuis v0.6 * r59a9a9b / (9 files in 2 dirs):
10:15:25  <CIA-95>node: buffer: add .read*() and .write*() methods to SlowBuffer prototype
10:15:25  <CIA-95>node: Fixes #2138. - http://git.io/1bFs8w
10:20:34  <bnoordhuis>piscisaureus_: ping
10:46:51  <piscisaureus_>bnoordhuis: pong
10:47:04  <piscisaureus_>bnoordhuis: sorry, have no notifications set up yet in my irc client. will do now
10:47:08  <bnoordhuis>piscisaureus_: tell me about srwlocks
10:47:17  <piscisaureus_>ok
10:47:35  <piscisaureus_>bnoordhuis: I think we should just implement the try functions ourselves
10:47:42  <piscisaureus_>or entirely
10:47:50  <bnoordhuis>for srwlocks? how?
10:48:01  <piscisaureus_>I looked at how windows implemented them and they are just spinlocks
10:48:12  <bnoordhuis>aargh, i don't want to write assembler
10:48:19  <piscisaureus_>You don't need to
10:48:33  <piscisaureus_>It can be done in c
10:48:53  <piscisaureus_>bnoordhuis: https://gist.github.com/1367367
10:49:14  <piscisaureus_>try lock exclusive is just one CAS operation
10:49:26  <bnoordhuis>lock bts dword ptr [eax],0 <- is that InterlockedSomething()?
10:49:51  <piscisaureus_>yep
10:50:22  <piscisaureus_>bnoordhuis:
10:50:37  <bnoordhuis>piscisaureus_: but that means you have to write the a full srwlock implementation
10:50:44  <piscisaureus_>https://gist.github.com/1376139
10:50:52  <piscisaureus_>it's easy
10:51:30  <bnoordhuis>i'm mildly sceptical about that
10:51:52  <bnoordhuis>but okay
10:51:55  <piscisaureus_>bnoordhuis: but for now we can just implement TryAcquireSRWLockExclusive and TryAcquireSRWLockShared for vista and server 2k
10:52:29  <piscisaureus_>*2k8
10:52:48  <bnoordhuis>how are AcquireSRWLockShared, ReleaseSRWLockShared and friends implemented?
10:52:48  <piscisaureus_>As long as they're compatible with vista's own AcquireSRWLockExclusive/Shared
10:53:03  <bnoordhuis>i mean, what makes them block?
10:53:07  <piscisaureus_>let's check
10:53:12  <piscisaureus_>I think they are spin locks
10:53:17  <piscisaureus_>Actually, I am sure :-)
10:53:30  <bnoordhuis>right
10:54:01  <bnoordhuis>in that case we should implement all of the srwlock API for consistency's sake
10:54:09  <piscisaureus_>heh
10:54:13  <bnoordhuis>would be nice to have one API for all the windowses
10:54:51  <piscisaureus_>well implementing AcquireSRWLockExclusive is easy: while (!TryAcquireSRWLockExclusive(...)) BackOff();
10:55:11  <piscisaureus_>the only thing we need to add is Initialize which is just initializing some value to 0
10:55:21  <piscisaureus_>and the Release functions
10:55:27  <bnoordhuis>i can do that, i've initialized things to zero before
10:56:20  <piscisaureus_>bnoordhuis: for completeness sake, let me disassemble all of that shit
10:56:22  <piscisaureus_>in x64 more
10:56:28  <bnoordhuis>cool
10:57:02  <piscisaureus_>SleepConditionVariableSRW may be a little more complicated
10:57:07  <piscisaureus_>But I think we do not need it
10:57:08  <bnoordhuis>i don't use it
10:57:17  <piscisaureus_>that
11:00:32  <bnoordhuis>i'm going to land the work so far first though
11:00:55  <bnoordhuis>using the fallback if TryAcquireSRWLockShared isn't available
11:01:09  <bnoordhuis>i'll look into implementing srwlocks ourselves next week
11:01:11  <bnoordhuis>piscisaureus_: ^
11:03:37  <piscisaureus_>bnoordhuis: yeah, sure
11:03:41  <piscisaureus_>bnoordhuis: no hurry ;-)
11:12:45  <bnoordhuis>piscisaureus_: critical sections aren't spinlocks, are they?
11:21:19  <piscisaureus_>bnoordhuis: nope they aren't
11:21:25  <bnoordhuis>piscisaureus_: cool, thanks
11:21:32  <piscisaureus_>bnoordhuis: actually, I just figured out that SRW locks are also not spin locks
11:21:51  <piscisaureus_>well they spin for a number of cycles, then they turn into wait locks
11:21:56  <piscisaureus_>just like critical sections
11:22:05  <bnoordhuis>right, that sounds plausible
11:22:20  <piscisaureus_>bnoordhuis: that massively complicates stuff :-/
11:22:50  <piscisaureus_>bnoordhuis: but the Try functions actually do not wait ....
11:22:58  <piscisaureus_>hmm
11:23:02  <piscisaureus_>scheisse
11:23:09  <piscisaureus_>bnoordhuis: let's do this later.
11:23:13  <bnoordhuis>let's :)
11:23:16  <piscisaureus_>importing things first
11:23:22  <piscisaureus_>*important
11:28:28  <piscisaureus_>bnoordhuis: another question
11:28:35  <bnoordhuis>piscisaureus_: another answer
11:28:58  <piscisaureus_>bnoordhuis: what would you want to do with this "list of handles", besides killing them all?
11:29:15  <bnoordhuis>piscisaureus_: print them
11:29:48  <bnoordhuis>piscisaureus_: maybe selectively walk them so i can check how many outstanding tcp connections there are
11:29:55  <bnoordhuis>that last one is not really essential though
11:30:06  <bnoordhuis>but printing would be sweet debugging goodness
11:44:07  <CIA-95>libuv: Ben Noordhuis master * r1fc1f28 / (12 files in 6 dirs):
11:44:07  <CIA-95>libuv: Wrap platform mutex and rwlock APIs.
11:44:07  <CIA-95>libuv: Read/write locks are emulated with critical sections on Windows XP and Vista
11:44:07  <CIA-95>libuv: because those platforms don't have a (complete) native read/write lock API. - http://git.io/pyVs-g
11:45:52  <piscisaureus_>bnoordhuis: I don't see where that code is
11:45:57  <piscisaureus_>where you detect that you're on vista
11:46:20  <bnoordhuis>piscisaureus_: if (pTryAcquireSRWLockShared)
11:46:32  <piscisaureus_>aah ok
11:46:34  <piscisaureus_>subtle :-)
11:48:23  <bnoordhuis>piscisaureus_: i can change it into something more explicit like -> if (HAVE_SRWLOCK_API())
11:48:46  <piscisaureus_>bnoordhuis: no this is fine actually
11:55:01  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/libuv/commit/17ac452 <- i kind of like it actually
12:00:39  <CIA-95>libuv: Ben Noordhuis master * re4680cc / src/win/thread.c : windows: make SRWLock API usage more explicit - http://git.io/vxzBjQ
12:00:56  <bnoordhuis>so i just went and did it :)
12:09:01  <CIA-95>node: Ben Noordhuis master * r1b2d333 / (19 files in 6 dirs): uv: upgrade to e4680cc - http://git.io/0HoBtQ
12:09:01  <CIA-95>node: Ben Noordhuis master * ra808ab4 / src/node_crypto.cc : crypto: use the libuv rwlock API - http://git.io/ZUUAvA
12:23:06  * AvianFluquit (Ping timeout: 260 seconds)
12:33:53  * AvianFlujoined
13:00:24  * paddybyersjoined
13:10:00  <CIA-95>node: Ben Noordhuis issue2061 * r1e34c85 / (6 files in 4 dirs): uv: upgrade to b297d36 - http://git.io/X8FHHA
13:10:00  <CIA-95>node: Ben Noordhuis issue2061 * r8bed0a3 / wscript :
13:10:01  <CIA-95>node: build: compile with -D__DARWIN_64_BIT_INO_T on OS X
13:10:01  <CIA-95>node: Fixes a struct stat size mismatch on 64 bits machines that made Node crash with
13:10:01  <CIA-95>node: a EXC_BAD_ACCESS on startup.
13:10:01  <CIA-95>node: Fixes #2061. Solution proposed by Paddy Byers. - http://git.io/vY8nlA
13:16:13  <paddybyers>bnoordhuis: that darwin flag is only relevant to 64 bit systems
13:16:23  <paddybyers>and only some compiler/linker setttings
13:16:31  <bnoordhuis>oh man
13:16:36  <bnoordhuis>what should i check for?
13:16:47  <paddybyers>that's the problem; I don't know :(
13:16:51  <bnoordhuis>does it hurt on 32 bit systems?
13:17:21  <paddybyers>I was thinking that we really need to get the gyp system in shape, in the sense that it currently forces everything to gcc 4.2 and ia32
13:17:30  <paddybyers>really it should detect what;
13:17:40  <paddybyers>what's appropriate for the host/target
13:17:52  <bnoordhuis>this is a patch for v0.6, that still uses waf
13:18:15  <paddybyers>so who maintains the homebrew package?
13:19:16  <CIA-95>node: Ben Noordhuis v0.6 * r8bdf6ba / doc/api/dgram.markdown : docs: document dgram error event - http://git.io/iVQM0g
13:19:45  <bnoordhuis>paddybyers: no idea
13:19:49  <bnoordhuis>not us anyway
13:21:10  <paddybyers>I think a blind patch is just as likely to cause more problems; but at a minimum you would want the flag only for x86_64
13:22:29  <paddybyers>but I think the way to fix it properly is for us to get our own builds in order (maybe only post-0.6) and then the package maintainer has something to go on
13:23:07  <paddybyers>but homebrew is using clang I think, not gcc, and I think that's where the problem comes from
13:23:32  <bnoordhuis>yes - this is not the first issue we've had with homebrew builds
13:23:57  <paddybyers>I'll comment on the issue.
14:36:09  * creationixjoined
15:50:34  * dapjoined
16:03:45  <piscisaureus_>bnoordhuis: do you think I can use stl?
16:03:52  <piscisaureus_>or will ryah shoot me?
16:28:41  * isaacsjoined
16:28:42  <creationix>piscisaureus_: he'll shoot you
16:29:59  <creationix>but what do I know
16:36:42  <piscisaureus_>creationix: okay I will do it myself then
16:36:57  <piscisaureus_>I tried to use STL and I hate it already
16:41:56  * kuebkpart
16:44:04  <isaacs>i've encountered nothing but anger coming from the stl
16:51:26  * ericktjoined
17:02:07  <bnoordhuis>piscisaureus_: not just shoot you but perform unholy acts on your corpse
17:07:08  <piscisaureus_>bnoordhuis: ryah: igorzi: I will skip the call today
17:07:23  <piscisaureus_>I might be only tonight, I might be not
17:07:53  <bnoordhuis>piscisaureus_: anything to report?
17:08:31  <piscisaureus_>bnoordhuis: yes I almost finished the watcher list
17:08:49  <bnoordhuis>piscisaureus_: source or it didn't happen
17:10:00  <piscisaureus_>bnoordhuis: okay, it didn't happen
17:10:29  <bnoordhuis>piscisaureus_: colour me curious
17:10:32  * bnoordhuisis off to dinner
17:13:38  * ericktquit (Ping timeout: 244 seconds)
17:14:08  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
17:18:41  * paddybyers_joined
17:21:42  * paddybyersquit (Ping timeout: 245 seconds)
17:21:42  * paddybyers_changed nick to paddybyers
17:35:40  * piscisaureus_joined
17:43:26  * rmustaccpart
17:56:06  * brsonjoined
18:00:26  * indexzerojoined
18:10:44  * piscisaureus_quit (Ping timeout: 252 seconds)
18:14:42  <igorzi>ryah bnoordhuis: i will most likely also skip the call today
18:23:54  <isaacs>well, it looks like i'll be the only one on the call then.
18:24:13  <isaacs>maybe we'll have it in person.
18:25:10  <bnoordhuis>i'll be there
18:25:15  <bnoordhuis>in the spirit, that is
18:26:12  * isaacschats with bnoordhuis's spirit
18:27:53  * mikealquit (Quit: Leaving.)
18:34:55  * ericktjoined
18:38:46  * piscisaureus_joined
19:53:58  <piscisaureus_>test
19:54:06  * creationixjoined
19:57:15  <bnoordhuis>i admit defeat - i can't reproduce that os x segfault
19:57:30  <bnoordhuis>v0.6 compiles and works with both apple gcc and homebrew clang
19:57:37  <bnoordhuis>paddybyers: ^
19:58:30  <paddybyers>bnoordhuis: are on Lion?
19:58:36  <bnoordhuis>paddybyers: yes
19:59:33  <paddybyers>bnoordhuis: that's the same as I found - homebrew worked for me
19:59:54  <bnoordhuis>so it's a leopard problem?
20:00:13  <paddybyers>maybe I can try to reproduce the conditions to make it fail
20:01:07  <paddybyers>I had the crash on Lion, but not with homebrew; it was in Xcode after manually changing the compiler to clang 64 bit
20:02:04  <bnoordhuis>paddybyers: maybe that's it, my clang is 64 bit
20:03:13  * bnoordhuisrecompiles with -m32
20:03:55  <CIA-95>node: Ryan Dahl master * r5859361 / Makefile : Add missing rm - http://git.io/3LhWSg
20:05:28  * AvianFlu_changed nick to AvianFlu
20:06:29  <paddybyers>bnoordhuis: I might try a standalone program and see which options recreate it; or you could be recompiling all night
20:07:15  <bnoordhuis>paddybyers: this macbook has 8 cores so it's pretty speedy :)
20:07:29  <bnoordhuis>`JOBS=8 make -j 8` takes less than a minute
20:10:46  * dapquit (Quit: Leaving.)
20:14:56  <bnoordhuis>32 bit builds work fine too...
20:15:10  <bnoordhuis>i give up, i'm going to merge that patch and cross my fingers
20:17:50  <CIA-95>libuv: Ben Noordhuis v0.6 * r01fbe7b / (config-unix.mk uv.gyp):
20:17:50  <CIA-95>libuv: build: compile with -D__DARWIN_64_BIT_INO_T on OS X
20:17:50  <CIA-95>libuv: Fixes a struct stat size mismatch on 64 bits machines that made Node crash with
20:17:50  <CIA-95>libuv: a EXC_BAD_ACCESS on startup.
20:17:50  <CIA-95>libuv: Solution proposed by Paddy Byers. - http://git.io/MtQmgg
20:18:04  <CIA-95>node: Ben Noordhuis v0.6 * r3f92031 / (6 files in 4 dirs): uv: upgrade to 01fbe7b - http://git.io/_5ozLw
20:18:05  <CIA-95>node: Ben Noordhuis v0.6 * rf004d5a / wscript :
20:18:05  <CIA-95>node: build: compile with -D__DARWIN_64_BIT_INO_T on OS X
20:18:05  <CIA-95>node: Fixes a struct stat size mismatch on 64 bits machines that made Node crash with
20:18:05  <CIA-95>node: a EXC_BAD_ACCESS on startup.
20:18:05  <CIA-95>node: Fixes #2061. Solution proposed by Paddy Byers. - http://git.io/mJMQ6w
20:19:53  <bnoordhuis>piscisaureus_: https://github.com/joyent/libuv/pull/222
20:24:42  <paddybyers>bnoordhuis: ok :)
20:25:31  <bnoordhuis>paddybyers: not that i won't appreciate it if you come up with a reproducible test case
20:25:54  <paddybyers>bnoordhuis: still trying but no luck yet
20:28:00  <ryah>im testing the build on my old mac...
20:35:11  <piscisaureus_>bnoordhuis: I will look at the pull request tomorrow
20:41:20  * AvianFluquit (Ping timeout: 260 seconds)
20:42:43  * AvianFlujoined
20:43:40  * indexzeroquit (Quit: indexzero)
20:43:48  * piscisaureus_quit (Ping timeout: 244 seconds)
20:53:37  * isaacsjoined
20:54:15  * creationixquit (Ping timeout: 244 seconds)
20:54:48  * dapjoined
21:04:40  <bnoordhuis>https://gist.github.com/36d8a02ed0f6e96986f0 <- changelog for 0.6.2, looks okay to everyone?
21:05:00  * bnoordhuistopic: v0.6.2 tonight. release manager: bnoordhuis
21:05:14  * brsonquit (Ping timeout: 276 seconds)
21:06:37  <ryah>bnoordhuis: lgtm
21:12:00  * brsonjoined
21:13:07  <bnoordhuis>ryah: is `make dist` supposed to create a node-v0.6.2.tar.gz when the current tag is v0.6.2-rc1?
21:13:40  <bnoordhuis>looking at tools/getnodeversion.py i suppose it is
21:21:08  <ryah>bnoordhuis: it's using getnodeversion.py now
21:21:20  <ryah>that just changed in the last release
21:21:30  <ryah>i found the git describe stuff too annoying
21:21:38  <bnoordhuis>okay, no worries - i just scp'ed the files by hand
21:21:39  <ryah>because i had to keep retagging shit
21:21:50  <ryah>bnoordhuis: if you do "make dist-upload" it will upload
21:22:00  <ryah>i guess it will name them without "RC" though
21:22:46  <ryah>bnoordhuis: okay im going to test the tarballs
21:22:49  <bnoordhuis>okay, here we go - http://nodejs.org/dist/v0.6.2/node-v0.6.2-rc1.tar.gz and http://nodejs.org/dist/v0.6.2/node-v0.6.2-rc1.pkg
21:23:55  <ryah>bnoordhuis: you forgot to update src/node_version.h
21:24:12  <bnoordhuis>nu-uh, i did update it
21:24:29  <ryah>% /usr/local/bin/node -v
21:24:29  <ryah>v0.6.2-pre
21:24:36  <ryah>^-- using the pkg install
21:24:42  <bnoordhuis>-#define NODE_VERSION_IS_RELEASE 0
21:24:42  <bnoordhuis>+#define NODE_VERSION_IS_RELEASE 1
21:25:06  <bnoordhuis>hrm, the pkg probably didn't get rebuilt
21:26:03  <igorzi>bnoordhuis: it might be worth mentioning the windows bug i fixed yesterday, where stdout.write caused exception if node was started with nul stdout
21:26:31  <bnoordhuis>igorzi: i didn't see it in the logs, was that a libuv fix?
21:27:47  <igorzi>bnoordhuis: yep
21:29:31  * dapquit (Quit: Leaving.)
21:29:34  <bnoordhuis>ryah: new pkg uploaded, this one reports as v0.6.2
21:29:51  <bnoordhuis>igorzi: i'll update the changelog - can you give me a 80 chars max blurb?
21:29:53  * dapjoined
21:32:24  <igorzi>bnoordhuis: windows: fix stdout writes when redirected to nul
21:32:43  <bnoordhuis>igorzi: thanks
21:38:17  <bnoordhuis>igorzi ryah: who is testing what version of windows?
21:39:51  <igorzi>bnoordhuis: i'll test it. is there an rc?
21:40:33  <ryah>bnoordhuis: im compiling
21:40:39  <bnoordhuis>igorzi: not as an msi, ryah is going to build those
21:42:57  <ryah>[00:51|% 100|+ 293|- 3]: Done <-- illumos
21:43:55  <ryah>[01:13|% 100|+ 294|- 2]: Done <-- linux
21:45:34  <ryah>bnoordhuis: did you update the pkg?
21:45:52  <bnoordhuis>ryah: yes
21:46:12  <ryah>cool - works
21:50:20  <bnoordhuis>[14:32|% 100|+ 638|- 8]: Done <- sunos
21:54:33  <bnoordhuis>[15:16|% 100|+ 643|- 3]: Done <- linux
21:57:48  <ryah>win still building...
21:58:54  <bnoordhuis>[23:17|% 100|+ 633|- 13]: Done <- os x lion
21:59:05  <bnoordhuis>worse than sunos...
21:59:11  <wankdanker>is the 0.6.2 tar package made from the 0.6 branch on github?
21:59:27  <wankdanker>i built and tested win xp based on that
21:59:41  <ryah>wankdanker: yes
22:00:17  <wankdanker>[17:50|% 100|+ 277|- 19]: Done <-- win xp (https://gist.github.com/1377895)
22:00:31  <wankdanker>let me know if i'm missing any deps to run the win xp tests
22:00:42  <wankdanker>more than happy to run these for you guys when you need them.
22:01:41  <bnoordhuis>wankdanker: thanks
22:01:51  <bnoordhuis>not great but okay-ish
22:05:32  <ryah>[03:38|% 100|+ 286|- 10]: Done <-- win7
22:05:59  <ryah>isaacs:
22:06:02  <ryah>=== release test-zlib-from-string ===
22:06:02  <ryah>Path: simple/test-zlib-from-string
22:06:02  <ryah>assert.js:104
22:06:02  <ryah> throw new assert.AssertionError({
22:06:02  <ryah> ^
22:06:04  <ryah>AssertionError: gzip encoded string should match
22:06:06  <ryah> at C:\Users\ryan\node-v0.6.2\test\simple\test-zlib-from-string.js:37:1
22:06:17  <isaacs>well, that's not right.
22:06:29  <ryah>let me uplaod the .exe and .msi
22:06:32  <isaacs>k
22:06:47  * bnoordhuisboots virtualbox
22:07:47  <isaacs>ryah: does the assertion error show what you're getting and what you're expecting?
22:11:05  <ryah>isaacs: http://nodejs.org/dist/v0.6.2/node-v0.6.2-rc1.msi
22:11:57  <ryah>i love how tightly packaged the .pkg and .msi are
22:12:02  <ryah>2mb
22:12:05  <ryah><3
22:13:09  <bnoordhuis>ryah igorzi: i get this warning when installing the msi: "An installation for Microsoft .NET framework 2.0 Service Pack 2 is currently suspended. You must undo the changes made by that installation to continue. Do you want to undo those changes?"
22:13:23  <bnoordhuis>i'm scared :(
22:14:11  <bnoordhuis>but i suspect it's benign? i probably did abort a .net upgrade some time ago
22:14:34  <igorzi>don't be scared bnoordhuis :) just undo. i doubt that you care about have .net framework sp2 :)
22:14:43  <igorzi>*having
22:15:01  <bnoordhuis>igorzi: very true :) it indeed installs fine when i click undo
22:16:48  * ryahbans this vlad person from the mailing list
22:21:14  <bnoordhuis>isaacs: Assertion failed: in_off + in_len <= Buffer::Length(in_buf), file src\node_zlib.cc, line 104 <- windows xp
22:24:22  <bnoordhuis>isaacs: sorry, belay that - the test runner was running the wrong node binary
22:25:19  <isaacs>ryah: lol
22:25:21  <isaacs>good all
22:25:22  <isaacs>*call
22:28:41  <ryah>i set him on moderate again
22:29:17  <wankdanker>is that regarding the ms sql thead? are people not aware that node-odbc exists?
22:30:36  <ryah>isaacs: are you able to figure out the zlib failure?
22:30:42  <isaacs>i'm digging into it
22:30:51  <isaacs>i'm not sure yet.
22:30:59  <isaacs>there's definitely one byte different, but i'm not sure if it's actually a valid failure.
22:32:23  <isaacs>yeah, when i unzip it, i get the input string back. so, i think this is just one of those cases where lib z is being a black box.
22:32:47  <isaacs>it should be all deterministic, but i dunno.
22:33:05  <isaacs>if i pack it on windows, then take that output, and unpack on mac, it works, and vice versa
22:34:36  <isaacs>yeah, the diff is in the gzip header. so, not even part of the actual input string.
22:34:41  <isaacs>?
22:34:55  <isaacs>maybe a different zlib version or something? some header thingie? i dunno. i'll make the test valid.
22:35:51  <ryah>isaacs: sounds good
22:36:34  <ryah>bnoordhuis: did you get igorzi's changelog item integrated?
22:37:06  <ryah>isaacs: you might want to leave a bit of a comment in the test
22:37:11  <ryah>isaacs: just in case this pops up again
22:37:12  <isaacs>yep, already on that :)
22:37:21  <isaacs>more to keep anyone from "fixing" it :)
22:37:31  <bnoordhuis>ryah: yes
22:37:36  <bnoordhuis>but... comment in the test?
22:37:43  <bnoordhuis>oh wait, that's for isaacs
22:37:48  <isaacs>want it in 0.6, or just master?
22:38:14  <ryah>isaacs: 0.6
22:38:48  <ryah>bnoordhuis: can i see the version bump diff?
22:39:29  <isaacs>look good? https://github.com/isaacs/node/commit/975f1930b1c6ec1ee1a9e52e603aec097a088720
22:40:18  <bnoordhuis>ryah: you can, you may - https://github.com/bnoordhuis/node/commit/a63e7bd
22:41:24  <ryah>isaacs: sure - i haven't tested it but seems fine
22:41:31  <isaacs>wfm
22:41:44  <isaacs>if it fails for you, at least it'll be a more useful failing test :)
22:42:07  <CIA-95>node: isaacs v0.6 * r975f193 / test/simple/test-zlib-from-string.js : zlib: Fix invalidly failing test - http://git.io/xZLF4w
22:42:24  <ryah>bnoordhuis: okay- can you put up a new tarball
22:42:29  * ircretaryjoined
22:42:30  <ryah>i think this will be the release tarball
22:43:18  <ryah>bnoordhuis: also prepare the blog.nodejs.org post
22:43:37  <bnoordhuis>ryah: will do
22:45:46  <igorzi>[08:58|% 100|+ 306|- 14]: Done <--- test-all on win2k8
22:47:32  <bnoordhuis>http://nodejs.org/dist/v0.6.2/node-v0.6.2-rc2.tar.gz and http://nodejs.org/dist/v0.6.2/node-v0.6.2-rc2.pkg
22:51:59  <ryah>signtool.exe is pretty great
22:52:02  <ryah>very easy
22:52:14  <ryah>seems a bit insecure - but whatever
22:52:44  <bnoordhuis>insecure in what sense?
22:53:27  <ryah>no password entry
22:53:46  <ryah>i dont even really understand how i got the certificates on my computer
22:53:54  <ryah>i just clciked a link someone sent me
22:53:55  <ryah>:)
22:54:12  <igorzi>ryah: it probably installed the cert into the windows cert store
22:54:13  <ryah>and i dont have to specify the joyent cert or anything
22:54:22  <ryah>i just do "signtool sign Release\node.exe"
22:54:25  <ryah>and it does it
22:54:44  <ryah>which is great - because i don't really give a fuck - but i could see how this would bother people
22:55:33  <ryah>or maybe it's just awesome - it knows i only have one cert
23:03:14  <bnoordhuis>https://blog.nodejs.org/?p=361&preview=true
23:05:32  <isaacs>anyone know how to create a file on windows with a path longer than 259 characters?
23:05:42  <isaacs>i'm getting crazy enoents and stuff.
23:05:43  * indexzerojoined
23:08:57  <bnoordhuis>so are we good to go?
23:11:43  <ryah>bnoordhuis: http://nodejs.org/docs/v0.6.2/
23:12:12  <bnoordhuis>ryah: i haven't actually done `make dist-upload` yet
23:12:40  <ryah>bnoordhuis: that's not going to upload the docs anyway
23:12:47  <ryah>bnoordhuis: i think those need to be done manually still
23:13:01  <bnoordhuis>okay, will do now
23:13:33  <ryah>okay i upload .exe .msi .pdb for rc2
23:13:39  <ryah>i think we're good to go
23:13:58  <ryah>[02:27|% 100|+ 290|- 6]: Done <-- win7
23:18:28  * indexzeroquit (Quit: indexzero)
23:21:06  <bnoordhuis>cool, i'll push the tag, publish the blog post and send a mail to the ML
23:23:08  <CIA-95>node: Ben Noordhuis v0.6 * ra4402f0 / (4 files in 3 dirs): Bump version to v0.6.2 - http://git.io/01ewpg
23:23:08  <CIA-95>node: Ben Noordhuis v0.6 * r61e796d / src/node_version.h : Now working on v0.6.3 - http://git.io/zxjsfQ
23:23:43  * piscisaureus_joined
23:24:04  <ryah>bnoordhuis: rename the rc2 files
23:24:14  <ryah>currently those links 404
23:24:26  <ryah>the msft files are already renamed
23:24:30  <bnoordhuis>working on it
23:26:03  <piscisaureus_>isaacs: You could try prepending \\?\
23:26:08  <piscisaureus_>but I woud not recommend it
23:26:19  <isaacs>piscisaureus_: i was just going through and trying that?
23:26:24  <isaacs>why would you not recommend it?
23:26:28  <piscisaureus_>it's something that node should do automatically if necessary
23:26:35  <piscisaureus_>isaacs: because the paths are unparsed
23:26:41  <piscisaureus_>so / is no longer equivalent to \
23:26:44  <piscisaureus_>and more weirdness
23:26:46  <isaacs>yeah
23:26:48  <piscisaureus_>but I don't know
23:26:52  <isaacs>i'm also replacing / with \
23:27:01  <piscisaureus_>you should probably make paths absolute
23:27:10  <isaacs>they're already absolute
23:27:16  <piscisaureus_>ok, good
23:29:28  <isaacs>i agree that it's something node should do automatically internally, but there's a bunch of business folks that are going to get ulcers if this isn't working soon.
23:30:43  <piscisaureus_>ACH SO
23:31:25  <isaacs>ach so?
23:32:05  <ryah>it's a stupid thing that germans say
23:32:13  <ryah>all the time
23:33:32  <piscisaureus_>ryah: do the next release on a thursday
23:33:50  <piscisaureus_>then I will be able to do it / participate for a change
23:34:20  <isaacs>piscisaureus_: should i use unc path for reads also?
23:34:31  <isaacs>piscisaureus_: or just when i have to write? it seems like i it just fails on file creation
23:34:43  <piscisaureus_>isaacs: that should not make a difference
23:34:57  <isaacs>so do em on reads, too?
23:35:00  <piscisaureus_>yep
23:35:59  <ryah>piscisaureus_: ok
23:37:53  <igorzi>wow.. i just read that ms sql thread :)
23:38:33  <bnoordhuis>blog post is up, mailing list notified
23:38:36  <ryah>bnoordhuis: good to go?
23:38:37  <bnoordhuis>there is no going back now guys
23:38:38  <ryah>ok
23:38:57  <ryah>only forward
23:39:50  <isaacs>this life goes one direction.
23:40:28  <piscisaureus_>the ms sql thread is retarded
23:40:49  <piscisaureus_>bnoordhuis++ ryah++
23:40:55  <ryah>bnoordhuis: nice!
23:41:46  <bnoordhuis>yay :)
23:42:11  * ryahuses the new realtime google analytics to watch people come to the site
23:42:23  <piscisaureus_>rly
23:42:24  <piscisaureus_>ghe
23:42:34  * piscisaureus_tries the msi installer for the first time
23:43:27  <piscisaureus_>I like that it's only 2.1mb
23:45:10  <ryah>yeah
23:45:12  <piscisaureus_>I would like the msi to support upgrading
23:45:25  <ryah>yeah, really - super annoying to uninstall it first
23:45:38  <piscisaureus_>I am going to do it.
23:45:52  <piscisaureus_>Logo, path questions, upgrade
23:46:12  <ryah>awesome
23:46:24  <ryah>maybe people will take us seriously then
23:47:32  * bnoordhuistopic: libuv, the little library that could
23:47:34  <ryah>we're handling about 7 req/min
23:47:39  <ryah>on nodejs.org
23:47:48  <piscisaureus_>nah
23:47:56  <piscisaureus_>wait for the twitter announce to spreak
23:47:59  <piscisaureus_>*spread
23:48:13  <ryah>getting hotter now
23:48:29  <ryah>17 req in the last minute
23:48:29  <piscisaureus_>we should also have an icon for node.exe
23:48:44  <piscisaureus_>I like the white n in the green hexagon that joyent uses on its site
23:49:04  <igorzi>piscisaureus_: msi upgrades are pretty annoying afaik.. there are 2 types of upgrades minor and major.. major is easy, it just auto-uninstalls the previous install before installing the new one
23:49:17  <piscisaureus_>igorzi: so that's good
23:49:26  <piscisaureus_>igorzi: we just make every upgrade a major :-)
23:49:51  <igorzi>yeah, that's good.. the only possible issue is node_modules folder.. it would be nice to keep it
23:50:00  <igorzi>between installs
23:50:07  <piscisaureus_>the node_modules folder should not be located in program files imho
23:50:14  <piscisaureus_>for now, that's prolly fine
23:50:22  <igorzi>yeah
23:50:22  <ryah>why not?
23:50:25  <piscisaureus_>but you get in trouble with UAC
23:51:00  <igorzi>i think it's actually nice that it gets in trouble with UAC.. that'll mean that people would need to run elevated to install anything --global
23:51:17  <ryah>yes
23:51:17  <piscisaureus_>hmm ok
23:51:46  <piscisaureus_>the problem is that elevating is painful
23:52:04  <piscisaureus_>because npm will not trigger an elevation prompt- it will just fail
23:52:26  <piscisaureus_>ryah: the exe is not signed or?
23:52:33  <ryah>piscisaureus_: it should be
23:52:50  <isaacs>piscisaureus_: it just fails in unix, too
23:52:54  <isaacs>if root owns the global dir.
23:54:00  <piscisaureus_>ryah: msi not signed here
23:54:03  <piscisaureus_>er, exe
23:54:07  <piscisaureus_>only the msi is signed
23:54:40  <ryah>piscisaureus_: http://nodejs.org/dist/v0.6.2/node.exe ?
23:54:46  <ryah>or the .exe installed with the msi?
23:54:55  <piscisaureus_>the exe installed with the msi
23:55:03  <ryah>how do you check that?
23:55:28  <piscisaureus_>ryah: oh the downloadable exe *is* signed
23:55:43  <ryah>ok - i sign it after i created the msi
23:55:46  <ryah>should i upload a new one?
23:55:47  <ryah>:/
23:56:01  <piscisaureus_>ryah: right click the file, "properties", "digital signatures"
23:56:04  <piscisaureus_>ryah: nah, it's fine
23:56:40  <piscisaureus_>not much of a deal, since the exe is installed via a safe route windows trusts it anyway
23:57:16  <igorzi>you can also run "signtool.exe verify /a node.exe"
23:58:44  <piscisaureus_>C:\Users\Bert Belder\Downloads>signtool verify /a node.exe
23:58:44  <piscisaureus_>SignTool Error: A certificate chain processed, but terminated in a root
23:58:44  <piscisaureus_> certificate which is not trusted by the trust provider.
23:58:44  <piscisaureus_>Number of errors: 1
23:59:07  <piscisaureus_>^-- is what I get for the downloaded exe
23:59:29  <piscisaureus_>same for the msi btw