00:00:09  <isaacs>oh, wait, right
00:00:20  <isaacs>for a second i thought you were shimming the world.
00:00:24  <ryah>DrPizza: has ms released win8 on arm?
00:00:29  <ryah>for preview?
00:00:30  <isaacs>'We'll make every package work everywhere!!'
00:00:43  <bnoordhuis>the plan9 port is next!
00:01:38  <isaacs>bnoordhuis: don't forget about putting gopher, xmpp, and soap support into libuv
00:01:46  * isaacstrollin since ryah's here
00:02:16  * ryahfigured :)
00:02:26  <bnoordhuis>i kinda like gopher, actually
00:03:49  <bnoordhuis>i like the rust community
00:04:05  <bnoordhuis>i wrote a module once, now people send me PRs every now and again to update it against the latest rust
00:04:12  <DrPizza>ryah: nope, no preview of ARM
00:04:42  <ryah>i assume surface is arm
00:10:20  <isaacs>bnoordhuis: hm, the int/void thing is still breaking
00:10:22  <isaacs>../nas.cc: In function ‘v8::Handle<v8::Value> DoSomethingAsync(const v8::Arguments&)’:
00:10:25  <isaacs>../nas.cc:46: error: invalid conversion from ‘int (*)(eio_req*)’ to ‘void (*)(eio_req*)’
00:10:28  <isaacs>../nas.cc:46: error: initializing argument 1 of ‘eio_req* eio_custom(void (*)(eio_req*), int, int (*)(eio_req*), void*, eio_channel*)’
00:10:44  <bnoordhuis>isaacs: you need to apply that patch i linked
00:13:34  <isaacs>oh, ok
00:14:28  <ryah>bnoordhuis: are you releaving libuv of its libc dependency?
00:14:41  <ryah>bnoordhuis: i noticed syscall.c today in there
00:14:42  * isufyjoined
00:14:59  <isaacs>bnoordhuis: i see. because the internal structure of eio changed.
00:15:02  <isaacs>yeah...
00:15:05  <bnoordhuis>ryah: yes
00:15:07  <isaacs>ok, this is far enough.
00:15:13  <ryah>bnoordhuis: awesome
00:15:18  <bnoordhuis>thanks :)
00:15:30  <isaacs>bnoordhuis: ok, lgtm, man. i'ts an awful dirty hack, but whatever.
00:15:41  <isaacs>it'll remove an excuse for people to upgrade :)
00:15:55  <isaacs>bnoordhuis: can you rebase and land it?
00:16:01  <bnoordhuis>sure
00:16:42  <CIA-108>node: Ben Noordhuis master * rf405517 / (src/node.h tools/installer.js src/ev-emul.h): build: install ev-emul.h - http://git.io/d-MOEw
00:16:43  <CIA-108>node: Ben Noordhuis master * r1b22f53 / (src/node.h tools/installer.js src/eio-emul.h): build: install eio-emul.h - http://git.io/96hM2A
00:16:43  <ryah>did anyone say how much MSFT surface costs?
00:23:57  <isaacs>ryah: no, but i heard it runs on node
00:24:02  <isaacs>ryah: or uses node on the server side
00:24:02  <isaacs>or somethign
00:24:33  <isaacs>someone was like, "nook microsoft tablet node, you hear about it?" and i said "is it a 0.8 blocker?" and they looked at me like i was a crazy person and i told htem to shut up.
00:24:47  <CIA-108>node: Trent Mick v0.6 * rdd1c3d7 / wscript : build: always link with -lz - http://git.io/4TgCsg
00:26:55  * TooTallNatejoined
00:34:55  * xaqquit (Remote host closed the connection)
00:34:59  <ryah> /quit
00:35:01  * ryahquit (Quit: leaving)
00:36:13  <isaacs>bnoordhuis: what would go into making eio-emul work with int (*)(eio_req*) functions?
00:36:44  <isaacs>bnoordhuis: we could conceivably have a function that takes a int (*)(eio_req*) and returns a void (*)(eio_req*), no?
00:37:39  <dap>I'd think you could just cast the function pointer on the way in
00:38:02  <isaacs>dap: that doesn't seem to work, at least in my naive attempt
00:38:09  <dap>compile error?
00:38:25  <isaacs>dap: # define eio_custom(a,b,c,d) eio_custom((void (*)(eio_req*))a,b,(void (*)(eio_req*))c,d,(&uv_default_loop()->uv_eio_channel))
00:38:28  <isaacs>yeah
00:38:44  <dap>where's the actual eio_custom signature?
00:38:52  <isaacs>../nas.cc:44: error: invalid conversion from ‘void (*)(eio_req*)’ to ‘int (*)(eio_req*)’
00:39:06  <isaacs>eio_req* eio_custom(void (*)(eio_req*), int, int (*)(eio_req*), void*, eio_channel*)
00:39:21  <isaacs>oh, wait, it wants 3 to be an int?
00:39:24  <isaacs>one sec
00:40:19  <isaacs>huh. yeah, that works.
00:40:23  <isaacs>i was just doing it wrong
00:40:25  <dap>cool.
00:40:40  <isaacs>ok, let's do this 100%. wth, why not.
00:41:11  <isaacs>dap: the only remaining delta, then, would be that it installs to build/Release instead of the 0.4-style build/default/
00:41:44  <isaacs>dap: you can make wscript do that, though, and gyp defaults to Release
00:42:13  * mcavagequit (Ping timeout: 244 seconds)
00:42:34  <dap>I think 0.6 uses build/Release too, doesn't it?
00:43:07  <TooTallNate>are we adding deprecation warnings to all these shims?
00:43:13  <isaacs>dap: yeah
00:43:21  <isaacs>TooTallNate: not atm
00:43:25  <isaacs>TooTallNate: the ev ones have them
00:43:31  <isaacs>because ev is actually gone, and re-implemented in terms of uv
00:43:33  <dap>isaacs: so I don't think that's any new breakage with that.
00:43:41  <isaacs>dap: k, kewl.
00:43:57  <dap>thanks!
00:45:20  <isaacs>wait a second... it looks like *only* eio_custom uses a void function
00:47:47  <CIA-108>node: isaacs reviewme * r1df222f / lib/net.js : Fix breakage introduced in de65ba7 - http://git.io/7xzsBg
00:47:47  <CIA-108>node: Ben Noordhuis reviewme * rf405517 / (src/node.h tools/installer.js src/ev-emul.h): build: install ev-emul.h - http://git.io/d-MOEw
00:47:48  <CIA-108>node: Ben Noordhuis reviewme * r1b22f53 / (src/node.h tools/installer.js src/eio-emul.h): build: install eio-emul.h - http://git.io/96hM2A
00:47:48  <CIA-108>node: isaacs reviewme * r8996b5e / src/eio-emul.h : eio-emul: Cast eio_custom method to void(*)(eio_req*) - http://git.io/w-bRbA
00:48:08  <isaacs>you know we're moving fast with the "reviewme" branch is a fast-forward ;)
00:49:11  <CIA-108>node: isaacs reviewme * r5cfe0b8 / src/eio-emul.h : eio-emul: Cast eio_custom method to void(*)(eio_req*) - http://git.io/idWgSQ
00:49:14  <isaacs>just removed that stray empty line
00:49:29  <isaacs>bnoordhuis: review? https://github.com/joyent/node/commit/5cfe0b86d5be266ef51bbba369c39e412ee51944
00:56:58  * mjr_joined
00:57:45  <bnoordhuis>isaacs: lgtm
00:59:10  <CIA-108>node: isaacs master * r5cfe0b8 / src/eio-emul.h : eio-emul: Cast eio_custom method to void(*)(eio_req*) - http://git.io/idWgSQ
01:00:17  * isaacstopic: ONE MOAR! https://github.com/joyent/node/issues?milestone=10&sort=updated&state=open
01:00:47  <isaacs>ok, i'm gonna branch 0.7.12, probably release it tomorrow morning
01:02:05  * \toothrotjoined
01:02:53  * toothrquit (Ping timeout: 244 seconds)
01:09:01  * bbenviejoined
01:09:51  <CIA-108>node: isaacs master * r10fa526 / LICENSE : Update npm's license - http://git.io/ysfiug
01:09:55  <CIA-108>node: isaacs master * redb62dd / (125 files in 19 dirs): npm: Upgrade to 1.1.29 - http://git.io/8ZGGRQ
01:10:35  * toothrjoined
01:12:08  * bnoordhuis_joined
01:12:54  * \toothrotquit (Ping timeout: 240 seconds)
01:13:10  * bnoordhuisquit (Ping timeout: 264 seconds)
01:13:10  * benviequit (Ping timeout: 264 seconds)
01:25:52  <isaacs>just tag push
01:26:01  <isaacs>oh, cia doesn't update on those, i guess
01:27:28  * bnoordhuis_quit (Ping timeout: 246 seconds)
01:29:38  * pieternquit (Quit: pietern)
01:36:41  * isufyquit
01:38:16  * isufyjoined
01:53:25  * ericktquit (Ping timeout: 245 seconds)
01:58:12  <CIA-108>node: isaacs v0.7.12-release * r02985a8 / tools/email-footer.md : email-footer: Label Windows x64 installer properly - http://git.io/pTGpsw
01:58:12  <CIA-108>node: isaacs v0.7.12-release * r5d6d4c1 / (ChangeLog src/node_version.h): 2012.06.19, Version 0.7.12 (unstable) - http://git.io/tcBxTA
02:03:50  * isaacstopic: Test, please: http://nodejs.org/dist/v0.7.12/
02:04:09  * isaacstopic: Test, please: http://nodejs.org/dist/v0.7.12/ (or branch v0.7.12-release)
02:13:46  * dapquit (Quit: Leaving.)
02:14:24  * brsonquit (Ping timeout: 245 seconds)
02:38:06  * abraxasjoined
02:41:47  * mjr_quit (Quit: mjr_)
03:22:00  * brsonjoined
03:22:00  * brsonquit (Client Quit)
03:22:21  * brsonjoined
03:23:56  * isufyquit
03:35:35  * \toothrotjoined
03:37:56  * toothrquit (Ping timeout: 276 seconds)
03:41:59  * isufyjoined
03:46:12  * xaqjoined
03:53:49  * xaqquit (Read error: Connection reset by peer)
03:55:38  * xaq_joined
04:08:29  * xaq_quit (Ping timeout: 276 seconds)
04:08:50  * xaqjoined
04:40:21  * xaqquit (Read error: Connection reset by peer)
04:40:39  * xaqjoined
04:51:23  * xaqquit (Ping timeout: 276 seconds)
05:30:19  * paddybyersjoined
05:31:05  * toothrjoined
05:33:24  * \toothrotquit (Ping timeout: 240 seconds)
05:46:01  * paddybyersquit (Quit: paddybyers)
05:46:59  * isufyquit
05:55:35  * \toothrotjoined
05:58:28  * toothrquit (Ping timeout: 248 seconds)
06:03:24  * AvianFlupart ("Leaving")
06:21:14  * paddybyersjoined
06:49:24  * stephankquit (Quit: *Poof!*)
07:00:16  * TooTallNatequit (Quit: Computer has gone to sleep.)
07:05:04  * brsonquit (Ping timeout: 244 seconds)
07:09:44  * AndreasMadsenjoined
07:13:18  * mmaleckijoined
07:20:47  * AndreasMadsenquit (Remote host closed the connection)
07:29:09  * mmaleckiquit (Ping timeout: 246 seconds)
07:41:32  * bbenviequit
08:06:19  * mmaleckijoined
08:25:37  * benviejoined
08:51:04  * hzjoined
09:15:57  * rendarjoined
09:51:31  * hzquit
09:56:37  * mmaleckiquit (Quit: Lost terminal)
09:59:36  * mmaleckijoined
10:07:57  * abraxasquit (Remote host closed the connection)
10:10:18  * paddybyersquit (Quit: paddybyers)
11:43:18  * bnoordhuis_joined
11:53:26  <CIA-108>node: Ben Noordhuis master * r7de6eba / (src/eio-emul.h src/node.h): node: preemptively unbreak windows build - http://git.io/89hvWg
11:56:55  * skomskijoined
11:57:14  <einaros>will 0.7.12 be the last 0.7 release?
12:03:13  * skomskiquit (Remote host closed the connection)
12:09:12  * \toothrotchanged nick to toothr
12:22:23  <bnoordhuis_>einaros: yes, provided no critical bugs pop up
12:24:32  <einaros>hopefully the v8 perf regression fix will be included by then
12:25:06  <bnoordhuis_>yep, that's the one blocking issue
12:28:23  <einaros>do you currently have a single box somewhere which you use to generate stable / comparable benchmark results?
12:37:10  * mmaleckiquit (Ping timeout: 265 seconds)
13:22:36  * xaqjoined
13:34:13  <indutny>morning guys
13:40:30  <indutny>bnoordhuis_: so what do we have?
13:40:46  <indutny>bnoordhuis_: nothing than v8 perf issue, right?
13:48:12  * dvv-androidjoined
13:49:20  * c4milojoined
13:49:29  * dvv-androidquit (Remote host closed the connection)
13:54:27  * dvv-androidjoined
13:54:36  <saghul>bnoordhuis_ I know you guys are busy now, but when you can, here is he ares stuff: https://github.com/joyent/libuv/pull/472
13:56:40  * dvv-androidquit (Remote host closed the connection)
13:58:02  * dvv-androidjoined
13:58:20  * dvv-androidquit (Remote host closed the connection)
14:02:05  * irajoined
14:02:11  * mcavagejoined
14:05:30  * `3rdEdenjoined
14:06:49  * irapart
14:07:01  * ericktjoined
14:10:01  <bnoordhuis_>indutny: yes, that's the last one
14:10:08  <bnoordhuis_>saghul: will look at it today
14:10:20  <saghul>no hurries :-)
14:19:01  * mcavagequit (Quit: Lost terminal)
14:24:52  * mmaleckijoined
14:27:43  * mcavagejoined
14:29:06  * TheJHjoined
14:46:39  <bnoordhuis_>[12:03|% 100|+ 845|- 1]: Done <- not bad
14:47:04  <bnoordhuis_>and the one failing test passed on the second run
14:49:51  <indutny>what was that?
14:51:29  * rendarquit (Ping timeout: 245 seconds)
14:53:36  * piscisaureus_joined
14:56:00  <isaacs>bnoordhuis_: which test?
14:56:11  <bnoordhuis_>isaacs: simple/test-domain-http-server
14:56:32  <bnoordhuis_>forgive me the spam but:
14:56:34  <bnoordhuis_>=== release test-domain-http-server ===
14:56:34  <bnoordhuis_>Path: simple/test-domain-http-server
14:56:34  <bnoordhuis_>listening on localhost:12346
14:56:34  <bnoordhuis_>server error { name: 'AssertionError',
14:56:34  <bnoordhuis_> message: 'Data should not be undefined',
14:56:35  <bnoordhuis_> actual: false,
14:56:37  <bnoordhuis_> expected: true,
14:56:39  <bnoordhuis_> operator: '==',
14:56:41  <bnoordhuis_> domain_thrown: true,
14:56:43  <bnoordhuis_> domain:
14:56:47  <bnoordhuis_> { members: [ [Object], [Object] ],
14:56:49  * bnoordhuis_changed nick to bnoordhuis
14:56:51  <bnoordhuis> _events: { error: [Function] } } }
14:56:54  <bnoordhuis>client error { [SyntaxError: Unexpected token A]
14:56:56  <bnoordhuis> domain_thrown: true,
14:56:57  <bnoordhuis> domain:
14:56:59  <bnoordhuis> { members: [ [Object], [Object] ],
14:57:01  <bnoordhuis> _events: { error: [Function], dispose: [Function] } } }
14:57:03  <bnoordhuis>requests=5 responses=1
14:57:05  <bnoordhuis>Command: out/Release/node /home/bnoordhuis/src/nodejs/master/test/simple/test-domain-http-server.js
14:57:07  <bnoordhuis>--- TIMEOUT ---
14:57:15  <isaacs>bnoordhuis: that's bonkers...
14:57:35  <isaacs>also: spamming test results in here when asked is 100% acceptable, and imo polite.
14:57:39  * tellnesjoined
14:57:49  <bnoordhuis>it was probably cosmic rays or solar flares
14:57:53  <indutny>bnoordhuis: I won't kick you for that
14:58:07  <isaacs>bnoordhuis: i think that test does a json.parse on the results
14:58:10  <bnoordhuis>you're too kind, fedor :)
14:58:21  <isaacs>bnoordhuis: so, there may be a http edge case bug somewhere. probably not news to anyone.
14:58:46  <bnoordhuis>isaacs: no, kind of odd that it only happens sometimes. i'm running it in a loop and it fails maybe 1 out of 5 times
14:58:59  <bnoordhuis>and so far only with the release build
14:59:54  <bnoordhuis>and now it's passing all the time...
14:59:54  <isaacs>oh, that's super odd
15:01:35  <isaacs>bnoordhuis: it might nice to put some logging in there before the JSON.parse so that we can see what the error is when it fails.
15:01:55  <isaacs>bnoordhuis: is https://github.com/joyent/node/commit/7de6eba0616c9333566e0fb815b589a4e4536b49 required for 0.7.12, you think?
15:02:03  * rendarjoined
15:03:22  <bnoordhuis>logging added. still continually passing though :)
15:03:42  <bnoordhuis>isaacs: yes
15:04:01  <bnoordhuis>i discovered afterwards that japj had filed a bug report about it
15:04:20  <isaacs>k. i haven't done the build yet for windows. i'll pull it in. thanks for saving me the time :)
15:06:49  <bnoordhuis>isaacs: it's choking on JSON.parse
15:06:56  <piscisaureus_>isaacs: you want me to land that v8 patch ?
15:06:58  <bnoordhuis>as to why...
15:07:20  * theColejoined
15:07:26  <isaacs>piscisaureus_: i have not investigated it. but sure. if you say it's good, i believe you
15:07:39  <piscisaureus_>isaacs: I think it is. I can try it now if you want to
15:07:58  <piscisaureus_>isaacs: (a WIP version of the patch fixed the perf issues with einaros/ws)
15:08:04  <isaacs>awesome.
15:08:24  <isaacs>yeah, let's land on 0.7.12. then we can at least see if it breaks anything before 0.8
15:08:36  <piscisaureus_>isaacs: btw, I am very happy that you fixed the pause/resume before connect issue.
15:08:56  <piscisaureus_>isaacs: doing that in libuv would have been quite some work and risky. We can do it for 0.8.x
15:09:36  <isaacs>sure. but now it's just a task we can get to whenever, not a thing that breaks programs.
15:09:48  <piscisaureus_>yes
15:14:22  * pieternjoined
15:14:47  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
15:16:44  <isaacs>piscisaureus_: lemme know when i should pull master into v0.7.12-release
15:18:21  <indutny>isaacs: no objections against release so far, right?
15:18:26  <indutny>I mean 0.8
15:20:48  <isaacs>indutny: no, i think when piscisaureus_ lands this perf reg thing, we're good to go
15:20:53  <isaacs>basically 0.7.12 will be 0.8
15:21:17  <piscisaureus_>isaacs: actually, there's still one test failure that I don't know the cause of
15:21:18  <indutny>cool
15:21:40  <piscisaureus_>isaacs: namely, test-cluster-disconnect
15:21:54  <isaacs>piscisaureus_: ok, so we might have SOME changes for 0.8
15:22:11  <isaacs>piscisaureus_: * Fix cluster disconnect behavior (Bert Belder) <-- entire changelog
15:22:18  <piscisaureus_>:-p
15:22:22  <indutny>hahaha
15:22:26  <isaacs>that should be the whole email.
15:22:34  <indutny>it's ok to pull fixes into 0.8
15:22:35  <isaacs>"That's what we did for 0.8. Bert fixed a bug. Enjoy."
15:22:44  <indutny>so feel free to broke anything just to show your name in changelog later
15:22:44  <isaacs>indutny: yes. same as 0.6
15:22:49  <isaacs>hahah
15:23:00  <indutny>I always do that
15:23:07  <indutny>this way I'm looking like an important person
15:23:17  <isaacs>indutny: so THAT's why you're so good at finding bugs.
15:23:21  <indutny>hahaha
15:23:22  <isaacs>you already know where tehy are!
15:23:42  <indutny>if I don't know - bnoordhuis knows
15:23:45  <indutny>bnoordhuis: ;)
15:23:47  <indutny>hahahaha
15:24:00  <indutny>and if something is broken on windows
15:24:05  <indutny>...
15:24:12  <indutny>I won't point my finger on that person
15:24:13  <isaacs>speedup loops.
15:24:17  <indutny>yeah
15:24:19  <isaacs>we need more speedup loops.
15:24:28  <indutny>core: speedup
15:24:29  <indutny>diff:
15:24:34  <indutny>- sleep(10)
15:24:41  <isaacs>yep :)
15:24:50  <indutny>controlled benchmarks, what could be better
15:29:04  <bnoordhuis>isaacs: in simple/test-domain-http-server, is JSON.parse expected to fail sometimes?
15:29:21  <isaacs>bnoordhuis: i don't think so...
15:29:54  <bnoordhuis>okay. i ask because it does sometimes
15:30:13  <bnoordhuis>specifically, it's trying to parse a stringified AssertionError
15:30:18  <isaacs>right
15:30:24  <isaacs>that's wrong :)
15:32:11  * AndreasMadsenjoined
15:35:06  <piscisaureus_>isaacs: ok, cherry picked all the patches from bleeding edge that I think we need
15:35:10  <piscisaureus_>isaacs: testing now
15:35:50  <isaacs>sweet
15:36:39  <piscisaureus_>isaacs: I am going throught the full test suite to be sure that nothing breaks (incl. v8 tests), so this is going to take a while
15:37:00  <isaacs>sure take your time
15:37:09  * tjfontainepictures the xkcd compiling comic
15:37:18  <bnoordhuis>isaacs: can you look with me at simple/test-domain-http-server?
15:37:29  <isaacs>bnoordhuis: sure, looking now
15:37:35  <bnoordhuis>i know what's happening but i don't know *why* it's happening :)
15:37:58  <bnoordhuis>isaacs: line 54, assert(data !== undefined, 'Data should not be undefined');
15:38:11  <isaacs>right
15:38:17  <bnoordhuis>that assertion triggers because req.url == '/' and objects['/'] doesn't exist
15:38:51  <bnoordhuis>line 71 does makeReq('/')
15:38:56  <piscisaureus_>did you guys add any tests
15:39:02  <piscisaureus_>?
15:39:10  <bnoordhuis>piscisaureus_: no, it's an existing test
15:39:19  <piscisaureus_>isaacs: bnoordhuis ^?
15:39:19  <piscisaureus_>because I now have 5 new failures wrt cluster
15:39:23  <piscisaureus_>those weren't there yesterday
15:39:35  <bnoordhuis>piscisaureus_: no, no new tests at all
15:41:00  <bnoordhuis>isaacs: re line 71, does that mean that the test expects at least some requests to fail and return a bogus json response?
15:41:38  <isaacs>bnoordhuis: i'm seeing "client error { [SyntaxError: Unexpected token A]" but the domain is catching it
15:41:50  <isaacs>bnoordhuis: so yeah, i guess it IS supposed to throw
15:41:59  <isaacs>but it's supposed to be handled in the domain
15:42:18  <bnoordhuis>okay, that makes sense
15:42:25  <bnoordhuis>because the test does pass most of the time
15:43:49  * elijah-mbppart
15:47:51  * stephankjoined
15:52:00  <CIA-108>node: Ben Noordhuis master * r8966480 / tools/installer.js : Revert "build: fix `make install DESTDIR=/path`" - http://git.io/7OGEOQ
15:52:35  <bnoordhuis>isaacs: ^ you should pull that into v0.8
15:52:49  <isaacs>bnoordhuis: haha
15:52:59  <isaacs>bnoordhuis: someone in the joyent jabber room was just bitching about that
15:53:07  <isaacs>i'll pull it into 0.7.12
15:53:48  * brsonjoined
15:55:30  <isaacs>bnoordhuis: that's for #3489, right?
15:57:01  <piscisaureus_>cluster is entirely broken on windows
15:57:23  <isaacs>piscisaureus_: really? was it ever working?
15:57:28  <piscisaureus_>isaacs: yes
15:57:42  <isaacs>piscisaureus_: maybe the stdio changes?
15:57:48  <piscisaureus_>isaacs: I will bisect
15:57:51  <CIA-108>node: danno@chromium.org master * rbf682af / deps/v8/src/mips/macro-assembler-mips.cc : MIPS: Fix packed-element.js test on ARM with no snap - http://git.io/6KJJ5A
15:57:52  <CIA-108>node: erik.corry@gmail.com master * rf94b85d / (5 files in 3 dirs):
15:57:52  <CIA-108>node: Revert part of 11727 as it sometimes tanked V8 benchmark (raytrace) performance
15:57:52  <CIA-108>node: for reasons that are not obvious. Now we make objects into fast-case objects
15:57:52  <CIA-108>node: when they are made prototypes for other objects, but we do not mark objects that
15:57:53  <CIA-108>node: are already fast case with a bit that helps keep them in fast case. Review URL:
15:57:53  <CIA-108>node: https://chromiumcodereview.appspot.com/10556004 - http://git.io/OKlDLQ
15:57:54  <CIA-108>node: jkummerow@chromium.org master * r7282ce9 / deps/v8/src/objects.cc : Remove unused-but-set variable from objects.cc. - http://git.io/JAuRjQ
15:57:54  <CIA-108>node: verwaest@chromium.org master * r8a6ef65 / (deps/v8/src/objects.cc deps/v8/src/objects.h):
15:57:55  <CIA-108>node: Fixing bugs in promotion of elements transitions (r1175). - Fixed invalid memory
15:57:55  <CIA-108>node: access when reading enum-cache from descriptor array with elements transitions
15:57:56  <CIA-108>node: but 0 real descriptors. - Fixed infinite recursion in the intrusive map iterator
15:57:56  <CIA-108>node: when visiting elements transitions. - Properly cached non-fastmode elements
15:58:45  <CIA-108>node: transitions. - http://git.io/uKhtyw
16:01:59  <isaacs>bnoordhuis: ok, I just ran that test 1000 times. zero failures.
16:02:10  <isaacs>bnoordhuis: what os are you running on?
16:02:25  <isaacs>bnoordhuis: ubuntu?
16:02:42  <bnoordhuis>isaacs: yes, ubuntu
16:04:40  <isaacs>50 passes on CentOS
16:04:43  <isaacs>trying ubuntu
16:05:18  <isaacs>it'll be a little while, I guess I don't have node checked out into my user dir here
16:09:27  <isaacs>bnoordhuis: aha!
16:09:33  <isaacs>something is getting hung up and laggy.
16:09:46  <isaacs>i don't think this is necessarily a domain bug (though it might be)
16:11:39  <piscisaureus_>isaacs: 0a89e8b838caf0769807c4e9a665a70d2e63b032 is the first bad commit
16:11:44  <piscisaureus_>isaacs: iow, add stdin to forks
16:12:19  <piscisaureus_>aah
16:14:04  <isaacs>piscisaureus_: that's a nice-to-have.
16:14:10  <isaacs>piscisaureus_: but it WOULD be nice to have.
16:14:18  <piscisaureus_>isaacs: it's easy to fix
16:14:22  <isaacs>we can revert it, though. i'm not overly attached to it for 0.8. we can clean it up for 0.9
16:14:26  <piscisaureus_>isaacs: indutny just forgot something
16:14:29  <isaacs>piscisaureus_: how easy? minutes, or hours?
16:14:32  * mmaleckiquit (Ping timeout: 276 seconds)
16:14:35  <piscisaureus_>isaacs: minutes
16:14:45  <isaacs>k. go forth and vanquish!
16:16:12  <piscisaureus_>there is one issue though.... node cannot detect which is the IPC channel's FD
16:16:18  <piscisaureus_>that's sort of unrelated to this bug tho
16:16:35  <isaacs>piscisaureus_: i thought we put it in the env
16:16:40  <piscisaureus_>do we?
16:16:44  <isaacs>NODE_IPC_FD or something?
16:16:44  <isaacs>yeah
16:16:52  <isaacs>NODE_CHANNEL_FD maybe?
16:17:06  <isaacs>bnoordhuis: so, how does one debug a hung program on linux?
16:17:08  <piscisaureus_>oh, kewl
16:17:22  <piscisaureus_>isaacs: that was exactly the solution I was thinking of :-)
16:17:55  <isaacs>piscisaureus_: oh, we probably need to make sure we remove it from other grandchildren. maybe that's handled already, i dunno
16:18:06  * piscisaureus_does not know either
16:19:00  <piscisaureus_>isaacs: yes, that's taken care of at startup time
16:19:50  <bnoordhuis>isaacs: attach gdb
16:20:32  <bnoordhuis>isaacs: more specifically, attach gdb and inspect the backtrace
16:23:09  * dapjoined
16:24:23  <isaacs>bnoordhuis: this what you'er seeing? https://gist.github.com/2955079
16:24:38  <isaacs>bnoordhuis: looks like on ubuntu we're getting into a case where we're hanging in epoll wait
16:24:46  <bnoordhuis>yes
16:25:06  <bnoordhuis>but... have to go, dinner is ready
16:26:10  <isaacs>k.
16:26:19  <isaacs>thanks for your help, ben
16:30:31  * brsonquit (Ping timeout: 252 seconds)
16:30:36  <piscisaureus_>isaacs: the problem wasn't what I though it would be - but I still have good hope I can fix it fast :-)
16:40:21  * AndreasMadsenquit (Remote host closed the connection)
16:47:35  * brsonjoined
16:48:27  <isaacs>ok
16:54:30  <CIA-108>node: isaacs master * r6ba1c62 / (105 files in 10 dirs): npm: Upgrade to 1.1.30 - http://git.io/I7uUwA
16:56:53  <CIA-108>node: isaacs v0.7.12-release * r86782f0 / (ChangeLog src/node_version.h): 2012.06.19, Version 0.7.12 (unstable) - http://git.io/hHeuxQ
16:56:55  <CIA-108>node: isaacs v0.7.12-release * r0e00d89 / tools/email-footer.md : email-footer: Label Windows x64 installer properly - http://git.io/gyiopA
16:58:29  <isaacs>piscisaureus_: i'm getting a little itchy to shove this out the door. how's it coming?
16:58:41  <piscisaureus_>isaacs: hey man, relax :-)
16:58:46  <isaacs>hahah
16:58:51  * isaacshad too much coffee, i think
16:58:57  <piscisaureus_>isaacs: yes, I have a tentative fix, I am testing it
16:59:03  <isaacs>fantastic
16:59:28  <isaacs>i'm gonna ride the trains around for a little bit, visit the office today.
16:59:31  <piscisaureus_>isaacs: please don't ship without. the ipc channel is completely broken on windows
16:59:37  <piscisaureus_>isaacs: do that
16:59:38  <isaacs>ok
16:59:51  <piscisaureus_>isaacs: in 30 minutes it will all be said and done
16:59:54  <isaacs>piscisaureus_: thanks. as usual you are the voice of reason for the poor downtrodden windows users.
17:00:07  <isaacs>great :)
17:02:19  * theColequit (Quit: theCole)
17:15:34  * \toothrotjoined
17:17:24  * toothrquit (Ping timeout: 240 seconds)
17:19:04  * toothrjoined
17:21:24  * brsonquit (Ping timeout: 240 seconds)
17:22:20  * \toothrotquit (Ping timeout: 265 seconds)
17:24:30  <CIA-108>libuv: Bert Belder reviewme * r6709065 / src/win/process.c : windows: set the child_pid property for all IPC pipes in uv_spawn - http://git.io/X2vYMg
17:24:32  <piscisaureus_>bnoordhuis: ^-- care to review ?
17:24:55  <piscisaureus_>I would have igorzi do it if he was still with us :-)
17:25:13  <piscisaureus_>japj: hey
17:31:54  * TooTallNatejoined
17:34:21  * paddybyersjoined
17:34:58  * mmaleckijoined
17:36:11  * brsonjoined
18:01:58  * mikealjoined
18:02:52  * brsonquit (Ping timeout: 246 seconds)
18:04:38  * mmaleckiquit (Quit: Reconnecting)
18:04:48  * mmalecki_joined
18:05:05  * brsonjoined
18:06:04  * \toothrotjoined
18:08:24  * toothrquit (Ping timeout: 240 seconds)
18:09:06  <CIA-108>node: isaacs v0.7.12-release * r9956bb6 / (doc/api/cluster.markdown lib/querystring.js): trivial: Doc typo and lint fix - http://git.io/BnXuIw
18:09:07  <CIA-108>node: isaacs v0.7.12-release * r077108f / (ChangeLog src/node_version.h): 2012.06.19, Version 0.7.12 (unstable) - http://git.io/XxcSTA
18:10:02  <isaacs>koichik is correcting my prose. What he lacks in english fluency, he more than makes up for in intense precision :)
18:10:45  <piscisaureus_>haha
18:11:00  <piscisaureus_>isaacs: I need someone to review http://git.io/X2vYMg
18:12:11  * mmalecki_quit (Quit: Reconnecting)
18:12:21  <piscisaureus_>isaacs: btw, I just came to the conclusion that this new stdio api does not work for one of these cases that we had in mind
18:12:22  <piscisaureus_>(the one where you write a passphrase to fd 4)
18:12:22  <piscisaureus_>since all fds > 0 are read-only
18:15:02  <piscisaureus_>and I found another bug in child_process.js
18:15:43  * rendarquit
18:17:56  * ericktquit (Quit: erickt)
18:18:11  <isaacs>piscisaureus_: http://git.io/X2vYMg looks good to me
18:18:29  <isaacs>piscisaureus_: but i dont' know if you want to take my word for it
18:18:30  <piscisaureus_>isaacs: :-)
18:18:41  <piscisaureus_>I'll take the risk
18:18:46  <isaacs>ok
18:19:00  <CIA-108>libuv: Bert Belder master * r6709065 / src/win/process.c : windows: set the child_pid property for all IPC pipes in uv_spawn - http://git.io/X2vYMg
18:19:05  <isaacs>if it makes the fork and cluster tests work on windows, then i'm all for it
18:19:07  <indutny>piscisaureus_: isaacs huh?
18:19:11  <indutny>what I forgot?
18:19:22  <isaacs>indutny: windows libuv was setting ipc on fd:0 only
18:19:24  * `3rdEdenjoined
18:19:32  <indutny>isaacs: oh
18:19:33  <piscisaureus_>indutny: yeah, it wasn't your fault after all
18:19:34  * mmaleckijoined
18:19:40  <indutny>piscisaureus_: you should document it!
18:19:41  <piscisaureus_>indutny: sorry for pointing fingers
18:19:41  <indutny>:)
18:19:45  <indutny>piscisaureus_: np
18:19:56  <indutny>you can change a letter in my nick and I won't know next time
18:19:56  <indutny>:D
18:20:00  <indutny>indutnyy
18:20:05  <isaacs>piscisaureus_: for the "connect fd4 to read a password" you'd have to open /dev/tty and pass the fd as the child proc's #4
18:20:11  <isaacs>piscisaureus_: i'm not sure how to even do that on windows
18:20:19  <isaacs>piscisaureus_: but it'll unblock those who could do it before
18:20:46  <piscisaureus_>isaacs: well, not necessarily... you could also just set stdio: [0, 1, 2, 'pipe']
18:20:54  <isaacs>piscisaureus_: right, or that
18:20:58  * pieternquit (Read error: Connection reset by peer)
18:21:05  <piscisaureus_>isaacs: except for the fact that that pipe would be read-only :-(
18:21:11  * pieternjoined
18:21:32  <isaacs>piscisaureus_: hm. on windows, or everywhere?
18:21:37  <piscisaureus_>isaacs: everywhere
18:21:47  <piscisaureus_>isaacs: I just tried, I cannot make the pipe duplex
18:21:59  <isaacs>RO meaning that the child cannot write to it? or the parent can't write to it?
18:22:07  <piscisaureus_>isaacs: the parent can't write to it
18:22:25  <isaacs>hm...
18:22:31  <piscisaureus_>isaacs: if I make all pipes duplex then other stuff breaks
18:22:41  <isaacs>what other stuff?
18:22:55  <piscisaureus_>isaacs: because if (readable==0) then stream.end == close but if (readable==1) then stream.end == send FIND
18:22:57  <piscisaureus_>*send FIN
18:23:14  <piscisaureus_>isaacs: but if you try to pipe to e.g. grep then sending FIN does not yield the desired effect
18:24:51  <indutny>piscisaureus_: that's wierd
18:25:34  <piscisaureus_>indutny: well, that part I tried only on windows
18:25:41  <indutny>brb
18:25:55  <piscisaureus_>окна
18:26:36  * mmaleckiquit (Ping timeout: 260 seconds)
18:26:47  <piscisaureus_>Федор Индутный скоро вернусь
18:28:15  <indutny>hahaha
18:28:17  <indutny>wtf
18:30:04  * toothrjoined
18:31:20  * \toothrotquit (Ping timeout: 245 seconds)
18:31:25  * pieternquit (Read error: Operation timed out)
18:33:58  * mmaleckijoined
18:34:01  <piscisaureus_>I do not know why your name ends with ый btw
18:34:12  <piscisaureus_>or rather, why that is not reflected in english
18:34:52  * mikealquit (Quit: Leaving.)
18:35:15  * mikealjoined
18:35:32  * CoverSlidejoined
18:37:48  <indutny>piscisaureus_: Indutnyy looks better to you?
18:38:07  <piscisaureus_>Indutney looks more sophisticated
18:38:11  * xaqquit (Read error: Connection reset by peer)
18:38:15  <indutny>indeed
18:38:27  <indutny>but my father was publishing under that transliterated name
18:38:35  <indutny>so I decided to continue that years ago
18:38:40  * xaqjoined
18:39:00  <indutny>anton indutny if you're interested in medicine
18:40:04  <piscisaureus_>But I think Indutnyi is more common
18:40:31  <indutny>actually, I've Indutnyy in my foreign passport
18:40:38  <indutny>and Fedor should be Fiodor then
18:41:32  <indutny>Фёдор - that's how it's really written in russian
18:41:54  <piscisaureus_>ah, right
18:42:01  <piscisaureus_>e -> io would be kinda surprising
18:42:12  <indutny>heh
18:42:13  <piscisaureus_>Fiodor Indootnyy
18:42:19  <indutny>hahaha
18:42:21  <indutny>indeed
18:44:34  * \toothrotjoined
18:44:48  <indutny>Vyacheslav Egorov is much simplier
18:45:16  <indutny>though somebody claim that pronouncing Vyacheslav is hard for non-russians
18:45:28  <piscisaureus_>Вячеслав Егоров
18:45:41  <piscisaureus_>straigtforward
18:45:46  <piscisaureus_>but I cannot pronounce indeed
18:46:24  * toothrquit (Ping timeout: 240 seconds)
18:46:43  <piscisaureus_>but that's also because his name reveals very little about the pronounciation
18:46:45  <isaacs>Aww... no unicode in my irssi.
18:46:47  <piscisaureus_>He should have used
18:46:49  <isaacs>all i see are ???
18:47:00  <piscisaureus_>vee-dzjez-lav je-goo-ruv
18:47:49  <isaacs>piscisaureus_: indutny: So, 0.7.12. what's the call? got something for me to pull in?
18:48:16  <indutny>isaacs: no
18:48:18  <indutny>should I?
18:48:32  <isaacs>indutny: mostly that's for piscisaureus_ :)
18:48:36  <indutny>ok
18:48:42  <indutny>windows is always late :P
18:49:02  <CIA-108>node: Bert Belder master * r5e1d844 / deps/uv/src/win/process.c : uv: upgrade to 67090653 - http://git.io/OXhfMw
18:49:07  <piscisaureus_>^-- isaacs: that
18:49:53  * \toothrotquit (Ping timeout: 276 seconds)
18:51:31  <einaros>websocket benchmark comparison of 0.6.19 and 0.7.12-release: https://gist.github.com/8ca3d7e8f8dadebe1787
18:51:41  <einaros>0.7.12 is significantly faster where it matters
18:52:08  <TooTallNate>einaros++
18:52:33  * \toothrotjoined
18:54:38  <isaacs>einaros: woot.
18:54:40  <isaacs>ok, great.
18:54:56  <isaacs>piscisaureus_: yeah, you can't write to the extra stdio's.
18:55:02  <isaacs>piscisaureus_: that's not really ideal, i guess.
18:55:25  <isaacs>piscisaureus_: but this is an unstable interface. we can either call this a bug in 0.8, and fix it there, or just see how much it burns, and fix it in 0.9.
18:55:38  <isaacs>i'm not gonna hold the release for it. let's not do any more surgery.
18:56:10  <isaacs>piscisaureus_: anything for the ChangeLog for that win/process.c change?
18:56:31  <indutny>brb
18:57:08  <piscisaureus_>isaacs: no
18:57:23  <piscisaureus_>isaacs: well, I agree that you should not hold the release for the stdio issue
18:57:35  <piscisaureus_>there is another issue as well :-0
18:57:40  <piscisaureus_>isaacs: we can also fix that later
18:57:48  <isaacs>kewl. post a bug :)
19:00:23  <piscisaureus_>isaacs: https://github.com/joyent/node/issues/3492 and https://github.com/joyent/node/issues/3493
19:01:05  <isaacs>thanks
19:01:48  <piscisaureus_>quite unsatisfying but yeah, we should release 0.8.0 first
19:02:28  * brsonquit (Ping timeout: 248 seconds)
19:03:18  <isaacs>piscisaureus_: those are really good bugs for bugfix releases :)
19:03:24  <isaacs>because they don't change semantics or API at all
19:03:33  * toothrjoined
19:03:39  <piscisaureus_>isaacs: before 0.8.0 I am going to make a slight change to the libuv ABI ...
19:03:44  <piscisaureus_>isaacs: so we can fix this later
19:04:10  <isaacs>kewl.
19:04:11  * \toothrotquit (Ping timeout: 276 seconds)
19:04:27  <isaacs>i'd much prefer not to change any behavior between 7.12 and 8.0
19:04:36  <isaacs>RC1 uploaded
19:05:17  <TooTallNate>bnoordhuis: you don't need to call uv_ref()/unref() when using uv_queue_work(), right?
19:06:42  <TooTallNate>piscisaureus_: ^?
19:06:48  <piscisaureus_>TooTallNate: no
19:07:07  * pieternjoined
19:07:19  <TooTallNate>piscisaureus_: just out of curiosity, where does that happen internally? (how does it keep the event loop alive?)
19:07:29  * brsonjoined
19:07:29  <piscisaureus_>TooTallNate: it calls uv_ref :-p
19:07:32  <piscisaureus_>oh hmm
19:07:33  <piscisaureus_>maybe not
19:07:38  * isufyjoined
19:08:44  <piscisaureus_>TooTallNate: I think the fact that there is a req keeps the loop alive
19:09:06  <piscisaureus_>TooTallNate: on unix, it could be that the uv_async that awaits thread pool jobs keeps it alive
19:09:20  <TooTallNate>piscisaureus_: good enough for me. just wanted to make sure that i wasn't bullshitting here :D https://github.com/developmentseed/node-sqlite3/pull/76#discussion_r1012427
19:10:08  * AndreasMadsenjoined
19:11:45  <CIA-108>libuv: Bert Belder master * r9dca9c3 / include/uv-private/uv-win.h : windows: reserve some room for zero-read minibuffer - http://git.io/UdwmvQ
19:12:09  * AndreasMadsenquit (Client Quit)
19:12:14  <CIA-108>libuv: Bert Belder master * rb496c12 / include/uv-private/uv-win.h : windows: reserve some room for zero-read minibuffer - http://git.io/kQDfoA
19:13:53  <CIA-108>node: Bert Belder master * r3f687db / deps/uv/include/uv-private/uv-win.h : uv: upgrade to b496c122 - http://git.io/lOQMQQ
19:15:01  <indutny>piscisaureus_: so forks won't have stdin on windows?
19:15:12  <piscisaureus_>indutny: huh, why?
19:15:36  <indutny>piscisaureus_: because IPC works only on stdin channel?
19:15:45  <piscisaureus_>indutny: no, you changed that, right?
19:15:53  <indutny>piscisaureus_: em...
19:16:01  <indutny>yes
19:16:06  <indutny>but you said that doesn't work on windows
19:16:37  <piscisaureus_>indutny: yes, but that was because there was a bug in libuv where the pid of the child wouldn't get stored properly when the ipc channel had an FD > 0
19:16:48  <piscisaureus_>indutny: so I fixed that in http://git.io/OXhfMw
19:16:52  <indutny>ah
19:17:02  <indutny>good
19:19:48  * isufyquit
19:21:45  * brsonquit (Ping timeout: 245 seconds)
19:26:52  * toothrquit (Ping timeout: 246 seconds)
19:28:33  * toothrjoined
19:33:50  * CoverSlidequit (Quit: leaving)
19:34:05  * mmaleckiquit (Ping timeout: 276 seconds)
19:34:10  * CoverSlidejoined
19:37:56  * travis-cijoined
19:37:56  <travis-ci>[travis-ci] joyent/libuv#427 (reviewme - 6709065 : Bert Belder): The build passed.
19:37:56  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/commit/67090653df47
19:37:56  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1657435
19:37:56  * travis-cipart
19:38:20  * brsonjoined
19:40:03  * toothr-joined
19:40:31  * toothrquit (Ping timeout: 246 seconds)
19:41:11  * sh1mmerquit (Quit: sh1mmer)
19:47:36  <indutny>that's really make me rad ^^
19:52:12  <indutny>I'm going to write some fast TLS Server with uv and C++ tomorrow
19:52:25  <indutny>to see how much handshake is openssl able to handle
19:53:59  <bnoordhuis>back
19:55:05  <bnoordhuis>TooTallNate: there's a couple of async and idle handles that keep the event loop alive
19:55:18  <bnoordhuis>TooTallNate: src/unix/uv-eio.c in case you're interested
19:57:29  * mmaleckijoined
20:06:00  <isaacs>back
20:06:04  <isaacs>time to build the binaries
20:06:13  <indutny>hehe
20:06:17  <indutny>for 0.7.12 or 0.8?
20:12:59  * indexzerojoined
20:13:45  <isaacs>bnoordhuis: review? https://gist.github.com/2956268
20:14:00  <isaacs>bnoordhuis: i'm told that this makes it compile if you're not doing 'using namespace node;'
20:16:19  <isaacs>indutny: for 0.7.12
20:16:26  <isaacs>0.8 will be on thursday :)
20:16:51  <indutny>you should build a big red button by then
20:17:15  <isaacs>:D
20:17:20  <isaacs>a big red button? what for?
20:17:29  <indutny>for release
20:17:35  <indutny>MOAR RELEASE!
20:17:55  <indutny>it'll post update on twitter/facebook/github/google plus simultaneously
20:18:03  <indutny>like guys at Facebook did
20:18:14  <bnoordhuis>isaacs: oh, there's a better solution, just wrap everything in an extern "C"
20:18:31  <bnoordhuis>i was planning to fix that tonight, realized at dinner i'd forgotten about that :)
20:18:39  <isaacs>oh, ok.
20:18:57  <bnoordhuis>wait, i'll write a quickie patch
20:19:11  <isaacs>sure.
20:19:28  <isaacs>please let me test it before landing
20:20:59  * `3rdEdenquit (Quit: Leaving...)
20:21:27  <bnoordhuis>isaacs: https://gist.github.com/e94c0e963e1779239b60
20:22:52  <isaacs>bnoordhuis: sweet, that fixes it also
20:22:57  <isaacs>and much prettier :)
20:23:12  <isaacs>wanna land it, or should I do it and assign blame to you?
20:24:37  <bnoordhuis>hah, no worries, i'll land it
20:24:55  <isaacs>i <3 git format-patch: https://gist.github.com/2956323
20:25:00  <isaacs>you can make any patch come from anywhere!
20:25:21  <isaacs>"Wait, I didn't say that... did i?"
20:26:25  <CIA-108>node: Ben Noordhuis master * r0fdee0a / src/ev-emul.h : node: fix namespacing issue in ev-emul.h - http://git.io/CzbWyA
20:26:37  <bnoordhuis>isaacs: should that go into 0.7.12-release?
20:26:49  <isaacs>awesome
20:26:57  <isaacs>yes
20:26:57  <CIA-108>node: isaacs v0.7.12-release * r6f0f1ee / tools/email-footer.md : email-footer: Label Windows x64 installer properly - http://git.io/YVYKow
20:26:58  <CIA-108>node: isaacs v0.7.12-release * r1e6f8b3 / (doc/api/cluster.markdown lib/querystring.js): trivial: Doc typo and lint fix - http://git.io/-zCCIg
20:26:58  <CIA-108>node: isaacs v0.7.12-release * rf37f0c6 / (ChangeLog src/node_version.h): 2012.06.19, Version 0.7.12 (unstable) - http://git.io/MNUlbA
20:27:01  <isaacs>boom
20:27:05  * `3rdEdenjoined
20:27:15  <isaacs>just rebased onto master
20:27:53  * TooTallNatequit (Ping timeout: 255 seconds)
20:28:21  <isaacs>RC2 online
20:31:09  <isaacs>binaries building
20:36:28  * TooTallNatejoined
20:37:03  <isaacs>indutny: yes, i like the BRB approach
20:40:18  <indutny>huh?
20:40:23  <indutny>BRB = Big Red Button?
20:43:17  * `3rdEdenquit (Quit: Leaving...)
20:45:16  <indutny>bbiab
20:49:42  <isaacs>bnoordhuis: do we do anything weird to strerror on purpose? https://github.com/joyent/node/issues/3494
20:49:52  <isaacs>bnoordhuis: or is that just some side effect somewhere?
20:51:15  <bnoordhuis>isaacs: os x?
20:52:54  <isaacs>bnoordhuis: yeah
20:52:58  <isaacs>that's where i'm seeing this anyway
20:53:15  <isaacs>haven't checked on linux
20:54:02  <bnoordhuis>ah, i can reproduce it
20:54:03  <bnoordhuis>odd
20:55:39  <indutny>what is stderror anyway?
20:56:38  <isaacs>indutny: strerror
20:56:49  <isaacs>indutny: it turns an errno into a string message, i think
20:56:52  <isaacs>or something
20:57:17  <indutny>oh
20:57:19  <indutny>magic
20:57:20  <indutny>bbiab
21:00:45  <isaacs>bnoordhuis: it has to have broken in the last day or so
21:00:51  <isaacs>or, no, actually, nvm
21:00:59  <isaacs>i don't know when root's node was built.
21:01:18  <tjfontaine>that sentence confuses me
21:03:13  <piscisaureus_>isaacs: ship it ?
21:03:18  <piscisaureus_>:-)
21:03:39  <piscisaureus_>isaacs: otherwise I am going to send more patches
21:03:45  <piscisaureus_>(this is a threat)
21:05:21  <isaacs>piscisaureus_: ha
21:05:25  <isaacs>going to ship soon.
21:05:35  <isaacs>i just wanna at least know the extent of this strerror weirdness
21:05:37  <bnoordhuis>found it
21:05:40  <isaacs>oh?!?
21:05:42  <isaacs>what's up?
21:06:13  <bnoordhuis>isaacs: https://github.com/bnoordhuis/node/commit/067be2b
21:06:26  <bnoordhuis>the #include was inside namespace node { }
21:06:40  <bnoordhuis>silly bug
21:07:02  <isaacs>yep. that is correct. land at will.
21:07:36  <CIA-108>node: Ben Noordhuis master * r067be2b / (src/ev-emul.h src/node.h): node: fix namespacing issue in ev-emul.h - http://git.io/JPcJ9A
21:07:47  <bnoordhuis>^ it's a force-push, i had already squashed it
21:07:57  * ericktjoined
21:08:00  <isaacs>ok.
21:08:51  <CIA-108>node: isaacs v0.7.12-release * r252f034 / tools/email-footer.md : email-footer: Label Windows x64 installer properly - http://git.io/-B86WA
21:08:51  <CIA-108>node: isaacs v0.7.12-release * rf105f2f / (doc/api/cluster.markdown lib/querystring.js): trivial: Doc typo and lint fix - http://git.io/6YBJDw
21:08:52  <CIA-108>node: isaacs v0.7.12-release * ra721201 / (ChangeLog src/node_version.h): 2012.06.19, Version 0.7.12 (unstable) - http://git.io/TMTYIw
21:09:00  <isaacs>ugh. rebuilding the binaries.
21:09:42  <isaacs>then we're shipping this thing
21:16:03  * mjr_joined
21:16:16  * japjjoined
21:23:47  <japj>heya, how's 0.7.12 doing? :)
21:24:55  <bnoordhuis>i'm creating a rackspace account and the form complains my completely non-alphanumeric password must contain at least one capital letter....
21:25:13  <tjfontaine>rackspace, spensive
21:25:22  * AndreasMadsenjoined
21:26:36  <bnoordhuis>the small instances are not so bad
21:27:09  <bnoordhuis>but, as it turns out, rackspace.nl won't let you pay with anything but creditcards
21:27:27  <tjfontaine>oh they changed their pricing since last I looked, more on par with linode
21:27:32  <bnoordhuis>i guess i won't be paying rackspace at all
21:28:46  <bnoordhuis>piscisaureus_: recommendations for dutch vps hosting?
21:29:10  <piscisaureus_>bnoordhuis: eh, do they exist?
21:29:34  * mmaleckiquit (Quit: leaving)
21:29:51  <bnoordhuis>piscisaureus_: they're numerous, what with our liberal laws
21:31:00  * travis-cijoined
21:31:00  <travis-ci>[travis-ci] joyent/libuv#428 (master - 6709065 : Bert Belder): The build passed.
21:31:00  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/382f2a262bc0...67090653df47
21:31:00  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1657877
21:31:00  * travis-cipart
21:31:13  <TooTallNate>piscisaureus_: so what would I have to do to this for windows support? https://github.com/TooTallNate/ttys/blob/master/index.js
21:31:19  <TooTallNate>i've been reading something about "CON:"
21:31:53  <isaacs>japj: building binaries (again)
21:31:57  <piscisaureus_>TooTallNate: yes, CON, CONIN$ or CONOUT$
21:32:04  <bnoordhuis>piscisaureus_: https://www.transip.nl/vps/prijzen-en-bestellen/ <- not a bad deal
21:32:04  <japj>piscisaureus_: should simple/test-cluster-http-pipe still fail (as normal user) on the v0.7.12-release branch?
21:32:08  <isaacs>japj: and making sure that the new joyent platform build finishes with it
21:32:15  <piscisaureus_>TooTallNate: but I don't think it is possible to open the thing in the proper mode with node right now
21:32:23  <isaacs>japj: there are still a few test failures on windows
21:33:32  * piscisaureus__joined
21:33:42  <japj>isaacs: ok, it would be nice if tracis would support windows at some point, but they're still working on vagrant (or something related) to support that I believe
21:33:49  <japj>*travis
21:34:57  <TooTallNate>piscisaureus_: ya, i guess not… I get "Error: ENOENT, no such file or directory 'C:\ttys\CONIN$'"
21:35:15  <piscisaureus__>TooTallNate: well, the reason for that is slightly different
21:35:34  <piscisaureus__>TooTallNate: you could try with process.binding('fs').open('con', 1, 0) or something
21:36:22  * piscisaureus_quit (Ping timeout: 246 seconds)
21:39:24  * brsonquit (Ping timeout: 245 seconds)
21:41:06  * mjr_quit (Quit: mjr_)
21:46:08  <TooTallNate>piscisaureus__: with "con" I got "not found" and with "CONIN$" I got "operation not permitted"
21:47:18  * piscisaureus__part
21:47:22  * piscisaureus__joined
21:47:59  * c4miloquit (Remote host closed the connection)
21:49:30  * sh1mmerjoined
21:51:31  * brsonjoined
21:52:49  <CIA-108>node: Bert Belder master * r41d5fef / src/res/node.ico : windows: update icon - http://git.io/h8zdag
21:54:26  <japj>new icon?
21:55:09  <piscisaureus__>well
21:55:20  <piscisaureus__>the previous one needed some attention
21:56:13  <TooTallNate>this new one's all fancy and 3d ;)
21:56:40  <piscisaureus__>TooTallNate: I made it slightly less 3d actually
21:56:44  <japj>wasn't that a release blocker for 0.7.12? ;)
21:56:55  <TooTallNate>piscisaureus__: what do you use btw?
21:57:07  <piscisaureus__>TooTallNate: well, paint.net
21:57:23  <piscisaureus__>that's for putting the ico together
21:57:29  <piscisaureus__>for the rendering I used Solidworks
21:57:45  <piscisaureus__>since I cannot work with any graphical software
21:57:56  <japj>is it a different shade of green?
21:58:08  <japj>I'm trying to see the differences in the icon
21:58:36  <piscisaureus__>japj you should download the old one and the new one
21:58:41  <piscisaureus__>and look at them in different sizes
21:59:18  <japj>I build a new binary and am looking at it from the explorer
21:59:59  <piscisaureus__>japj: new: https://github.com/joyent/node/raw/41d5fefc10ac238c74c6c47fe5f6eee728d205c8/src/res/node.ico
21:59:59  <piscisaureus__>japj: old: https://github.com/joyent/node/raw/v0.7.11/src/res/node.ico
22:00:27  * xaqquit (Remote host closed the connection)
22:01:48  <japj>strange.. maybe it doesnt rebuild the binary if only the icon changes? checking...
22:02:02  <japj>that definately looks different ;)
22:09:40  <isaacs>omg building an operating system takes a long ass time
22:10:06  * isaacswaiting to see if 0.7.12-RC3 works in a JPC build, which will be a huge hurdle to get over
22:10:09  <japj>isaacs: is it a javascript os? ;)
22:10:19  <isaacs>er, not JPC build, SDC.
22:10:26  <isaacs>Smart DataCenter, joyent's product
22:10:57  <japj>bah, I was hoping you were building a javascript os (which would also take a long time to build ;)
22:11:12  <japj>anyway, it's getting late for jokes... time to hit the bed
22:11:44  <japj>good luck on the release process of 0.7.12!
22:12:19  * voodootikigodjoined
22:12:27  * travis-cijoined
22:12:27  <travis-ci>[travis-ci] joyent/libuv#429 (master - 9dca9c3 : Bert Belder): The build was broken.
22:12:27  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/67090653df47...9dca9c38f64b
22:12:27  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1658297
22:12:27  * travis-cipart
22:13:35  * japjquit (Remote host closed the connection)
22:15:37  * travis-cijoined
22:15:37  <travis-ci>[travis-ci] joyent/libuv#430 (master - b496c12 : Bert Belder): The build was fixed.
22:15:37  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/9dca9c38f64b...b496c1221c38
22:15:37  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/1658304
22:15:37  * travis-cipart
22:19:41  <piscisaureus__>http://www.oracle.com/javaone/lad-en/session-presentations/clientside/24821-enok-1439095.pdf
22:19:44  <piscisaureus__>see page 27
22:22:22  * mikealquit (Quit: Leaving.)
22:25:56  <indutny>piscisaureus__: "reuse your Java middleware" meh
22:26:09  <piscisaureus__>I am not saying you should like it
22:26:34  <indutny>piscisaureus__: I like Maxine VM
22:26:54  <indutny>piscisaureus__: I think it's probably the most important thing that's happening with JVM now
22:30:58  <piscisaureus__>jvm, bah
22:31:29  <piscisaureus__>ok, ok, the jvm is actually good
22:32:49  <piscisaureus__>why are there 2 pummel tests for watch file?
22:33:29  <piscisaureus__>ah my
22:34:43  * AndreasMadsenquit (Remote host closed the connection)
22:38:54  <piscisaureus__>bnoordhuis: hey
22:39:04  <bnoordhuis>piscisaureus__: ho
22:39:15  <piscisaureus__>bnoordhuis: does fs.watchFile now actually use uv_stat ?
22:39:29  <bnoordhuis>piscisaureus__: no, i haven't landed that yet
22:39:36  <piscisaureus__>ah, that explains
22:40:41  <piscisaureus__>bnoordhuis: so, does it work. I noticed it is using ev_ref and ev_unref...
22:40:45  <indutny>a question
22:40:55  <indutny>how much event queues does uv has?
22:40:58  <indutny>s/has/have
22:40:59  <bnoordhuis>piscisaureus__: does what work?
22:41:05  <bnoordhuis>indexzero: as many as you want
22:41:07  <piscisaureus__>bnoordhuis: fs.watchFile on unix
22:41:33  <bnoordhuis>piscisaureus__: you probably misread a call to uv_unref
22:41:49  <bnoordhuis>indexzero: sorry, indutny
22:41:59  <bnoordhuis>autocomplete fail
22:42:14  <indexzero>:)
22:42:20  <piscisaureus__>bnoordhuis: https://github.com/joyent/node/blob/master/src/node_stat_watcher.cc#L102
22:44:16  <bnoordhuis>piscisaureus__: remember how i said i haven't landed it yet?
22:44:28  <bnoordhuis>what's currently in node is in various states of broken
22:44:32  * mikealjoined
22:44:38  <piscisaureus__>aha
22:44:52  <piscisaureus__>daar komt de aap uit de mauw
22:44:55  <piscisaureus__>*mouw
22:53:03  * toothrjoined
22:54:01  <CIA-108>node: Bert Belder master * r3795d77 / (2 files in 2 dirs): windows: `listen(pipe_name)` is not supported when running cluster - http://git.io/vYn0SQ
22:54:14  * mikealquit (Read error: Connection reset by peer)
22:54:25  * mikealjoined
22:55:31  * toothr-quit (Ping timeout: 248 seconds)
22:57:09  * paddybyersquit (Quit: paddybyers)
22:57:52  <CIA-108>node: Bert Belder master * r7a4dfb6 / test/simple/test-cluster-worker-kill.js : test-cluster-worker-kill: use SIGKILL instead of SIGHUP - http://git.io/m4Iapg
23:08:56  <CIA-108>node: Bert Belder master * r26b1191 / test/simple/test-domain.js : test-domain: fix the test to work on Windows - http://git.io/bTFI1w
23:22:11  * TheJHquit (Ping timeout: 248 seconds)
23:28:53  * mikealquit (Quit: Leaving.)
23:29:46  * sh1mmerquit (Ping timeout: 246 seconds)
23:35:54  <bnoordhuis>isaacs: add stat() based file watcher (Ben Noordhuis) <- that hasn't landed yet
23:36:26  <isaacs>bnoordhuis: oh, whoops. i saw it in the changelog.
23:36:39  <isaacs>bnoordhuis: oh, it's in libuv, but not node?
23:36:43  <bnoordhuis>yes
23:36:46  <isaacs>oh, ok
23:36:52  <isaacs>bnoordhuis: is it going to land in node soon?
23:37:18  <isaacs>i mean, it's not exactly incorrect. there's just no JS api for it.
23:37:27  <isaacs>but the libuv in node does have that
23:37:35  <bnoordhuis>yes
23:37:46  <isaacs>the changelog doesn't SAY it's exposed in any relevant way :)
23:37:53  <isaacs>if it gets in soon, that's fine.
23:37:59  <bnoordhuis>cool :)
23:40:04  <bnoordhuis>simple/test-eio-race fails pretty consistently on my vps...
23:42:21  <piscisaureus__>one test failure left
23:42:31  <piscisaureus__>an actual bug, as far as I can see
23:48:30  <isaacs>i really want to re-do the node blog.
23:48:35  <piscisaureus__>yes
23:48:39  <isaacs>like, just have some markdown dump into html, and upload it
23:48:47  <isaacs>wordpress is such bullshit.
23:48:51  <isaacs>and comments! jesus! why??
23:48:54  <piscisaureus__>also, fix the brown crap
23:49:02  <piscisaureus__>into nicer brown ...
23:49:02  <isaacs>we have TWO mailing lists and IRC channels!
23:49:14  <isaacs>well, if i do it, i'm gonna match the rest of the nodejs.org site
23:49:16  <piscisaureus__>if you really care you could just add a disqus box
23:49:20  <isaacs>actualy, maybe i'll do that right now
23:49:25  <piscisaureus__>zero pain
23:49:42  <isaacs>piscisaureus__: or, for even less than zero pain, i could REMOVE the disqus box we already have
23:49:54  <isaacs>piscisaureus__: oh, i guess we have wordpress.com comments.
23:49:59  <isaacs>whatever. jsut as bad.
23:50:12  <isaacs>people post bug reports for express against blog posts from last year.
23:50:15  <isaacs>it's just so terrible.
23:50:24  * mjr_joined
23:50:45  <isaacs>i don't think i've seen one useful comment on a blog post since 2004
23:51:09  <isaacs>email? sure. forums? definitely. even hacker news has its moments, though they are few and far between
23:51:17  <piscisaureus__>isaacs: maybe just add a "news" section to the site
23:51:29  <piscisaureus__>isaacs: I don't know, it is not even a blog
23:51:42  <piscisaureus__>isaacs: except for that one article dap wrote
23:51:48  <isaacs>we've had a few articles
23:51:52  <isaacs>mostly it's the release feed.
23:52:07  <isaacs>i've written some articles about random stuff. so has ryan.
23:52:08  <piscisaureus__>Node.js Blog
23:52:08  <piscisaureus__>Insight, perspective and events
23:52:20  <isaacs>writing blog posts is so tedious, though
23:52:33  <isaacs>if i'm gonna do that, i'm gonna do it on izs.me so that i can write about whatever, and not just node.
23:52:57  <isaacs>but, it is more accessible than NodeUp
23:53:02  <piscisaureus__>yeah
23:53:10  <piscisaureus__>also nodeup is all about portland incrowd
23:53:17  <isaacs>portland incrowd?
23:53:29  <isaacs>you mean oakland?
23:53:32  <piscisaureus__>er, yeah
23:53:34  <piscisaureus__>that's what I meant
23:53:44  <piscisaureus__>where you live and mikeal
23:53:44  <isaacs>a few of the regulars are from other places.
23:54:00  <isaacs>but yeah, me, mikeal and substack are the most frequently available
23:54:12  <isaacs>cjm is actually in new york, i think. the other folks are all scattered about.
23:55:04  <piscisaureus__>anyway
23:55:13  <piscisaureus__>what I was going to say... the current state of the blog is not so nice
23:55:17  <isaacs>anyway, yes, blog will be just a bunch of dated markdowns in doc/
23:55:20  <piscisaureus__>there are actually interesting articles there
23:55:23  <isaacs>yes.
23:55:29  <piscisaureus__>but they get pushed out by release announcements
23:55:35  <piscisaureus__>you should do something about that
23:55:39  <isaacs>hm.
23:55:48  <isaacs>maybe it should just show each most recent release announcement?
23:55:49  <piscisaureus__>I mean, who cares about 7.1 .. 7.11 now that we have 7.12
23:55:54  <isaacs>right
23:55:55  <piscisaureus__>yes
23:56:06  <piscisaureus__>or just the latest release on every branch
23:56:10  <isaacs>in fact, why do I even HAVE to post a release announcement?
23:56:12  <piscisaureus__>and the rest can be hidden deep
23:56:23  <isaacs>we have the changelog in the source tree already.
23:56:27  <piscisaureus__>well
23:56:31  <piscisaureus__>not everyone reads that stuff
23:56:39  <isaacs>the blog builder should just build the latest ones.
23:56:41  <piscisaureus__>it is good to prominently state it if there's a new version
23:56:48  <isaacs>yes.
23:57:00  <piscisaureus__>but it doesn't have to include all the changes, actually, esp. on unstable branches
23:57:02  <isaacs>but only the most recent for each branch should be shown
23:57:06  <piscisaureus__>yes
23:57:34  <isaacs>so you'd see 0.7.12, dap's thing, 0.6.19, bunch of articles, 0.4.12, etc.
23:57:50  <piscisaureus__>yeah
23:58:02  <isaacs>also: 0.even should be considered the continuation of 0.even-1
23:58:13  <isaacs>0.5 is meaningless once 0.6 lands
23:58:26  <piscisaureus__>yes indeed
23:58:36  <piscisaureus__>also, people don't care that much about unstable releases
23:58:39  <isaacs>i think doing just that, pagination, and putting it in the same template as the rest of the site, would be awesome.
23:58:47  <piscisaureus__>yeah
23:58:55  <isaacs>piscisaureus__: i think we managed to get a lot of people interested in 0.7 by teh end.
23:59:03  <piscisaureus__>because 0.7.x is good
23:59:10  <isaacs>if our release trains were shorter, it'd help a lot with that.
23:59:10  <piscisaureus__>I actually installed it on my machine now
23:59:14  <isaacs>hah
23:59:27  <isaacs>yeah, at this point, i get annoyed that 0.6 is so unstable.
23:59:30  <isaacs>that's a good sign.
23:59:31  <piscisaureus__>Never did that with the 0.5 series
23:59:44  <piscisaureus__>Now I have to fix one actual refcount bug in libuv
23:59:45  <piscisaureus__>hah
23:59:52  <isaacs>awesome.