03:33:40  <piscisaureus_>had to restart slurp
03:40:21  <tjfontaine>heh
03:44:36  <piscisaureus_>k.goodbye, really now :-)
03:44:39  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
03:47:44  * ericktjoined
04:06:35  * mikealjoined
04:06:53  * mikealquit (Client Quit)
04:27:49  * avalanche123|hjoined
04:47:02  * kristatejoined
04:56:11  * avalanche123|hquit (Quit: Computer has gone to sleep.)
04:59:28  * AvianFluquit (Ping timeout: 265 seconds)
05:03:07  * AvianFlujoined
05:10:54  * AvianFluquit (Read error: Connection reset by peer)
05:11:09  * AvianFlujoined
05:20:48  * AvianFluquit (Ping timeout: 248 seconds)
05:26:39  * AvianFlujoined
05:48:03  * avalanche123|hjoined
05:54:36  * avalanche123|hquit (Quit: Computer has gone to sleep.)
05:57:02  * avalanche123|hjoined
06:08:17  * AvianFluquit (Quit: AvianFlu)
06:18:07  * stagasjoined
06:32:48  * V1joined
06:35:48  * V1changed nick to `3rdEden
06:37:50  * avalanche123|hquit (Quit: Computer has gone to sleep.)
06:50:13  * rendarjoined
06:50:45  * stephankquit (Quit: *Poof!*)
07:02:42  * stephankjoined
07:18:02  <indutny>hoya
07:37:06  * AvianFlujoined
07:40:51  * stagasquit (Quit: ChatZilla 0.9.89-rdmsoft [XULRunner 1.9.0.17/2009122204])
07:59:45  * AvianFluquit (Quit: AvianFlu)
09:02:24  * Benviequit (*.net *.split)
09:02:41  * rvaggquit (*.net *.split)
09:02:49  * mitsuhikoquit (*.net *.split)
09:02:52  * Benviejoined
09:03:05  * rvaggjoined
09:05:20  * mitsuhikojoined
09:06:25  * russell_hquit (*.net *.split)
09:06:38  * russell_hjoined
09:26:56  * mmalecki[zzz]joined
09:52:57  * mmalecki[zzz]changed nick to mmalecki
09:53:13  * mmaleckiquit (Quit: leaving)
09:53:32  * mmaleckijoined
10:14:11  * kristatequit (Ping timeout: 244 seconds)
10:17:21  * Raltjoined
11:07:02  * abraxas_quit (Remote host closed the connection)
11:48:37  * kristatejoined
12:01:42  * Raltquit (Read error: Connection reset by peer)
12:08:02  * bnoordhuisjoined
12:10:25  * sergimjoined
12:18:55  <indutny>bnoordhuis: hoya
12:26:26  <mmalecki>bnoordhuis: hey, I was supposed to ping you yesterday, but got shangai'd into refactor
12:26:46  <mmalecki>indutny: did you solve those availability problems?
12:27:05  <indutny>mmalecki: nope, not yet
12:27:11  <indutny>looking into it
12:27:42  <mmalecki>okay :)
12:27:49  <mmalecki>when it's fixed, let me know
12:28:06  <mmalecki>I'm investigating http-proxy performance regression when using `Stream#pipe` btw
12:28:15  <mmalecki>gotta do some cool flamegraphs yo
12:30:23  <indutny>cool
12:34:59  <mmalecki>indutny: I'm getting around 10 req/s slowdown with Stream#pipe, wondering what could it be
12:40:18  <bnoordhuis>indutny: heya
12:41:53  <CIA-131>node: Adam Blackburn v0.8 * rd5e9895 / doc/api/_toc.markdown : doc: put API table of contents in alphabetical order - http://git.io/sjgMjA
12:43:47  * sergimquit (Quit: Computer has gone to sleep.)
12:45:01  <CIA-131>node: Shigeki Ohtsu master * ra1ba29d / test/simple/test-http-client-timeout-agent.js : test: fix failing due to too early timeout - http://git.io/E3xvww
12:50:19  * TheJHjoined
13:04:28  <mmalecki>actually, what I said about pipe refers to 0.8 branch only...
13:04:47  <indutny>bnoordhuis: have you got any luck yesterday?
13:04:50  <indutny>with tlsnappy
13:05:20  <mmalecki>*and* I'm getting some failed transactions on master...
13:09:31  * sergimjoined
13:12:03  <bnoordhuis>indutny: no, i did some other things when you went offline
13:12:18  * piscisaureus_joined
13:15:12  <indutny>bnoordhuis: np
13:15:16  <indutny>thank you anyway
13:15:51  <piscisaureus_>bnoordhuis: https://github.com/joyent/node/compare/openssl-exports <-- works on linux and windows
13:17:49  <bnoordhuis>piscisaureus_: openssl/crypto/dso/dso_vms.c ?
13:18:06  <bnoordhuis>longing for the 90s?
13:20:02  <piscisaureus_>bnoordhuis: I think that is an empty function
13:20:19  <piscisaureus_>bnoordhuis: I set it up such that it builds the same stuff on all platforms
13:20:38  <piscisaureus_>bnoordhuis: very close to what "./config; make" would build
13:20:47  <piscisaureus_>bnoordhuis: from here we can selectively disable stuff
13:21:05  <piscisaureus_>but dso_vms.c really doesn't matter
13:22:01  <piscisaureus_>bnoordhuis: I think we can indeed drop camellia and idea for example
13:23:20  * AvianFlujoined
13:24:32  <piscisaureus_>bnoordhuis: indutny: I think we should reorganize the openssl config headers so people can include the same file on all platforms
13:25:13  <piscisaureus_>this should really not be necessary -> https://github.com/indutny/tlsnappy/blob/d90e7f3e82d762a57ec47d7a879bc16d8210e510/binding.gyp#L16-26
13:25:23  <bnoordhuis>sergim: ping
13:25:43  <sergim>bnoordhuis: hey
13:25:54  <piscisaureus_>ha, sergim
13:26:05  <bnoordhuis>sergim: still looking for a speaker are you?
13:26:45  <sergim>bnoordhuis: yeah, I replied to you before
13:27:11  <bnoordhuis>sergim: ah, did you? i don't have any messages
13:33:44  <indutny>piscisaureus_: agreed
13:36:26  <piscisaureus_>indutny: good
13:36:35  <piscisaureus_>indutny: I will go hack on this
13:38:01  <bnoordhuis>piscisaureus_: a single header with #if defined(__i386__) #elif defined(__x86_64__) etc?
13:42:42  <piscisaureus_>bnoordhuis: yes that was the plan
13:42:44  <piscisaureus_>:-)
13:43:11  <piscisaureus_>bnoordhuis: we have 5 config files now but the differences are quite minimal
13:45:15  * ericktquit (Quit: erickt)
13:48:30  <bnoordhuis>piscisaureus_: sound good
14:32:47  * AvianFluquit (Quit: AvianFlu)
14:33:12  * AvianFlujoined
14:37:55  <indutny>bnoordhuis: so, when this stuff with SSL_read happens
14:38:04  <indutny>bnoordhuis: I see SSL errors on my side too
14:38:20  <bnoordhuis>indutny: sorry, i'm working on a presentation
14:38:32  <indutny>oh
14:38:33  <indutny>np
15:08:36  * ericktjoined
15:10:46  <CIA-131>node: isaacs v0.8.10-release * rc2879c9 / (AUTHORS ChangeLog src/node_version.h): wip - http://git.io/0H89Rw
15:12:46  * stagasjoined
15:16:24  * `3rdEdenquit (Remote host closed the connection)
15:22:16  <isaacs>0.8.10 coming very soon. anyone care to review changelog? ^
15:22:23  <isaacs>or anything else that needs to get in?
15:25:02  <mmalecki>isaacs: are indutny's SSL changes in?
15:25:39  <mmalecki>I think they were supposed to go in in this release
15:26:43  <isaacs>ah, no, openssl-asm is not in
15:26:56  <isaacs>piscisaureus_: ping
15:27:03  <isaacs>piscisaureus_: can we merge openssl-asm into v0.8?
15:27:43  <indutny>oh
15:27:58  <indutny>isaacs: there're also slab allocator
15:28:03  <indutny>and I'm not sure about openssl-asm
15:28:11  <indutny>but better ask bertje
15:28:15  <isaacs>yep
15:28:24  <isaacs>indutny: how big of an effect is the slab allocator?
15:28:26  <indutny>last time he told me that he's still fighting with it
15:28:32  <isaacs>hm, ok
15:28:37  <indutny>isaacs: you want numbers? :)
15:28:39  <isaacs>i'm running it on npmjs.org, seems to be fine.
15:28:39  <indutny>isaacs: ask mmalecki
15:28:45  <mmalecki>it'd mean tons for us if you could get that in
15:29:01  <isaacs>mmalecki: right, but how much is due to the the SA, and how much due to asm in the build?
15:29:29  <bnoordhuis>isaacs: changelog lgtm
15:29:54  * bentkusjoined
15:30:20  <mmalecki>isaacs: I don't remember exact numbers, unfortunately, but whole thing gave us ~3 x speedup in https benchmark
15:30:47  <mmalecki>I remember however that SA gave me a noticable performance bump
15:30:48  <isaacs>bnoordhuis: we ought to get openssl-asm in this as well.
15:30:58  <isaacs>bnoordhuis: how do you feel about the tls slab allocator?
15:31:07  <isaacs>bnoordhuis: f210530f46e8ddbd9e7cc0d0c37778888c27f526
15:31:25  <isaacs>https://github.com/joyent/node/commit/f210530f46e8ddbd9e7cc0d0c37778888c27f526
15:31:37  <mmalecki>this has been running on production load balancers for some time now with no crashes
15:33:04  <bnoordhuis>isaacs: slab allocator seems harmless enough
15:33:12  <isaacs>bnoordhuis: k.
15:33:15  <isaacs>i'm gonna pull it in.
15:33:18  <bnoordhuis>but merging the openssl-asm branch hours before the release is probably less than wise
15:33:26  * AndreasMadsenjoined
15:33:31  <isaacs>well... i've been running it, and so has nodejitsu.
15:33:45  <isaacs>it's not as if it hasn't seen the light of day
15:34:13  <isaacs>of course, we're both on similar architectures, etc.
15:34:13  <bnoordhuis>okay, merge it
15:34:35  <isaacs>i'd like to know how nervous it makes bertje :)
15:34:43  <indutny>isaacs: nodejitsu was running my patch, I suppose
15:34:53  <isaacs>indutny: ah, yes.
15:34:54  <indutny>mmalecki: have you pulled openssl-asm branch?
15:35:19  <mmalecki>indutny: let me double check, but I'm 90 % sure we have
15:36:42  * stagasquit (Quit: ChatZilla 0.9.89-rdmsoft [XULRunner 1.9.0.17/2009122204])
15:36:50  <mmalecki>* openssl-asm 4634126 openssl: use optimized asm code on x86 and x64
15:36:53  <mmalecki>that's it, correct?
15:38:59  <CIA-131>node: isaacs v0.8.10-release * r83d39c8 / (556 files in 108 dirs): npm: upgrade to 1.1.62 - http://git.io/3d994g
15:39:00  <CIA-131>node: Fedor Indutny v0.8.10-release * r7651228 / lib/tls.js : tls: use slab allocator - http://git.io/9LIgHw
15:39:00  <CIA-131>node: isaacs v0.8.10-release * r420b8dc / (AUTHORS ChangeLog src/node_version.h): wip - http://git.io/npOZbw
15:39:07  <isaacs>mmalecki: there are quite a few more commits in the openssl-asm branch
15:39:33  <indutny>we need bert for this
15:39:34  <tjfontaine>that mmap v8 fix went in .9 or is it going in .10?
15:39:36  <indutny>piscisaureus_: ^^
15:39:41  <indutny>oh, btw
15:39:51  <indutny>releasing without it seems to be a bad idea too
15:39:56  <indutny>brb
15:39:57  <mmalecki>hmm
15:40:02  <mmalecki>last one I'm seeing has this SHA
15:40:06  <mmalecki>https://github.com/joyent/node/tree/openssl-asm
15:40:18  <mmalecki>https://github.com/joyent/node/commit/46341267496c43955e01a35e9955783f903bdbfb
15:40:52  <isaacs>tjfontaine: it's in 0.8.10 as well
15:41:06  <tjfontaine>isaacs: right I just wasn't sure if it made .9
15:41:24  <isaacs>tjfontaine: yeah, it's not in 0.8.9
15:41:33  <tjfontaine>k
15:41:52  <CIA-131>node: isaacs v0.8.10-release * r48598c7 / (AUTHORS ChangeLog src/node_version.h): wip (+9 more commits...) - http://git.io/0LF0jA
15:42:25  <isaacs>ok... let's try it. it's easy enough to prune off the merge commit if it's not good :)
15:42:43  <isaacs>or release 0.8.11 tomorrow witout openssl-asm ;)
15:42:52  <mmalecki>isaacs: ++
15:43:37  <tjfontaine>I know piscisaureus_ was worried slightly about a benchmark regression in AES on windows with openssl-asm
15:43:38  <CoverSlide>gogogo
15:44:00  <tjfontaine>but I don't think it was too drastic
15:44:29  <isaacs>also, i'm totally zeroing in on this weird login/password/signup bug in npmjs.org
15:44:41  <isaacs>i can reproduce it now, but it still looks impossible.
15:45:17  <tjfontaine>hah, 2012.09.15 time machine time
15:45:24  <tjfontaine>isaacs: what is the behavior?
15:46:59  <isaacs>tjfontaine: tjfontaine thanks :)
15:47:02  <CIA-131>node: isaacs v0.8.10-release * r8451e5e / (AUTHORS ChangeLog src/node_version.h): wip - http://git.io/NzJ1bQ
15:47:13  <tjfontaine>isaacs: I can't claim full credit, mscdex got it
15:47:14  <isaacs>tjfontaine: the behavior is that if someone resets their password, then password stuff is broken on that worer.
15:47:24  <tjfontaine>oh, interesting
15:47:29  <isaacs>tjfontaine: so there's some state changing in that flow that makes signups fail.
15:48:05  <isaacs>tjfontaine: i can reproduce it in dev, but in order to reproduce it, you have to have a way to send email, so that you can actually run through the password reset.
15:48:25  <piscisaureus_>isaacs: i was pretty sure that I landed the openssl-asm branch in 0.8
15:48:28  <isaacs>tjfontaine: so i'm going to change around that bit, so that if NODE_ENV=='dev' and there's no email sending configs, then it'll just print out on the secreen what the email would be.
15:48:29  <piscisaureus_>isaacs: did it get forced out?
15:48:33  <isaacs>piscisaureus_: not sure.
15:48:36  <isaacs>but i don't see it
15:49:41  <piscisaureus_>isaacs: let's land it okay?
15:49:59  <isaacs>piscisaureus_: k, i'm on it.
15:50:04  <isaacs>piscisaureus_: landed in v0.8.10-release
15:50:23  <isaacs>should probably mention it on the changelog :)
15:50:37  <piscisaureus_>isaacs: I think I would just rebase instead of merge
15:50:40  <piscisaureus_>but whatever
15:51:16  <piscisaureus_>isaacs: enter the magic words "git rebase -f 7651228ab2149c252eaa3ed"
15:51:58  * sergimquit (Quit: Textual IRC Client: http://www.textualapp.com/)
15:53:35  <isaacs>inlined.
15:53:38  <CIA-131>node: isaacs v0.8.10-release * rd9fc7b6 / (AUTHORS ChangeLog src/node_version.h): wip (+8 more commits...) - http://git.io/adR32Q
15:54:12  <isaacs>oh, forgot the changelog
15:54:14  <CIA-131>node: isaacs v0.8.10-release * r8cb3c2e / (AUTHORS ChangeLog src/node_version.h): wip - http://git.io/HPGPCw
15:54:17  <isaacs>there it is
15:54:23  * dapjoined
15:54:50  <piscisaureus_>yay
15:54:55  <isaacs>yay indeed!
15:55:01  <bentkus>why bash, when you can have node
15:55:16  <isaacs>bentkus: because node's child process api is stupidly verbose.
15:55:35  <isaacs>bentkus: and you have to use operators to concatenate strings in javascript, which is dumb, when you could just put them next to each other to concatenate them.
15:55:40  * piscisaureus_bbl
15:55:47  <isaacs>bentkus: also, heredoc!
15:56:38  <creationix>piscisaureus_, I updated luvit to the latest libuv and my issues seems to have gone away
15:58:23  <mmalecki>node is pretty sweet for system administration tho
15:58:49  <mmalecki>it gets even better when you use node to execute bash scripts
15:58:58  <mmalecki>on thousands of machines :)
16:00:16  * piscisaureus_quit (Ping timeout: 248 seconds)
16:04:10  * TooTallNatejoined
16:06:20  <bentkus>isaacs: THIS IS MADNESS, BUT I AM USING NODE INSTEAD OF BASH
16:06:27  <bentkus>or at least i'm trying
16:09:55  <bentkus>because of node.js and npm js has become my favourite scripting language
16:10:16  <isaacs>bentkus: that's the idea :)
16:10:29  <isaacs>ok, tests are pretty acceptable on smartos and linux
16:10:31  <isaacs>and osx
16:10:48  <isaacs>time to commute, i'll proceed with the rest of the build in an hour or so
16:17:07  * CIA-131quit (Ping timeout: 272 seconds)
16:29:52  <tjfontaine>TooTallNate: hmm, is there a way for ref to report a nicer error if I mistakenly pass `0` when the type is pointer and it's expecting `null`, right now the error comes back as "TypeError: writePointer: Buffer instance expected as third argument"
16:30:39  <TooTallNate>tjfontaine: so you'd want it to be more ffi-centric there?
16:31:00  <TooTallNate>what do you want the error to say?
16:31:12  <bentkus>my god, isn't there an easy to use liquid template engine for node.js :/
16:31:29  <tjfontaine>TooTallNate: I'm not sure, perhaps an assert, I need to check the stack more closely to see where that would fit
16:33:56  <TooTallNate>tjfontaine: granted, it would be better if node-ffi threw the error, "vlc_open: number expected as 1st argument"
16:34:18  <TooTallNate>but with the type system, there's not really a "check" kind of thing
16:34:26  <TooTallNate>so doing that is a little hard
16:35:12  <tjfontaine>ya I'm trying to determine if I want it to be that high level, or handling it in writePointer is better
16:35:14  <bentkus>ok, I am cursed or something, installed liquid, require('liquid') doesnt work
16:35:41  <CoverSlide>you sure it's not 'liquid-node
16:36:06  <CoverSlide>'oh there is a liquid, i searched google before i searched npm
16:36:13  <TooTallNate>tjfontaine: are you expecting to pass 0 and write a null pointer?
16:36:28  <bentkus>liquid-node didn't work at all
16:36:32  <bentkus>the api was batshit crazy
16:37:10  <isaacs>bentkus: "doesn't work" is very vauge. you should probably ask in #node.js, and gist the actual code and output.
16:37:21  <CoverSlide>btw I think module questions should go in #Node.js
16:37:22  <isaacs>s/vauge/vague/
16:37:25  <tjfontaine>TooTallNate: I'm fine with 0 not working, it's just that the exception saying it was expecting the third arg to be a buffer is bit confusing because I thought initially the third argument of the ffi function was wrong
16:37:53  <TooTallNate>right, right, that's happened to me i think too :)
16:38:01  <TooTallNate>the being confused part
16:38:23  <bentkus>http://tinyurl.com/c98xzqs
16:38:26  <bentkus>here is my gist
16:38:52  <tjfontaine>I'm thinking a generic type checking would be ideal, but it should probably be switchable like buffer asserts so people who don't want it can live without it
16:39:28  <tjfontaine>anyway, I'm going to chew on it, solved the problem and have moved on :P
16:41:25  * piscisaureus_joined
16:42:51  * brsonjoined
16:45:33  <piscisaureus_>bnoordhuis: hey
16:46:03  <mmalecki>piscisaureus_: what's your flight number again?
16:46:18  <mmalecki>my flight got rerouted, now we might have ended up on the same flight XD
16:46:40  <piscisaureus_>mmalecki: TAP
16:46:43  <piscisaureus_>663
16:47:28  <mmalecki>nope, no luck, mine is TP 6720
16:48:55  * ericktquit (Quit: erickt)
16:54:38  * ericktjoined
16:55:42  * brsonquit (Ping timeout: 264 seconds)
17:00:19  * brsonjoined
17:03:02  <indutny>so
17:03:05  <indutny>isaacs: released?
17:20:29  <bentkus>i just discoevered that i dont like eating ice anymore
17:24:41  * V1joined
17:28:07  <piscisaureus_>indutny: not yet, later today
17:32:32  * Raltjoined
17:32:38  * AndreasMadsenquit (Read error: Connection reset by peer)
17:36:53  * bentkusquit (Ping timeout: 256 seconds)
17:46:26  * V1changed nick to `3rdEden
17:54:42  <isaacs>indutny: not yet.
18:06:28  * Raltquit (Remote host closed the connection)
18:10:32  * mikealjoined
18:11:07  <isaacs>passive-agressive commit messages ftw.
18:14:31  <indutny>:)
18:15:00  <indutny>isaacs: https://github.com/isaacs/node/commit/411d46087f4027db69a5906985109a2b16f9ff52#commitcomment-1903295
18:27:10  <piscisaureus_>I though nodesummit was months ago
18:27:16  <piscisaureus_>why does it keep showing up in my timeline
18:29:56  * mikealquit (Quit: Leaving.)
18:34:09  <indutny>piscisaureus_: it was yesterday?
18:34:34  <indutny>piscisaureus_: http://nodesummit.com/
18:34:43  <piscisaureus_>oh, they had another one
18:35:27  <`3rdEden>piscisaureus_: indutny: it wasn't the same as the last nodesummit
18:35:48  <piscisaureus_>`3rdEden: what do you mean?
18:36:02  <indutny>I don't understand it too
18:36:03  <`3rdEden>This nodesummit was part of a bigger conference
18:36:20  <`3rdEden>aka Dataweek
18:36:22  <piscisaureus_>ah, right
18:36:36  <`3rdEden>(as far as I understand it)
18:37:51  <`3rdEden>This version was also a half day instead of the 2 days of the last nodesummit
18:42:50  <piscisaureus_>bnoordhuis: https://docs.google.com/a/c9.io/document/d/1BQqShTmo8LG9IL8Iuz0-xLe4umAqfbGuMofdTwteBOI/edit
18:46:46  * creationixpart
18:47:48  * creationixjoined
18:48:13  <isaacs>piscisaureus_: https://gist.github.com/3783666
18:48:32  <isaacs>piscisaureus_: building x64 on SmartOS is failing because of an openssl thing. any ideas before i dig into it?
18:48:40  <piscisaureus_>isaacs: master? or 0.8?
18:48:52  * `3rdEdenquit (Quit: derp)
18:49:10  <isaacs>piscisaureus_: v0.8.10-release
18:49:31  <piscisaureus_>what's the ip of the smartos64 drone?
18:49:59  <isaacs>165.225.128.184
18:50:37  <isaacs>looks like both wp_block.o and wp-x86_64.o define it
18:50:44  <piscisaureus_>ah
18:50:50  <piscisaureus_>wp_block should be removed then
18:50:57  <piscisaureus_>isaacs: lemme check to be sure
18:51:39  <isaacs>k
18:51:46  * AndreasMadsenjoined
18:58:14  <piscisaureus_>isaacs: testing patch
18:58:23  <piscisaureus_>isaacs: have to be careful to not break x86 ...
19:00:31  <TooTallNate>i'll test arm ;)
19:00:45  <piscisaureus_>oh, it won't affect arm :-)
19:01:03  <tjfontaine>famous last words
19:01:06  <piscisaureus_>but testing arm always +1
19:01:11  <TooTallNate>piscisaureus_: did you land the openssl export stuff?
19:01:16  <piscisaureus_>no
19:01:19  <piscisaureus_>not yet
19:01:25  <piscisaureus_>it probably won't work on mac
19:01:27  <TooTallNate>ok, that's what i need to test on arm
19:01:33  <TooTallNate>piscisaureus_: at all?
19:01:37  <piscisaureus_>well
19:01:38  <tjfontaine>wouldn't that count as an abi change? :)
19:01:51  <TooTallNate>tjfontaine: it's fine for master imo
19:02:00  <piscisaureus_>tjfontaine: the openssl export stuff is going into master
19:02:06  <piscisaureus_>TooTallNate: it will work on mac but it doesn't atm
19:02:06  <tjfontaine>right ok
19:02:13  <piscisaureus_>also the config has to be cleaned up
19:02:18  <TooTallNate>piscisaureus_: ok kewl
19:02:56  <piscisaureus_>isaacs: how do I compile node for an x64 target?
19:03:10  <piscisaureus_>isaacs: the build script says it's a 32 bit machine
19:03:12  <piscisaureus_>(on smartos)
19:03:18  * mjr_joined
19:03:27  <TooTallNate>piscisaureus_: --dest-cpu=x64
19:03:50  <piscisaureus_>so lame
19:04:01  <piscisaureus_>why not fix uname?
19:04:20  <TooTallNate>well ./configure sniffs the gcc defines
19:04:33  <TooTallNate>and solaris' gcc builds for 32-bit by default apparently
19:04:45  <TooTallNate>unless you pass -m64
19:04:46  <piscisaureus_>I have to admit that on windows the same happens
19:06:08  <isaacs>DESTCPU=x64 make -j32
19:06:11  <isaacs>piscisaureus_: ^
19:06:12  * Raltjoined
19:09:02  <piscisaureus_>scheisse
19:09:17  <piscisaureus_>now I get undefined symbol "fcrypt_body"
19:10:17  <piscisaureus_>very odd
19:11:37  <piscisaureus_>right
19:11:54  <piscisaureus_>the sunos linker isn't braindead, it's just zealous
19:14:44  <mjr_>Solaris 32/64 linking issues make me sad. So much lost time because of that.
19:16:07  <piscisaureus_>mjr_: actually it's my fault
19:16:41  <piscisaureus_>mjr_: but many linkers just shrug if you reference non-existing symbols from functions that are not used themselves.
19:17:03  <piscisaureus_>[[email protected] ~/node]$ ./node
19:17:03  <piscisaureus_>> process.arch
19:17:03  <piscisaureus_>'x64'
19:17:31  <mjr_>Speaking of 64 bit, we found that 64 bit node uses quite a bit more memory than 32 bit node, and is somewhat slower at most things.
19:17:38  <mjr_>Perhaps it's better at openssl though.
19:20:57  <mmalecki>mjr_: it is noticably better at proxying https traffic
19:21:11  <mjr_>On SmartOS, or everywhere?
19:21:20  <mmalecki>smartos
19:21:50  <mjr_>BTW, we are about to roll out stud to see if that makes our lives better. I patched node's http server to parse the haproxy PROXY protocol line and set the remote IP correctly.
19:22:07  <mjr_>Currently using nginx, but nginx doesn't support chunked request bodies.
19:22:09  <piscisaureus_>mjr_: it's somewhat slower indeed. I never noticed anything shocking about memory usage, but in theory depends on what your're using your memory for
19:22:49  <piscisaureus_>mjr_: x64 is significantly better at openssl.
19:23:17  <piscisaureus_>mjr_: but it may change with 8.10, because the asm patches have a big impact on x86 and small impact on x64
19:23:38  <piscisaureus_>TooTallNate: cross compiling guru, how do I build node-x86 on 64 bit ubuntu
19:23:50  <piscisaureus_>TooTallNate: i have ia32-libs-multiarch installed
19:24:00  <TooTallNate>piscisaureus_: ./configure --dest-cpu=ia32 :D
19:24:02  <piscisaureus_>TooTallNate: but it complains:
19:24:03  <piscisaureus_>In file included from ../deps/http_parser/http_parser.c:24:0:
19:24:03  <piscisaureus_>../deps/http_parser/http_parser.h:30:23: fatal error: sys/types.h: No such file or directory
19:24:16  <piscisaureus_>that's what I did of course :-)
19:24:45  * stagasjoined
19:24:52  <TooTallNate>piscisaureus_: i'd guess you're missing some other package
19:25:01  <piscisaureus_>um ya
19:25:06  <TooTallNate>one sec, lemme try to find a list
19:26:50  <TooTallNate>piscisaureus_: ubuntu?
19:27:00  <piscisaureus_>TooTallNate: ya
19:27:26  * Raltquit (Remote host closed the connection)
19:28:29  <tjfontaine>maybe libc6-dev-i386
19:28:49  <tjfontaine>oh wait
19:29:11  <piscisaureus_>tjfontaine: well, it works
19:29:32  <TooTallNate>that worked?
19:29:49  <piscisaureus_>tjfontaine: gets much further atleast.
19:30:01  <piscisaureus_>tjfontaine: it's now compiling
19:30:24  <piscisaureus_>CIA is MIA again
19:30:35  <TooTallNate>piscisaureus_: ubuntudrone has installed:
19:30:38  <TooTallNate>[email protected]:~# dpkg --get-selections | egrep "32|multilib"
19:30:38  <TooTallNate>g++-4.4-multilib install
19:30:38  <TooTallNate>g++-multilib install
19:30:38  <TooTallNate>gcc-4.4-multilib install
19:30:38  <TooTallNate>gcc-multilib install
19:30:39  <TooTallNate>lib32gcc1 install
19:30:39  <TooTallNate>lib32gomp1 install
19:30:39  <TooTallNate>lib32ncurses5 install
19:30:40  <TooTallNate>lib32ncurses5-dev install
19:30:41  <TooTallNate>lib32ncursesw5 install
19:30:41  <TooTallNate>lib32ncursesw5-dev install
19:30:41  <TooTallNate>lib32stdc++6 install
19:30:50  <TooTallNate>ignore the "install" bit at the end
19:30:56  <piscisaureus_>tjfontaine: I also need stdc++ for ia32 iI think
19:31:05  <piscisaureus_>TooTallNate: thanks
19:31:11  <mmalecki>fedora works much better for this kind of stuff, you just install one package and it pulls everything
19:31:57  <TooTallNate>mmalecki: i think there's one of those for ubuntu, we're just being lame
19:32:18  <TooTallNate>[email protected]:~# dpkg --get-selections | egrep 386
19:32:18  <TooTallNate>libc6-dev-i386 install
19:32:18  <TooTallNate>libc6-i386 install
19:32:24  <piscisaureus_>I thought that ia32-libs-multiarch would do that
19:32:26  <TooTallNate>^ those too, so tjfontaine gets cookie :)
19:32:30  <mmalecki>heh :)
19:32:31  <piscisaureus_>well
19:32:34  <tjfontaine>heh
19:32:39  <piscisaureus_>I ran aground again
19:32:39  <TooTallNate>piscisaureus_: ya i thought that was the one
19:32:43  <mmalecki>bnoordhuis: ping
19:32:46  <piscisaureus_>now it complains about libstdc++ is not found
19:33:07  <piscisaureus_>usr/bin/ld: skipping incompatible /usr/lib/gcc/x86_64-linux-gnu/4.6/libstdc++.a when searching for -lstdc++
19:33:28  <TooTallNate>piscisaureus_: did you install lib32stdc++6?
19:33:50  <mmalecki>I'll never wrap my head around ubuntu naming scheme, I guess
19:34:44  <piscisaureus_>lib32stdc++6 is already the newest version.
19:35:11  <mmalecki>does dpkg have something like `yum provides`?
19:35:20  <tjfontaine>mmalecki: there's apt-file
19:35:37  <tjfontaine>but mostly I use packages. for that style searching
19:36:15  <tjfontaine>for instance to solve piscisaureus_'s initial issue I did http://packages.ubuntu.com/search?searchon=contents&keywords=sys%2Ftypes.h&mode=&suite=precise&arch=any
19:36:32  <piscisaureus_>apt-cache search
19:36:42  <tjfontaine>apt-cache search doesn't do by contents of file though
19:36:46  <tjfontaine>s/file/package/
19:36:48  <piscisaureus_>ah, right
19:37:35  <mmalecki>heh, using browser to search for packages
19:37:37  <tjfontaine>95% of my needs are satisified by apt-cache, the rest are handled by packages.(d.o|u.c) or apt-file
19:38:07  <tjfontaine>mmalecki: already indexed on the interwebs :)
19:38:22  <TooTallNate>piscisaureus_: do you have /usr/lib32/libstdc++.so.6?
19:38:59  <tjfontaine>oh neat " If you would prefer not to download the large files used by apt-file you can
19:39:02  <tjfontaine> run rapt-file, which calls a remote server to do the searches."
19:39:26  <piscisaureus_>][email protected]:~/node$ file $(realpath /usr/lib32/libstdc++.so.6)
19:39:26  <piscisaureus_>/usr/lib32/libstdc++.so.6.0.16: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, BuildID[sha1]=0x45c8ceb195067642635eddecce8a0b39308d9812, stripped
19:40:03  <piscisaureus_>TooTallNate: I think the problem is that ld is looking in the wrong dir
19:40:21  <mmalecki>piscisaureus_: LIBRARY_PATH=/usr/lib32 ?
19:40:37  <mmalecki>try and export that, might accidentally work XD
19:40:47  <tjfontaine>LD_LIBRARY_PATH?
19:40:57  <piscisaureus_>no
19:41:07  * Raltjoined
19:41:31  <piscisaureus_>. /usr/lib32/libstdc++.so doesn't exist
19:41:39  <piscisaureus_>so maybe I just create a symlink
19:46:46  <TooTallNate>piscisaureus_: try it :)
19:47:23  <TooTallNate>piscisaureus_: also, try copy and pasting the link command that's failing, and running it manually
19:47:29  <TooTallNate>with -V as the first argument
19:48:13  <piscisaureus_>g++: error: unrecognized option ā€˜-Vā€™
19:48:16  <piscisaureus_>ENOLUCK
19:48:28  <piscisaureus_>small v works :-)
19:49:25  <TooTallNate>oh that's it :p
19:49:53  * mmaleckiquit (Quit: Reconnecting)
19:50:31  <TooTallNate>piscisaureus_: are you able to see the paths it's trying to use?
19:50:35  * mmaleckijoined
19:50:37  <piscisaureus_>yeah
19:51:46  <piscisaureus_>g++-4.6-multilib
19:51:51  <piscisaureus_>^-- tjfontaine TooTallNate
19:52:23  <TooTallNate>piscisaureus_: that fixed it?
19:52:27  <piscisaureus_>yup
19:52:31  <TooTallNate>niiceee
19:52:42  <piscisaureus_>thanks guys
19:54:44  <tjfontaine>oh it only deps on gcc-multilib
19:54:45  <tjfontaine>interesting
19:56:12  <piscisaureus_>you need to find what depends on it, not the other way around
19:56:55  <tjfontaine>hmm?
19:57:34  <tjfontaine>I should have realized after seeing:
19:57:34  <tjfontaine>The following NEW packages will be installed:
19:57:35  <tjfontaine> gcc-4.6-multilib gcc-multilib lib32gcc1 lib32gomp1 lib32quadmath0 libc6-dev-i386 libc6-i386
19:57:42  <tjfontaine>that you'd need g++-multilib as well
19:58:04  <piscisaureus_>isaacs: the openssl-fix isaacs: https://github.com/joyent/node/commit/openssl-fix
19:58:49  <piscisaureus_>isaacs: tested everything except mac and arm
19:59:42  <TooTallNate>arm compiled fine before the openssl-fix branch
20:00:15  <indutny>mjr_: heya! are you going to give a try node 0.8.x after release?
20:00:41  <mmalecki>indutny: if that branch is ready I can deploy it to few of LBs
20:00:49  <mjr_>indutny: yeah, you bet. We were waiting for that mmap/bad_alloc change to land, which it just recently did.
20:01:27  <TooTallNate>mjr_: damn, you guys are still on v0.6?
20:01:35  <mjr_>TooTallNate: yeah, sadface.
20:01:53  <mjr_>Normally we are always running the latest, but this bad_alloc thing killed us.
20:04:57  <mmalecki>isaacs: any ETA on the actual release? not sure if I should take a nap or not.
20:11:50  <indutny>sadface
20:11:51  <indutny>true
20:12:07  <indutny>I've almost finished my secret project
20:12:12  <indutny>not sure it it'll boom
20:12:18  <indutny>but i'll definitely bam for someone
20:12:28  <piscisaureus_>isaacs is hanging out with CIA-141
20:14:18  <indutny>k
20:14:19  <indutny>time to sleep
20:14:20  <indutny>ttyl
20:16:00  <mmalecki>night Fedor
20:16:09  <piscisaureus_>later, indutny
20:16:17  * AndreasMadsenquit (Remote host closed the connection)
20:23:44  * brsonquit (Ping timeout: 248 seconds)
20:32:29  * Raltquit (Remote host closed the connection)
20:34:23  <bnoordhuis>back
20:34:53  <bnoordhuis>piscisaureus_: you rang?
20:34:54  <bnoordhuis>mmalecki: pong
20:35:09  <piscisaureus_>bnoordhuis: we have to do this stuff for jos
20:35:14  <piscisaureus_>bnoordhuis: are you coming to 020 tomorrow?
20:35:17  <bnoordhuis>piscisaureus_: yes
20:35:47  <piscisaureus_>bnoordhuis: ok, good. I think it has to be finished by the end of the month, but I am gone after tomorrow. So we either finish it then, or you're on your own :-)
20:35:59  <bnoordhuis>piscisaureus_: we'll do it tomorrow
20:36:08  <bnoordhuis>i'll go an finish my talk now
20:36:17  <piscisaureus_>bnoordhuis: good luck
20:36:28  <bnoordhuis>thanks
20:36:38  <bnoordhuis>it's not hard, just talk interminably until the 30 minutes are up
20:36:49  <bnoordhuis>anyone can do it
20:37:01  <piscisaureus_>bnoordhuis: well your subject is so broad that that should not be a problem :-)
20:39:11  <mmalecki>bnoordhuis: resume ;)
20:39:42  <mmalecki>piscisaureus_: ooc, you finished your lxjs slides already?
20:39:54  <piscisaureus_>mmalecki: not completely
20:39:58  <mmalecki>I got pretty far with mine
20:40:06  <mmalecki>I have `git init` done
20:40:12  <piscisaureus_>mmalecki: but its based on my nodeconf talk
20:40:32  <piscisaureus_>mmalecki: so, I have to practice rather than make slides
20:40:49  <mmalecki>yeah, I did a talk on my topic already
20:40:58  <mmalecki>but my slides were, ugh.
20:41:15  <mmalecki>I finished them on the previous talk, while being sick and hangover
20:41:33  <piscisaureus_>mmalecki: maybe I'm going to ask you to swap slots
20:41:40  <piscisaureus_>mmalecki: not sure yet
20:41:47  <mmalecki>piscisaureus_: right, I'm still confused about order
20:42:00  <piscisaureus_>mmalecki: I think I should be after you
20:42:01  <piscisaureus_>but then
20:42:30  <mmalecki>piscisaureus_: I actually wanted to make a short introduction for your topic
20:42:33  <mmalecki>just like, one slide
20:42:47  <mmalecki>since I'm planning to go from high level to low level
20:43:21  * TheJHquit (Ping timeout: 256 seconds)
20:43:27  <piscisaureus_>mmalecki: so if we don't swap then you have to do your talk backwards
20:43:29  <piscisaureus_>that'd be fun
20:43:57  <mmalecki>yes, yes it would
20:44:21  <piscisaureus_>start practicing already
20:45:10  <mmalecki>meh
20:47:13  <piscisaureus_>I also requested an "entrance tune"
20:47:45  <piscisaureus_>a very long one with a climax at the end, which is where I enter the stage
20:47:49  <piscisaureus_>and the crowd cheers
20:47:49  <piscisaureus_>so
20:47:56  <piscisaureus_>that works much better when there's a break before my talk
20:48:24  * isaacsback
20:48:27  <piscisaureus_>otherwise I'll overlap with the last 15 minutes of your talk
20:48:30  <isaacs>piscisaureus_: thanks, pulling now
21:00:50  * brsonjoined
21:03:21  * dapquit (Quit: Leaving.)
21:03:54  * dapjoined
21:03:56  * stagasquit (Quit: ChatZilla 0.9.89-rdmsoft [XULRunner 1.9.0.17/2009122204])
21:17:55  * rendarquit
21:27:17  <bnoordhuis>mmalecki: i'll mail it tomorrow
21:28:14  <txdv>I LIKE TURTLES
21:30:25  <mmalecki>bnoordhuis: awesomesauce
21:36:01  <txdv>found a perl binding for libuv
21:36:08  <txdv>is it called perl bindings or binding?
21:36:15  <txdv>whats more appropriate english?
21:37:04  <mjr_>txdv: if you say, "found *a* perl binding", then it would be singular.
21:37:25  <mjr_>But if you say "I found perl bindings", you'd probably go plural, with an s.
21:37:31  <txdv>I always read bindings and bindings refering to one library, so I got used to it
21:43:53  <TooTallNate>it could be plural as in "bindings to the multiple functions within the library"
21:44:07  <TooTallNate>but that just makes things confusing :p
21:52:43  <CoverSlide>hmm nothing in CPAN
21:53:10  <CoverSlide>search for perl and libuv everything is in japanese
21:58:17  <txdv>cpan the dumping ground for perl libs?
21:58:38  <txdv>I don't understand why these people don't join this channel.
21:59:04  <txdv>Like I do, I joined here, started to cry about how much it sucks to write bindings for c#.
21:59:50  <txdv>that perl thing has only support for one loop.
22:00:08  <txdv>One loop to rule all the connections.
22:14:56  <piscisaureus_>These people don't speak English
22:14:57  <piscisaureus_>sadly
22:18:24  <txdv>i don't get the japanese people on 2 things
22:18:31  <txdv>why the fuck are they still hunting whales
22:18:34  <txdv>and how can't they learn english
22:23:51  <isaacs>alright, all tests passing.
22:24:02  <isaacs>binaries built and uploaded
22:24:05  <isaacs>shasums calculated.
22:24:08  <isaacs>moving forward.
22:24:25  <isaacs>oh, wait, the pkg. whoops :)
22:24:25  <txdv>shazam!
22:24:29  <isaacs>forgot that one
22:33:06  * dapquit (Quit: Leaving.)
22:35:45  <piscisaureus_>I am under the impression that no one ever looked at the openssl config files for 10 years or so
22:36:13  <piscisaureus_>there's a *lot* of comments about vax and dec
22:36:30  <piscisaureus_>not a single one about x64
22:38:29  * dapjoined
22:45:04  <isaacs>and we're out
22:47:50  <piscisaureus_>I wonder why test-fs-largefile suddenly started failing...
22:47:55  <CoverSlide>\o/ wheee!!
22:48:02  <piscisaureus_>we didn't even get close to that
22:48:09  <piscisaureus_>s/close to/near/
22:48:56  * dapquit (Quit: Leaving.)
22:49:10  <isaacs>piscisaureus_: it's been kinda flaky for me for a while on linux
22:49:14  <isaacs>works 100% on smartos and osx
22:49:22  <piscisaureus_>isaacs: I suppose that it times out right?
22:49:35  <isaacs>piscisaureus_: yep.
22:49:40  <isaacs>or the numbers don't match
22:49:50  <piscisaureus_>isaacs: well, the latter is a serious isse
22:50:07  <piscisaureus_>isaacs: a timeout happens when your hard drive doesn't manage to write a 5gb file within the deadline
22:50:51  <piscisaureus_>weird enough
22:50:57  <piscisaureus_>the file that it creates is actually big enough
22:51:07  <isaacs>hmm...
22:51:15  <isaacs>yeah, i'm getting a fail on that on all three unixes right now..
22:51:31  <piscisaureus_>something is broken I think
22:51:36  <piscisaureus_>I hope it's not really serious
22:53:09  <isaacs>hm. looks like the truncate isn't making the file big.
22:53:22  <piscisaureus_>I think truncate may be broken
22:53:27  <piscisaureus_>or, not largefile aware
22:53:28  <isaacs>could e
22:54:06  <txdv>who needs x64 if one has vax
22:54:45  <isaacs>piscisaureus_: it looks like truncateSync is returning before it's actually done.
22:54:58  <isaacs>piscisaureus_: it's creating a 5Gb file
22:55:05  <isaacs>piscisaureus_: but it's not returning synchronously
22:55:08  <isaacs>it's returning too fast.
22:55:09  <piscisaureus_>isaacs: yeah, so maybe stat is the one that's failing
22:55:11  <isaacs>so thte stat happens before it's done
22:55:30  <piscisaureus_>isaacs: maybe we're accidentally running the async version?
22:55:30  <isaacs>when i run it on os x, it actually hangs for a bit, and i can watch the file getting bigger in that time
22:55:36  <isaacs>looks that way
22:55:42  <piscisaureus_>hah
22:55:43  <piscisaureus_>ouch
22:56:32  <piscisaureus_>isaacs: yeah, here too
22:56:37  <piscisaureus_>isaacs: it also happens on windows
22:56:44  <piscisaureus_>so the bug is most likely on the node end of things
22:57:09  <isaacs>we're calling the sync binding function
22:57:38  <isaacs>it's definitely doing SYNC_CALL(ftruncate, 0, fd, len)
22:57:45  <isaacs>maybe it's in libuv?
22:58:13  <piscisaureus_>maybe it's a stat bug
22:58:26  <piscisaureus_>if I use it with settimeout it still fails
22:58:35  <isaacs>how much of a timeout?
22:58:40  <piscisaureus_>10 s
22:58:48  <piscisaureus_>I have an ssd btw
22:58:52  * dapjoined
22:59:06  <piscisaureus_>it takes virtually no time to create that file
22:59:16  <piscisaureus_>(it doesn't actually zero 5gb or something)
22:59:38  <isaacs>yeah, i have an ssd, too
22:59:41  <isaacs>i don't get why it's hanging
22:59:48  <isaacs>using fs.truncate(..., cb) doesn't make it go away, either
23:00:19  <isaacs>why the hang?
23:00:21  <isaacs>i don't get it
23:00:31  * Raynos_quit (Excess Flood)
23:00:35  <isaacs>yeah, i'm becoming suspicious of stat now
23:00:36  <piscisaureus_>07804c7c9a4c2eb50eb897f ?
23:00:39  <isaacs>fstat also says it's 1gb
23:01:01  <piscisaureus_>which is basically the actual filesize minus pow(2, 32)
23:01:12  <piscisaureus_>s/minus/modulo/
23:01:16  <piscisaureus_>integer overflow
23:01:23  <piscisaureus_>which is exactly what this test is supposed to catch :_)
23:02:22  <piscisaureus_>isaacs: yup, identified
23:02:25  <piscisaureus_>isaacs: it's that patch
23:02:38  <piscisaureus_>bnoordhuis: !
23:03:07  <bnoordhuis>piscisaureus_: ?
23:03:24  <piscisaureus_>bnoordhuis: you broke stat
23:03:44  <bnoordhuis>go on
23:04:13  <piscisaureus_>bnoordhuis: files > 4gb don't have the correct size reported
23:04:23  <piscisaureus_>bnoordhuis: because you swapped Number::New for Integer::new
23:04:26  * mjr_quit (Quit: mjr_)
23:04:35  <piscisaureus_>enough name calling for today :-)
23:04:40  <bnoordhuis>ah
23:04:45  <bnoordhuis>well, easy to fix
23:04:53  <piscisaureus_>yup
23:05:12  * mjr_joined
23:05:35  <piscisaureus_>I hope nobody uses node to work with big files
23:05:55  <txdv>from which version till what version was this bug in
23:06:08  <piscisaureus_>the bug was released 1 hour ago
23:06:11  <piscisaureus_>or thereabout
23:06:23  <txdv>innn 0.8?
23:07:12  <piscisaureus_>0.8.10
23:07:47  <CoverSlide>oh who needs stat?
23:08:21  <txdv>well, looks like a 0.8.11 is needed immediatelyh
23:09:34  <isaacs>yeah. seems that way.
23:09:44  <isaacs>well, we can release it tomorrow.
23:09:50  * dscapequit (Remote host closed the connection)
23:09:50  * indutnyquit (Remote host closed the connection)
23:10:59  <chilts>heh, I just installed v0.8.10 ... good job for having nvm :)
23:11:34  <txdv>I install them manually
23:12:47  <piscisaureus_>https://github.com/joyent/node/commit/970b011bc277ce8376d9814f5ba4c234947ed55e
23:12:56  <piscisaureus_>^-- isaacs, bnoordhuis: review?
23:12:59  <piscisaureus_>either
23:13:35  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/node/commit/f1440b3
23:13:52  <piscisaureus_>bnoordhuis: that will spit out a ton of warning on my machine
23:13:59  <bnoordhuis>why?
23:14:02  <piscisaureus_>bnoordhuis: Number::New takes a double
23:14:09  <bnoordhuis>so?
23:14:16  <piscisaureus_>so it complains about implicit conversion from int64 to double
23:14:36  <bnoordhuis>msvc...
23:14:47  <piscisaureus_>bnoordhuis: with msvc this would not have happened :-)
23:15:06  <piscisaureus_>bnoordhuis: btw - this conversion between double and int... let's not do it if not necessary
23:15:11  <piscisaureus_>we're not javascript
23:15:29  <bnoordhuis>okay, land your patch
23:15:45  <piscisaureus_>it still doesn't pass test-fs-largefile here
23:15:47  <piscisaureus_>let's figure that out
23:15:53  <bnoordhuis>but i guess it needs to be done for ino and blocks as well
23:15:53  <piscisaureus_>D:\node42>Release\node.exe test\pummel\test-fs-largefile.js
23:15:53  <piscisaureus_>[TypeError: Not an integer]
23:16:03  <piscisaureus_>bnoordhuis: probably
23:16:20  <piscisaureus_>bnoordhuis: it's not a regression tho - those always were converted with Integer::New
23:16:35  <bnoordhuis>ah okay. well, let it be for now
23:16:45  <piscisaureus_>bnoordhuis: I wouldn't mind to fix that
23:16:52  <bnoordhuis>btw, the test is passing, it's just chatty
23:16:53  <piscisaureus_>bnoordhuis: we could even claim we made node better, then :-)
23:17:05  <piscisaureus_>oh
23:17:08  <piscisaureus_>you're right
23:17:29  <piscisaureus_>bnoordhuis: maybe we can have another x macro for 64-bit sized numbers
23:17:35  <bnoordhuis>sure
23:17:37  <bnoordhuis>one sec
23:17:49  <piscisaureus_>bnoordhuis: and static_cast to double ;-)
23:18:01  <piscisaureus_>bnoordhuis: or I will refuse to clap tomorrow
23:19:25  <bnoordhuis>you're coming to the talk?
23:22:32  <bnoordhuis>piscisaureus_: https://github.com/bnoordhuis/node/commit/ec03c47
23:22:56  <piscisaureus_>sure
23:23:19  <piscisaureus_>bnoordhuis: lgtm - testing noew
23:23:44  <bnoordhuis>passes tests here
23:29:00  <piscisaureus_>bnoordhuis: here too - ship it
23:29:27  <bnoordhuis>c'mon CIA
23:29:40  <piscisaureus_>CIA is MIA
23:32:13  <txdv>stable official release with such a bug ... how can you!?
23:32:18  * piscisaureus_changed nick to CIA-1313
23:32:19  <CIA-1313>node: bnoordhuis v0.8 * r970b01 / (src/node_file.cc): fs: fix fs.stat reports the wrong size for files > 4GB - https://github.com/joyent/node/commit/970b011bc277ce8376d9814f5ba4c234947ed55e
23:32:25  * CIA-1313changed nick to piscisaureus
23:32:39  <bnoordhuis>har har
23:35:24  <bnoordhuis>is there a reason `make test` doesn't run the pummel tests?
23:35:42  <piscisaureus>bnoordhuis: because it takes long. It also skips internet tests
23:35:44  <piscisaureus>bnoordhuis: run test-all
23:36:02  <bnoordhuis>yeah, but test-all also runs tests in debug mode
23:36:02  <piscisaureus>speaking of which
23:36:11  <piscisaureus>bnoordhuis: oh, not on windows
23:36:20  <piscisaureus>bnoordhuis: maybe we should test to run all tests
23:36:33  <piscisaureus>and have "test-fast" or something
23:36:37  <bnoordhuis>yes
23:36:59  <piscisaureus>bnoordhuis: so blame the messenger, eh?
23:37:01  <piscisaureus>:-p
23:37:22  <bnoordhuis>well, i would've caught that if `make test` ran the pummel tests
23:37:29  <piscisaureus>we also need to update test-internet to not query google's CNAME record
23:37:34  <piscisaureus>since there isn't any
23:37:53  <piscisaureus>bnoordhuis: if you had asked for review it would also be caught
23:38:21  <bnoordhuis>yeah, maybe
23:38:24  <txdv>its my fault
23:38:52  <piscisaureus>now who uses a cname record
23:38:55  <piscisaureus>c9.io ?
23:39:31  <bnoordhuis>i have a couple of cname records
23:41:32  <piscisaureus>microsoft.com
23:41:44  <piscisaureus>unfortunately the test still fails because ipv6 not working
23:46:08  <txdv>did you guys gut sendfile from nodejs?
23:49:00  <bnoordhuis>txdv: no, it's there, just not very useful
23:49:13  * dscapejoined
23:49:24  <txdv>why is it not useful?
23:49:40  <bnoordhuis>txdv: because it needs a socket file descriptor that you can't get at
23:49:51  <piscisaureus>you can make an educated guess
23:50:13  <bnoordhuis>somewhere between 1 and 2^31-1
23:51:43  <txdv>Don't I have a fd if I open a file?
23:58:10  <piscisaureus>you can't sendfile() to a file :-)
23:58:51  <txdv>o yeah, it takes a path to a file and a socket fd
23:58:54  * indutnyjoined