00:00:29  <bnoordhuis>sh1mmer: was that meant for #node.js?
00:00:43  <sh1mmer>Not really.
00:01:40  <bnoordhuis>oh, in that case i don't have an opinion
00:01:40  <sh1mmer>I don't mean at a JS level, his "benchmarks" show Silk doing an order of magnitude faster response than Node.js and I'm wondering why.
00:02:13  <sh1mmer>It seems like that's going to have more to do with libuv than Node in general.
00:06:59  <piscisaureus_>it seems it has more to do with it doing a prefork server
00:07:19  <piscisaureus_>and comparing that to a single node process
00:07:31  <sh1mmer>I'm rerunning tests now with 0.8.2
00:07:33  <piscisaureus_>but we gotta make fs faster
00:07:34  <sh1mmer>and cluster
00:07:59  <sh1mmer>but his benchmark compared the node helloworld to a png transfer
00:08:37  <sh1mmer>and still did it in 1/13th of the time using 4 cores instead of 1
00:09:06  <sh1mmer>I don't know about the caching in his sample and how much of that was in memory
00:09:07  * EhevuTov_quit (Quit: Leaving)
00:09:28  <piscisaureus_>so where is that benchmark?
00:10:33  <sh1mmer>https://github.com/mschwartz/SilkJS/wiki/HTTPD-Benchmarks
00:10:45  <sh1mmer>I'm rerunning on my MBA now
00:12:29  <bnoordhuis>seems he's running a single node instance and multiple silkjs instances
00:12:37  <bnoordhuis>apples, oranges
00:13:07  <bnoordhuis>also v0.6.5
00:13:13  <sh1mmer>actually 0.4.8
00:13:23  <bnoordhuis>hah, that's even worse
00:13:25  <sh1mmer>right
00:13:28  <sh1mmer>it's not just 1/4 though
00:13:32  <sh1mmer>which is why I'm rerunning
00:13:47  <sh1mmer>if it's a really large gap I'm curious why
00:13:58  <piscisaureus_>doh
00:14:04  <piscisaureus_>in his node test keepalive is not used
00:14:12  <sh1mmer>:/
00:14:16  <sh1mmer>"benchmarks"
00:14:19  <piscisaureus_>that's weird btw
00:14:20  <piscisaureus_>well
00:14:27  <piscisaureus_>he says he is using ab -k
00:14:55  <piscisaureus_>but his results are more consistent with non-keepalive numbers that I'm used to
00:15:19  <piscisaureus_>with keepalive node should typically do 12000-16000 rps
00:15:31  <piscisaureus_>he's seeing 6000
00:15:41  <piscisaureus_>but maybe something is broken about node
00:15:48  <piscisaureus_>so ab -k is not working
00:16:04  <sh1mmer>I think there was also a broken version of ab that shipped with mac
00:16:09  <sh1mmer>but I forget what was broken
00:16:27  <sh1mmer>I just remember I have a self compiled version on my mac for some reason
00:16:32  <piscisaureus_>node 0.4.8
00:16:34  <piscisaureus_>hahaha
00:16:54  <piscisaureus_>sh1mmer: that's correct, but you'll see soon enough
00:17:29  <piscisaureus_>sh1mmer: faulty abs fail with apr_socket_something: fuckup(42)
00:17:53  <sh1mmer>piscisaureus_: I had lots of fun benchmarking at joyent for telcos
00:17:54  <sh1mmer>bleh
00:18:04  <piscisaureus_>ghe
00:19:34  <piscisaureus_>so I think that node will actually beat silk to it when keepalive is used
00:19:58  <bnoordhuis>keepalive and cluster
00:20:21  <sh1mmer>https://gist.github.com/3094660
00:20:32  <sh1mmer>3x faster
00:20:41  <bnoordhuis>hah, owned
00:21:02  <bnoordhuis>good one, sh1mmer
00:21:07  <sh1mmer>low amount of concurrency but I couldn't be bothered to fiddle with FDs and ephemeral port settings
00:21:18  <sh1mmer>on machine benchmarking is bullshit anyway
00:21:39  <bnoordhuis>benchmarking on os x too
00:21:49  <sh1mmer>Also that.
00:22:49  <sh1mmer>in fairness silk was sending a larger response but whatever. Close enough. I can't be bothered to make a "fair" comparison
00:30:20  * hij1nxjoined
00:36:40  <isaacs>Stupid build/git/branch process bs quesion: should we merge v0.6.20 back into v0.8? or back into master? or just let it go to /dev/null?
00:37:01  <isaacs>i honestly have no opinion
00:37:14  <bnoordhuis>i vote /dev/null
00:37:28  <bnoordhuis>anything else becomes madness fast
00:37:41  <isaacs>it's basically just the changelog
00:37:58  <isaacs>and it's not a very exiciting changelog
00:38:10  <isaacs>very boring, in fact.
00:40:00  * hij1nxquit (Quit: hij1nx)
00:42:10  <isaacs>ok, i'm just gonna let it wither and die then. merging v0.8 into master now
00:42:19  <isaacs>just running make test first.
00:42:56  * c4miloquit (Remote host closed the connection)
00:43:03  <ryah>can someone email me at ry@tinyclouds.org - as a test?
00:43:47  <sh1mmer>ryah: done
00:44:25  <piscisaureus_>isaacs: I would merge it, but I suppose there's no hurry :-)
00:44:57  <bnoordhuis>piscisaureus_: what's in v0.6 that we want in v0.8 or master?
00:45:02  <isaacs>piscisaureus_: are you saying that you will merge it, but you won't hurry?
00:45:04  <isaacs>;P
00:45:06  <piscisaureus_>the changelog
00:45:14  <piscisaureus_>isaacs: yes
00:45:31  <piscisaureus_>isaacs: sure, I will do it, no worries
00:45:38  <isaacs>kewl
00:46:03  <piscisaureus_>would we merge into 0.8 or into master?
00:46:07  <piscisaureus_>that's the question :-)
00:47:09  <isaacs>indutny: please make jslint pass (all tests, really) before landing.
00:47:10  <isaacs>https://github.com/joyent/node/commit/f210530f#L0R83
00:48:02  <CIA-108>node: isaacs master * r3ad07ed / lib/tls.js : lint (+44 more commits...) - http://git.io/H9hh8w
00:48:04  <isaacs>but that was the only failure merging v0.8 into master, which is nice.
00:48:20  <isaacs>now to make npm actually work on master :)
00:56:24  <CIA-108>node: isaacs master * reb9b8f5 / (121 files in 16 dirs): npm: upgrade to 1.1.39 - http://git.io/xFieRg
00:56:26  <isaacs>done
00:56:54  <ryah>sh1mmer: thanks
00:57:02  <ryah>can someone else try to email me? bnoordhuis?
00:57:08  <ryah>(someone who doens't use gmail)
00:57:17  <ryah>fucking configuring exim
00:57:24  <sh1mmer>I can use yahoo if you like
00:57:24  <ryah>remembering why i hate computers
00:57:29  <ryah>ok
00:58:12  <sh1mmer>done
00:58:17  <sh1mmer>are you coming out for a drink tonight?
00:58:21  <sh1mmer>benders in a couple of hours
01:01:48  <bnoordhuis>ryah: is 97.107.132.72 your email server? port 25 seems to be either closed or filtered
01:02:05  * dapquit (Quit: Leaving.)
01:03:07  <tjfontaine>I concur
01:03:18  <ryah>bnoordhuis: yes.. hm
01:10:07  <CIA-108>node: Bert Belder merge * rfeaf468 / (4 files in 3 dirs): Merge branch 'v0.6' - http://git.io/zs7WzA
01:10:10  <piscisaureus_>^-- isaacs ?
01:10:19  <piscisaureus_>the http cluster thing is funny
01:10:26  <piscisaureus_>you gotta explain that to me
01:12:33  <bnoordhuis>piscisaureus_: that's the cluster benchmark
01:12:38  <piscisaureus_>I know
01:12:48  <piscisaureus_>but how did the for loop end up being different in 0.6 ?
01:12:55  <bnoordhuis>hah, we have nodejs.es now? awesome!
01:13:33  <bnoordhuis>piscisaureus_: i backported the bencmark to v0.6 a while ago, then fixed something in master
01:13:39  <bnoordhuis>i probably didn't backport the fix
01:13:49  <piscisaureus_>bnoordhuis: so should the loop start with 0 or with 1 ?
01:13:53  <bnoordhuis>0
01:14:01  <piscisaureus_>bnoordhuis: ok
01:14:06  <piscisaureus_>so then that wasn't fixed in master
01:14:09  <piscisaureus_>but now it is
01:14:19  <piscisaureus_>so maybe I should also merge into v0.8
01:14:25  <bnoordhuis>or maybe i fixed it in v0.6 and not in master :)
01:14:30  <piscisaureus_>yes, exactly
01:14:48  <piscisaureus_>so probably I should merge v0.6 into v0.8 and not into master
01:15:00  <bnoordhuis>yes, that seems most logical
01:15:01  * piscisaureus_sighs
01:15:40  <piscisaureus_>It also seems that some changes were made in punicode.js that were later reverted by upgrading punicode to 1.something
01:16:52  <bnoordhuis>yes, that sounds about right
01:17:36  * pieternquit (Quit: pietern)
01:19:13  <ryah>bnoordhuis: can you try again?
01:19:40  <bnoordhuis>ryah: port is open now
01:20:08  <bnoordhuis>ryah: 220 localhost ESMTP Exim 4.76 Thu, 12 Jul 2012 01:19:01 +0000
01:20:11  <bnoordhuis>seems to work
01:20:29  <bnoordhuis>i didn't actually send an email though :)
01:21:49  <sh1mmer>ryah: why not use gmail?
01:22:23  <ryah>sh1mmer: i do use gmail
01:22:29  <ryah>setting up forwarding
01:24:02  <sh1mmer>why not just point the DNS to goog's mail servers?
01:26:19  <piscisaureus_>bnoordhuis: do you know why punycode was downgraded in 0.6?
01:26:59  <bnoordhuis>piscisaureus_: yes... or rather, i did
01:27:40  <bnoordhuis>piscisaureus_: iirc we botched a punycode.js upgrade
01:27:48  <piscisaureus_>right, ok
01:28:17  * abraxasjoined
01:28:21  <piscisaureus_>bnoordhuis: I find it odd that the upgrade is reverted on 0.6 but not on v0.8
01:29:14  <piscisaureus_>maybe because it was an api change
01:29:22  <piscisaureus_>and someone figured out we don't want that
01:29:37  <ryah>smtp is so horrible
01:29:50  <bnoordhuis>piscisaureus_: that was probably it
01:29:53  <bnoordhuis>let me dig up the PR
01:30:48  <bnoordhuis>piscisaureus_: https://github.com/joyent/node/pull/3359
01:31:20  <piscisaureus_>bnoordhuis: yeah, I found that too, but I found it really unclear what happened there
01:32:20  <CIA-108>node: Bert Belder v0.8 * rd2e40f6 / (4 files in 3 dirs): Merge branch 'v0.6' into v0.8 (+22 more commits...) - http://git.io/BDyQLQ
01:40:04  <ryah>sh1mmer: email me again
01:48:59  * chobi_echanged nick to chobi_e_
01:54:40  * mikealquit (Quit: Leaving.)
01:57:36  <piscisaureus_>ok, I'm off
01:57:42  <piscisaureus_>talk to you later people
01:59:26  * c4milojoined
01:59:45  <bnoordhuis>sleep tight, bertje
02:03:59  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:24:00  * TooTallNatequit (Quit: Computer has gone to sleep.)
02:24:51  <bnoordhuis>woo, https://patchwork.kernel.org/ is back online - now i can watch all patches on v9fs-devel@kernel.org again
02:29:40  * mikealjoined
02:32:21  * bnoordhuisquit (Quit: Leaving)
02:53:52  * beachdogjoined
02:59:36  * ericktquit (Ping timeout: 244 seconds)
03:03:19  * ericktjoined
03:03:21  * Ariajoined
03:13:11  * ericktquit (Quit: erickt)
03:15:13  * c4miloquit (Remote host closed the connection)
03:18:59  * brsonquit (Quit: leaving)
03:39:37  * mikealquit (Quit: Leaving.)
03:40:10  * TooTallNatejoined
03:45:49  * beachdogquit (Read error: Connection reset by peer)
03:48:18  * hamzajoined
03:50:27  * brsonjoined
03:53:14  * beachdogjoined
04:24:51  * philips_quit (Excess Flood)
04:28:34  * philips_joined
04:36:46  * ericktjoined
04:37:51  * TooTallNatequit (Quit: Computer has gone to sleep.)
04:58:08  * ericktquit (Quit: erickt)
05:16:26  * Ariaquit (Remote host closed the connection)
05:28:11  * beachdogquit (Remote host closed the connection)
05:36:47  * brakerjoined
05:49:17  * paddybyersjoined
06:03:56  * mikealjoined
06:28:10  <indutny>morning everyone
06:44:13  * EhevuTovjoined
06:58:47  * hzjoined
07:02:49  * hzquit (Disconnected by services)
07:02:52  * hzjoined
07:04:07  * stephankquit (Quit: *Poof!*)
07:19:04  * AvianFluchanged nick to AvianusAsleepus
07:24:50  * loladirojoined
07:25:00  * brakerquit (Ping timeout: 245 seconds)
07:25:35  * TooTallNatejoined
07:29:29  * TooTallNatequit (Client Quit)
07:35:30  * brsonquit (Ping timeout: 252 seconds)
07:55:26  * loladiroquit (Quit: loladiro)
07:59:22  * AlbireoXquit (Quit: IRC crashed.)
08:04:42  * rendarjoined
08:12:29  * AlbireoXjoined
08:27:06  * ggreerquit (Excess Flood)
08:31:09  * ggreerjoined
09:06:58  * AndreasMadsenjoined
09:34:33  * EhevuTovquit (Quit: This computer has gone to sleep)
09:39:27  * hamzaquit (Quit: Page closed)
10:05:28  * mmaleckijoined
10:06:46  * piscisaureus_joined
10:31:03  * AndreasMadsenquit (Remote host closed the connection)
10:31:10  * abraxasquit (Remote host closed the connection)
10:58:53  * piscisaureus_quit (Ping timeout: 246 seconds)
11:10:00  * loladirojoined
11:23:54  * piscisaureus_joined
11:40:40  * piscisaureus__joined
11:46:08  * piscisaureus_quit (Ping timeout: 246 seconds)
11:46:08  * ggreerquit (Ping timeout: 246 seconds)
11:46:08  * piscisaureus__quit (Ping timeout: 246 seconds)
11:46:29  * einarosquit (Ping timeout: 246 seconds)
11:47:40  * einarosjoined
11:53:56  * AngryParsleyjoined
12:12:10  * loladiroquit (Ping timeout: 250 seconds)
12:17:07  * piscisaureus_joined
12:18:55  <piscisaureus_>hello9
12:19:15  <mmalecki>hello10
12:23:23  <indutny>hello0xa
12:23:52  <mmalecki>lol
12:29:17  * piscisaureus__joined
12:30:56  * piscisaureus_quit (Ping timeout: 246 seconds)
12:37:22  <piscisaureus__>my asm parser is almost done :-0
12:37:30  * piscisaureus__changed nick to piscisaureus
12:38:40  * xaqjoined
12:42:51  * loladirojoined
12:44:09  <indutny>wow
12:44:12  <indutny>piscisaureus: for openssl?
12:44:16  <piscisaureus>yes
12:44:24  <indutny>really nice!
12:44:27  <piscisaureus>well
12:44:31  * bnoordhuisjoined
12:44:35  <piscisaureus>it won't be good for anything
12:44:36  <mmalecki>piscisaureus: assembler parser?
12:44:40  <piscisaureus>except for getting this done
12:44:41  <piscisaureus>yes
12:44:49  <piscisaureus>it parses at&t asm
12:44:53  <mmalecki>ok I don't even
12:44:55  <piscisaureus>and generates other format
12:44:58  <indutny>piscisaureus: hahaha
12:45:04  <indutny>other, even more strangier
12:45:24  <piscisaureus>maybe I can also just emit bytecode
12:45:39  <piscisaureus>but it's written in javascript
12:46:01  <piscisaureus>so it's probably not useful for any purpose
12:46:58  <mmalecki>bnoordhuis: o hai
12:47:04  <bnoordhuis>mmalecki: sup
12:47:09  <mmalecki>bnoordhuis: shots?
12:47:10  <indutny>piscisaureus: bytecode?
12:47:21  <piscisaureus>well yeah like, object files or something
12:47:23  <indutny>piscisaureus: you can reuse v8's codegen
12:47:33  <piscisaureus>that's also an option :-)
12:47:49  <piscisaureus>I don't think I will do that
12:47:53  <indutny>well, yes
12:47:56  <piscisaureus>just create the files that I noeed
12:47:57  <indutny>because it's jitting code
12:48:02  <indutny>.S files would be good
12:48:16  <indutny>how are you going to integrate it with gyp?
12:48:20  <piscisaureus>I am not
12:48:21  <bnoordhuis>mmalecki: i've had them except for influenza
12:48:28  <piscisaureus>I am just going to generate the asm files that I need
12:48:36  <indutny>piscisaureus: for what?
12:48:51  <piscisaureus>indutny: for doing bignums in asm
12:48:52  <mmalecki>bnoordhuis: lol
12:48:58  <indutny>well
12:49:05  <indutny>ok, cool :)
12:49:13  <indutny>ping me once you'll finish it
12:49:14  <piscisaureus>indutny: what's your reservations?
12:49:23  <indutny>piscisaureus: hm?
12:49:26  * hij1nxjoined
12:49:35  <piscisaureus>[14:48] <@indutny> well
12:49:35  <mmalecki>piscisaureus: hey, Nuno said we're supposed to coordinate or something
12:49:42  <bnoordhuis>why is gyp so full of bugs?
12:49:47  <indutny>piscisaureus: well?
12:50:02  <bnoordhuis>i try to suppress -O3 and use -O2, instead it compiles with neither...
12:50:04  <indutny>piscisaureus: what are you talking about?
12:50:10  <piscisaureus>nvm
12:50:15  <piscisaureus>mmalecki: sure
12:50:26  <mmalecki>piscisaureus: what's your topic actually?
12:54:02  * c4milojoined
12:57:32  <bnoordhuis>gah, i have it working but now it compiles v8 without -O2 or -O3
12:57:36  * bnoordhuisbangs head against table
12:57:47  <piscisaureus>bnoordhuis: what are you doing?
12:58:16  <piscisaureus>bnoordhuis: fix gcc 4.0.2
12:58:20  <piscisaureus>?
12:58:28  <bnoordhuis>piscisaureus: and 4.4.6
12:58:37  <bnoordhuis>and 4.4.3 apparently
12:58:40  <piscisaureus>oh my
12:58:41  <piscisaureus>the horror
12:58:44  <bnoordhuis>also 4.5.2, i believe
12:59:09  <mmalecki>bnoordhuis: well, tell people to upgrade?
12:59:20  <mmalecki>it's like, not your fault
12:59:24  <bnoordhuis>that's hardly an option
12:59:26  <mmalecki>#kthxbai
12:59:32  <mmalecki>why?
13:03:36  * joe______joined
13:03:41  * hij1nxquit (Quit: hij1nx)
13:03:58  <joe______>I really thought it was Unicorn Velociraptor
13:04:10  * joe______changed nick to joeandaverde
13:04:30  <mmalecki>unicorn velociraptor would be pretty cool
13:06:30  * chobi_e_changed nick to chobi_e
13:07:48  <bnoordhuis>son of a... v8 has 'cflags!': ['-O2','-Os'] in its common.gypi
13:08:07  <mmalecki>son of a v8?
13:08:09  <piscisaureus>haaaaaaaaaahahahaha
13:08:26  <piscisaureus>^-- sardonic laugh
13:09:28  <bnoordhuis>looks like we'll be floating patches on top of v8 after all >:(
13:11:31  <piscisaureus>]oh
13:11:35  <mmalecki>so sad
13:11:36  <piscisaureus>we have small patches for the build
13:12:01  <mmalecki>hey, anyone of you guys going to jsconf.eu?
13:12:12  <piscisaureus>I am not
13:12:43  <mmalecki>Bem, you, maybe?
13:12:44  <piscisaureus>bnoordhuis: we're floating this already -> https://github.com/joyent/node/commit/f60def5e9a72fbfe087
13:13:02  <piscisaureus>felixge will, probably
13:13:24  <mmalecki>cool. I'm submitting my talk proposals right now
13:13:59  <bnoordhuis>mmalecki: no, i'm not going
13:14:24  <mmalecki>bnoordhuis: #lame
13:14:53  <bnoordhuis>i don't really have an affinity with javascript
13:16:10  <mmalecki>bnoordhuis: yeah, I imagine you're not a good javascript programmer
13:17:04  <mmalecki>FUCK
13:17:09  <mmalecki>vodka keeps disappearing
13:17:53  <mmalecki>it's not like I'm drinking since 11
13:18:17  <bnoordhuis>i guess it's true, my js has a very distinct c flavor to it
13:18:30  <mmalecki>bnoordhuis: c js is the best kind of js
13:18:58  <mmalecki>bnoordhuis: (really. tends to be the fastest.)\
13:19:41  <mmalecki>bnoordhuis: oh hey, I might be coming to Amsterdam soon
13:19:55  <mmalecki>due to some reasons
13:20:00  <bnoordhuis>mmalecki: yeah? when?
13:20:01  <mmalecki>including my life being a mess
13:20:03  <mmalecki>but anyway
13:20:13  <mmalecki>bnoordhuis: when I get sober and get to the airpot
13:20:21  <mmalecki>airport, even
13:20:33  <bnoordhuis>as in this week?
13:20:53  <mmalecki>I hope, but I have 4 bottles left in my freezer
13:21:03  <mmalecki>so MAYBE
13:21:20  * AndreasMadsenjoined
13:22:21  * joeandaverdequit (Quit: joeandaverde)
13:24:17  <mmalecki>bnoordhuis: you're going to be around next week?
13:32:05  <bnoordhuis>mmalecki: maybe
13:47:07  <mmalecki>bnoordhuis: thats notvery precise
13:48:36  <indutny>bnoordhuis: looks like yesterday's statSync stuff wasn't stack overflow problem
13:49:04  <indutny>bnoordhuis: see http://code.google.com/p/v8/issues/detail?id=2232 for details
13:49:51  <indutny>well, it's related to allocation failurs
13:49:57  <indutny>but not our fault
13:50:01  <indutny>surprisingly!
13:56:03  <mmalecki>I hate urls which are not restful
13:56:15  <indutny>mmalecki: you may not open it then
13:56:33  <mmalecki>they could simply make it http://code.google.com/p/v8/issues/2232
13:58:40  <bnoordhuis>ah, interesting
13:59:06  <indutny>yeah
13:59:16  <indutny>so I just got bogus stack trace, and was confused because of it
13:59:46  <bnoordhuis>mmalecki: when are you in 020?
14:00:01  <indutny>020?
14:00:08  <mmalecki>bnoordhuis: no idea
14:00:11  <indutny>http://en.wikipedia.org/wiki/020
14:00:16  <mmalecki>bnoordhuis: sorry, I'm, uhm
14:00:35  <bnoordhuis>indutny: amsterdam
14:00:43  <indutny>oh
14:00:46  <indutny>I want to get there :)
14:06:50  <piscisaureus>in gas, numerical immediates like $123 are decimal, right ?
14:07:04  <piscisaureus>^-- bnoordhuis - o grote goeroe, verlicht mij
14:07:23  <indutny>hahah
14:07:24  <indutny>hex gas
14:08:37  <mmalecki>indutny: oh, you have a new twitter avatar!
14:09:53  <indutny>mmalecki: we all have now :P
14:11:27  <indutny>I like this trend, actually
14:21:47  * hzquit
14:38:01  <CIA-108>node: Ben Noordhuis v0.8 * r202df30 / (common.gypi configure deps/v8/build/common.gypi): build: disable unsafe optimizations - http://git.io/NP-lHA
14:38:02  <CIA-108>node: Ben Noordhuis v0.8 * rc6bb361 / configure : build: partially fix configure on ARM - http://git.io/3baXlQ
14:38:04  <bnoordhuis>piscisaureus: yes
14:40:32  <indutny>bnoordhuis: I guess there is just no other clean solution to this than returning an empty handle. Generally, any API call can return an empty handle as an error sentinel and it's the embedder's job to check against it.
14:40:34  <indutny>basically
14:40:36  <indutny>we're fucked up
14:41:03  <indutny>bnoordhuis: that's a quote
14:41:05  <indutny>actually
14:41:17  <indutny>of yangguo@chromium.org
14:41:31  <indutny>about that v8 stuff that we was debugging yesterday
14:42:14  <bnoordhuis>oh well
14:42:28  <bnoordhuis>let's just let node crash for now
14:43:11  <indutny>haha
14:47:53  <indutny>well, I think v8 should handle that
14:47:57  <indutny>or at least return Undefined
14:48:53  <bnoordhuis>Undefined is a valid value
14:49:01  <bnoordhuis>i can see why they return an empty handle
14:49:17  <bnoordhuis>it'd be kind of annoying to have to handle that everywhere though
14:49:30  <bnoordhuis>and when i say 'kind of' i mean 'very'
14:49:40  <piscisaureus>Can you guys explain to me what this is about?
14:49:47  <piscisaureus>when does v8 return an empty handle?
14:50:04  <indutny>well
14:50:10  <indutny>piscisaureus: if v8 fails to allocate any stuff on heap
14:50:15  <indutny>piscisaureus: it returns empty handle
14:50:19  <piscisaureus>oh
14:50:21  <indutny>piscisaureus: yes
14:50:23  <piscisaureus>shrug
14:50:30  <piscisaureus>in that case the heap is full
14:50:40  <piscisaureus>crashing is the only viable option anyway
14:50:42  <indutny>well, I think there're other stuff that may cause it
14:50:50  <indutny>piscisaureus: no
14:50:55  <indutny>piscisaureus: it should throw js exception
14:51:04  <piscisaureus>eh wut?
14:51:09  <indutny>Maximum call-stack size exceeded
14:51:22  <piscisaureus>oh
14:51:30  <indutny>yeah
14:51:31  <piscisaureus>I really don't get it
14:51:34  <indutny>me too
14:51:48  <piscisaureus>So you're saying we should say stack overflow when the heap is full?
14:51:55  <indutny>indeed
14:51:55  * piscisaureusgets whisky
14:51:59  <piscisaureus>indutny: why?
14:52:07  <indutny>oh wait a minute
14:52:08  <mmalecki>piscisaureus: ++ for whisky
14:52:10  <kohai>piscisaureus has 3 beers
14:52:14  <indutny>why have I said "heap is full"
14:52:17  <indutny>stack is full :D
14:52:22  <piscisaureus>aha
14:52:26  <piscisaureus>well
14:52:26  <indutny>it fails to anything because stack limit reached
14:52:32  <indutny>we can't check this in C++ land
14:52:36  <piscisaureus>well
14:52:37  <indutny>because it reaches limit in v8's side
14:52:41  <piscisaureus>you can just set the v8 limit lower
14:52:46  <indutny>won't work
14:52:49  <indutny>it'll crash
14:52:49  <piscisaureus>why not?
14:52:53  <piscisaureus>huh
14:52:53  <piscisaureus>now
14:52:54  <indutny>go and try :)
14:52:56  <piscisaureus>*no
14:53:01  <indutny>it'll crash because of empty handle
14:53:08  <piscisaureus>no
14:53:19  <piscisaureus>if you make it lower the stack overflow will happen in js land
14:53:37  <piscisaureus>we just need to make sure that there is always enough stack space left for whatever we want to do in c++ land
14:53:39  <piscisaureus>then there is no problem
14:54:00  <indutny>рфрф
14:54:03  <indutny>hahah
14:54:13  <indutny>we can do that check
14:54:16  <indutny>but it'll still fail randomly
14:54:21  <piscisaureus>why?
14:54:25  <indutny>depending on where stack overflow will happen
14:54:28  <piscisaureus>it's not about a check
14:54:35  <piscisaureus>we don't add a check
14:54:46  <piscisaureus>v8 checks for stack overflows all the time
14:55:00  <piscisaureus>we just have to make it trip slightly earlier
14:55:18  <indutny>piscisaureus: well
14:55:21  <piscisaureus>so we never get to the situation where we don't have enough stack space left in c++ land
14:55:26  <indutny>hm...
14:55:28  <indutny>sounds good
14:55:36  <indutny>I can implement it as floating patch to v8
14:55:39  <piscisaureus>no
14:55:47  <piscisaureus>the v8 stack overflow limit is configurable
14:55:53  <piscisaureus>you can just change the limit
14:55:55  <isaacs>good morning
14:55:59  <piscisaureus>anb *poof* magic
14:56:04  <piscisaureus>^-- indutny
14:56:09  <indutny>isaacs: morning
14:56:12  <mmalecki>isaacs: morning
14:56:18  <indutny>piscisaureus: you may want to try
14:56:25  <indutny>we have --stack-trace-limit option for node
14:56:34  <indutny>probably, I don't understand the whole problem
14:56:37  <piscisaureus>indutny: dude - that's not the stack
14:56:41  <indutny>btw, link: http://code.google.com/p/v8/issues/detail?can=2&start=0&num=100&q=&colspec=ID%20Type%20Status%20Priority%20Owner%20Summary%20HW%20OS%20Area%20Stars&groupby=&sort=&id=2232
14:56:48  <piscisaureus>indutny: that's the maximum length of a stack trace
14:56:51  <indutny>errrr
14:56:53  <indutny>sorry
14:56:57  <indutny>that was a typo
14:57:10  <indutny>what's up with me
14:57:14  <indutny> " --max-stack-size=val set max v8 stack size (bytes)\n"
14:57:19  <indutny>piscisaureus: that's it ^
14:59:37  <piscisaureus>sure
14:59:55  * chobi_echanged nick to chobi_e_
15:01:11  <piscisaureus>I don't think the --max-stack-size option actually does anything
15:02:21  <piscisaureus>indutny: that argument is just not working
15:02:48  <indutny>piscisaureus: wat?!
15:02:56  <piscisaureus>indutny: no matter what I set there, I get stack overflow after approx. 20000 recursive function calls
15:03:02  <piscisaureus>(empty function)
15:03:10  <indutny>are you sure?
15:03:12  <piscisaureus>so there's ur answer
15:03:15  <piscisaureus>sure
15:03:18  <indutny>well, yes
15:03:22  <indutny>it doesn't work here
15:03:28  <indutny>but it's definitely calling API
15:03:31  <piscisaureus>ok
15:03:36  <piscisaureus>so there you have your bug :-)
15:03:55  <piscisaureus>maybe the stack limit actually gets compiled into the snapshot
15:04:00  <piscisaureus>that wouldn't suprise me
15:04:08  <piscisaureus>you should try with a --without-snapshot build
15:04:30  <indutny>well, it shouldn't
15:04:44  <piscisaureus>doesn't it get stashed in the roots array?
15:04:52  <piscisaureus>I mean, sure, it shouldn't
15:04:54  <piscisaureus>:-)
15:05:10  <indutny>hm...
15:05:18  <indutny>well, all check that I've found are in runtime
15:05:44  <piscisaureus>yeah
15:05:56  <piscisaureus>but where does the variable that it checks against sit?
15:06:13  <piscisaureus>oh btw I am sure it also does these things in generated code
15:06:19  <indutny>yes, it does
15:06:27  <indutny>it should be doing it
15:06:38  <piscisaureus>It emits a check at every loop edge
15:06:43  <piscisaureus>I would hope that wasn't a runtime call
15:06:53  * piscisaureusbrb
15:07:18  <indutny>k
15:10:26  <isaacs>call in one hour
15:11:37  <indutny>ohhh
15:11:38  <indutny>right
15:11:46  <indutny>I almost forgot about it
15:11:57  <isaacs>indutny: i know, you're probably a few days in either direction by now :)
15:12:04  <indutny>haha
15:14:45  * pieternjoined
15:19:02  * philips_quit (Excess Flood)
15:21:41  * philips_joined
15:23:33  * ericktjoined
15:27:00  * dapjoined
15:42:16  * hij1nxjoined
15:44:59  * mikealquit (Quit: Leaving.)
15:45:53  * EhevuTovjoined
15:47:09  * AndreasMjoined
15:49:30  <piscisaureus>might be ~ 10 minutes late
15:50:00  <piscisaureus>isaacs: ^
15:50:05  <isaacs>k
15:50:14  <piscisaureus>bnoordhuis: don't forget the call
15:50:20  <piscisaureus>(we have to start pinging bnoordhuis)
15:51:34  * piscisaureusquit (Quit: ~ Trillian Astra - www.trillian.im ~)
15:55:00  <indutny>bnoordhuis: hey ben
15:55:02  * loladiroquit (Ping timeout: 246 seconds)
15:55:19  <indutny>bnoordhuis: I just created a script to translate callgrind's logs to add js function names here
15:55:20  <indutny>:D
15:56:22  <bnoordhuis>here
15:56:38  <bnoordhuis>indutny: neat
15:57:20  * stephankjoined
15:58:07  * AndreasMquit (Remote host closed the connection)
15:58:11  <indutny>heat
15:58:33  * piscisaureus_joined
15:59:00  <piscisaureus_>bnoordhuis: ?
15:59:12  <bnoordhuis>piscisaureus_: ?
15:59:19  <piscisaureus_>ah you are here
15:59:27  <piscisaureus_>i will be in 10
15:59:48  <indutny>?
15:59:55  <indutny>:)
16:09:34  <indutny>piscisaureus_: ping pong?
16:15:08  <isaacs>piscisaureus_: lost you
16:16:35  * piscisaureus__joined
16:16:57  * piscisaureus_quit (Read error: Connection reset by peer)
16:18:09  * mikealjoined
16:21:17  <bnoordhuis>review anyone? https://github.com/bnoordhuis/node/compare/joyent:v0.8...v0.8
16:21:29  <bnoordhuis>related issue -> https://github.com/joyent/node/issues/3699
16:25:34  * AndreasMadsenquit (Remote host closed the connection)
16:38:38  * xaqquit (Read error: Connection reset by peer)
16:38:41  <piscisaureus__>bnoordhuis: lgtm
16:38:55  <bnoordhuis>thanks
16:39:08  * xaqjoined
16:39:21  <CIA-108>node: Ben Noordhuis v0.8 * r5d97d72 / lib/net.js : net: fix bogus errno reporting - http://git.io/DTpjHQ
16:39:21  <CIA-108>node: Ben Noordhuis v0.8 * r3a6314d / (3 files in 3 dirs): net: fix net.Server.listen({fd:x}) error reporting - http://git.io/JiO37A
16:39:33  <isaacs>bnoordhuis: lgtm
16:39:37  <isaacs>oh, bert beat me to it
16:39:41  <isaacs>thanks
16:39:41  <bnoordhuis>heh
16:40:17  <bnoordhuis>okay, i'm afk again
16:48:32  * piscisaureus__quit (Ping timeout: 246 seconds)
16:50:05  * arlolrajoined
16:55:12  <indutny>soooo
16:58:55  <indutny>bnoordhuis: npm install -g callgrind.js
16:59:22  * ericktquit (Quit: erickt)
17:00:13  <indutny>bnoordhuis: basically you need to run `callgrind ./node --log-code script.js` first
17:00:32  <indutny>then callgrind.js --clog callgrind.out.pid --vlog v8.log > callgrind.out.js
17:02:35  <indutny>though, info doesn't look very useful, unfortunatelly
17:02:48  <indutny>as the most of the time js code is spending in runtimes
17:03:03  <indutny>and stubs
17:03:22  <indutny>and tree view isn't working for unknown reason (for me)
17:05:02  * arlolraquit (Quit: Linkinus - http://linkinus.com)
17:05:14  * AvianusAsleepuschanged nick to AvianFlu
17:06:45  * joe______joined
17:06:49  * joe______changed nick to joeandaverde
17:09:32  * piscisaureus_joined
17:15:27  * hij1nxquit (Quit: hij1nx)
17:19:21  <indutny>em...
17:19:31  <indutny>piscisaureus_: how does server.listen(0) works?
17:21:07  * ericktjoined
17:26:29  * mikealquit (Quit: Leaving.)
17:26:37  * mikealjoined
17:28:51  * bitprobejoined
17:30:04  * hij1nxjoined
17:33:10  * loladirojoined
17:36:14  * theColejoined
17:54:32  <piscisaureus_>indutny: it listens on a random port. why?
17:54:39  <indutny>nvm
17:54:41  <indutny>:)
17:54:46  <piscisaureus_>ok
17:54:51  <indutny>a lot of time passed, and I realized all the things
17:55:02  <piscisaureus_>oh
17:55:04  <piscisaureus_>good
18:02:13  * AngryParsleychanged nick to ggreer
18:02:14  * ggreerquit (Changing host)
18:02:14  * ggreerjoined
18:05:25  * brsonjoined
18:07:40  * TooTallNatejoined
18:12:41  <indutny>bnoordhuis: yt?
18:13:01  * joeandaverdequit (Quit: joeandaverde)
18:20:03  * hzjoined
18:25:50  * piscisaureus_quit (Ping timeout: 246 seconds)
18:27:02  <bnoordhuis>indutny: ih
18:28:06  <indutny>bnoordhuis: do you know why functions may not have callers in callgrind's log?
18:29:35  <bnoordhuis>indutny: no. when/where does that happen?
18:30:19  <bnoordhuis>indutny: i guess longjmp() style code could cause that but that shouldn't be the case in node
18:32:39  * hij1nxquit (Quit: hij1nx)
18:32:51  * hzquit
18:35:11  * hij1nxjoined
18:35:23  * `3rdEdenjoined
18:36:01  * mikealquit (Quit: Leaving.)
18:48:40  * hij1nxquit (Quit: hij1nx)
18:59:27  * loladiroquit (Ping timeout: 248 seconds)
19:08:56  * mikealjoined
19:17:58  * loladirojoined
19:30:09  * piscisaureus_joined
19:35:08  * piscisaureus_quit (Ping timeout: 246 seconds)
19:35:52  * EhevuTovquit (Ping timeout: 249 seconds)
19:39:22  * `3rdEdenquit (Quit: Leaving...)
19:39:51  <indutny>bnoordhuis: well, probably this is the cause on the boundary of v8 and node
19:40:54  * EhevuTovjoined
19:42:45  * `3rdEdenjoined
19:49:04  * mitsuhikojoined
19:49:38  <mitsuhiko>hey everybody
19:50:02  <mitsuhiko>if i cancel a timer, do i still need to uv_close it?
19:50:15  <creationix>mitsuhiko, I believe so
19:50:21  <creationix>I think nothing auto-closes
19:50:27  <mitsuhiko>hmm
19:50:29  <creationix>bnoordhuis ^?
19:50:42  <mitsuhiko>that's quite annoying because i have a tcp socket and a timer to handle connection timeouts
19:50:47  <mitsuhiko>which obviously want to die at the same time on timeout :)
19:51:05  <creationix>why not call cancel and close at the same time?
19:51:12  <creationix>close is async
19:51:17  <creationix>it waits till all pending requests are done
19:51:25  <mitsuhiko>creationix: because someone needs to free the memory
19:51:33  <mitsuhiko>which currently the on_close of the tcp connection does
19:51:50  <mitsuhiko>since i don't know in which order the closes are called i would have to refcount it
19:51:53  <mitsuhiko>or have one close call another close
19:52:18  <creationix>mitsuhiko, are you using a scripting engine or just plain C
19:52:21  <mitsuhiko>plain c
19:52:39  <creationix>chaining close calls seems sane
19:52:53  <creationix>but I don't know your structs
19:53:15  <mitsuhiko>it's basically a struct with a timer and a tcp handle on it :)
19:53:20  <mitsuhiko>for the sake of this discussion
19:53:42  <creationix>and either the tcp or timer may close first
19:54:08  <creationix>isn't the on_close callback optional?
19:54:18  <mitsuhiko>i am using the on_close callback for the free of the struct
19:54:33  <creationix>right, but which on_close
19:54:35  <creationix>the timer or the tcp
19:54:44  <mitsuhiko>currently of the tcp socket
19:54:56  <mitsuhiko>right now i just cancel the timer (stop) before the free
19:55:09  <creationix>I'm not sure, I'm still new to C
19:55:12  <creationix>I've just used libuv a lot
19:55:31  <creationix>but I think if you free the struct and then call close on the timer you should be safe
19:57:06  <creationix>hmm, not sure that would work
19:57:11  <creationix>sorry I'm not more help
19:57:29  <mitsuhiko>no, that's fine :)
19:57:36  <mitsuhiko>maybe someone has an answer later on
20:01:08  <indutny>mitsuhiko: you can just have a counter inside your structure
20:01:18  <indutny>and on on_close_cb handler
20:01:33  <indutny>which will decrement counter and free object once it's equal to zero
20:01:43  <mitsuhiko>indutny: right, see my refcounting remark above
20:01:54  <mitsuhiko>was just hoping i could avoid that :)
20:02:06  <indutny>well, why not?
20:02:27  <mitsuhiko>well, because if the timer would not have to be cv_close'd then it would just be easier :)
20:03:01  <indutny>well, may be it's possible to have timer binded to some specific handle
20:03:05  <indutny>that seems to be sane for me
20:03:08  <indutny>bnoordhuis: ^
20:03:19  <indutny>everyone is asking ben today :)
20:03:21  <indutny>and everyday
20:04:20  * hzjoined
20:05:47  * paddybyers_joined
20:07:23  * paddybyersquit (Ping timeout: 255 seconds)
20:07:23  * paddybyers_changed nick to paddybyers
20:08:03  * loladiroquit (Ping timeout: 246 seconds)
20:08:51  * loladirojoined
20:10:43  * piscisaureus_joined
20:12:36  * EhevuTovquit (Quit: This computer has gone to sleep)
20:12:39  * AndreasMadsenjoined
20:12:52  * AndreasMadsenquit (Remote host closed the connection)
20:26:47  * EhevuTovjoined
20:27:20  * loladiroquit (Ping timeout: 248 seconds)
20:37:30  * hij1nxjoined
20:42:26  * hij1nxquit (Ping timeout: 265 seconds)
20:45:06  * perezdquit (Quit: perezd)
20:51:29  * mmaleckiquit (Quit: leaving)
21:14:23  * `3rdEdenquit (Quit: Linkinus - http://linkinus.com)
21:17:30  * hij1nxjoined
21:17:37  * arlolrajoined
21:19:39  * xaqquit (Read error: Connection reset by peer)
21:20:09  * xaqjoined
21:24:15  * EhevuTovquit (Quit: This computer has gone to sleep)
21:32:59  * perezdjoined
21:34:41  * EhevuTovjoined
21:47:20  * TooTallNatequit (Ping timeout: 248 seconds)
21:55:00  * bitprobequit (Quit: Computer has gone to sleep.)
22:08:29  * hzquit
22:14:05  <bnoordhuis>here
22:14:58  <bnoordhuis>mitsuhiko: you need to close the handles separately
22:15:10  <bnoordhuis>mitsuhiko: so yes, you'll have to keep track of what's close and what not
22:15:26  <mitsuhiko>:-(
22:15:26  <mitsuhiko>oh well, i guess i will find a nice solution for that
22:15:41  <mitsuhiko>bnoordhuis: i suppose there is no built in support to set up a shutdown chain or something?
22:15:41  * TooTallNatejoined
22:15:48  <bnoordhuis>mitsuhiko: alas, no
22:16:24  <mitsuhiko>before i jump onto changing my code, maybe i am doing something wrong to start with: how do i handle timeouts ideally :)
22:16:33  <ik>robably
22:16:40  <mitsuhiko>as in: if the connection fails to yield useful results in n seconds i want to close it
22:16:41  <ik>p
22:17:35  <bnoordhuis>mitsuhiko: you need to use a timer
22:17:43  <bnoordhuis>so i guess you were on the right track already ::)
22:19:40  <mitsuhiko>heh. alright, then i guess i'm going to refcount this and use the same callback for both
22:22:05  * pieternquit (Ping timeout: 246 seconds)
22:28:59  <TooTallNate>isaacs: are you aware that http://nodejs.org/api/ has the 0.8.1 docs still?
22:29:13  <isaacs>TooTallNate: they're not actually 0.8.1 docs.
22:29:18  <isaacs>but yes. it's dumb.
22:29:20  <isaacs>lemme fix
22:30:27  <isaacs>TooTallNate: fixx0red!
22:30:42  <TooTallNate>isaacs: awesomeeeee thanx :)
22:34:47  * chobi_e_changed nick to chobi_e
22:35:30  * c4miloquit (Remote host closed the connection)
22:35:31  * hij1nxquit (Quit: hij1nx)
22:36:52  <bnoordhuis>TooTallNate: https://github.com/joyent/node/issues/3702
22:36:57  <bnoordhuis>revert, i guess?
22:37:13  <TooTallNate>ya i saw...
22:37:36  <TooTallNate>it could go either way, right?
22:37:44  <TooTallNate>but i guess since we already reverted the other one...
22:37:44  * chobi_echanged nick to chobi_e_
22:37:46  <TooTallNate>fuck
22:38:07  <TooTallNate>i'd argue that he should just have 1 event name, with an identifier as an argument
22:38:21  <TooTallNate>rather than keeping an infinite number of event names
22:38:56  <bnoordhuis>TooTallNate: i don't know, that doesn't seem like unreasonable use
22:39:12  <bnoordhuis>also, it breaks something that used to work fine
22:39:29  <TooTallNate>ya, i mean i'm ok with reverting
22:40:17  <bnoordhuis>TooTallNate: i'll let you do the honors, it's always nicer when your own name is on the revert commit :)
22:40:37  <TooTallNate>sure thing :)
22:44:40  <CIA-108>node: Nathan Rajlich v0.8 * r713b924 / (2 files in 2 dirs): Revert "events: don't delete the listeners array in removeListener()" - http://git.io/wVGpJA
22:45:59  <isaacs>TooTallNate: what's this about?
22:46:22  <isaacs>oh, i see.
22:46:23  <TooTallNate>oh just people complaining
22:46:23  <isaacs>yeah
22:46:28  <isaacs>kinda fair, i guess
22:46:41  <TooTallNate>no, it's fine since the removeAllListeners() patch was already reverted
22:48:35  * pieternjoined
22:54:29  * rendarquit
22:57:52  <piscisaureus_>movq %rax, 32(%rcx)
22:57:52  <piscisaureus_> movq %rdx, 40(%rcx)
22:57:52  <piscisaureus_>.L26:
22:57:52  <piscisaureus_> rep
22:57:52  <piscisaureus_> ret
22:58:01  <piscisaureus_>^-- wtf does that 'rep' mean ?
22:58:40  <bnoordhuis>more context?
22:59:00  <piscisaureus_>that is code generated by gcc
22:59:19  <bnoordhuis>never seen it do that
22:59:20  <piscisaureus_>you could try it yourself with gcc -S x86_64-gcc.c
22:59:32  <bnoordhuis>let me test that
23:00:05  <bnoordhuis>piscisaureus_: no rep in my output
23:00:09  <bnoordhuis>what version of gcc is it?
23:00:13  <piscisaureus_>bnoordhuis: -O3 ?
23:00:30  <bnoordhuis>oh wait, with -O3 i do have it
23:00:37  <piscisaureus_>bnoordhuis: 4.6.1
23:01:21  <piscisaureus_>bnoordhuis: so there are 2 weird things: "do they mean rep ret" (they are on separate lines, but rep is not an instruction)
23:01:33  <bnoordhuis>$ objdump -d x86_64-gcc.o | grep rep
23:01:34  <bnoordhuis> fd: f3 c3 repz retq
23:01:34  <bnoordhuis> 18f: f3 c3 repz retq
23:01:37  <piscisaureus_>and if so, why the heck would you want to do rep ret ?
23:01:38  <bnoordhuis>how awesome is that?
23:01:53  <piscisaureus_>insanity
23:02:02  <bnoordhuis>i guess it's a conditional return
23:04:08  <piscisaureus_>well
23:04:15  <piscisaureus_>masm refuses to assemble that
23:04:24  <piscisaureus_>1>openssl\crypto\bn\asm\x86_64-win32-masm.asm(237): error A2068: instruction prefix not allowed
23:05:10  <bnoordhuis>from https://github.com/mozilla/rust/blob/release-0.3/RELEASES.txt - Extensive work on libuv interface
23:05:15  <bnoordhuis>pfox__: ^ is that your work?
23:05:44  <pfox__>i am to blame, yes.
23:05:49  <bnoordhuis>congrats :)
23:05:54  <pfox__>brson was the adult in the room.
23:06:10  <piscisaureus_>bnoordhuis: btw, how do you disassemble on unix?
23:06:17  <bnoordhuis>piscisaureus_: objdump -d
23:06:22  <piscisaureus_>ok, thanks
23:07:45  <brson>I am trying to write a http request in for servo, in rust, using uv, right now
23:07:48  <brson>thanks, pfox__!
23:10:16  * pieternquit (Quit: pietern)
23:12:04  * pieternjoined
23:16:22  <piscisaureus_>bnoordhuis: ok, so after the "ret" there is garbage, no code
23:16:34  <piscisaureus_>bnoordhuis: so it's not a conditional return
23:16:41  <bnoordhuis>piscisaureus_: define garbage? i see nops
23:16:45  <piscisaureus_>bnoordhuis: right
23:16:56  <piscisaureus_>bnoordhuis: so, not code that should be reached
23:17:01  <piscisaureus_>bnoordhuis: I think its this: http://mikedimmick.blogspot.nl/2008/03/what-heck-does-ret-mean.html
23:17:03  <bnoordhuis>that's not garbage, that's for alignment
23:17:17  <piscisaureus_>sure
23:17:50  * theColequit (Quit: theCole)
23:20:01  * theColejoined
23:22:13  <bnoordhuis>piscisaureus_: that blog makes a convincing case though
23:22:36  <piscisaureus_>bnoordhuis: yeah, the only problem is that masm refuses to do "rep ret" or "rep; ret"
23:22:41  <piscisaureus_>because apparently it's not legal
23:22:43  <piscisaureus_>aaah
23:22:48  <piscisaureus_>so I think I will do "ret 0"
23:23:18  * arlolraquit (Quit: Leaving...)
23:33:46  <bnoordhuis>piscisaureus_: https://gist.github.com/742c6d86d57c6eafa900 <- the repz is indeed a no-op, the retq always executes
23:34:14  <mitsuhiko>does anyone of you use msgpack with libuv?
23:34:29  <mitsuhiko>it claims to have a streaming interface but i fail to see how it works without filling the buffer up
23:34:49  <piscisaureus_>bnoordhuis: ok, cool
23:35:01  <piscisaureus_>mitsuhiko: I don't
23:35:05  <bnoordhuis>mitsuhiko: never used it
23:37:06  <mitsuhiko>heh. thanks though :)
23:47:54  <piscisaureus_>Ok, build succeeded. The question now is: will it not crash.
23:47:59  <piscisaureus_>WILL IT NOT CRASH
23:48:00  <LOUDBOT>DO PEOPLE EAT THEM RAW OR WHAT
23:55:36  <tjfontaine>HOW CAN IT CRASH THE HERO WROTE IT
23:55:37  <LOUDBOT>YOU ARE NOT DOWN WITH THE BEEF BUS
23:55:58  * paddybyersquit (Quit: paddybyers)
23:58:03  <CIA-108>node: Ben Noordhuis v0.8 * racbfc40 / node.gyp : build: fix add-on loading on freebsd - http://git.io/7Bvjjg
23:59:03  <TooTallNate>^ nice one ben!
23:59:25  <bnoordhuis>thanks :)