00:00:13  <piscisaureus>bnoordhuis: you should also check to what extent String::NewSymbol already has lazy init functionality
00:00:15  <piscisaureus>I think it hase
00:00:27  <piscisaureus>s/e$//
00:00:41  <bnoordhuis>yeah, i realize that
00:00:52  <bnoordhuis>but i want to keep the diffs as simple as possible
00:01:21  <piscisaureus>k
00:33:04  <igorzi>piscisaureus: does v8/node build for 64bit?
00:33:17  <piscisaureus>igorzi: yes but not on windows atm
00:33:29  <piscisaureus>igorzi: its not difficult to support I think
00:33:47  <igorzi>piscisaureus: is it v8 that's not building on windows?
00:33:47  <piscisaureus>igorzi: but v8.gyp has no script for an x64 build
00:34:10  <piscisaureus>igorzi: it's v8.gyp that's missing it
00:34:16  <igorzi>piscisaureus: got it
00:34:36  <piscisaureus>Sorry, I thought the first time I said that it was black holed :-)
00:35:28  <igorzi>there are no 64bit chrome builds?
00:35:39  <igorzi>(on windows)
00:35:53  <piscisaureus>igorzi: nope
00:36:03  <piscisaureus>I think there are no x64 chrome builds at all
00:36:44  <piscisaureus>it could be that the latest v8 does have x64 btw - you should check it
00:41:35  <piscisaureus>igorzi: yt?
00:42:09  <piscisaureus>igorzi: do you know if KeWaitForSingleObject accepts the same type of handle as WaitForSingleObject?
00:45:41  <igorzi>piscisaureus: i think so.. i can find out
01:24:40  * dapquit (Quit: Leaving.)
01:27:38  * pieternquit (Quit: pietern)
02:01:44  * ericktquit (Quit: erickt)
02:23:21  * AvianFluquit (Quit: Leaving)
02:34:57  * bnoordhuisquit (Read error: Connection reset by peer)
02:36:07  * ericktjoined
02:41:26  * piscisaureusquit (Ping timeout: 255 seconds)
02:45:36  * piscisaureus_joined
02:51:22  * brsonquit (Quit: leaving)
02:53:20  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:54:24  * piscisaureus_joined
03:12:04  <isaacs>I think i got the fs.open invalid arguments bug fixed.
03:12:22  <isaacs>at least, the bit of it that was crashing linux.
03:20:03  * isaacsquit (Quit: isaacs)
03:46:13  * AvianFlujoined
05:43:03  * isaacsjoined
06:11:54  * ericktquit (Quit: erickt)
06:15:32  <isaacs>any of you guys use travis? this thing is pretty rad
06:35:18  * felixgejoined
06:44:32  * mikealquit (Quit: Leaving.)
06:47:17  * paddybyersquit (Quit: paddybyers)
06:48:03  * paddybyersjoined
07:10:58  * paddybyersquit (Quit: paddybyers)
07:13:34  * mikealjoined
07:34:28  * luxigoquit (Read error: Operation timed out)
07:34:43  * luxigojoined
07:46:39  * paddybyersjoined
07:49:40  * mralephjoined
07:54:12  <CIA-95>node: seebees v0.6 * raab958b / (2 files in 2 dirs):
07:54:12  <CIA-95>node: OutgoingMessage.prototype.write does not take Array
07:54:12  <CIA-95>node: Changed the type checking for OutgoingMessage.prototype.write so it only accepts string and Buffer.
07:54:12  <CIA-95>node: And test.
07:54:12  <CIA-95>node: Fixes #2162
07:54:12  <CIA-95>node: Fixes #2208 - http://git.io/rQocKw
08:22:47  * alessioalexjoined
08:22:53  * alessioalexpart
08:27:02  * kuebkjoined
08:37:12  * mralephquit (Quit: Leaving.)
08:45:53  * piscisaureus_quit (Ping timeout: 252 seconds)
09:51:37  * mikeal1joined
09:51:58  * mikealquit (Read error: Connection reset by peer)
10:01:33  * f00barzjoined
10:09:35  * f00barzchanged nick to eddyb
10:24:21  <eddyb>https://github.com/joyent/node/issues/2147
11:23:25  * sh1mmerquit (Quit: sh1mmer)
11:28:56  <eddyb>is that going to be fixed soon?
11:29:01  <eddyb>or should I attempt it?
11:38:59  * piscisaureus_joined
11:40:01  <eddyb>I love github
11:41:37  <eddyb>piscisaureus_: hi
11:42:02  <piscisaureus_>eddyb: is that #2147 fix you just landed ready for prime time?
11:42:39  <eddyb>it should be usable in 0.6
11:42:47  <eddyb>I'm not very sure if it works, btw
11:43:06  <piscisaureus_>eddyb: why are you not sure?
11:43:27  <eddyb>it was 3-click patch, let me test it real quick and see how it behaves
11:45:17  <eddyb>oh crap
11:45:27  <eddyb>my first github pull request and I failed
11:47:30  <eddyb>no, node is broken
11:47:48  <CIA-95>node: Bert Belder v0.6 * r6ed721a / (deps/uv/src/win/core.c deps/uv/src/win/req.c): Upgrade libuv to fce90652ed38ab0925131ed141972354a494933b - http://git.io/zcK8pQ
11:49:00  <eddyb>bad makefile, let me cleanup and try again
11:50:06  <piscisaureus_>eddyb: you'll also have to sign the CLA if you didn't already
11:50:27  <eddyb>I wanted just to show the fix
11:50:41  <eddyb>but don't use it, it doesn't work the way I thought it does
11:56:14  <eddyb>ok, so now it should be working
12:19:03  <eddyb>piscisaureus_: adammw said toplevel_dir
12:19:07  <eddyb>but I think it's bogus
12:19:13  <eddyb>I can't find a good value
12:29:32  <eddyb>darn gyp
12:32:01  <eddyb>piscisaureus_: toplevel set to .. puts Makefile in . instead of out
12:33:39  <eddyb>YES
12:33:43  <eddyb>I think I got it :)
12:34:06  <eddyb>nope, dammit
12:42:38  <eddyb>piscisaureus_: I say gyp is bugged
12:43:15  <eddyb>gyp_binary = gyp.common.FixIfRelativePath(params['gyp_binary'], options.toplevel_dir)
12:43:54  <eddyb>so in order to get it to call gyp from node root, you need to set toplevel_dir to that
12:44:03  <eddyb>but, only relative paths matter, so you set '..'
12:44:14  <eddyb>and then, the Makefile gets put there
12:44:22  <eddyb>so it either shouldn't care about relativity
12:44:31  <eddyb>or not put Makefile in the toplevel-dir
12:48:11  <eddyb>actually, gyp shouldn't put Makefile in toplevel-dir, it's a bug that it does
12:54:24  <eddyb>FFS
13:03:25  <eddyb>AHAHAHAHAHAHAHAHA
13:03:45  <eddyb>piscisaureus_: gyp's functions are poorly written to return path relative to current directory
13:04:10  <eddyb>so we should, in fact, mkdir out && cd out && call gyp from them
13:05:02  <eddyb>or use Unrelative
13:16:50  <eddyb>YESSS
13:16:55  <eddyb>piscisaureus_: I've done it
13:26:34  <eddyb>crap, after first regeneration, I get this
13:26:35  <eddyb>cmd_regen_makefile = /home/eddy/Projects/node/tools/gyp_node -fmake --ignore-environment "--toplevel-dir=/home/eddy/Projects/node" -I/home/eddy/Projects/node/common.gypi -I/home/eddy/Projects/node/options.gypi -I/home/eddy/Projects/node/common.gypi -I/home/eddy/Projects/node/options.gypi "--depth=/home/eddy/Projects/node" "-Goutput_dir=/home/eddy/Projects/node/out" "-Goutput_dir=/home/eddy/Projects/node/out" "--generator-output=/home/eddy/Projects/node/out"
13:26:37  <eddyb>"-Dtarget_arch=ia32" "-Dcomponent=static_library" "-Dlibrary=static_library" "-Dtarget_arch=ia32" "-Dcomponent=static_library" "-Dlibrary=static_library" /home/eddy/Projects/node/out/node.gyp /home/eddy/Projects/node/out/node.gyp
13:26:51  <eddyb>while the original one was
13:26:53  <eddyb>cmd_regen_makefile = /home/eddy/Projects/node/tools/gyp_node -fmake --ignore-environment "--toplevel-dir=/home/eddy/Projects/node" -I/home/eddy/Projects/node/common.gypi -I/home/eddy/Projects/node/options.gypi "--depth=." "-Goutput_dir=/home/eddy/Projects/node/out" "--generator-output=/home/eddy/Projects/node/out" "-Dtarget_arch=ia32" "-Dcomponent=static_library" "-Dlibrary=static_library" /home/eddy/Projects/node/node.gyp
13:33:31  * piscisaureus__joined
13:34:57  * piscisaureus_quit (Ping timeout: 244 seconds)
13:38:24  * bnoordhuisjoined
13:41:27  * bnoordhuisquit (Client Quit)
13:41:41  * bnoordhuisjoined
13:45:01  <bnoordhuis>piscisaureus__: isolates again, which approach do you like best? https://github.com/bnoordhuis/node/commit/57f4b9e or https://github.com/joyent/node/commit/386178e
13:45:52  <bnoordhuis>the first one uses macro magic, it's actually not that intrusive and it let's you compile a single-loop version of node without the performance penalty of tls
13:46:05  <bnoordhuis>the second one is more explicit (and more intrusive)
13:46:16  <piscisaureus__>bnoordhuis: do you have a handle on how big the performance penalty actually is?
13:46:48  <bnoordhuis>piscisaureus__: not really, no hard numbers
13:47:29  <bnoordhuis>in the grand scheme of things it's probably not that ba
13:47:30  <bnoordhuis>*bad
13:48:47  <piscisaureus__>bnoordhuis: I don't know what the EMPTY macro is for
13:49:03  <bnoordhuis>piscisaureus__: to check if a handle is empty / not initialized
13:49:10  <bnoordhuis>don't get too hung up on the names just yet
13:49:18  <piscisaureus__>+# define EMPTY(name) (__tls_##name)
13:49:26  <piscisaureus__>looks like NOT_EMPTY to me
13:49:40  <bnoordhuis>piscisaureus__: with HAVE_ISOLATES the handle is actually a pointer to a handle
13:49:49  <bnoordhuis>so that's a !NULL check
13:50:07  <bnoordhuis>oh, right - i botched that one :)
13:50:27  <bnoordhuis>that's what you get when you do last-minute rebasing at 4 am in the morning
13:50:27  <piscisaureus__>bnoordhuis: but yeah forget about the naming
13:51:02  <piscisaureus__>bnoordhuis: but honestly, I like the second patch better
13:51:16  <bnoordhuis>piscisaureus__: okay, why?
13:51:28  <piscisaureus__>bnoordhuis: and I rethought this last night, I don't really think we want to publish non-isolate versions
13:51:50  <piscisaureus__>bnoordhuis: but the upside of patch 1 is that we can easily compare isolate and nonisolate performance
13:52:02  <piscisaureus__>bnoordhuis: so I would suggest to go for option 1
13:52:27  <bnoordhuis>piscisaureus__: yes, that - and it'd be nice to have a baseline non-isolate version for debugging ("does the crash also happen when you compile with -DHAVE_ISOLATES=0")
13:53:21  <eddyb>hmm, why are all those static symbols so important for you?
13:53:30  <piscisaureus__>bnoordhuis: but actually I'd like a struct-based solution even better :-)
13:54:02  <bnoordhuis>piscisaureus__: i suppose that's another advantage of the macro solution, we can move stuff into structs later on if we so choose
13:54:12  <piscisaureus__>yeah - that's good too
13:54:30  <piscisaureus__>and after that we can get rid of the macros again :-)
13:54:31  <piscisaureus__>maybe
13:54:39  <eddyb>seriously?
13:54:44  <eddyb>all those static symbols
13:54:44  <piscisaureus__>after the euro has collapsed
13:54:51  <piscisaureus__>hmm maybe not
13:55:27  <piscisaureus__>the patch is actually not that intrusive indeed
13:55:31  <bnoordhuis>eddyb: simple - i don't want to rewrite half of node
13:55:52  <piscisaureus__>bnoordhuis: btw - I have one remark
13:56:01  <eddyb>I mean, I would understand if you would initialize them to some v8::SymbolThatStoresAPrecomputedHash instance...
13:56:09  <eddyb>bnoordhuis: oh, so they aren't important?
13:56:15  <bnoordhuis>piscisaureus__: speak to me
13:56:32  <bnoordhuis>eddyb: you can't init non-primitive thread locals
13:56:38  <piscisaureus__>bnoordhuis: you may want to rewrite node::Start so that it barely does anything other than set argv
13:56:47  <piscisaureus__>bnoordhuis: (while you're at it)
13:57:03  <bnoordhuis>piscisaureus__: yes, that's the plan - most of that logic moves into the functions that set up isolates
13:57:07  <eddyb>bnoordhuis: no, I mean... not in this context. in node/master
13:57:26  <eddyb>are they a huge factor in performance?
13:57:42  <bnoordhuis>eddyb: yes-ish
13:57:48  <piscisaureus__>bnoordhuis: because, for windows we're going to have a node::Start(int argc, WCHAR* argv[])
13:58:12  <piscisaureus__>bnoordhuis: also, we need to move chdir and cwd to libuv
13:58:33  <bnoordhuis>piscisaureus__: sure
13:58:36  <piscisaureus__>vanwege tekensetproblemen
13:59:06  <bnoordhuis>om niet te zeggen: karaktercollectieuitdagingen
13:59:40  <eddyb>niet german for me :P
13:59:50  <bnoordhuis>eddyb: not german, dutch!
14:00:02  <bnoordhuis>i love how in dutch you can string words together ad infinitum
14:00:02  <piscisaureus__>dat is geen Duits vriend!
14:00:06  <bnoordhuis>haha
14:00:07  <eddyb>~same thing
14:00:44  <eddyb>bnoordhuis: so thaaaat's where microsoft got that from...
14:00:55  <piscisaureus__>hottentottensoldatententententoonstelling
14:01:00  <eddyb>ffs
14:01:23  <bnoordhuis>try doing that in that english of yours
14:01:53  <piscisaureus__>eh
14:01:55  <piscisaureus__>Hottentot soldiers tents exhibition ?
14:02:01  <eddyb>bnoordhuis: but what's the overhead of not having static symbols, but initializing them every time they're used?
14:02:19  <eddyb>ascii->utf16?
14:02:35  <eddyb>uhm, that would suck
14:03:00  <bnoordhuis>eddyb: yes, and lots of malloc'ing - unless you use String::NewSymbol but that means v8 has to hash the string every time you call it
14:03:22  <bnoordhuis>so caching the result is faster
14:03:33  <eddyb>wait, NewSymbol caches the hash?
14:03:49  <eddyb>the actual hash used later when accessing object properties?
14:04:27  <bnoordhuis>eddyb: no, String::NewSymbol is like an atom in erlang or a symbol in ruby
14:04:52  <bnoordhuis>it's created once and then reused - so if you call String::NewSymbol("foo") again, you get a reference to the original "foo"
14:05:00  <eddyb>ah
14:05:09  <eddyb>so the overhead is just hash computation
14:05:16  <eddyb>that's perfect
14:05:18  <bnoordhuis>but v8 needs to hash the string and look it up in the atom table
14:05:30  <eddyb>can I hook into v8's hash function?
14:05:53  <eddyb>because, if you didn't know already, in C++11 you can have compile-time hashing :)
14:06:51  <eddyb>so, technically, it would be even faster than static symbols
14:07:30  <piscisaureus__>looks like runner-mei is working late today
14:07:35  <piscisaureus__>https://github.com/runner-mei/libuv/commit/f56d9e0699b477a74a58430c8b8551ca66eab28b#diff-6
14:08:05  <eddyb>just add _Sym after the string and it computes the hash at compile-time :D
14:08:25  <piscisaureus__>wut?
14:09:04  <piscisaureus__>I bet you need Boost for that
14:09:07  <eddyb>piscisaureus__: what, are you not familiar with C++11 a.k.a. C++AWESOME
14:09:24  <eddyb>I wrote a compile-time hash in #geordi the other day
14:09:36  <eddyb>you don't even need a standard library plugged in
14:09:58  <eddyb>lookup user-defined literals
14:10:05  <eddyb>and constexpr functions
14:11:02  <eddyb>bnoordhuis: btw, I started fixing https://github.com/joyent/node/issues/2147
14:11:09  <eddyb>after about an hour, I gave up
14:11:13  <eddyb>gyp sucks
14:11:40  <eddyb>it's one of those things that works for what it was built, but becomes a PITA for anything else
14:12:02  <eddyb>I suggest a regen rule in the main Makefile
14:13:01  <eddyb>out/Makefile: common.gypi deps/uv/uv.gyp deps/http_parser/http_parser.gyp deps/zlib/zlib.gyp deps/v8/build/common.gypi deps/v8/tools/gyp/v8.gyp node.gyp options.gypi
14:15:14  <eddyb>and make ./configure just output options.gyp
14:23:01  <eddyb>bnoordhuis: https://github.com/joyent/node/issues/2147
14:23:09  <eddyb>there, simple fix
14:23:23  <eddyb>I'm sick of gyp and its python
14:30:21  * bnoordhuisquit (*.net *.split)
14:31:57  <eddyb>piscisaureus__: proper patch now
14:32:00  <eddyb>coming your way
14:32:31  <eddyb>" An unexpected error seems to have occurred. Why not try refreshing your page? Or you can contact us if the problem persists." - nice timing...
14:34:31  <eddyb>and I got #2222: https://github.com/joyent/node/pull/2222
14:36:20  * bnoordhuisjoined
14:39:18  <eddyb>bnoordhuis: if you missed my messages, I fixed that bugger with a simple squash
14:39:26  <eddyb>and opened a pull request
14:39:45  <bnoordhuis>i think there was a net split
14:39:52  <bnoordhuis>so i don't know what we're talking about
14:40:43  <eddyb>yeah, I saw the netsplit
14:40:50  <eddyb>bnoordhuis: https://github.com/joyent/node/pull/2222
14:43:04  <bnoordhuis>eddyb: lgtm, seems like it should work
14:44:10  <eddyb>I tried it
14:44:13  <eddyb>and it works fine
14:44:47  <eddyb>the best gyp-hack-based fix I could manage failed to write a sane regen command after its first regen
14:45:06  <eddyb>this one is gyp-proof :)
14:48:29  <eddyb>omg...
14:48:43  <eddyb>the github tab is eating half a gigabyte
14:56:16  <bnoordhuis>yeah, i think github is b0rked
14:59:13  <eddyb>http://html5demos.com/history
14:59:16  <eddyb>mind=blown
14:59:24  <eddyb>so that's how it was doing it
15:10:24  * AvianFluquit (Quit: Leaving)
15:31:58  <CIA-95>node: Ben Noordhuis master * ra033261 / configure : build: DRY configure script - http://git.io/c08L8A
15:35:31  * pieternjoined
15:36:19  <eddyb>bnoordhuis: does jscc get triggered every time make is run?
15:37:12  <bnoordhuis>eddyb: no (i think)
15:38:02  <eddyb>because I have an ACTION setup just like jscc in node
15:38:11  <eddyb>and it gets triggered no-matter-what
15:38:16  <eddyb>and it pisses me off
15:39:50  <bnoordhuis>eddyb: you probably need to set the right type and toolsets
15:40:04  <bnoordhuis>and make sure your inputs and outputs are constant
15:40:06  <eddyb>uhm, let me check
15:40:50  <eddyb>it's just like in node
15:41:01  <eddyb>except other inputs, outputs and actions
15:43:48  <eddyb>also, there's no way to have source-file-like behaviour?
15:44:01  <eddyb>so each of the input file gets a call to the action?
15:44:09  <eddyb>or it's just another type?
15:54:20  * eddybquit (Ping timeout: 255 seconds)
15:54:33  <bnoordhuis>just when i want to reply >:(
15:55:59  * eddybjoined
15:55:59  * eddybquit (Changing host)
15:55:59  * eddybjoined
15:57:54  <eddyb>bnoordhuis: is there a way to see errors from tests?
15:58:01  <eddyb>could be a good TODO list
15:59:15  <bnoordhuis>eddyb: i usually run the tests with `make test-all 2>&1 | tee test.log`
15:59:26  <bnoordhuis>we also have buildbots btw
16:00:25  <eddyb>bnoordhuis: but why don't I see anything if I run the tests?
16:00:37  <eddyb>there's a number with + next to it and one with - next to it
16:01:09  <eddyb>it one of those "if(stdout.isPipe())...." tricks?
16:01:20  <eddyb>*it's
16:01:21  <bnoordhuis>eddyb: does it tell you what test it's currently running?
16:01:26  <eddyb>yeah
16:01:30  <bnoordhuis>okay, that's good
16:01:39  <bnoordhuis>it only prints more if the test fails
16:01:53  <eddyb>it doesn't print more
16:02:06  <eddyb>and no, I haven't implemented huge parts of node
16:02:10  <eddyb>it can't just... work
16:03:00  <bnoordhuis>eddyb: i'll let you in on a secret
16:03:25  <bnoordhuis>major parts of node are simply -> for (i=0; i < 10*1024*1024; i++)
16:03:33  <bnoordhuis>so it looks like we're doing something
16:03:38  <eddyb>you mean of the tests
16:04:07  <eddyb>oooh
16:04:13  <bnoordhuis>i'm teasing you ;)
16:04:13  <eddyb>it prints when it finishes
16:04:15  <eddyb>niiice
16:04:34  <bnoordhuis>you're working on a mac?
16:04:42  <isaacs>eddyb: it's so that when someone posts a benchmark that makes node look bad, we can just take out some loops, and say "Look! Now it's faster!!"
16:04:43  <eddyb>linux
16:05:02  <bnoordhuis>and be the heroes of the day!
16:05:16  <eddyb>I love the humor
16:05:28  <eddyb>but I have a node.js rewrite to finish :D
16:05:28  <bnoordhuis>eddyb: some people have reported that their console buffers everything until `make test-all` completes
16:05:34  <bnoordhuis>never seen it myself though
16:05:42  <eddyb>are you kidding me?
16:05:52  <bnoordhuis>eddyb: do i ever?
16:05:54  <eddyb>it's test.py that does that
16:06:06  <eddyb>it's *very* obvious
16:06:20  <bnoordhuis>only it doesn't on my linux or sunos machines
16:07:01  <eddyb>bnoordhuis: well, in the log it looks like you say
16:07:22  <eddyb>but I'm guessing without tee it just has a nice interface
16:09:26  * dapjoined
16:12:10  <eddyb>bnoordhuis: any ideas on that gyp stuff?
16:12:30  <bnoordhuis>eddyb: no, sorry
16:12:32  <eddyb>I can't test if all my generated stuff regens every time
16:12:39  <eddyb>Error: NativeModule fs doesn't exist
16:12:49  <eddyb>just because the fs module didn't get to compile
16:16:06  <eddyb>omg
16:16:16  <eddyb>gyp.... gyp is so feature-less
16:16:21  <eddyb>I can't use that crap
16:16:37  <eddyb>I'll have to hand-craft makefiles to make this work
16:16:51  * ericktjoined
16:34:18  * felixgequit (Quit: felixge)
16:35:25  * kuebkpart
16:36:16  * felixgejoined
16:36:16  * felixgequit (Changing host)
16:36:16  * felixgejoined
16:42:54  <CIA-95>node: Ben Noordhuis isolates * rf867220 / (src/node.cc src/node.h): core: isolate-ify (+6 more commits...) - http://git.io/p9NH6Q
16:56:50  <piscisaureus__>bnoordhuis: http://groups.google.com/group/nodejs/browse_thread/thread/d36c630d350c375b <-- you know about that?
17:35:32  <eddyb>what's __thread?
17:35:36  <eddyb>part of C++11?
17:38:00  <dap>it's a thread-local specifier.
17:38:56  <dap>http://en.wikipedia.org/wiki/Thread-local_storage#C.2B.2B
17:39:05  <dap>unfortunately, it's under "C++", but it's a C-ism.
17:44:34  <eddyb>dap: and for visual C++, #define __thread __declspec(thread)
17:44:36  <eddyb>?
17:45:12  <eddyb>oh, thread_local is the new way of doing it, in C++11
17:45:34  <dap>sorry, I've only used __thread, and only with gcc and Sun Studio.
17:57:41  * sh1mmerjoined
18:22:32  <CIA-95>node: Ben Noordhuis v0.6 * r5fea005 / lib/sys.js : sys: print stack trace if NODE_DEBUG=sys - http://git.io/UG6a7A
18:23:39  <bnoordhuis>eddyb: it's a gcc-ism, __declspec(thread) is the msvc analog
18:27:08  <isaacs>call today? or no?
18:27:50  <bnoordhuis>isaacs: sure - don't have much to tell though
18:32:48  <eddyb>bnoordhuis: I got that :)
18:37:38  <CIA-95>node: [email protected] master * reda1538 / Makefile :
18:37:38  <CIA-95>node: build: regenerate out/Makefile from within the main Makefile
18:37:38  <CIA-95>node: Fixes #2147. - http://git.io/IlFV4A
18:37:42  <CIA-95>node: Ben Noordhuis isolates * rd4f1972 / (src/node.cc src/node.h): core: isolate-ify (+6 more commits...) - http://git.io/n4mR9w
18:38:03  * eddybfeels better :)
18:42:05  * igorziquit (Ping timeout: 265 seconds)
18:46:52  * igorzijoined
18:50:19  * AvianFlujoined
18:51:04  <isaacs>my update: travis-ci is awesome.
18:51:14  <isaacs>found and fixed the core issue, this uncovered another one
18:53:52  * AvianFlu_joined
18:54:14  <rmustacc>Should node master be building openssl by default?
18:54:15  * piscisaureus__quit (Ping timeout: 244 seconds)
18:54:44  * AvianFluquit (Ping timeout: 260 seconds)
18:55:01  * piscisaureus__joined
18:58:05  * AvianFlu_changed nick to AvianFlu
18:59:10  * brsonjoined
19:01:41  <igorzi>piscisaureus__ bnoordhuis: are we having a call today? or?
19:02:02  <bnoordhuis>rmustacc: no, only if there is no system openssl
19:02:22  <bnoordhuis>igorzi: if you want to - the only news i have is that i want to postpone the release to thursday
19:02:36  <piscisaureus__>igorzi: having a call with claudio?
19:02:40  <piscisaureus__>or just the 3 of us?
19:04:00  <rmustacc>bnoordhuis: Okay, well, I definitely have a system openssl.
19:04:06  <piscisaureus__>bnoordhuis:: get online
19:04:17  <bnoordhuis>piscisaureus__: woof
19:04:30  <bnoordhuis>rmustacc: do you have pkg-config installed?
19:04:59  <rmustacc>Why would I have pkg-config installed on OS X?
19:05:19  <rmustacc>I would hope pkg-config isn't now a build dep for node.
19:05:55  <bnoordhuis>rmustacc: not so much a dependency as a nice-to-have
19:06:23  <bnoordhuis>but it'll build the bundled openssl without pkg-config
19:07:15  <rmustacc>Okay, that's a bit unfortuntae.
19:07:17  <rmustacc>*unfortunate
19:07:47  <rmustacc>Did we get rid of all detection of it?
19:08:05  * mikeal1quit (Quit: Leaving.)
19:08:09  <bnoordhuis>rmustacc: kind of - the openssl check in the new configure script is kind of rudimentary
19:08:48  <rmustacc>bnoordhuis: Gotcha. Mind if I file an issue on it then?
19:09:07  <bnoordhuis>rmustacc: sure - patches are even more welcome :)
19:10:23  <rmustacc>bnoordhuis: I'll try and get some cycles to do it. I'm mired in lots of kernel woes right now.
19:31:03  * mikealjoined
19:34:23  * mikealquit (Client Quit)
19:36:11  * bnoordhuisis back in a couple of hours
19:36:41  <isaacs>npm passing on linux: http://travis-ci.org/#!/isaacs/npm/builds/354953
19:40:33  * bnoordhuisquit (Ping timeout: 252 seconds)
19:43:40  * mikealjoined
19:47:30  * piscisaureus__quit (Ping timeout: 260 seconds)
20:03:29  <isaacs>igorzi: this is odd. when i try to fs.stat or fs.lstat a symlink on a folder mapped into vmware, it gives me an EACCES
20:03:41  <isaacs>running as administrator
20:04:27  <isaacs>may well be a vmware bug.
20:07:23  <igorzi>isaacs: "on a folder mapped into vmware".. is that a local folder?
20:07:33  <isaacs>yeah
20:07:37  <isaacs>well, it's on the host
20:07:56  <isaacs>it's just odd that lstat is giving me eacces
20:08:37  <isaacs>i can stat other files just fine
20:10:37  <igorzi>isaacs: can you read/write from/to it?
20:11:06  <isaacs>i can read other files, but not write to them (actually don't have write perms)
20:11:11  <isaacs>but i can't read the symlink, no
20:13:32  * pieternquit (Quit: pietern)
20:40:55  <eddyb>anyone here with knowledge of the v8 internals?
20:41:08  <eddyb>I'm trying to do something tricky
20:41:24  * isaacsquit (Quit: isaacs)
20:42:30  * piscisaureus_joined
20:45:30  * mikealquit (Quit: Leaving.)
20:48:07  * piscisaureus_quit (Ping timeout: 244 seconds)
20:49:43  * piscisaureus_joined
21:03:52  * pieternjoined
21:04:25  * pieternquit (Remote host closed the connection)
21:05:05  * pieternjoined
21:06:43  * pieternquit (Client Quit)
21:09:46  * pieternjoined
21:11:47  * eddybquit (Quit: Konversation terminated!)
21:12:45  * pieternquit (Client Quit)
21:15:24  * pieternjoined
21:17:39  * mralephjoined
21:25:25  * pieternquit (Quit: pietern)
22:20:02  * pieternjoined
22:20:33  * bnoordhuisjoined
22:30:50  * felixgequit (Quit: felixge)
22:39:34  * mikealjoined
22:42:45  * isaacsjoined
22:42:59  * mralephquit (Quit: Leaving.)
22:48:08  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
22:48:35  <rmustacc>Can someone confirm that I read msdn right and Windows does have abort()?
22:48:36  * piscisaureus_joined
22:49:04  <CoverSlide>windows is pro-choice
22:52:57  <piscisaureus_>rmustacc: yes, windows does have abort()
22:53:16  <piscisaureus_>rmustacc: I think it throws an SEH exception
22:54:23  <rmustacc>piscisaureus_: Okay. Thanks. We were talking about something here and I wanted to make sure I wasn't crazy.
23:34:02  * sh1mmerquit (Quit: sh1mmer)
23:34:17  * sh1mmerjoined
23:37:29  * sh1mmerquit (Client Quit)
23:38:01  * sh1mmerjoined
23:38:46  * AvianFlu_joined
23:39:12  * isaacsquit (Read error: Connection reset by peer)
23:39:36  * isaacsjoined
23:41:45  * AvianFluquit (Ping timeout: 260 seconds)
23:45:23  * AvianFlu_changed nick to AvianFlu
23:49:17  * isaacs_joined
23:49:18  * isaacsquit (Read error: Connection reset by peer)
23:49:18  * isaacs_changed nick to isaacs
23:53:29  * benviequit