00:48:20  * benviejoined
00:50:01  * indutnychanged nick to indutny_away
01:03:17  * mikealquit (Quit: Leaving.)
01:35:51  * brsonquit (Read error: Operation timed out)
02:08:12  * abraxasjoined
02:37:43  * mikealjoined
02:57:55  * mikealquit (Quit: Leaving.)
03:46:58  * bnoordhuisquit (Ping timeout: 260 seconds)
04:51:52  * mikealjoined
04:55:20  * mikealquit (Client Quit)
04:57:12  * mikealjoined
06:34:36  * paddybyersjoined
07:52:13  * stephankquit (Quit: *Poof!*)
07:56:50  * orlandovftwjoined
07:56:50  * orlandovftwquit (Client Quit)
07:57:05  * orlandovftwjoined
08:04:05  * sh1mmerquit (Quit: sh1mmer)
08:46:20  * orlandovftwquit (Ping timeout: 250 seconds)
09:06:00  * einaroschanged nick to einaros-
09:19:27  * indutny_awaychanged nick to indutny
09:28:43  * rendarjoined
09:29:03  * txdvjoined
10:55:19  * mikealquit (Ping timeout: 246 seconds)
10:58:03  * abraxasquit
11:59:32  * mmalecki[zzz]changed nick to mmalecki
12:29:19  * skomskijoined
12:40:48  * skomskiquit (Quit: skomski)
12:40:55  * skomskijoined
12:50:57  * skomskiquit (Quit: skomski)
12:59:52  * bnoordhuisjoined
13:26:42  <CIA-99>node: Your Name master * rfb47a33 / test/common.js : test: changed instances of == to === - http://git.io/qTHl3w
13:27:26  <mmalecki>Your Name is a good guy!
13:32:58  * AndreasMadsenjoined
13:41:09  * travis-cijoined
13:41:09  <travis-ci>[travis-ci] joyent/node#613 (master - fb47a33 : Your Name): The build is still failing.
13:41:09  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/70e6889...fb47a33
13:41:09  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/900023
13:41:09  * travis-cipart
13:43:12  <tjfontaine>cute
15:04:55  <CIA-99>libuv: Ben Noordhuis master * r49d4e18 / src/unix/linux/core.c :
15:04:55  <CIA-99>libuv: linux: use CLOCK_BOOTTIME if available
15:04:55  <CIA-99>libuv: Use CLOCK_BOOTTIME to calculate the system uptime. Fall back to CLOCK_MONOTONIC
15:04:55  <CIA-99>libuv: if CLOCK_BOOTTIME is not available (all pre-2.6.39 kernels).
15:04:55  <CIA-99>libuv: The problem with CLOCK_MONOTONIC is that it doesn't increase when the system is
15:04:55  <CIA-99>libuv: suspended, making the uptime differ from what the uptime(1) tool reports. - http://git.io/-j5JHA
15:06:56  * travis-cijoined
15:06:56  <travis-ci>[travis-ci] joyent/libuv#143 (master - 49d4e18 : Ben Noordhuis): The build is still failing.
15:06:56  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/379ca42...49d4e18
15:06:56  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/900410
15:06:56  * travis-cipart
15:26:28  * txdvquit (Quit: No Ping reply in 180 seconds.)
15:26:53  * txdvjoined
15:34:50  <AndreasMadsen>tjfontaine: clearly remember bnoordhuis that he didn't take pull requests there where only esthetic
15:35:03  <AndreasMadsen>* bnoordhuis say that
15:36:36  <bnoordhuis>oh hah, 'Your Name' - awesome
15:37:34  <bnoordhuis>AndreasMadsen: that's right, but this was a minor commit that made common.js a little more consistent
15:37:51  <bnoordhuis>also, it's a good gateway drug
15:39:02  <AndreasMadsen>bnoordhuis: I'm sure it was. I clearly lack the experience to tell the difference between "more consistent" and "only esthetic"
15:40:00  <bnoordhuis>oh, snap!
15:41:55  * dshaw_joined
15:42:44  <AndreasMadsen>So what test would be appropriate to this one https://github.com/joyent/node/pull/2925/files
15:45:28  <bnoordhuis>AndreasMadsen: one that triggers the code path?
15:45:40  <indutny>hi
15:45:52  <indutny>bnoordhuis: review please https://github.com/joyent/node/pull/2958/files
15:46:15  <bnoordhuis>indutny: hey, still alive?
15:46:24  <indutny>bnoordhuis: I hope I am
15:46:25  <indutny>:)
15:46:44  * indutnyjust resurrected
15:46:53  <bnoordhuis>indutny: is that going to fix test-debugger-repl and test-debugger-repl-utf8?
15:47:04  <bnoordhuis>oh, apparently not
15:47:10  <indutny>bnoordhuis: nope it doesn't
15:47:11  <AndreasMadsen>bnoordhuis: yeah, but what will. I made the pull request because there has been edge cases where I wondered "was .kill() there failed". But I don't know what will tricker it in a consistent way.
15:47:11  <indutny>:(
15:51:00  <bnoordhuis>indutny: maybe split 63fbf178 into two commits
15:52:03  <indutny>bnoordhuis: I don't mind
16:05:24  * perezdjoined
16:10:26  * dapjoined
16:11:17  <indutny>bnoordhuis: updated
16:12:29  <indutny>bnoordhuis: ooops
16:12:31  <indutny>something get wrong
16:15:30  <indutny>bnoordhuis: ok, updated
16:16:04  * dapquit (Quit: Leaving.)
16:16:20  * dapjoined
16:18:02  * pieternjoined
16:18:20  * TooTallNatejoined
16:18:41  <CIA-99>node: Ben Noordhuis master * rd8c4ece / test/simple/test-cluster-master-error.js :
16:18:41  <CIA-99>node: test: fix race in simple/test-cluster-master-error
16:18:41  <CIA-99>node: Said test checks that the workers shut down when the master errors but it failed
16:18:41  <CIA-99>node: intermittently. Insert a small delay before doing the 'is dead?' check to give
16:18:41  <CIA-99>node: the workers a chance to shut down. - http://git.io/nMcMCw
16:19:43  <bnoordhuis>indutny: lgtm if it works as advertised but shouldn't it have a test?
16:20:25  <indutny>bnoordhuis: hm...
16:20:44  <indutny>bnoordhuis: do you think that's required?
16:20:58  <indutny>bnoordhuis: I don't really like functional testing for such things
16:21:07  <bnoordhuis>indutny: *shrug*
16:21:08  <bnoordhuis>land it
16:21:22  <indutny>bnoordhuis: cool
16:22:51  <CIA-99>node: Fedor Indutny master * rf61d4b7 / lib/_debugger.js :
16:22:51  <CIA-99>node: debugger: exit process on repl exit
16:22:51  <CIA-99>node: * When entering repl - clone 'SIGINT' listeners array (instead of using
16:22:51  <CIA-99>node: existing), as it will be spliced in .removeAllListeners() call later. - http://git.io/RgYFkA
16:22:51  <CIA-99>node: Fedor Indutny master * rb6cb6ce / (lib/_debugger.js lib/repl.js):
16:22:52  <CIA-99>node: debugger: remove 'repl' command from builtinLibs
16:22:52  <CIA-99>node: * It was displaying useless warning - http://git.io/XwDiYA
16:22:52  <CIA-99>node: Fedor Indutny master * r7418905 / lib/_debugger.js :
16:22:53  <CIA-99>node: debugger: breakOnException
16:22:53  <CIA-99>node: Do not break automatically on exception, fixes #2926 - http://git.io/7VtRww
16:22:58  <indutny>bnoordhuis: done :)
16:24:47  * dappart
16:24:57  * dapjoined
16:34:14  * travis-cijoined
16:34:14  <travis-ci>[travis-ci] joyent/node#614 (master - d8c4ece : Ben Noordhuis): The build is still failing.
16:34:14  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/fb47a33...d8c4ece
16:34:14  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/900809
16:34:14  * travis-cipart
16:38:39  * travis-cijoined
16:38:39  <travis-ci>[travis-ci] joyent/node#615 (master - 7418905 : Fedor Indutny): The build is still failing.
16:38:39  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/d8c4ece...7418905
16:38:39  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/900839
16:38:39  * travis-cipart
16:44:08  <TooTallNate>indutny: nice usage of the "exit" event ;)
16:45:31  * orlandovftwjoined
16:49:30  <indutny>TooTallNate: hehe
16:52:55  * dshaw_quit (Quit: Leaving.)
17:01:10  <CIA-99>node: isaacs master * r89653cb / (100 files in 7 dirs): Upgrade npm to 1.1.10 - http://git.io/pU8HqQ
17:01:19  <bnoordhuis>so... to STL or not to STL?
17:01:30  <bnoordhuis>https://github.com/joyent/node/pull/2962
17:02:00  <indutny>bnoordhuis: -1 for STL
17:02:28  <creationix>don't do it
17:02:53  <indutny>bnoordhuis: that's almost equal to rewriting node in coffeescript to me
17:04:11  <bnoordhuis>okay, give me reasons
17:04:53  <creationix>bloat
17:05:02  <mmalecki>-1 as well. don't let too much C++ism into core
17:05:03  <creationix>scares away *real* C developers
17:05:21  <indutny>yeah, node was always in the middle between C++ and C
17:05:27  <indutny>autoptr, smartptr
17:05:30  <bnoordhuis>i'm a real c developer and c++ templates don't scare me
17:05:31  <indutny>is not what I want to see in code
17:05:43  <indutny>bnoordhuis: I think I'm quite real too :)
17:05:43  <bnoordhuis>oh, i'm not really talking about autoptr
17:05:55  <bnoordhuis>i'm thinking of containers: vector, set, map
17:06:01  <creationix>ok, so it scares away Linus
17:06:08  <bnoordhuis>and banishing manual memory management
17:06:10  <indutny>bnoordhuis: why should we rely on someone's code?
17:06:23  <creationix>just write the parts you want from STL as a library
17:06:26  <indutny>bnoordhuis: lets use boehm GC
17:06:28  <indutny>:)
17:06:44  <bnoordhuis>indutny: yeah, that's the major drawback
17:06:57  <bnoordhuis>you can't rely on the STL being available or decent on all platforms
17:07:01  <indutny>yes
17:07:06  <bnoordhuis>then again, we only target gcc and msvc
17:07:15  <bnoordhuis>both have high-quality STL implementations
17:07:19  <creationix>keep in mind that luvit is pure C
17:07:22  <creationix>no C++ anywhere
17:07:38  <indutny>creationix: how is that related to luvit?
17:07:39  <bnoordhuis>creationix: i'm talking about node, not libuv
17:07:43  <creationix>so you don't need g++ to build it (not sure how much that matters)
17:07:53  <creationix>ahh, for node
17:08:05  <bnoordhuis>libuv will always be pure c
17:08:20  <creationix>ok, now that I understand the question
17:08:29  <creationix>I still say -1, but ignore my comments about C++ in general
17:08:34  <bnoordhuis>you should've read the PR, tim :)
17:08:56  <creationix>I did, but I didn't notice it was node
17:09:35  <creationix>anyway, I'm a C/C++ newb, ignore my ranting
17:10:08  <bnoordhuis>mmalecki: not too much, just enough
17:10:17  <bnoordhuis>it's mostly the manual memory management that annoys me
17:10:33  <bnoordhuis>node_crypto.cc for example is full of memory leaks, i'm sure
17:10:33  <indutny>bnoordhuis: tradeoff is too high
17:10:44  <mmalecki>bnoordhuis: I like manual memory management in C and C++
17:10:45  <bnoordhuis>indutny: how so?
17:10:54  <mmalecki>I'd go crazy if I had to do it in JS tho.
17:11:12  <indutny>bnoordhuis: well, not always. but you may expect 10% slowdown if we'll move to GC
17:11:19  <mmalecki>but really, most C++ programmers don't even use these containers. you're making the entry barrier higher
17:11:26  <bnoordhuis>indutny: wut? who's talking about garbage collection?
17:11:39  <indutny>bnoordhuis: what's automatic memory management to you then?
17:11:44  <bnoordhuis>indutny: RAII
17:12:17  <bnoordhuis>mmalecki: if a higher barrier to entry means higher quality code from the few people that make it through, well, i'm all for it
17:12:49  <mmalecki>bnoordhuis: I don't see how this code is better
17:13:12  <bnoordhuis>mmalecki: there's no chance of memory leaks, for one
17:13:24  <mmalecki>`std::vector<uint16_t> bs(len, 0); ` is more cryptic than `uint16_t * twobytebuf = new uint16_t[len];`
17:13:34  <indutny>bnoordhuis: I think you can replace https://github.com/joyent/node/pull/2962/files#L0R956 and others with stack allocated variables
17:13:38  <indutny>bnoordhuis: and classes
17:13:44  <mmalecki>also, why does he comment out code?
17:14:06  <bnoordhuis>mmalecki: don't get too hung up on the PR itself
17:14:56  <bnoordhuis>indutny: stack-allocated VLAs are blasphemous, we will speak no more of them
17:15:02  <mmalecki>bnoordhuis: ok. so in general, I'm -1 on using STL bullshit in node C++ code.
17:15:11  <TooTallNate>what does it do to performance?
17:15:16  <bnoordhuis>mmalecki: so i've noticed but i want reasons :)
17:15:38  <indutny>bnoordhuis: what about classes
17:15:41  <bnoordhuis>indutny: i half-kid but on-stack VLAs are dangerous for the same reason that alloca() is dangerous
17:15:58  <indutny>bnoordhuis: I didn't meant 15mb on-stack variables
17:16:11  <indutny>bnoordhuis: you should use class wrappers for that kind of staff
17:16:15  <indutny>s/staff/stuff
17:16:16  <bnoordhuis>TooTallNate: there's no clear cut answer - might be faster, might be slower, might be exactly the same
17:16:30  <mmalecki>bnoordhuis: they introduce bloat, they are slower than arrays and they make code more complicated
17:16:31  <indutny>bnoordhuis: RAII is about design, not STL
17:16:36  <TooTallNate>depends on the STL impl
17:16:43  <indutny>TooTallNate: exactly
17:17:03  * travis-cijoined
17:17:03  <travis-ci>[travis-ci] joyent/node#616 (master - 89653cb : isaacs): The build is still failing.
17:17:03  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/7418905...89653cb
17:17:03  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/901067
17:17:03  * travis-cipart
17:17:20  * orlandovftwquit (Ping timeout: 246 seconds)
17:17:53  <bnoordhuis>mmalecki: std::vector<> at -O2 comes down to direct array indexing
17:18:13  <bnoordhuis>mmalecki: iow, "slower than arrays" is false
17:18:20  <bnoordhuis>more complicated... yes, maybe
17:18:50  * paddybyers_joined
17:19:03  <bnoordhuis>iterating over containers has always been the bane of my existence
17:19:28  <TooTallNate>lol
17:19:31  <bnoordhuis>for (std::map<foo, bar>::const_iterator i = m.begin(), end = m.end(); i != end; ++i) ...
17:19:58  <DrPizza>for(auto it(m.begin()), end(m.end()); it != end; ++it) {
17:19:58  <DrPizza>}
17:20:06  <indutny>bnoordhuis: hahaha
17:20:17  <bnoordhuis>DrPizza: yes... but we target c++98 compilers
17:20:17  <indutny>bnoordhuis: yees, iterating is a 5% of candor's code
17:20:36  <indutny>well, not literally
17:20:40  <DrPizza>bnoordhuis: balls to C++98!
17:20:55  <DrPizza>I think c++11 is pretty transformational
17:21:15  <indutny>I think c++11 is pretty another language
17:21:19  <DrPizza>heh
17:21:24  <DrPizza>though until Apple ships a clang that does lambdas
17:21:26  * paddybyersquit (Ping timeout: 276 seconds)
17:21:26  * paddybyers_changed nick to paddybyers
17:21:27  <bnoordhuis>it is - c++11 is great to work with again
17:21:29  <DrPizza>it's annoying
17:21:47  <bnoordhuis>but until all the world is c++11, we target c++98
17:22:36  <bnoordhuis>okay, i understand everyone is in favor of switching to the STL?
17:22:44  * bnoordhuisducks
17:22:47  <tjfontaine>I am :)
17:23:13  <tjfontaine>also stack allocated, the threadpool stack is tiny so you can't do it there
17:24:40  <indutny>bnoordhuis: who are everyone to you?
17:24:51  <indutny>:)
17:25:00  <tjfontaine>indutny: "me myself and I"
17:25:01  <indutny>I've seen at least 3 people who was against that
17:25:02  <bnoordhuis>indutny: me and my 3 split personalities
17:25:08  <tjfontaine>the "royal we"
17:25:12  <indutny>bnoordhuis: well, ok then :)
17:25:38  <bnoordhuis>indutny: i don't intend to switch right away
17:25:44  <bnoordhuis>but i want people to think it over
17:28:51  <TooTallNate>does v8 use stl?
17:29:51  <indutny>TooTallNate: I think no
17:32:55  * einaros-changed nick to einaros
17:40:25  * sh1mmerjoined
17:42:02  * dshaw_joined
17:42:19  * pfox___joined
17:43:54  * mikealjoined
17:46:37  <bnoordhuis>TooTallNate: google is in general quite stl adverse
17:46:50  <TooTallNate>interesting
17:47:15  <bnoordhuis>TooTallNate: their main motivation is that stl classes can throw exceptions
17:48:04  <TooTallNate>and we compile with -fno-exceptions :p
17:48:14  <bnoordhuis>well, yes - but that's not it
17:48:37  <bnoordhuis>i assume that most of google's code is long-lived and needs to handle OOM conditions (that's when the STL throws)
17:49:08  <bnoordhuis>but for us, OOM is the end of the road anyway
17:56:32  <bnoordhuis>mmalecki: why didn't you tell you were in nodeup #13?
17:57:01  * igorzijoined
17:57:12  * dshaw_quit (Quit: Leaving.)
17:57:13  * AndreasMadsenquit (Remote host closed the connection)
17:57:33  <TooTallNate>bnoordhuis: hey i redid this pull request https://github.com/bnoordhuis/node-buffertools/pull/23
17:57:55  <TooTallNate>npm + node-gyp makes things a lot simpler to migrate now
17:58:15  * AndreasMadsenjoined
17:58:56  <mmalecki>bnoordhuis: oh, hey, I was in nodeup #13, btw
17:59:16  <mmalecki>bnoordhuis: I don't know, actually :)
18:02:42  <bnoordhuis>TooTallNate: i'll try to check it sometime this week (honestly)
18:03:19  <bnoordhuis>i suppose i could just merge it and blame you if people report problems
18:03:31  <TooTallNate>:p
18:06:52  <mmalecki>that's what responsible people do!
18:07:04  <mmalecki>bnoordhuis: also, I'm pretty sure it sounds terrible
18:07:09  <mmalecki>I didn't even listen to it
18:07:33  <bnoordhuis>mmalecki: i'll let you know how embarrassing it is :)
18:07:55  <mmalecki>bnoordhuis: it'd be appreciated, thanks! :)
18:08:23  * isaacsjoined
18:11:01  * paddybyersquit (Quit: paddybyers)
18:11:44  <AndreasMadsen>isaacs: yo
18:12:10  <AndreasMadsen>isaacs: Just trying to be a pain again - but
18:13:30  * mikealquit (Quit: Leaving.)
18:13:30  <AndreasMadsen>when are these cluster patches landing (or rejected)
18:14:49  <AndreasMadsen>and this guy -> https://github.com/joyent/node/pull/2650
18:14:52  <isaacs>AndreasMadsen: i'm reviewing them now.
18:15:34  <AndreasMadsen>isaacs: oh, cool - didn't know
18:24:43  * orlandovftwjoined
18:24:43  * orlandovftwquit (Client Quit)
18:25:07  * orlandovftwjoined
18:25:54  * AndreasMadsenchanged nick to Andreas[away]
18:32:31  * brsonjoined
18:33:01  * mikealjoined
18:39:53  * `3rdEdenjoined
18:49:35  <igorzi>AndreasMadsen: tcp connections patch for windows is coming along.. we discovered some issues with it last week, so hopefully it should be done in the next few days
18:52:10  <Andreas[away]>igorzi: that sounds lovely - you are a robot
18:52:22  * Andreas[away]changed nick to AndreasMadsen
18:52:54  * indutnychanged nick to indutny_sleeping
18:53:29  <bnoordhuis>AndreasMadsen: that's not comforting, i just watched terminator 1...
18:54:33  <AndreasMadsen>bnoordhuis: never seen it
18:54:37  <benvie>hmm buffertools
18:57:11  <AndreasMadsen>http://upload.wikimedia.org/wikipedia/commons/5/53/Reagan%2BSchwarzenegger1984.jpg
18:57:39  <TooTallNate>the governator
18:58:38  <AndreasMadsen>That must be the next best picture I ever seen, right after http://upload.wikimedia.org/wikipedia/commons/a/a2/Steve_Jobs_and_Bill_Gates_%28522695099%29.jpg
19:22:35  * dshaw_1joined
19:22:47  * k-s[AWAY]changed nick to k-s
19:22:49  * dshaw_1changed nick to dshaw_
19:45:28  * mikealquit (Quit: Leaving.)
20:08:31  * mikealjoined
20:12:22  <TooTallNate>readline woes :\ https://github.com/joyent/node/issues/2959
20:18:47  <benvie>you should take a look at the current reified api when you get a chance ;)
20:22:03  <TooTallNate>benvie: ya i was checking it out the other day. looks really nice!
20:22:17  <TooTallNate>i wanna start seeing what it would take to refactor node-ffi using that soon
20:23:13  <benvie>I improved the api a lot and exposed it mostly as a single function
20:24:27  <benvie>so stuff like reified('Int32[4][4][2]'), or Int32[4][4][2] or new reified('Int32[4][4][2]') work
20:24:36  <TooTallNate>benvie: maybe you have a separate addon part that adds native indirection
20:24:42  <TooTallNate>that's optional
20:24:48  <TooTallNate>or i just add it to node-ffi
20:24:55  <benvie>yeah I've been thinking on it
20:25:12  <benvie>wondering if there's really two concerns
20:25:22  <benvie>one is the file format type with dynamic structures/offsets
20:25:23  <TooTallNate>benvie: so you're doing like string parsing now?
20:25:30  <AndreasMadsen>TooTallNate: "perhaps just monkey-patch child._internal.kill to always return -1 for that test." is that really allowed. That will be the worst testcase I ever wrote.
20:25:33  <benvie>which is similar to pointers but not the samw
20:25:42  <benvie>no just two things
20:25:54  <benvie>one is matching the array syntax, and also any type can be referenced by it's name
20:26:08  <TooTallNate>AndreasMadsen: well at least it tests the code path. if there's no other way to force kill() to fail then idk
20:26:10  <benvie>and specifying a type's name is now while not required very much recommended
20:26:13  <benvie>you could do like
20:26:54  <benvie>reified('RGB', { r: 'UInt8', g: 'UInt8', b: 'UInt8' }); new reified('RGB')
20:27:14  <benvie>basically there's so many types it would be annoying to required direct references to all of the constructors
20:27:16  <AndreasMadsen>TooTallNate: yeah, okay I will do that :p
20:28:04  <TooTallNate>AndreasMadsen: i agree it's a little clumsy :p
20:28:04  <benvie>var RGB = reified('RGB', { r: 'UInt8', g: 'UInt8', b: 'UInt8' }); new RGB
20:28:07  <benvie>would also work
20:28:34  <TooTallNate>benvie: is the name part optional?
20:28:51  <benvie>yeah names and the actual structs themselves are itnerchangingable when defining stuff
20:28:54  <benvie>you don't need to name anything
20:29:28  <TooTallNate>sounds rather flexible :)
20:29:29  <benvie>all the apis are exposed as they were also
20:29:55  <AndreasMadsen>TooTallNate: I just have this princip about always testing code from a user perspective. And if I see one rewrite ._internal.kill I will snap them.
20:29:55  <TooTallNate>benvie: try to get some real examples in an examples/ dir maybe
20:29:55  <benvie>just now that base require('reified') is the catch-all function that can do basically anything
20:30:10  <benvie>yeah I'm going to redo my ttf font lib
20:30:33  <benvie>I'll explore the file offset/pointer stuff while doing that too, and I'll just include the whole thing as one big ass example
20:30:41  <benvie>ttf is kind of complex
20:31:48  <benvie>I think it can work as it is in browsers too, assuming it was packaged in one file. I made a buffer abstraction that works across
20:40:19  <isaacs>AndreasMadsen: you around?
20:40:32  <AndreasMadsen>isaacs: I think so
20:40:50  <isaacs>AndreasMadsen: can you rebase 840369058fac1597811f8b521da9d5a57633defe (step 7) on top of 6c383c61611fb78d0961a332c1ad55d1d5514ba6 (step 6)
20:40:52  * k-schanged nick to k-s[AWAY]
20:40:55  <isaacs>i'm getting a merge conflict
20:41:08  <isaacs>i can land step 6 into master first if that's easier
20:41:23  <AndreasMadsen>isaacs: yes please do that
20:41:40  <isaacs>step 6 = disconnect stuff. step 7 = kill workers when master dies
20:41:41  <isaacs>ok
20:42:03  <AndreasMadsen>isaacs: yes that's correct
20:42:07  <CIA-99>node: Andreas Madsen master * rab32e9e / lib/child_process.js : child_process: emit 'channel closed' error instead of throwing - http://git.io/JECcsQ
20:42:08  <CIA-99>node: Andreas Madsen master * rd927fbc / (4 files in 3 dirs): (log message trimmed)
20:42:08  <CIA-99>node: cluster: add graceful disconnect support
20:42:08  <CIA-99>node: This patch add a worker.disconnect() method there will stop the worker from accepting
20:42:08  <CIA-99>node: new connections and then stop the IPC. This allow the worker to die graceful.
20:42:08  <CIA-99>node: When the IPC has been disconnected a 'disconnect' event will emit.
20:42:08  <CIA-99>node: The patch also add a cluster.disconnect() method, this will call worker.disconnect() on
20:42:09  <CIA-99>node: all connected workers. When the workers are disconneted it will then close all server
20:42:11  <isaacs>AndreasMadsen: ^
20:42:28  <isaacs>AndreasMadsen: i've got a few little english-language spelling and doc stuff, but i'll land that as a separate patch.
20:42:41  <isaacs>AndreasMadsen: it's nothing very major, but it'll cause unnecessary conflicts to do it before landing your stuff.
20:43:47  <AndreasMadsen>isaacs: I will rebase now then :D
20:44:19  <isaacs>kewl
20:44:58  * paddybyersjoined
20:58:01  * pieternquit (Quit: pietern)
20:58:25  * travis-cijoined
20:58:25  <travis-ci>[travis-ci] joyent/node#617 (master - d927fbc : Andreas Madsen): The build is still failing.
20:58:25  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/89653cb...d927fbc
20:58:25  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/902440
20:58:25  * travis-cipart
20:58:57  * pieternjoined
20:59:49  <AndreasMadsen>is it just my internet connection or is github static file server down?
21:00:16  <tjfontaine>AndreasMadsen: thigns seem kosher here
21:02:02  <AndreasMadsen>tjfontaine: I was afraid of that
21:02:33  <AndreasMadsen>isaacs: I have pull the merge fix, I hope you can see them
21:02:43  <isaacs>ok
21:07:04  <AndreasMadsen>isaacs: got connection, it looks fine to me
21:11:17  <isaacs>AndreasMadsen: i'm still getting merge conflict. can you do this? git checkout cluster2-fixes; git fetch -a upstream; git rebase upstream/master; git push origin +cluster2-fixes
21:13:02  <AndreasMadsen>isaacs: no will try fixing it
21:13:27  * k-s[AWAY]changed nick to k-s
21:15:02  <AndreasMadsen>rebuliding yet another time
21:15:19  <isaacs>k
21:15:33  <TooTallNate>i've started doing --shared-zlib --openssl-use-sys when building
21:15:36  <TooTallNate>saves a lot of time
21:16:06  <AndreasMadsen>TooTallNate: it is v8 there kills me
21:17:24  <TooTallNate>isaacs: how do you feel about adding `process.config.argv`?
21:17:26  <TooTallNate>> process.config.argv
21:17:31  <TooTallNate>[ '--debug',
21:17:31  <TooTallNate> '--shared-zlib',
21:17:31  <TooTallNate> '--shared-v8' ]
21:18:00  <TooTallNate>not strictly useful, but i always kinda liked how gcc shows you the configure input :p
21:18:01  <tjfontaine>hmm, how does that differ from the defines?
21:18:28  <isaacs>TooTallNate: meh. it's nice, but what does it get you that isn't in process.config already?
21:18:31  <TooTallNate>tjfontaine: the rest of process.config would be more useful, so i'm leaning towards no actually :p
21:18:39  <TooTallNate>isaacs: nothing :D, so nvm
21:18:42  <isaacs>hahah
21:18:49  <isaacs>fair enough, i'm convinced :)
21:18:52  <TooTallNate>hahaha
21:19:43  * dshaw_quit (Ping timeout: 246 seconds)
21:20:59  * dshaw_joined
21:21:38  * pfox___quit (Ping timeout: 246 seconds)
21:21:40  <AndreasMadsen>isaacs: try again
21:23:11  <isaacs>AndreasMadsen: bingo, thanks :0
21:23:13  <isaacs>:)
21:24:50  * mikealquit (Quit: Leaving.)
21:29:53  * dshaw_1joined
21:31:25  <AndreasMadsen>TooTallNate: https://github.com/joyent/node/pull/2925 shoud be fixed now too
21:31:47  <AndreasMadsen>isaacs: can I go to bed now? - :)
21:31:57  <isaacs>AndreasMadsen: running tests now
21:32:02  <isaacs>i'll land in a minute if it's working
21:32:08  <isaacs>or comment. but you can go to bed, yes :)
21:32:10  <isaacs>thanks
21:32:28  <isaacs>yep, only simple/test-eio-limit is failing now
21:32:31  <isaacs>which is... weird?
21:32:40  <tjfontaine>I've been getting that one a lot
21:32:46  * dshaw_quit (Ping timeout: 260 seconds)
21:33:02  <isaacs>simple/test-child-process-fork2 fails for me sometimes, too
21:33:14  <tjfontaine>I haven't seen that one
21:34:58  <AndreasMadsen>isaacs: I can't see that a cluster patch can have anything to do with zlib, so its properly just a glitch.
21:35:18  <isaacs>yep
21:35:43  <isaacs>also, some other random failures still.
21:35:50  <isaacs>fork2 and debugger-repl
21:36:08  <AndreasMadsen>isaacs: crypto and jslint fails for me
21:36:12  <AvianFlu>isaacs, I see test-eio-limit fail on linux sometimes
21:36:16  <AvianFlu>maybe 1/6 of the time
21:36:46  <AvianFlu>I also occasionally see test-dgram-pingpong crash the test runner
21:36:47  <tjfontaine>it's been about 1/3 on my lion install
21:39:08  <AndreasMadsen>night, night
21:39:52  * AndreasMadsenquit (Remote host closed the connection)
21:41:44  <CIA-99>node: Andreas Madsen master * r94d337e / (lib/cluster.js test/simple/test-cluster-master-kill.js):
21:41:44  <CIA-99>node: cluster: kill workers when master dies
21:41:44  <CIA-99>node: This patch will kill the worker once it has lost its connection with the parent.
21:41:44  <CIA-99>node: However if the worker are doing a suicide, other measures will be used. - http://git.io/3gn6Aw
21:41:45  <CIA-99>node: isaacs master * rdce8682 / (doc/api/cluster.markdown lib/cluster.js): cluster: English language fixing - http://git.io/Jt4-VQ
21:55:06  * skomskijoined
21:57:43  * travis-cijoined
21:57:43  <travis-ci>[travis-ci] joyent/node#618 (master - dce8682 : isaacs): The build is still failing.
21:57:43  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/d927fbc...dce8682
21:57:43  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/902858
21:57:43  * travis-cipart
22:02:25  * skomskiquit (Quit: node-irc says goodbye)
22:02:33  * pieternquit (Quit: pietern)
22:02:35  * dshaw_joined
22:03:42  * pieternjoined
22:04:10  * dshaw_1quit (Ping timeout: 244 seconds)
22:04:41  * skomskijoined
22:13:31  * skomskiquit (Quit: Lost terminal)
22:16:15  * skomskijoined
22:16:24  * skomskiquit (Remote host closed the connection)
22:17:34  * skomskijoined
22:32:44  * rendarquit
22:34:39  * paddybyersquit (Quit: paddybyers)
22:48:47  * paddybyersjoined
22:58:42  * mikealjoined
23:00:10  * paddybyersquit (Quit: paddybyers)
23:01:47  * skomskiquit (Read error: Connection reset by peer)
23:04:53  * bnoordhuisquit (Ping timeout: 246 seconds)
23:15:12  * elijah-mbpquit (Remote host closed the connection)
23:16:21  * elijah-mbpjoined
23:19:33  * k-schanged nick to k-s[AWAY]
23:27:23  * mjr_quit (Quit: mjr_)
23:27:38  * mjr_joined
23:33:36  * brsonquit (Ping timeout: 240 seconds)
23:33:50  * brsonjoined
23:50:14  * mikealquit (Quit: Leaving.)
23:52:30  * sh1mmerquit (Quit: sh1mmer)
23:54:54  <igorzi>isaacs: hey yt?
23:55:16  * `3rdEdenquit (Quit: Zzz)