00:11:15  * mikealjoined
00:18:13  * `3rdEdenquit (Quit: ZZZZzzzz)
00:22:50  * mikealquit (Quit: Leaving.)
00:49:33  <bnoordhuis>ryah: https://github.com/bnoordhuis/node/compare/wip%5E%5E...wip
00:49:47  <bnoordhuis>don't review it just yet, it's still WIP :)
00:50:02  <bnoordhuis>but test-isolates, test-isolates2 and test-isolates-ping-pong all work
00:56:53  * mikealjoined
00:57:34  <bnoordhuis>signing off, sleep tight all
00:57:52  <mmalecki>night bnoordhuis
00:58:52  * mikealquit (Client Quit)
01:02:23  * bnoordhuisquit (Ping timeout: 252 seconds)
01:04:33  * mikealjoined
01:04:42  * mikealquit (Client Quit)
01:24:20  * brsonquit (Ping timeout: 252 seconds)
01:30:59  * brsonjoined
01:54:13  * dapquit (Quit: Leaving.)
01:54:23  * perezdjoined
01:54:31  <perezd>hey paddybyers !
02:08:53  * CoverSlidechanged nick to Cover|Away
02:10:30  <ryah>bnoordhuis++
02:10:30  <kohai>bnoordhuis has 1 beer
02:10:43  <ryah>mjr_, dannycoates: any crashes?
02:11:27  <mjr_>oh shit, we got a couple
02:12:08  <mjr_>looking
02:12:16  <ryah>:~
02:12:25  <dannycoates>no crashes
02:12:33  <dannycoates>er
02:12:45  <dannycoates>on 1009
02:13:07  <mjr_>https://gist.github.com/8fb5158f6b4cede3fb65
02:13:16  <mjr_>not in http parser!
02:14:13  <mjr_>updated gist with the second one
02:14:27  <mjr_>identical perhaps
02:18:23  <ryah>hm
02:18:43  <mjr_>seems way less serious though, because the old crash rate was way higher.
02:21:34  <ryah>strange crash
02:23:09  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
02:23:14  <ryah>well - glad to hear we seem to have squashed the http-parser problem
02:23:31  <ryah>that was a bitch
02:28:12  <dannycoates>mjr_: those cores aren't nrs
02:28:18  <mjr_>oh, ha
02:28:28  <mjr_>So totally new problem then.
02:28:56  <dannycoates>i've seen this one before
02:30:12  <mjr_>Interesting that it's crashing in buffer stuff. That process doesn't ever use Buffer explicitly.
02:30:38  <mjr_>It is a very busy process though.
02:33:39  <ryah>it's during the slab alloc for libuv
02:34:50  * pieternquit (Quit: pietern)
02:36:51  <CIA-111>node: Andreas Madsen master * r5f08c3c / (11 files in 4 dirs):
02:36:51  <CIA-111>node: cluster improvements: Worker class and isolate internal messages
02:36:51  <CIA-111>node: Fixes #2388 - http://git.io/p_2KLA
02:48:29  * travis-cijoined
02:48:29  <travis-ci>[travis-ci] joyent/node#178 (master - 5f08c3c : Andreas Madsen): The build is still failing.
02:48:29  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/e21643d...5f08c3c
02:48:29  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/479793
02:48:29  * travis-cipart
02:49:27  * brsonquit (Quit: leaving)
02:59:22  * dshaw_1quit (Quit: Leaving.)
05:38:12  <indutny>bnoordhuis: w00t! your WIP branch looks really neat
05:54:01  * igorziquit (Ping timeout: 258 seconds)
06:04:45  * mralephjoined
06:42:05  * mralephquit (Quit: Leaving.)
06:43:01  * TkTechjoined
06:44:14  <TkTech>Hello Gents. Are there any minimal examples of using libuv?
06:44:25  <TkTech>For example, the classic echo server.
06:46:31  <ryah>TkTech: check out the test folder
06:46:35  <ryah>TkTech: there are a few examples
06:46:58  <ryah>TkTech: EG https://github.com/joyent/libuv/blob/master/test/echo-server.c
06:47:27  <TkTech>Bawh, and I was looking right at it. Many thanks.
06:47:59  <ryah>mjr_, dannycoates: one last check - any crashes?
07:01:39  * isaacsquit (Quit: isaacs)
07:37:34  * perezdquit (Quit: perezd)
07:38:25  * paddybyersquit (Quit: paddybyers)
07:39:50  * paddybyersjoined
07:45:39  * mikealjoined
08:23:02  * mikealquit (Quit: Leaving.)
08:33:08  <CIA-111>node: Andreas Madsen master * re2f1e50 / (3 files in 2 dirs):
08:33:09  <CIA-111>node: typos
08:33:09  <CIA-111>node: fixes #2465 - http://git.io/PyfyPg
08:42:31  * mikealjoined
08:44:21  * mikealquit (Client Quit)
08:44:30  * travis-cijoined
08:44:31  <travis-ci>[travis-ci] joyent/node#179 (master - e2f1e50 : Andreas Madsen): The build is still failing.
08:44:31  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/5f08c3c...e2f1e50
08:44:31  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/480640
08:44:31  * travis-cipart
08:44:55  <CIA-111>node: koichik v0.6 * rbaebd30 / lib/http.js : http: use `self` insted of `this` - http://git.io/SkeMQw
08:52:38  * travis-cijoined
08:52:39  <travis-ci>[travis-ci] joyent/node#180 (v0.6 - baebd30 : koichik): The build is still failing.
08:52:39  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/3452477...baebd30
08:52:39  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/480650
08:52:39  * travis-cipart
09:00:10  * raggi_quit (Quit: ZNC - http://znc.sourceforge.net)
09:03:13  * raggijoined
09:27:34  * benviequit
09:29:45  * benviejoined
09:49:16  <indutny>bnoordhuis: yt?
11:41:16  * indexzeroquit (Quit: indexzero)
13:48:26  * bnoordhuisjoined
14:01:57  <indutny>bnoordhuis: ping ping!
14:02:27  <bnoordhuis>indutny: pong pong
14:03:12  <indutny>bnoordhuis: why libuv isn't fork-safe?
14:03:40  <bnoordhuis>indutny: mutexes, one-time init guards
14:04:42  <indutny>bnoordhuis: but child process will have same memory and etc
14:04:48  <indutny>bnoordhuis: ah, multi-threaded env
14:04:55  <indutny>bnoordhuis: ok
14:16:28  <mmalecki>I wonder how did Charlie manage to make daemon.node work
14:21:32  <indutny>mmalecki: haha
14:21:35  <indutny>we managed
14:21:40  <indutny>it just worked fine
14:21:41  <indutny>:D
14:21:47  <mmalecki>indutny: oh, awesome. yeah, next version
14:21:49  <indutny>dunno why you labeled it as broken
14:21:52  <indutny>yep
14:21:57  <mmalecki>indutny: many people labeled it as broken
14:22:20  <mmalecki>indutny: it was an issue with closing a unix socket after forking. please try it with previous version
14:22:45  <mmalecki>indutny: nc -U .forever/sock/forever.sock will be empty after `forever start`
14:22:56  <mmalecki>at least on 0.6
14:23:01  <mmalecki>gotta run, ttyl
14:24:42  <indutny>mmalecki: ok!
14:24:44  <indutny>thanks
14:36:15  <indutny>bnoordhuis: how is it going with child_process.fork?
14:36:56  <bnoordhuis>indutny: check out my wip branch, it mostly works
14:37:32  <indutny>bnoordhuis: so, can I rebase my debugger stuff over it?
14:37:46  <indutny>bnoordhuis: ( I seen branch, it's really neat)
14:38:08  <bnoordhuis>indutny: hold off a little longer, i'm going to clean it up a little
14:53:48  <indutny>bnoordhuis: ok
15:05:21  * dshaw_joined
15:13:59  * voodootikigod__quit (Remote host closed the connection)
15:14:14  * voodootikigod_joined
15:45:17  <CIA-111>node: Ben Noordhuis v0.6 * r760928b / README.md : docs: mention that python 2.6 or 2.7 is required - http://git.io/50NGVg
15:48:29  * isaacsjoined
15:54:01  * travis-cijoined
15:54:01  <travis-ci>[travis-ci] joyent/node#181 (v0.6 - 760928b : Ben Noordhuis): The build was fixed.
15:54:01  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/baebd30...760928b
15:54:01  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/481555
15:54:01  * travis-cipart
15:57:44  <CIA-111>node: Shigeki Ohtsu master * ra993d22 / tools/addon.gypi : build: set product_prefix to empty for Linux - http://git.io/Y9bKWw
15:57:45  <CIA-111>node: Shigeki Ohtsu master * r55c2197 / common.gypi :
15:57:45  <CIA-111>node: build: compile without -fvisibility=hidden
15:57:45  <CIA-111>node: Fixes symbol lookup errors when loading an addon module on Linux. - http://git.io/j4VT6g
16:04:52  * voodootikigod_quit (Read error: Connection reset by peer)
16:04:59  * voodootikigod_joined
16:09:22  * travis-cijoined
16:09:23  <travis-ci>[travis-ci] joyent/node#182 (master - 55c2197 : Shigeki Ohtsu): The build is still failing.
16:09:23  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/e2f1e50...55c2197
16:09:23  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/481623
16:09:23  * travis-cipart
16:11:43  <indutny>more I work on node-spdy - more I think that I should write article on implementing servers with node's http module compatible API
16:11:57  * voodootikigod__joined
16:12:25  * voodootikigod_quit (Read error: Connection reset by peer)
16:12:25  * voodootikigod__quit (Read error: Connection reset by peer)
16:12:39  * voodootikigod_joined
16:26:54  * voodootikigod_quit (Read error: Connection reset by peer)
16:27:27  * voodootikigod_joined
16:32:39  * rmustaccjoined
16:38:52  * dshaw_quit (Quit: Leaving.)
17:12:22  * perezdjoined
17:19:39  * dapjoined
17:30:59  * AndreasMadsenjoined
17:49:30  * mikealjoined
17:51:11  * mikealquit (Client Quit)
17:56:49  * dshaw_joined
18:12:58  * voodootikigod__joined
18:16:37  * voodootikigod_quit (Ping timeout: 240 seconds)
18:18:47  * voodootikigod_joined
18:20:04  * voodootikigod__quit (Ping timeout: 240 seconds)
18:22:44  * `3rdEdenjoined
18:23:46  * brsonjoined
18:29:09  * TooTallNatejoined
18:38:51  * Cover|Awaychanged nick to CoverSlide
18:44:57  * isaacsquit (Quit: isaacs)
19:00:26  <bnoordhuis>call?
19:02:51  <ryah>let's skip im not really here
19:03:29  <bnoordhuis>cool
19:05:25  <TooTallNate>ryah: sick again/still?
19:05:49  <ryah>yes
19:09:05  <TooTallNate>sorry to hear that
19:09:08  * AndreasMadsenquit (Ping timeout: 252 seconds)
19:09:29  <ryah>bnoordhuis: your patch is looking good - going to land that today?
19:09:39  <indutny>bnoordhuis: ryah +1!!!
19:09:41  <indutny>:)
19:09:45  <ryah>indutny wants to do his debugger isolate stuff on top of it
19:10:03  <ryah>mjr_, dannycoates: one last check - any crashes?
19:10:09  <mjr_>checking
19:10:19  <dannycoates>ryah: crash free!
19:10:33  <dannycoates>we have libuv crashes
19:10:40  <ryah>awesome
19:10:45  <ryah>can we start a new issue for those?
19:10:48  <mjr_>Yeah, that libuv one keeps happening.
19:11:03  <ryah>the libuv one = the buffer one?
19:11:03  <mjr_>But thankfully that http parser ones seems gone.
19:11:12  <ryah>how often is the libuv one?
19:11:14  <mjr_>#0 0x0000000000596b14 in v8::Object::GetIndexedPropertiesExternalArrayDataLength() ()
19:11:18  <mjr_>that one
19:11:27  <ryah>ok
19:11:59  <dannycoates>it might be partially self inflicted. we found another crasher in that code this morning
19:12:36  <mjr_>Node shouldn't ever segfault though.
19:14:21  <mjr_>We crashed 31 times in the last 12 hours.
19:14:51  <mjr_>With this latest issue. So it's still a serious thing.
19:15:28  <mjr_>But perhaps we can work around it somewhat, as our code in that process was doing something dumb.
19:15:34  <bnoordhuis>ryah: sure
19:16:20  <ryah>mjr_: okay - we'll definitely want to have a look at it
19:16:36  <ryah>i can't imagine how that is actually failing though
19:16:58  <ryah>at the moment though i need the latest npm and v8 patch so we can push out v0.6.7
19:17:03  <ryah>ErikCorryV8: ping
19:17:20  <mjr_>We might be able to come up with a better repro for it. That crash happens in a different process, not router.
19:17:37  <mjr_>Great, I say ship a new release.
19:18:20  <bnoordhuis>ryah: i can do a release tonight or tomorrow night if you want
19:18:28  <bnoordhuis>so you can go back to your chicken broth
19:18:44  <ryah>bnoordhuis: i want to get isaacs to do it again so he's up to speed
19:18:51  <bnoordhuis>even better
19:18:55  <ryah>unfortunately he's MIA
19:19:05  <rmustacc>ryah: He just showed up to SFO.
19:19:17  <rmustacc>He'll be online in a min or so.
19:19:48  <ryah>SFO?
19:20:06  <ryah>SF office?
19:20:08  <rmustacc>The sf office.
19:20:10  * ErikCorryV8quit (Ping timeout: 258 seconds)
19:20:15  <ryah>bad acronym :)
19:20:15  <rmustacc>Though I used the airport code accidentally at first.
19:21:04  <ryah>rmustacc: can you tell him to push the latest npm into v0.6?
19:21:13  <ryah>im AFK for a bit
19:21:31  * AndreasMadsenjoined
19:21:50  <rmustacc>ryah: Sure
19:23:53  <CIA-111>node: isaacs v0.6 * r78dbb4b / (204 files in 15 dirs): [email protected] - http://git.io/qoclzg
19:24:02  * isaacsjoined
19:31:43  * travis-cijoined
19:31:44  <travis-ci>[travis-ci] joyent/node#183 (v0.6 - 78dbb4b : isaacs): The build passed.
19:31:44  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/760928b...78dbb4b
19:31:44  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/482453
19:31:44  * travis-cipart
19:55:00  * mralephjoined
19:55:06  <ryah>mraleph: hey
19:55:16  <mraleph>ryah: hi
19:55:29  <ryah>mraleph: i need that patch backported to 3.6
19:55:52  <ryah>or a nod that my attempt is correct: https://gist.github.com/d5fcb57e00e1a32475b4
19:56:05  <mraleph>ryah: I understand that, there were some issues with it today.
19:56:35  <mraleph>ryah: also does node work for you correctly with the patch?
19:56:35  <ryah>any chance we'll get it today?
19:57:00  <ryah>yeah, seems to work
19:57:56  <mraleph>ryah: if by "get" you mean we commit backported patch to 3.6 then no, there is no chance.
19:58:14  <ryah>mraleph: a gist is fine :)
20:01:09  <mraleph>ryah: there are still issues with it. and I just realized that there is a severe bug.
20:01:22  <ryah>ok
20:02:02  <ryah>then we'll push the release back. any idea of how big of an issue there is?
20:03:21  <mraleph>you mean with the current patch? random hash mismatches.
20:03:33  <CIA-111>node: Mikael Bourges-Sevenier master * r5b05429 / (3 files in 2 dirs):
20:03:34  <CIA-111>node: typed arrays: add Buffer -> TypedArray constructor
20:03:34  <CIA-111>node: - create a typed array from a node::Buffer object
20:03:34  <CIA-111>node: - update TypedArray::set() to spec
20:03:34  <CIA-111>node: - add TypedArray::get() method - http://git.io/8PgZUw
20:03:53  <ryah>that sounds bad
20:04:00  <isaacs>yikes
20:05:01  <mraleph>they are very low probability. will happen only for hashes that have all 30 bits at the bottom equal to zero… one place in generated code was not updated to match C++ code, I noticed it only when I started looking at your
20:05:13  <mraleph>gist
20:05:56  <mraleph>theoretically it's one line (x number of arches) change; but who can guarantee that we did not miss the other one?
20:05:56  <ryah>well glad my extremely blind rebase was helpful :P
20:06:13  <AndreasMadsen>bnoordhuis: Thanks for looking intro #2462
20:06:24  <mraleph>i need to talk to Erik if we can test it more extensively.
20:06:32  <ryah>ok
20:07:32  <ryah>mraleph: there's a lot of people bugging me about this DoS thing - can you tell me tomorrow what the status is?
20:08:20  * sh1mmerjoined
20:10:41  <ryah>mraleph: also - is it correct that if we turn off snapshotting the seed will be generated on startup?
20:10:49  <mraleph>yep
20:11:06  <mraleph>I will update you tomorrow.
20:11:25  <ryah>infinite thanks
20:13:52  <ryah>AndreasMadsen: how's it going to with the cluster 2.0 stuff?
20:14:07  <AndreasMadsen>ryah: I'm writeing the text to cluster2 step 5 :)
20:14:13  <mraleph>actually it seems that generated code was out of sync with C++ code. and _now_ when we forgot to update it — they are in sync again. this is madness.
20:14:28  <AndreasMadsen>ryah: https://github.com/joyent/node/pull/2470
20:15:40  <AndreasMadsen>ryah: I wasn't as fast as usually since I found a really pain bug in the cluster 2.0 final: https://github.com/joyent/node/pull/2038#issuecomment-3374323
20:15:41  * travis-cijoined
20:15:41  <travis-ci>[travis-ci] joyent/node#184 (master - 5b05429 : Mikael Bourges-Sevenier): The build is still failing.
20:15:41  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/55c2197...5b05429
20:15:41  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/482724
20:15:41  * travis-cipart
20:15:44  <ryah>AndreasMadsen: im still not sold on autoFork
20:16:16  * AndreasMadsentopic: enoght
20:16:22  <AndreasMadsen>ryah: Also I have about 8 pull request running, you guys should have enough to do.
20:16:40  <mraleph>ryah: I think if you merge both http://code.google.com/p/v8/source/detail?r=103380 and http://code.google.com/p/v8/source/detail?r=10338 you will be fine.
20:16:59  <mraleph>ryah: I checked the generated code on all arches.
20:17:05  <AndreasMadsen>ryah: Okay, the current pull request is not about that but we can discuess it now.
20:17:06  * bnoordhuistopic: #libuv
20:19:53  <mraleph>ryah: your gist looks fine, just put r10338 on top of that; run all your and V8 tests in debug/release with and without snapshots and I think it is good to go.
20:20:10  <ryah>mraleph: yeah?
20:20:36  <ryah>mraleph: will the tests catch this hash mismatch problem?
20:21:17  <mraleph>ryah: it is madness but generated code was not in sync with C++ code before and now it is actually in sync. so the problem that I though is there is actually not there.
20:23:55  <ryah>uh -ok
20:28:26  <AndreasMadsen>ryah: fixed styling :)
20:37:04  <ryah>mraleph: which two revisions should i pull? one of those links was messed up
20:37:14  <ryah>r10338 i got
20:37:44  <mraleph>no i think you got r10330
20:38:18  <mraleph>so it's two revs in total: r10330, r10338
20:38:32  <mraleph>r10330 you already merged
20:41:41  <AndreasMadsen>bnoordhuis: just restored errno magic :)
20:41:54  <bnoordhuis>AndreasMadsen: good :)
20:42:22  <mmalecki>bnoordhuis: poke poke pull requests poke
20:42:41  <bnoordhuis>mmalecki: i'm running the benchmarks actually
20:43:06  <AndreasMadsen>really :)
20:43:39  <mmalecki>bnoordhuis: good for you :)
20:44:02  <AndreasMadsen>mmalecki: did you see my comment https://github.com/joyent/node/pull/2454#issuecomment-3353822
20:44:32  <mmalecki>AndreasMadsen: sounds good to me!
20:44:39  <mmalecki>bnoordhuis: does ^ work?
20:45:01  <mmalecki>AndreasMadsen: (sorry, I've like 70 notifications now, if I miss something, feel free to ping me)
20:45:29  <CIA-111>node: Emerson Macedo master * raa67b1f / (4 files in 3 dirs): fs: add appendFile() and appendFileSync() functions - http://git.io/DI1JEg
20:45:33  <bnoordhuis>only 70 :/
20:45:46  * piscisaureus_joined
20:46:02  <mmalecki>bnoordhuis: I try to clear it everyday
20:46:23  <mmalecki>(like in, actually go through them)
20:46:39  <bnoordhuis>mmalecki: as do i.... Select All, Mark Read
20:46:41  <mmalecki>but then I see some forever issues and just click Mark all as read
20:46:48  <bnoordhuis>exactly :)
20:47:35  <mmalecki>I kinda want to write something that'd behave sanily
20:47:46  <mmalecki>like, priorities, threading, listing mentions
20:48:34  <AndreasMadsen>mmalecki: I have about 8 patches on node, so I have enough to do already. It's not like I'm waiting for your reponse :)
20:48:58  <AndreasMadsen>+ school work
20:49:03  <piscisaureus_>mraleph: the bug that http://code.google.com/p/v8/source/detail?r=10338 solves also existed before we did the hash randomization stuff
20:49:08  <mmalecki>AndreasMadsen: haha :). no problem. I'll try to keep up
20:49:31  <mmalecki>AndreasMadsen: I also have some school work. I deal with it by ignoring it - still getting good grades
20:49:38  <mraleph>piscisaureus_: yep.
20:49:41  <mmalecki><success-kid>
20:50:02  <AndreasMadsen>mmalecki: I guess danish school is harder then :)
20:50:14  <mmalecki>AndreasMadsen: likely :)
20:50:22  <mraleph>piscisaureus_: but I did not realize that the logic in generated code is correct.
20:51:23  <piscisaureus_>mraleph: btw I think there is another code path that might need that change
20:51:26  <mraleph>AndreasMadsen: kids in Denmark learn to pronounce soft d that is already hardcore in my book.
20:51:48  <mraleph>piscisaureus_: point it :-0
20:52:02  <mmalecki>mraleph: ever tried speaking polish :) ?
20:52:33  <mraleph>mmalecki: dobre
20:52:55  <AndreasMadsen>mraleph: what book?
20:52:56  <mmalecki>mraleph: nieźle :)
20:53:28  <mraleph>mmalecki: yeah, you have some combinations of consonants that are definetely hard. [ok… no more offtopic]
20:53:54  <mraleph>piscisaureus_: so I am really very concerned that we missed something.
20:55:22  <piscisaureus_>mraleph: what exactly did we miss :-)
20:56:40  * travis-cijoined
20:56:41  <travis-ci>[travis-ci] joyent/node#185 (master - aa67b1f : Emerson Macedo): The build is still failing.
20:56:41  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/5b05429...aa67b1f
20:56:41  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/482863
20:56:41  * travis-cipart
20:56:57  <mraleph>piscisaureus_: the places that should be synced.
20:58:16  <piscisaureus_>mraleph: https://github.com/joyent/node/blob/master/deps/v8/src/objects.cc#L11422
20:58:58  <piscisaureus_>mraleph: merely a correctness issue though, if you don't fix that nobody will notice
21:00:09  <ryah>so we're going to turn off snapshots for the binary builds in v0.6.7, piscisaureus_
21:00:19  <piscisaureus_>ryah: ok, +1
21:00:40  <ryah>it would be very nice to be able to turn those on again at some point
21:00:49  <ryah>it's a 17ms hit on my laptop
21:00:52  <piscisaureus_>ryah: by how much will that increase startup time
21:00:54  <piscisaureus_>ah
21:01:15  <piscisaureus_>ryah: I am all for it
21:01:31  * ryahlooks longingly at the V8 team
21:01:39  <ryah>(puppy dog eyes)
21:02:14  <mraleph>ryah: Erik wants to fix it, the thing is engineering effort required for that is tremendous.
21:02:53  <mraleph>piscisaureus_: oh, yeah. thank you I thought that place was fixed.
21:03:01  <ryah>https://www.google.com/search?q=puppy+dog+eyes&tbm=isch
21:03:15  <`3rdEden>i'll throw in a free imaginary beer as well ;)
21:03:39  <AndreasMadsen>good night all of you :)
21:04:18  * AndreasMadsenquit (Remote host closed the connection)
21:04:22  <mmalecki>I'll throw in http://i.imgur.com/XgYDl.jpg
21:04:26  <ryah>here's a good visualization of the node team looking at the V8 team http://www.freewebs.com/xxcute-stuffxx/puppy%20dog%20eyes.jpg
21:04:37  <mmalecki>bnoordhuis: hey, opinions on what Andreas proposed?
21:04:39  * igorzijoined
21:05:17  <mmalecki>why did we disallow customFds at all?
21:05:32  <piscisaureus_>lubelska cytrynówka
21:06:13  <ryah>mmalecki: because windows doesn't have fds
21:06:37  <mmalecki>ryah: http://d37nnnqwv9amwr.cloudfront.net/entries/icons/original/000/003/617/okayguy.jpg
21:07:00  <mmalecki>ryah: have you looked at my patches, btw?
21:07:04  <ryah>no
21:07:08  <ryah>what patches?
21:07:09  <piscisaureus_>mmalecki: because on windows you can't just give an fd to a process (for stdio) and just expect it to work
21:07:43  <mmalecki>https://github.com/joyent/node/pull/2454 and https://github.com/joyent/node/pull/2455
21:07:47  <mmalecki>actually:
21:07:56  <mmalecki>!gh joyent/node#2454
21:07:56  <kohai>mmalecki, https://github.com/joyent/node/issues/2454
21:08:08  <mmalecki>!gh joyent/node#2455
21:08:08  <kohai>mmalecki, https://github.com/joyent/node/issues/2455
21:08:16  <ryah>i have a new official song for libuv
21:08:48  * ryahtopic: http://www.youtube.com/watch?v=i3Jv9fNPjgk
21:09:07  * ryahtopic: libuv song http://www.youtube.com/watch?v=i3Jv9fNPjgk
21:09:25  <mmalecki>I dig it.
21:10:26  <piscisaureus_>wut
21:10:44  <piscisaureus_>when did we drop korsakoff - my empty bottle?
21:10:52  <piscisaureus_>ryah, we've been over this before
21:10:59  <mmalecki>btw, http://www.youtube.com/watch?v=0znwavPhRc8 is nice #justsayin
21:12:42  <TooTallNate>ryah: so random, someone showed me that song this morning
21:13:36  <ryah>piscisaureus_: :D
21:13:45  <ryah>TooTallNate: it's the winter 2012 hit
21:14:03  <TooTallNate>yup yup :p
21:14:28  <TooTallNate>ryah: btw, i was able to get libffi compiled on my windows VM, so node-ffi on windows is close
21:15:02  <TooTallNate>i'm going the route evan suggested, by just using libffi's build system (autoconf) and just ensuring that the static lib file is present when building the module
21:18:06  <ryah>awesome
21:20:34  <igorzi>ryah bnoordhuis: so the I"P"C between isolates will be done through shared mem? (looking at bnoordhuis' https://github.com/bnoordhuis/node/compare/wip%5E%5E...wip patch)
21:24:12  <piscisaureus_>hmm
21:24:17  <piscisaureus_>I am somewhat surprised
21:24:27  <piscisaureus_>I thought ryah was working on a pipe for IPC
21:24:33  <piscisaureus_>now what's it gonna be
21:25:06  <piscisaureus_>ryah bnoordhuis : ^^
21:25:06  <igorzi>piscisaureus_: right, that's what i thought
21:25:54  <mmalecki>isolates are doing to be run in context of current process, right? I mean, when I, for example, add a property to `process`, i'll be able to see it in an isolate?
21:26:43  * `3rdEdenquit (Read error: Connection reset by peer)
21:32:07  <piscisaureus_>Unfortunately my battery is running out and the charger is in this pile of boxes somewhere. Looks like I'll have to start unpacking again :/
21:35:21  * `3rdEdenjoined
21:36:00  <ryah>piscisaureus_, mraleph: https://gist.github.com/1567443 unfortunately this is crashing on a lot of tests
21:36:17  * piscisaureus_quit (Ping timeout: 240 seconds)
21:36:42  <ryah>i had to change some stuff in code-stubs-ia32.cc
21:36:52  <ryah>i wonder if i did it incorrectly?
21:37:57  <mraleph>looks fine
21:42:34  <ryah>=== release array-constructor ===
21:42:34  <ryah>Path: mjsunit/array-constructor
21:42:34  <ryah>Command: /Users/ryan/projects/node-v0.6/deps/v8/d8 /Users/ryan/projects/node-v0.6/deps/v8/test/mjsunit/mjsunit.js /Users/ryan/projects/node-v0.6/deps/v8/test/mjsunit/array-constructor.js --test
21:42:38  <ryah>--- CRASHED ---
21:42:41  <ryah>for example
21:42:45  <ryah>sigh :(
21:43:04  <mraleph>very strange
21:43:30  <mraleph>it might be that the patch is incompatible with 3.6 unfortunately.
21:43:48  <mraleph>though I see no reason for it.
21:44:20  * piscisaureus_joined
21:44:23  <mraleph>I will ask Erik to try it with 3.6 tomorrow. I have to go sleep now, I can hardly think.
21:44:28  <ryah>ok
21:44:31  <ryah>mraleph: night
21:46:06  * mralephquit (Quit: Leaving.)
21:46:52  * piscisaureus__joined
21:48:37  * piscisaureus_quit (Ping timeout: 240 seconds)
21:51:23  <TooTallNate>bnoordhuis: ping
21:52:47  <TooTallNate>or ryah, or piscisaureus__, or anyone who knows the v8 API :p
21:53:24  * ircretaryjoined
21:53:25  <TooTallNate>i'm basically just wondering if there's a way to create an instance of a FunctionTemplate without invoking the associated C++ function
21:54:08  <TooTallNate>template->InstanceTemplate()->NewInstance() still ends up calling the C++ function
21:54:14  * piscisaureus___joined
21:54:26  * piscisaureus___changed nick to piscisaureus_
21:55:24  * piscisaureus__quit (Ping timeout: 240 seconds)
21:55:48  <paddybyers>TooTallNate: you can omit the C++ function in the constructor
21:56:08  * piscisaureus_quit (Read error: Connection reset by peer)
21:56:23  <TooTallNate>paddybyers: but i do want the function to invoke when the user calls from JS
21:56:34  <mmalecki>oh. https://github.com/joyent/node/pull/2472
21:56:49  * piscisaureus_joined
21:56:51  <mmalecki>I don't even.
21:57:52  <paddybyers>TooTallNate:sorry, I misunderstood
21:57:54  <CoverSlide>holy fuck
21:58:07  <benvie>wow I hope that was an automated conversion
21:58:13  <paddybyers>you can instance the function template (and get a function)
21:58:30  <paddybyers>and your c++ function is only called when that function instance is called from JS
21:59:00  <TooTallNate>paddybyers: well that seems to not be the case
21:59:02  <paddybyers>calling FunctionTemplate::GetFunction() doesn't invoke the function itself; it just instantiates it in the current context
21:59:14  <CoverSlide>I am hoping that was a joke
21:59:38  <TooTallNate>paddybyers: ya the GetFunction() part is fine, that goes to JS-land
21:59:47  <TooTallNate>but I'm trying to create an instance of one of these in C++ land
21:59:55  <TooTallNate>without invoking the associated JS constructor
22:00:27  <TooTallNate>I would think funcTemplate->InstanceTemplate()->NewInstance() would to the trick, but that still invokes the C++ function
22:01:26  <paddybyers>TooTallNate: yes, it will do .. but that's what you defined the FunctionTemplate to do
22:01:44  <paddybyers>…so I don't understand what you're trying to do
22:02:12  <TooTallNate>i'm trying to create an instance of said FunctionTemplate
22:03:09  <paddybyers>TooTallNate: terminology: an instance of a FunctionTemplate is a Function object corresponding to that function
22:03:30  <paddybyers>I think you're talking about an instance of the class for which that function is the contructor
22:03:42  <TooTallNate>paddybyers: yes, that's what I mean
22:04:30  <paddybyers>so yes, you call FunctionTemplate->GetFunction()->NewInstance()
22:04:47  <paddybyers>and it will call the C++ function that you specified when you did FunctionTemplate::New()
22:05:30  <TooTallNate>ok, i guess this module just does something kinda weird i suppose
22:05:30  <TooTallNate>https://github.com/rbranson/node-ffi/blob/master/src/pointer.cc#L111-117
22:05:59  <TooTallNate>instances created in C++ land don't need to allocate memory like instances from JS land do
22:06:42  <TooTallNate>so as a workaround I suppose I can pass some arbitrary Object into the constructor
22:06:56  <TooTallNate>which i can strictEquals() against in the New() function
22:07:24  <TooTallNate>pretty hacky, :\ but I can't think of anything better
22:09:22  <`3rdEden>mmalecki I tried the same conversion while mikesago, ian and I where at Nodeconf.it, we managed to convert convert a large part of the code automatically using JS2Coffee, only the HTTP and net modules needed hand conversion
22:10:58  <paddybyers>TooTallnate: OK I still don't understand the concern .. that call to NewInstance() will call Pointer::New() (at L124), and it will be entered with a new JS Object as args.This()
22:12:20  <TooTallNate>paddybyers: the concern is that i'm attempting to alter the New() function such that it throws an Error if the user doesn't pass any args.
22:12:30  <TooTallNate>then the NewInstance() call calls the constructor with no arguments
22:12:35  <TooTallNate>so it throws an exception
22:12:44  <TooTallNate>that's why I was trying to make it not be called in the first place
22:13:00  <TooTallNate>since everything that happens in that constuctor is redundant in the WrapPointer() context
22:14:29  <benvie>can be done in JS using harmony Proxies =X
22:15:38  <paddybyers>TooTallNate: ok, so you want to distinguish between the two cases that it's called from WrapPointer() or it's called from JS
22:15:49  <TooTallNate>correct
22:16:47  <paddybyers>ok, I don't know how to do that, sorry :( except with some nasty hack as you suggest
22:17:19  <paddybyers>you can distinguish between the case that it's called as a function or constructor
22:17:55  <paddybyers>but not constructor-from-JS vs constructor-from-native
22:18:07  <TooTallNate>paddybyers: ya but in this case they're both called as constructor
22:18:15  <TooTallNate>nasty hack it is I guess :\
22:20:26  <paddybyers>are these two cases really instances of the same class? or could you have two different constructors?
22:24:07  <TooTallNate>technically i could probably pull out the malloc'd version of the Pointer class into a subclass
22:25:58  * pieternjoined
22:29:53  * benviequit (Read error: Operation timed out)
22:30:17  * piscisaureus_quit (Ping timeout: 240 seconds)
22:46:27  <ryah>dannycoates, mjr_, dshaw_: http://dtrace.org/blogs/dap/2012/01/05/where-does-your-node-program-spend-its-time/
22:46:42  <ryah>you guys should really read that one
22:46:49  <dap>thanks ryah :) you're fast!
22:46:56  <mjr_>just reading that, dap just emailed me as well.
22:47:25  <dannycoates>ryah: thanks, reading
22:47:43  * sh1mmer_joined
22:48:19  <mjr_>Can't wait to get our stuff on SmartOS.
22:48:46  * sh1mmer_quit (Remote host closed the connection)
22:50:29  * sh1mmer_joined
22:50:46  * sh1mmerquit (Ping timeout: 255 seconds)
22:50:47  * sh1mmer_changed nick to sh1mmer
22:54:18  * Raynosquit (Write error: Connection reset by peer)
22:59:24  <`3rdEden>to bad it doesn't run on Mac OSX, anyways good excuse to install Smart OS on a VM again
22:59:40  <ryah>yes - it really SUCKS that it doesn't work on OSX
23:18:57  * paddybyersquit (Quit: paddybyers)
23:37:22  * Raynosjoined
23:52:19  * piscisaureus_joined
23:52:31  * indexzerojoined
23:53:01  <CIA-111>node: Ben Noordhuis master * r075acfa / (9 files in 5 dirs): uv: upgrade to 56a31f0 - http://git.io/gmOEzQ
23:53:02  <CIA-111>node: Ben Noordhuis master * rdadc303 / (6 files in 2 dirs):
23:53:03  <CIA-111>node: isolates: implement message passing
23:53:03  <CIA-111>node: Parent and child isolates can now pass arbitrary binary messages between each
23:53:03  <CIA-111>node: other. The messages are sent and received through a thread-safe queue that
23:53:03  <CIA-111>node: wakes up the event loop of the receiving thread. - http://git.io/6Q30fg
23:53:03  <CIA-111>node: Ben Noordhuis master * r1e73e4c / (src/node.cc src/node_isolate.cc src/node_isolate.h):
23:53:04  <CIA-111>node: isolates: remove global isolates list
23:53:04  <CIA-111>node: No longer necessary, each isolate now waits until its subordinate isolates have
23:53:05  <CIA-111>node: exited. - http://git.io/TnJnMQ
23:53:05  <CIA-111>node: Ben Noordhuis master * r4228112 / (lib/child_process.js src/node.js):
23:53:06  <CIA-111>node: child_process: add isolates support
23:53:06  <CIA-111>node: Passing an options object with {thread:true} to .fork() or .spawn() will run the
23:53:07  <CIA-111>node: target script in a thread instead of a separate process. - http://git.io/oUaD8g
23:55:01  <piscisaureus_>bnoordhuis: hey, yt?
23:55:08  <bnoordhuis>piscisaureus_: yep
23:55:40  <piscisaureus_>bnoordhuis: so what are you using for inter isolate communication
23:55:47  <piscisaureus_>shared memory or pipes?
23:55:57  <bnoordhuis>piscisaureus_: an in-memory message queue
23:56:06  <bnoordhuis>easier, faster
23:56:09  <piscisaureus_>bnoordhuis: igor and me got confused because ryah seems to work on pipes and you use shared memory
23:56:30  <bnoordhuis>yeah, i started out with ryah's uv_pipe_pair() bindings
23:56:40  <piscisaureus_>bnoordhuis: is the queue implemented in node or in libuv?
23:56:49  <bnoordhuis>node, it's in node_isolate.cc
23:57:22  <piscisaureus_>hmm
23:57:23  <piscisaureus_>ok
23:57:57  <bnoordhuis>it should Just Work(TM) on windows, if that's what you're asking
23:58:05  <bnoordhuis>how did the move go btw?
23:58:30  <piscisaureus_>bnoordhuis: yeah but it's kind of backwards since we use a message queue to emulate ev_async and then we use that to build a message queue :-/
23:58:45  <piscisaureus_>bnoordhuis: the move went well although my house now is a total mess
23:58:47  <piscisaureus_>moving is painful
23:59:41  <bnoordhuis>yeah... put stuff in box, carry over box, unpack box
23:59:59  <bnoordhuis>in my case, step 3 usually happens six months later