00:12:39  * lohkey_joined
00:12:39  * lohkey_quit (Client Quit)
00:13:28  <piscisaureus>I just got reminded that I need to make libuv releases
00:15:05  <piscisaureus_>"As a reminder, on October 31, 2012, Joyent will bring support for our no.de platform as a service (PaaS) to an end." <-- slurp's days are numbered :-(
00:16:21  * lohkeyquit (Ping timeout: 252 seconds)
00:19:04  <piscisaureus>TooTallNate: install indutny's library, set a breakpoint in all these funtions and enter "disas" when you hit them
00:19:47  <piscisaureus>TooTallNate: (I'm just repeating what indutny told me earlier tonight)
00:20:05  <TooTallNate>piscisaureus: thanks, i'll give it a shot
00:20:36  <TooTallNate>piscisaureus: it was very exciting to see this working on 10.6 though! https://gist.github.com/3751570
00:21:11  <piscisaureus>TooTallNate: ya, indutny showed me that. Nice huh :-)
00:21:38  <TooTallNate>i don't at all understand his event loop code, but it seems like it works
00:21:57  <TooTallNate>i mean, i understand the libuv parts at least :)
00:22:02  * loladirojoined
00:25:12  * loladiroquit (Client Quit)
00:26:14  <piscisaureus>TooTallNate: it's not complicated. He just blocks on the CF mach ports on another thread and then signals the main thread when there are events available
00:26:41  <piscisaureus>TooTallNate: infact, his solution is a little lame ... I will discuss this with indutny later
00:26:54  <TooTallNate>haha
00:26:57  * loladirojoined
00:27:04  * loladiroquit (Client Quit)
00:30:57  * EhevuTovquit (Quit: This computer has gone to sleep)
00:38:27  <piscisaureus>mikeal: don't shut down the list. I like it as a place where people can just ask for help.
00:38:34  * avalanche123joined
00:38:56  * stagasquit (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
00:39:16  <mikeal>piscisaureus: i don't think it's doing that well enough
00:39:23  <mikeal>we need to find a better way for people to get help
00:39:31  <piscisaureus>mikeal: stack overflow?
00:39:34  <mikeal>and after we do, and we know it works, we should shut down the list
00:39:50  <mikeal>not enough people invested in node are sitting on SO questions
00:40:05  <mikeal>to answer them, whatever we do has to be integrated heavily with GitHub
00:40:17  <mikeal>because everyone has some kind of workflow for managing notifications on GitHub
00:40:39  <piscisaureus>mikeal: well... github is sort of my daily todo list
00:40:46  <piscisaureus>mikeal: and even more so, ben's
00:41:10  <piscisaureus>I don't think I would be really happy to have support questions there
00:41:53  <mikeal>you don't have to subscribe :)
00:42:22  <piscisaureus>mikeal: well, I do, since fixing node issues is my job. Providing support isn't
00:42:54  <mikeal>node issues, like bugs or comments or features requests, should be github issues on joyent/node
00:43:00  <piscisaureus>yes
00:43:04  <piscisaureus>to that I agree
00:43:09  <mikeal>what the mailing list is suppose to be for, that i would like to replace, are higher level "help" questions
00:43:22  <mikeal>like "i have an exception, no idea what its in cause i have 30 modules"
00:43:32  <piscisaureus>mikeal: yes. i don't want to have that on github
00:43:32  <mikeal>or "what module should i use for x,y,z"
00:43:46  <piscisaureus>this is also where the community can really help out.
00:43:49  <mikeal>you're missunderstanding
00:43:56  <mikeal>on github, in a different repo
00:44:06  <mikeal>with a different interface, easier for newbies
00:44:16  <piscisaureus>right
00:44:19  <mikeal>but, you login with GitHub
00:44:36  <CoverSlide>what about the important stuff: industry gossip?
00:44:46  <piscisaureus>hacker news, please
00:44:52  <mikeal>CoverSlide: that's what Twitter is for
00:46:09  <piscisaureus>mikeal: if people really have something interesting to say I generally direct them to this channel.
00:46:27  <mikeal>IRC is too much for new users
00:46:34  <mikeal>many of them have never gone on IRC
00:46:39  <piscisaureus>true
00:46:43  <piscisaureus>which is good
00:46:56  <piscisaureus>since I don't want this to be a support channel :-)
00:48:42  <piscisaureus>although when you're intelligent and you're finding interesting issues (like voxer and nodejitsu) you're welcome here I'd say
00:49:12  <mikeal>those people know where to go and are getting what they need, they are not my concern :)
00:49:24  <piscisaureus>mikeal: but is the mailing list really that bad for new users?
00:49:46  <piscisaureus>mikeal: it seems that most of them get an answer if they ask a reasonable question
00:53:23  <isaacs>true, most threads do actually get resolved quickly
00:53:36  <mikeal>most of them don't ask
00:53:36  <isaacs>that's the problem - the long-running threads are always the hazards.
00:53:50  <isaacs>and those long-running threads make it seem like that's what the mailing list is for
00:53:56  * ericktquit (Ping timeout: 268 seconds)
00:54:01  <mikeal>compare our user list to a project with less traction
00:54:02  <isaacs>it'd be nice if google groups gave us a way to limit the number of responses that would be accepted.
00:54:07  <mikeal>and less users
00:54:17  <mikeal>and we don't have that many new user questions coming in
00:54:21  <mikeal>not like we should be getting
00:55:15  <piscisaureus>mikeal: so what's the problem you're seeing? People are not asking questions?
00:55:28  <mikeal>https://groups.google.com/forum/?fromgroups#!forum/selenium-users
00:55:31  <mikeal>30 threads active today
00:55:37  <mikeal>just user questions and issues
00:55:47  <mikeal>we have WAY more users than Selenium
00:58:34  * isaacsgoing back to headcold-land.
00:58:38  <mikeal>questions simply aren't being asked
00:58:41  <isaacs>have fun plotting the demise of the mailing list.
00:58:49  <isaacs>lmk what you guys come up with.
00:58:53  <piscisaureus>isaacs: haha
00:59:04  <piscisaureus>isaacs: see you tomorrow. Do you think the standup will be with you?
00:59:09  <isaacs>yes.
00:59:12  <piscisaureus>cool
00:59:15  <isaacs>i'll be better by then. have to be.
00:59:18  <isaacs>big plans tomorrow night.
00:59:24  * dapquit (Quit: Leaving.)
00:59:33  <piscisaureus>get better isaacs
00:59:43  <isaacs>thanks
00:59:45  * isaacszzzz
01:00:43  <piscisaureus>mikeal: but do you think questions aren't being asked because of these long and acid discussions?
01:01:05  <piscisaureus>mikeal: I mean, if those weren't happening then the ML would just be more silent.
01:03:41  <mikeal>i think that people on both sides have checked out of the ML
01:03:47  <mikeal>people that would ask questions, and those that would answer them
01:04:11  <mikeal>because of the volume of meaningless garbage that doesn't belong there
01:04:49  * joshthecoderquit (Quit: Leaving...)
01:08:06  <piscisaureus>mikeal: well, if you have practical ideas how to do better than you're welcome. As long as it doesn't come down to me and ben and isaac and nate having to answer a pile of support questions every day. This should be a community thing.
01:08:18  <piscisaureus>s/than/then/
01:08:22  <mikeal>definitely
01:08:27  <mikeal>writing some code now :)
01:08:34  <piscisaureus>mikeal: that's also good
01:09:42  <mikeal>i'm starting to really love writing little apps that live on top of other people's APIs and have no database or filesystem of their own :)
01:10:35  <piscisaureus>I totally get that.
01:11:06  <piscisaureus>dealing with persistent data is a chore. You have to be careful and stuff.
01:11:09  <tjfontaine>bonus points if you store the issues in a git repo, and your app is just a pretty way to display and manage etc :)
01:12:28  <piscisaureus>the challenge is to engage the community enough to actually answer support questions
01:12:34  <piscisaureus>stack overflow figured this out
01:14:03  * xaqquit (Remote host closed the connection)
01:22:11  <piscisaureus>mikeal: also, Marco Rogers had a good point. Lame discussions on ML won't derail the node project any more.
01:22:38  <mikeal>they derail other people from asking questions
01:22:51  <piscisaureus>mikeal: true
01:22:58  <mikeal>they don't effect people working on core because they's mostly stopped watching the ML because of it
01:23:46  <piscisaureus>mikeal: true. But let's face it: nobody asks questions to the php mailing list, yet everyone is perfectly capable of using php.
01:24:28  <mikeal>because they have docs with comments on them where people ask questions :)
01:24:34  <piscisaureus>yeah
01:24:37  <piscisaureus>I would dig that
01:24:45  <mikeal>i would hate that :)
01:24:52  <mikeal>i like our docs :)
01:25:14  <piscisaureus>ok. We disagree, then :-)
01:26:21  <piscisaureus>mikeal: I just had a look at the nodejs ML again and most topics there are okay, it's just this threads discussion that always comes up.
01:28:35  <piscisaureus>mikeal: unfortunately we get trolled to easily into actually engaging in this discussion. I'd probably have helped if someone sent this rektide dude a private email explaining the state of isolates.
01:29:44  <piscisaureus>I plea guilty
01:31:03  <mikeal>per threads percentage is good
01:31:08  <mikeal>per message percentage is aweful
01:32:04  <ryah>yo
01:32:06  <piscisaureus>I have to admit I look at the web interface, since I unsubscribed a long time ago :-)
01:32:10  <piscisaureus>yo ryah
01:32:26  <ryah>ball so hard
01:32:35  <piscisaureus>ummm .... ?
01:32:48  <ryah>nothign
01:33:16  <bnoordhuis>piscisaureus: see you in a couple of hours, bertje
01:33:18  <bnoordhuis>i'm off to bed
01:33:24  <piscisaureus>bnoordhuis: oh, fuck
01:33:30  <piscisaureus>bnoordhuis: yeah, cu
01:35:33  <ryah>what's new bert?
01:36:16  <piscisaureus>ryah: speed is the new pot
01:36:30  <ryah>:)
01:37:08  <piscisaureus>ryah: nothing really. i'm distracted by some internal chores that are over soon
01:37:59  <piscisaureus>ryah: node-wise optimizing tls is now high on the list.
01:38:18  <piscisaureus>ryah: and closing the gap with http.sys
01:38:35  * bnoordhuisquit (Ping timeout: 256 seconds)
01:38:49  <piscisaureus>ryah: so when can I come to nyc again?
01:40:34  <piscisaureus>also, corporate is the new hipster
01:43:25  * abraxasjoined
02:00:21  * toothrchanged nick to toothrot
02:00:33  <tjfontaine>TooTallNate: is there an existing solution where an extension isn't appended to a library name, or will I just have to do it all manually with DynamicLibrary?
02:01:37  <tjfontaine>statSync would work, but I have a strong suspicion that no one else will like that idea
02:02:59  <piscisaureus_>I'm off peepz
02:03:03  <piscisaureus_>goodnight all
02:03:17  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:06:43  * mmaleckiquit (Ping timeout: 252 seconds)
02:26:42  <TooTallNate>tjfontaine: you're saying you want Library, but without the extension appending behavior?
02:26:54  <TooTallNate>(just clarifying)
02:28:42  <tjfontaine>TooTallNate: yes, I'm specifying a full path, and in this case no extension is in the filename, for a workaround for now in the extension check I just added a try{ fs.statSync() }
02:29:16  <tjfontaine>TooTallNate: though, if someone were to want to (for whatever reason) try and get bindings.node they would hit the same issue
02:31:02  <TooTallNate>we could add an option maybe
02:31:31  <tjfontaine>an option would work for me, I'm already doing the stat external to creating the library
02:35:34  <TooTallNate>DynamicLibrary is essentially doing sync io under the hood so ideally you do it at boot-up anyways
02:38:36  <tjfontaine>if that's the case, then I don't feel so bad about a statSync in Library
02:42:37  * ericktjoined
02:43:28  * Benviequit (*.net *.split)
02:43:28  * toothrotquit (*.net *.split)
02:44:04  * toothrjoined
02:49:00  * xaqjoined
02:49:31  * Benviejoined
02:49:50  * joshthecoderjoined
02:51:16  * loladirojoined
02:59:16  * brsonquit (Quit: leaving)
03:13:51  * mmaleckijoined
03:19:00  * mmaleckiquit (Ping timeout: 244 seconds)
03:41:47  * brsonjoined
04:19:52  * TooTallNatequit (Quit: ["Textual IRC Client: www.textualapp.com"])
04:29:40  * loladiroquit (Quit: loladiro)
04:30:17  * loladirojoined
04:38:27  * xaqquit (Remote host closed the connection)
04:41:54  * AvianFluquit (Quit: AvianFlu)
04:43:46  * ericktquit (Quit: erickt)
04:57:06  * loladiroquit (Quit: loladiro)
05:05:46  * jerroquit (Remote host closed the connection)
05:11:47  * loladirojoined
05:13:13  * xaqjoined
05:14:56  * xaqquit (Remote host closed the connection)
05:19:44  * paddybyersjoined
05:20:20  * philips_quit (Excess Flood)
05:24:02  * philips_joined
05:36:53  * xaqjoined
05:44:41  * xaqquit (Remote host closed the connection)
05:45:14  * mikealquit (Quit: Leaving.)
05:59:46  * paddybyersquit (Quit: paddybyers)
05:59:54  * mikealjoined
06:03:00  * xaqjoined
06:24:29  * loladiroquit (Quit: loladiro)
06:39:17  * stagasjoined
06:40:59  * turddasheedonjoined
06:45:11  * brsonquit (Read error: Connection reset by peer)
06:45:44  * turddasheedonchanged nick to `3rdEden
06:50:00  * bulatshakirzyanojoined
07:03:39  <indutny>piscisaureus: heya
07:17:10  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
07:22:15  * paddybyersjoined
07:53:38  * rendarjoined
08:38:52  <indutny>piscisaureus: ping
08:51:17  * joshthecoderquit (Quit: Linkinus - http://linkinus.com)
10:39:28  * stagasquit (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
10:40:48  * ki9ajoined
10:40:58  <ki9a>are there up to date bindings for C#/LibUV?
10:41:08  <ki9a>the only ones I could find were on github and > 1 year old.
10:52:31  <ki9a>this is probably obvious, maybe even a faq (Though I couldn't find it); I'm using loop = uv_loop_new(); uv_ref(loop); uv_run(loop); ; uv_run exits immediately
10:52:39  <ki9a>how do I keep it running until I call unref
10:52:56  <ki9a>(I've got uv_run in a thread)
11:24:53  <saghul>uv_ref applies to handles now
11:25:22  <saghul>you need to have an active handle or an inactive handle which refs the loop to prevent it from exiting
11:25:25  <saghul>ki9a ^
11:25:38  <ki9a>ah ke
11:25:42  <ki9a>an async handle will work
11:26:49  <ki9a>thanks saghul.
11:27:02  <saghul>ki9a welcome
11:27:06  <ki9a>I'm surprised how well things work on Windows (even when using pinvoke from C#)
11:27:13  <ki9a>thats unusual for open source projects.
11:39:46  <piscisaureus>heya indutny
11:39:47  <piscisaureus>sup?
11:40:06  <indutny>piscisaureus: uv_run_embedded
11:40:09  <indutny>piscisaureus: what do you think
11:40:13  <piscisaureus>well
11:40:14  <indutny>piscisaureus: I've implemented that API
11:40:16  <piscisaureus>what does it do?
11:40:17  <piscisaureus>ah
11:40:32  <indutny>piscisaureus: yeah, it allows you to embed uv_loop_t inside any other event loop
11:40:39  <piscisaureus>is that that library that you pointed me to yesterday>
11:40:46  <indutny>piscisaureus: huh, what lib?
11:40:51  <piscisaureus>ok, nvm
11:40:53  <indutny>was that me?
11:40:54  <indutny>:)
11:40:54  <piscisaureus>indutny: where's the code?
11:41:01  <indutny>piscisaureus: one sec, I'll push it
11:41:03  <piscisaureus>link link link
11:41:46  <indutny>piscisaureus: https://github.com/joyent/libuv/pull/561
11:41:58  <indutny>piscisaureus: there's a test with example
11:43:52  <indutny>piscisaureus: it's basically polling in another thread
11:44:05  <indutny>piscisaureus: and asking you to invoke uv_poll_embedded in main thread when needed
11:44:26  <indutny>piscisaureus: works fine on darwin
11:44:32  <indutny>and has no ABI changes
11:44:33  <indutny>;)
11:44:51  <indutny>piscisaureus: can this be implemented on windows?
11:49:28  <mitsuhiko>is there a limit to how large uv_write's individual buffers can be?
11:50:27  <indutny>mitsuhiko: it's usually limited by kernel
11:50:44  <mitsuhiko>so basically, i should split it up myself into small chunks
11:51:29  <ki9a>How long do I have to keep around the data passed to uv_write ?
11:52:14  * bnoordhuisjoined
11:52:24  <CIA-131>libuv: Fedor Indutny master * r73cf360 / src/unix/fsevents.c : darwin: fsevents: emit UV_CHANGE on metadata change - http://git.io/g7nAnQ
11:52:36  <indutny>piscisaureus: thanks
11:52:56  <piscisaureus>ki9a: until the write callback is made
11:53:01  * abraxasquit (Remote host closed the connection)
11:53:13  <indutny>bnoordhuis: hoya
11:53:16  <ki9a>thanks.
11:53:22  <indutny>Benvie: have you seen this https://github.com/joyent/libuv/pull/561/files ?
11:53:24  <indutny>oops
11:53:25  <indutny>bnoordhuis: ^
11:53:28  <indutny>Benvie: sorry
11:53:35  <piscisaureus>indutny: so, can you give me an executive summary?
11:53:43  <bnoordhuis>indutny: yes. apparently you broke the build
11:53:56  <indutny>bnoordhuis: haha, yes
11:54:05  * travis-cijoined
11:54:05  <travis-ci>[travis-ci] joyent/libuv#723 (master - 73cf360 : Fedor Indutny): The build passed.
11:54:05  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/2916a1606738...73cf3600d75a
11:54:05  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/2516505
11:54:05  * travis-cipart
11:54:05  <indutny>bnoordhuis: that's because this feature is not implemented for other platforms than darwin
11:54:15  <indutny>piscisaureus: what do you expect from me?
11:54:28  <bnoordhuis>indutny: but i don't want to land that now anyway, it means i'll have to redo it when libev is removed
11:54:39  <piscisaureus>indutny: so this allows you to poll another loop on the main thread and libuv will wait for events in the side thread?
11:54:40  <indutny>bnoordhuis: I don't think so
11:54:50  <indutny>piscisaureus: yeah
11:55:00  <indutny>piscisaureus: and interrupt another loop to invoke callbacks
11:55:10  <indutny>bnoordhuis: it's not related to libev at all
11:55:19  <indutny>bnoordhuis: it's only taking backend_fd out of it
11:55:51  <bnoordhuis>oh, right
11:56:19  <indutny>I honestly tried to implement it in userland
11:56:23  <indutny>but it doesn't seem to be possible
11:56:36  <indutny>even with hacks
11:57:11  <indutny>bnoordhuis: this will allow people running Cocoa applications on node.js
11:57:28  <indutny>and having all TCP and timers stuff working
11:57:38  <piscisaureus>indutny: right
11:57:48  <piscisaureus>indutny: looks interesting. And yes, it can be implemented on windows
11:57:52  <indutny>piscisaureus: great
11:57:54  <indutny>bnoordhuis: epoll?
11:57:59  <indutny>will it work with it?
11:58:08  <piscisaureus>indutny: i think you can even make it generic
11:58:15  <piscisaureus>indutny: just call select on the backend fd
11:58:24  <piscisaureus>indutny: I think that works on any platform except windows
11:58:29  <indutny>piscisaureus: hm... really?
11:58:32  <indutny>let me test
11:58:51  <piscisaureus>or whatever, poll()
11:59:03  <piscisaureus>poll seems more convenient
11:59:51  <indutny>yeah
11:59:53  <indutny>lets try
12:00:03  <indutny>I'm not sure if one call poll() on kqueue's fd
12:00:15  <ki9a>can I get an uv_tcp_t from an existing socket handle /
12:00:16  <ki9a>?
12:00:16  <piscisaureus>i don't know either
12:00:20  <piscisaureus>but it's worth trying
12:02:29  <piscisaureus>ki9a: not atm. there is a patch pending for that.
12:02:40  <indutny>something odd is happening there
12:02:40  * mmaleckijoined
12:02:40  <indutny>:)
12:02:46  <ki9a>oh. I just came accross uv_tcp_open << that's not it then?
12:02:54  <piscisaureus>ki9a: does that work>
12:02:59  <ki9a>I have no idea :)
12:03:02  <piscisaureus>ki9a: that's supposed to do it indeed
12:03:08  <ki9a>still got 100 lines of infrastructure to write before i get ther
12:03:12  <piscisaureus>I was not aware that patch already landed
12:03:39  <indutny>piscisaureus: apparently, no
12:03:43  <ki9a>the implementation looks good.
12:05:16  <indutny>piscisaureus: EPOLLNVAL
12:05:24  <indutny>err
12:05:26  <indutny>POLLNVAL
12:06:15  <indutny>bnoordhuis: so, I'll implement it for other unixes then
12:06:24  <indutny>what do you guys think about APIs?
12:06:26  <indutny>any suggestions?
12:06:37  <piscisaureus>indutny: well, there's something
12:06:48  <piscisaureus>indutny: this looks useful in situations where event loops can't be joined
12:07:15  <piscisaureus>indutny: but we also need another embed api that allows you to replace the poll function
12:07:53  <indutny>piscisaureus: hm... I hadn't thought about it
12:08:02  <indutny>piscisaureus: and this seems like a custom poll
12:08:05  <indutny>not really an embedding
12:08:19  <indutny>embedding is an application of this API
12:08:19  <piscisaureus>indutny: well, it achieves the same (embed another loop) but with a different strategy
12:08:29  <indutny>yeah
12:08:34  <indutny>piscisaureus: btw, it's not needed for node-cf
12:08:54  <indutny>piscisaureus: so if you don't know any other application, I doubt it'll be needed by anyone else
12:09:07  <bnoordhuis>piscisaureus: https://gist.github.com/3a283a6e865dfafbe7a5
12:09:14  <piscisaureus>indutny: well, there are more events loops
12:09:23  <piscisaureus>e.g. CEF, qt etc
12:09:38  <indutny>ok
12:09:43  <indutny>:)
12:09:56  <indutny>still I think that embedding libuv with my method is more convenient
12:10:21  <piscisaureus>indutny: I dont disagree
12:10:29  <piscisaureus>indutny: although that method is probably more efficient
12:10:46  <indutny>yeah
12:10:49  <piscisaureus>indutny: with your method you end up switching threads all the time.
12:11:09  <indutny>yeah, I understand
12:11:16  <indutny>but sometimes it's the only possible way to do it
12:11:19  <indutny>unfortunatelly
12:11:41  <indutny>NSApplication run - is running CFLoop and getting events and processing them inside of the call
12:11:50  <indutny>that's why I did this API
12:12:14  <indutny>NSApplication:run it's like following loop:
12:12:25  <indutny>while (event = blockingGetEvent()) processEvent(event)
12:12:45  <indutny>and blockingGetEvent is calling __CFRunLoopRun after all
12:13:21  <piscisaureus>indutny: yes - I know
12:13:28  <piscisaureus>indutny: so I am not apposed to your API
12:13:39  <piscisaureus>indutny: I was just thinking about how this all fits in
12:14:10  <piscisaureus>indutny: I like this api because it offers a solution for doing windows gui aps
12:14:24  <piscisaureus>which was extremely hairy before
12:15:41  <piscisaureus>indutny: the only thing I am not sure about is whether we should all this api surface to libuv. eventually the only interesting function that people could not write themselves is "block on the backend fd on another thread"
12:16:55  <indutny>piscisaureus: yeah, you're right
12:17:04  <indutny>this is what I wasn't able to implement
12:17:19  * mmaleckiquit (Remote host closed the connection)
12:17:35  <indutny>I'm ok with moving it to userland
12:17:39  * mmaleckijoined
12:17:44  <indutny>but it may be useful for people
12:17:57  <indutny>and it doesn't add a lot of complexity inside
12:18:07  <indutny>(i.e. it's very straightforward)
12:19:06  <piscisaureus>indutny: sure
12:19:15  <piscisaureus>indutny: I will try to do this for windows
12:19:28  <piscisaureus>indutny: on windows there is no function to block on the completion port without dequeueing
12:19:28  <indutny>piscisaureus: lets expose uv_wait then
12:19:41  <indutny>I'll update my pull request
12:20:02  <piscisaureus>indutny: so I have to think about the implementation strategy on windows.
12:20:15  * paddybyersquit (Ping timeout: 240 seconds)
12:21:10  <indutny>bnoordhuis: btw, probably you're right about libev
12:21:23  <indutny>bnoordhuis: it seems to be a big PITA to implement uv_wait for poll() or select() engine
12:21:32  <ki9a>uv_close is used for graceful tcp disconnect too ?
12:21:38  <bnoordhuis>ki9a: yes
12:21:41  <indutny>bnoordhuis: though, it's straightforward for both epoll and kqueue
12:21:45  <ki9a>perfect. thakns.
12:21:51  <bnoordhuis>indutny: yes. and i don't plan to support poll or select in faio :)
12:22:19  <indutny>bnoordhuis: huh?
12:22:25  <indutny>bnoordhuis: what about cygwin?
12:22:35  <indutny>bnoordhuis: are we dropping support for it?
12:22:37  <bnoordhuis>indutny: i don't care
12:22:41  <indutny>ah ok
12:22:42  <indutny>:)
12:22:46  <bnoordhuis>i think we have exactly zero cygwin users
12:22:57  <indutny>ok
12:23:06  <indutny>bnoordhuis: and you're going to use completion ports for sunos, right?
12:23:33  <bnoordhuis>indutny: you mean event ports?
12:23:39  <indutny>the same thing...
12:23:50  <piscisaureus>ki9a: it will gracefully disconnect, but if you want to do proper 1-way close you should use uv_shutdown
12:24:03  <ki9a>1 way close?
12:24:12  <bnoordhuis>indutny: i think you'll get blank stares from sunos people if you mention completion ports
12:24:13  <indutny>bnoordhuis: can I help you? I really would like to "boost" development of it
12:24:17  <piscisaureus>ki9a: send a FIN packet but still allow reading
12:24:24  <indutny>bnoordhuis: well, AIX is essentially the same thing AFAIK
12:24:31  <bnoordhuis>indutny: no, aix has pollset
12:24:37  <ki9a>oh oke
12:24:39  <piscisaureus>indutny: AIX has it's own special way right?
12:24:39  <bnoordhuis>which we don't support right now but can
12:24:40  <ki9a>thanks.
12:24:51  <indutny>ah
12:24:53  <indutny>ok, I was wrong
12:25:13  <indutny>piscisaureus: pushed update to my pull request
12:25:21  <bnoordhuis>we can even support hp-ux, it has /dev/poll :)
12:25:27  <indutny>piscisaureus: exposed uv_wait and copy pasted uv_wait for all bsd platforms
12:25:28  <piscisaureus>indutny: another downside of this method is that prepare and check callbacks are made at the wrong point
12:25:33  <piscisaureus>but I think we can just ignore that
12:25:45  <indutny>piscisaureus: yeah, we can
12:25:54  <indutny>piscisaureus: so far
12:26:02  <indutny>piscisaureus: that's a good reason of having this API implemented internally
12:26:06  <indutny>we can always reshape it
12:27:29  * mmaleckiquit (Ping timeout: 248 seconds)
12:28:34  <piscisaureus>what happens if someone wants to embed 2 event loops? :-)
12:29:12  <indutny>piscisaureus: two threads will be running
12:29:18  <indutny>piscisaureus: and each will be polling it's own event loop
12:29:28  <indutny>piscisaureus: and interrupt main event loop when poll needs to be done
12:29:28  <indutny>:)
12:29:28  * TheJHjoined
12:29:30  <piscisaureus>ububub
12:29:31  <indutny>the same thing
12:29:44  <piscisaureus>ik snap er geen zak meer van
12:30:04  <indutny>piscisaureus: you're not a bag?
12:30:10  <indutny>congratulations
12:30:31  <piscisaureus>indutny: dutch course lesson 2
12:30:46  <piscisaureus>indutny: it means "I totally don't get it anymore"
12:30:54  <indutny>that's what I thought
12:30:57  <indutny>russian translation was awful
12:31:01  <indutny>english was much better
12:31:21  <indutny>dutch is something between german and french
12:31:24  <indutny>I suppose
12:31:26  <piscisaureus>indutny: so, the *uv* loop is polled on a separate thread right?
12:31:32  <indutny>piscisaureus: yes
12:31:40  <indutny>bnoordhuis: works on linux now https://github.com/joyent/libuv/pull/561 ;)
12:31:48  <piscisaureus>indutny: so, if you have 2 embed loops then there are 2 threads that poll the *uv* loop
12:31:54  <indutny>yes
12:31:54  <piscisaureus>I don't think that will be possible in windows
12:32:10  <indutny>piscisaureus: why not?
12:32:27  <piscisaureus>indutny: well, because you can't poll an iocp port without dequeuing packets
12:32:41  <piscisaureus>indutny: so 1 packet would have to wake up 2 loops
12:32:44  <piscisaureus>but that won't work
12:33:42  <indutny>ah
12:33:54  <indutny>em...
12:33:57  <indutny>I don't understand that
12:33:59  <piscisaureus>indutny: ah - wait
12:34:16  <indutny>piscisaureus: iocp is edge-triggered?
12:34:20  <piscisaureus>indutny: so the idea is that on every loop iteration you call uv_run_embedded manually
12:34:29  <piscisaureus>indutny: you could call it edge triggered yes
12:34:51  <indutny>piscisaureus: oh, that's not good :)
12:35:13  <piscisaureus>indutny: so if you have 2 embedded loops (say A and B) and you're doing uv_run_embedded(A) and an event arrives on B then it doesn't work
12:35:16  <piscisaureus>that's probably fine
12:35:22  <piscisaureus>not something for us to worry about
12:35:56  * mmaleckijoined
12:36:50  <indutny>ok
12:36:54  <indutny>you're getting API wrong
12:36:58  <indutny>and this is probably my fault
12:37:16  <indutny>think of uv_run_embedded as of uv_embed()
12:37:21  <indutny>it's not what uv_run() does
12:37:26  <indutny>it just starts a thread
12:37:41  <indutny>and invokes callback in uv_embed_t when there're events to process
12:37:56  <indutny>uv_poll_embedded() is a wrapper around uv_run_once()
12:38:18  <indutny>which should be called in main thread when callback in uv_embed_t was fired
12:38:24  <indutny>in the house that jack built
12:38:59  <piscisaureus>ghe
12:39:22  <indutny>namings are bad
12:39:24  <indutny>I know
12:39:54  <piscisaureus>indutny: do you mind if I sit on this for a while. I think it's a good way of looking at the problem
12:40:07  <indutny>sure
12:40:15  <indutny>bnoordhuis: does epoll_wait blocks when event count is zero?
12:40:37  <piscisaureus>indutny: yes
12:40:41  <indutny>good
12:41:39  <piscisaureus>indutny: so infact what we need is a function that 1) waits for events to arrive but doesn't process the events and 2) is threadsafe
12:42:06  <indutny>piscisaureus: oh, thread safety...
12:42:09  <indutny>good point
12:42:15  * mmaleckiquit (Ping timeout: 240 seconds)
12:42:26  <indutny>I need more synchronization
12:42:48  <piscisaureus>indutny: I think that if we offer that function then people can do the synchronization tricks etc themselves
12:43:06  <indutny>piscisaureus: it's not so hard actually
12:43:11  <indutny>I just need a semaphore
12:43:19  <piscisaureus>indutny: since you are already asking jack to build a house, you might as well ask jack to feature some plumbing
12:43:27  <indutny>that's it
12:43:30  <indutny>oh not
12:43:31  <indutny>it isn't
12:43:32  * mmaleckijoined
12:44:16  * `3rdEdenquit (Ping timeout: 244 seconds)
12:44:29  * paddybyersjoined
12:46:00  * paddybyersquit (Client Quit)
12:46:11  <indutny>oh crap
12:47:10  <indutny>piscisaureus: ok, look at this https://github.com/joyent/libuv/pull/561/files#L3R293
12:47:14  <indutny>this one should be correct
12:47:44  <indutny>it's not stopping immediately now, but I'll fix it in a bit
12:47:58  * paddybyersjoined
12:52:15  * mmaleckiquit (Ping timeout: 240 seconds)
12:53:43  * mmaleckijoined
12:54:20  <indutny>piscisaureus: so it doesn't need to be thread safe now
12:56:55  <saghul>indutny what about void uv_poll(uv_loop_t *loop, int timeout)
12:57:14  <saghul>indutny and then the user is responsible for spawning a thread and blocking there
12:57:36  <indutny>saghul: it's not really a poll
12:57:49  <indutny>saghul: but that's indeed the question we're trying to solve
12:58:01  <indutny>provide APIs that people using uv_wait will have to build anyway
12:58:08  <indutny>or let those people create their own APIs
12:58:12  <indutny>or do that at the same time
12:58:39  <tjfontaine>ki9a: btw txdv has some newer c# bindings https://github.com/txdv/LibuvSharp
12:59:58  <ki9a>tjfontaine: nice! shame i'm almost done with my own :(
13:00:35  <tjfontaine>ki9a: que sera sera, unfortunately I had just reading backlog
13:02:16  <saghul>indutny can you send an email to the mailing list?
13:02:28  <indutny>saghul: I'm not really good at writing
13:02:38  <indutny>and do we have mailing list for libuv?
13:02:48  <saghul>indutny yeah, we do
13:03:01  <indutny>piscisaureus: bnoordhuis: why I hadn't heard of it before? :)
13:03:02  <saghul>indutny I'll comment in the issue on github if you prefer
13:04:36  <piscisaureus>indutny: I think I sent you an invite
13:05:31  <piscisaureus>indutny: groups.google.com/group/libuv
13:07:10  <indutny>piscisaureus: ok, joined
13:07:42  <piscisaureus>let's add it to the topic
13:08:25  * piscisaureustopic: liberal utopian vacation ~ http://groups.google.com/group/libuv ~ http://piscisaureus.no.de/libuv/
13:11:35  <indutny>piscisaureus: writing post
13:11:50  <txdv>ki9a: your own?
13:12:07  <ki9a>txdv: Just the tcp parts.
13:12:30  <txdv>almost done ... with the tcp part ... muahahahaha
13:13:12  <txdv>I have to work on them though some more
13:13:18  <txdv>probably ill do some coding tonight
13:13:28  * `3rdEdenjoined
13:14:13  <txdv>ki9a: is the source online? btw I have been spamming every possible way about my bindings and manos and mono
13:14:19  <txdv>i cant believe you just ignored all that
13:14:36  <ki9a>txdv: I didn't notice it in the mono mailling list. I left manos behind ages ago.
13:14:49  <txdv>by the way
13:14:58  <txdv>I have CAresSharp running on my LibuvSharp
13:15:02  <txdv>so dns stuff is there
13:15:12  <ki9a>I'm getting fed up with mono's crappy implementation for socket.s
13:15:41  <txdv>if thats all you njeed, just fix the implementation
13:15:49  <ki9a>If i could I would.
13:15:51  <txdv>if you need an awesome event loop
13:18:03  <indutny>piscisaureus: huh, you need to approve topics? ;)
13:18:25  <piscisaureus>indutny: done
13:18:42  <piscisaureus>indutny: just like at nodejs, your first message will be moderated
13:18:52  <indutny>ah, first message
13:18:53  <indutny>ok
13:18:54  * `3rdEdenquit (Remote host closed the connection)
13:18:56  <indutny>saghul: https://groups.google.com/forum/?fromgroups=#!topic/libuv/qSSaBRBwFoY
13:18:57  <indutny>bnoordhuis: ^
13:19:04  * V1joined
13:19:11  <saghul>indutny thanks
13:19:35  * V1changed nick to `3rdEden
13:21:43  <txdv>piscisaureus: will anyone take a look at the pull request where I fix around the struct layout a bit?
13:24:27  * mmaleckiquit (Ping timeout: 268 seconds)
13:26:02  * mmaleckijoined
13:28:37  * TheJHquit (Quit: goodbye)
13:29:04  * piscisaureustopic: liberemus ut volat ~ http://groups.google.com/group/libuv ~ http://piscisaureus.no.de/libuv/
13:29:09  <piscisaureus>txdv: ok
13:29:39  <txdv>https://github.com/joyent/libuv/pull/555
13:29:42  <txdv>this one to be precise
13:29:43  * EhevuTovjoined
13:29:51  <txdv>the other one was another approach, turned out to be a bummer
13:31:16  <piscisaureus>txdv: looks good. you removed uv_req_private_fields but you never put it back anywhere. Is it not needed
13:31:25  <piscisaureus>?
13:31:30  <piscisaureus>txdv: oh nvm ...
13:31:37  <txdv>No, I left them
13:31:51  <txdv>I wanted to move all the private fields at the bottom of each struct, but that turned out to be a bummer
13:31:55  * mmaleckiquit (Ping timeout: 246 seconds)
13:32:45  <txdv>so I added the possibility to get the size of the generic 'classes' so one could jump over the offset
13:34:54  <txdv>I rearanged it so it would be possible to retrieve with at least some extra logic using ffi all the public fields
13:35:21  <txdv>dunno, if you have any questions, just leave a comment there
13:35:31  <indutny>txdv: interesting
13:36:31  <txdv>for example the uv_loop_t started with private fields, which were of dynamic size
13:36:48  <txdv>if you haven't got the c definitions, you are basically stranded to retrieve the public fields
13:47:53  * EhevuTovquit (Quit: This computer has gone to sleep)
13:48:55  * EhevuTovjoined
13:53:19  <piscisaureus>txdv: why did you move loop->data up?
13:54:01  <txdv>so the 2 engine x queues would be in the end
13:54:06  <txdv>these are 4 pointers
13:54:15  <txdv>that are not very important
13:55:18  <txdv>more convenience then necessity
13:57:01  <txdv>data, error, reference counting int, and then the 2 ngx queue
13:57:21  <txdv>data is public writable and readable
13:57:25  <txdv>so it goes first
13:57:33  <txdv>the rest is public but only readable
13:57:40  <txdv>so I thought this layout would make more sense
14:05:20  <piscisaureus>txdv: ok, looks good, landing
14:05:41  <txdv>yoohoo, close the other pull request then
14:14:48  * piscisaureus_joined
14:14:56  <piscisaureus_>bnoordhuis: https://github.com/torvalds/linux/blob/master/net/ipv4/inet_connection_sock.c#L236-278
14:20:27  <CIA-131>libuv: Bert Belder master * r46bd5fb / include/uv.h : Rearrange struct layouts to make FFI bindings easier - http://git.io/IGE2XA
14:20:42  <tjfontaine>keep up the good work cia
14:22:03  <txdv>github bug
14:22:08  <txdv>it says that it will be merged in an hour
14:22:13  <txdv>http://tinyurl.com/8v9bbjc
14:22:38  <ki9a>https://github.com/carlokok/LibuvSharp < looks even scarier
14:22:42  <ki9a>The backend storage is temporarily offline. Usually this means the storage server is undergoing maintenance. Please¬†contact support¬†if the problem persists.
14:22:44  * travis-cijoined
14:22:44  <travis-ci>[travis-ci] joyent/libuv#729 (master - 46bd5fb : Bert Belder): The build passed.
14:22:44  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/73cf3600d75a...46bd5fb3924c
14:22:44  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/2517527
14:22:44  * travis-cipart
14:36:14  * TheJHjoined
14:42:30  * `3rdEdenquit (Ping timeout: 264 seconds)
14:42:57  * AvianFlujoined
14:45:49  * `3rdEdenjoined
14:51:01  * c4milojoined
15:10:12  <ki9a>when using uv_tcp_open and then doing an uv_read_start I get "not connected" because handle->flags & UV_HANDLE_READING isn't true
15:10:18  <ki9a>what else do I need to call to take over a handle?
15:10:23  <ki9a>(uv_tcp_open succeeds)
15:13:45  * ericktjoined
15:14:00  * mmaleckijoined
15:15:43  * xaqquit (Remote host closed the connection)
15:16:01  * hzjoined
15:17:17  * mikealquit (Quit: Leaving.)
15:17:36  * paddybyersquit (Quit: paddybyers)
15:17:58  * hzpart
15:18:35  <piscisaureus_>ki9a: you should channel your complaints to saghul and bnoordhuis :-)
15:19:24  <saghul>ki9a uv_tcp_open is not supported for connected sockets
15:19:25  <ki9a>I wasnt' complaining!
15:19:27  <ki9a>:)
15:19:42  <piscisaureus_>ki9a: you should. and complain to saghul and bnoordhuis
15:19:53  <piscisaureus_>ki9a: what you want to do is not supported atm
15:19:59  <saghul>steal the fd, use uv_tcp_open and then uv_tcp_connect it
15:20:00  <ki9a>So I hear
15:20:29  <ki9a>the handle I have is already open (connected)
15:20:33  <ki9a>it came from accept from somewhere else
15:20:48  <saghul>then better use uv_poll_t
15:21:13  <saghul>actually, piscisaureus_ it does work on Unices, but don't tell anyone
15:21:22  <saghul>;-)
15:21:27  <piscisaureus>ki9a: are you using windows?
15:21:31  <ki9a>yes
15:22:07  <piscisaureus>aha
15:22:20  <ki9a>on windows a c fd does not match a socket handle afaik
15:22:21  * mikealjoined
15:22:21  <piscisaureus>ki9a: where did you get that socket from? Who created it for you?
15:22:29  <piscisaureus>ki9a: that's correct.
15:22:31  <ki9a>.net's socket
15:22:58  <ki9a>I have a fairly large chunk of infrastructure from earlier I was wondering if I could just take over the socket from that
15:23:15  <ki9a>but I can probably change that to use libev for listening too
15:23:54  <piscisaureus>ah, right, that's the problem
15:23:59  <piscisaureus>.net gives you a connected socket
15:24:09  <ki9a>yes
15:24:17  <piscisaureus>ya umm
15:24:23  <piscisaureus>it doesn't work atm ...
15:24:28  <ki9a>the call made me thing it worked
15:24:33  * mmaleckiquit (Ping timeout: 248 seconds)
15:24:33  <ki9a>it's not that important really
15:24:48  <piscisaureus>you could try to set socket->flags |= UV_HANDLE_READABLE|UV_HANDLE_WRITABLE|UV_HANDLE_BOUND or something
15:24:49  * dapjoined
15:25:02  <piscisaureus>but i think these flags are all private
15:25:10  <ki9a>they are
15:25:18  <ki9a>and iirc the "type" isn't set either
15:25:29  <ki9a>so it aborts in the loop because it has an unknown kind of event
15:25:39  <piscisaureus>hmmm
15:25:43  <piscisaureus>ah
15:25:56  <piscisaureus>oh shit you have to init the read req
15:26:00  <piscisaureus>yeah, don't go there...
15:26:35  <txdv>will uv_tcp_open support opened sockets in the future?
15:26:49  <bnoordhuis>txdv: definitely maybe
15:26:55  <piscisaureus>it could be done, it's not that hard
15:27:03  <piscisaureus>it's just a matter of someone caring enough
15:27:10  <piscisaureus>and hammering it out :-)
15:27:14  <txdv>don't force me to care enough
15:27:30  <txdv>I dont like windows :(
15:27:34  <txdv>but it works on unix, right?
15:27:42  <piscisaureus>well
15:27:43  <ki9a>the api itself doesn't work for this
15:27:53  <ki9a>probably not on unix either
15:28:03  <piscisaureus>saghul just said it does work
15:28:10  <piscisaureus>but i wonder if it sets the correct flags etc
15:28:12  <ki9a>hmm
15:28:22  <saghul>piscisaureus There are no flags
15:28:28  <saghul>for the connected state
15:28:34  <piscisaureus>saghul: ah, right
15:28:36  <saghul>so you can uv_read_start it and it will work
15:28:52  <piscisaureus>saghul: so listening sockets won't work but connected sockets do
15:29:03  <saghul>I notices this while reading the code, but it's "undocumented behavior" :-)
15:29:24  <saghul>piscisaureus I didn't check listening sockets
15:30:10  <piscisaureus>unix is so easy sometimes
15:31:35  * mikealquit (Quit: Leaving.)
15:32:00  * loladirojoined
15:36:54  * xaqjoined
15:40:14  * loladiroquit (Quit: loladiro)
15:41:33  * loladirojoined
15:43:50  * mikealjoined
15:45:09  * ericktquit (Quit: erickt)
15:47:04  * paddybyersjoined
15:47:42  * loladiroquit (Quit: loladiro)
15:50:22  <isaacs>good morning
15:50:36  <isaacs>standup in 0:10? /cc piscisaureus bnoordhuis indutny tootallnate
15:50:42  <indutny>oooh
15:50:44  <indutny>standup
15:52:04  <isaacs>call
15:52:09  <isaacs>you can sit down for it if you want :)
15:52:42  * TooTallNatejoined
15:52:55  <indutny>isaacs: har har har
15:56:12  <bnoordhuis>isaacs: yep
15:56:23  <piscisaureus_>isaacs: 5 minutes
15:56:46  <indutny>10 minutes :)
15:56:53  <indutny>who gives more
15:57:36  <mitsuhiko>god. who designed websockets. seriously
15:58:44  * loladirojoined
16:00:08  <txdv>saghul: make uv_tcp_open useable on open sockets on windows plz!
16:01:12  <saghul>txdv I have no idea how to do that :-S is there a way to detect if a socket is connected or listening?
16:01:26  <txdv>someone just said that it is possible
16:01:42  <txdv>[18:20:41] <piscisaureus> it could be done, it's not that hard
16:01:42  <txdv>[18:20:48] <piscisaureus> it's just a matter of someone caring enough
16:03:22  <TooTallNate>call!
16:03:23  <saghul>txdv nice to know!
16:03:41  <txdv>you have to care more!
16:03:45  * saghuldoesn't know much about windows
16:04:07  <indutny>call
16:04:16  <indutny>isaacs: ?
16:04:17  <txdv>me neither
16:04:18  <indutny>piscisaureus_: ?
16:04:20  <indutny>bnoordhuis: ?
16:06:12  * mikealquit (Quit: Leaving.)
16:06:54  <ki9a>i don't tink you can find out if it's listening vs stream (socket)
16:11:22  * loladiroquit (Quit: loladiro)
16:15:53  <bnoordhuis>ki9a: set O_NONBLOCK and call accept(). if you get EINVAL, it's not a listen socket
16:16:30  <ki9a>ah
16:16:40  * loladirojoined
16:16:42  <ki9a>would tcp_opn ever really be used for listening sockets?
16:17:36  <bnoordhuis>ki9a: i don't see why not
16:19:34  * loladiroquit (Client Quit)
16:20:00  * stagasjoined
16:26:31  * `3rdEdenquit (Quit: gbai)
16:34:43  * loladirojoined
16:55:15  * paddybyersquit (Ping timeout: 240 seconds)
16:57:00  <indutny>bnoordhuis: piscisaureus: so about embedding
16:57:08  <indutny>we've all read that email
16:57:10  <indutny>and seen that patch
16:57:12  <indutny>lets decide something
16:57:19  <indutny>;)
17:00:44  * xaqquit (Remote host closed the connection)
17:02:22  * loladiroquit (Quit: loladiro)
17:06:41  * loladirojoined
17:09:05  * xaqjoined
17:16:27  * loladiroquit (Quit: loladiro)
17:23:15  * EhevuTovquit (Ping timeout: 240 seconds)
17:25:04  * EhevuTovjoined
17:36:20  * loladirojoined
17:36:56  * loladiroquit (Client Quit)
17:42:02  * loladirojoined
17:42:36  * loladiroquit (Client Quit)
17:43:02  * loladirojoined
17:43:02  * loladiroquit (Client Quit)
17:54:48  * lohkeyjoined
17:58:16  * loladirojoined
17:59:27  * loladiroquit (Client Quit)
18:02:49  * xaqquit (Remote host closed the connection)
18:02:52  * brsonjoined
18:15:03  * loladirojoined
18:19:42  * loladiroquit (Ping timeout: 264 seconds)
18:24:22  * bnoordhuisquit (Ping timeout: 245 seconds)
18:25:16  * loladirojoined
18:25:43  * loladiroquit (Client Quit)
18:25:45  * dapquit (Quit: Leaving.)
18:27:56  * brson_joined
18:30:53  * brsonquit (Ping timeout: 260 seconds)
18:43:10  * ki9aquit (Read error: Connection reset by peer)
18:45:37  * xaqjoined
18:45:56  * stagas_joined
18:47:00  * stagasquit (Ping timeout: 246 seconds)
18:47:05  * stagas_changed nick to stagas
18:54:24  * paddybyersjoined
19:03:37  * EhevuTovquit (Quit: This computer has gone to sleep)
19:03:56  * dapjoined
19:30:40  * lohkeyquit (Quit: lohkey)
19:34:02  * CoverSlidepart
19:44:40  * joshthecoderjoined
19:55:08  * xaqquit (Remote host closed the connection)
20:09:11  * xaqjoined
20:15:04  * isaacs_mobilejoined
20:26:04  * lohkeyjoined
20:34:32  * isaacs_mobilequit (Remote host closed the connection)
20:44:54  * CoverSlidejoined
20:46:47  * loladirojoined
20:54:37  * philips_quit (Excess Flood)
20:56:02  * philips_joined
21:02:27  * paddybyersquit (Read error: Connection reset by peer)
21:02:46  * paddybyersjoined
21:19:14  * bulatshakirzyanojoined
21:27:10  <saghul>indutny after thinking about it, my proposal wouldn't work in the scenario you need
21:40:34  * loladiroquit (Quit: loladiro)
21:42:00  * hzjoined
21:47:02  * hzquit (Ping timeout: 268 seconds)
21:51:28  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
21:52:08  <CIA-131>libuv: Bert Belder master * r45931f8 / README.md : Add mailing list to readme - http://git.io/l7DiKw
21:53:55  * bulatshakirzyanojoined
21:56:31  <piscisaureus>'lo
22:00:16  * xaqquit (Remote host closed the connection)
22:03:32  * kristatequit (Ping timeout: 240 seconds)
22:05:11  * loladirojoined
22:06:42  * bulatshakirzyanoquit (Quit: Computer has gone to sleep.)
22:15:21  * loladiroquit (Quit: loladiro)
22:27:44  * loladirojoined
22:29:49  * loladiroquit (Client Quit)
22:33:22  * kuebkjoined
22:54:15  * TheJHquit (Ping timeout: 268 seconds)
22:55:44  <CIA-131>libuv: Luigi Grilli master * r5bfb7c9 / (common.gypi gyp_uv uv.gyp vcbuild.bat): windows: build system improvements - http://git.io/z7e9Ug
23:00:14  * stagasquit (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
23:06:16  * piscisaureus_quit (Ping timeout: 246 seconds)
23:06:41  * piscisaureusquit (Ping timeout: 260 seconds)
23:12:11  * piscisaureus_joined
23:12:57  <piscisaureus_>UPC is doing unannounced nighly maintenance again
23:16:04  * kuebkquit
23:16:36  * loladirojoined
23:18:30  <tjfontaine>$1 = <value temporarily unavailable, due to optimizations>
23:18:49  <tjfontaine>such useless are people who when --enable-debug is passed they still build with optimizations.
23:19:33  <piscisaureus_>isaacs: is bryan around?
23:19:43  <isaacs>piscisaureus_: don't know
23:19:47  <isaacs>one sec, lemme see
23:19:49  <piscisaureus_>oh
23:20:00  <isaacs>seems to be offline
23:20:02  <isaacs>(on jabber)
23:20:07  <piscisaureus_>isaacs: ask him if he sent the postmortem patch upstream and/or if he wants me to do it
23:20:20  <isaacs>ircretary: tell bcantrill <@piscisaureus_> isaacs: ask him if he sent the postmortem patch upstream and/or if he wants me to do it
23:20:20  <ircretary>isaacs: I'll be sure to tell bcantrill
23:20:39  <piscisaureus_>isaacs: I don't see bcantrill around here often
23:20:43  <isaacs>no, indeed :)
23:20:52  <piscisaureus_>so ircretary may not help
23:21:07  <isaacs>i'd say, go ahead and send it upstream.
23:21:14  <isaacs>if it's not a lot of trouble to do so
23:21:21  <piscisaureus_>it's not
23:27:46  * loladiroquit (Quit: loladiro)
23:28:26  * rendarquit
23:30:08  * loladirojoined
23:44:52  * loladiroquit (Quit: loladiro)
23:46:03  * xaqjoined
23:54:37  <CIA-131>node: Bryan Cantrill master * rcc1b09d / (3 files): test: add tests for postmortem and DTrace support - http://git.io/hAsG6Q
23:55:44  <CIA-131>node: isaacs v8-3.13.7.1 * r3411a03 / (381 files in 23 dirs): V8: Upgrade to 3.13.7.1 - http://git.io/gBGPIQ
23:55:44  <CIA-131>node: isaacs v8-3.13.7.1 * rd77c24d / (deps/v8/build/common.gypi deps/v8/tools/gyp/v8.gyp): V8: reapply floating patches - http://git.io/ntaXCg
23:55:45  <CIA-131>node: Bryan Cantrill v8-3.13.7.1 * rbd0e45b / deps/v8/src/platform-posix.cc : Loosen artificial mmap constraint - http://git.io/rCCVWg
23:55:45  <CIA-131>node: Bryan Cantrill v8-3.13.7.1 * r2796068 / deps/v8/tools/gen-postmortem-metadata.py : v8: fix postmortem metadata generation - http://git.io/yMtkKg
23:56:15  * brson_quit (Ping timeout: 240 seconds)
23:58:30  * brsonjoined