00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:07  * ircretaryjoined
00:01:08  <isaacs>trevnorris: orly?
00:01:10  <isaacs>trevnorris: howso?
00:01:23  * isaacsexcited.
00:01:39  <trevnorris>just a few optimizations in startup.processNextTick area
00:01:57  * jmar777joined
00:02:04  <trevnorris>dropping splice, simplifying the check logic. small stuff.
00:03:26  * bnoordhu1sjoined
00:05:49  <isaacs>awesome.
00:05:54  <isaacs>pull req that shiz!~
00:06:00  <isaacs>fa sersly
00:06:19  <trevnorris>can I include your branch changes? it's built on top of that.
00:07:48  * bnoordhu1squit (Ping timeout: 248 seconds)
00:12:42  * mikealquit (Quit: Leaving.)
00:14:15  <TooTallNate>isaacs: is v0.9.9 still today?
00:14:21  <isaacs>TooTallNate: maybe
00:14:29  <isaacs>trevnorris: of course.
00:14:47  <isaacs>TooTallNate: might be an evening release.
00:14:56  <isaacs>trying to work out what actually changed in v0.8's node_zlib.cc
00:15:21  <TooTallNate>what's the problem?
00:15:26  <TooTallNate>with zlib?
00:15:35  <isaacs>well... it looks like bnoordhuis's changes sorta rename "close" to "clear"
00:15:43  <isaacs>but on top of v0.8
00:15:53  <isaacs>and it's sort of half there already in v0.9, but with a different name
00:15:59  <isaacs>so it's just a tricky merge, that's all
00:17:44  <mmalecki>isaacs: thanks for merging that pull request, it means a lot to me :)
00:18:04  <isaacs>mmalecki: no problem, thanks for fixing the bug :)
00:18:36  * mikealjoined
00:19:54  <TooTallNate>isaacs: api change in v0.8?
00:19:58  <TooTallNate>of what that internal?
00:20:02  <TooTallNate>s/of/or
00:20:07  <TooTallNate>s/what/was
00:20:15  <bnoordhuis>saghul: connect() never blocks with udp sockets, it just sets a default dst address
00:20:32  <saghul>bnoordhuis, ok, as I suspected then, thanks!
00:20:52  <bnoordhuis>isaacs: i back-ported two patches from master
00:21:19  <bnoordhuis>hm, i forgot to mention that in the commit logs
00:21:24  <isaacs>bnoordhuis: oh, ok
00:21:35  <isaacs>yeah, when i renamed Clear to Close, the diffs were all rather minimal
00:22:31  <isaacs>but the rename confused git
00:22:59  <isaacs>making test before pushing v0.8-merge
00:28:05  <indutny>TooTallNate: interesting
00:28:07  <bnoordhuis>god, twitter...
00:28:19  <bnoordhuis>each read tweet costs you 1 iq point
00:28:30  <TooTallNate>hahaha
00:28:36  <TooTallNate>bnoordhuis: are you even on twitter?
00:28:45  <bnoordhuis>thank god no
00:28:52  <TooTallNate>didn't think so :p
00:28:53  <indutny>TooTallNate: I guess its something that have changed
00:29:00  <indutny>TooTallNate: I'm not sure if its a tls bug, really
00:29:13  <TooTallNate>indutny: well we'll see what einaros comes up with
00:29:16  <indutny>but I'll look tomorrow
00:29:18  <indutny>yeah
00:29:30  <TooTallNate>i don't really wanna look into it
00:29:30  <indutny>you know why I'm telling this?
00:29:37  <indutny>because we've this kind of connections in tests
00:29:42  <indutny>like one long running connection
00:29:47  <indutny>and it seems to be working properly
00:29:57  <TooTallNate>ya i mean i don't have a standalone tls test case
00:29:57  <indutny>so either tests are wrong :)
00:29:58  <CoverSlide>https://twitter.com/bnoordhuis
00:30:01  <TooTallNate>so it could be ws
00:30:02  <MI6>joyent/node: isaacs master * 5cc3569 : Merge remote-tracking branch 'ry/v0.8' into master Conflicts: AUTHORS (+19 more commits) - http://git.io/6cnlug
00:30:06  * paddybyersquit (Ping timeout: 264 seconds)
00:30:09  <indutny>TooTallNate: yes, that's a good thing :)
00:30:13  <indutny>TooTallNate: thank you for testing
00:30:15  <CoverSlide>0 tweets :(
00:30:17  <isaacs>bnoordhuis: i'm not going to merge that fs patch into master just yet.
00:30:21  <isaacs>bnoordhuis: thanks for the +1, though
00:30:34  <isaacs>bnoordhuis: i want 0.9.10 to be HOLY MOLY OMG ALL OF A SUDDEN MASTER IS FAST AGAIN!
00:30:44  <isaacs>bnoordhuis: every perf improvement is going in there :)
00:30:48  <isaacs>it's going to be SHOCKING!
00:30:54  <bnoordhuis>haha, color me curious
00:30:54  <isaacs>people will be SHOCKED
00:31:12  <indutny>:)
00:31:21  <indutny>so have you released 0.9.9?
00:31:25  <isaacs>not yet
00:31:27  <isaacs>workin on it :)
00:31:46  <indutny>ok :)
00:32:01  <indutny>I'm not sure about tls thing, considering TooTallNate's issue with ws
00:32:13  <indutny>and I'm definitely not willing to investigate what's happening in it right now
00:32:20  <indutny>only 7-8 hours later
00:32:28  <indutny>isaacs: so if you feel like we better revert it
00:32:32  <indutny>I'll probably be up for it
00:32:54  <indutny>other choice would be if you'll look into it yourself :)
00:32:59  <TooTallNate>indutny: if the tests are passing for tls then we probably shouldn't revert it
00:33:06  <indutny>it might be some straightforward problem
00:33:08  <TooTallNate>indutny: but hopefully we'll have more information about it tomorrow
00:33:08  <isaacs>hmm.
00:33:09  <isaacs>ok
00:33:12  <indutny>with compatibility layer or something like this
00:33:14  <isaacs>let's put off 0.9.9 until tomorrow
00:33:21  <indutny>isaacs: that's third choice too
00:33:25  <isaacs>that's good.
00:33:30  <isaacs>that means i can not be late to go drink with mikeal and oleg.
00:33:31  <indutny>ok
00:33:32  <indutny>ttyl guys
00:33:35  * indutny&
00:33:35  <LOUDBOT>I'M NOT LEARNING CANTONESE ANYTIME SOON SO WHY DO I NEED IT
00:33:38  <isaacs>g'nite, indutny
00:33:50  <isaacs>TooTallNate: what's the issue? got a link to a gist or something?
00:34:05  <TooTallNate>isaacs: https://github.com/einaros/ws/issues/159
00:34:26  <isaacs>TooTallNate: i love how you post bugs for versions of node that don't exist yet :)
00:34:26  * perezdquit (Quit: perezd)
00:34:32  <TooTallNate>hahaha
00:34:35  <TooTallNate>bleeding-edge
00:34:37  <TooTallNate>and beyond
00:34:54  <TooTallNate>i guess v0.9.9-pre would be more accurate
00:35:30  <isaacs>ounces of prevention, pounds of cure.
00:35:32  <isaacs>you know how it goes
00:36:29  * perezdjoined
00:38:57  * karupaneruraquit (Excess Flood)
00:40:40  * karupanerurajoined
00:43:57  * papertigersjoined
00:46:26  * mikealquit (Quit: Leaving.)
00:48:02  <isaacs>ok, i'm out, also
00:48:13  <trevnorris>isaacs: one sec.
00:48:34  <trevnorris>just tested your changes off latest master and "test/message/stdin_messages.js" fails.
00:48:45  <trevnorris>were there some test changes you made that I don't see?
00:49:16  * bnoordhu1sjoined
00:50:09  <trevnorris>isaacs: wait... think I have a rouge change. nm for now.
00:51:24  <CoverSlide>rouge change? you're putting on blush?
00:51:25  * pooyajoined
00:52:03  <trevnorris>CoverSlide: no. i just have so many stashs/branches i'm switching between that I get lost sometimes.
00:52:29  <trevnorris>like, I use my benchmarks from my PR that haven't been merged yet. so I always do a `git checkout benchmark-refactor -- benchmar/`
00:52:31  <trevnorris>so I can run them
00:52:46  * mikealjoined
00:52:48  <trevnorris>but then I have to `git reset --hard` to revert them before I switch branches.
00:53:36  * bnoordhu1squit (Ping timeout: 244 seconds)
00:55:06  * mikealquit (Client Quit)
00:55:19  <papertigers>Hey guys, we are seeing a node process randomly grow in rss. The process with grow its heap to a large size and then eventually scale back leaving a a large free slab. ::vmem will report something like 78Mb used out of 750mb
00:55:46  <papertigers>I managed to catch a process in this growth phase and have captured some user stacks that may be useful/meaningful to someone here
00:56:36  <papertigers>https://gist.github.com/papertigers/e259e9d65e3f9d6aa6db
00:56:40  <trevnorris>papertigers: can you check if it's an issue similar to GH-4525
00:57:33  <papertigers>trevnorris: looking
00:57:40  <indutny>isaacs: I guess I know what is causing ws to fail
00:57:44  <indutny>papertigers: looking
01:01:22  <TooTallNate>indutny: that was fast :p
01:01:38  <indutny>TooTallNate: and probably wrong :)
01:01:48  <TooTallNate>indutny: which part?
01:01:50  <indutny>from debugging output
01:01:59  <indutny>it seems that socket emits 'end' event
01:02:03  <indutny>like its closed by other side
01:02:43  <indutny>yes
01:02:44  <indutny>EOF
01:02:45  <indutny>in net.js
01:03:25  <indutny>TooTallNate: so it seems like server dislikes something from what you send to it :)
01:04:09  <TooTallNate>indutny: i mean it's just an echo server i'm connecting to :p
01:04:14  <indutny>well
01:04:27  <indutny>idk
01:04:34  <TooTallNate>hahah
01:04:35  <TooTallNate>ok
01:04:58  <indutny>HTTP/1.1 101 Web Socket Protocol Handshake
01:04:58  <indutny>Upgrade: WebSocket
01:04:58  <indutny>Connection: Upgrade
01:04:58  <indutny>Sec-WebSocket-Accept: NbP1bDOunociMXUUSx26rz+Xy78=
01:04:58  <indutny>Server: Kaazing Gateway
01:04:58  <indutny>Date: Thu, 07 Feb 2013 00:56:25 GMT
01:05:07  <indutny>that's the only thing we receive from the socket
01:05:10  <indutny>before it emits EOF
01:05:35  <TooTallNate>indutny: i have a hunch that data being written isn't actually being written
01:05:44  <indutny>what data?
01:05:47  <TooTallNate>indutny: so the server closes the socket since no handshake was sent
01:05:50  <indutny>from ws to tls?
01:05:53  <TooTallNate>indutny: websocket protocol data
01:05:55  <TooTallNate>ya
01:05:57  <indutny>yes, it might be true
01:06:00  <indutny>let me check logs
01:06:51  <indutny>doesn't seem like so
01:07:13  <indutny>it would be simplier if I had SSL private key for that server
01:07:13  <indutny>:)
01:07:18  <indutny>can I setup a local one somehow?
01:09:40  <indutny>TooTallNate: ^
01:09:56  <TooTallNate>indutny: ya let me try to make a gist
01:10:37  <indutny>may be handshake it writes is invalid?
01:10:42  <indutny>because I see its being written
01:10:43  <indutny>anyway
01:10:47  <indutny>I need to compare wireshark data
01:10:53  <indutny>and I need it to run on my own key for that
01:11:12  <TooTallNate>ya i understand
01:11:19  <TooTallNate>einaros doesn't seem to have any good server examples
01:11:21  <TooTallNate>except for https://github.com/einaros/ws#server-example
01:11:39  <TooTallNate>i guess you have to attach it to an https server
01:11:46  <indutny>oooh
01:11:49  <indutny>too hard for 5am
01:11:51  <indutny><Buffer fd fd fd fd 51 fd fd fd 3d b6 fd 26 fd fd fd 35>
01:11:57  <indutny>if you're into wss thing
01:12:04  <indutny>does it seems to be a valid handshake?
01:12:12  <TooTallNate>no clue :p
01:12:17  <TooTallNate>we'll have to ask einaros
01:12:28  <TooTallNate>indutny: get sleep man :D
01:12:35  <indutny>oook
01:12:36  <indutny>ttyl
01:12:37  * indutny&
01:12:37  <LOUDBOT>IS THAT SERIOUSLY THE ONLY THING YOU HAVE EVER SAID
01:15:03  * EhevuTovquit (Quit: This computer has gone to sleep)
01:20:11  * lohkeypart
01:20:33  * lohkey_joined
01:20:33  * lohkey_quit (Client Quit)
01:25:18  * mikealjoined
01:27:22  * trevnorrisquit (Quit: Leaving)
01:35:50  * CoverSlide&
01:35:51  <LOUDBOT>BONDAGE PORN MEANS THAT YOU'RE A HIGH ORDER SEXUAL PREDATORY SERIAL KILLER ACCORDING TO THIS SHOW
01:36:43  * abraxasjoined
01:39:56  * c4miloquit (Remote host closed the connection)
01:44:45  * sblomquit (Ping timeout: 244 seconds)
01:49:22  * cjdjoined
01:50:07  <cjd>for those who were keeping track, false alarm with the possible apple bug I mentioned earlier, the uv_inref() trick works perfectly, the issue was something else entirely
01:50:14  <cjd>*unref
01:54:58  * jmar777quit (Remote host closed the connection)
01:57:24  * pooyaquit (Quit: pooya)
01:58:22  * AvianFluquit (Remote host closed the connection)
01:58:49  * jmar777joined
02:10:27  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
02:17:52  * mmaleckichanged nick to mmalecki[zzz]
02:18:43  * pooyajoined
02:19:03  * c4milojoined
02:19:05  * pooyaquit (Client Quit)
02:27:10  * TooTallNatequit (Quit: Computer has gone to sleep.)
02:30:03  * kazuponjoined
02:39:18  * brsonquit (Quit: leaving)
02:43:30  * kazuponquit (Remote host closed the connection)
02:48:10  * dapquit (Quit: Leaving.)
02:48:58  * garbagecollectioquit (Quit: garbagecollectio)
02:50:46  * garbagecollectiojoined
02:53:55  * jmar777quit (Remote host closed the connection)
03:04:11  * bnoordhuisquit (Ping timeout: 245 seconds)
03:07:22  * jmar777joined
03:10:21  * jmar777quit (Remote host closed the connection)
03:34:09  * shigekijoined
03:34:18  * c4miloquit (Remote host closed the connection)
03:37:01  <shigeki>Hi, is Bert here?
03:37:25  <tjfontaine>he left about an hour ago, but that doesn't mean he won't be back
03:37:48  <shigeki>Oh, I see. Thanks.
03:38:27  <tjfontaine>you can leave him a message by saying something like "ircretary: tell piscisaureus_ ..."
03:38:43  <tjfontaine>and when he gets back he'll see the message and respond
03:39:01  <shigeki>Okay, I will.
03:41:21  * lohkeyjoined
03:42:58  <shigeki>ircretary: tell piscisaureus_ Thanks. Comment on #701. I had a mistake. Do you have any idea to compare the same timeout handles rather than using loop->timer_counter?
03:42:59  <ircretary>shigeki: I'll be sure to tell piscisaureus_
03:43:54  * kazuponjoined
03:49:39  * kazuponquit (Ping timeout: 256 seconds)
03:57:38  * loladiroquit (Quit: loladiro)
04:05:21  * paddybyersjoined
04:09:28  * loladirojoined
04:15:02  * TooTallNatejoined
04:17:03  * garbagecollectioquit (Quit: garbagecollectio)
04:17:21  * TooTallNatequit (Client Quit)
04:17:31  * paddybyersquit (Ping timeout: 245 seconds)
04:23:37  * piscisaureus_joined
04:24:39  * piscisaureus_quit (Client Quit)
04:29:56  * piscisaureus_joined
04:32:12  * piscisaureus_quit (Client Quit)
04:33:19  * piscisaureus_joined
04:33:25  * piscisaureus_quit (Client Quit)
04:45:17  * kazuponjoined
04:47:57  * AvianFlujoined
04:49:41  * piscisaureus_joined
04:50:39  * kazuponquit (Ping timeout: 252 seconds)
04:54:49  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
05:05:57  * lohkeyquit (Quit: lohkey)
05:13:51  * kazuponjoined
05:21:36  * piscisaureus_joined
05:21:42  * kazuponquit (Remote host closed the connection)
05:21:48  * piscisaureus_quit (Client Quit)
05:22:39  * c4milojoined
05:23:32  * c4miloquit (Remote host closed the connection)
05:40:58  <shigeki>tell piscisaureus_ I come across if we really need to compare timer handles of which timeouts are equal. Is return a->timeout < b->timeout ? -1: 1; wrong solution? Please review https://github.com/shigeki/libuv/compare/shigeki_20130207_2
05:41:46  * perezd_joined
05:41:55  * perezdquit (Read error: Connection reset by peer)
05:42:01  * perezd_quit (Read error: Connection reset by peer)
05:42:11  * perezdjoined
05:52:26  * TheJHjoined
05:56:55  * piscisaureus_joined
05:59:15  * loladiroquit (Quit: loladiro)
05:59:15  * piscisaureus_quit (Client Quit)
06:07:57  * TheJHquit (Ping timeout: 240 seconds)
06:19:44  * piscisaureus_joined
06:28:36  * wolfeidauquit (Remote host closed the connection)
06:29:03  * mikealquit (Quit: Leaving.)
06:46:23  * mikealjoined
06:50:38  * piscisaureus_quit (Read error: Connection reset by peer)
06:51:16  * piscisaureus_joined
06:51:25  * piscisaureus_quit (Client Quit)
06:56:53  * piscisaureus_joined
06:57:38  * piscisaureus_quit (Client Quit)
06:58:41  * piscisaureus_joined
06:58:49  * piscisaureus_quit (Client Quit)
07:33:37  * toothrotquit (Ping timeout: 246 seconds)
07:37:44  * toothrjoined
07:41:14  * `3rdEdenjoined
07:42:54  * paddybyersjoined
07:52:35  * Raltquit (Ping timeout: 248 seconds)
07:56:26  * rendarjoined
08:30:11  * shigekipart
08:31:16  * hzjoined
09:01:56  * mmalecki[zzz]changed nick to mmalecki
09:12:21  * roxlu_quit (Quit: leaving)
09:44:34  <saghul>txdv yt?
09:51:25  * Raltjoined
10:10:00  * wolfeidaujoined
10:43:00  <indutny>isaacs: yt? :)
10:43:01  <indutny>morning
11:09:55  * `3rdEdenchanged nick to `3E|LUNCH
11:18:48  * abraxasquit (Remote host closed the connection)
11:39:35  * `3E|LUNCHchanged nick to `3rdEden
11:51:42  * bnoordhuisjoined
12:17:28  <bnoordhuis>indutny: is https://github.com/joyent/node/issues/4727 a bug report or ?
12:17:38  <indutny>RFC
12:17:47  <indutny>basically
12:17:53  <indutny>I told it in a wrong way
12:17:55  <indutny>the problem is
12:18:05  <indutny>that socket.write(buffer, 'binary') works
12:18:10  <indutny>but stream.write(buffer, 'binary') not
12:18:16  <indutny>after streams2
12:18:19  <bnoordhuis>ah
12:18:27  <bnoordhuis>you might want to update the issue to clarify that
12:18:45  <indutny>yes
12:18:50  <indutny>posted a comment there
12:19:55  <indutny>which is odd, btw
12:20:04  <bnoordhuis>that you posted a comment?
12:20:16  <indutny>erm
12:20:21  <indutny>thoughts and messages
12:20:22  <indutny>all mixed
12:20:47  <indutny>basically, I don't fully get it
12:20:56  <indutny>this should not work with sockets
12:21:39  <indutny>aah
12:21:44  * c4milojoined
12:21:49  <indutny>socket has decodeStrings = false
12:21:57  <indutny>yep, this explains it
12:23:54  <indutny>but its a bit retarded to do .write(buffer, 'binary')
12:23:55  <indutny>you know
12:24:07  <indutny>so the good news is that tls seems to be working properly after my patch
12:24:17  <indutny>but some underlying semantics have changed a bit
12:25:35  <indutny>btw
12:25:41  <indutny>why ain't we using pgp for release tags? :)
12:26:55  <bnoordhuis>ask isaacs, i believe i've suggested it in the past
12:27:26  <indutny>yep, will do later
12:28:14  <indutny>what we've in priorities now?
12:28:18  <indutny>anything except performance?
12:30:27  * loladirojoined
12:31:22  <bnoordhuis>indutny: there's a ton of pull requests that need reviewing
12:39:58  * sgallaghjoined
12:43:29  * loladiroquit (Quit: loladiro)
12:44:55  <bnoordhuis>indutny: you're the expert on all things streaming right?
12:45:03  <bnoordhuis>https://gist.github.com/bnoordhuis/f45ceb227f190331a5c1 <- why does that work with master but not v0.8
12:45:25  <bnoordhuis>it's a simple passthrough stream but Faux.prototype.write never gets called in v0.8
12:45:29  <bnoordhuis>i must be missing something obvious...
12:46:33  <indutny>one sec
12:49:24  <indutny>aah
12:49:26  <indutny>this.writable = true
12:49:28  <indutny>this.readable = true
12:49:41  <indutny>bnoordhuis: ^
12:49:44  <indutny>bnoordhuis: 150$
12:49:50  <bnoordhuis>haha
12:49:56  <indutny>129$
12:50:05  <indutny>per incident
12:50:42  <bnoordhuis>thanks fedor, indeed obvious in hindsight
12:50:50  <bnoordhuis>curious it works with master though
12:51:02  <bnoordhuis>stream.Stream is a duplex stream now?
12:52:31  <indutny>yes
12:52:39  <indutny>and it sets .writable .readable to true
12:52:41  <indutny>by default
12:52:44  <indutny>now
12:52:49  <indutny>bnoordhuis: you're welcome
13:00:27  * garbagecollectiojoined
13:00:57  * loladirojoined
13:12:01  * toothrchanged nick to toothrot
13:16:58  * loladiroquit (Quit: loladiro)
13:28:22  <indutny>bnoordhuis: what do you think about this?
13:28:23  <indutny>https://github.com/joyent/node/pull/4709/files
13:28:27  <indutny>comment is pretty offensive
13:28:27  * piscisaureus_joined
13:29:30  <bnoordhuis>indutny: commented
13:29:50  <indutny>interestingly enough
13:29:54  <indutny>he has created another pull request
13:29:57  <indutny>after that one
13:29:58  <indutny>https://github.com/joyent/node/pull/4710/files
13:30:29  <bnoordhuis>ah, that's the guy i had to tell 3x to add tests and documentation
13:30:39  <indutny>:)
13:30:53  <bnoordhuis>i'll admit that by the third time i was rather curt
13:32:32  <bnoordhuis>now for something completely different
13:32:34  <bnoordhuis> var a = new Uint8Array(2);
13:32:34  <bnoordhuis> var b = new Uint16Array(a);
13:32:34  <bnoordhuis> assert.equal(a.length, 2);
13:32:34  <bnoordhuis> assert.equal(b.length, 2);
13:32:44  <bnoordhuis>^ passes in ff and chrome, not in node
13:33:05  <bnoordhuis>but why the hell is b.length 2?
13:33:39  <indutny>:)
13:33:48  <indutny>yeah, it's very odd
13:33:57  <indutny>let me test it :)
13:34:05  <indutny>I don't beleive you
13:34:10  <bnoordhuis>i guess we'll follow the browsers here but it doesn't make sense to me
13:34:55  <indutny>I know why it is like this
13:34:59  <bnoordhuis>indutny: https://gist.github.com/bnoordhuis/44994f7a3f366b0dceaa
13:35:02  <indutny>it converts it's argument to array
13:35:32  <bnoordhuis>i don't think that's what happens, see the full test case
13:35:46  <bnoordhuis> var a = new Uint8Array(2);
13:35:46  <bnoordhuis> var b = new Uint16Array(a);
13:35:46  <bnoordhuis> a[0] = 0;
13:35:46  <bnoordhuis> a[1] = 0;
13:35:46  <bnoordhuis> b[0] = 257;
13:35:48  <bnoordhuis> assert.equal(a[0], 0);
13:35:51  <bnoordhuis> assert.equal(a[1], 0);
13:35:54  <bnoordhuis> assert.equal(b[0], 257);
13:35:56  <bnoordhuis> assert.notEqual(a, b.buffer);
13:35:58  <bnoordhuis>^ for posterity
13:36:21  <indutny>I don't get it
13:36:27  <indutny>what are you trying to say?
13:36:43  <bnoordhuis>that the two arrays don't share a backing store
13:36:58  <indutny>should they?
13:37:06  <indutny>ah
13:37:07  <bnoordhuis>the spec leaves that open :)
13:37:14  <indutny>believe me
13:37:17  <bnoordhuis>but a.length == b.length even though they're different types
13:37:19  <indutny>it converts argument to array
13:37:23  <indutny>if it isn't
13:37:23  <indutny>:)
13:38:20  <indutny>searching chromium sources for more information
13:38:30  <MI6>joyent/node: Bert Belder v0.8 * 53282ec : http: protect against response splitting attacks This patch is a back-po - http://git.io/KNsFbg
13:38:39  <indutny>piscisaureus_: nice job, bro :)
13:39:12  <piscisaureus_>indutny: :-p it's practically just a cherry-pick. I am not sure why wpreul made all these weird commits ..
13:39:20  <indutny>yes, I agree
13:39:27  <indutny>just wanted to bump you a little :)
13:39:44  <bnoordhuis>well, i guess the typed array -> array thing indeed happens
13:40:10  <bnoordhuis>who are the people who think up these standards?
13:40:20  <MI6>joyent/node: Bert Belder v0.8 * 255bc94 : http: protect against response splitting attacks This patch is a back-po - http://git.io/AbOmZQ
13:40:27  <indutny>:)
13:40:30  <indutny>mainacs
13:40:36  <indutny>piscisaureus_: force push?
13:40:36  * garbagecollectioquit (Ping timeout: 248 seconds)
13:40:42  <indutny>maniacs*
13:40:53  <piscisaureus_>indutny: yes
13:40:56  <indutny>why?
13:41:04  <piscisaureus_>indutny: because I am a maniac
13:41:19  <indutny>I don't see any difference
13:41:19  <bnoordhuis>it's true
13:41:23  <bnoordhuis>you should see him laugh
13:41:32  <indutny>he's a CEO
13:41:33  <bnoordhuis>dr. evil can't hold a candle to it
13:41:38  <indutny>of course he's a maniac
13:41:46  <indutny>piscisaureus_: ;)
13:41:53  * loladirojoined
13:43:52  * garbagecollectiojoined
13:46:32  <indutny>shit, I can't find any links to typed array in chrome's sources
13:49:56  <piscisaureus_>npmjs.org down?
13:52:49  <indutny>nope
13:53:02  <indutny>why are you asking?
13:55:12  <indutny>bnoordhuis: it seems that they ain't converting to arrays literally
13:55:19  <indutny>but doing some per-index copying
13:55:27  <indutny>like in for loop
14:03:01  * garbagecollectioquit (Quit: garbagecollectio)
14:06:44  <indutny>bnoordhuis: can you sign off this, please https://github.com/joyent/node/pull/4628 ?
14:07:03  <indutny>I'm going to push it to master if you've nothing to say about it :)
14:07:05  <indutny>in next 1 minute
14:08:02  <indutny>ok, landing...
14:08:19  <MI6>joyent/node: Nikhil Marathe master * 6bcd9a4 : Update http docs with http.IncomingMessage. http.ServerRequest and http. - http://git.io/KasFfQ
14:25:00  * garbagecollectiojoined
14:35:30  * indexzeroquit (Quit: indexzero)
14:37:11  * indexzerojoined
14:37:50  * indexzeropart
14:37:53  * jmar777joined
14:38:25  * jmar777quit (Remote host closed the connection)
14:42:28  * dsantiagoquit (Ping timeout: 252 seconds)
14:45:36  * dsantiagojoined
14:48:42  * TheJHjoined
14:51:21  * loladiroquit (Quit: loladiro)
14:51:55  * dapjoined
14:55:15  * hzquit
14:55:27  * garbagecollectioquit (Quit: garbagecollectio)
14:58:43  * bradleymeckjoined
15:00:51  * hzjoined
15:06:49  * dsantiagoquit (Ping timeout: 256 seconds)
15:25:08  * hzquit (Disconnected by services)
15:25:12  * hzjoined
15:25:51  * indexzerojoined
15:30:28  * dapquit (Quit: Leaving.)
15:31:10  * dapjoined
15:33:36  * jmar777joined
15:37:19  <bnoordhuis>indutny: aw
15:37:30  <bnoordhuis>for one, the commit message should have been reworded
15:39:17  * dsantiagojoined
15:43:17  <isaacs>indutny: i dont' get that bug: https://github.com/joyent/node/issues/4727
15:46:33  <piscisaureus_>isaacs: I actually like how the NPM internals are laid out. Kudos!
15:47:26  <isaacs>piscisaureus_: thanks
15:48:39  * TheJHquit (Quit: goodbye)
15:52:17  <piscisaureus_>isaacs: I wonder though how you got the rewrite rule set up in iriscouch
15:52:25  <piscisaureus_>er, the vhost
15:53:00  <isaacs>piscisaureus_: it's just a config in couchdb
15:53:12  <isaacs>piscisaureus_: log into /_utils and click "configuration"
15:53:26  <isaacs>piscisaureus_: or there's some http way to set it, i guess
15:53:30  <piscisaureus_>ah
15:53:35  <piscisaureus_>ok will try, thnx
15:57:51  <isaacs>indutny: we ignore the encoding if the thing is a buffer: var enc = Buffer.isBuffer(data) ? 'buffer' : encoding;
15:58:10  <isaacs>indutny: then in createWriteReq: switch (encoding) {
15:58:10  <isaacs> case 'buffer':
15:58:10  <isaacs> return handle.writeBuffer(data);
15:59:06  <bnoordhuis>call in 2?
15:59:08  <bnoordhuis>1
16:00:29  * sblomjoined
16:01:48  <isaacs>oh, right
16:01:49  <isaacs>one sec
16:01:58  * isaacswill be a few minutes late, sorry, forgot and didn't pour coffee :)
16:02:11  <indutny>call?
16:03:05  <isaacs>5 minutes
16:03:06  <isaacs>sorry
16:03:13  <indutny>np
16:03:29  <bnoordhuis>another np from me
16:03:54  <bnoordhuis>god... skype on the mac is such a POS
16:03:59  <indutny>POS?
16:04:19  <tjfontaine>piece of shit
16:04:36  * bnoordhu1sjoined
16:04:37  <indutny>looks like TOS
16:04:38  <indutny>to me
16:04:52  <cjd>maybe it's a PointOfSale but most likely not
16:05:58  <cjd>skype's architecture is so similar to that of malware, it's really disconcerting
16:06:35  <tjfontaine>things have started to change under ms though, super nodes aren't elected anymore they're delegated to dedicated hardware by ms
16:06:58  <indutny>yeah
16:07:00  <bnoordhuis>and powered by node!
16:07:04  <indutny>well, that system has failed
16:07:05  <indutny>once
16:07:12  <indutny>delegation of super-nodes
16:07:33  <indutny>I guess running it on dedicated hardware is much simplier than monitoring p2p network
16:07:51  <sblom>Do any of you know the history behind why we install per-user on Windows instead of system-wide?
16:08:10  <tjfontaine>so the peons who don't have admin access still can use it?
16:08:18  * TheJHjoined
16:08:40  <indutny>yeah
16:08:47  <indutny>system-wide requires additional privileges
16:08:56  * `3rdEdenchanged nick to `3E|SHOPPING
16:09:04  <cjd>that big christmas supernode failure was an attack
16:09:13  <sblom>Why didn't I think of that. Makes perfect sense.
16:09:31  <sblom>I'll double-check with isaacs, but I bet that's how he managed to get 2 installs.
16:09:53  <cjd>somebody regged a user called diplomas_here and mass msg'd people with a message which looked like spam but crashed them
16:10:37  <cjd>people talked about being crashed on forums and stuff..
16:11:06  <cjd>And it came a few months before the MS aquisition so I'd bet you a cup of coffee where that attack came from :)
16:11:45  <tjfontaine>heh
16:22:47  * AvianFluquit (Remote host closed the connection)
16:23:21  * AvianFlujoined
16:25:58  * perezdquit (Quit: perezd)
16:31:17  * bradleymeck_joined
16:31:19  * bradleymeckquit (Read error: Connection reset by peer)
16:31:19  * bradleymeck_changed nick to bradleymeck
16:45:36  <isaacs>indutny: fix for that bug coming up shortly
16:45:52  <bnoordhuis>where was nathan? i wanted to sing happy birthday for him
16:48:16  <indutny>ok
16:51:21  <isaacs>indutny: https://github.com/joyent/node/pull/4729
16:51:30  <isaacs>bnoordhuis: i don't know, he wasn't on skype
16:51:45  <indutny>interesting
16:51:52  <indutny>isaacs: are tests passing
16:51:53  <indutny>also
16:51:58  <indutny>why was this check in place at all?
16:52:10  <indutny>I mean, it might be worth adding another branch for buffer input
16:53:07  <indutny>isaacs: ^
16:53:18  <isaacs>indutny: i'm not sure.
16:53:20  <isaacs>it's clearly wrong
16:53:24  <indutny>why?
16:53:40  <indutny>input should be converted to string
16:53:44  <indutny>if itsn't string
16:54:08  <isaacs>indutny: if it's not a string or a buffer, then it's an objectMode stream
16:54:09  <indutny>or will new Buffer() will do it for us?
16:54:12  <isaacs>userland monkey busines
16:54:26  <isaacs>if it's a buffer, great, do nothing. if it's a string, make it a buffer.
16:54:29  <indutny>ok
16:54:32  <indutny>if tests are passing
16:54:38  <isaacs>unless decodeStrings is set to false
16:54:44  <isaacs>in which case, just pass it through as-is
16:54:46  <indutny>but I still think that it might be better to do another check
16:55:00  <isaacs>the check is rihgt above
16:55:01  <indutny>like if (b instanceof Buffer) { }
16:55:02  <isaacs> if (!state.objectMode &&
16:55:02  <isaacs> typeof chunk !== 'string' &&
16:55:02  <isaacs> chunk !== null &&
16:55:02  <isaacs> chunk !== undefined &&
16:55:05  <isaacs> !Buffer.isBuffer(chunk))
16:55:07  <isaacs> state.objectMode = true;
16:55:10  <indutny>oook :)
16:55:13  <indutny>land it then
16:55:17  <indutny>LGTM
16:55:20  <isaacs>if it's not a string or a buffer or undefined or null, then it's an objectmode
16:55:22  <isaacs>kewl
16:55:35  <indutny>I see your point now
16:55:44  <isaacs>and actually... null/undefined shouldn't even be here...
16:55:48  <isaacs>since write(null) is an error
16:56:03  <isaacs>but whatever. we can fix it when it matters. just another speedup loop
16:57:19  * indexzeroquit (Quit: indexzero)
16:58:24  * qmx|awaychanged nick to qmx
16:58:33  <indutny>ok
16:58:34  <indutny>got it
16:58:42  <indutny>so I think nothing more prevents us from doing a release
16:58:46  <indutny>0.9.9
17:02:01  <isaacs>yep
17:02:11  <MI6>joyent/node: isaacs master * 33b2aeb : stream: Writable should ignore encoding for buffers Fix #4727 Fix einaro - http://git.io/IZB5jw
17:02:15  <isaacs>gonna just rebase the release branch onto master now and then get it rolling
17:03:27  * bradleymeckquit (Read error: Connection reset by peer)
17:03:42  * bradleymeckjoined
17:10:32  <isaacs>review changelog, please: https://github.com/joyent/node/commit/v0.9.9-release
17:11:04  <isaacs>oh.. that's odd. it didn't put the \ chars in the commit message.
17:11:08  <MI6>joyent/node: isaacs created branch v0.9.9-release - http://git.io/XUUU3g
17:11:39  * sblomquit (Ping timeout: 244 seconds)
17:11:49  <MI6>joyent/node: isaacs v0.9.9-release * 4b9f0d1 : 2013.02.07, Version 0.9.9 (Unstable) * tls: port CryptoStream to streams - http://git.io/Tw1yJA
17:15:19  <indutny>lgtm
17:29:20  * sblomjoined
17:35:13  * hzquit
17:36:22  * trevnorrisjoined
17:39:00  * Raltquit (Ping timeout: 248 seconds)
17:42:13  * jmar777quit (Remote host closed the connection)
17:52:22  <isaacs>ok, waiting on windows slowpoke builds
17:52:26  <isaacs>and then we're good
18:01:59  <indutny>isaacs: agree to land it https://github.com/joyent/node/pull/4720 ?
18:09:07  <trevnorris>isaacs: pr-4687, not sure why that error is happening. i'll take a look today.
18:09:32  * loladirojoined
18:15:08  <ryah>when is the next 0.9 release?
18:15:14  * jmar777joined
18:15:53  <ryah>isaacs: --^
18:16:24  <ryah>oh i see the branch there .. cool
18:18:00  * qmxchanged nick to qmx|away
18:22:09  <isaacs>ryah: minutes :)
18:22:37  <isaacs>indutny: i want to review that more carefully
18:22:46  <isaacs>we may wnat to change how that works. it's a bit weird now
18:23:23  <indutny>yes
18:23:30  <indutny>have you read my last comment to it?
18:24:07  <isaacs>indutny: yeah
18:24:15  <isaacs>indutny: it might be worthwhile to have the lowWaterMark do this
18:24:26  <isaacs>like, it'll automatically start reading up to that level if it's not zero
18:24:36  <indutny>hm...
18:24:38  <isaacs>but it gets all weird for stdin
18:24:43  <isaacs>and other lwm=0 streams
18:27:29  <isaacs>hm... this is weird. the windows x64 build made a node-v0.8.19-x64.msi file
18:27:36  <isaacs>that seems... not possible
18:28:12  <isaacs>but the node.exe it created says Release\node.exe -v v0.9.9
18:28:36  * mikealquit (Quit: Leaving.)
18:29:00  <isaacs>hm. rebuild fixed it.
18:29:08  <isaacs>windows is magical and strange.
18:29:24  * mikealjoined
18:29:28  * `3E|SHOPPINGchanged nick to `3rdEden
18:31:16  <ryah>is this going to be the last 0.9 ?
18:31:18  * Raltjoined
18:31:24  <indutny>ryah: I'm not sure
18:31:30  <indutny>I think no
18:31:35  <indutny>we should match up 0.8 performance
18:31:36  <indutny>before 0.10
18:31:42  <indutny>so at least there'll be 0.9.10
18:31:43  <ryah>indeed
18:31:54  <indutny>and after - who knows
18:33:42  <MI6>joyent/node: isaacs created tag v0.9.9 - http://git.io/VqFZ0Q
18:34:03  * stagas_joined
18:34:22  <MI6>joyent/node: isaacs master * c3c8eee : Now working on v0.9.10 (+2 more commits) - http://git.io/J8uLBA
18:35:01  <MI6>joyent/node: isaacs v0.8 * e4d97b1 : blog: v0.9.9 - http://git.io/lmggLw
18:35:39  * stagasquit (Ping timeout: 252 seconds)
18:35:42  * stagas_changed nick to stagas
18:36:00  <MI6>joyent/node: isaacs v0.8 * 2810b1a : blog: v0.9.9 is unstable, not stable - http://git.io/ODBVxg
18:38:28  <isaacs>ryah: there'll be a v0.9.10 next week. then 0.10 will be after that, hopefully.
18:38:36  <isaacs>The name of the game at this point is performance.
18:39:05  <isaacs>we have no features to add, that's it. just have to get better benchmarks, and then make those benchmarks faster.
18:40:12  * `3rdEdenquit (Quit: brb, k)
18:40:57  * Raltquit (Remote host closed the connection)
18:41:19  <isaacs>trevnorris: do you have your makecallback/tickcallback stuff in a branch somewhere i can check out?
18:41:37  <trevnorris>isaacs: pr-4687
18:41:40  <isaacs>kewl
18:41:42  <isaacs>thanks
18:41:48  <trevnorris>np. still in progress.
18:41:59  <trevnorris>and your change is causing one test to fail. mentioned in pr.
18:44:05  <isaacs>trevnorris: yeah, i think because some of the messages changed.
18:44:14  <isaacs>trevnorris: ie, the stack doesn't include process._makeCallback any more.
18:44:22  <isaacs>i'll investigate.
18:44:46  <trevnorris>cool. should have my part done soon.
18:44:48  <trevnorris>ish
18:45:08  * bradleymeckquit (Quit: bradleymeck)
18:46:48  <isaacs>trevnorris: fixed: https://github.com/isaacs/node/commit/make-callback-in-js-revert
18:47:13  <trevnorris>ah, heh. ok.
18:47:18  <trevnorris>i'll get that
18:47:36  <isaacs>also rebased onto master
18:47:43  <trevnorris>will do
18:49:09  <trevnorris>isaacs: dude, you have a mother load of branches.
18:49:49  * bnoordhu1squit (Ping timeout: 244 seconds)
18:50:16  * bnoordhu1sjoined
18:50:56  <isaacs>trevnorris: i do
18:51:01  <isaacs>i use branches very liberally
18:51:08  * TooTallNatejoined
18:51:21  <isaacs>TooTallNate: i have a question for you..
18:51:31  <isaacs>TooTallNate: if i make a binary addon and want to use openssl, does gyp make that easy?
18:51:42  <isaacs>TooTallNate: like, can i just load in openssl and get node's openssl?
18:51:50  <indutny>TooTallNate: happy birthday! :)
18:51:53  <TooTallNate>isaacs: except for windows, yes
18:52:00  <TooTallNate>indutny: thanks!
18:52:28  <isaacs>TooTallNate: what about windows?
18:52:29  <TooTallNate>isaacs: https://github.com/TooTallNate/node-gyp/wiki/Linking-to-OpenSSL
18:53:01  <TooTallNate>isaacs: https://github.com/joyent/node/issues/4051
18:53:48  <isaacs>TooTallNate: ok. we should fix that, probably
18:53:52  <isaacs>thanks :)
18:53:59  <TooTallNate>agree'd
18:54:03  <isaacs>these dtls guys really do not believe that it can be done in userland.
18:54:38  <isaacs>also, they seem to be hesitant to put the code on github, but if we give them permission to do so (and i think maybe agree in advance to accept it?) will send us a pull request.
18:54:45  <isaacs>sorta... not how it works.
18:57:28  * Raltjoined
18:57:42  <TooTallNate>hahah, weird
19:03:24  * brsonjoined
19:04:21  * `3rdEdenjoined
19:04:29  <isaacs>it'd be lovely to get dtls, but like, even if we do it as a builtin, it's not going to land before 0.12
19:04:38  <isaacs>so that'll suck anyway. they need to do this in userland.
19:05:04  * qmx|awaychanged nick to qmx
19:05:49  * Raltquit (Remote host closed the connection)
19:07:21  <isaacs>TooTallNate: btw, I thwarted your future-bug report! https://github.com/einaros/ws/issues/159#issuecomment-13253723
19:07:36  <TooTallNate>isaacs: that's awesome!
19:07:39  * isaacsin ur timelinz cauzin ur paradoxes
19:10:18  <TooTallNate>isaacs: wow, i'm surprised einaros was/is using binary strings
19:16:10  <indutny>yep
19:16:13  <indutny>I was surprised too
19:21:06  <tjfontaine>https://github.com/joyent/node/issues/4315#issuecomment-13252950 ... could do it himself of course
19:21:18  <tjfontaine>but why break his streak on waiting on others to do it for him
19:26:22  * bradleymeckjoined
19:30:04  <TooTallNate>i have no idea what ECDHE even is
19:30:15  * loladiroquit (Quit: loladiro)
19:30:30  <tjfontaine>eliptic curve stuff
19:30:38  <tjfontaine>+l
19:30:46  <cjd>I know ECDH but I'm not sure about ECDHE
19:31:23  <indutny>I know it
19:31:31  <indutny>so it isn't supported by node.js?
19:31:36  <indutny>interesting :)
19:31:39  <indutny>ah
19:31:43  <indutny>we only support PEM keys
19:36:51  * sblomquit (Ping timeout: 260 seconds)
19:37:04  * Raltjoined
19:37:09  <isaacs>piscisaureus_: there's a few errors in your talk :)
19:37:23  <isaacs>piscisaureus_: streams have to inherit from stream.Readable or stream.Writable (or stream.Duplex)
19:37:31  <isaacs>piscisaureus_: inheriting from Stream is not so useful.
19:40:45  <trevnorris>isaacs: should users have the ability to override nextTick and/or _tickCallback?
19:40:57  <isaacs>trevnorris: they should not do that :)
19:41:02  <trevnorris>cool
19:41:05  <isaacs>trevnorris: if they do, and crazy undefined errors etc happen, then fin
19:41:10  <tjfontaine>only if they want to never get support
19:41:11  <isaacs>but we dno't have to actively prevent it
19:41:56  <trevnorris>just doing some tests if local references to all the process functions will speed things up.
19:43:14  <isaacs>ah, i see
19:43:17  <isaacs>yeah, that's fine
19:43:23  <trevnorris>cool
19:43:24  <isaacs>i think just assume that users will not break that.
19:43:54  * mikealquit (Remote host closed the connection)
19:44:10  * mikealjoined
19:48:47  <trevnorris>isaacs: dude, _tickCallback is like the arc de triomphe
19:51:25  * dapquit (Quit: Leaving.)
19:53:55  * Raltquit (Remote host closed the connection)
19:55:32  * bnoordhu1squit (Ping timeout: 248 seconds)
19:56:52  * EhevuTovjoined
20:01:18  <isaacs>orly??
20:01:30  <isaacs>in the sence that it's massive and made of stone, and in france?
20:01:40  <isaacs>or that it's the center where all things flow through?
20:01:40  <trevnorris>lol, exactly.
20:01:45  <trevnorris>yes. that
20:02:01  <isaacs>most stuff hangs directly off of MakeCallback, I thought
20:02:08  <isaacs>but every MakeCallback hits tickCallback, so that makes sense.
20:03:43  * philips_quit (Excess Flood)
20:04:10  <isaacs>trevnorris: what benches are you seeing diffs on?
20:04:48  <trevnorris>isaacs: right now just 3% above yours. but i'm a spit away from getting another 3-5% at least.
20:04:56  <isaacs>nice
20:05:28  <trevnorris>isaacs: you can see what changes i've made so far to src/node.js on pr-4687
20:05:35  <trevnorris>right now just some logic cleanup.
20:08:47  * philips_joined
20:09:04  * EhevuTovquit (Quit: This computer has gone to sleep)
20:09:24  * philips_quit (Excess Flood)
20:09:48  * EhevuTovjoined
20:10:53  * Raltjoined
20:13:37  * qmxchanged nick to qmx|away
20:14:35  * mikealquit (Quit: Leaving.)
20:14:48  * philips_joined
20:14:53  * qmx|awaychanged nick to qmx
20:18:54  <isaacs>kewl
20:43:36  * mikealjoined
20:46:01  * mikealquit (Client Quit)
20:46:18  * stephankquit (Quit: *Poof!*)
20:52:09  * trevnorris&
20:52:09  <LOUDBOT>I AM OK WITH ONE OF THOSE IDEAS
20:58:46  * stephankjoined
20:59:30  * dapjoined
21:01:42  * Raltquit (Remote host closed the connection)
21:01:47  * EhevuTovquit (Quit: This computer has gone to sleep)
21:02:29  * EhevuTovjoined
21:07:26  * hzjoined
21:09:47  * stagasquit (Ping timeout: 252 seconds)
21:14:22  * qmxchanged nick to qmx|away
21:16:12  * mikealjoined
21:18:06  * qmx|awaychanged nick to qmx
21:22:25  * stagasjoined
21:22:51  * bradleymeckquit (Quit: bradleymeck)
21:23:09  * loladirojoined
21:31:51  * isaacslunching
21:37:52  * trevnorrisfg
21:39:23  <indutny>isaacs: have you thought about signing release tags with pgp key?
21:39:31  <isaacs>indutny: i do
21:39:41  <indutny>and, why ain't we doing it?
21:39:44  <isaacs>indutny: git show v0.9.9
21:40:03  <isaacs>indutny: https://gist.github.com/4734495
21:40:22  <indutny>oh
21:40:23  <indutny>right
21:40:25  <indutny>interesting
21:40:50  <isaacs>if everyone had the latest version of git, we could also be signing all commits.
21:40:52  <indutny>and where is open part of key?
21:41:01  * jmar777quit (Remote host closed the connection)
21:41:22  <indutny>I can sign my commits :)
21:41:28  <indutny>just in case
21:41:32  <isaacs>indutny: http://pgp.mit.edu:11371/pks/lookup?op=vindex&search=0xB0A78B0A6C481CF6
21:41:43  <isaacs>http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xB0A78B0A6C481CF6
21:42:00  <indutny>kewl
21:42:07  <indutny>how could one register himself there?
21:42:12  <indutny>should I send my passport to them?
21:42:17  <isaacs>i don't remember, exactly
21:42:24  <isaacs>i think you just have to verify ownership of the email address
21:42:25  <indutny>meh
21:42:29  <indutny>I too lazy :)
21:42:38  <indutny>I don't think it really fits github flow
21:42:53  <isaacs>indutny: well, we don't really use github's flow anyway. we use git's flow.
21:43:05  <isaacs>we wouldn't have all contributors sign off, just the committer
21:43:08  <indutny>well, we accept pull requests
21:43:11  <isaacs>right
21:43:13  <tjfontaine>do we use gitflow? :)
21:43:14  <isaacs>but we don't use the green button
21:43:18  <indutny>oh
21:43:21  <isaacs>tjfontaine: ugh. no.
21:43:24  <indutny>you call this git flow? :)
21:43:26  <tjfontaine>heh
21:43:30  <indutny>I don't trust this green button
21:43:33  <indutny>who knows what it'll do
21:43:36  <isaacs>indutny: i mean, we don't use the github features, really
21:43:41  <isaacs>github is just a pretty way to email patches
21:43:45  <trevnorris>indutny: don't have to. works just as well if you grab their remote, merge it and push. github auto detects that.
21:43:46  <indutny>time to move to other server
21:43:50  <indutny>and setup email for patches
21:44:09  <trevnorris>indutny: i use pr's just for keeping track. but I never use the "apply pr" button.
21:44:53  <isaacs>yes, the green button is not allowed in node.
21:45:01  <isaacs>you must apply the patch, run tests, then push
21:45:06  <trevnorris>we could start asking contributors to use "git format-patch" ;-)
21:45:18  <isaacs>trevnorris: well, i mean, slap a .patch on the end of the url, and you've got exactly that.
21:45:23  <isaacs>trevnorris: but that IS how we used to do it
21:45:30  <trevnorris>ah ok
21:45:33  <isaacs>email a .patch to [email protected]
21:45:38  <isaacs>then [email protected]
21:45:43  <isaacs>then eventually we just moved to github
21:46:08  <tjfontaine>and if it ever becomes a problem there's always things like gerrit
21:46:33  * wolfeidauquit (Remote host closed the connection)
21:48:00  <isaacs>ok, srsly, heading out. laters
21:48:11  <trevnorris>would someone mind checking this out: https://github.com/joyent/node/pull/4687/files#r2934197
22:00:24  * roxlujoined
22:00:25  * EhevuTovquit (Quit: This computer has gone to sleep)
22:00:26  <roxlu>hi
22:01:57  <roxlu>I'm working on some networking code and wondering how to approach this. I've got a [provider]-->[server]-->{many clients}. When the [provider] sends data, it's distributed to the {many clients}.
22:02:34  <roxlu>but when one of the {many_clients} disconnects and [provider] sends something, my [server] tries to write data to this {client} which just disconnected
22:02:38  * lohkeyjoined
22:03:07  <roxlu>I'm calling "uv_shutdown/uv_close" but I keep getting: Assertion failed: (!uv__io_active(&stream->io_watcher, UV__POLLIN) && "stream->read_cb(status=-1) did not call uv_close()"), function uv__read, file src/unix/stream.c, line 995.
22:04:12  <roxlu>I'm closing a connection when I receive a -1 or EOF in my read callback
22:05:41  * wolfeidaujoined
22:07:03  * rendarquit
22:12:49  * sblomjoined
22:20:04  * hzquit
22:26:41  <indutny>isaacs: I've another idea
22:26:47  <indutny>isaacs: we could distribute sources by email
22:26:48  <indutny>on-demand
22:27:28  <indutny>roxlu: so uv_close() is failing when received EOF?
22:28:14  <roxlu>indutny: I think the problem is that uv_write is called before I get a chance to call uv_close
22:28:25  <indutny>this should not be possible
22:29:19  * roxlushould create a test case
22:29:53  <indutny>yep
22:29:59  <indutny>this would be highly welcome
22:33:40  <roxlu>i need to recode some stuff for which I only have time this weekend, but I'll let you know if I got time to sit down for it
22:34:34  <indutny>ok
22:34:35  <indutny>np
22:34:46  <indutny>I'm pretty sure there's probably a bug in your code :)
22:34:50  <indutny>but I need to see it first :P
22:34:51  <indutny>brb
22:35:21  <roxlu>yeah I hope so : )
22:36:17  * sgallaghquit (Remote host closed the connection)
22:41:25  <trevnorris>indutny: dude. i'm having some of the weirdest ass issues with the ticker. mind being a sanity check?
22:43:04  * qmxchanged nick to qmx|away
22:45:09  <CoverSlide>ass issues are the worts
22:45:12  <CoverSlide>*worst
22:45:31  <CoverSlide>http://xkcd.com/37/
22:46:16  <trevnorris>heh, awesome
22:46:51  <trevnorris>CoverSlide: but seriously. some how the test, and only the test, test-process-next-tick.js is failing
22:47:07  <trevnorris>but when I put a console log in the execution chain to observe what is going on it succeeds.
22:50:37  <indutny>trevnorris: sure
22:52:53  * chrisdickinsonquit (Excess Flood)
22:53:31  * chrisdickinsonjoined
22:54:56  <trevnorris>indutny: https://github.com/joyent/node/pull/4687/files#r2934197
22:55:07  * TooTallNatequit (Quit: Computer has gone to sleep.)
22:55:17  * loladiroquit (Quit: loladiro)
22:55:22  <trevnorris>indutny: but when I put a console.log in that test to see the counter, it succeeds.
22:55:29  <indutny>oh
22:55:33  <indutny>I guess IO
22:55:53  <trevnorris>what I really don't get is how only that test could fail.
22:56:09  <trevnorris>it seems that failure would cause other tests to as well.
22:59:55  <indutny>its pretty odd
22:59:57  <indutny>but it's late
23:00:10  <indutny>and I can't really figure out anything now
23:00:11  <indutny>sorry
23:00:15  <trevnorris>ah yeah. forgot. no worries.
23:01:08  <trevnorris>just going to pummel it into submissions.
23:01:22  * `3rdEdenquit (Remote host closed the connection)
23:02:29  * indutny&
23:02:30  <LOUDBOT>HO NOES ITS THE CAREBEAR STARE
23:02:45  * TooTallNatejoined
23:05:36  * TheJHquit (Ping timeout: 264 seconds)
23:07:37  * bnoordhu1sjoined
23:16:01  * jmar777joined
23:18:29  * piscisaureus_&
23:18:30  <LOUDBOT>CONTROL IS NOT CONVINCED, BUT THE COMPUTER HAS TEH EVIDENCE!
23:21:53  * garbagecollectiojoined
23:22:04  * c4miloquit (Remote host closed the connection)
23:29:43  * c4milojoined
23:31:49  * shigekijoined
23:34:12  * c4miloquit (Ping timeout: 248 seconds)
23:41:36  <mmalecki>isaacs: our npm error rates have drastically decreased on smartos since that patch, fyi
23:41:45  <isaacs>mmalecki: i'll bet :)
23:45:05  <trevnorris>isaacs: dude, i'm going to rip this ticker to bits. keep getting little perf gains just to have a random test fail.
23:46:02  <isaacs>hahah
23:46:09  <isaacs>those tests are for reasons :)
23:46:16  <isaacs>the semantics have to remain contstant
23:46:25  <trevnorris>heh, yeah.
23:46:49  <cjd>you guys use tests which are shipped with random npm packages to test node?
23:47:10  <trevnorris>isaacs: problem is that the obvious logic flow is consistent, but little timing issues come up.
23:47:41  <trevnorris>isaacs: and when I mean little, in one test just by placing a console.log in the function to monitor the outcome the test would pass.
23:47:46  <trevnorris>but w/o the console.log the test fails...
23:48:25  <isaacs>yep
23:48:49  <isaacs>especially when dealing with timing issues around net sockets, that happens a LOT
23:48:52  <isaacs>since stdout is a net socket
23:49:16  <isaacs>i sometimes have a debug function shove messages in an array, then dump them on process exit
23:49:44  <trevnorris>aahhh. good idea.
23:50:21  <bnoordhu1s>cjd: no, not really
23:50:40  <trevnorris>i really want to get this sucker smooth. tracked how many times _tickCallback ran in 5 seconds of net2_c2s. ran about 1 million times.
23:51:57  <trevnorris>isaacs: anyways. the changes posted in 4687 are the best I can do right now.
23:52:05  <isaacs>ok
23:52:45  <trevnorris>it just "feels" like there's something right there, but can't see it.