00:00:35  <isaacs>updating npm, too
00:00:39  <CIA-115>node: isaacs v0.6 * r5c0f039 / (95 files in 5 dirs): Update npm to 1.1.0-3 - http://git.io/edpmvQ
00:00:40  <isaacs>let's see how long it takes for me
00:00:47  <isaacs>yeah, github just hates europe, i think
00:00:55  <isaacs>they're racist
00:01:32  <bnoordhuis>i'm going to start my own github
00:01:37  <bnoordhuis>with blackjack and hookers
00:01:54  * AvianFlu_joined
00:02:02  * AvianFluquit (Disconnected by services)
00:02:07  * AvianFlu_changed nick to AvianFlu
00:03:36  <isaacs>bnoordhuis: not gonna start cheering yet, but the test just blasted past the dgram stuff without a hiccup
00:03:53  <isaacs>so i'm definitely getting ready to cheer :)
00:05:51  <piscisaureus_>isaacs: the stat improvement patch that landed in 0.6.8 seems to have caused a regression where stat("c:\") errs
00:05:53  <isaacs>best osx test i've seen in a while: [01:52|% 100|+ 323|- 3]: Done
00:06:16  <piscisaureus_>isaacs: do you want to hold the release on that? I am unable to fix it right now :-/
00:06:34  <bnoordhuis>piscisaureus_: "c:\" or "c:\\"?
00:06:35  * travis-cijoined
00:06:36  <travis-ci>[travis-ci] joyent/node#317 (v0.6 - 352febe : Ben Noordhuis): The build passed.
00:06:36  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/ccf7b41...352febe
00:06:36  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/581905
00:06:36  * travis-cipart
00:06:43  <piscisaureus_>bnoordhuis: the latter
00:07:24  <piscisaureus_>bnoordhuis: or "c:/" or "../../../../../.." or whatever
00:07:31  <bnoordhuis>right
00:08:23  <isaacs>piscisaureus_: that's important.
00:08:42  <isaacs>piscisaureus_: do you know why it fails? how big a task it is to fix?
00:08:56  * travis-cijoined
00:08:56  <travis-ci>[travis-ci] joyent/node#318 (v0.6 - 5c0f039 : isaacs): The build passed.
00:08:56  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/352febe...5c0f039
00:08:56  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/581908
00:08:56  * travis-cipart
00:08:56  <piscisaureus_>isaacs: I know, yes
00:09:14  <piscisaureus_>isaacs: it's not completely trivial to fix
00:09:20  <isaacs>piscisaureus_: if we delay until tomorrow, is that enough time?
00:09:36  <piscisaureus_>isaacs: probably, yes
00:09:53  <isaacs>how many people are screaming about this, and how loudly?
00:09:54  <piscisaureus_>yeah that'll work
00:09:55  <isaacs>:)
00:10:02  <piscisaureus_>isaacs: nobody is screaming
00:10:07  <piscisaureus_>isaacs: I just ran into it
00:10:09  <isaacs>sure, let's delay until tomorrow.
00:10:28  <isaacs>it's better to fix it before too many people start being affected by it
00:11:20  <piscisaureus_>ok
00:14:33  * lwillejoined
00:19:08  <piscisaureus_>The only udp problem in my test failures is test-dgram-multicast-multi-process
00:19:54  <wankdanker>piscisaureus_: is it failing or hanging?
00:20:02  <piscisaureus_>hanging
00:20:10  <piscisaureus_>bnoordhuis: why is wankdanker's patch not landed?
00:20:54  <bnoordhuis>piscisaureus_: i thought isaacs was going to do that
00:22:01  <piscisaureus_>oh, I was wrong
00:22:03  <piscisaureus_>release test-dgram-broadcast-multi-process also fails
00:22:16  <wankdanker>what platform?
00:22:18  <piscisaureus_>EADDRINUSE
00:22:20  <piscisaureus_>windows
00:22:51  <wankdanker>you will definitely want https://github.com/joyent/node/pull/2616, at least it shouldn't hang then
00:23:02  <wankdanker>has work been done on windows dgram?
00:23:52  <piscisaureus_>apparently
00:24:04  <isaacs>i'll land it, one sec
00:24:19  <piscisaureus_>oh wait it hasn't
00:24:24  <piscisaureus_>they are just stubs
00:24:29  <bnoordhuis>piscisaureus_: uv_udp_set_broadcast hasn't been implemented
00:24:43  <bnoordhuis>ah, beaten to the punch
00:24:49  <piscisaureus_>bnoordhuis: set_multicast_loop also not
00:25:14  <piscisaureus_>isaacs: bnoordhuis: I will fix these tomorrow as well
00:25:22  <piscisaureus_>this is like 10 minutes work
00:25:51  <bnoordhuis>i was going to but it seems xp sp2 doesn't support multicast
00:26:12  <piscisaureus_>unlikely
00:26:22  <isaacs>k
00:26:24  <piscisaureus_>bnoordhuis: which function?
00:26:56  <bnoordhuis>piscisaureus_: i don't remember
00:27:39  <bnoordhuis>piscisaureus_: oh wait, i do - setsockopt(IP_MULTICAST_LOOP)
00:28:32  <piscisaureus_>this is fun
00:28:34  <piscisaureus_>Note The Winsock version of the IP_MULTICAST_LOOP option is semantically different than the UNIX version of the IP_MULTICAST_LOOP option:
00:28:34  <piscisaureus_> In Winsock, the IP_MULTICAST_LOOP option applies only to the receive path.
00:28:34  <piscisaureus_> In the UNIX version, the IP_MULTICAST_LOOP option applies to the send path.
00:28:53  <bnoordhuis>yeah, i saw that on msdn
00:31:49  <piscisaureus_>bnoordhuis: afaict it is supported on xp
00:31:55  <piscisaureus_>so shouldn't be a problem
00:32:53  <bnoordhuis>good
00:32:55  <isaacs>back, sorr
00:32:57  <isaacs>y
00:33:59  * xaqjoined
00:34:27  <CIA-115>node: Dan VerWeire v0.6 * ra0119af / (2 files):
00:34:28  <CIA-115>node: test: handle unhandled dgram scenarios
00:34:28  <CIA-115>node: - watch for the death of child processes and fail the test if they all die
00:34:28  <CIA-115>node: - use setTimeout to fail the test if responses are not received and processed in 5000ms - http://git.io/ih8wcw
00:34:51  * bbenviejoined
00:35:19  <piscisaureus_>Ok I am gone again. I know what my queue looks like for tomorrow.
00:36:08  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
00:38:45  * isaacsquit (Read error: Connection reset by peer)
00:39:07  * isaacsjoined
00:41:06  <isaacs>bnoordhuis: hey, that issue that rm brought up yesterday on irc, with EPIPE and uncatchable errors trying to close stdout.
00:41:09  <isaacs>you remember that?
00:41:20  <bnoordhuis>isaacs: yes
00:41:44  <isaacs>do you think we can fix that in 0.6.9, or should we just tell him to work around it?
00:42:13  <bnoordhuis>you mean have it fixed by tomorrow evening?
00:42:44  * travis-cijoined
00:42:44  <travis-ci>[travis-ci] joyent/node#319 (v0.6 - a0119af : Dan VerWeire): The build was broken.
00:42:44  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/5c0f039...a0119af
00:42:44  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/582048
00:42:44  * travis-cipart
00:43:23  <isaacs>bnoordhuis: yeah
00:43:28  <isaacs>i'm not familiar with the issue.
00:43:56  <isaacs>"can't be done in 0.6" is an perfectly acceptable answer. "can't be done by tomorrow" is also acceptable.
00:44:20  <bnoordhuis>isaacs: not tomorrow anyway. stdio in node is complex in a spaghetti-like way, it's too easy to introduce regressions
00:44:39  <isaacs>is there an issue that has some details and whatnot?
00:45:09  <bnoordhuis>i don't think he opened one
00:45:21  <isaacs>ok
00:45:26  <isaacs>he said that yodu' referenced one.
00:45:34  <bnoordhuis>oh right, that's true
00:45:44  <isaacs>he doesn't have logs, though, and the joyent outage in AM1 seems to have prevented bert's bot from logging it
00:45:46  * bnoordhuisdigs in his logs
00:47:12  <tjfontaine>https://github.com/joyent/libuv/blob/master/src/win/util.c#L488 (uv_interfaces_address) makes a call to GetAdaptersAddresses which includes defined dns servers per adapter, I presume the probabilty of accepting a patch that passed that information to os.networkInterfaces would be quite slim?
00:48:10  <tjfontaine>(since it's not a pattern followed on the other platforms)
00:48:29  <bnoordhuis>isaacs: https://github.com/joyent/node/issues/2507
00:48:56  * mikealquit (Quit: Leaving.)
00:49:21  <bnoordhuis>tjfontaine: quite slim yes
00:49:51  <isaacs>bnoordhuis: ahh, right.
00:49:52  <tjfontaine>bnoordhuis: sad, but quite predictable, thanks
00:49:53  <isaacs>i know this one
00:51:17  <isaacs>i think throwing in that case is inappropriate, really
00:51:17  * paddybyersquit (Ping timeout: 276 seconds)
00:51:36  <isaacs>the problem is that we .close() when you .pipe() from one thing toa nother
00:52:46  <bnoordhuis>i can live with removing the throw
00:54:09  <isaacs>yeah, i wouldn't want to do too much more in there.
00:55:21  <isaacs>igorzi's point about making sure we don't give the wrong impression that it's actually closing when it isn't is valid, but if we wanted to be that strict about it, then we should rejigger pipe/etc so that it doesn't try to close it in so many different ways.
01:01:18  * mikealjoined
01:13:39  * bradleymeckjoined
01:19:35  * mikealquit (Quit: Leaving.)
01:20:25  <isaacs>bnoordhuis: https://github.com/isaacs/node/commit/e7926339b8dc19e645e740c2d4b5a76a664dcf72
01:20:41  <isaacs>bnoordhuis: not sure that's the ideal fix, but it seems like the minimum possible change to retain the best of 0.4 and 0.6 use cases
01:24:04  <bnoordhuis>isaacs: why is .end a no-op?
01:24:28  <isaacs>bnoordhuis: because there's no case i could found where we call it
01:24:38  <isaacs>and the semantics are not such that .end() makes the handle get destroyed
01:24:41  <isaacs>that's what .close() does
01:24:50  <isaacs>so that's the only one we have to warn about
01:25:45  <bnoordhuis>isaacs: net.Stream.end calls .destroySoon
01:26:03  <bnoordhuis>i'm sure that's used (maybe implicitly) in some places
01:26:15  <isaacs>bnoordhuis: sure.
01:26:26  <isaacs>hm.
01:26:42  <isaacs>so maybe we should leave .end() alone, and just hijack destroy{,Soon}?
01:27:00  <bnoordhuis>i think so, yes
01:29:42  <wankdanker>good night, guys. good luck with 69
01:31:28  <isaacs>bnoordhuis: https://github.com/isaacs/node/commit/40f8d5570f064f8d232cb6a72e0309e7ef751454
01:31:31  <isaacs>wankdanker: thanks!
01:33:05  <bnoordhuis>isaacs: looks good, i think
01:33:16  <bnoordhuis>i'm kind of undecided on .emit
01:33:31  <bnoordhuis>i guess you could just console.trace if you want to warn people
01:33:42  <bnoordhuis>but i don't know
01:33:54  <isaacs>the problem is that we call .destroy() when there's an underlying error
01:34:01  <isaacs>like if you hit an EPIPE
01:34:06  <isaacs>you have to be able to handle that
01:34:47  <isaacs>bnoordhuis: https://gist.github.com/1686286
01:35:07  <bnoordhuis>oh sure, it's a sensible thing to do on errors
01:36:31  <bnoordhuis>isaacs: land it if you want
01:36:54  <isaacs>yeah, and there's no change for people who aren't already trying to catch it, since emit("error") throws
01:37:09  * orlandovftwquit (Ping timeout: 248 seconds)
01:37:29  <isaacs>k, landing it. i'll make a note to bug piscisaureus or igorzi about it tomorrow to at least get a +1
01:37:31  <isaacs>thansk
01:37:53  * mikealjoined
01:56:51  <CIA-115>node: isaacs v0.6 * rff0f0ae / (3 files in 3 dirs):
01:56:51  <CIA-115>node: Fix #2507 Raise errors less agressively when destroying stdio streams
01:56:51  <CIA-115>node: Also, if an error is already provided, then raise the provided
01:56:51  <CIA-115>node: error, rather than throwing it with a less helpful 'stdout cannot
01:56:51  <CIA-115>node: be closed' message.
01:56:52  <CIA-115>node: This is important for properly handling EPIPEs. - http://git.io/aNmEpQ
01:57:09  <isaacs>bnoordhuis: that's what you reviewed, plus a test.
01:57:15  <isaacs>sorry, didn't get the test reviewed.
01:57:31  * isaacsdoesn't live up to his own standards sometimes.
01:57:35  <bnoordhuis>that's okay, i have faith in my co-workers
01:58:24  <tjfontaine>famous last words
02:00:24  * bnoordhuissigns off
02:04:53  * bnoordhuisquit (Ping timeout: 248 seconds)
02:05:05  * travis-cijoined
02:05:05  <travis-ci>[travis-ci] joyent/node#320 (v0.6 - ff0f0ae : isaacs): The build is still failing.
02:05:05  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/a0119af...ff0f0ae
02:05:05  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/582430
02:05:05  * travis-cipart
02:25:03  * TooTallNatequit (Quit: Linkinus - http://linkinus.com)
02:26:42  * xaqquit (Remote host closed the connection)
02:36:15  * AvianFluquit (Quit: Leaving)
02:40:01  * brsonquit (Quit: leaving)
02:49:01  * mikealquit (Quit: Leaving.)
02:50:32  * dshaw_quit (Ping timeout: 240 seconds)
02:54:54  * mikealjoined
02:55:17  * mikealquit (Client Quit)
03:26:00  * sh1mmerquit (Quit: sh1mmer)
03:46:29  * mikealjoined
04:07:54  * sh1mmerjoined
04:10:22  * mikealquit (Quit: Leaving.)
04:11:59  * skabbespart
04:20:17  * isaacsquit (Remote host closed the connection)
04:31:05  * sh1mmerquit (Quit: sh1mmer)
04:32:47  * mikealjoined
04:35:50  * mikealquit (Client Quit)
04:39:30  * dapquit (Quit: Leaving.)
04:58:57  * mikealjoined
05:01:16  * mikealquit (Client Quit)
05:21:03  * orlandovftwjoined
05:25:52  * dshaw_joined
05:33:19  * xaqjoined
05:46:24  <indutny_sleeping>pquerna: hehe
05:46:25  <indutny_sleeping>pquerna: ok
05:46:40  * indutny_sleepingchanged nick to indutny
05:55:46  * mikealjoined
06:03:39  * mralephjoined
06:20:57  * mikealquit (Quit: Leaving.)
06:25:49  * mikealjoined
06:26:03  * mikealquit (Client Quit)
06:33:19  * mikealjoined
06:34:40  * mikealquit (Client Quit)
06:38:15  * mikealjoined
07:00:25  * mralephquit (Quit: Leaving.)
07:02:57  * xaqquit (Remote host closed the connection)
07:04:46  * mikealquit (Quit: Leaving.)
07:10:44  * mikealjoined
07:20:12  * bradleymeckquit (Ping timeout: 240 seconds)
07:28:25  * mikealquit (Quit: Leaving.)
07:53:45  * paddybyersjoined
07:55:31  * dshaw_quit (Quit: Leaving.)
08:49:35  * mikealjoined
08:55:29  * AvianFlujoined
09:52:25  * orlandovftwquit (Ping timeout: 252 seconds)
12:47:39  * piscisaureus_joined
13:25:34  <piscisaureus_>bnoordhuis: yt?
13:26:01  * bnoordhuisjoined
13:26:07  <piscisaureus_>ircretary tell bnoordhuis Why is there no ipv6 counterpart of uv_udp_set_membership?
13:26:08  <ircretary>piscisaureus_: I'll be sure to tell bnoordhuis
13:26:58  <bnoordhuis>ircretary: tell piscisaureus_ that ipv6 is a fiction
13:26:58  <ircretary>bnoordhuis: I'll be sure to tell piscisaureus_
13:27:36  <piscisaureus_>ircretary tell bnoordhuis Why is there no room in the API for it?
13:28:26  <bnoordhuis>ircretary: tell piscisaureus_ that the multicast support kind of got rushed through
13:28:41  <bnoordhuis>so yeah
13:30:08  <bnoordhuis>hmm... https://github.com/joyent/node/commit/5c0f039c9ca9f36dfc05b5d38d8a20343d02f9f0#commitcomment-906819
14:10:21  <piscisaureus_>bnoordhuis: so do we have any tests for udp multicast support in libuv?
14:10:46  <bnoordhuis>piscisaureus_: no
14:10:51  * piscisaureus_sighs
14:10:55  <bnoordhuis>i wasn't joking when i said it got rushed through
14:11:06  <piscisaureus_>but yeah I can imagine you don't wanna write one
14:12:04  <bnoordhuis>don't want to, no, but i will eventually
14:12:43  <bnoordhuis>writing ttl tests that work on localhost is kind of hard though :/
14:22:23  <indutny>bnoordhuis: yahooo!!!
14:22:27  <indutny>bnoordhuis: I fixed node-waf
14:23:00  <indutny>bnoordhuis: do you have time to review?
14:23:07  <bnoordhuis>indutny: sure
14:24:42  <indutny>bnoordhuis: https://github.com/indutny/node/tree/feature-node-waf-works-on-osx
14:24:46  <indutny>https://github.com/indutny/node/commit/710bc8e1f41de8f48a73465d893af11d4a6f0660
14:32:53  <bnoordhuis>indutny: reviewed
14:33:24  <bnoordhuis>so google is trying to hire people for the v8 team
14:33:29  <bnoordhuis>they're sending me job offers
14:33:33  <bnoordhuis>or interview offers rather :)
14:33:51  <bnoordhuis>piscisaureus_: ^ you'd probably dig it
14:34:23  <piscisaureus_>for the v8 team, that's rather cool
14:35:40  <indutny>bnoordhuis: can I land it?
14:35:45  <indutny>oh
14:35:45  <indutny>I see
14:36:51  <bnoordhuis>piscisaureus_: they're looking for people to work on dart, too
14:36:57  <bnoordhuis>want me to forward it?
14:40:18  <indutny>bnoordhuis: review - https://github.com/indutny/node/commit/85a86b5fd64ee4f5502552096dcc5f616158d7d6 ?
14:40:33  * bradleymeckjoined
14:45:33  <bnoordhuis>indutny: lgtm, assuming it works as advertised :)
14:46:01  <indutny>bnoordhuis: hahaha
14:46:07  <indutny>bnoordhuis: it wfm on osx 10.6
14:46:24  <indutny>anyone with 10.7 here?
14:46:39  <bnoordhuis>indutny: yeah, me - but my laptop's battery is dead
14:46:40  <piscisaureus_>bnoordhuis: can you look up struct ipv6_mreq on linux for me?
14:46:58  <bnoordhuis>i left the charger at the c9 offices last week :/
14:47:01  <indutny>bnoordhuis: ok, pushing
14:47:03  <indutny>:)
14:47:06  <indutny>that's not good
14:47:13  <bnoordhuis>piscisaureus_: sure, what in particular do you need?
14:47:42  <piscisaureus_>bnoordhuis: well on windows in6amr_interface is an integer
14:47:43  <CIA-115>node: Fedor Indutny master * r85a86b5 / tools/wafadmin/Tools/node_addon.py : waf: predict target arch for OSX - http://git.io/8CveZg
14:47:47  <piscisaureus_>it's supposed to be the interface intex
14:47:49  <piscisaureus_>index
14:47:55  <piscisaureus_>*sigh*
14:48:06  <piscisaureus_>bnoordhuis: I wonder if linux is equally retarded
14:48:12  <bnoordhuis>piscisaureus_: yep, struct in6_addr ipv6mr_multiaddr + unsigned int ipv6mr_interface is all it contains
14:48:29  <piscisaureus_>gnarf
14:48:33  <indutny>bnoordhuis: does libuv in 0.6 has in differences in uv_work_t with master?
14:48:47  <bnoordhuis>indutny: i don't think so. why?
14:49:00  <bnoordhuis>piscisaureus_: what specifically do you need?
14:49:18  <indutny>bnoordhuis: dunno, my code works on 0.7.x
14:49:23  <indutny>bnoordhuis: and fails on node 0.6.8
14:49:29  <bnoordhuis>indutny: define 'fails'
14:49:39  <piscisaureus_>bnoordhuis: so we supply an ipv6 address for the interfcace to uv_udp_set_membership
14:49:56  <piscisaureus_>then we have to translate that to an interface index
14:50:37  <indutny>bnoordhuis: process exits, with code 0
14:50:42  <indutny>bnoordhuis: but in the middle of js code
14:50:56  <bnoordhuis>piscisaureus_: https://github.com/joyent/libuv/issues/271 <- like that?
14:51:16  <bnoordhuis>indutny: hm, sounds like a refcount issue
14:51:21  <indutny>bnoordhuis: oooh
14:51:23  <indutny>nooo!!!
14:51:26  <indutny>:(
14:51:37  <indutny>bnoordhuis: I can show you the code if you want
14:51:45  <bnoordhuis>indutny: what happens if you put a breakpoint at exit or _exit and inspect uv_default_loop()->ev->activecnt?
14:51:52  <piscisaureus_>bnoordhuis: yes, that. bleh
14:52:06  <indutny>bnoordhuis: one sec, I'm going to get code out of mocha test
14:52:54  <bradleymeck>indutny for starting/stopping the debugger any complaints if I just add a process.debugger.start/stop(port) ? and do you know if there is a way to expose the raw sendmsg for the debugger from a different isolate/child (ie. have the debugger on but still piping data from a source that is not tcp)?
14:53:43  <indutny>bradleymeck: you can start isolate manually
14:54:05  <indutny>bradleymeck: .fork(.., { thread: true, debug: function(d) {}, debugBrk: true })
14:54:14  <indutny>bradleymeck: no need in process.debugger
14:54:35  <bradleymeck>yea but if I am piping to Isolate A's debugger from Isolate B and do not want to open up a TCP port?
14:55:48  <bradleymeck>indutny, i think i get you, but for child processes?
14:56:24  <indutny>bradleymeck: ah, for child processes
14:56:32  <indutny>bradleymeck: you'll need to create your own thread as v8 do
14:57:02  <indutny>bradleymeck: and your own server in that thread, because you can't do things with debugger in isolate's thread
14:57:10  <bradleymeck>indutny sounds fine, this sounds like its just going to end up a C++ addon then
14:57:34  <piscisaureus_>bnoordhuis: hey
14:57:49  <bnoordhuis>piscisaureus_: ho
14:58:10  <piscisaureus_>bnoordhuis: so how exactly does the interface address relate to the interface
14:58:23  <piscisaureus_>bnoordhuis: does it specify its unicast address, or its multicast address?
14:58:29  <piscisaureus_>bnoordhuis: or either?
14:58:41  <bnoordhuis>piscisaureus_: define 'interface address' - do you mean its name, e.g. 'eth0' or the assigned ip address?
14:59:14  <indutny>bradleymeck: yeah
14:59:20  <piscisaureus_>bnoordhuis: IP_MULTICAST_JOIN takes an multicast interface address that specifies which interface to use
14:59:30  <piscisaureus_>bnoordhuis: that address is in ipv4 format
14:59:38  <indutny>bradleymeck: that's odd #0 0x00000001000622f0 in v8::TryCatch::HasCaught ()
14:59:42  <indutny>SIGABORT here ^
14:59:54  <bnoordhuis>piscisaureus_: you have similar addresses for ipv6
15:01:29  * travis-cijoined
15:01:29  <travis-ci>[travis-ci] joyent/node#321 (master - 85a86b5 : Fedor Indutny): The build is still failing.
15:01:29  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/f89beaf...85a86b5
15:01:29  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/584885
15:01:29  * travis-cipart
15:02:07  <indutny>bnoordhuis: is AfterWork callback of uv_work_t running in event-loop's thread?
15:02:19  <bnoordhuis>indutny: yes
15:02:44  <bradleymeck>what will sigabort do there? idk what the signal handler is set up for
15:03:39  <indutny>bnoordhuis: ok, that's odd but it fails in another place now
15:04:54  <bnoordhuis>bradleymeck: i found the cause of that bug you reported but so far i don't know how to fix it without breaking other things
15:06:16  <bradleymeck>bnoordhuis, yea not without massive duplication :/
15:07:32  <bradleymeck>bnoordhuis, i have been thinking though, if there is a way to check if stdin is dead why not make child.stdin null?
15:07:48  <bradleymeck>force them to use new fds
15:09:07  <indutny>bnoordhuis: ok, that was uv_mutex_t polyfill
15:09:34  <bnoordhuis>bradleymeck: you don't know if the other end has closed the pipe until you try to read from or write to it
15:10:13  <bnoordhuis>indutny: polyfill?
15:11:09  <indutny>bnoordhuis: yep, for 0.6.0
15:14:05  <indutny>bnoordhuis: woohoo, it works
15:14:32  <indutny>bnoordhuis: btw, when are we going to release 0.7.3?
15:14:34  <indutny>what's blocking it?
15:14:49  <indutny>oops
15:14:49  <indutny>0.7.2
15:16:52  <bnoordhuis>indutny: piscisaureus_ is
15:17:01  <bnoordhuis>i kid, there was a last-minute regression on windows
15:17:32  <indutny>oh, ok
15:18:43  * bnoordhuisstarts writing libuv multicast/broadcast tests...
15:19:31  <bnoordhuis>piscisaureus_: btw, there's one test already: udp_multicast_join
15:25:01  <piscisaureus_>bnoordhuis:
15:25:06  <piscisaureus_>bnoordhuis: should that be valid on my machine?
15:27:29  <bnoordhuis>piscisaureus_: yes, it's part of the ipv4 multicast block
15:27:44  <piscisaureus_>bnoordhuis: binding to that address fails for me
15:28:45  <bnoordhuis>piscisaureus_: maybe try an an address in the or /24 range?
15:30:29  <CIA-115>libuv: Bert Belder v0.6 * rf925275 / test/test-fs.c : Fix windows build - http://git.io/j3LpzQ
15:32:09  * travis-cijoined
15:32:10  <travis-ci>[travis-ci] joyent/libuv#54 (v0.6 - f925275 : Bert Belder): The build is still failing.
15:32:10  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/9c76d0d...f925275
15:32:10  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/585044
15:32:10  * travis-cipart
15:32:44  <CIA-115>libuv: Bert Belder mcast-win * r7bc6f3a / src/win/udp.c :
15:32:44  <CIA-115>libuv: Implement udp multicast methods on windows.
15:32:44  <CIA-115>libuv: Test is failing.
15:32:44  <CIA-115>libuv: uv_udp_set_membership is not supported for UDP/IPv6 sockets yet. - http://git.io/1wzxpg
15:33:01  <piscisaureus_>bnoordhuis: I am giving up for now. I have 1 hour left to fix the stat regression.
15:33:09  <bnoordhuis>piscisaureus_: go for it
15:33:45  <piscisaureus_>bnoordhuis: if you feel like it you could try to fix udp for windows. the mcast-win branch is supposed to work but the test fails and I didn't try in node
15:34:24  * travis-cijoined
15:34:24  <travis-ci>[travis-ci] joyent/libuv#55 (mcast-win - 7bc6f3a : Bert Belder): The build failed.
15:34:24  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/commit/7bc6f3a
15:34:24  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/585054
15:34:24  * travis-cipart
15:42:27  * xaqjoined
16:13:12  * xaqquit (Remote host closed the connection)
16:32:29  * xaqjoined
17:07:12  * bradleymeckquit (Ping timeout: 240 seconds)
17:15:47  <CIA-115>libuv: Bert Belder stat-regress * rcf8038e / (src/win/fs.c test/test-fs.c test/test-list.h): Windows: fix stat("c:\\") regression - http://git.io/Dv_JVA
17:16:19  <piscisaureus_>isaacs: igorzi: ^-- can you review / test that? it should fix the stat('c:\\') regression in node
17:16:37  <piscisaureus_>ircretary: tell isaacs http://git.io/Dv_JVA
17:16:37  <ircretary>piscisaureus_: I'll be sure to tell isaacs
17:17:12  <piscisaureus_>ircretary: tell isaacs I have to run, won't be online tonight.
17:17:28  * travis-cijoined
17:17:28  <travis-ci>[travis-ci] joyent/libuv#56 (stat-regress - cf8038e : Bert Belder): The build failed.
17:17:28  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/commit/cf8038e
17:17:28  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/585507
17:17:28  * travis-cipart
17:17:44  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
17:20:04  * dapjoined
17:24:24  * orlandovftwjoined
17:32:56  * isaacsjoined
17:42:40  * dshaw_joined
18:13:39  * sh1mmerjoined
18:15:10  * perezdjoined
18:16:19  * TooTallNatejoined
18:25:55  * xaqquit (Remote host closed the connection)
18:28:10  * mikealquit (*.net *.split)
18:28:10  * paddybyersquit (*.net *.split)
18:28:10  * elijah-mbpquit (*.net *.split)
18:28:10  * wankdankerquit (*.net *.split)
18:28:10  * ljacksonquit (*.net *.split)
18:28:11  * igorziquit (*.net *.split)
18:28:19  * mikealjoined
18:28:31  * elijah-mbpjoined
18:28:36  * wankdankerjoined
18:28:37  * ljacksonjoined
18:29:26  * xaqjoined
18:46:27  * skabbesjoined
19:19:02  * mralephjoined
19:21:35  * brsonjoined
19:22:06  * mikealquit (Quit: Leaving.)
19:22:36  * mikealjoined
19:39:57  * mikealquit (Quit: Leaving.)
19:43:17  <isaacs>hello
19:47:53  * pieternjoined
19:49:17  * xaqquit (Remote host closed the connection)
19:50:18  * kuebkjoined
19:50:55  * kuebkquit (Client Quit)
19:58:58  <TooTallNate>isaacs: hello :)
20:35:47  * pieternquit (Quit: pietern)
20:39:32  <CIA-115>libuv: Ben Noordhuis master * r9db39bb / src/unix/freebsd.c : unix: don't use C99 comments in C89 source - http://git.io/uHthCQ
20:39:32  <CIA-115>libuv: Ben Noordhuis master * rceff85d / src/unix/freebsd.c :
20:39:32  <CIA-115>libuv: unix: fix compile error in freebsd.c
20:39:32  <CIA-115>libuv: Fixes joyent/node#2606. - http://git.io/VpvekA
20:39:32  <CIA-115>libuv: Ben Noordhuis master * rc89a75f / (src/unix/kqueue.c src/unix/tcp.c src/unix/udp.c):
20:39:33  <CIA-115>libuv: unix: fix compiler warning in kqueue.c, tcp.c, udp.c
20:39:33  <CIA-115>libuv: Include <unistd.h>, it contains the definition of close(). - http://git.io/_uQ6Zg
20:39:34  <CIA-115>libuv: Ben Noordhuis master * r9974825 / (config-unix.mk uv.gyp):
20:39:34  <CIA-115>libuv: unix: link to libkvm on freebsd
20:39:35  <CIA-115>libuv: Fixes 'undefined reference' linker errors to kvm_open(), kvm_close(), etc. - http://git.io/gPybcA
20:41:26  * travis-cijoined
20:41:26  <travis-ci>[travis-ci] joyent/libuv#57 (master - 9974825 : Ben Noordhuis): The build is still failing.
20:41:26  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/1cca230...9974825
20:41:26  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/586470
20:41:26  * travis-cipart
20:47:54  * AndreasMadsenjoined
20:55:06  <isaacs>bnoordhuis: you know if piscisaureus is around?
20:55:28  <bnoordhuis>isaacs: i do and he's not
20:56:19  <isaacs>bnoordhuis: i see. how do you feel about his changes re stat("c:\\") and windows dgram stuff?
20:56:41  <isaacs>looks like it fixes the issues on windows.
20:56:55  <bnoordhuis>isaacs: windows udp support isn't complete afaik
20:57:06  <isaacs>right, but it's not any *worse* off for these changes.
20:57:12  <isaacs>and not likely to get done today otherwise, anyway.
20:57:19  <bnoordhuis>probably not
20:57:26  <bnoordhuis>to both
20:57:38  <isaacs>i'd very much like to release 0.6.9. sounds like you can't see any reason not to, and there are no other objectors, so i'll move forward on it :)
20:57:41  <bnoordhuis>i think igor commented on bert's stat patch
20:57:51  * mikealjoined
20:57:56  <bnoordhuis>isaacs: https://github.com/joyent/libuv/commit/cf8038ef1d81acf5340a1f2c47d0b8d3fb6178a6#commitcomment-909433
20:58:09  <isaacs>ah, i see
20:58:28  <bnoordhuis>i can probably fix that
20:58:44  * bnoordhuisboots windows xp
20:58:54  <isaacs>if you could, that'd be great.
20:59:09  <isaacs>we haven't made promises about stat(c:\) working, but we have made promises about dgram landing.
20:59:25  <isaacs>so if it gets too tricky, we can push that out until 0.6.10
21:00:50  * mikealquit (Client Quit)
21:02:44  <bnoordhuis>hope you're not in a rush, compiling takes bloody forever with this vm
21:05:01  <isaacs>no problem, i'm about to go get some lunch
21:05:08  <isaacs>had to jump on a no.de issue.
21:05:18  <bnoordhuis>oh, was it broken again?
21:05:24  <isaacs>you know that scene in Godfather 3? Every time i get out, they pull me back in.
21:05:30  <bnoordhuis>heh
21:05:31  <isaacs>yeah, i broke it.
21:05:45  <isaacs>that's the thing, it's usually my fault, because i'm doing some patch for it, and not fully invested or focused.
21:05:49  <isaacs>wanting to get back ot other things.
21:05:58  <isaacs>so i feel guilty, and need to make it work again
21:06:51  <bnoordhuis>isn't the problem with no.de that nobody really wants to work on it?
21:06:58  <bnoordhuis>i always get / got that impression
21:06:58  <isaacs>that's one of the problem.s
21:07:07  <isaacs>the other problem is that PaaS is a retarded model.
21:07:17  <bnoordhuis>hah, don't tell your boss
21:07:19  <isaacs>IaaS is great. SaaS is great. PaaS is the worst of all possible worlds.
21:07:25  * mikealjoined
21:07:39  <isaacs>i think they know i feel that way: https://twitter.com/#!/izs/status/162742890852913155
21:07:46  <tjfontaine>you just like iaas because it looks like your name
21:07:52  <isaacs>tjfontaine: inorite?!
21:08:30  <isaacs>PaaS includes letters that are not in isaacs
21:08:35  <isaacs>that's the real problem.
21:09:26  <isaacs>but seriously, PaaS is one of those things like a Singleton. I mean, sure, you need to use it sometimes, maybe, but really, what you want is probably something else, and the fact that you're using it means that something else is lacking.
21:09:57  <isaacs>i've never seen a PaaS that didn't wed you to some kind of absurdly complicated black box.
21:10:00  <isaacs>usually, a single vendor
21:10:33  <isaacs>we just need better solutions for single-command deploy with a client-side tool. ie, the opposite of paas.
21:10:35  <tjfontaine>I'm too much of a control freak to like PaaS
21:10:42  <isaacs>tjfontaine: exactly
21:10:45  <isaacs>anywya, we're off topic.
21:10:50  <isaacs>should take this ranting to #node.js
21:10:53  <isaacs>:)
21:11:01  <tjfontaine>#node.js makes me so depressed :/
21:13:06  <CIA-115>node: isaacs v0.6 * rf98999c / (3 files in 2 dirs):
21:13:06  <CIA-115>node: Properly update npm to 1.1.0-3
21:13:06  <CIA-115>node: Thanks @mscdex for spotting this. - http://git.io/XjnoJA
21:15:54  <CIA-115>libuv: Bert Belder v0.6 * r92b260c / (src/win/fs.c test/test-fs.c test/test-list.h): windows: fix stat("c:\\") regression - http://git.io/8FrlsA
21:15:54  <CIA-115>libuv: Ben Noordhuis v0.6 * r3f1bad2 / test/test-fs.c : test: add missing return statement in fs_stat_root - http://git.io/qBd94g
21:17:39  * travis-cijoined
21:17:39  <travis-ci>[travis-ci] joyent/libuv#58 (v0.6 - 3f1bad2 : Ben Noordhuis): The build is still failing.
21:17:39  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/f925275...3f1bad2
21:17:39  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/586641
21:17:39  * travis-cipart
21:17:56  <isaacs>bnoordhuis: lgtm.
21:18:05  <isaacs>want to merge into node, or shall i?
21:18:21  <bnoordhuis>isaacs: i have the commit ready
21:18:25  <isaacs>sweet
21:18:44  <isaacs>go for it
21:18:51  <CIA-115>node: Ben Noordhuis v0.6 * r8a413b0 / (3 files in 2 dirs): uv: upgrade to 3f1bad2 - http://git.io/3idDeg
21:19:09  <isaacs>awesome
21:19:24  <isaacs>bnoordhuis: so, that's just the stat thing, but dgram on windows is still busted, yes?
21:19:41  <bnoordhuis>isaacs: that's correct
21:20:03  <isaacs>so, what's your opinion? better to half-fix it, or make windows dgram wait for 0.6.10?
21:20:36  <bnoordhuis>isaacs: i say let it wait
21:20:52  <isaacs>ok. that's reasonable.
21:21:12  <isaacs>i'll call it out in the changelog as replaced in osx and linux
21:21:20  <isaacs>er, in Unix, rather
21:21:25  <isaacs>since it's there now for solaris, too
21:21:41  * travis-cijoined
21:21:41  <travis-ci>[travis-ci] joyent/node#322 (v0.6 - f98999c : isaacs): The build is still failing.
21:21:41  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/ff0f0ae...f98999c
21:21:41  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/586627
21:21:41  * travis-cipart
21:24:34  * pieternjoined
21:25:32  <isaacs>bnoordhuis: can you take a look at https://github.com/joyent/libuv/pull/297?
21:26:03  <CIA-115>libuv: Ben Noordhuis master * r3de0411 / (9 files in 4 dirs): Merge remote-tracking branch 'origin/v0.6' (+149 more commits...) - http://git.io/B_CVLA
21:26:20  <bnoordhuis>isaacs: looked at and noted
21:26:52  <isaacs>k
21:27:21  <isaacs>he needs to sign the cla anyway
21:27:24  <isaacs>not today :)
21:27:36  * travis-cijoined
21:27:36  <travis-ci>[travis-ci] joyent/node#323 (v0.6 - 8a413b0 : Ben Noordhuis): The build is still failing.
21:27:36  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/f98999c...8a413b0
21:27:36  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/586654
21:27:36  * travis-cipart
21:27:51  <bnoordhuis>... why does the cia bot think there are another 149 commits in that last push?
21:27:56  * travis-cijoined
21:27:56  <travis-ci>[travis-ci] joyent/libuv#59 (master - 3de0411 : Ben Noordhuis): The build is still failing.
21:27:56  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/9974825...3de0411
21:27:56  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/586688
21:27:56  * travis-cipart
21:28:27  <isaacs>whoa... weird?
21:29:00  <bnoordhuis>pretty weird
21:29:05  <isaacs>bnoordhuis: because it's a merge commit maybe?
21:29:28  <bnoordhuis>hmm, github recognizes that it's only 11 commits
21:29:35  <isaacs>yeah, that's odd
21:29:45  <bnoordhuis>i blame the software
21:29:57  <isaacs>any other major libuv changes besides dgram and stat(C:\)?
21:30:24  <bnoordhuis>no, just lots of platform bug fixes
21:30:55  <bnoordhuis>oh, maybe better ansi support on windows
21:31:20  <isaacs>that's pretty minor, though. (in the sense that not a lot of people were screaming aboutit, afaik)
21:31:53  <bnoordhuis>in that case you're probably not interested in the uv_fs_stat performance improvements either :)
21:32:33  <CIA-115>node: Ben Noordhuis master * r14b7c43 / (14 files in 5 dirs): uv: upgrade to 3de0411 - http://git.io/Mh543g
21:34:31  <isaacs>bnoordhuis: there were mre?
21:34:55  <isaacs>bnoordhuis: i though that big change went in for 0.6.8
21:35:02  <bnoordhuis>isaacs: you're right
21:35:08  <bnoordhuis>i thought 0.6.8 was longer ago
21:35:10  <isaacs>:)
21:35:17  <isaacs>weekly releases are a workout.
21:41:17  * dshaw_quit (Read error: Connection reset by peer)
21:43:29  <CIA-115>node: isaacs v0.6 * r3865152 / test/simple/test-tty-stdout-end.js : stdout ending test message correction - http://git.io/skL-9w
21:46:24  * travis-cijoined
21:46:24  <travis-ci>[travis-ci] joyent/node#324 (master - 14b7c43 : Ben Noordhuis): The build is still failing.
21:46:24  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/85a86b5...14b7c43
21:46:24  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/586716
21:46:24  * travis-cipart
21:47:59  <isaacs>test, please: http://nodejs.org/dist/v0.6.9-RC0/
21:48:20  <isaacs>i'm about to start building/testing on windows, mac looks good already
21:48:33  <isaacs>if you want to check linux, would be helpful
21:51:44  * travis-cijoined
21:51:44  <travis-ci>[travis-ci] joyent/node#325 (v0.6 - 3865152 : isaacs): The build is still failing.
21:51:44  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/8a413b0...3865152
21:51:44  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/586757
21:51:44  * travis-cipart
21:57:02  <isaacs>ah, forgot to update the version properly.
21:57:08  <isaacs>rc1 coming right up :)
22:03:06  * sh1mmerquit (Quit: sh1mmer)
22:04:23  <isaacs>test, please: http://nodejs.org/dist/v0.6.9-RC1/
22:04:46  <AvianFlu>you got that out before RC0 was done compiling for me >.<
22:04:50  <AvianFlu>still running the tests
22:05:42  <isaacs>:)
22:06:38  <AvianFlu>https://gist.github.com/7492a3d805310f238ded
22:06:40  <AvianFlu>two failures
22:08:17  * bnoordhuistests
22:08:18  <wankdanker>wish i had a debian machine to work out why the multicast and broadcast tests are reporting failures
22:08:24  <bnoordhuis>isaacs: i can test linux and solaris
22:08:41  <bnoordhuis>wankdanker: how/where?
22:08:55  <wankdanker>https://gist.github.com/7492a3d805310f238ded
22:09:14  <isaacs>AvianFlu: what os?
22:09:23  <wankdanker>description says debian 6.02
22:09:40  <isaacs>oh, right :)
22:09:43  <wankdanker>from the output i can see that the child processes are receiving the responses
22:10:01  <wankdanker>the child processes then use process.send to report back to their parent
22:10:03  <AvianFlu>kernel is 2.6.32-5
22:10:05  <wankdanker>it's almost as if that is failing.
22:11:01  <isaacs>yeah, the actual dgram sending looks like it's wokring.
22:11:10  <isaacs>which is good :)
22:11:18  <wankdanker>rc1 on linux 3.0.0-13-generic: [00:55|% 100|+ 327|- 0]: Done
22:12:03  <bnoordhuis>you people with your fast computers, you're making me feel insecure
22:12:09  <isaacs>haha
22:12:25  <wankdanker>i noticed that the test failed this same way on the travis-ci last night
22:12:29  <wankdanker>size does matter
22:17:06  * mikealquit (Quit: Leaving.)
22:17:28  <bnoordhuis>multicast ttl is borked on sunos
22:22:30  * mralephquit (Quit: Leaving.)
22:23:46  <isaacs>yeah, hearing the same thing from some joyent folks
22:24:10  <isaacs>similar to debian, it looks like.
22:24:16  <isaacs>kids get the message, but process.send fails.
22:24:25  <isaacs>also breaks some cluster tests.
22:24:36  <isaacs>http://pastie.org/3266029
22:25:07  <isaacs>er, no, there's another thing, too: Error: setMulticastTTL EINVAL
22:25:12  <isaacs>yeah, that's borked for sure.
22:25:40  <wankdanker>does process.send block?
22:25:59  <AndreasMadsen>wankdanker: yes
22:26:04  <AndreasMadsen>:(
22:26:18  <wankdanker>AndreasMadsen: ty
22:27:25  <AndreasMadsen>^-- https://github.com/joyent/node/issues/2598
22:28:32  <bnoordhuis>[14:13|% 100|+ 701|- 7]: Done <- SunOS 5.11 joyent_20110802T190624Z
22:28:38  <bnoordhuis>a.k.a. no.de
22:29:34  <bnoordhuis>nothing unexpected except for simple/test-dgram-multicast-multi-process
22:32:26  <isaacs>bnoordhuis: want to try taking a look at multicast on sunos?
22:32:30  <bnoordhuis>isaacs: sure
22:32:34  <bnoordhuis>[16:01|% 100|+ 706|- 2]: Done <- ubuntu 10.04
22:32:34  <isaacs>k
22:32:46  * isaacsneeds food desperately. will be back in a bit.
22:33:30  <bnoordhuis>pummel/test-timers and pummel/test-keep-alive failing on account of my system being overloaded
22:39:31  <AvianFlu>I ran the tests again, they all passed; a third time, and the test-dgram-{broadcast,multicast}-multi-process tests both failed the same way - race condition?
22:41:09  <AvianFlu>earlier gist updated
22:52:01  <wankdanker>i have an idea what might be going on in that test if it is a race condition.
22:52:48  <wankdanker>i think that the child process is exiting before the parent process receives the last message sent from the child.
22:58:02  <wankdanker>AvianFlu: can you run this test a couple times to see if it is more consistent? https://gist.github.com/1691411
22:58:13  <AvianFlu>yeah, sure
22:59:18  <TooTallNate>bnoordhuis: quick little pull: https://github.com/joyent/node/pull/2626
22:59:28  <TooTallNate>since the guy never submitted one back in october
22:59:50  <TooTallNate>which is a bummer; could have been in 0.6.x already :\
23:01:43  * mikealjoined
23:06:34  * AndreasMadsenquit (Remote host closed the connection)
23:10:05  <AvianFlu>wankdanker, I'm still seeing it fail, but less often than the other one does
23:10:06  <AvianFlu>https://gist.github.com/bfcdcda65308665c97a8
23:10:24  <AvianFlu>it fails a little differently than the other - no worker death messages it looks like
23:11:25  <wankdanker>thanks AvianFlu. very interesting.
23:14:24  <bnoordhuis>isaacs: https://github.com/bnoordhuis/libuv/commit/16124bb
23:14:56  <isaacs>wait, it needs the arg to be a char?
23:15:14  <isaacs>sunos, why you gotta be different all the time?
23:15:32  <bnoordhuis>i know right?
23:16:05  <isaacs>business as usual, i guess.
23:16:06  <bnoordhuis>anyway, want me to merge it and upgrade libuv (again)?
23:16:06  <isaacs>lgtm
23:16:10  <isaacs>yes, please.
23:16:12  * isaacssigh
23:16:35  <CIA-115>libuv: Ben Noordhuis v0.6 * r16124bb / src/unix/udp.c :
23:16:35  <CIA-115>libuv: sunos: fix uv_udp_set_ttl and uv_udp_multicast_ttl
23:16:35  <CIA-115>libuv: The argument to setsockopt(IP_TTL|IP_MULTICAST_TTL) should be a char, not an int
23:16:35  <CIA-115>libuv: like on other Unices. - http://git.io/8xdSkQ
23:17:30  <isaacs>we need a better CI sytem.
23:17:39  <isaacs>every push should run every test on every OS.
23:18:15  * travis-cijoined
23:18:15  <travis-ci>[travis-ci] joyent/libuv#60 (v0.6 - 16124bb : Ben Noordhuis): The build is still failing.
23:18:15  <travis-ci>[travis-ci] Change view : https://github.com/joyent/libuv/compare/3f1bad2...16124bb
23:18:15  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/libuv/builds/587091
23:18:15  * travis-cipart
23:18:35  <CIA-115>node: Ben Noordhuis v0.6 * r22e2347 / deps/uv/src/unix/udp.c : uv: upgrade to 16124bb - http://git.io/kKt4hA
23:18:52  <bnoordhuis>isaacs: i imagine joyent has a couple of spare servers
23:19:02  <isaacs>bnoordhuis: yeah, it's not hte lack of servers.
23:19:07  <isaacs>it's the lack of time to set them all up
23:19:09  <bnoordhuis>i think tom was going to work on that
23:19:15  <bnoordhuis>until he got fired, that is
23:19:20  <isaacs>he's putting jetpacks on dinosaurs now.
23:19:33  <bnoordhuis>god's work
23:19:37  <isaacs>indeed.
23:21:04  <bnoordhuis>TooTallNate: whose code is that? yours or axel's?
23:21:21  <TooTallNate>bnoordhuis: mine technically
23:21:29  <bnoordhuis>technically?
23:21:55  <TooTallNate>i looked at his proposed patch and copied it
23:22:00  <TooTallNate>idk the technicality of that :p
23:22:12  <TooTallNate>but there's some whitespace differences for whatever that matters
23:22:18  <isaacs>TooTallNate: it doesn't.
23:22:22  <isaacs>axel has to sign the CLA to accept it.
23:22:36  <isaacs>(or whoever wrote whatever teh change is)
23:22:49  <TooTallNate>well how would i make that "mine"?
23:23:03  <isaacs>TooTallNate: you'd have to never have seen his patch :)
23:23:58  <TooTallNate>well then I didn't :)
23:24:12  <TooTallNate>seriously though, it's a 1-liner, how "diverse" can that get?
23:24:21  <TooTallNate>should we change the key name or something?
23:25:00  <isaacs>TooTallNate: i know, it's annoying.
23:25:18  <isaacs>but if you describe it to someone else, and then have never seen the implementation, and end up doing the same thing, then that's "clean"
23:25:38  <TooTallNate>well that's how I would consider this case
23:25:39  <isaacs>because IP law has allowances for two people having the same idea.
23:25:45  <isaacs>right... but.. it's not.
23:26:03  <TooTallNate>well then what's the solution?
23:26:14  <isaacs>TooTallNate: what's the thing we're talking about?
23:26:15  <wankdanker>i'll submit it
23:26:21  <wankdanker>someone describe what i should do
23:26:25  <TooTallNate>and, side note, what's the difference in this case?
23:26:25  <wankdanker>:)
23:26:32  * travis-cijoined
23:26:32  <travis-ci>[travis-ci] joyent/node#326 (v0.6 - 22e2347 : Ben Noordhuis): The build is still failing.
23:26:32  <travis-ci>[travis-ci] Change view : https://github.com/joyent/node/compare/3865152...22e2347
23:26:32  <travis-ci>[travis-ci] Build details : http://travis-ci.org/joyent/node/builds/587098
23:26:32  * travis-cipart
23:26:46  <isaacs>TooTallNate: if you describe the bug or whatever, and someone else does it, they haven't been "tainted" by having seen the original IP.
23:27:02  <isaacs>you and i and everyone else know that sometimes, that's just silly
23:27:07  <isaacs>but lawyers are not reasonable.
23:27:13  <isaacs>(or rather, laws are not reasonable)
23:27:29  <TooTallNate>wankdanker: the "tty" module should emit "keypress" events even for unidentified keys
23:27:39  <isaacs>lawyers are like programmers, and the law is the computer. you can argue all you want that the arg to set_ttl should be an int and not a char, but you won't convince sunos
23:27:51  <bnoordhuis>it's on moments like these that i appreciate living in europe
23:28:01  <TooTallNate>lol, nice one isaacs
23:28:03  * mikealquit (Quit: Leaving.)
23:28:21  <wankdanker>TooTallNate: ok. let me take a look and see what clever solution i can come up with.
23:28:25  <TooTallNate>wankdanker: i guess that's the only description i can give without making you tainted ;)
23:28:28  <isaacs>bnoordhuis: are eu's IP laws less retardeed?
23:28:38  <wankdanker>haha
23:28:42  <bnoordhuis>isaacs: yes. that is, still retarded but less so
23:28:47  <TooTallNate>i want my damn commit count++ though :(
23:29:00  <isaacs>wankdanker: there's a big switch in the tty module that looks at the key value and decides what type of key it is.
23:29:19  <isaacs>i'm guessing you just need to add a default or some thing.
23:29:25  <isaacs>haven't seen that code in a long time, though
23:31:17  <isaacs>test, please: http://nodejs.org/dist/v0.6.9-RC2/
23:33:52  <wankdanker>so it looks like the issue is that there is no default in the switch
23:34:36  <bnoordhuis>wankdanker: correct
23:35:12  <TooTallNate>ugh, this is hella silly
23:35:20  <isaacs>hahahahh
23:35:27  <bnoordhuis>at least you'll have something to tweet about
23:35:46  <TooTallNate>all cause i said "technically" :p
23:36:08  <isaacs>TooTallNate: you're doing the right thing
23:36:08  <bnoordhuis>you live, you learn
23:36:19  <bnoordhuis>i mean, what isaacs said
23:36:30  <isaacs>node is used by big companies that are highly sueable, and part of what we offer them is assurance that node is "clean" from an IP pov.
23:36:31  <TooTallNate>hahaha
23:38:07  <isaacs>i knew a guy once who wrote code for hardware drivers, and he was a former lawyer. he told me, if you think operating systems are awful to work on, go to law school.
23:38:46  <TooTallNate>ya man, not a field I want to venture down :p
23:38:49  <TooTallNate>too much paperwork
23:38:52  <isaacs>srsly
23:39:01  <bnoordhuis>there are girls singing outside my window
23:39:06  <bnoordhuis>i must be doing something right
23:39:35  <wankdanker>so, since "s" is an undefined key i assume that we should append it to the key object for later inspection
23:39:47  <wankdanker>anyone care what this attribute is called? or is there already something defined?
23:39:52  <wankdanker>(i don't use tty often)
23:40:09  <wankdanker>(ever)
23:40:20  <TooTallNate>there's not one currently defined, so you can pick something
23:40:43  <isaacs>key.code like in the browser, maybe?
23:40:51  <isaacs>or .keyCode
23:40:54  * isaacsshrug
23:40:55  * dshaw_joined
23:41:14  <TooTallNate>I like key.code
23:41:45  <isaacs>oh, i see, you already have a code... yeah, they should probably all get key.code to be the thing that the switch is keying off of
23:42:01  <isaacs>like, put that up at the top
23:42:07  <isaacs>this is in master, right?
23:42:25  <TooTallNate>ya
23:44:11  <isaacs>wankdanker: so, on line 156, after the sequence has been made into a string, I'd add key.sequence = s
23:44:31  <isaacs>wankdanker: then on line 215, before going into the switch, key.code = code
23:44:59  <isaacs>it's not quite the same as keyCode on the browser, since that's a number.
23:45:00  <isaacs>but whatever.
23:45:28  <wankdanker>got it.
23:45:31  <isaacs>then for unknown ones, you can just set key.name = 'unknown'
23:45:33  <wankdanker>do you want this on master?
23:45:38  <isaacs>and you'll be able to dig into it if you care.
23:45:42  <isaacs>yes, this is a master thing
23:45:46  <wankdanker>k
23:45:46  <isaacs>it changes api
23:45:53  <isaacs>thanks :)
23:45:54  <wankdanker>10-4
23:46:27  <isaacs>oh, looks like we explicitly prefer not to emit on unknown keys...
23:46:29  <isaacs> // Don't emit a key if no name was found
23:46:30  <isaacs> if (key.name === undefined) {
23:46:30  <isaacs> key = undefined;
23:46:31  <isaacs> }
23:46:52  <TooTallNate>isaacs: so is the difference here that you've already signed the cla, so wankdanker can use your IP?
23:47:03  <isaacs>TooTallNate: yessir
23:47:11  <TooTallNate>ok
23:47:20  <isaacs>it's retarded, i know, i know
23:47:37  <isaacs>wankdanker: please give @TooTallNate props in the commit message :)
23:47:45  <TooTallNate>:p
23:47:47  <wankdanker>:) will do.
23:48:21  <isaacs>oh, ok, so you can still get name === undefined in some cases.
23:48:26  <isaacs>nevermind my objection before.
23:48:31  <wankdanker>ok.
23:48:49  <isaacs>this will just make it handle unknown functionKeyCodeRe key sequences
23:48:54  <isaacs>thanks
23:53:08  <bnoordhuis>Error: setTTL EINVAL <- whaaat?
23:53:36  <isaacs>new record: compiling and testing on 4 vm's at once.
23:53:42  <isaacs>(if you include os x)
23:54:01  <TooTallNate>isaacs: what VM software you use?
23:54:02  <isaacs>i'm guessing most errors will be due to timeouts
23:54:04  <isaacs>vmwar
23:54:07  <isaacs>*vmware
23:54:29  <TooTallNate>:\ geez, my MBP slows to a crawl with just 1 vm running!
23:55:00  <isaacs>TooTallNate: i've got a new laptop with an i7
23:55:09  <isaacs>TooTallNate: if you haev a core2duo, it's too terrible
23:55:23  <TooTallNate>i have the i7
23:55:31  <TooTallNate>an older one though
23:56:50  <isaacs>bnoordhuis: you're getting an EINVAL? whereabouts?
23:57:00  <bnoordhuis>on solaris again...
23:58:19  <wankdanker>https://github.com/joyent/node/pull/2627