Index

Show enters and exits. Hide enters and exits.

00:03:38crafterm leaves the room.
00:04:31manverudrbrain: http://p.ramaze.net/639
00:05:05manveruwith 1.8.6 i could've iterated the whole IPv4 space, but rubinius is still too slow to finish within a month :|
00:06:50drbrainmanveru: the truncation is probably a separate unrelated issue
00:06:57drbrainbut, I wonder what's up with our getaddrinfo() call
00:07:10drbrainmaybe we've got a flag mis-set?
00:07:38Vagabondmanveru: IPAddr.new('192.168.0.0') works for me on freebsd now
00:08:20manveruheh
00:08:50manveruVagabond: it works here too as long as i run nscd
00:08:58Vagabondncsd?
00:09:22manveruhttp://linux.about.com/cs/linux101/g/nscd.htm
00:10:27drbrainwell, I'm going to watch a movie or something while my sistem flushes anticoagulant
00:10:55VagabondI see
00:11:03manveruwell, good luck with that :)
00:11:35drbrainI gave platelets this morning
00:11:45drbrainusually I can't taste it after the third return
00:11:55drbrainbut today is different for some reason
00:13:33manverujust learned 'anticoagulant' and 'platelets'
00:13:59Vagabondmanveru! platelet attack, now!
00:15:19manveruoO
00:15:28manveruno, work, now :|
00:15:33manverucu later guys
00:15:37Vagabondlater
00:15:50crafterm enters the room.
00:21:00ezmobius_ enters the room.
00:22:07_ADSI am having a really hard time understanding git
00:22:21_ezmob_ enters the room.
00:22:42_ADShow do I revert everything except 1 file to the HEAD
00:23:25_ADSor even revert everything - I can redo the changes..
00:24:04ctennis_ADS: git reset --hard HEAD
00:24:25ctennis_ADS: that only changes files that git already knows about. if you want to get rid of other files, use git-clean
00:26:20crossblaim enters the room.
00:27:45AndrewO enters the room.
00:34:33defunkt enters the room.
00:38:16rueYep, `checkout` is usually good for that. It is recursive, though
00:41:07hornbeck enters the room.
00:44:40KirinDav leaves the room.
00:48:49seydar enters the room.
00:49:56seydardrbrain: i think i got most of syslog.rb done. my problems lie with understanding syslog method and using it.
00:54:12seydarvoila: http://pastie.caboo.se/156778
00:56:09Vagabondseydar: nice job
00:56:20seydarbut its not finished yet!
00:56:36seydari hope it gets accepted, because i had a paper to write...
00:57:03seydari want to write rubinius bindings to LLVM
00:57:23seydarbecause FFI is sweet, and so is LLVM, and i think i'm a lot better of a coder than I actually am
00:59:32Vagabondllvm?
01:00:06tarcieriseydar: you need portable explicitly managed stack frames
01:00:28seydarfor llvm? i should hope not syslog.rb!
01:00:36tarcieriFor LLVM
01:00:48seydaroh thats better.
01:01:05seydarand that sounds complicated and difficult, reasoning why it hasn't been done before?
01:01:11tarcieriPyPy manages without it because Python doesn't have real closures
01:01:54seydarhow do they have "fake" closures?
01:02:20tarcieriYou can't mutate closed over variables, iiuc
01:03:13tarcierihttp://www.nondot.org/sabre/LLVMNotes/ExplicitlyManagedStackFrames.txt
01:03:22tarcierithere's that approach, but evan says it isn't portable...
01:03:22seydaris that often done in Ruby? i have no clue what that means
01:04:23tarcieriwell, I use them a lot
01:05:12seydarah
01:05:48tarcieria trivial example:
01:05:49tarcierifoo = 42; proc { foo = 69 }.call; p foo
01:06:39seydarooooh. ruby 1.9 gets rid of those. so why not adhear to 1.9 specs?
01:06:51tarcieriRuby 1.9 doesn't change that
01:06:56tarcieriIt changes how block arguments work
01:07:18tarcieriIf that proc had an argument |foo|, it'd warn you that it was shadowing the local
01:07:26tarcieriAnd it wouldn't mutate the local
01:07:55tarcieriBut if you explicitly mutate locals inside a lambda, they stay mutated...
01:07:56kevwil leaves the room.
01:08:15seydarah
01:08:20tarcieriThe lambda "closes over" those locals
01:08:23seydarkompreneble
01:08:40manveruand back
01:09:50drbrainseydar: the arguments for Syslog#open are all optional
01:10:22seydardrbrain: whoops. had that there originally
01:10:31seydarhow should I do log?
01:11:24drbrainit looks the same as printf()
01:11:35drbrainso, def log(priority, format, *args)
01:11:37seydarkk
01:11:45drbrainor even
01:11:54drbraindef log(priority, *args)
01:12:09drbrainbut, the former will make argument checking free
01:12:16drbrainI'm not sure how to FFI call that though
01:12:23drbrainis there a string_array type?
01:12:52drbrainor a pointer array?
01:15:25cuppr leaves the room.
01:15:38seydari can splat it
01:15:39drbrainseydar: if you can't get it to work, just raise NotImplementedError and leave a comment
01:15:47seydark
01:16:57seydari'll just go with NotImplementedError because I phale
01:17:44drbrainI'm going back to watching my movie in a second, but submit that as a patch and I'll give it a full review
01:18:16_ADS leaves the room.
01:18:38obvio leaves the room.
01:19:35seydarhurray!
01:21:16seydaris it wrong to use instance eval?
01:21:27seydari'll guess yes
01:21:46drbrainsometimes it is wrong, sometimes it is right
01:21:57drbrainheads back to his movie
01:22:30Vagabondit's often wrong though
01:23:07tarcieriabuses instance_eval quite a bit sometimes :/
01:24:14VagabondI try to avoid it wherever I can
01:24:22seydarit was for method generation from an array. it seemed appropriate
01:24:43Vagabondwhy not define_method?
01:25:06seydarbecause i didn't know how to say "i want it to take 2 args"
01:27:07seydarhuzzah! i'm set now
01:27:16Vagabonddefine_method(:foo) {|arg1,arg2| .. } ?
01:28:50crafterm leaves the room.
01:29:46crafterm enters the room.
01:29:49crafterm leaves the room.
01:34:40hornbeck leaves the room.
01:38:56seydarfork. i totatlly forgot about that
01:39:26seydaranyways, drbrain i posted the patch. so.... i'm in with the in-crowd now
01:39:29perdix leaves the room.
01:42:52wifelette leaves the room.
01:50:05seydar leaves the room.
01:57:35boyscout1 commit by Charles Nutter
01:57:36boyscout * Add a spec for Enumerable#inject with a *arg; JRUBY-2162 exposed it.; d87df0b
02:04:14d2dchat leaves the room.
02:07:08crossblaim leaves the room.
02:07:38KirinDav enters the room.
02:09:04rubuildius_ppcCharles Nutter: d87df0b76; 1616 files, 5542 examples, 19080 expectations, 0 failures, 12 errors; http://pastie.caboo.se/paste/156800
02:10:52rubuildiusCharles Nutter: d87df0b76; 1616 files, 5541 examples, 19088 expectations, 0 failures, 0 errors; http://rafb.net/p/9nYbK254.html
02:11:48hornbeck enters the room.
02:13:09cored leaves the room.
02:14:34perdix enters the room.
02:16:25imajes_ leaves the room.
02:19:06lopex leaves the room.
02:28:54AndrewO_ enters the room.
02:35:47AndrewO leaves the room.
02:47:17VVSiz_ enters the room.
02:51:08_mutle enters the room.
02:56:09d2dchat enters the room.
03:02:06d2dchat leaves the room.
03:05:24VVSiz leaves the room.
03:07:07mutle leaves the room.
03:10:20_mutle leaves the room.
03:10:24mutle enters the room.
03:14:31crafterm enters the room.
03:15:08boyscout1 commit by Nikolai Lugovoi
03:15:09boyscout * Fixes for String#to_sub_replacement:; 9f3e252
03:16:24imajes leaves the room.
03:19:53ezmobius leaves the room.
03:19:54ezmobius_ leaves the room.
03:19:54_ezmob_ leaves the room.
03:25:54d2dchat enters the room.
03:28:42rubuildius_ppcNikolai Lugovoi: 9f3e25289; 1616 files, 5545 examples, 19083 expectations, 0 failures, 12 errors; http://pastie.caboo.se/paste/156818
03:30:56rubuildiusNikolai Lugovoi: 9f3e25289; 1616 files, 5544 examples, 19091 expectations, 0 failures, 0 errors; http://rafb.net/p/L2Lj7G54.html
03:37:33mernen leaves the room.
03:43:46rubyconsumer enters the room.
03:44:21headius leaves the room.
03:44:54headius enters the room.
03:50:15mae enters the room.
04:00:34mae leaves the room.
04:20:34brweber2_ enters the room.
04:25:02boyscout1 commit by Charles Nutter
04:25:03boyscout * Add a second case for truncating IO buffers that specifies too-small size.; f8e6200
04:27:40brweber2_ leaves the room.
04:27:40headius leaves the room.
04:30:30headius enters the room.
04:32:29KirinDave leaves the room.
04:33:45rubuildius_ppcCharles Nutter: f8e620027; 1616 files, 5545 examples, 19084 expectations, 0 failures, 12 errors; http://pastie.caboo.se/paste/156831
04:36:01rubuildiusCharles Nutter: f8e620027; 1616 files, 5544 examples, 19092 expectations, 0 failures, 0 errors; http://rafb.net/p/e8Y7yh75.html
04:57:15srbaker_ enters the room.
05:00:50hoopy leaves the room.
05:03:49kwatch leaves the room.
05:10:39ruePhew
05:10:58rueNext week I will actually be at the computer instead of the den
05:11:29rueevan: What is your seat-of-pants on just the VM performance in comparison to 1.9?
05:11:41srbaker leaves the room.
05:12:17rubyconsumer leaves the room.
05:14:33d2dchat leaves the room.
05:26:45benburkert leaves the room.
05:31:29wifelette enters the room.
05:38:28AndrewO_ leaves the room.
05:40:59RyanTM leaves the room.
05:41:54hoopy enters the room.
05:42:25dodecaphonic leaves the room.
06:12:04defunkt enters the room.
06:20:50_ADS_ enters the room.
06:22:05headiushttp://headius.blogspot.com/2008/02/rubys-threadraise-threadkill-timeoutrb.html
06:22:11headiusmost of you know this already
06:24:06evanheadius: btw, you might consider pulling in my new timeout.rb to jruby
06:24:16headiusyeah, I was going to have a look at it
06:24:18evanit removes half the insanity of of it
06:24:28evanie, there is only one watcher thread
06:24:30headiusI had an impl that used a thread-local variable to store a timeout thread for future uses
06:24:32evannot one per call
06:24:40headiusone watcher wouldn't work for us
06:24:45headiusbut one per calling thread would
06:24:56evanah, true.
06:25:04evanyou could make a pretty easy adjustment to do that
06:25:26headiusI may reimplement it in Java with just a Java thread and Java concurrency primitives, which are a lot saner than Ruby's
06:25:32evanalso, while we haven't extensively tested
06:25:41headiuseither way it's not safe though, so I'm starting the battle
06:25:45evanwe've got Thread#kill just raising a Die exception in the target thread
06:25:51evanwhich removes some of the madness of #kill
06:25:59evanie, runs ensure blocks, etc.
06:26:11headiuskill and raise are essentially identical for us anyway, since we have to unroll Java stack
06:26:15headiusThreadKill
06:26:24headiusgets raised for kill
06:26:28headiusunrescuable
06:26:49evanah
06:26:51evanyeah, very similar
06:27:00evancurrently you can rescue Thread::Die
06:27:22headiusyeah, you'll need to find a way to fix that
06:27:30headiusrescue Exception would produce an unkillable thread
06:27:48evanyep
06:28:00headiusnot that threads ought to be killable in the first place, of course :)
06:28:08evanyeah
06:29:07agardiner leaves the room.
06:29:17evanyou make a very good point about the fact that no one rescues Timeout::Error when using net/http
06:30:22evando you use checked exception in java much?
06:30:33evani was reading an article that called them a failed experiment
06:32:56rueheadius: Hm, no killable threads? No suspendable threads, sure
06:33:33headiusevan: we generally are forced to use them
06:33:39headiusI go back and forth on whether they're a good thing
06:34:01headiusmostly they're an irritation, but it's much easier to build reliable systems with them if you use them right
06:34:23headiustrouble is, they're so cumbersome most people don't use them right
06:34:28headiusheadius: eh?
06:34:44headiuser
06:34:47headiusrue: eh?
06:35:30evanheadius: right, the article talked about their cumbersome nature
06:35:38rueheadius: You are opposed to killable threads?
06:36:00rueChecked exceptions are a programmer usability nightmare. I am sure they meant well though
06:36:02headiusabsolutely
06:36:17headiusyou *can not* safely kill a parallel thread
06:36:37headiusunless you intend to bring down the whole process
06:36:38evanthats kinda why i like ours
06:36:44evanit being a nice request
06:36:48evan'could you please die?'
06:36:53evanand it can just say no.
06:36:59headiusyeah, that's how ours work...that's how they had to work
06:37:14headiusdamn, I didn't even get into the fact that kill and raise and timeout don't work when calling native libs either
06:37:21headiussomeone add a comment :)
06:37:40evani do wonder how many bizarre ruby problems are from the use of timeout and thus Thread#kill relating to Mutex#syncronize and ensure
06:37:58headiusI know we have to constantly battle such issues
06:38:07evanoh bingo
06:38:12evanjust got a great idea
06:38:14headiusand we incur a bunch of overhead to make threads constantly check for events
06:38:29evanclass Thread::Die; def self.===(o); false; end; end
06:38:32evanyay! uncatchable!
06:38:33evan:D
06:38:42headiushah, nice
06:40:46_ADS enters the room.
06:49:20_ADSrue: are you there?
06:51:07cuppr enters the room.
06:52:27_ADSIt seems like the ticket I submitted a spec for got closed before the fix
06:52:39_ADSI have the fix now, should I attatch it anyway?
06:53:21brixen_ADS: what #?
06:53:32_ADSbug #333
06:54:32brixen_ADS: sure, attach it, I'll take a look
06:55:38_ADSbrixen: ok, done
06:58:00brixen_ADS: I've reopened #333 and assigned it to evan
06:58:07wifelette leaves the room.
06:58:14_ADSbrixen: thanks
06:58:23brixennp
07:12:53wifelette enters the room.
07:13:29wifelette leaves the room.
07:14:06wifelette enters the room.
07:17:59crafterm leaves the room.
07:19:53cris_kiev enters the room.
07:20:09imajes enters the room.
07:29:58dbussinkmorning
07:31:22brixenmorning
07:32:21dbussinkbrixen: been doing some lighthouse cleanup i see
07:37:04brixenjust a bit
07:40:48wycats leaves the room.
07:41:37wycats enters the room.
07:42:07dbussinkevery bit helps :)
07:42:25dbussinkbrixen: btw, you've been working on some 0.0 / -0.0 issues?
07:43:21brixendbussink: last I looked, seemed that libc version was the issue
07:43:33brixenit's not anything rbx is doing afaics
07:43:36dbussinkyeah, well, openbsd suffers from the same problem
07:43:47brixenthere's a ticket on it, one sec..
07:43:49dbussinkand it's a strtod problem there too
07:43:58dbussinkyeah, but that ticket was closed
07:44:07dbussinkbut mri exposes the exact same behavior
07:44:08wycats leaves the room.
07:44:32brixen#307
07:44:36brixenstill open
07:44:59dbussinkah yeah
07:45:05brixenI was going to check this on ubuntu, but not sure I feel like upgrading libc from source
07:45:09brixenon gutsy that is
07:45:25dbussinkbut if we mimic mri behavior, it's actually not a problem
07:47:00brixenwell, not sure what you mean by mimic. you mean use whatever code mri uses?
07:47:18dbussinki mean that mri has the same 0.0 / -0.0 problem
07:47:21brixenthe idea of passing the specs is mimicking ;)
07:47:24dbussinkat least on openbsd that is
07:47:28brixennot on gutsy
07:47:33brixenrbx doesn't do the same as mri
07:47:33dbussinkthat's weird
07:47:59brixenhowever, I need to install 1.8.6p111 on gutsy
07:48:02brixenle'me try with that
07:48:39dbussinkah wait, openbsd has a bit older ruby, maybe they implemented a work-around
07:56:43wifelette leaves the room.
07:58:10be9 enters the room.
08:00:13imajes leaves the room.
08:03:54_ADS leaves the room.
08:15:04perdiy enters the room.
08:17:32evanbrixen: i wouldn't try upgrading libc from source on any linux
08:17:47evanwhile you might get it to work, if you're testing things, you'll have a system NO ONE else has.
08:19:25brixenevan: yeah
08:20:04dbussinkbrixen: hmm, one of the socket specs depends on an assumption about /etc/hosts that is not always true
08:20:20dbussinkthat there is a 127.0.0.1 localhost definition
08:20:24brixendbussink: feel free to fix it ;)
08:20:38dbussinkit's nasty :)
08:20:51brixenI can imagine
08:21:25brixendbussink: the thing with libc is probably a red herring anyway, since MRI shows one behavior on gutsy and rbx another
08:22:19dbussinkbrixen: strange, because on openbsd it definitely is a problem with libc, wrote a tiny program to verify it
08:22:39brixenbut on openbsd, mri behaves the same as rbx, yes?
08:23:06dbussinkyeah
08:23:13dbussinkbut that means it fails the 0.0 spec
08:23:19brixenok, still waiting for p111 to compile on gutsy
08:23:24defunkt enters the room.
08:23:35dbussinkbut it's a older version of ruby, have to compile p111 still to verify
08:23:40dbussinkbrixen: http://pastie.caboo.se/156881
08:23:54dbussinkgives three times 0.0 on openbsd, never a -0.0
08:25:35defunkt_ enters the room.
08:25:42defunkt leaves the room.
08:25:53maduyb__ enters the room.
08:26:00drbrainI get -0.0, 0.0, -0.0 on FreeBSD 6.2
08:26:13drbrainso, I bet open's libc is different
08:27:16dbussinkdrbrain: yeah, that's what os x gives too
08:27:38brixendbussink: I get -, +, - on gutsy with libc 2.6.1
08:27:47dbussinkbrixen: hmm, that is ok
08:28:00dbussinkbrixen: so that means there is also something broken in rubinius
08:28:03brixendbussink: yeah, but rbx fails
08:28:05brixenyeah
08:28:08drbrainopenbsd has a completely different manpage for strtod than freebsd
08:28:19drbrainso, different implemenation
08:28:33dbussinkdrbrain: well, i didn't write the infinity / nan workaround for nothing ;)
08:29:09drbraininteresting
08:29:22drbrainopen: The strtod() function conforms to ANSI X3.159-1989 (``ANSI C'').
08:29:42drbrainfree: The strtod() function conforms to ISO/IEC 9899:1999 (``ISO C99''), with the exception of the bug noted below.
08:29:45perdix leaves the room.
08:29:46drbrainos x says the same
08:30:05drbrainso, that's your problem
08:31:01dbussinkyeah, but we're having two different problems now
08:32:21drbrainMRI doesn't use libc strtod, it seems
08:32:25drbrainsee util.c
08:33:01drbrainutil.h:#define strtod(s,e) ruby_strtod(s,e)
08:33:16dbussinkyeah i see, they wrote their own
08:33:24manveruoi
08:33:29drbrainno, imported Sun's
08:34:06dbussinkah yeah, i found the copyright notice
08:34:58manveru-0.0 + 0.0 anyone? :)
08:36:04manveruah, bsd
08:36:17manveruand C...
08:36:41manverudrbrain: did you have a sudden epiphany about my problem already?
08:36:50drbrainmanveru: no
08:36:51thehcdreamer enters the room.
08:36:55crossblaim enters the room.
08:38:04manveruheadius: maybe you should have looked at 'ri Kernel#select' yourself... it points to +Kernel#select+ - gotta love ri...
08:38:21headiusyeah I know
08:38:27headiusI don't know what's up with that
08:38:47tarcieriyeah, that's awesome
08:38:48tarcieriheh
08:41:04manveruso, how exactly do you look at that information? :)
08:41:18tarcierirecursively?
08:42:15manveru^^
08:42:32manveruwell, i have problems figuring out where exactly this 'Kernel' is in ruby source
08:42:47manverufor IO there is io.c
08:43:40manverubut can't find any kernel.c
08:47:48manveruthe docs in rubinius for IO::select are quite nonexistent as well...
08:48:32drbrainRubinius' IO::select is broken
08:49:01drbrainI think it only checks one read socket :(
08:49:26rueIO.select
08:49:39evanoh, i need to do that
08:49:49evanputs 'wire up full IO.select' on the top of his todo
08:51:04defunkt_ leaves the room.
08:55:03radarek enters the room.
08:56:41dbussinkevan: how long is that list? ;)
08:56:56headiusseeya all in about 13 hours
08:56:59headius leaves the room.
08:58:25evandbussink: i try to keep it fairly short
09:04:24w1rele55 enters the room.
09:04:25aotearoa leaves the room.
09:07:01langenberg enters the room.
09:07:29langenberg leaves the room.
09:08:10langenberg enters the room.
09:11:26octopod enters the room.
09:16:31imajes enters the room.
09:27:33boyscout1 commit by Evan Phoenix
09:27:34boyscout * Add arity checks to all primitives, add TURBO option; f67d9ac
09:28:08dbussinkevan: does the turbo help?
09:28:35dbussinkas in, do you have some benchmarks?
09:34:32maduyb__ leaves the room.
09:36:06dbussinkevan: the addition of the -Winline causes openbsd to throw warnings everywhere: rubinius.c: In function `rbs_class_new_instance': ../../shotgun/lib/object_memory-inline.h:9: warning: inlining failed in call to `_om_inline_new_object'
09:36:24dbussinkbut i have to go, back later
09:41:06rubuildiusEvan Phoenix: f67d9ace4; 1616 files, 5544 examples, 19092 expectations, 0 failures, 0 errors; http://rafb.net/p/jgdLJl18.html
09:41:18rubuildius_ppcEvan Phoenix: f67d9ace4; 1616 files, 5545 examples, 19084 expectations, 0 failures, 12 errors; http://pastie.caboo.se/paste/156894
09:43:31ttmrichter enters the room.
09:59:55perdiy leaves the room.
10:06:01d2dchat enters the room.
10:07:13cavalle enters the room.
10:09:05obvio enters the room.
10:09:34radarek leaves the room.
10:10:36perdix enters the room.
10:20:12radarek enters the room.
10:34:59GMFlash leaves the room.
10:35:03GMFlash enters the room.
10:42:21crossblaim leaves the room.
10:56:57obvio171 enters the room.
10:58:08ttmrichter leaves the room.
10:59:16radarek leaves the room.
11:00:04radarek enters the room.
11:01:06radarek leaves the room.
11:11:26obvio171 leaves the room.
11:13:17obvio leaves the room.
11:15:40crossblaim enters the room.
11:16:16radarek enters the room.
11:16:46ttmrichter enters the room.
11:19:53zimbatm enters the room.
11:35:31d2dchat leaves the room.
11:44:40dodecaphonic enters the room.
11:45:07cremes leaves the room.
12:01:18dodecaphonic leaves the room.
12:07:45kwatch enters the room.
12:12:46crossblaim leaves the room.
12:13:13msimbartl enters the room.
12:33:14dctanner enters the room.
12:39:09FFighter leaves the room.
12:47:08cavalle leaves the room.
12:49:41hornbeck leaves the room.
12:50:09crossblaim enters the room.
13:14:04hornbeck enters the room.
13:30:41cavalle enters the room.
13:33:05lifo leaves the room.
13:33:16geekounet leaves the room.
13:39:09obiejuan enters the room.
13:46:18RyanTM enters the room.
13:57:53anonuser enters the room.
14:07:03cremes enters the room.
14:07:06cavalle leaves the room.
14:07:23cavalle enters the room.
14:19:57dctanner leaves the room.
14:20:21dctanner enters the room.
14:21:39lifo enters the room.
14:27:15AndrewO enters the room.
14:32:50ragnard leaves the room.
14:34:02ragnard enters the room.
14:39:44dbussinkdid we ever look at libffi for our ffi purposes?
14:40:58rueThink so
14:42:08dbussinkbut it was dismissed?
14:42:31dbussinkbecause i was looking at a way we could implement something like va_list, or is that already possible?
14:44:07kwatch enters the room.
14:44:49rueI think it was dismissed because there had been no development
14:45:25agile leaves the room.
14:45:29rueI cannot imagine variable arguments would be terribly hard to do, it is just a matter of the calling convention
14:47:32probablycorey enters the room.
14:47:34dbussinkrue: no development of libffi? they just released a version yesterday :)
14:50:06skaa1 enters the room.
14:51:47ragnard leaves the room.
14:55:45rueMm, this was back then
15:00:22moofbong enters the room.
15:02:44imajes leaves the room.
15:04:03probablycorey leaves the room.
15:06:03probablycorey enters the room.
15:08:32obiejuan leaves the room.
15:12:39imajes enters the room.
15:13:19skaa1 leaves the room.
15:13:38skaar enters the room.
15:23:28pietia enters the room.
15:25:21dbussinkrue: dunno how usable it though, just stumbled into it
15:25:29dbussinkdid see that for example python uses it
15:26:09benburkert enters the room.
15:27:01hooopy enters the room.
15:30:53obiejuan enters the room.
15:34:39chris2 enters the room.
15:36:11rubymaverick enters the room.
15:38:54hooopy('a'..'z').join
15:39:05hooopyoops
15:39:14hooopywrong terminal
15:40:58rueNah, this interpreter is just really slow
15:41:21hooopyi seriously did think "why is this so slow?!"
15:41:30hooopymondays...
15:41:48ruedbussink: Yep, I have vague recollection that it was due to that, could have been lack of platform support
15:42:12wmoxam enters the room.
15:42:13DefilerHrm.. libffi seems to have made some nice progress
15:42:31DefilerMight be worth looking at again. They support a bunch of nice stuff, like Darwin64 and FreeBSD x86-64
15:42:56dbussinkDefiler: they even got some openbsd stuff in
15:43:17dbussinkthey already support more than we do
15:43:44DefilerYeah, and more than CoCo does
15:43:53dodecaphonic enters the room.
15:44:13DefilerNeed coffee. Back soon
15:44:31DefilerThe Windows support looks pretty good too
15:45:17dbussinkonly see cygwin in the list, or do they support other things too?
15:46:35rueWell, technically Cygwin would probably count as "good" against Lightning
15:46:41dbussinkhehe
15:47:19dbussinklooks like the mailing list also has some increased activity
15:49:43dbussinkhttp://sourceware.org/ml/libffi-announce/2008/msg00001.html
15:51:20ttmrichter leaves the room.
15:54:22Defilerdbussink: I downloaded the source, and it supports a bunch of things that aren't on that list
15:54:30Defilerdbussink: Apparently that list is just what they tested this release against
15:54:41dbussinkah, that's really nice
15:56:01Defilerhttp://sources.redhat.com/cgi-bin/cvsweb.cgi/libffi/src/x86/?cvsroot=libffi
15:56:52d2dchat enters the room.
15:57:02macournoyer enters the room.
15:57:53dbussinkguess it's probably worth testing this stuff
15:58:00dbussinkwe need something for ppc anyway
16:00:15DefilerLooks like it has a fair number of contributors in the changelog, as well
16:00:21DefilerSo I guess it is less dead than we thought
16:00:58dbussinkwell, like that announcement says, it has been hiding inside the gcc tree
16:03:15FoobarWidget enters the room.
16:06:43rueHm, looks good
16:16:15wycats enters the room.
16:17:46obiejuan leaves the room.
16:20:46benburkert leaves the room.
16:24:46wycats leaves the room.
16:25:22wycats enters the room.
16:26:46octopod_ enters the room.
16:29:11obiejuan enters the room.
16:29:43octopod leaves the room.
16:38:02lopex enters the room.
16:38:46_ADS enters the room.
16:44:13cavalle leaves the room.
16:45:51cavalle enters the room.
16:48:07twbray enters the room.
16:49:21cavalle leaves the room.
16:50:42msimbartl leaves the room.
16:50:48joachimm enters the room.
16:56:43langenberg leaves the room.
16:57:03geekounet leaves the room.
16:57:13geekounet enters the room.
17:00:42agile enters the room.
17:05:53dewd leaves the room.
17:07:13crayz__ enters the room.
17:07:37fabiokung enters the room.
17:09:47fabiokung leaves the room.
17:09:50agile leaves the room.
17:10:54agile enters the room.
17:11:23DefilerOK, NOW I can get coffee
17:13:58rueThat was a long pre-coffee event
17:14:07DefilerYeah, electricians showed up
17:14:23DefilerTurns out my house didn't have a ground whee
17:14:35DefilerA number of things work better now. :)
17:20:52crayz_ leaves the room.
17:25:22msimbartl enters the room.
17:26:04msimbartl leaves the room.
17:26:15msimbartl enters the room.
17:28:18joachimm leaves the room.
17:28:52dbussinkDefiler: that's nasty :)
17:37:55cavalle enters the room.
17:39:05thehcdreamer leaves the room.
17:42:35hooopynot having a ground... wow
17:42:36rueDefiler: Yep, got plugs like that here. Only two with the ground prong to begin with and neither is actually /connected with the ground/
17:43:16rueYay, today is a day off. Tomorrow, the work begins
17:43:31rueWell, the implementation work anyway
17:43:35olabini enters the room.
17:43:39benburkert enters the room.
17:44:19Fullmoon enters the room.
17:44:23hooopyat what time are the daily source tarballs generated?
17:46:45djwhittrue: implementing what?
17:49:06rueMonkeys!
17:49:19djwhittmonkeys?
17:49:22rueNot sure which question that answered
17:49:27djwhittheh
17:49:33rueI was watching a zoo thingy
17:49:58djwhittbacks away slowly
17:50:00ruedjwhitt: I can finally start poking around the mod_rubinius implementation this week
17:50:16rueLast two weeks I mostly spent reading
17:50:35dbussinkhooopy: dunno, just pull the source from the repository
17:50:40rueRe-read all of Stevens and anything else I could get my hands on
17:50:44djwhittyeah, I remember you mentioning something about 8k pages or some such absurdly large number
17:50:45dbussinkrue: ah, any ideas for it?
17:51:46ruehooopy: If you look at the directory listing, looks like 0:00 each day, I would venture PST
17:52:09rueSo about 10 hours ago
17:52:41hooopycool... i think i'll just set up a cron task to download/compile/etc it at 1:00 AM CST
17:52:44ruedjwhitt: Around 8900~ altogether in the end
17:52:48obiejuan leaves the room.
17:53:03dbussinkhooopy: you're gonna do some build testing?
17:53:05rueAlthough admittedly I skimmed parts of stuff I had read before
17:53:12hooopyi could do that
17:53:27dbussinkhooopy: because djwhitt was working on that
17:53:30ruehooopy: Hehe, does it matter when it is generated if you grab it once a day only anyway?
17:53:43hooopywell it's a daily tarball :P
17:53:59dbussinkhooopy: you can also git pull every day :)
17:54:16djwhitthooopy: I'll be looking for people to help me get some buildbot slaves going sometime this week if you're interested
17:54:17hooopyis install git right now
17:54:24hooopyi could do that
17:54:40dbussinki off for a trip to barcelona tomorrow morning :)
17:54:43hooopyi have an intel mac running 10.5.2, not sure how much that info helps but there ya have it
17:55:03ruedbussink: Not sure if there is anything revolutionary (apart from what Rubinius brings in.) Mostly it will just be targetted, super-easy to configure. Going the daemon route
17:55:15dbussinkrue: gonna use multi vm?
17:55:36djwhitthooopy: that'll be good. that's what a lot of the developers use, but having an automated build on it wouldn't hurt
17:55:54djwhitthooopy: here's what I have setup so far: http://spcom.org/rbx/buildbot/
17:55:56dbussinki thought you said you wanted to leave uid's / gid's to apache?
17:56:12djwhitthooopy: I'm running this: http://buildbot.net/trac
17:56:26probablycorey leaves the room.
17:56:42joachimm enters the room.
17:56:46ruedbussink: Best we can, multi-VM may wait for a bit because I do not have a clear picture of what all will be possible directly through it. Initially probably just threading several interpreters
17:56:51djwhitthooopy: I'm using the latest version pulled from CVS because it has better git support, but seeing as they use buildbot to test buildbot I imagine it's pretty stable
17:57:00hooopyhow often are builds requested?
17:57:03ruedbussink: It ends up being simpler
17:57:18djwhitthooopy: that can be determined by you if you like
17:57:36djwhittfor my machine I request them whenever the source tree updates
17:57:41djwhittwell, almost
17:57:41dbussinkah ok, well i think one of the major points should be leveraging proper copy on write, so memory usage is kept low
17:58:09djwhittbut if you wanted to do daily build or something it wouldn't be a problem
17:58:17djwhittbuildbot supports configureable scheduling
17:58:47dbussinkdjwhitt: is there easy documentation on how to register a bot with this?
17:58:54hooopywell if you all already have my platform...
17:59:16hooopywhat if i put something obscure on another box
17:59:38djwhitthooopy: that would probably be the most helpful
17:59:58ruedbussink: Yep, that would definitely be a handy thing
18:00:03hooopywishes he still had a ppc machine laying around
18:00:14djwhittright now I think we have OS X PPC and i386 (32bit) Linux accounted for
18:00:30obiejuan enters the room.
18:00:34djwhittand Linux amd64 via my machine
18:00:40djwhittthough it's not alerting IRC yet
18:01:16djwhittdbussink: I'm going to put something together about how to setup the buildslave
18:01:29djwhittbut you have to have a name + password from me to connect to the server
18:01:30dbussinkdjwhitt: ah, that would be nice
18:01:45dbussinkshould go for public / private key
18:01:58dbussinkthat you add that to a web interface or something like that
18:02:11djwhittyeah, that would be sweet
18:02:14hooopyi think i might throw together a box and put freebsd on it
18:02:15djwhittstarting with the basics though
18:02:27dbussinkwell, public / private key is pretty easy
18:02:47djwhittwell, using SSH it is
18:02:52djwhittbut this is all through buildbot
18:03:04djwhittI haven't checked to see what other authentication mechanisms they support
18:03:08wycats leaves the room.
18:03:14dbussinkyeah ok, but then i wouldn't need to give you a password :)
18:03:37djwhittwell, don't make it a password you use for anything else
18:03:43djwhittit's not a password for an account or anything
18:03:55djwhittjust to connect to the buildmaster
18:03:55dbussinki know, but if it does ssh, it does public / private key
18:04:02djwhittit doesn't do ssh
18:04:08dbussinkah ok
18:04:12dbussinksome daemon?
18:04:16djwhittyeah
18:04:22djwhittactually let me check...
18:05:21djwhittnah, I don't think it supports it
18:05:43djwhittI'm not married to buildbot either just seemed like a fairly easy solution
18:09:56joachimm leaves the room.
18:11:32Fullmoon leaves the room.
18:11:40dbussinkdjwhitt: is it possible to send triggers so the bot builds on a git push?
18:11:51w1rele55 leaves the room.
18:12:10djwhittdbussink: yes
18:12:37alxm enters the room.
18:12:58djwhittright now I'm just using the RSS feed, but it would be easy to allow the git server to send a command to the build server
18:15:26langenberg enters the room.
18:17:34joachimm enters the room.
18:25:05probablycorey enters the room.
18:25:20wifelette enters the room.
18:29:59wycats enters the room.
18:30:15luke_bkk enters the room.
18:31:07scooprwhat do guys think, is it wrong or correct for $? (status of last run childprocess) to be nil on windows after system("nonexistent"), but does have something sensible on non-windows systems (where I presume it does execute bash so there is sensible returncode etc.)
18:31:12rbx_buildbot enters the room.
18:33:17TheVoice enters the room.
18:35:09rbx_buildbot leaves the room.
18:35:52TheVoicemacournoyer: you around?
18:36:03macournoyeryeah
18:36:22macournoyerwazup
18:36:36TheVoicecouple quick questions
18:36:58macournoyersure
18:37:26TheVoicewhats the default for the amount of keep-alive connections?
18:37:52macournoyer512
18:38:13macournoyershould be in thin -h
18:38:18macournoyeryes
18:39:24_ADS leaves the room.
18:39:36octopod leaves the room.
18:40:37TheVoiceyou might want to state each option's default your release emails for easy googling in the future
18:40:46lifo leaves the room.
18:41:09macournoyeryeah sure
18:41:20probablycorey leaves the room.
18:41:28TheVoiceare you doing another naming contest for the next release candidate?
18:41:48macournoyersure, next minor version
18:41:49_ADS enters the room.
18:42:05macournoyerthis time ever bigger prices to win
18:42:18rbx_buildbot enters the room.
18:43:03wifelette leaves the room.
18:43:12djwhittok, testing buildbot irc stuff...
18:43:35probablycorey enters the room.
18:45:27wifelette enters the room.
18:46:55thehcdreamer enters the room.
18:48:06defunkt enters the room.
18:50:24djwhittrbx_buildbot: list builders
18:50:25rbx_buildbotConfigured builders: amd64-gentoo-full
18:50:40djwhittrbx_buildbot: status all
18:50:40rbx_buildbotamd64-gentoo-full: idle, last build 80 secs ago: failed test
18:52:34djwhittrbx_buildbot: help commands
18:52:34rbx_buildbotUsage: commands - List available commands
18:52:42djwhittrbx_buildbot: commands
18:52:42rbx_buildbotbuildbot commands: commands, dance, destroy, excited, force, hello, help, last, list, source, status, stop, version, watch
18:52:51djwhittrbx_buildbot: help watch
18:52:52rbx_buildbotUsage: watch <which> - announce the completion of an active build
18:54:55langenberg_ enters the room.
18:55:01langenberg leaves the room.
19:00:59probablycorey leaves the room.
19:01:23benburkert_ enters the room.
19:01:48enebo enters the room.
19:02:06probablycorey enters the room.
19:05:17langenberg_ leaves the room.
19:06:11twbray leaves the room.
19:12:21_ADS leaves the room.
19:17:13hornbeck leaves the room.
19:18:24geekounet leaves the room.
19:19:17probablycorey leaves the room.
19:20:36probablycorey enters the room.
19:21:01benburkert leaves the room.
19:21:41thehcdreamer leaves the room.
19:22:26hooopywonders why nobody has taken the suck out of mod_ruby
19:23:20djwhittI'm not 100% sure, but I think there are inherent problems with MRI that make a good mod_ruby difficult to implement
19:23:58djwhittothers in here could probably tell you the specifics of what those problems are
19:24:11hooopyphp would have a cooked goose by now if not mod_ruby.suck?
19:24:49Defilerrbx_buildbot: help status
19:24:50rbx_buildbotUsage: status [<which>] - List status of a builder (or all builders)
19:25:00Defilerrbx_buildbot: status
19:25:00rbx_buildbotamd64-gentoo-full: idle, last build 2139 secs ago: failed test
19:25:06DefilerAwesome
19:25:12djwhittyeah, getting there
19:25:33Defilerrue: How does an apache module 'wire up' to something that is meant to be persistent, like a ruby interpreter?
19:25:47DefilerI found myself wondering that earlier
19:26:23rueApache has persistent allocation pools, you can stash objects away and grab them when the handler is called
19:26:42pietia leaves the room.
19:26:49DefilerSensible
19:27:13FoobarWidgetDefiler: writing mod_rubinius?
19:27:31pietia enters the room.
19:27:40dbussinkFoobarWidget: rue is our man for that
19:27:51DefilerFoobarWidget: Yeah, rue is
19:27:52FoobarWidgetthe typical apache setup uses the prefork MPM, so be careful if you want to share stuff between different apache worker processes
19:28:07FoobarWidgetI had to invent a whole client/server architecture to get around that
19:28:27geekounet enters the room.
19:29:37rueFoobarWidget: Well, we will not have to worry about that part fortunately
19:29:48rueAlthough I did see a module that used shared memory for that :)
19:30:22FoobarWidgetthat works fine until you have to share file descriptors :(
19:30:38FoobarWidgetor implementing a hash table/binary tree in shared memory
19:30:40FoobarWidgetouch
19:31:02boyscout6 commits by Brian Ford
19:31:02boyscout * Add the rest of spec/* directories to CI process.; f26bb0c
19:31:04boyscout * Added --dry-run option to mspec-tag.; 1900456
19:31:04boyscout * Excludes for spec/kernel, spec/debugger to run with CI.; 29f3683
19:31:06boyscout * Excludes for subtend specs so they will run with CI.; 12bbdf7
19:31:06boyscout * Remove specs for removed Compression::ZLib.; 0cbc2b1
19:31:08boyscout ...
19:31:23rueHm, the prefork MPM uses a semaphored shared accept socket by default
19:31:38brixenok, mspec ci should be running all our specs, ~5750 of them
19:31:44rueFDs are easy to pass through unix sockets though if they are completely separate
19:32:50FoobarWidgetyes. what I meant was that file descriptors can't be shared via shared memory
19:33:04brixenDefiler: did you see the instance_eval branch?
19:33:31Defilerbrixen: Yeah, though I haven't looked at what it contains yet
19:33:37DefilerI think I can guess, though. Heh
19:33:42djwhittrbx_buildbot: watch
19:33:42rbx_buildbottry 'watch <builder>'
19:33:57brixenDefiler: yeah, just the change to mspec and ~109 failures on ci run :)
19:34:03djwhittrbx_buildbot: watch amd64-gentoo-full
19:34:03rbx_buildbotthere are no builds currently running
19:34:08brixenDefiler: also, your TOPLEVEL_BINDING spec
19:34:10rueFoobarWidget: Oh, sure. At some point the Apache model just gets in the way
19:34:28brixenDefiler: I'm guessing fixing that will have a positive impact on running rails stuff
19:34:36DefilerYeah
19:35:15brixenDefiler: fyi, #354 and I assigned it to evan, but I forgot he's traveling
19:35:19ragnard enters the room.
19:35:39FoobarWidgetrubinius development looks pretty fast
19:35:46FoobarWidgetyou guys have made great progress in the past few months
19:35:58rueI wish I could take credit for it
19:36:07dc_ leaves the room.
19:36:10rueTruth is, though, I could never have done it without the crack
19:36:17djwhittlol
19:36:56dbussinkbrixen: i was planning on asking evan why he put in those optimizations :)
19:37:07dbussinkbrixen: because i can't seem to find a significant difference with them
19:37:10brixendbussink: you mean TURBO ?
19:37:16dbussinkbrixen: yeah
19:37:17langenberg enters the room.
19:37:22dbussinkbrixen: and he added the -Winline back too
19:37:38dbussinkbrixen: to the default build, that spews out a ton of warnings on openbsd and freebsd
19:37:38brixendbussink: naturally aspirated ftw anyday, seen the new m3? :D
19:37:51DefilerFoobarWidget: rue is right. The network of disembodied brains in the back room helps too
19:38:00dbussinkbrixen: new m3?
19:38:11Defilerhttp://www.angryflower.com/uncomm.html
19:38:18brixendbussink: bmw m3 2008 naturally aspirated v8 ;)
19:38:27brixendbussink: no turbo
19:38:32FoobarWidgethow far is mod_rubinius by the way?
19:38:33rueI thought you were talking about the razor
19:38:45dbussinkbrixen: i'm going for the audi r8
19:38:56hooopyaudi = secksi
19:38:59FoobarWidgetI have been developing an apache module for the last 4 weeks and I know how much of a pain it can be
19:38:59brixendbussink: ohh nice
19:39:00DefilerI am holding out for the Skyline GT-R
19:39:03rueFoobarWidget: In fantasyland or reality?
19:39:07FoobarWidgetboth :)
19:39:08DefilerThough I am worried that it is too fast for me
19:39:23dbussinkbut i kinda need to build up some savings for that, minor issue though :P
19:39:45brixendbussink: so, yeah, don't know about the TURBO stuff, but evan will probably be around at some point
19:40:00hooopyhell, i'd just like a 2008 Accord
19:40:02Fullmoon enters the room.
19:40:04rueFoobarWidget: Heh, actually I just completed my research/planning/overarching design phase. Starting on the code today basically
19:40:07ruehooopy: Haha
19:40:12rueMe, I want a Bianchi
19:40:25hooopyhas a 2004 Accord Coupe EX-L V6
19:40:36hooopyit's nice... but i shouldn't have gone with the coupe
19:40:43dbussinki'm happy enough that i'm able to buy my first car somewhere in the next couple of weeks :P
19:40:48DefilerThat is a nice car, but I need a manual transmission
19:41:02DefilerI don't like how automatics push against me
19:41:08FoobarWidgetrue: what I do is moving most logic outside the apache module, into units/classes that can be tested independently
19:41:25dbussinkdefinitely manual, luckily that's the default here in europe
19:41:26FoobarWidgetso that the apache module is a lightweight wrapper for everything else
19:41:33rueFoobarWidget: Right now the plan is to use the lightest possible Apache module to bridge requests over to a daemon
19:42:00FoobarWidgetah
19:42:03dbussinkDefiler: automatics are more expensive and harder to get here :)
19:42:03FoobarWidgetwill it be C or C++?
19:42:19Defilerdbussink: Awesome. As they should be
19:43:03rueFoobarWidget: I 'c' no reason to use C++
19:43:21benburkert_ leaves the room.
19:43:25FoobarWidgetnot even smart pointers and exceptions? :)
19:43:33FoobarWidget(oh, and classes)
19:44:09Fullmoon_ enters the room.
19:44:28rueNah
19:44:36dbussinkif it's an ultra thin wrapper, it's probably overkill
19:44:42rubuildius_ppcBrian Ford: f26bb0c4e; 1712 files, 5759 examples, 19564 expectations, 0 failures, 12 errors; http://pastie.caboo.se/paste/157122
19:44:50FoobarWidgetyeah, that's a good point
19:45:44rueMaybe I should write it in Erlang
19:45:52Fullmoon_ leaves the room.
19:46:37dbussinkDefiler: good thing about the usa when it comes to cars is that they are a lot cheaper, and gas is too
19:46:41rubuildiusBrian Ford: f26bb0c4e; 1712 files, 5758 examples, 19572 expectations, 0 failures, 0 errors; http://rafb.net/p/qBJflp49.html
19:47:56FoobarWidgetrue: nah. assembly FTW
19:48:20drbrainI love my 1978 Mercury
19:48:51rueCaptain Planet hates your ride, drbrain
19:49:12dbussinkbrixen: hmm, ./bin/mspec ci -t ruby now also runs rbx specific specs
19:49:20lifo enters the room.
19:51:03drbrainrue: actually, he doesn't, as I drive all of 3,000 miles/year
19:51:50brixendbussink: bin/mspec ci is not for running -t ruby
19:52:04dbussinkbrixen: well, it does do something :)
19:52:21brixendbussink: we're not doing development on ruby, why would we run ci specs?
19:52:34brixenby definition, all specs should pass on ruby, those are in spec/ruby
19:52:52dbussinkbrixen: true, but how do we keep an eye on that?
19:52:59brixeneye on what?
19:53:08dbussinkthat all specs are still compliant with mri
19:53:18drbrainbrixen: it makes it easy to do command-line editing when I'm working on exactly one thing
19:53:20dbussinkand that not some spec is wrong or fixed wrong
19:53:21brixenwell, we need an mri bot that runs bin/mspec spec/ruby
19:53:27drbrainbin/mspec ci spec/blah => bin/mspec ci -tr spec/blah
19:53:44brixendrbrain: and it's confusing to everyone else
19:53:49brixenso use bin/mspec -G fails
19:54:24drbrainthere is no -G
19:54:38brixenok, lunchtime bbiab...
19:54:58brixendrbrain: bin/mspec run -h
19:55:05drbrainI did
19:55:15drbraintTIrnXDAwvh
19:55:19drbrainno G
19:55:29brixen -G, --excl-tag TAG Exclude examples with descriptions matching ones tagged with TAG
19:55:40drbrainis this new?
19:55:49brixencame with the mspec rewrite
19:55:58drbrainwhich one?
19:56:08brixentry this exactly: bin/mspec run -h
19:56:13brixenI'm sure you'll see it
19:56:26brixenk, lunch..
19:56:32rbx_buildbot leaves the room.
19:56:35drbrainoh, gah
19:57:07drbrain*that* is confusing
19:57:21drbrainif run is the default, why don't I see its help by default?
19:57:26drbrainalso lunches
20:00:23rbx_buildbot enters the room.
20:01:34rbx_buildbot leaves the room.
20:01:37rbx_buildbot enters the room.
20:07:47rbx_buildbot leaves the room.
20:07:51rbx_buildbot enters the room.
20:11:26rbx_buildbot leaves the room.
20:12:22dctanner enters the room.
20:13:05context leaves the room.
20:15:22nemerle leaves the room.
20:18:57cored enters the room.
20:20:18seydar enters the room.
20:25:43cored leaves the room.
20:38:27imajes enters the room.
20:41:50drbrain-G isn't what I want
20:42:21context enters the room.
20:46:21seydardrbrain: did i please the master?
20:46:36drbrainseydar: haven't looked yet
20:46:41drbrainand don't call me that
20:46:52seydaryes.... sir
20:53:26hooopyhttp://pastie.caboo.se/157169
20:53:57hooopythere is surely a more formal way to report stuff like that and i apologize for not being famialiar with it
20:54:42therealadam enters the room.
20:55:43drbrainhooopy: a) file a bug on lighthouse
20:55:55drbrainb) just fix it and submit a patch
20:56:16seydardrbrain: i finished my homework. is there something else that needs to be done?
20:56:30drbrainseydar: ok, I will go look
20:56:42VVSizhooopy: damn, fails on JRuby too: inject.rb:10:in `to_proc': wrong # of arguments(0 for 1) (ArgumentError) :)
20:56:49seydar:-)
20:56:54cavalle leaves the room.
20:57:28cavalle enters the room.
20:59:01drbraininternet is slow
20:59:33seydarOMG host
20:59:40seydarsorry, halo thing
21:01:50srbakerdrbrain: my chair was sitting on the line. kinked it
21:04:07drbrainseydar: can you turn stdlib/ext/syslog/syslog.txt into RDoc?
21:04:26seydaruh, sure
21:04:38seydarah! now i know what you mean
21:05:01ctennis enters the room.
21:05:16imajes leaves the room.
21:05:56imajes enters the room.
21:06:07drbrainMRI does something different with ::write
21:06:39Phoop leaves the room.
21:06:48seydarfigured
21:06:56seydarwhat does it do?
21:07:19Phoop enters the room.
21:08:13drbrainI think you need a def write(pri, format, *args) message = format % args; Foreign.syslog pri, "%s", message
21:08:40drbrainthat extra String#% in the middle
21:08:54Defilerhooopy: Please file a ticket for that, yeah.
21:09:37hooopyokay
21:09:44seydargotcha
21:10:01hooopynot sure what to title it
21:10:21seydarDefiler: i figured out how to use `git format-patch`!!
21:11:38dbussinkhooopy: and as addition, write a spec that tests this
21:11:56hooopyi'm still nubbish :O
21:12:00hooopyi'll look into it though
21:13:05seydardrbrain: can you post syslog.txt please? Mine has disappeared in the deletion
21:13:37dbussinkseydar: you can also look at the gitweb interface
21:13:47seydardbussink: was unaware of that. thanks!
21:13:55drbrainhttp://git.rubini.us/?p=code;a=blob;f=stdlib/ext/syslog/syslog.txt;h=8764b2dbf3cc758f4879027509aba 62e8e56cd14;hb=HEAD
21:14:01drbrainwow, that's one hell of a URL
21:14:40rueUse the raw link for a cleaner HEAD URL
21:15:23luke_bkk leaves the room.
21:17:53hooopy leaves the room.
21:18:05seydardrbrain: i fixed up write and implemented log. easier than i imagined!
21:18:25hooopy enters the room.
21:18:41gnufied enters the room.
21:19:06hooopyinject isn't calling to_proc on the symbol... it thinks it is a proc and is attempting to call it
21:19:11benburkert enters the room.
21:19:29rbx_buildbot enters the room.
21:19:48twbray enters the room.
21:19:51djwhittrbx_buildbot: notify
21:19:51rbx_buildbottry 'notify on|off|list <EVENT>'
21:20:39djwhittrbx_buildbot: notify list failed
21:20:39rbx_buildbotThe following events are being notified: []
21:20:49djwhittrbx_buildbot: notify on failed
21:20:49rbx_buildbotThe following events are being notified: ['failed']
21:21:29aotearoa enters the room.
21:21:40ragnard leaves the room.
21:22:35Defilerhooopy: It needs to already be a proc by that point
21:23:18Defilerhooopy: Somewhere in the code invoked by the 'send_stack_with_block' instruction, that needs to happen
21:23:34DefilerTrying to figure out where now, though.. all of this code has been reworked in the last few weeks
21:23:41zenspiderDefiler: lemme know when you want that ppc box up
21:23:59zenspiderI've got plans to use that IP in the nearish future, but it is clear for now
21:24:01Defilerzenspider: Who got picked to work on the botnet again?
21:24:05rueThe code is trying to get into its yellow polka-dot bikini
21:24:11seydari've got one running (Tiger/ppc) right now, if you want me to test something
21:24:14boyscout2 commits by Ryan Davis
21:24:14zenspiderthat, or you can work on/with it when you're here (you're coming up, right?)
21:24:15boyscout * Fixed autotest churn by removing empty.txt and moving to /tmp; cee0888
21:24:17boyscout * Fixed 'rake todo', broken after last mspec change; d348928
21:24:39Defilerzenspider: Yeah. When does the australia trip thing start? 7th?
21:24:55zenspiderDefiler: botnet itself is eric. the various systems, whichever suckers step up
21:25:05zenspiderDefiler: I think so, but he leaves the 3rd
21:25:43DefilerGotcha
21:25:54DefilerI'm asking if EY wants to cover the ticket now
21:26:21zenspidercool
21:27:14evanDefiler: cover which ticket?
21:27:28rbx_buildbotbuild amd64-gentoo-full #7 is complete: Failure [failed test] Build details are at http://spcom.org/rbx/buildbotbuilders/amd64-gentoo-full/builds/7
21:27:39djwhittw00t
21:27:42djwhittit works
21:27:57rueCool
21:28:01djwhitt