Index

Show enters and exits. Hide enters and exits.

00:00:04evandelete lines 314 to 322
00:00:41Defilerwell, only 315-318 are the if+body, right?
00:00:52Defilermy line 314 is 'val = lookuptable_find(...'
00:01:04evanmy line numbers are off
00:01:08evanstarting at
00:01:08Defileraah
00:01:13evan if(object_kind_of_p(state, under, state->global->module)) {
00:01:20evandelete that AND the body
00:01:32Defileraah, gotcha
00:01:45DefilerOh, right.. bug is pretty visible, reading that. Heh
00:01:50evanyep
00:01:53evanuse your head!
00:01:54evan:)
00:02:10rubuildius_amd64Ryan Davis: b7ba50232; 2090 files, 6570 examples, 22949 expectations, 0 failures, 0 errors; http://rafb.net/p/W3ul8e72.html
00:02:11rubuildius_amd64Wilson Bilkovich: 046ba6228; 2090 files, 6570 examples, 22949 expectations, 0 failures, 0 errors; http://rafb.net/p/vjSUb553.html
00:02:48DefilerWell, in my copy, the 315+ lines are also about returning if the constant is found
00:02:51Defilerso it was confusing :)
00:03:56Defileroh snap
00:03:57DefilerUnable to send const_missing on #<Module> (NoMethodError)
00:04:40Defiler(when attempting to compile loader.rbc)
00:05:54agile leaves the room.
00:08:06rubuildius_ppcRyan Davis: b7ba50232; 2090 files, 6573 examples, 22978 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184523
00:08:07rubuildius_ppcWilson Bilkovich: 046ba6228; 2090 files, 6573 examples, 22978 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184522
00:11:43fbuilesv leaves the room.
00:14:15Defilerevan: huh. I put it back along with some print statements, and that codepath is what is fired to load RUBY_CONFIG, RUBY_PLATFORM, etc
00:14:47evanhm
00:14:54evanshouldn't the loop above that pick those up?
00:15:24DefilerRUBY_CONFIG and RUBY_RELEASE_DATE are in the same boat. wacky
00:15:48boyscout2 commits by Marnen Laibow-Koser
00:15:49boyscout * Remove a bunch of conflict markers that somehow got committed.; 0754709
00:15:50benburkert leaves the room.
00:15:50boyscout * Merge branch 'master' of git@git.rubini.us:code; 24b2501
00:16:03evanmarnen: please use rebase.
00:16:15marnenI did.
00:16:19benburkert enters the room.
00:16:27evanno, ya didn't.
00:16:33evansee the merge commit?
00:16:58marnenYeah. Hmm.
00:17:15Defilerline 137 in loader.rb appears to be what triggers it
00:17:15marnenSorry, still feeling my way around git. If I rebased to the wrong thing, I'm sorry.
00:17:23Defileraha.. yeah.. that is at toplevel
00:17:42DefilerSo maybe we should put this back, but avoid running it if 'under' is set?
00:17:42evanit must be confused
00:17:51evanwait wait.
00:18:14evanwhose shouldn't be using const_get_from
00:18:16evanthose
00:18:30evanthats const_get_in_context
00:19:14evanmarnen: no problem.
00:19:44evanmarnen: merge commits are created when you run 'git pull' into a branch that contains local-only commits
00:19:54evanso if you want update your branch
00:19:58evanyou should use
00:20:00evanrake git:pull
00:20:00radarek leaves the room.
00:20:05evanit will do the right thing.
00:20:10marnenaha, thanks.
00:20:16marnenI hadn't found that yet.
00:20:43evanoh
00:20:54evanthose should be highlight in giant text on the howto use git page on lighthouse
00:21:00evanare they not? did you see that page?
00:21:25evandamnit!
00:21:28evanit's not even on there
00:21:30evan*sigh*
00:21:35evancorundum: git is http://rubinius.lighthouseapp.com/projects/5089/using-git
00:21:36corundumcan do!
00:22:13fbuilesv enters the room.
00:23:43flori leaves the room.
00:23:52marnenYes, I read the page. No, not all the info is on there. Sorry again. Oy.
00:23:56Defilerevan: hrm.. how could these toplevel ones not end up doing 'push_const'?
00:24:03cored enters the room.
00:24:09marnenI suppose this is payback for me being the SVN guru at my last couple of jobs. :D
00:24:10evanDefiler: use gdb
00:24:17evanDefiler: set a breakpoint where you print
00:24:20evanlook at the gdb backtrace
00:24:23evanfigure out the flow to that point
00:24:30evanyou'll be able to tell which instruction was used.
00:24:32Defileryeah.. rebuilding now
00:25:31zimbatm leaves the room.
00:26:16ruemarnen: No worries; you may want to get into the habit of checking git-log before pushing just in case something odd happens
00:26:42rubuildius_amd64Marnen Laibow-Koser: 0754709ca; 2090 files, 6613 examples, 23131 expectations, 2 failures, 0 errors; http://rafb.net/p/DqyOls40.html
00:26:52marnenThanks for the reminder. I had been doing that.
00:27:02marnenbut somehow got out of the habit.
00:27:12ruemarnen: The merge commit in itself is not harmful
00:27:37marnenOK.
00:27:42rubuildius_ppcMarnen Laibow-Koser: 0754709ca; 2090 files, 6616 examples, 23160 expectations, 2 failures, 0 errors; http://pastie.caboo.se/paste/184528
00:27:58rueLooks like you need to research some failures though :)
00:27:58marnenOK, what did I break this time? :)
00:28:19Fullmoon enters the room.
00:28:19marnenEek. Those weren't supposed to be in CI yet.
00:29:37rueJust tag them, they do look like existing issues
00:31:22Defilerevan: so yeah.. find_const is what is being emitted here
00:31:43evanthats odd.
00:32:09evanfind_const takes a parent
00:32:11evanwhat is the parent?
00:32:59Defiler<Module:Rubinius>
00:33:03Defileris the _inspect output
00:33:10evanoh
00:33:12evanof course
00:33:19evani thought we were talking about RUBY_PLATFORM
00:33:41evanis there a Rubinius::RUBY_VERSION constant even?
00:33:45evanmaybe thats a type
00:33:48Defiler puts "rubinius #{Rubinius::RBX_VERSION} (ruby #{Rubinius::RUBY_VERSION} compatible) (#{Rubinius::BUILDREV[0..8]}) (#{RUBY_RELEASE_DATE}) [#{RUBY_PLATFORM}]"
00:34:00Defilerthat is a pretty weird line, really
00:34:15DefilerI mean, there should be a 'RUBY_VERSION' toplevel constant, right?
00:34:25evanis there a Rubinius::RUBY_VERSION though?
00:34:28evannot the toplevel one.
00:34:29evanif not
00:34:31evanthen thats the bug.
00:34:38evanit should be just RUBY_VERSION, no Rubinius::
00:34:50DefilerYeah, it exists
00:35:04evanhow are you cheking
00:35:07evanif it's in irb
00:35:09evanthen thats invalid
00:35:16evanbecause of the broken constant lookup rules.
00:35:19evanyou need to check the source
00:35:41Defilerdoing it with -e on the command-line
00:35:50evandouble bad.
00:36:01DefilerRUBY_VERSION = Rubinius::RUBY_VERSION
00:36:04Defileris in the source, though
00:36:04evanyou can't test a broken lense by looking through it.
00:39:24DefilerOK.. so..
00:39:47Defilerreferencing Rubinius::FOO, when a toplevel FOO exists..
00:39:54Defilerthat is meant to hit const_missing, right?
00:40:01evanyes.
00:40:04evanthat should NOT work.
00:40:06boyscout1 commit by Ryan Davis
00:40:07boyscout * Added dynamic platform abbreviation; 262ce6a
00:40:08evanbut it does because of this code.
00:40:11Defilergotcha
00:43:24headiusyou sure about that
00:43:24headius?
00:43:57Defilerevan: Can you take a look at shotgun/lib/machine.c line 580 and tell me what that is all about?
00:44:38evanwhat about it?
00:44:41evanthats platform detection
00:44:52Defiler machine_set_const(m, "RUBY_CONFIG", m->s->global->config);
00:44:56evanfuck
00:44:59evanmy lines are really off.
00:45:12evanok, what about that?
00:45:38DefilerWhy do we want a toplevel constant called RUBY_CONFIG?
00:45:39rype leaves the room.
00:46:28evancause i coded it that way.
00:46:48DefilerI should change this to set_const_under, it seems to me
00:47:30evanseems like you're getting really off the original bug
00:48:26DefilerDoesn't seem that way to me
00:48:35DefilerI have to make the kernel not use this code path before I can remove it
00:48:41evani have no clue what you're looking for in machine
00:48:55DefilerYep. changing that fixed it
00:49:33evanwhat does RUBY_CONFIG have to do with it?
00:49:48evani think you've moved on in the debugging and left me back a bit
00:49:51evanso i'm out of context.
00:49:52evananyway
00:49:59evani'm going to keep working on C++
00:50:01evanyou go
00:51:23rubuildius_amd64Ryan Davis: 262ce6ad1; 2090 files, 6613 examples, 23131 expectations, 2 failures, 0 errors; http://rafb.net/p/SbRWMi77.html
00:51:31evanmarnen: are you fixing those failures?
00:51:32evanwhats up?
00:51:35Defilerk. sorry to be confusing
00:52:29dewd leaves the room.
00:52:57rubuildius_ppcRyan Davis: 262ce6ad1; 2090 files, 6616 examples, 23160 expectations, 2 failures, 0 errors; http://pastie.caboo.se/paste/184542
00:54:27headiusevan: you know that XXX::YYY will still access toplevel constants, right?
00:54:33zenspiderI didn't do it!
00:54:35headiusi.e. a toplevel YYY
00:54:42evannot with the tests i did
00:55:15headiusin both 111 and 114 it does for me
00:55:15FoobarGuy leaves the room.
00:55:20headiuswhat tests did you do?
00:55:22evanhttp://pastie.caboo.se/184543
00:55:48benstiglitz leaves the room.
00:56:03headiuscome now evan
00:56:05headiusyou should know you can't trust irb's nesting
00:56:07headiusneither of those are defined at toplevel
00:56:19headiustry it in a script
00:56:40evannope.
00:56:46evanexactly the same thing in a script
00:56:56evanand i get a const_missing
00:57:00headiusinteresting...I'm seeing a different effect in -e
00:57:08evanyou should know not to trust -e!
00:57:11zenspiderirb's nesting?
00:57:14headiushah
00:57:16headiusit's only modules
00:57:19headiusclasses work
00:57:45headiusawesome
00:57:46headiuspastie
00:57:55pastiehttp://pastie.org/184544 by headius.
00:58:24evani fucking give up.
00:58:28evani'm going to become a gardiner.
00:58:48Defilerhaha that is the most appropriate typo ever
00:59:06headiuslooks like they work right in JRuby, though I couldn't tell you why
00:59:06zenspiderhahaha
00:59:12DefilerThe agardiner post is already taken, sadly
00:59:16marnen leaves the room.
00:59:20agardinerhehe
01:00:24evani'm bringing that up tonight.
01:01:07marnen enters the room.
01:01:14headiusenebo worked on our constant scoping stuff...what's up with this enebo
01:01:53eneboI am not sure :) I do remember there being a difference...but it confuses me too
01:02:03eneboProbably inheritance walking in TC.getconstant
01:02:04wmoxam enters the room.
01:02:50headiusI'll see if I can find it
01:03:00Defilerzenspider: This is for you: http://www.youtube.com/watch?v=x5idyptK15w
01:03:17headiusoh yeah
01:04:48headiusI found it
01:05:08zenspiderDefiler: omg I hate you. :P
01:05:11eneboguessing module does not check Object
01:05:13headiusafter the search runs, if the containing scope is a module it also checks object
01:05:14headiuser
01:05:32zenspiderI'm gettin' LAID TONIGHT!!!... thanks to Instant Fresh Folgers!
01:05:41headiushmm
01:05:53DefilerThis sounds like a job for... const_missing!
01:06:55enebowow...fresh-perked will never sound the same to me
01:07:24boyscout2 commits by Marnen Laibow-Koser
01:07:25boyscout * Update tags on failing specs.; 645784c
01:07:26boyscout * Put the decimal point into to_s.; d7be3ab
01:07:26zenspiderhaha
01:07:28headiushmm
01:07:38zenspiderI've got a perk for ya, right here!
01:08:14flori enters the room.
01:08:34marnen leaves the room.
01:11:00headiusok, I found it
01:11:24headiuscolon2 against a module does not search nesting
01:11:53headiuswait, no
01:11:57headiusblast
01:12:09Defilerbadass. I got this working
01:12:18DefilerAnother amusing bootstrap puzzle :)
01:12:26ezmobius_ enters the room.
01:12:27evanheadius: thats not it
01:12:36evanheadius: they both use rb_const_get_from()
01:12:43evanClass and Module
01:13:21flori leaves the room.
01:13:37rueevan: Check in stupid.c
01:13:42rueI think it is in there
01:13:55ruerb_do_retarded_stuff()
01:13:59zenspiderhaha
01:14:10evanfrankly my dear, I don't give a damn.
01:14:10headiusjruby's logic isn't going to match ruby's because we don't have cref anymore
01:14:21evandoesn't matter
01:14:24evancolon2 doesn't use cref.
01:14:49headiusexactly, doesn't use nesting
01:14:54evanexactly.
01:14:54headiusbut that's not why this is different
01:15:04headiusI see that much
01:15:35evanperhaps it's the superclass search.
01:15:45evanObject is located in the superclass search of a Class
01:15:47evanbut not in a Module.
01:15:55eneboevan: I think that is it
01:16:01evansuperclass search of a Module is only includes into that Module
01:16:20rubuildius_amd64Marnen Laibow-Koser: 645784c3d; 2090 files, 6619 examples, 23154 expectations, 0 failures, 0 errors; http://rafb.net/p/GrB1u185.html
01:16:29headiusyes
01:16:32evanall the cases we've given fit that conclusion
01:16:34headiusjust saw it here too
01:16:36evanand so does MRIs code.
01:17:03headiusso it makes perfect sense! :
01:17:06headius:D
01:17:16headiusgo forth and conquer
01:17:25evanand thus why line 314'ish should be removed.
01:18:12headiuswell, as long as the great 314 doesn't break it for classes, we're golden
01:18:17evanit doesn't
01:18:18flori enters the room.
01:18:23evanbecause the code above that does a superclass search
01:18:27evanjust like rb_const_get_0
01:19:03headiusexcellent...glad we've cleared up that Bar::Foo can reach top-level constants then
01:19:47evani can remove it from the agenda now.
01:20:07rubuildius_ppcMarnen Laibow-Koser: 645784c3d; 2090 files, 6621 examples, 23180 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184556
01:20:11evanthe agenda is empty btw.
01:20:13evanstill
01:20:23evani wonder if it's because pbwiki requires you to log in to edit
01:20:39headiusyeah, I'm on it now
01:20:45headiusand I saw laurent just mentioned pbwiki
01:21:31evanyep
01:22:12headiushmm, my pbwiki confirmation never showed
01:22:35evani just invited you
01:22:56headiusthanks, in now
01:23:03dblackdoes rubuildius report every rubinius build everywhere?
01:23:09trythil leaves the room.
01:23:18headiusgoing to take your item off the agenda?
01:23:31evannah
01:23:33evani'll leave it on
01:23:43evanwe can still discuss it briefly
01:23:46evansince there is a warn there
01:23:51evanmaybe the idea is to remove that
01:23:55evanwe should verify with matz.
01:24:47rueI would recommend taking the first one to just come up with topics
01:25:03ruedblack: No, there are two dedicated buildbots right now, _ppc and _amd64
01:25:11Defilerevan: So I should leave this in?
01:25:22ruedblack: Plus I think the dynamic duo were working on a tinderbox so probably more soonish
01:25:23Defilerevan: I just got CI totally happy without warnings with support for it removed ha ha
01:25:25evanDefiler: sure
01:25:31evanDefiler: huzzah!
01:25:34evanyou have worth!
01:25:35evan:D
01:25:37dblackrue: i hate to ask this, but who's the dynamic duo?
01:25:43dblackis kind of new around here :-)
01:25:46ruedblack: drbrain and zenspider
01:25:47evanzenspider and drbrain
01:25:51rueHehe
01:26:02evanzenspider and I got some new CI code up today
01:26:02dblackoh THEM :-)
01:26:07headiusI added a few
01:26:14evanci.rubini.us is going to have stats
01:27:14tarcieriare you going to have a hall of shame for who breaks the build the most? :)
01:27:20evancourse
01:27:22tarcieriyay
01:27:33evanit's going to be the doom leader board
01:27:49evanin all 8bit glory
01:28:29tarcierinice
01:28:35ezmobius leaves the room.
01:29:15rueI reserve RUE
01:29:27evanconsider it reserved
01:30:13Defilerevan: I'm going to go ahead and push this, and then add back support for this craziness (once we have a spec) via const_missing, rather than C code
01:30:29DefilerSound OK?
01:30:30evanwhich craziness?
01:30:43evanthere is no new craziness
01:30:43Defilerthe accessing toplevels via YYY::ZZZ
01:30:47evanthats already done.
01:30:58evanvia the superclass search that code already does
01:31:01DefilerYeah but I just took it out heh
01:31:14Defiler(assuming you mean the code you indicated earlier)
01:31:51evanno you didn't
01:31:53evanif you did
01:31:58evanconstant lookup is completely broken
01:32:01evanin which case
01:32:03evanDO NOT CHECK IN.
01:32:03DefilerOK, then I didn't
01:32:05Defilerbecause CI passes
01:32:09evanthen you're golden.
01:32:31DefilerI only skimmed the conversation you had with headius etc, so I probably missed something
01:32:49evanif CI passes, then you're fine.
01:32:52boyscout2 commits by Wilson Bilkovich
01:32:53boyscout * Rebuild stable archives without incorrect constants; 14b43f7
01:32:54boyscout * Remove support for incorrect constant references, and fix existing sites; 1e131ca
01:32:56evanok, i've got to run some errands.
01:35:24ezmobius_ leaves the room.
01:35:52fbuilesvruby-core's meeting is at 9 PM PST?
01:35:57fbuilesv7*
01:36:28evanyeah
01:36:32evan11am japan time
01:37:01d2dchat leaves the room.
01:37:57dblack leaves the room.
01:41:20rubuildius_amd64Wilson Bilkovich: 14b43f738; 2090 files, 6619 examples, 23154 expectations, 0 failures, 0 errors; http://rafb.net/p/aHcXow85.html
01:42:38flori leaves the room.
01:42:44flori enters the room.
01:42:50twbray leaves the room.
01:43:11yugui enters the room.
01:43:38antares leaves the room.
01:45:08rubuildius_ppcWilson Bilkovich: 14b43f738; 2090 files, 6621 examples, 23180 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184569
01:45:35binary42 enters the room.
01:46:57DefilerHrm.. somehow there is still a code path that can cause Thing::Foo::Baz to refer to ::Baz
01:50:03wmoxam_ enters the room.
01:50:32wmoxam leaves the room.
01:50:41wmoxam_ leaves the room.
01:50:52wmoxam enters the room.
01:52:21Defileraah, I see it.
01:53:00DefilerThis is an interesting behavior. I think headius and crew are correct about how it needs to be implemented. const_missing cannot be relied upon to fire
01:53:50Defilersince module Top;end AnyClass::Top will resolve the toplevel Top
01:54:04headiushehe
01:54:06headiuswe went through that
01:54:16headiusit should work with classes but not modules
01:54:16DefilerI see what you mean now, though
01:54:29Defilerautoload is there to provide a 'more specific' version of that constant
01:54:32headiusbecause classes resolve all the way up to Object in their superclass chain while modules do not
01:54:34headiusyeah
01:54:38headiusit really *is* the constant
01:54:39Defilerand there is no way for const_missing to inject itself into the process
01:55:09agardinerhas learnt more than he ever thought he wanted to know about constant lookup!
01:55:12eneboDefiler: Another aspect to behavior like this is it did break real apps for us. We largely spent most of our time just fixing broken libraries for a long time...corner cases working came second
01:55:36eneboThough I guess you could call this a corner case perhaps :)
01:55:45DefilerYeah, this is definitely not optional behavior, due to legacy horror
01:56:02DefilerPersonally I would prefer to kill off Something::Top, rather than just warning about it
01:56:13Defilerthen you really could use const_missing to implement autoload
01:59:02bricolage leaves the room.
02:00:23benny leaves the room.
02:01:14headiusin general I worry about any atttempt to use const_missing for autoload
02:01:16cored leaves the room.
02:01:18headiusbecause that's basically what we had
02:06:54dblack enters the room.
02:09:45aotearoa enters the room.
02:10:35AndrewO enters the room.
02:11:21wyhaines leaves the room.
02:12:45agile enters the room.
02:16:08GMFlash leaves the room.
02:16:18GMFlash enters the room.
02:19:16MenTaLguY enters the room.
02:19:22MenTaLguYhowdy
02:20:23tarcieriwhat's up MenTaLguY
02:20:33MenTaLguYnot too much
02:20:44tarcieriMenTaLguY: is there a suitable channel for IO event reception in Actors now, or should I add another?
02:20:47rueHey, MenTaLguY
02:20:54MenTaLguYtrying to figure out how to stay up for trhe ruby design meeting without being dead tomorrow
02:21:00zico enters the room.
02:21:47rueHow early do you have to be up?
02:22:09rueStarts at 10 unless I got myself confused now
02:22:43MenTaLguYit's 11pm pst, isn't it?
02:23:15agile leaves the room.
02:24:46drbraincurrently it is 18:25 Pacific Time
02:25:00headiusit's 7PM PDT
02:25:02drbrainthe #ruby-core thing starts at 19:00 Pacific Time
02:25:04dlee leaves the room.
02:25:06headiusfor the meeting
02:25:12headius11:00 JST
02:25:16MenTaLguYohh
02:25:18MenTaLguYgreat
02:25:22headiusso you're good
02:25:33MenTaLguYno multiphasic sleep experiments for me then
02:25:39rueHehe
02:25:52dblackhi MenTaLguY
02:26:01MenTaLguYhowdy david
02:26:04agile enters the room.
02:26:05headiusI've been off those lately myself, and oddly I'm getting less done
02:26:08rueMenTaLguY: Trust me, it is NOT as much fun as they make it out to be. My sleep schedule has been fuX0red for months now
02:26:12headiusso I think the myths about normal sleep schedules and productivity are a bunch of nonsense
02:26:19dblackMenTaLguY: one of the tests in the Actor class that i think you wrote got me into an infinite loop or hang or something... i haven't tracked it down but i will unless you've found it already
02:26:53MenTaLguYdblack: I've not written any of the specs for it.. but I can have a look maybe
02:27:05rueheadius: I would venture to say that by an large an average specimen of the human race will do better with a schedule
02:27:08MenTaLguYrue: oh, so's mine. but it's been simple fuXation so far
02:27:10dblackMenTaLguY: i'll try to reproduce it if it doesn't happen spontaneously for you
02:27:27MenTaLguYdblack: all the ci specs pass for me
02:27:29tarcieridblack: I think all the Actor tests were written by rosejn
02:27:48MenTaLguYyes, I didn't write any of the actor specs
02:28:01MenTaLguY(should have, but didn't)
02:28:17tarcierisooo
02:28:32tarcieriMenTaLguY: the easiest way to hook in for IO events is... through IO itself :/
02:28:41tarcieridelivering messages through the existing mechanisms
02:28:45tarcieribut that requires changes to IO
02:28:45tarcieriugh
02:28:56yugui enters the room.
02:28:56headiusrue: human race...piffle!
02:29:18rueWell, you know how the average goes
02:29:42ko1_ leaves the room.
02:29:44MenTaLguYtarcieri: I still don't see why really
02:30:07tarcieriMenTaLguY: multiplexing messages generated via IO with messages from other Actors
02:30:29MenTaLguYI think the correct answer is to have an "IO actor" to which actual IO can be delegated
02:30:41tarcieriI suppose
02:30:44ko1_away enters the room.
02:30:46tarcieriThat's certainly not the solution Erlang took
02:30:56MenTaLguYit is, kind of
02:31:02tarcieriKind of...
02:31:03dblackof course now i can't remember what i did to get it to hang
02:31:05MenTaLguYalthough it's less explicit about it
02:31:12tarcieriThat's originally what I wanted to do
02:31:23tarcieriHave an Actor/Thread/whatever running a Reactor loop
02:31:25VVSiz_ enters the room.
02:31:28tarcierithat's how Revactor does it
02:31:40tarcieriBut in an architecture like Rubinius it seems unnecessary
02:32:15ttmrichter_ enters the room.
02:32:18dblackit was this: bin/mspec spec/library/actor/
02:32:24dblackwhich i perhaps have no business doing
02:33:22rueOh, are we allowed to idle in #ruby-core? I am sure the conversation will go much better if just had by the head honchos but I am nosey
02:33:51headiusI don't see why not, but obviously we don't want a free-for all
02:34:33tarcieridblack: I encounter errors doing the same thing
02:35:14dblacktarcieri: does it hang?
02:35:24rueYep, I figure any more than you lot actively participating will just be detrimental
02:35:31tarcieridblack: yep
02:35:39dblacki shall leave the honchos in splendid isolation
02:35:40MenTaLguYhm, what must one do to get access to the pbwiki?
02:36:00dblackand look forward to hearing reports from all quarters
02:36:36antares enters the room.
02:36:38rueMenTaLguY: You need to have an account on pbwiki and then be invited to the particular one if I recall correctly
02:36:51ruedblack: No quarter!
02:36:56MenTaLguYah
02:37:01MenTaLguYwell, I've got the account anyway
02:37:18MenTaLguYnot entirely sure if I rate wiki access or not
02:37:35VVSiz leaves the room.
02:38:20twbray enters the room.
02:38:30rueI propose that we submit a motion to consider appointing a committee to perform preliminary planning for a panel to evaluate a hierarchical approach to these design meetings
02:38:50tarcieriheh
02:39:21dblackrue: your proposal will be duly considered by the proposal consideration taskforce
02:39:57rueI shall fill out the request for an expression of joy
02:46:30ttmrichter leaves the room.
02:52:32MenTaLguYI like pwiki; it's really slick.
02:52:36MenTaLguY(never used it before)
02:53:00antaresMenTaLguY: I cannot open it up :( DNS issues or something
02:53:01evanyeah
02:53:04evanit's nice
02:53:12evanthey've made it nicer recently
02:53:18evanthe templates make it extra nice
02:53:35evannothing ground breaking though (the case with all wikis)
02:54:19MenTaLguYno, but nice to use is more important than groundbreaking as long as it does the important things well
02:54:58evanyep
02:55:53aotearoa leaves the room.
02:57:12rueUnless you are a geologist
02:57:25lopex leaves the room.
03:09:09macournoyer enters the room.
03:10:30seydar enters the room.
03:11:18dblack leaves the room.
03:11:32boyscout1 commit by Ryan Davis
03:11:33boyscout * Added latest_incremental_hash.txt; 1907e0a
03:20:14wycats leaves the room.
03:20:31chop3 enters the room.
03:21:20rubuildius_amd64Ryan Davis: 1907e0a66; 2090 files, 6619 examples, 23154 expectations, 0 failures, 0 errors; http://rafb.net/p/LZoSK369.html
03:21:35seydarAnd1
03:21:37seydar leaves the room.
03:23:51rubuildius_ppcRyan Davis: 1907e0a66; 2090 files, 6621 examples, 23180 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184618
03:25:59jero5 leaves the room.
03:26:50macournoyer leaves the room.
03:27:28macournoyer enters the room.
03:27:54marnen enters the room.
03:28:54boyscout1 commit by MenTaLguY
03:28:54boyscout * fix up registration spec; d7a7d0c
03:35:09boyscout1 commit by MenTaLguY
03:35:11boyscout * fix linked actors spec (sort of); 054582f
03:36:33rubuildius_amd64MenTaLguY: d7a7d0c4d; 2090 files, 6619 examples, 23154 expectations, 0 failures, 0 errors; http://rafb.net/p/5Y5M3r47.html
03:37:30mernen leaves the room.
03:38:01wmoxam_ enters the room.
03:38:19wmoxam leaves the room.
03:39:11wycats enters the room.
03:46:23rubuildius_amd64MenTaLguY: 054582f3b; 2090 files, 6619 examples, 23154 expectations, 0 failures, 0 errors; http://rafb.net/p/dLY5UL63.html
03:50:10rubuildius_ppcMenTaLguY: 054582f3b; 2090 files, 6621 examples, 23180 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184631
03:50:11rubuildius_ppcMenTaLguY: d7a7d0c4d; 2090 files, 6621 examples, 23180 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184629
03:53:37wdperson enters the room.
03:59:57macournoyer leaves the room.
04:02:23twbray leaves the room.
04:04:30ttmrichter_ leaves the room.
04:04:43benny enters the room.
04:05:15ttmrichter_ enters the room.
04:07:41imajes leaves the room.
04:09:51zenspider leaves the room.
04:11:51fbuilesv leaves the room.
04:17:24d2dchat enters the room.
04:19:42jtoy enters the room.
04:23:48ttmrichter__ enters the room.
04:30:23wycats leaves the room.
04:34:33AndrewO leaves the room.
04:34:58enebo leaves the room.
04:38:35ttmrichter_ leaves the room.
04:45:40agile leaves the room.
04:56:23kirindav_ enters the room.
05:02:20elight enters the room.
05:03:25headius leaves the room.
05:10:41zico leaves the room.
05:11:30antares leaves the room.
05:12:59evanmore cpp code is up
05:13:06evancompiles clean on linux again
05:13:09evanclean on valgrind too
05:15:23benny leaves the room.
05:24:13KirinDave leaves the room.
05:32:01RyanTM leaves the room.
05:32:21d2dchat leaves the room.
05:37:03dlee enters the room.
05:41:48be9 enters the room.
05:45:28pastie leaves the room.
05:47:46ttmrichter__ leaves the room.
05:48:56KirinDave enters the room.
05:50:36wycats enters the room.
05:56:49jennyw enters the room.
06:02:47agardiner leaves the room.
06:06:48eventualbuddha enters the room.
06:11:42elight leaves the room.
06:16:30ttmrichter enters the room.
06:16:55eventualbuddha leaves the room.
06:18:59twbray enters the room.
06:21:04benburkert leaves the room.
06:29:27headius enters the room.
06:31:58headius_ enters the room.
06:33:29boyscout4 commits by Marnen Laibow-Koser
06:33:31boyscout * Continue implementing bits of #inspect and updating spec tags.; e1fc7c6
06:33:32boyscout * Implement BigDecimal#==/eql? and the beginnings of #inspect.; e4371f1
06:33:33boyscout * Start implementing BigDecimal#sign and #zero?.; 0494c1c
06:33:33boyscout * Get BigDecimal#abs working.; 3c071b5
06:34:06evango marnen go!
06:34:09marnen:)
06:34:37marnenThis is what I do when I should be sleeping.
06:35:10wycats leaves the room.
06:37:57evanheh
06:38:00marnenOf course, this is the easy stuff. I sense a refactoring coming on as my naïve first stab starts to crack under its own weight.
06:38:01evanyou're not the only one.
06:38:08marnenI'd be surprised if I were.
06:38:38evanrefactor is encourgage with specs
06:38:43evanencouraged
06:38:59marnenI know. That's one of the reasons I like [TB]DD.
06:39:09ezmobius_ enters the room.
06:39:28ezmobius_ leaves the room.
06:39:37marnenor rather, not that refactoring is encouraged so much as that it's easier to tell if you refactored it right!
06:39:43evanyeah, having the new VM have lots of tests is great
06:39:47evani can really feel the code better now.
06:40:36marnenI probably wouldn't be contributing to this project if that weren't the case...I've never really done compiler- or VM-type stuff before. Having the specs made me a lot less afraid to dive in.
06:40:58evanwonderful!
06:41:03evanhows BigDecimal?
06:41:10marnenComing along.
06:41:11evanjust poke, trying to fill it out a bit.
06:41:23rubuildius_amd64Marnen Laibow-Koser: e1fc7c6dc; 2090 files, 6635 examples, 23217 expectations, 0 failures, 0 errors; http://rafb.net/p/06omWK95.html
06:41:28marnenThat's how.
06:41:33marnen:)
06:41:35evansweet.
06:41:44bricolage enters the room.
06:41:50evanjust tested and implement raising exceptions in rubinius
06:41:54evannice and tight now.
06:41:58marnenlovely!
06:42:02evan15 lines
06:42:06marnen!
06:42:10marnenWhat is this, Lisp?
06:42:15evan:)
06:42:20evani'll push, you can see.
06:42:46marnenor APL? :D
06:42:58evanhehe
06:43:03marnenoh, wait, no, it uses letters
06:43:05evanyeah 15 lines, but 80 columns per line
06:43:07evan:)
06:43:16evanuh oh
06:43:18evanis github down?
06:43:30marnenmaybe...my push hung when it tried to go there.
06:43:41evanhanging for me too
06:43:44evanit went ok though?
06:43:50marnento origin, yes.
06:43:55evandid you let it time out?
06:43:58evanor ^C it?
06:44:08evani should put a timeout on it
06:44:09marnenI aborted it after waiting a good long while.
06:45:19marnenshit, it's late here in NY. Gotta go to sleep.
06:45:22marnenGood night!
06:45:43marnen leaves the room.
06:46:00evannite!
06:46:01evanoh well.
06:46:39atmos leaves the room.
06:46:43atmos enters the room.
06:47:59Maledictus enters the room.
06:48:41headius leaves the room.
06:49:09rubuildius_ppcMarnen Laibow-Koser: e1fc7c6dc; 2090 files, 6637 examples, 23243 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184685
06:58:02ctennis_ enters the room.
06:58:39ctennis leaves the room.
07:07:46benburkert enters the room.
07:12:15dbussinkmorning
07:12:17dbussinkare there logs from the meeting yet?
07:12:27sudoer enters the room.
07:13:33evanyep
07:13:44evanhttp://ruby-design.pbwiki.com/Design20080421
07:15:56headius_ leaves the room.
07:16:15sudoer leaves the room.
07:18:17ingsoc_ enters the room.
07:21:56dbussinkah, lets read up on that then
07:29:22sudoer enters the room.
07:29:42jtoy leaves the room.
07:31:08agardiner enters the room.
07:34:23KirinDave leaves the room.
07:38:49jtoy enters the room.
07:47:40rueevan: Core all seemed very open to it although I was a bit surprised they seemed somewhat unfamiliar with recent developments in testing
07:48:54rueErm, "it" being the regression testing and specs
07:50:46evanyeah
07:50:48evan*shrug*
07:50:52dbussinkyeah, was kinda surprised they didn't even investigate themselves
07:50:54evanthe good thing is there is movement.
07:51:02dbussinkmore ignorance than non willingness
07:57:32sudoer leaves the room.
07:58:55dlee leaves the room.
08:00:58ingsoc_ leaves the room.
08:14:27VVSiz leaves the room.
08:15:19dctanner enters the room.
08:24:05mutle enters the room.
08:25:38zimbatm enters the room.
08:26:00twbray enters the room.
08:30:56brixendbussink: (supposed to be asleep, but)... there's another good use of toplevel shared dir
08:31:16dbussinkme asleep?
08:31:18brixendbussink: it_mixes_in Enumerable, File
08:31:21dbussinkor are you? :P
08:31:23brixenno me :P
08:31:27dbussinkhehe
08:31:48brixendbussink: that way you can write a set of specs for what an Enumerable should do, but be able to run them in the context of that class
08:31:50lstoll leaves the room.
08:32:12brixenI found a couple days ago that File was not an Enumerable, and realized we need a simple way to spec that without repeating everything
08:33:51dbussinkyeah, that's probably a useful use then
08:35:22TheVoice leaves the room.
08:38:12agardiner leaves the room.
08:43:28twbray leaves the room.
08:44:49benburkert leaves the room.
08:44:49twbray enters the room.
08:45:58twbray leaves the room.
08:47:51benburkert enters the room.
08:49:56zimbatm_ enters the room.
08:50:03zimbatm leaves the room.
08:54:47thehcdreamer enters the room.
09:02:04octopod enters the room.
09:07:11dctanner leaves the room.
09:10:01dlee enters the room.
09:19:29benburkert leaves the room.
09:20:58yaroslav enters the room.
09:25:51sudoer enters the room.
09:29:02olabini enters the room.
09:34:35jtoy leaves the room.
09:37:05kw leaves the room.
09:40:00obvio leaves the room.
09:43:59SpakmanHas anyone tried running ruby-gtk2 with Rubinius? I would, but I'm having trouble building Rubinius just now..
09:44:08dlee leaves the room.
09:49:58BlackEdder enters the room.
09:53:51GMFlash leaves the room.
10:00:10dctanner enters the room.
10:12:50chris2 enters the room.
10:16:52olabini leaves the room.
10:36:53Arjen_ enters the room.
10:42:35qwert666 enters the room.
10:46:05yaroslav leaves the room.
10:46:11riffraff enters the room.
10:48:27yaroslav enters the room.
10:51:29riffraffhi
11:00:28kAworu enters the room.
11:11:31VVSiz enters the room.
11:12:16lstoll enters the room.
11:18:37be9 leaves the room.
11:28:17peeja leaves the room.
11:37:00rueHeya
11:37:19rueSpakman: I doubt the extension builds without some tweaking
11:40:35Spakmanthanks rue
11:41:30zimbatm__ enters the room.
11:43:24zimbatm_ leaves the room.
11:52:49antares enters the room.
11:58:51antares leaves the room.
12:02:34yaroslav leaves the room.
12:04:35qwert666 leaves the room.
12:07:58dblack enters the room.
12:08:15radarek enters the room.
12:09:31sudoer leaves the room.
12:16:36Fullmoon leaves the room.
12:19:47qwert666 enters the room.
12:23:44danlucraft enters the room.
12:24:17danlucraftSpakman, rue: I am even as we speak trying to get Ruby-Gnome2 building for Rubinius (well, last night)
12:24:59hornbeck leaves the room.
12:25:25hornbeck enters the room.
12:25:36danlucraftI'm trying to get a report that will automatically tell me what's missing from subtend.
12:25:41danlucraftThere's quite a lot.
12:26:00danlucraftRuby-Gnome2 seems to exercise the entire Ruby-C API
12:27:31Spakmandanlucraft: I must go for lunch now, but I'd love to hear how you get on :)
12:33:15foysavas leaves the room.
12:33:36foysavas enters the room.
12:33:51foysavas leaves the room.
12:34:04foysavas enters the room.
12:35:48danlucraftOK, I'll report progress here. :)
12:35:55danlucraft leaves the room.
12:36:25yaroslav enters the room.
12:43:54ctennis_ leaves the room.
12:48:24rueShould probably go to bed
12:55:43mutle leaves the room.
12:58:07chris2 leaves the room.
13:00:40wdperson enters the room.
13:01:49jennyw leaves the room.
13:02:28yaroslav leaves the room.
13:03:07yaroslav enters the room.
13:04:47mutle enters the room.
13:06:47chop3 leaves the room.
13:08:15ctennis enters the room.
13:18:46wycats enters the room.
13:24:02yaroslav leaves the room.
13:24:24yaroslav enters the room.
13:27:40RyanTM enters the room.
13:33:54dblack leaves the room.
13:34:06dctanner leaves the room.
13:35:14be9 enters the room.
13:38:50riffraff leaves the room.
13:48:05yaroslav leaves the room.
13:48:37marnen enters the room.
13:48:53headius enters the room.
13:49:17srbaker enters the room.
13:53:39rueOK, now I go.
13:57:12headius leaves the room.
14:03:55srbaker leaves the room.
14:05:54srbaker enters the room.
14:14:19qwert666 leaves the room.
14:17:58fbuilesv enters the room.
14:18:09rueOr maybe now
14:18:15rueBe back in a little
14:21:21AndrewO enters the room.
14:33:48radarek leaves the room.
14:37:45radarek enters the room.
14:39:44kw enters the room.
14:43:38boyscout3 commits by Marnen Laibow-Koser
14:43:39boyscout * Implement BigDecimal#precs.; d6503e8
14:43:40boyscout * Implement BigDecimal#precs.; d0171de
14:43:41boyscout * Get precs working better, especially for Infinity and NaN.; bfdd091
14:50:38kw leaves the room.
14:51:20rubuildius_amd64Marnen Laibow-Koser: d6503e8eb; 2090 files, 6638 examples, 23245 expectations, 0 failures, 0 errors; http://rafb.net/p/SMi69A36.html
14:55:18moofbong enters the room.
14:55:25boyscout1 commit by Marnen Laibow-Koser
14:55:27boyscout * Fix BigDecimal#zero, update spec tags. Also make #precs deal correctly with lowercase ...; f1b2bf5
14:59:52smparke1 leaves the room.
15:00:40fbuilesvmarnen: have you been able to find something about the Singleton.new spec error you were getting?
15:00:58marnenNo. It still happens about every second time.
15:01:22fbuilesvmarnen: does it happen only running rbx or does it happen with MRI too?
15:01:36marnenHaven't checked with MRI yet. Let me try that.
15:04:20marnenSeems to be just with Rubinius.
15:04:44Illocution enters the room.
15:05:22marnenOTOH, MRI has some problems with the REXML specs.
15:05:45fbuilesvmarnen: some of those specs were tagged as failing
15:06:05marnenI wonder if I screwed up my tag files, then.
15:06:20rubuildius_amd64Marnen Laibow-Koser: f1b2bf510; 2090 files, 6642 examples, 23257 expectations, 0 failures, 0 errors; http://rafb.net/p/Pmnnce45.html
15:06:45fbuilesvmarnen: check which ones are failing, they should be inside a ruby_bug guard
15:07:04marnenyou mean the REXML ones?
15:07:14fbuilesvmarnen: yes
15:07:23marnenok
15:07:55rubuildius_ppcMarnen Laibow-Koser: f1b2bf510; 2090 files, 6644 examples, 23283 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184800
15:07:56rubuildius_ppcMarnen Laibow-Koser: d6503e8eb; 2090 files, 6640 examples, 23271 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184792
15:08:40marnenSure doesn't look like it. /spec/ruby/1.8/library/rexml/document/new_spec.rb has no guards.
15:10:13fbuilesvmarnen: that's pretty weird, those ones run just fine for me
15:10:19fbuilesvmarnen: what version of Ruby are you using?
15:10:29marnen1.8.6
15:10:41marnenMac OS X 10.5.2, Intel
15:10:46fbuilesvp114?
15:10:51marnenno, 111
15:10:53obvio enters the room.
15:10:55obvio leaves the room.
15:10:57marnenmaybe that's the issue
15:11:04obvio enters the room.
15:11:19fbuilesvcould you upload the errors you're getting? I'll try to get a p111 to try
15:11:54marnenhttp://pastie.caboo.se/184803
15:12:41fbuilesvmarnen: that's an error with REXML, pre 1.9 that library's pretty fucked up
15:12:43dctanner enters the room.
15:12:46marnenoic
15:12:59marnenThen I won't worry about it.
15:16:27jennyw enters the room.
15:18:53d2dchat enters the room.
15:19:39bricolage leaves the room.
15:21:35wmoxam enters the room.
15:29:29GMFlash enters the room.
15:31:09dblack enters the room.
15:35:45macournoyer enters the room.
15:37:14twbray enters the room.
15:40:20foysavas leaves the room.
15:42:10trythil enters the room.
15:45:21probablycorey enters the room.
15:45:40foysavas enters the room.
15:45:50marnen_ enters the room.
15:46:07twbray leaves the room.
15:52:09marnen leaves the room.
15:52:44fbuilesv leaves the room.
15:53:50smparke1 enters the room.
15:56:20benstiglitz enters the room.
15:57:28wycats leaves the room.
15:57:54benstiglitzOn the bus to work I had a terrible thought about Rubinius + L4, and the benefits of implementing stdlib in Ruby.
15:58:17scoopradd llvm to the mix too :)
15:58:33benstiglitzMm.
16:07:53KirinDave enters the room.
16:20:00obvio171 enters the room.
16:23:23lopex enters the room.
16:25:22obvio leaves the room.
16:28:06w1rele55 leaves the room.
16:29:28benburkert enters the room.
16:37:28TheVoice enters the room.
16:38:12dblack leaves the room.
16:40:59jlindley enters the room.
16:41:49therealadam enters the room.
16:45:51dlee enters the room.
16:47:26w1rele55 enters the room.
16:50:08agile enters the room.
16:50:17jennyw leaves the room.
16:54:27dctanner leaves the room.
16:55:53enebo enters the room.
16:56:45evanmorning
16:57:18djwhittmorning
16:57:33djwhittread ola's blog post about the meeting last night
16:57:38djwhittsounds like things went well
16:59:05anteaya enters the room.
17:08:00headius enters the room.
17:08:25KirinDave leaves the room.
17:08:27twbray enters the room.
17:09:06mutle leaves the room.
17:15:16benburkert leaves the room.
17:17:34jennyw enters the room.
17:20:59Fullmoon enters the room.
17:29:55twbray leaves the room.
17:31:55boyscout1 commit by Marnen Laibow-Koser
17:31:56boyscout * Change to a significand-and-exponent implementation.; 3e6f16c
17:32:56KirinDave enters the room.
17:36:28dgtizedevan: test_task.hpp:858: undefined reference to `rubinius::Exception::create(rubinius::VM*)'
17:36:38evanhrm.
17:36:45evani may have gotten a file.
17:36:55evanyep
17:36:57evanone sec.
17:38:03evanok, in.
17:38:49thehcdreamer leaves the room.
17:39:17boyscout1 commit by Wilson Bilkovich
17:39:18boyscout * Rename ambiguously-worded autoload spec; 527a4b6
17:39:34DefilerI am totally going to get Merb running today
17:39:45DefilerI have a radar lock on it
17:40:18dleeis rubinius' C++ VM done?
17:40:20evanawesome
17:40:22evandlee: no
17:40:35dleehow will Defiler get Merb running?
17:41:04dbussinkhe aims, locks and shoot
17:41:05evanusing the current one.
17:41:05djwhittmagic (and the C base VM)
17:41:07dbussinkshoots
17:41:32Defilerdlee: Are you saying that Merb has an evaluation order dependency?
17:41:42DefilerIf it does, I haven't run into it yet
17:42:54headiusyou guys find that blocking/mysql thing?
17:43:04evanyep
17:43:12evani'll let Defiler explain.
17:43:36dbussinkevan: that's cruel ;)
17:43:40DefilerWell, you see.. I am a primitive ape-like creature, and evan is a cyborg sniper from the future.
17:43:47DefilerSo I made him fix it.
17:43:48dleei'm not sure if merb uses it... i would assume not because ezmobius insists strongly on clarity
17:44:32gnufied enters the room.
17:44:35dleetechnical question... what do you call the layer on top of the VM that implements everything in Ruby?
17:44:48evankernel
17:44:48Defiler'the kernel'
17:45:16dleeis there a layer on top of that, or is that the layer that Ruby programmers see?
17:45:21evanheadius: when @sock.read(4) was run, it pulled in like 70 some bytes into it's buffer
17:45:24DefilerRubinius is (currently) 'shotgun' (the vm and grammar), 'the kernel' (the ruby stuff) and 'the compiler' (lib/compiler)
17:45:40evanheadius: but IO#read didn't simple return data from it's buffer, it was always trying to fill the buffer more
17:45:43evanand thus the block.
17:45:55headiusnice
17:45:58evanso the 2nd @sock.read(len) blocked, even though the buffer had +len+ bytes
17:46:08headiusright
17:46:10DefilerI read that code and thought "well sure it makes sense to fill up the buffer if you can"
17:46:13Defilerbut durh
17:46:27headiusyeah, you only fill it if you need more than it has
17:47:05rubuildius_amd64Wilson Bilkovich: 527a4b663; 2090 files, 6645 examples, 23321 expectations, 0 failures, 0 errors; http://rafb.net/p/Ibuvnq33.html
17:47:06rubuildius_amd64Marnen Laibow-Koser: 3e6f16c41; 2090 files, 6645 examples, 23321 expectations, 0 failures, 0 errors; http://rafb.net/p/hefl4132.html
17:50:35nicksieger leaves the room.
17:51:40nicksieger enters the room.
17:52:16yaroslav enters the room.
17:52:18srbaker leaves the room.
17:52:40DefilerIs there a trick to adding a new file to kernel and referencing it on a 'depends' line?
17:52:48boyscout1 commit by Vladimir Sizikov
17:52:49boyscout * More detailed specs for BigDecimal#sqrt and fixes for old ones.; e7894fb
17:53:00DefilerI'm getting a can't depend on non-existent file blah blah kind of error
17:53:14dleehas all of Kernel been implemented already?
17:53:22dleeor are there still some parts missing?
17:53:23evanDefiler: when are you getting that?
17:53:32boyscout1 commit by Marnen Laibow-Koser
17:53:33boyscout * Fix a poorly written condition in BigDecimal#eql?.; e30a039
17:53:38evandlee: there are still parts missing
17:53:38Defilerre-running the build now to see exactly
17:53:55dleeevan: where can i get a list of what's missing?
17:54:11evanthere is no such list.
17:54:20evanyou can run the specs and see what doesn't work
17:54:21VVSizwork on BigDecimal in parallel (in JRuby and Rubinius) :)
17:54:31evanbin/mspec run spec/ruby/1.8/core
17:54:42evanany failures or errors are things missing/broken in the kernel
17:54:44Defilerhuh. never mind. rake clean fixed it
17:55:45gnufiedevan, what was the context of @sock.read(4) ? Dunno, but I filed a bug report some days ago, about them.
17:56:08evangnufied: i don't follow, please rephrase
17:56:40KirinDave leaves the room.
17:56:42headius leaves the room.
17:56:59headius enters the room.
17:57:53gnufiedevan, oh. never mind, i will read the archives. sorry to bother. BTW,I was talking about IO.read issue that you and charles were discussing sometime ago.
17:58:06evanwhich one was that?
18:00:11benburkert enters the room.
18:00:20rueBraaaaaaainnnsss
18:01:14headius_ enters the room.
18:01:22rubuildius_amd64Marnen Laibow-Koser: e30a039ca; 2090 files, 6645 examples, 23321 expectations, 0 failures, 0 errors; http://rafb.net/p/QC4ulG93.html
18:02:17benstiglitzWhile we’re hitting on IO, can we have a quick discussion about non-buffered IO?
18:02:30anonuser enters the room.
18:02:36rueY
18:02:38rueE
18:02:39dbussink#330 also does some nasty things to io
18:02:40rues
18:03:05benstiglitzI think this came up quickly before, but things like pipes are not seekable.
18:03:19benstiglitzWhich really breaks with the current implementation of buffered IO, which does some seeking-around junk.
18:03:40benstiglitzDoes implementing a private #seekable? sound good? I don’t see why we shouldn’t buffer, but we definitely don’t want to seek.
18:03:55headius leaves the room.
18:03:57benstiglitzThings like sockets and pipes would return NO, and get the right behavior.
18:04:19benstiglitzPipe pairs are a little tricky only because they’re not a subclass of IO, but we can have a setter, too.
18:04:22benstiglitzThoughts?
18:04:46evanhm
18:04:49evanyeah, i think thats ok
18:04:56evanthe seek() stuff is to reset the buffer
18:04:58headius enters the room.
18:05:02evanbut agreed, it's very File centric
18:05:09headius leaves the room.
18:05:29evansince it's File centric
18:05:37benstiglitz…maybe it should go there.
18:05:39evanwhy not move part of that implementation to the File class
18:05:48evanand have it call super into IO
18:05:59evanso File#write would be
18:06:07evanseek_and_reset_buffer
18:06:08evansuper
18:06:09evanend
18:06:13boyscout1 commit by Vladimir Sizikov
18:06:14boyscout * One more test case, for BigDecimal#sqrt with nil.; 8eb5451
18:06:22benstiglitzSounds good. I’ll take that chunk, then.
18:06:28evanok
18:07:44srbaker enters the room.
18:09:25dgtizedevan: how many failures on cpp vm test suite?
18:09:41evanon linux, 3 i think.
18:09:46dgtizedk, same here
18:11:18yaroslav leaves the room.
18:13:54marnen_ leaves the room.
18:16:24rubuildius_amd64Vladimir Sizikov: 8eb5451f8; 2090 files, 6645 examples, 23321 expectations, 0 failures, 0 errors; http://rafb.net/p/93TSlt18.html
18:19:15headius_ leaves the room.
18:20:05rubuildius_ppcVladimir Sizikov: 8eb5451f8; 2090 files, 6647 examples, 23347 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184911
18:20:06rubuildius_ppcMarnen Laibow-Koser: e30a039ca; 2090 files, 6647 examples, 23347 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/184905
18:20:07rubuildius_ppcVladimir Sizikov: e7894fb