Index

Show enters and exits. Hide enters and exits.

00:00:39evanDefiler: sup?
00:01:11Defilerevan: OK. Let's say I have a line of code on line 1 of a file.. "spec do 5 end"
00:01:22Defilerand spec is just something that captures the block, as per the earlier example
00:01:49DefilerBefore that block pass, when we've just got this BlockEnvironment we created..
00:02:04anonuserbah.
00:02:26DefilerWhat is the relationship between that BlockEnv instance and the context object that would show 'line 1' if asked to print in a backtrace right that second
00:02:34mass anonuser: are you anonymous user or a non-user?
00:02:55evanbetween which context object
00:03:04evanit has 3 relationships to context objects
00:03:18evan1) home: the MethodContext it was created in (where self is)
00:03:19DefilerRight. It has a 'home' and 'home' block, which could be different
00:03:24nexcastellananonuser: Actually, I'd like to hold off my commit but if you give me your email address, I'll email you a diff.
00:03:25Defilerand then it has a BlockContext later when it is run
00:03:31Defilerany others?
00:03:35evan2) home_block: the contents of MethodContext.current when it was ceated
00:03:37anonusermass, neither.
00:04:00evan3) created context: when #call is sent to the BlockEnv, a new context is created to run the code it contains
00:04:04atduskgreg leaves the room.
00:04:13DefilerOK, I need the other one that isn't on that list.. the one that MRI prints. Heh
00:04:22evanthere is no other one.
00:04:35DefilerSomething has to be pointing at 'this_file.rb:1'
00:04:43DefilerBecause that is where we are executing right now
00:04:51evanin rubinius or MRI?
00:04:54DefilerBoth
00:05:07evanwell, in rubinius, it's one of those 3
00:05:18evanin MRI, as headius said, it's duping frames internally
00:05:26DefilerSee, that turns out not to be it
00:05:34anonuserokay I'm seg faulting on swapping contexts;
00:05:35evanso there is no mapping between the 3 I gave you
00:05:37Defilerbecause we don't implement the thing MRI is duping, as far as I can see
00:05:53evanDefiler: that doesn't change that there are 3 contexts in question
00:05:56Cosmos95 enters the room.
00:06:09DefilerMRI dupes something that says "you are defining a block at line 1 in file1.rb and the method call this is a block for is 'spec'"
00:06:18dbussinkbrixen: gah, i can't even get a complete full spec run on my 64 bit box
00:06:19anonuserhpricot is trying to define an attribute "buffer_size" and the subtend layer just does not like that.
00:06:19Defilerbut it seems to me we don't have anything that retains that state
00:06:32Defilereven immediately, as the BlockEnvironment is instantiated
00:06:39dbussinkbrixen: it's hanging an insane amount of time
00:06:54brixendbussink: yeah, noticed that too with trying to sync the specs
00:06:56nexcastellananonuser, give me your email, I'll send you my patch. Seems to make RMagick compile but then it craps out on "TypeError", almost certainly because of my RData bit.
00:07:03brixendbussink: are you on the old specs or did you try to sync them?
00:07:17dbussinkbrixen: just running rake spec:full on the latest head
00:07:25brixenhmm
00:07:32DefilerWe hang on spec:full, right?
00:07:36anonusernexcastellan, sean@cyberwang.net or sean@hackinggibsons.com
00:07:53anonuserpick one :)
00:07:55brixendbussink: yep, same here
00:07:58dbussinkDefiler: well, probably it will get there eventually
00:08:12dbussinkjust like rubuildius_amd64, that one usually takes a huge amount of time too
00:08:14DefilerI thought that readlines("") thing just spun forever?
00:08:22anonusernexcastellan, run it in gdb and follow the backtrace
00:08:28DefilerMaybe I am mixing up all the rake tasks we have now
00:08:32DefilerSince we have 100
00:08:32brixenDefiler: we haven't been hanging on spec:full
00:08:37brixenthat's what the bots run
00:08:38nexcastellananonuser, yes but my bus comes in ten minutes. :)
00:08:43anonuseralright
00:08:46anonuserI'll see about it.
00:09:13nexcastellananonuser, emailed.
00:09:15DefilerI guess where I am going with this, evan, is this..
00:09:20nexcastellanG'night, all.
00:09:23ezmobius_ enters the room.
00:09:26Defilerfoo { insert_code_here }
00:09:33anonuserbut first to see why rb_define_attribute's failing :(
00:09:38anonusernexcastellan, thanks
00:09:58dbussinkbrixen: but i really need to get some sleep
00:10:01DefilerSorry, wrong. Mixing up passes. Ignore what I just said
00:10:03dbussink1 am over here
00:10:12brixendbussink: n/p
00:10:53evanDefiler: perhaps you should draw a picture in omni graffle
00:10:58evanof the object relationships
00:11:00DefilerYeah, I drew one on paper
00:11:04evanand another of the control flow
00:11:06Defilerbut if you want to see it too I can graffle it
00:11:15evanmight be nice
00:11:19evansince i'm sure we'll forget this
00:11:24drbrainwe should do a crayon one
00:11:32DefilerWhile I'm doing that, did I at least get across my question? Heh.
00:11:38evanno
00:11:44evanstill don't see where you're going
00:11:51evani know that we're missing information
00:11:58evanwe figured that out yesterday
00:12:00evanor the day before.
00:12:08DefilerOK, I'll give it to you in terms of kernel code
00:12:17evanbecause home_block continues to run, destroying the ip that the block was created at
00:12:24Defilertake a look at kernel/core/proc.rb and look at def self.__from_block__ or from_environment
00:12:46evank, i see it.
00:12:56Defilerif line 1 of file1.rb is: foo { baz }.. when that code runs, which context frame contains 'file 1, line 1' as its position information
00:13:07Defilersorry.. by "that code" I mean __from_block__
00:13:09evanwhich context frame.
00:13:15DefilerAny context frame
00:13:15evangive them names
00:13:18evanok
00:13:32DefilerThat has line 1 of file 1 as its instruction pointer
00:13:35evanare you asking me?
00:13:38DefilerYes
00:13:39evanoh.
00:13:41DefilerThat is the question
00:13:51evanwhen the block is executed
00:13:54evanright?
00:13:59DefilerNo, when it is block_passed
00:14:04evanie, yield or Proc#call is done.
00:14:22Defiler__from_block__ is called potentially far in advance of the actual Proc#call, if any
00:14:24evanok, so when 'def foo(&block)' is executed
00:14:25evanyes?
00:14:27Defileryes
00:14:31Defilerwell, no
00:14:39evanthere is no context yet for executing 'baz'
00:14:40Defiler__from_block__ runs when the block is actually passed, right?
00:14:44evanit doesn't exist yet.
00:14:54evanit runs both places
00:15:12evanon the caller side for "foo(&blah)"
00:15:21evanand on the reciever side for
00:15:27evan"def foo(&block)"
00:15:35Defiler__from_block__ runs when the block is actually passed, right?
00:15:39Defilerargh
00:15:40Defilerfreaking irc
00:15:41evanboth.
00:15:43evancall me.
00:15:45DefilerThere's a compile pass, and then there's a "foo is running now, and has a block associated with it to call __from_block__ on", right?
00:16:11evanjust call me
00:21:09anonuserheh. when the internet communication lines break down.
00:23:05qwert666_ leaves the room.
00:25:29loincloth enters the room.
00:33:47atduskgreg enters the room.
00:41:05headiusevan: where are you guys storing backref right now
00:41:27headiusbackref is the goddamn bane of my existence
00:42:38vertiginous leaves the room.
00:43:04imajes leaves the room.
00:43:31headius_ enters the room.
00:44:42headiusgrr
00:45:19headius leaves the room.
00:48:00ShayArnett enters the room.
00:51:21mernen enters the room.
00:59:18robertpostill enters the room.
00:59:20robertpostill leaves the room.
01:05:12lopex leaves the room.
01:07:31atduskgreg leaves the room.
01:07:53wmorganpreliminary inliner benchmarks: http://all-thing.net/2008/06/preliminary-rubinius-inliner-benchmarks.html
01:08:41drbrainwmorgan: sweet
01:13:19tarcierinice
01:13:28tarcierican't wait for some JIT :)
01:14:45ezmobiuswmorgan: killer
01:15:24evanwmorgan: cool!
01:18:42evanwmorgan: one nice thing about the way I'm structuring the LLVM integration is that we can just use your inliner straight up and feed the output to LLVM to create new machine code
01:18:55tarcieriw00t
01:19:28ezmobiushows the c++ vm coming?
01:20:05evangood
01:20:15wmorganevan: inliner+llvm is a match made in heaven. especially if you can do something with those meta_send_plus opcodes...
01:20:16evandid ya see the sample output for the while loop
01:20:24ezmobiusyeah
01:20:34ezmobiusexciting
01:21:00wmorganevan: also the inlined methods are full of redundant stack motion that llvm will optimize out
01:21:26evanyeah, i'm doing experiments right now to
01:21:40evanto figure out how to express the stack to LLVM to make it optimize the most
01:21:58evani think that letting LLVM see it as a OBJECT** is the trick
01:22:04evanbut still verifing that
01:25:22ruivaldo enters the room.
01:25:33ruewmorgan: Nice work
01:26:47tomy_ enters the room.
01:27:26wmorganthanks all!
01:27:26rueAlso with Sup :)
01:27:37bitsweat leaves the room.
01:27:55wmorganrue: sup on rubinius is an experiment i'm too scared to try
01:28:09wmorganbut one day...
01:28:20ruewmorgan: Hehe, maybe it would finally rid all the Redwood stuff ;)
01:28:31wmorganthen i won't have to benchmark it. it will just be super fast!
01:28:46wmorganrue: that is the least of my worries
01:30:02anonuseris back
01:30:10yugui enters the room.
01:42:28AndrewO enters the room.
01:43:00anteaya leaves the room.
01:43:26krisps leaves the room.
01:47:17anteaya enters the room.
01:48:10benny enters the room.
01:48:21Arjen__ leaves the room.
01:54:23naeu enters the room.
01:56:08rueset_ffi_lib is such a hack, man
01:56:22tomy_ leaves the room.
01:57:18evanshall we do
01:57:42evanclass IConvLib < FFI.library "libiconv"
01:57:46evan:)
01:57:55rueGo to your room, young man
01:58:09evanok!
01:58:16evancan I take a book and my ipod?
01:58:44rueNo, you will listen to this Zune
01:58:57evanNOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOO
01:59:06evanruns!
01:59:11evanYOU CAN'T MAKE ME
01:59:19rueHeh
01:59:57rueWell, I do not think I agree on not using keyword arguments (I am still not sure what the problem is), we could just make it a defaulting last parameter
02:00:45rueI think the #attach_foreign format was generally received OK since it is easy to see what is what
02:02:13evanno one used it
02:02:19evanso i'm not sure how it was received OK.
02:03:58rueIt was decreed that the original was to be used
02:04:12evanwe can revise it
02:04:14evani'm fine with that.
02:04:29rueWe could have a poll, I suppose, but either way we can add the library to the call itself
02:04:31evanwell
02:04:39evani'm headed out for dinner, then a show.
02:04:48rueOkaydoke
02:05:02evanattach_foriegn, if I recall, would have required that every line have the name of the library
02:05:09evanvery unDRY
02:05:37evanwe can talk
02:05:42evani think i was being fussy that day
02:05:49evani'm open to revisement.
02:06:12rueNo, I used keyword arguments :)
02:06:27evanstill means you'd have to pass it on every line
02:06:37evanattach_foreign 'blah', :library => "libiconv"
02:06:40evanattach_foreign 'blah2', :library => "libiconv"
02:07:33rueAh, if everything comes from a separate lib? Yeah.. there is a use-case for an set_ffi_lib-like method certainly
02:07:51evanok.
02:07:55evanwell, ponder that.
02:07:59evani'm off.
02:07:59rueNeeds to be thread-safe though
02:08:02rue*wave
02:08:08wmoxam enters the room.
02:14:56naeu leaves the room.
02:15:33ruivaldo leaves the room.
02:23:27ruivaldo enters the room.
02:24:22jicksta leaves the room.
02:24:36jicksta enters the room.
02:26:44mernen leaves the room.
02:31:05wmoxam leaves the room.
02:37:34VVSiz_ enters the room.
02:38:54yipstar leaves the room.
02:39:07blakewatters enters the room.
02:51:25pastie enters the room.
02:55:33bitsweat enters the room.
02:55:51VVSiz leaves the room.
03:02:37ezmobius_ leaves the room.
03:07:14ShayArnett leaves the room.
03:08:39ruivaldo leaves the room.
03:12:39twbray enters the room.
03:22:42boyscout3 commits by Eero Saynatkari
03:22:43boyscout * Iconv uses FFI::USE_THIS_PROCESS_AS_LIBRARY instead of nil.; 325b236
03:22:44boyscout * Module#set_ffi_lib allows multiple library names, documented same.; 812cba1
03:22:45boyscout * FFI::USE_THIS_PROCESS_AS_LIBRARY is the nil FFI library argument.; 8b1a22c
03:31:58jimbarnett enters the room.
03:34:38stcredzero enters the room.
03:36:43rubuildius_ppcEero Saynatkari: 325b23628; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
03:51:44twbray leaves the room.
03:52:35Maledictus leaves the room.
03:54:00twbray enters the room.
03:55:30fork_ enters the room.
03:58:26webmat enters the room.
03:59:43twbray leaves the room.
04:01:02rubuildius_ppc leaves the room.
04:01:42rubuildius_ppc enters the room.
04:01:43Cosmos95 leaves the room.
04:01:46stcredzerocompress.c:2:18: error: zlib.h: No such file or directory
04:02:07rubuildius_ppc leaves the room.
04:02:19AndrewO leaves the room.
04:02:29rubuildius_ppc enters the room.
04:04:31wycats leaves the room.
04:05:40dschn leaves the room.
04:07:03dysinger leaves the room.
04:07:30stcredzero leaves the room.
04:17:48stcredzero enters the room.
04:18:32michaellatta_ enters the room.
04:19:22stcredzeroI'm getting a bunch of undeclareds in readline.c. Is there a gem I need to install?
04:21:17shame enters the room.
04:24:39wycats enters the room.
04:25:48mernen enters the room.
04:29:04stcredzeroreadline.c:8:31: error: readline/readline.h: No such file or directory
04:29:24stcredzeroI've installed libreadline-ruby. What else?
04:30:21bitsweathow about libreadline?
04:30:51stcredzerono such package
04:31:05stcredzerodoesn't seem to be a readline gem either
04:32:01drbrainreadline is built-in to ruby
04:32:28benburkert_ enters the room.
04:32:36stcredzeroI installed ruby from Debian. Was that a bad idea?
04:33:35drbrainif you installed debian ruby and debian ruby readline, I would think it wold work
04:33:36jimbarnettThe docs for ubuntu say you need libreadline5-dev
04:33:39drbrain+u
04:33:55stcredzeroI just found that too!
04:34:05sambo82 enters the room.
04:34:33michaellatta leaves the room.
04:35:58stcredzeroThat's it. Apparently, we just finished a build
04:36:09stcredzeroThanks
04:36:39jimbarnettcool. Congrats
04:36:58stcredzeroAny parser gurus here?
04:37:59stcredzeroI want to know if it's futile for me to write a program get a complete AST from Ruby Source.
04:38:19drbrainit is
04:38:22drbrainit's been done
04:38:26drbrainParseTree
04:39:26stcredzeroThen not futile. So that's not limited to methods anymore? I need a full parse tree for entire source files, including random toplevel stuff.
04:39:55drbrainyou can get the AST for a method or a string
04:39:59drbrainmethod/block
04:40:14drbrainparse_tree_for_string File.read(your_file)
04:40:17drbrainI think
04:40:43stcredzeroThanks again
04:44:58stcredzeroStill, I'd rather get the AST from the parser in Rubinius. I want to see the AST it makes of itself
04:44:59nicksieger leaves the room.
04:46:18rubuildius_amd64Eero Saynatkari: 325b23628; 2189 files, 7319 examples, 26077 expectations, 1 failure, 2 errors; http://rafb.net/p/aV571B43.html
04:46:19drbrainwe're working on that
04:46:29drbrainyou can get the bytecode, though
04:47:56stcredzeroHmm, so there's still some oddities from the 1.8 MRI you're working through as far as having full access to the AST?
04:48:17benburkert leaves the room.
04:48:33drbrainactually, it's probably exposed in there somewhere
04:49:44stcredzeroI can't just write my own visitor on the Rubinius nodes?
04:50:05drbrainnot sure
04:51:26stcredzeroThen we'll be trying it and finding out.
04:54:14drbrainlib/compiler is where you want to start
04:56:46stcredzeroI been perusing that. Ciao
04:56:53stcredzero leaves the room.
05:16:20dlee leaves the room.
05:19:08yipstar enters the room.
05:23:44benny leaves the room.
05:50:26webmat leaves the room.
05:52:26RyanTM leaves the room.
05:54:02anonuseranyone here?
05:54:19willl enters the room.
05:56:00anonuserwell when someone gets back, is rb_funcall broken in subtend?
05:56:46edwardam leaves the room.
06:04:37Yurik enters the room.
06:04:58twbray enters the room.
06:05:56yipstar leaves the room.
06:33:14rueanonuser: Could well be, what are you running into?
06:36:02headius_ enters the room.
06:39:47twbray leaves the room.
06:51:58headius leaves the room.
07:01:42anteaya_ enters the room.
07:11:22yugui leaves the room.
07:11:56gnufied enters the room.
07:13:07anonuserrue, I'm back
07:13:38rue?
07:13:44anonuserwell i'm tracking down a problem after I'm implemented the needed crap to get hpricot to compile it segfaults the VM
07:13:58anonuserI'm running through what I added to see if things are wrong
07:14:08anonuserrb_singleton_class might be the culprit
07:14:40anonuserhttp://pastie.org/210477
07:14:49anonuserignore everything but rb_singleton_class
07:15:01rubuildius_ppc leaves the room.
07:15:03anonuserI have a breakpoint on the val = ....
07:15:42rubuildius_ppc enters the room.
07:16:07anonuserand just to see if the singleton class was getting defined i decided to try this
07:17:31anonuser rbx_string_as_cstr(module_get_name(val))
07:17:48anteaya leaves the room.
07:17:58anonuserthat created another segfault :(
07:18:24anonuserbut that makes sense because it hasn't been created :(
07:22:27anonusertried again and to no avail
07:22:40anonuserbasically when I do module_get_name(val)
07:23:08anonuserI get back an obj but it seems to be trash
07:24:54anonuserhttp://code.whytheluckystiff.net/hpricot/browser/trunk/ext/hpricot_scan/hpricot_scan.rl#L264
07:25:02anonuserbasically that's where hpricot freaks out.
07:25:33anonuserI thought it might be rb_define_attr but that has specs, so what I'm passing to it is obviously wrong
07:26:52anonuserany ideas? Or did you just explode and disperse into the internet?
07:27:38ruePacking
07:28:17anonuseroh sorry to bother you then
07:28:36rueNah, 's fine. Just in and out
07:28:56anonuserah
07:29:10jimbarnett leaves the room.
07:30:37krisps enters the room.
07:31:59anonuserif ya want to comment let me know, I'll be in and out as well
07:32:01rueHave you stepped into object_metaclass?
07:35:00anonusersadly I've ignored it for the most part. I'd like to think it's doing to right then and I'm just doing something wrong. That and I have no idea how it works.
07:35:40anonuserbasically I'm throwing grenades into the dark and hoping they go off.
07:39:43evanrue: that new ffi constant is crazy long.
07:40:11evanit's for use with set_ffi_lib, yes?
07:47:54anonuserevan, I've got a question about ID's in the subtend layer, http://github.com/evanphx/rubinius/tree/master/shotgun/lib/subtend/ruby.h#L36 is that correct just casting to a VALUE?
08:00:58michaellatta_ leaves the room.
08:01:31michaellatta enters the room.
08:23:08anteaya_ leaves the room.
08:24:26krisps leaves the room.
08:25:34JonathanT enters the room.
08:27:13blakewatters leaves the room.
08:28:36blakewatters enters the room.
08:43:20random8r enters the room.
08:49:46TheVoice leaves the room.
08:51:43antares_ enters the room.
08:58:42random8r leaves the room.
09:03:15xmlhacker enters the room.
09:03:25NoKarma enters the room.
09:05:20rueevan: I figure it saves a separate comment ;)
09:06:10rueWe do need to resolve the whole thing once and for all though
09:19:04rueanonuser: Looking into it now. Seeing if it can be reproduced in a smaller case
09:23:52Maledictus enters the room.
09:40:45VVSizdrbrain: thanks for the logs for "design meeting"! :)
09:42:27anonuserrue, if it helps module_get_name returns a NilClass
09:45:17rueVVSiz: There was a design meeting?
09:45:28VVSizyeah, two days ago
09:46:10rueVVSiz: Meh, missed it! Do you have the link?
09:46:41VVSizsure, one sec
09:47:16VVSizrue: http://ruby-design.pbwiki.com/Design20080604
09:49:19rueTHankees
09:51:53bitsweat leaves the room.
10:09:35lopex enters the room.
10:15:23thehcdreamer enters the room.
10:40:34antares_ leaves the room.
10:59:36qwert666 enters the room.
11:04:32rueanonuser: The metaclass extraction is working fine so far as I can figure out
11:06:21dbussinkrue: ah, i see you cleaned up my ugly hack a bit?
11:26:21loincloth leaves the room.
11:30:18ruedbussink: Just gave it a name. Need to do something proper
11:30:44dbussinkrue: yeah, that's true
11:32:11rueI thought evan pushed his code already but looks like not
11:33:48imajes enters the room.
11:49:32rueI wonder if there would be a better way to do platform resolution for an extension
11:57:05lopex leaves the room.
12:19:11Fullmoon enters the room.
12:24:46yugui enters the room.
12:54:35loop enters the room.
12:56:09Cosmos95 enters the room.
13:04:53jimbarnett enters the room.
13:06:41jero5 enters the room.
13:06:46kAworu leaves the room.
13:08:39kAworu enters the room.
13:11:19anteaya enters the room.
13:20:42sr leaves the room.
13:37:16VVSiz leaves the room.
13:37:28yugui leaves the room.
13:38:25VVSiz enters the room.
13:42:42yugui enters the room.
13:49:01sr enters the room.
13:50:33michaellatta_ enters the room.
13:50:39anteaya leaves the room.
13:58:10michaellatta leaves the room.
14:01:42michalw enters the room.
14:19:14RyanTM enters the room.
14:26:59blakewatters leaves the room.
14:35:08headius enters the room.
14:36:57jero5 leaves the room.
14:42:11headius_ leaves the room.
14:59:50Arjen_ enters the room.
15:01:42JonathanT leaves the room.
15:17:05moofbong enters the room.
15:24:58imajes leaves the room.
15:32:50twbray enters the room.
15:34:14headius_ enters the room.
15:35:05fbuilesv enters the room.
15:36:22enebo enters the room.
15:37:48moofbong leaves the room.
15:40:55headius leaves the room.
15:47:37imajes enters the room.
15:52:10twbray leaves the room.
15:58:43mjwhitt leaves the room.
16:06:04massgood morning
16:10:04shoemass: g'morning
16:11:39rueAhoy, strangers
16:12:04massI'm not strange, am I?
16:12:10rueYou are
16:12:31massnooooo
16:12:47rueFor all I know it could have been an evil twin last time
16:13:21massmy evil twin is moss
16:16:09DefilerOh hellll yeah
16:16:16Defilerlooks like I just got caller working
16:16:28dbussinkgoooo Defiler!
16:17:32DefilerYou thought 'caller' was supposed to list what called you. hahaha
16:17:33Defilerno
16:18:00DefilerYou can ask for the caller of things that have never been called. Rad.
16:18:00cored enters the room.
16:19:16atmos leaves the room.
16:21:59dbussinkDefiler: otherwise it would have been too easy of course ;)
16:22:17atmos enters the room.
16:22:59moofbong enters the room.
16:23:09boyscout1 commit by Wilson Bilkovich
16:23:10boyscout * Support correct 'caller' output for: eval("caller", some_proc.binding); 2467d2b
16:23:35DefilerKinda hacktastic at the moment, but I need to write more specs before I can merge the 'caller' and 'backtrace' code paths
16:24:28Defilerheadius_: Got it without duping any stack frames muhaha
16:30:24moofbong leaves the room.
16:35:04kAworu leaves the room.
16:35:15kAworu enters the room.
16:36:18anteaya enters the room.
16:37:38rubuildius_ppcWilson Bilkovich: 2467d2b10; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
16:38:49chris2 enters the room.
16:40:19rubuildius_amd64Wilson Bilkovich: 2467d2b10; 2189 files, 7319 examples, 26077 expectations, 2 failures, 2 errors; http://rafb.net/p/16ybOw87.html
16:41:26Defilerugh, we really need to sync up the CI specs
16:42:36mernen leaves the room.
16:42:44mernen enters the room.
16:46:13anonuserrue, thanks.
16:59:02DefilerHe and the engineers at Gemstone have gotten the full Ruby language running on the Gemstone VM in just three months.
16:59:10DefilerUgh.. having difficulty agreeing with that choice of words, chad =(
17:03:59DefilerMaglev is good for the Ruby ecosystem just as MRI/YARV, Macruby, JRuby, and IronRuby are good for the Ruby ecosystem.
17:04:09DefilerNotice anybody missing from that list? Ooh I should stop reading this article
17:09:56dbussinkDefiler: which article?
17:10:59anonuserDefiler, I just wish we knew what the target audience of Gemstone was
17:12:17rueThe paying audience :)
17:13:14Defilerdbussink: http://www.chadfowler.com/2008/6/5/maglev
17:14:18dbussinkDefiler: he did include rubinius is the paragraph before it though
17:14:52ch0wda enters the room.
17:17:12yugui leaves the room.
17:18:47twbray enters the room.
17:21:20twbray leaves the room.
17:29:52jimbarnett leaves the room.
17:30:22JonathanT enters the room.
17:32:33twbray enters the room.
17:40:51jimbarnett enters the room.
17:44:19twbray leaves the room.
17:48:37enebo leaves the room.
17:52:00lstoll leaves the room.
18:01:35ShayArnett enters the room.
18:10:01ch0wda leaves the room.
18:17:06cezarsa enters the room.
18:21:37thehcdreamer leaves the room.
18:25:25trythil enters the room.
18:30:08atmos leaves the room.
18:30:34atmos enters the room.
18:34:01lopex enters the room.
18:34:46fbuilesvis there any reason for Symbol not implement to_i yet?
18:35:34thehcdreamer enters the room.
18:39:10jp_tix:1 ?
18:39:38jp_tixwhat's the use case?
18:40:37jp_tixuh, :'1'?
18:40:55fbuilesvjp_tix: leaving aside compatibility, do stuff like trap(:INT)
18:41:44dbussinkwycats: can i bother you with a short dm question?
18:41:49jp_tixhmm, i don't even understand :'1'.to_i on MRI
18:42:07jp_tixseems to return the first five digits from the object id
18:42:12jp_tixor something
18:42:19dbussinkfbuilesv: there is, because it's pretty arbitraty
18:42:30dbussinkfbuilesv: we've had some discussion about it here
18:42:40mjwhitt enters the room.
18:42:59fbuilesvdbussink: let me check the logs, could help to fix this: http://rubinius.lighthouseapp.com/projects/5089/tickets/601
18:44:09fbuilesvthe socket error won't happen on os x but it'll still complain about not finding Symbol#to_i
18:44:34brixenwe don't support using fixnums as symbols
18:44:38brixenno will we ever
18:46:36fbuilesvbrixen: the issue's not using the number as the symbol but the conversion to integer, for support at least.
18:47:03brixenbut what does it do with the int after it converts it
18:47:16brixenI'm guessing it doesn't just do it for fun ;)
18:48:14stepheneb enters the room.
18:48:36DefilerIt didn't make any sense to implement one side of a two-way API
18:48:42cezarsafrom mri rdoc: "sym.to_i => fixnum: Returns an integer that is unique for each symbol within a particular execution of a program."
18:48:43chopmo enters the room.
18:48:44DefilerLike having send but not receive
18:48:57cezarsareturning __id__ wouldn't be enough?
18:49:18DefilerI guess it would, yeah
18:49:24DefilerMight as well define it that way
18:49:34shoeI once had a use for Symbol#to_i... It gave me consistent sort-order for symbols, and even though it wasn't platform-independent, it suited my purpose.
18:50:28shoein MRI, a symbol will always #to_i the same Fixnum.
18:50:30chopmoHi all. I'm having trouble running a clean build of rubinius on a MacBook Pro: http://pastie.org/210801
18:50:41chopmoDoes that ring a bell with anyone? TIA.
18:53:27trythil leaves the room.
18:53:47octopod enters the room.
18:56:05brixenchopmo: unfortunately, pastie is having a moment
18:56:19brixenchopmo: as soon as it loads, I might have some feedback :)
18:56:36chopmoWorks for me right now :-)
18:56:45chopmoThis may be easier to read: http://pastie.org/210801.txt
18:57:08chris2 leaves the room.
18:57:08brixenchopmo: you could paste here: http://rafb.net/paste/
18:57:36brixenchopmo: you must be in the special sector of the universe that pastie likes at the moment
18:57:58chopmoYeah - place called Denmark :)
18:58:02chopmohttp://rafb.net/p/IYmWnS43.html
18:58:37brixenhmm, and how did you get to this unfortunate exception? what command?
18:59:14chopmoWait, I'm sorry, it's working now. I thought I'd done a *clean* rebuild, but apparently not.
18:59:23brixenahh cool
18:59:38chopmoI think I missed "clean:rbc"
18:59:49brixenrake clean build works well for me
18:59:57chopmocool, thanks
19:00:14qwert666_ enters the room.
19:00:16brixenyou typically won't need that, just rake (rake build) is fine
19:00:37brixenif you get really messed up, rake distclean build :)
19:01:36fbuilesvactually I think the issue's inside trap, not with the to_i
19:02:06fbuilesvMRI's trap can receive a symbol and return the string from it and trap that, so :INT => "INT" and that'd work
19:02:20chopmoOK, I see. By the way, the "common build problems" page mentions a "pristine" rake task, but that seems to be outdated.
19:02:34brixenchopmo: ok, I'll update that
19:08:25enebo enters the room.
19:12:50wycatsdbussink: yessir
19:13:38ruebrixen: There?
19:14:19stepheneb_ enters the room.
19:15:14benny enters the room.
19:16:19brixenrue: yah
19:16:56m7d enters the room.
19:17:13fbuilesvwould you mind if I push http://rafb.net/p/1zqO2O32.html to forget about that to_i stuff then? The issue wasn't there
19:17:16ruebrixen: http://pastie.org/210814
19:17:57brixenfbuilesv: seems sane to me
19:17:58m7d leaves the room.
19:18:15brixenrue: pastie no workie for me
19:19:47rueTiming out here too
19:20:26qwert666 leaves the room.
19:21:28rueBetter
19:21:52ruebrixen: Reattempt
19:22:02brixenEREATTEMPTFAILED :)
19:22:48anonuserpastie, seems to be working now
19:23:02brixeny'all just like rubbing it in huh?
19:23:04brixen:P
19:23:19brixenthere's a pastie blackout in the pacific NW
19:23:49rueCould be
19:23:49brixenpastie should host with engine yard
19:23:55anonuserAh, I'm south east so it's all good here.
19:24:34evanmorning
19:24:50brixenmorning evan
19:24:57evanif anyone is curious
19:24:58evanhttp://www.ccs.neu.edu/scheme/pubs/stackhack4.html
19:25:15evanthat paper details a very similar technique i'm using to convert our bytecode into LLVM
19:25:24evanthey break the chunks into seperate funcitons
19:25:36boyscout1 commit by Federico Builes
19:25:37boyscout * Converts symbols to strings inside Kernel#trap to process them as normal; 2f32660
19:25:44evanwhere as I have them all in one function, and use a dispatch table at the top to move between chunks
19:26:00fbuilesvboyscout: normal strings :P
19:26:15evan2nd paragraph of step 4 indicates they could have done the same
19:26:21brixensweet
19:26:36brixenfbuilesv: commit summary line < 67 chars :P
19:26:58brixenor 72
19:27:04fbuilesvbrixen: I know, usually failt at it :P
19:27:10brixenheh
19:27:18evanalso
19:27:22evanid2name is deprecated
19:27:24evanuse #to_s
19:27:27fbuilesvouch
19:27:30fbuilesvlet me fix that
19:28:39krisps enters the room.
19:29:12boyscout1 commit by Federico Builes
19:29:13boyscout * Replace id2name (deprecated) with to_s in signal.rb; 0ac8d99
19:30:17rue< 67 now? Even tilman would not demand it to be less than 73 characters :)
19:30:24evan:)
19:30:27Defiler"Added spec for 0**-1 to return Infinity."
19:30:31DefilerRuby is wild times
19:31:29brixenrue: using git says 72, I thought it was less
19:32:13fbuilesvDefiler: shouldn't 0**-1 => 1/ 0 **1 => ZeroDivisionError?
19:32:55brixenDefiler: ruby is crazy, but apparently parsing webrick is equivalent to parsing all ruby syntax ;)
19:33:11brixenDefiler: oh the ignorance under which we labor
19:33:30octopod leaves the room.
19:33:30mjwhitt leaves the room.
19:33:30anteaya leaves the room.
19:33:30michaellatta_ leaves the room.
19:33:30sr leaves the room.
19:33:30Cosmos95 leaves the room.
19:33:30Yurik leaves the room.
19:33:30fork_ leaves the room.
19:33:30lypanov leaves the room.
19:33:30zenspider leaves the room.
19:33:35brixenI should have just copy-pasted webrick into spec/ruby/1.8/language
19:36:00sr enters the room.
19:36:06lypanov enters the room.
19:36:15evanbrixen: now now, no need to get fussy :)
19:36:21brixenhehe
19:37:09rueWhoa
19:37:14brixennerf-bats duels with rubyspec emblazon on them
19:37:33rueThey just applied a patch to Ruby from Moonwolf :D
19:37:48stepheneb leaves the room.
19:38:02brixengood lord, checkout #ruby-core
19:38:43brixenthey just rolled out about a dozen new patchlevels across versions
19:38:57ruebrixen: Try again, I re-routed some of the traffic to the Information County Highways
19:39:26ruebrixen: Oh, the same ones on #jruby?
19:39:56brixenrue: finally loaded
19:40:02brixenrue: one sec..
19:40:42michaellatta enters the room.
19:41:49zenspider enters the room.
19:42:00brixenrue: http://rafb.net/p/cLJDuE17.html
19:43:25zenspider leaves the room.
19:43:27octopod enters the room.
19:43:27mjwhitt enters the room.
19:43:27anteaya enters the room.
19:43:27michaellatta_ enters the room.
19:43:27Cosmos95 enters the room.
19:43:27fork_ enters the room.
19:43:27zenspider enters the room.
19:43:39zenspider leaves the room.
19:43:58zenspider enters the room.
19:44:50rubuildius_amd64Federico Builes: 0ac8d999a; 2189 files, 7319 examples, 26077 expectations, 2 failures, 2 errors; http://rafb.net/p/dDptGK97.html
19:45:19brixenrue: so, what you wrote makes sense
19:45:36brixenrue: unfortunately, 1.8.7 ensures that changes will occur in patchlevels
19:45:56brixenI like the idea of making RUBY_VERSION contain the patchlevel explicitly, but I'd leave out the 'p'
19:46:01evanlike the 110 patch level craziness
19:46:04brixen1.8.6.114 is sane enough
19:46:08evanor whatever one completely busted rexml
19:46:12brixenevan: yeah
19:46:16dschn enters the room.
19:46:21brixenthe ruby version is just a mess now
19:46:34evanright
19:46:42evanno one can depend on what will be in ANY version change now.
19:46:46brixenrue: and if you want an array, I think split('.') is enough
19:46:58rubuildius_ppcFederico Builes: 0ac8d999a; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
19:46:59rubuildius_ppcFederico Builes: 2f3266024; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
19:47:23brixenrue: what I really dislike is having to pull in rbconfig to accurately get the OS
19:47:35brixenthat should go into RUBY_PLATFORM somehow
19:48:54evani'm going to go get some food
19:48:56evanbbiab.
19:49:14ruebrixen: `RUBY_VERSION.split('.').map {|s| s.to_i }`, using the Array gives `RUBY_VERSION.join '.'`
19:49:37brixenyeah
19:49:51brixenwell, if you put it like that, I suppose another const would be in order :P
19:49:54fork_ leaves the room.
19:49:54Cosmos95 leaves the room.
19:49:54michaellatta_ leaves the room.
19:49:54mjwhitt leaves the room.
19:49:54anteaya leaves the room.
19:49:54octopod leaves the room.
19:50:24fbuilesvmmm, why's rubuildus amd64 runnign that old hash spec?
19:50:44brixenfbuilesv: haven't synced the CI specs yet
19:50:54brixenthere's a couple big problems I have to address first
19:51:03Yurik enters the room.
19:51:06fbuilesvbrixen: cool, but then why's ppc running fine?
19:51:14fbuilesvhand sync?
19:51:23brixenno, just doesn't fail on that platform
19:51:56jp_tix leaves the room.
19:51:59fbuilesvI see, didn't think that sort would be dependo n the platform
19:52:00wmorgan leaves the room.
19:52:15ruebrixen: Plus you can use <=> on the Arrays
19:52:31wmorgan enters the room.
19:52:32brixenrue: true, what would that const name be?
19:53:58rueI would actually just replace it and make a RUBY_VERSION_STRING but could give it some name too I suppose
19:54:44mjwhitt enters the room.
19:54:44anteaya enters the room.
19:54:44Cosmos95 enters the room.
19:54:44fork_ enters the room.
19:54:51rueRUBY_VERSION_{NUMBER,DATA,INFO}, something like that
19:54:55brixenwell, I wouldn't replace it, since way too much code will expect RUBY_VERSION to be a string
19:54:55anteaya leaves the room.
19:55:44anteaya enters the room.
19:55:45brixencould make it a special const, so RUBY_VERSION => '1.8.6.111'; RUBY_VERSION.to_a => [1,8,6,111]
19:55:51brixenbut that probably won't fly
19:56:08rueNah, might as well do a separate one
19:56:34brixenclass RubyVersion; def to_s ..; def to_a ..
19:57:21brixenRUBY_VERSION.{major|minor|teeny|tiny|patch}
19:57:23rueI suppose it would be easier to keep updated
19:57:44brixenjust needs to behave like a String in most circumstances
19:58:09benburkert_ leaves the room.
19:58:09jp_tix enters the room.
19:58:13rueWell, I shall give both options. I have no idea which they would be more receptive to
19:58:17brixenI'm strongly opposed to the idea of making RUBY_VERSION a hash though
19:58:28rueYeah, that is pointless
19:58:34rueIt has an ordering
19:58:47brixenyeah
19:59:05ruebrixen: Re: RUBY_PLATFORM, I think we should specifically _not_ constrain it too much
19:59:20brixenyeah, I see that perspective
19:59:38brixenat the same time, if we could convince on the RUBY_VERSION is an object idea
19:59:45brixenwe could do the same for RUBY_PLATORM
19:59:49brixener PLATFORM
19:59:57brixenand it could be much more useful
20:00:01rueMy approach would be to do the minimum possible needed to avoid ambiguity
20:00:37brixenR_P.{os|vendor|os_version|library|...}
20:00:45brixenand you could provide what you wish
20:00:54brixenbut, we should definitely have some basic fields
20:01:09brixeneverything runs on an OS
20:01:18rueRegexps give that power anyway
20:01:36brixenyou could regexp on the string rep, sure
20:01:55brixenbut then you'd have these standard regexp out there, why not put those into the object as fields?
20:02:18brixenagain, the rubyspec guards illustrate the issues involved here
20:03:34rueI dunno, I just think we should require as little as possible. The user can then do `do_some_complex_jvm_version_detection if R_P =~ /java/`
20:03:56Fullmoon leaves the room.
20:03:56brixenthe RUBY_PLATFORM object would just return '' if you ask for something it doesn't provide
20:04:07elight enters the room.
20:04:23brixenbut I think that there is a good case for some standard detection
20:04:33brixenif there weren't, the rubyspec guards would be pointless
20:04:40RyanTM leaves the room.
20:04:49brixenthey exist precisely because those distinctions need to be name for ruby behavior
20:04:56rueMy only reason to avoid that is the following question:
20:05:08brixens/name/made/
20:05:14rueWhat is the method name that returns "java" from the platform
20:05:31brixenTBD, but the idea is concrete
20:05:37RyanTM enters the room.
20:05:38brixenjvm, clr, gemstone/s, etc
20:05:53brixenthe object is there, what to refer to it as is the question
20:06:03brixen'os' is clear
20:06:08brixen'vendor' is pretty clear
20:07:02rueEssentially, I think the Java folks generally view the platform as "java" with a little asterisk for the platform used whereas we would tend to view the platform as arch + OS with a little asterisk for the language used
20:07:05brixenlet's say the identifier is 'subsystem', then R_P.subsystem => 'java'
20:07:32brixenyeah, I'd agree with that
20:07:49brixenbut we don't have to assign relative importance, just name the thing so you can get to it
20:08:49jazen enters the room.
20:10:22rueYeah. It seems like a banal issue but unless you have a good idea, I think we would be better off sidestepping it for now
20:10:46brixenwell, I really like turning these into objects
20:10:58brixenand for R_P, it should return '' for anything it doesn't know
20:11:11brixenR_P.who_shot_jfk => 'cia'
20:11:14brixenoops
20:11:26brixenR_P.who_is_the_best_vm => 'rubinius'
20:11:38brixenack, this object isn't working like I envisioned :P
20:12:02brixenR_P.is_microsoft_dead => ''
20:12:05brixenahh, there we go
20:12:18rueSee? You get problems with objects ;)
20:12:23brixenheh
20:12:27boyscout2 commits by Cezar Sa Espinola
20:12:28boyscout * Adding LONG2NUM and NUM2LONG to subtend.; 50bb348
20:12:29boyscout * Adding RDATA and DATA_PTR to subtend, with specs.; 04b9851
20:12:51rueI suppose it could be made transparent enough at least. The C side might be an issue with MRI
20:13:01brixenyeah, true
20:13:04brixenbut not much
20:13:14brixenyou could code this in C right quick actually
20:13:29brixen#to_str, #method_missing
20:13:33fork_ leaves the room.
20:14:00brixen#to_a might not work well, but you don't need it for R_P
20:14:26brixengrumbles about Ruby written in C :P
20:16:32benburkert enters the room.
20:16:44dbussinkhmm, who's cezar here?
20:16:57rue--> cezarsa
20:17:05dbussinkdog!
20:17:06cezarsayep
20:17:07dbussinkdoh!
20:17:13rueBastard!
20:17:16brixenhah
20:17:30brixendbussink: did you mean dawg!
20:18:08anteaya leaves the room.
20:18:08Cosmos95 leaves the room.
20:18:08mjwhitt leaves the room.
20:18:53dbussinkcezarsa: do you mind if i complain about some specs for subtend you appear to have written?
20:19:11dbussinksome warnings i get on 64 bit
20:19:42cezarsayeah i saw then, i think the problem is the typedef of VALUE
20:19:51cezarsaon mri it's a unsigned long
20:20:04cezarsaon rubinius it's a void*
20:22:05cezarsaand I was too lazy to change it and correct all the other warnings that would appear on ruby.c :)
20:22:18dctanner enters the room.
20:22:28rueI am futzing around with VALUE currently too
20:22:34ruedbussink: What were you seeing?
20:23:47dbussinkhttp://pastie.org/210836
20:24:28anteaya enters the room.
20:24:28mjwhitt enters the room.
20:24:28Cosmos95 enters the room.
20:25:44rubuildius_ppcCezar Sa Espinola: 50bb34830; 2189 files, 7322 examples, 26114 expectations, 0 failures, 0 errors
20:28:49cezarsaMRI is confuse, rb_const_defined and rb_respond_to return a Qtrue/Qfalse inside a int, rubinius was formerly returning a VALUE then I changed both functions to return a int, just like MRI
20:29:56dc_ leaves the room.
20:29:56RyanTM leaves the room.
20:29:56krisps leaves the room.
20:29:56mernen leaves the room.
20:29:56imajes leaves the room.
20:29:56jacen_ leaves the room.
20:29:56dbussink leaves the room.
20:29:56Illocution leaves the room.
20:30:09dc_ enters the room.
20:30:09RyanTM enters the room.
20:30:09krisps enters the room.
20:30:09mernen enters the room.
20:30:09imajes enters the room.
20:30:09jacen_ enters the room.
20:30:09Illocution enters the room.
20:30:09dbussink enters the room.
20:30:14rubuildius_amd64Cezar Sa Espinola: 50bb34830; 2189 files, 7321 examples, 26079 expectations, 2 failures, 2 errors; http://rafb.net/p/xXjADI66.html
20:30:44cezarsashould I change then to return a VALUE again? that would get rid of the warnings but the functions signature would be different to MRI's
20:30:59cezarsas/then/them
20:31:49rueWe can just fix the spec extension. It should be wrapping the int into objects
20:34:30cezarsaactually the ints should not be wrapped, they are already VALUEs (Qtrue/Qfalse)
20:35:46twbray enters the room.
20:36:33seydar enters the room.
20:37:20jacen_ leaves the room.
20:39:03jimbarnett leaves the room.
20:42:07benny leaves the room.
20:42:07sholden leaves the room.
20:42:07cschneid leaves the room.
20:42:07joachimm leaves the room.
20:42:07rue leaves the room.
20:42:07shoe leaves the room.
20:42:07Vagabond leaves the room.
20:42:07Jzalae leaves the room.
20:42:07ixx leaves the room.
20:42:07Chrononaut leaves the room.
20:42:07dmpk2k leaves the room.
20:42:07TheProkrammer leaves the room.
20:42:36benny enters the room.
20:42:36sholden enters the room.
20:42:36cschneid enters the room.
20:42:36joachimm enters the room.
20:42:36rue enters the room.
20:42:36shoe enters the room.
20:42:36Vagabond enters the room.
20:42:36Jzalae enters the room.
20:42:36dmpk2k enters the room.
20:42:36ixx enters the room.
20:42:36TheProkrammer enters the room.
20:42:36Chrononaut enters the room.
20:45:30joachimm leaves the room.
20:45:30benny leaves the room.
20:45:30dmpk2k leaves the room.
20:45:30Vagabond leaves the room.
20:45:30Chrononaut leaves the room.
20:45:30Jzalae leaves the room.
20:45:30ixx leaves the room.
20:45:30cschneid leaves the room.
20:45:30TheProkrammer leaves the room.
20:45:30sholden leaves the room.
20:45:30shoe leaves the room.
20:45:30rue leaves the room.
20:48:17benny enters the room.
20:48:17sholden enters the room.
20:48:17cschneid enters the room.
20:48:17joachimm enters the room.
20:48:17rue enters the room.
20:48:17shoe enters the room.
20:48:17Vagabond enters the room.
20:48:17Jzalae enters the room.
20:48:17dmpk2k enters the room.
20:48:17ixx enters the room.
20:48:17TheProkrammer enters the room.
20:48:17Chrononaut enters the room.
20:51:20rueWell hooray for the internet
20:53:17joachimm leaves the room.
20:53:17benny leaves the room.
20:53:17dmpk2k leaves the room.
20:53:17Vagabond leaves the room.
20:53:17Chrononaut leaves the room.
20:53:17Jzalae leaves the room.
20:53:17ixx leaves the room.
20:53:17cschneid leaves the room.
20:53:17TheProkrammer leaves the room.
20:53:17sholden leaves the room.
20:53:17shoe leaves the room.
20:53:17rue leaves the room.
20:53:52benny enters the room.
20:53:52sholden enters the room.
20:53:52cschneid enters the room.
20:53:52joachimm enters the room.
20:53:52rue enters the room.
20:53:52shoe enters the room.
20:53:52Vagabond enters the room.
20:53:52Jzalae enters the room.
20:53:52dmpk2k enters the room.
20:53:52ixx enters the room.
20:53:52TheProkrammer enters the room.
20:53:52Chrononaut enters the room.
20:59:36joachimm leaves the room.
20:59:36benny leaves the room.
20:59:36dmpk2k leaves the room.
20:59:36Vagabond leaves the room.
20:59:36Chrononaut leaves the room.
20:59:36Jzalae leaves the room.
20:59:36ixx leaves the room.
20:59:36cschneid leaves the room.
20:59:36TheProkrammer leaves the room.
20:59:36sholden leaves the room.
20:59:36shoe leaves the room.
20:59:36rue leaves the room.
21:00:01benny enters the room.
21:00:01sholden enters the room.
21:00:01cschneid enters the room.
21:00:01joachimm enters the room.
21:00:01rue enters the room.
21:00:01shoe enters the room.
21:00:01Vagabond enters the room.
21:00:01Jzalae enters the room.
21:00:01dmpk2k enters the room.
21:00:01ixx enters the room.
21:00:01TheProkrammer enters the room.
21:00:01Chrononaut enters the room.
21:00:48webmat enters the room.
21:04:05antares_ enters the room.
21:07:07webmat leaves the room.
21:07:07joachimm leaves the room.
21:07:07benny leaves the room.
21:07:07dmpk2k leaves the room.
21:07:07Vagabond leaves the room.
21:07:07Chrononaut leaves the room.
21:07:07Jzalae leaves the room.
21:07:07ixx leaves the room.
21:07:07cschneid leaves the room.
21:07:07TheProkrammer leaves the room.
21:07:07sholden leaves the room.
21:07:07shoe leaves the room.
21:07:07rue leaves the room.
21:07:36webmat enters the room.
21:07:36benny enters the room.
21:07:36sholden enters the room.
21:07:36cschneid enters the room.
21:07:36joachimm enters the room.
21:07:36rue enters the room.
21:07:36shoe enters the room.
21:07:36Vagabond enters the room.
21:07:36Jzalae enters the room.
21:07:36dmpk2k enters the room.
21:07:36ixx enters the room.
21:07:36TheProkrammer enters the room.
21:07:36Chrononaut enters the room.
21:08:06cezarsadbussink: do you have an avaliable 64 environment where you could test this patch: http://rafb.net/p/x2dh4F62.html ?
21:08:23cezarsait does a "typedef unsigned long VALUE", and should get rid of the warnings
21:10:40benburkert leaves the room.
21:10:48dbussinkcezarsa: yes, that seems to fix it
21:10:57benburkert enters the room.
21:11:04dbussinkrb_require doesn't work yet, does it?
21:11:11cezarsagreat, i'll push it then
21:11:16cezarsanot yet AFAIK
21:11:41cezarsai think i'll take a look at it, see if I can figure out what's happening
21:11:53twbray leaves the room.
21:12:07benburkert leaves the room.
21:12:41benburkert enters the room.
21:14:51benburkert leaves the room.
21:14:56benburkert enters the room.
21:16:51boyscout1 commit by Cezar Sa Espinola
21:16:52boyscout * Change VALUE's typedef to unsigned long, plus some subtend fixes.; 92129e6
21:16:58ruecezarsa, dbussink: Keep it a pointer type, please
21:17:07cezarsaoops
21:17:22cezarsaI can revert it, but why rue?
21:17:30dbussinkwell, do we want to mimic mri behavior here or not?
21:18:28joachimm leaves the room.
21:18:28benny leaves the room.
21:18:28dmpk2k leaves the room.
21:18:28Vagabond leaves the room.
21:18:28Chrononaut leaves the room.
21:18:28Jzalae leaves the room.
21:18:28ixx leaves the room.
21:18:28webmat leaves the room.
21:18:28cschneid leaves the room.
21:18:28TheProkrammer leaves the room.
21:18:28sholden leaves the room.
21:18:28shoe leaves the room.
21:18:28rue leaves the room.
21:19:00webmat enters the room.
21:19:00benny enters the room.
21:19:00sholden enters the room.
21:19:00cschn