Index

Show enters and exits. Hide enters and exits.

00:01:04corp leaves the room.
00:04:16sambo enters the room.
00:05:04tarcierinice
00:08:59tarcieriyeah, that was another weird thing
00:09:04corp enters the room.
00:09:04tarcieriBEAM originally emitted C code
00:10:19evanyeah, i see that.
00:11:17sambo82 leaves the room.
00:15:08wycatstarcieri: heh
00:15:11dctanner enters the room.
00:15:46corp leaves the room.
00:16:55tarcieriErlang could really use a jar sort of thing for .beam files
00:18:13drbrainoh, I am curious
00:18:20drbrainwhy do we have Numeric#==?
00:18:46drbrain(and the rest of the Comparable methods)
00:19:43sambo82 enters the room.
00:20:04MenTaLguYhm, is the tmp rspec helper something rubinius-specific?
00:20:14shame leaves the room.
00:20:22drbrainit may be mspec-specific
00:20:45MenTaLguYso yes
00:21:58drbrainI don't know rspec, but I know it is in mspec
00:22:03drbrainbut not in rubinius
00:23:27MenTaLguYyikes, it can create files in USERPROFILE
00:24:34MenTaLguYI guess I need to write my own helper anyway :/
00:24:50MenTaLguYwhere is the official mspec tree kept, OOC?
00:25:55krisps leaves the room.
00:26:27djwhitthttp://github.com/brixen/mspec/tree/master
00:26:40djwhittpretty sure that's the official one
00:26:52drbrainit is
00:27:01drbraincorundum: mspec?
00:27:01corundumbeats me
00:27:04evandrbrain: not sure why we have Numeric#==
00:27:11evanbut we should have Fixnum#== and Bignum#==
00:27:15evanbecause of how often they're used
00:27:15sambo leaves the room.
00:27:19evanrather than going via <=>
00:27:27drbrainyeah, and just use spaceship on Numeric
00:27:32evanyeah
00:27:35evannot sure why
00:28:35sambo enters the room.
00:37:44NoKarmaactually, we should only have <=> on Numeric and include comparable, which should give us <, <=, ==, => and >
00:39:09kw leaves the room.
00:40:07drbrainyup
00:40:08manverubut that's an operation that you can optimize... a lot?
00:40:21NoKarmawell
00:40:32drbrainmanveru: no
00:40:36NoKarma<=> should just check equality
00:40:49drbrainFloat#== and Fixnum#== and Bignum#== is what you do a lot of
00:40:55drbrainNumeric#==, almost never
00:41:05drbrainMRI is designed this way
00:41:24NoKarma leaves the room.
00:41:47manveruyeah...
00:41:58manverui guess it's just a translation of MRIs API
00:42:14drbrainwhich is why I was puzzled
00:42:31sambo82 leaves the room.
00:43:51drbrainif somebody else doesn't have spare cycles to fix it, I'll file myself a mental note
00:44:32MenTaLguY<=> is the general comparison operator, it can't (correctly) just check equality
00:45:57evandrbrain: file an LH note instead.
00:46:07drbrainooh! yes
00:46:12drbrainpart of my quota
00:46:13cezarsa enters the room.
00:47:02jeremydurham_ enters the room.
00:49:09dfg59speaking of equality, Array#== breaks for recursive arrays. stumbled across that working on File#join. Should I write a ticket?
00:50:48drbrainyes
00:51:06jeremydurham_ leaves the room.
00:51:18dfg59will do...looks like the behavior has actually changed recently. before it was hanging, now it's actually crashing shotgun/rubinius
00:51:31dctanner leaves the room.
00:52:19dfg59also interestingly enough, Array#eql? behaves as expected with recurssive arrays.
00:52:22dfg59*recursive
00:54:13cezarsa enters the room.
00:54:51olabini leaves the room.
00:55:02mernen enters the room.
00:56:03corp enters the room.
00:57:34sambo82 enters the room.
00:59:40samruby enters the room.
01:03:00samruby_ enters the room.
01:03:58enebo enters the room.
01:05:49samruby__ enters the room.
01:05:51drbrainI have found the source of my Dir.glob problem
01:05:57evanhuzzah!
01:08:29drbrainI'm not sure of the solution yet
01:12:47shame enters the room.
01:14:14cezarsahello guys
01:14:26cezarsai'm new on rubinius and need some help from u
01:14:52cezarsai'm trying to modify some of the .c files on /spec/subtend/ext/ but 'rake build' isn't recompiling the files, are there any extra steps?
01:18:45sambo leaves the room.
01:19:43evanhm.
01:19:53evanoh, spec ones
01:20:00evan'rake spec' should take care of that
01:20:39samruby leaves the room.
01:21:29sambo82 leaves the room.
01:22:17sambo enters the room.
01:24:30samruby_ leaves the room.
01:24:45olabini enters the room.
01:25:32cezarsaevan, thanks, that did the trick
01:25:45evanno problem
01:25:47cezarsai'm working on adding more functions to subtend (things like rb_str_cat2, rb_inspect,...), is this ok?
01:27:05enebo leaves the room.
01:30:27zenspidercezarsa: absolutely, just please get a code review (or three)
01:31:25anteaya_ enters the room.
01:32:51probablycorey enters the room.
01:34:18drbrainYAY!!!
01:34:25drbrainI have really discovered the problem
01:34:47drbrainour glob only allows for one pattern
01:35:01drbrainMRI glob allows for multiple patterns
01:35:27drbrainfiguring it out from their hybrid array/linked-list structure only took three readings of the C code
01:35:31evanwas it a mis translation of the C code?
01:35:36evanyeah
01:35:36evandude
01:35:44evanthey're hybrid array/list thing is NUTS.
01:36:31drbrainfor (cur = beg; cur < end; ++cur) { struct glob_pattern *p = *cur; /* ... */ p = p->next; /* ... */ }
01:36:36drbrainthat's tricky
01:36:44crafterm enters the room.
01:36:50evanvery
01:36:51evanfunny
01:36:56evani stared at that some code
01:36:59evangoing... "wtf."
01:37:05evan"eric will know."
01:37:06evan:)
01:37:42sambo82 enters the room.
01:38:42drbrainso, now I need to figure out how to make ours work right
01:39:24probablycorey leaves the room.
01:42:37samruby__ leaves the room.
01:45:36boyscout1 commit by William Morgan
01:45:37boyscout * initial alioth cross-language benchmark infrastructure; f5ec29e
01:45:46yugui enters the room.
01:46:05sambo leaves the room.
01:47:00jtoy enters the room.
01:49:30anteaya leaves the room.
01:49:43enebo enters the room.
01:53:03sambo enters the room.
01:53:04benburkert leaves the room.
01:54:44Cosmos95 leaves the room.
01:54:54luislavena enters the room.
01:55:00dfg59 leaves the room.
01:59:07sambo82 leaves the room.
01:59:14rubuildius_amd64William Morgan: f5ec29e7a; 2175 files, 7035 examples, 25542 expectations, 0 failures, 1 error
01:59:25evancrap!
01:59:50evanwhy isn't it spitting out a url?
02:00:06lopex leaves the room.
02:00:09rubuildius_ppcWilliam Morgan: f5ec29e7a; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
02:00:49drbrainhaha
02:01:12evansweet.
02:07:00djwhitthmm... thought I fixed those error pasties
02:07:05djwhitthang on. I'll get the error
02:07:42sambo82 enters the room.
02:09:07djwhittbleh, it's something messed up on my system
02:09:49dlee leaves the room.
02:09:50djwhittI'm getting stuff like this: http://rafb.net/p/yCoyKL65.html
02:11:34imajes_ leaves the room.
02:12:57trythil enters the room.
02:14:54lstoll enters the room.
02:16:03sambo leaves the room.
02:23:44luislavenadjwhitt: it seems you don't have write permission in /tmp ?
02:24:02sambo enters the room.
02:24:04luislavenaI've seems this on sudoless users trying to run specs that create temporary files.
02:24:13luislavena(or access them)...
02:29:02Fullmoon enters the room.
02:29:26VVSiz_ enters the room.
02:30:29yipstar leaves the room.
02:32:20mernenI'm about to commit specs for iconv, which rubinius doesn't support yet
02:32:25mernenare there any special steps I should take?
02:33:01tarcieriyou need to like, mark the specs as expected to fail, somehow
02:33:24drbrainread the URL in the topic
02:33:33tarcierihaha
02:34:23drbrainyou should not need to update tags
02:34:42drbrainas that'll happen on the next update of the rubyspec module
02:35:04yipstar enters the room.
02:35:16sambo82 leaves the room.
02:37:17mernenthanks, I'll have a look
02:39:32sambo82 enters the room.
02:45:40sambo leaves the room.
02:47:16VVSiz leaves the room.
02:47:32luislavenamernen: if you do specs for iconv, let me know, since we don't use GNU iconv in Windows, and MRI lacks test for it.
02:47:45luislavena(we don't use GNU iconv, we use a native windows iconv).
02:48:39djwhittluislavena: I think maybe I had a mock directory generated by a different user hanging out in tmp
02:48:47mernenI'm almost done
02:49:29mernenI tried to rely on a minimum of encodings: us-ascii, utf-8, utf-16 (be, le), iso-8859-1
02:50:03mernenno japanese encodings for now... I'm not sure what I can assume to be supported on other implementations
02:50:24AndrewO enters the room.
02:51:26luislavenadjwhitt: that could be too :-)
02:51:56Fullmoon leaves the room.
02:51:56luislavenadjwhitt: just checking, since I think I played with the chdir specs a while back and used the #tmp() helper.
02:52:28enebo leaves the room.
02:54:24ezmobius leaves the room.
02:54:57sambo enters the room.
02:56:04samruby enters the room.
03:02:40sambo82 leaves the room.
03:11:10enebo enters the room.
03:12:23sambo82 enters the room.
03:15:55djwhittbrixen: should mspec be automatically cleaning up the files it creates /tmp ?
03:16:15enebo leaves the room.
03:18:06sambo leaves the room.
03:23:15samruby leaves the room.
03:27:46sambo enters the room.
03:31:46enebo enters the room.
03:34:16enebo leaves the room.
03:35:24sambo82 leaves the room.
03:43:13sambo82 enters the room.
03:49:14nemerle_afk enters the room.
03:50:30sambo leaves the room.
03:50:35michaellatta_ enters the room.
03:50:55RyanTM leaves the room.
03:52:01sambo enters the room.
03:54:24sambo leaves the room.
03:54:28djwhittok, tested and the error pastie issue should really be fixed this time
03:54:51djwhittI also resolved the permissions issue on that machine
03:55:21djwhittI'm pretty sure that's what was causing some of the specs to not run as well
03:56:01michaellatta leaves the room.
03:57:10obvio171 enters the room.
03:58:31djwhittgoing to sleep now, but ping me if anything goes wrong with rubuildius_amd64 and I'll check it in the morning
03:59:26Ferdinand enters the room.
04:04:46sambo82 leaves the room.
04:06:54nemerle leaves the room.
04:07:00obvio enters the room.
04:07:45moofbong enters the room.
04:09:38michaellatta enters the room.
04:11:35robin_dewd leaves the room.
04:11:52michaellatta__ enters the room.
04:12:03michaellatta__ leaves the room.
04:15:41Ferdinand leaves the room.
04:17:22michaellatta_ leaves the room.
04:19:05lstoll leaves the room.
04:21:32Ferdinand enters the room.
04:25:26trythil_ enters the room.
04:26:17mernenluislavena: just tested my specs on windows
04:26:34mernenran into very few issues
04:26:55mernenmost notably, the platform guard for :linux is triggering on windows
04:27:21obvio171_ enters the room.
04:28:05michaellatta leaves the room.
04:33:20michaellatta enters the room.
04:36:36RyanTM enters the room.
04:37:28luislavenamernen: do you have a pastie with your specs?
04:37:43trythil leaves the room.
04:38:47mernenjust a sec
04:41:03corp leaves the room.
04:42:14obvio leaves the room.
04:44:51mernenluislavena: http://pastie.org/200007
04:45:35mernenthe only real problem I faced with the iconv library on windows was that it didn't react to //translit as I expected
04:46:13AndrewO leaves the room.
04:46:53mernenother than that, I discovered what I feared was true: there's no real rule for the byte order of utf-16
04:47:11mernenwhich makes it a somewhat tricky encoding to test (but still necessary)
04:48:16luislavenamernen: did you tested that against one-click installer?
04:48:21mernenyes
04:48:35mernenjust downloaded the latest one
04:49:04luislavenamernen: 186-26, correct?
04:49:11mernenyeah
04:49:40mernenI've guarded the //translit test for now, everything else runs fine
04:50:21mernenI just have to find a better way to test for utf-16 while ignoring endianness
04:54:20luislavenaI'll love if you can test that against the mingw build, since we are not using gnu iconv there.
04:54:23michaellatta_ enters the room.
04:54:41mernenoh
04:54:44Defileris it platform_is :linux these days?
04:54:52Defileror platform_is :os => :linux?
04:55:08brixenmernen: got a pastie for the platform guard triggering on windows?
04:55:08mernenDefiler: not sure, just grepped through the specs and found both ways
04:55:14mernenand also :os => [:linux]
04:55:26mernenbut both of them failed me on windows, IIRC
04:55:42brixenmernen: what file?
04:55:47mernenbrixen: platform_is :linux do puts "OMG" end :)
04:55:55luislavenabrixen: was was the discussion about platform_is :windows and :os => windows :-P
04:56:12brixenluislavena: both are supported
04:56:20brixenmernen: what version of windows?
04:56:23mernenluislavena: it's quite a hassle for me to open the windows vm right now
04:56:28mernenwindows XP, running on VirtualBox
04:56:43mernenand using the ruby one-click installer
04:56:45brixenmernen: ok, I've got xp up in parallels right now, I'll take a look at the guard
04:57:11luislavenamernen: ok no problem, was asking if you can also test mingw build, since we are using a native iconv implementation instead of gnu.
04:57:12brixenluislavena: how are you testing multiple installs of one-click?
04:57:22luislavenawas a true pain get gnu tools build on windows.
04:57:32brixenluislavena: I have the old one-click installed (pre mingw)
04:57:40atmos leaves the room.
04:57:41luislavenabrixen: yep, also against linux and rbx ;-)
04:57:43brixenluislavena: should I just replace it?
04:57:48mernenluislavena: I didn't know it was the mingw build, but I'll see if I can do it tomorrow
04:58:11luislavenabrixen: there are few bits missing, mostly some stdlibrary extensions.
04:58:24luislavenamernen: let me give you the link for more info.
04:59:09luislavenashameless self promotion: http://blog.mmediasys.com/
04:59:13mernennow, since there's plenty of people more experienced on the specs here, I'll ask for advice :)
04:59:44benburkert enters the room.
05:00:03luislavenaI'll also let you know there is a mailing list, is hard to me keep work focused with irc open ;-)
05:00:06mernenproblem is: there's no rule defining the endianness of the utf-16 encoding, as long as you include a BOM
05:00:16kw enters the room.
05:00:33mernenso, it makes things tricky to test
05:00:44michaellatta leaves the room.
05:01:25luislavenabrixen: back on mingw build, is safe to use it, as long you don't plan run Rails with it :-D
05:01:33brixenluislavena: gasp!
05:01:41luislavenaknows that rubinius can run rails, yada yada :-P
05:01:43brixenluislavena: how can you not support rails?!
05:01:43fowlduck enters the room.
05:01:50brixenluislavena: :D
05:01:52mernenI was originally using something like this to test: http://pastie.org/200013
05:02:04mernennot sure which is more hackish though, heh
05:02:09luislavenafights against RUBY_PLATFORM regexp, I'm tired of it :-P
05:02:50brixenluislavena: fighting against for what?
05:03:00luislavenabrixen: rails will work, what needs tweaks are sqlite, mysql and postgres packages :-)
05:03:11brixenluislavena: I only just learned saturday that jruby has 'java' in RUBY_PLATFORM on *everything* :P
05:03:23luislavenabrixen: we don't have alternative now, is a bummer, we need something like Rubinius::COMPILER or something like that ;-)
05:03:34luislavenabrixen: yeah, a platform to rule them all :-P
05:03:47brixenluislavena: oh, question for ya, what is Config::CONFIG['host_os'] under mingw build?
05:03:50mernenbrixen: any thoughts on how to test for correctness while ignoring endianness?
05:03:56mernenin a way, I'd need a specialized matcher
05:04:15brixenmernen: hmm, truth is, I hardly understand iconv at this point, let me read up a bit
05:04:23brixenmernen: a custom matcher is certainly possible
05:04:41mernenit's not really something related to iconv, but to utf-16 in general
05:04:48brixenmernen: does it need to process input read through IO, or just deals with String?
05:04:54mernenyou can specify the encoding as utf-16be or utf-16le
05:05:03mernenbrixen: just string, as far as I can tell
05:05:07brixenok
05:06:04brixenmernen: do you have a pastie of the specs that use Utf16Comparator?
05:06:06mernenwhen you pass just "utf-16", though, it's up to the implementation to decide the byte order, as long as the resulting string is prefixed witha byte-order mark
05:06:36mernenbrixen: http://pastie.org/200007 -- look for IconvSpecs.utf16
05:07:50fowlduck leaves the room.
05:08:25brixenmernen: I see, yeah a custom matcher would be quite useful there
05:08:38brixenshould equal_utf16 ?
05:09:00mernencould be, I guess
05:09:04roo enters the room.
05:09:47brixenmernen: also, zenspider did an extension to raise_error that takes a block to do stuff with the exception
05:10:13mernenoh, that'd be very useful
05:10:15brixenzenspider: where's a usage example of your raise_error extension
05:10:26brixenmernen: yeah, if only I could remember what specs :)
05:11:03brixenI'm thinking language specs. le'me look for it
05:11:05luislavenabrixen: http://pastie.org/private/uoity1mfp5ckzm6iovv6g
05:11:30brixenluislavena: thank you. I had assumed as much :)
05:13:30mernenby the way, could you tell me the output for this on osx? ruby -riconv -e 'p Iconv.conv("us-ascii//translit", "utf-16be", "\xff\xfd")'
05:13:49mediogre enters the room.
05:14:07obvio171_ leaves the room.
05:14:15RyanTM leaves the room.
05:14:17mernenthat's an attempt to transliterate the infamous � character into us-ascii... returns "?" on linux, fails on windows
05:14:52Ferdinand-e:1:in `conv': "\377\375" (Iconv::IllegalSequence)
05:14:52Ferdinand from -e:1
05:15:25mernendang, so far //translit seems to be a gnu-only option
05:15:26michaellatta enters the room.
05:15:58mernenI'll guard it appropriately then
05:17:31Ferdinand leaves the room.
05:17:49rue leaves the room.
05:19:29Defileroh cool. kent beck is a keynote speaker at RailsConf
05:20:49michaellatta_ leaves the room.
05:24:02mernenbrixen: btw, just found out mspec -v is failing here
05:24:34moofbong leaves the room.
05:25:13brixenmernen: ok, what does "failing" mean?
05:25:35mernenmspec/lib/mspec/utils/options.rb:306:in `add_version': uninitialized constant MSpecOptions::MSpec (NameError)
05:26:03Rich_Morin____ enters the room.
05:26:50brixenmernen: ok, got it
05:27:43Rich_Morinrbx is willing to parse some Ruby code and then let me inspect the result (eg, to find method definitions and calls). Can MRI be used in a similar manner? If so, how?
05:28:28DefilerYou can install the ParseTree gem on ruby 1.8.x
05:29:33nicksieger enters the room.
05:32:14Rich_MorinThat looks quite interesting, as does ruby_parser. Nothing in MRI itself, tho, I take it?
05:34:19tarcierino, and YARV removes the ability to implement ParseTree, afaik
05:35:12Rich_MorinSo I can use Parse Tree while I'm waiting for rbx to go golden, then switch over :)
05:35:48tarcieriyeah that's pretty much what I'm doing with Actors
05:35:59tarcieriYARV for now, rbx when it's ready
05:37:27trythil enters the room.
05:37:36Rich_MorinFWIW, my use case is analyzing method definitions and uses (etc) in Rails files (including erb) without actually running them.
05:38:51Rich_Morintnx for the info!
05:40:58smparkes leaves the room.
05:43:56mernenbrixen: sorry for constantly bothering, but what about equal_utf16? I can do that if you're busy
05:44:08trythil__ enters the room.
05:45:28brixenmernen: yeah, that seems like a reasonable thing to me. the matchers are really easy to add
05:45:34brixenmernen: please do if you feel like it
05:45:47brixenmernen: I've almost got the platform guard fixed
05:45:59luislavenabrixen: you're my man :-)
05:46:29brixenluislavena: it was the height of dumb. my brain on no coffe or green tea :)
05:46:45trythil___ enters the room.
05:47:00wmoxam_ enters the room.
05:49:45luislavenabrixen: I knwo that kind of stuff... I've mastered that old art...
05:50:01luislavenabrixen: also, I keep telling me that I shouldn't do any kind of code on mondays.
05:50:41luislavenatelling myself, argh...
05:51:32scooprluis :)
05:51:32brixenheh
05:51:44trythil_ leaves the room.
05:52:20obvio171 enters the room.
05:55:19mernenwell, gotta go... almost 2am here
05:55:23mernenthanks for all the help, guys
05:55:50brixenmernen: sure, thank you!
05:56:06mernen leaves the room.
05:56:06brixenmernen: I'll have -v and the guard fix pushed
05:56:20jazen enters the room.
05:56:23jacen_ leaves the room.
05:56:23dschn leaves the room.
05:56:33dschn enters the room.
05:57:12trythil leaves the room.
05:58:58luislavenascoopr: hey there, been a bit away, ping me if you need something...
05:59:29luislavenabrixen: can I propose something to the next design meeting? I'll be in an airplane at that time, but...
05:59:43luislavena I'm kind of sick of RUBY_PLATFORM, mostly because:
06:00:01luislavena1) java is java and not java-windows or java-linux or something like that.
06:00:10luislavena2) I dunno what will IronRuby report :-P
06:00:38brixenyeah, I feel the pain, believe me! :)
06:01:55brixenluislavena: I'll be your proxy :) http://ruby-design.pbwiki.com/Design20080604
06:02:28luislavenabrixen: I owe you a beer.
06:02:28trythil__ leaves the room.
06:02:36brixenlooks like somebody is forgetting to own their agenda items ;)
06:02:43luislavenabrixen: or a glass of good wine :-D
06:02:58brixenhehe
06:03:36wmoxam_ leaves the room.
06:03:39wmoxam leaves the room.
06:04:41gramos leaves the room.
06:06:32dfg59 enters the room.
06:06:44dfg59is rake spec broken for everyone else?
06:06:47gramos enters the room.
06:07:01rubuildius_ppc leaves the room.
06:07:16brixendfg59: I just ran it ok
06:07:22brixendfg59: what does "broken" mean?
06:07:31dfg59brixen: let me pastie the output
06:07:34rubuildius_ppc enters the room.
06:07:49boyscout1 commit by Brian Ford
06:07:50boyscout * Updated MSpec submodule to 3fc06169; 3b48d1b
06:07:59dfg59http://pastie.caboo.se/200033
06:08:21brixendfg59: rake clean build and try again
06:08:28dfg59k
06:08:41dfg59i did a rake build, should i always run clean?
06:08:59brixenyeah, some stuff changed recently that requires a full build
06:09:12brixenthose errors look suspect to me
06:09:30brixenbut in general, rake build is sufficient
06:09:45brixenbrb..
06:10:28michaellatta_ enters the room.
06:10:55dfg59brixen: that did the trick, thanks
06:11:09dfg59brixen: although, i am getting errors, interesting
06:12:59dfg59strange...getting a lot of errors about eql not being defined on Object
06:13:06dfg59should those be eql?
06:15:27luislavenadfg59: eql or be_eql? maybe there is a problem with MSpec be_equal matcher.
06:15:52dfg59luislavena: i'm not really sure what's going on here, i'm getting 40+ errors/failures
06:16:29michaellatta leaves the room.
06:19:06rubuildius_amd64Brian Ford: 3b48d1b7b; 2175 files, 7166 examples, 25730 expectations, 0 failures, 0 errors
06:21:54rubuildius_ppcBrian Ford: 3b48d1b7b; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
06:24:00dfg59 leaves the room.
06:24:27obvio171 leaves the room.
06:27:21drbrain leaves the room.
06:28:06drbrain enters the room.
06:28:12brixenluislavena: what do you use for av on windows?
06:28:26luislavenabrixen: NOD32
06:28:27brixenpaying for av on windows is like kissing bill gates ass
06:28:32brixenluislavena: link?
06:28:43luislavenabrixen: http://www.eset.com
06:29:08corp enters the room.
06:29:13brixenluislavena: thanks
06:29:30michaellatta enters the room.
06:29:40luislavenabrixen: since I start using it 3 years ago I dont deal with viruses :-)
06:29:58luislavenabrixen: except for those pen drives that sometimes jump into my usb ports :-P
06:30:54luislavenabrixen: I recommend you stay away from smart security and go only for the antivirus stuff.
06:31:46brixenluislavena: I was using kaspersky, but then they wanted money from me
06:31:49brixenfat chance :P
06:32:34luislavenabrixen: I think is the only software, besides windows I payed for.
06:32:35brixenI wonder if M$ gets kickbacks from AV vendors for keeping the market so rich :)
06:33:04luislavenabrixen: I think virus crators works for AV companies... can be a world conspiracy? ;-)
06:33:10brixenyeah, I payed for XP and OS X, but I certainly don't pay for AV on OS X
06:33:10luislavenas/cratos/creators
06:33:33luislavenabrixen: nod is kind of cheap for what it delivers.
06:33:47luislavenaI wouldn't spend a penny on Vista, fyi :-P
06:33:57brixenhhe
06:34:10rue leaves the room.
06:34:39luislavenagot it with my dell, after 3 hours bluetooth stop working :-P
06:35:36massis watching wycats speak
06:35:39michaellatta_ leaves the room.
06:37:03wycatsmass: what am I speaking about?
06:39:38atmos enters the room.
06:40:00massdatamapper
06:48:12massmwrc2008
06:50:17headius enters the room.
06:50:32michaellatta_ enters the room.
06:55:00tarcieriDataMapper does have the ability to iterate large collections without mapping in the whole result set at once, right?
06:55:25tarcierihalfway through converting something from DBI to AR... should've really looked at DataMapper instead
06:55:52michaellatta leaves the room.
06:59:01headius leaves the room.
07:06:14mass leaves the room.
07:06:22mass enters the room.
07:08:58headius enters the room.
07:12:07yipstar leaves the room.
07:14:54Maledictus enters the room.
07:15:02w1rele55 enters the room.
07:15:34michaellatta enters the room.
07:17:39twbray enters the room.
07:22:12michaellatta_ leaves the room.
07:26:22gramos leaves the room.
07:30:58kw leaves the room.
07:31:08chillywilly leaves the room.
07:31:10Skip enters the room.
07:31:13chillywilly enters the room.
07:34:37michaellatta_ enters the room.
07:35:02headius leaves the room.
07:36:08rue enters the room.
07:37:16lstoll enters the room.
07:40:53michaellatta leaves the room.
07:47:59michaellatta enters the room.
07:55:00headius enters the room.
07:56:34random8r enters the room.
07:57:53dbussinkmorning!
07:57:58random8rheyz
07:58:09evanmorning.
07:58:25random8ror afternoon ;-)
07:58:35tim_w enters the room.
07:58:48headius leaves the room.
07:58:59gnufied enters the room.
07:59:06brixenrandom8r: or midnight :)
07:59:19dbussinkevan: i refactored some more stuff in bignum / fixnum last night, now uses function overloading
07:59:23dbussinkmakes it look a lot cleaner
07:59:31evansweet.
08:01:00dbussinkprobably gonna add some float support in there too tonight
08:01:16random8reww.floats
08:01:23tim_w leaves the room.
08:01:28random8rhow good is flaming thunder's rational number support lol
08:01:35random8rbleh anyway. :)
08:02:56dbussinkeveryone expects floats to work ;)
08:03:20random8rthey do
08:03:30random8rit's just that when they work, they're not very mathematical, right?
08:04:13dbussinkyou just need to prepare for rounding errors
08:04:18random8ryup. lol
08:04:20dbussinka lot of people don't do that
08:04:25michaellatta_ leaves the room.
08:04:31random8rbetter to use those fatarsedreal nums or something right
08:04:54dbussinkwell, floats are usually a lot faster, fpu's are not there for nothing
08:07:00be9 enters the room.
08:07:30random8rjust not a good idea to use floats for currency
08:07:43dbussinkthat's definitely true
08:08:05dbussinkevan: got any new stuff for the cpp vm?
08:09:09twbray leaves the room.
08:12:05thehcdreamer enters the room.
08:13:02evandbussink: been working on rail stuff
08:13:12dbussinkyeah i saw that
08:13:15evanprobably a little tomorrow though
08:13:19evanthought a lot about llvm today
08:13:22dbussinkare those changes still usable?
08:13:38dbussinkhaven't seen to much vm changes lately, so i guess they are
08:13:53evanwhich changes?
08:15:01michaellatta_ enters the room.
08:17:33dbussinkneeded for rails, but i guess the vm didn't need any changes for it to work?
08:17:38dbussink(shotgun that is)
08:18:03dbussinkjust wondering whether there is a lot of work that needs to be redone for the new vm
08:18:13evannope
08:18:24evanone bug in the grammar stuff
08:18:29evanother than that, no vm changes.
08:18:38random8rhow would one run rails, out of curiosity? would one shotgun the server first? it's a little confusing.
08:19:12tarcierievan: you're looking at llvm again?
08:19:22evantarcieri: yep
08:19:30tarcieriand is that what brought up paghetti stacks?
08:19:31evanrandom8r: rbx gem install --no-ri --no-rdoc rails
08:19:38tarcierispaghetti, bleh
08:19:39evanrandom8r: rbx rails my_app
08:19:45evanrbx my_app/script/server
08:19:54evankinda
08:19:55evanyeah
08:20:08evanbeen thinking about execution models again.
08:20:13dbussinktarcieri: spaghetti, yum!
08:20:40tarcieriso you keep closed-over stack frames on the heap, managed by the garbage collector but interacting with LLVM sanely somehow?
08:21:26random8revan sweeet :)
08:22:10evantarcieri: my thinking is to not use LLVM 'correctly' at first
08:22:25michaellatta leaves the room.
08:22:27evanit's perfectly valid, just not as straight forward as one might expect
08:22:34tarcieriaah
08:22:46tarcieriwell sweet
08:22:50evanyou don't use a machine stack frame for locals and such
08:22:56evanyou use a normal MethodContext
08:23:00evan(this is my phase 1 idea)
08:23:30evanand each bytecode method is compiled into an LLVM function which, when it wants to call out to another ruby method, sets up some state, and returns
08:23:44evanit returns to a primary controller, which reads the state that was setup, runs the method
08:24:19tarcieriokay, so you effectively have a sort of shim function between LLVM and MethodContexts?
08:24:30evanon return, the LLVM function is called again, passing in a 'segment id', which is used in jump table, to restore control flow to right after the call
08:24:44evanright
08:24:58tarcierithat's cool
08:24:59evanthe idea being that we'd eliminate bytecode dispatch entirely
08:25:11evanbut still operate in a stackless manner
08:25:41tarcieriyeah, guess I could've said that much about BEAM... it's totally stackless
08:26:59evanmy other thoughts today were on method specialization of closures
08:27:32evanbut i think i've got them figured out
08:34:03michaellatta enters the room.
08:41:06michaellatta_ leaves the room.
08:44:32mutle enters the room.
08:47:56octopod enters the room.
08:51:18sudoer enters the room.
08:52:48benny leaves the room.
08:55:06michaellatta_ enters the room.
09:03:31michaellatta leaves the room.
09:04:22_mk_ enters the room.
09:06:59lstoll leaves the room.
09:09:12jtoy leaves the room.
09:12:46SpakmanI've got a rubyspecs patch which Rubinius fails just now. But, I've also got a Rubinius patch to fix it. I've got commit access to the Rubyspecs repository, but not yet the Rubinius repository. What's the best thing to do? Just commit the spec and create a Rubinius bug? Should I wrap the failing test in a block of some sort (JRuby ddon't want me to)?
09:14:08michaellatta enters the room.
09:16:07krisps enters the room.
09:16:45dbussinkyou can commit it to rubyspec
09:16:51dbussinkand put a patch up on lighthouse
09:16:55dbussinkfor the rubinius patch
09:17:21dbussinkif the test fails on jruby (and won't be fixed there) it should be marked as failing
09:17:43dbussinkif the spec should work on all ruby implementations, then don't wrap it, jruby guys should fix it then
09:21:14michaellatta_ leaves the room.
09:21:39_sk enters the room.
09:24:53imajes enters the room.
09:25:07esfam enters the room.
09:26:38esfamhi.. i'm trying to install rubinius but i get the following error during rake:build -> Unable to send initialize on #<FFI::NotFoundError> (NoMethodError)
09:27:39dbussinkesfam: what platform are you running?
09:27:50esfamos x tiger
09:28:35random8resfam is this a clean build?
09:28:45esfamyes
09:29:14dbussinkcan you try commenting out attach_function 'lchmod', [:string, :mode_t], :int in kernel/platform/posix.rb ?
09:29:21dbussinkjust for a quick test / fix
09:29:28esfamok..
09:30:15dbussinkah wait, that probably won't work because the stables are broken too then
09:30:26dbussinkrue: you there?
09:30:45ruedbussink: Yeah?
09:30:59dbussinkrue: what do you think of rebuilding stables without lchmod for now?
09:31:11dbussinkbecause it breaks on some systems, so those people can't even compile anymore
09:32:57rueJust conditionalise it for PPC and OpenBSD. It does look like PPC is missing it too, so that is the likely candidate.
09:33:18dbussinktiger is missing it
09:33:20dbussinknot ppc
09:33:39rueAh.
09:33:39mutle leaves the room.
09:33:55dbussinkwe would have noticed ppc :)
09:34:41rueI do not think it is used anywhere, FileUtils checks for it
09:34:41dbussinklchown is probably also not present on tiger (from what i've found)
09:34:51dbussinki've written specs for lchmod
09:35:00dbussinkbut there's no internal stuff that needs it
09:35:06rueRight
09:35:10michaellatta_ enters the room.
09:35:29rueJust attach_function ... unless RUBY_PLATFORM and so on unless evan commits his code.
09:36:11dbussinkyeah, well, the problem is that i don't know all idiosyncrasies and rubinius now fails
09:36:19dbussinkand people can't fix it then on their own platform
09:36:57rueI suppose you can just comment it out altogether and make sure we still run. Tag the specs I guess
09:37:13dbussinksomething like that yeah
09:38:27esfamthanks guys :)
09:39:05dbussinkesfam: if you have a recent linux / leopard system available you can test there for now ;)
09:39:07ruedbussink: Change FFI to print the error and `exit 1` instead of raising for all cases
09:39:35dbussinkrue: well, sitting at work atm, so not really time for it
09:40:07dbussinkbut i can check tonight, if people like esfam are patient :P
09:42:15dbussinkrue: or maybe check which ones are really necessary and print a big fat warning if one is missing
09:42:27dbussinkif another one is missing (a non critical one)
09:42:58michaellatta leaves the room.
09:43:21rueNah
09:43:34mutle enters the room.
09:43:57rueI will have it break immediately. If someone has a problem with it, they can fix it.
09:45:03rueTired of playing around with this.
09:45:37dbussinkwell, problem with breaking immediately is that people can't fix it themselves
09:45:45dbussinkbecause it will fail during compile
09:46:02JimMc enters the room.
09:46:05ruePeople cannot fix it themselves now
09:47:22dbussinktrue, and i don't really like that :)
09:48:13rueAny finesse requires building an actual framework to handle it and I am not taking another round of doing that
09:50:14_sk leaves the room.
09:50:30BlackEdder enters the room.
09:50:40_sk enters the room.
09:54:10random8r leaves the room.
09:54:12michaellatta enters the room.
10:02:22michaellatta_ leaves the room.
10:04:37corp leaves the room.
10:08:06TheVoice leaves the room.
10:08:16crafterm leaves the room.
10:08:33imajes leaves the room.
10:09:44jammi leaves the room.
10:11:09jammi enters the room.
10:14:53imajes enters the room.
10:15:14michaellatta_ enters the room.
10:17:28jammi leaves the room.
10:17:31jammi enters the room.
10:18:59imajes leaves the room.
10:20:07imajes enters the room.
10:22:01imajes leaves the room.
10:22:45michaellatta leaves the room.
10:27:06xhanjian enters the room.
10:34:17michaellatta enters the room.
10:39:43_sk leaves the room.
10:40:09_sk enters the room.
10:41:28michaellatta_ leaves the room.
10:55:19michaellatta_ enters the room.
10:58:43boyscout5 commits by Eero Saynatkari
10:58:44boyscout * Updated platform stable to omit lchmod() and use new reporting.; 188c513
10:58:45boyscout * Removed exception handling from attach_method. Print + exit or bypass.; 076c5d6
10:58:46boyscout * Extremely crude "backtrace" in bootstrap raise for some context.; 15af542
10:58:47boyscout * Tagged File.lchmod to fail until lchmod() issue solved.; d2bbdc3
10:58:48boyscout * Commented out attaching lchmod().; 563178e
10:59:00rueesfam: Update, we will see if this works.
11:02:50michaellatta leaves the room.
11:03:47crafterm enters the room.
11:07:31rueTwo minutes for a bot, then I am out.
11:08:22rubuildius_amd64Eero Saynatkari: 188c51352; bin/ci failed! http://rafb.net/p/W3XnOt73.html
11:12:25rubuildius_ppcEero Saynatkari: 188c51352; 2175 files, 7167 examples, 25748 expectations, 0 failures, 0 errors
11:13:03rueHuh. I cannot decide if that is good or bad.
11:14:22michaellatta enters the room.
11:20:50krisps leaves the room.
11:22:13michaellatta_ leaves the room.
11:24:07sudoer leaves the room.
11:27:03Yurikwow, rubinius is definitely cool, ported rbmodexcl to rubinius: http://github.com/yrashk/rbmodexcl/tree/master/rbxmodexcl.rb
11:30:47rueMeh, I have no idea.. I am getting the regular error trying to reproduce here
11:32:06rueUnless it tries to load iconv first.. I dunno.
11:32:33radarek enters the room.
11:34:57qwert666 enters the room.
11:35:24michaellatta_ enters the room.
11:38:56boyscout1 commit by Eero Saynatkari
11:38:57boyscout * Fall back to "normal" errors after Core is loaded for attach_function.; 82bfeb1
11:39:27rueThat should do it or then there is something weird about the run
11:42:24rueEither way, I am going to find out tomorrow.
11:43:32michaellatta leaves the room.
11:48:56rubuildius_amd64Eero Saynatkari: 82bfeb1ff; 2175 files, 7165 examples, 25722 expectations, 0 failures, 0 errors
11:52:08rubuildius_ppcEero Saynatkari: 82bfeb1ff; 2175 files, 7167 examples, 25748 expectations, 0 failures, 0 errors
11:54:25michaellatta enters the room.
12:00:59michaellatta_ leaves the room.
12:02:46Yurik leaves the room.
12:03:56_sk leaves the room.
12:11:35Yurik enters the room.
12:19:29michaellatta_ enters the room.
12:26:30michaellatta leaves the room.
12:32:01atmos leaves the room.
12:32:52crafterm leaves the room.
12:44:32michaellatta enters the room.
12:49:53RyanTM enters the room.
12:53:57michaellatta_ leaves the room.
12:59:10NoKarma enters the room.
13:05:33michaellatta_ enters the room.
13:05:44benny enters the room.
13:11:38michaellatta leaves the room.
13:18:20yugui leaves the room.
13:18:43yugui enters the room.
13:22:47ShayArnett enters the room.
13:24:35michaellatta enters the room.
13:27:44thehcdreamer_ enters the room.
13:32:40michaellatta_ leaves the room.
13:41:42thehcdreamer leaves the room.
13:45:37michaellatta_ enters the room.
13:53:19michaellatta leaves the room.
14:04:39michaellatta enters the room.
14:08:37fmardini enters the room.
14:10:42michaellatta_ leaves the room.
14:12:35AndrewO enters the room.
14:17:55blakewatters enters the room.
14:20:08ShayArnett leaves the room.
14:20:17gramos enters the room.
14:25:42michaellatta_ enters the room.
14:32:37michaellatta leaves the room.
14:33:08gramosHi all
14:33:34gramos I sent a patch to rubyspec (http://rubyspec.org/issues/show/9)
14:34:12gramosbut a I can see that is the number 9
14:35:34gramosIs the system ticket's active?
14:37:48fbuilesvgramos: Rubyspec was forked from Rubinius like 2 weeks ago, so it's normal that the number's so low
14:38:08gramosOkis
14:38:19gramosthanks fbuilesv_
14:38:29fbuilesvgramos: de nada :-)
14:38:46gramos:D
14:38:47fbuilesvalso, for next time try hitting #rubyspec, there's a chan for that now
14:38:49gramosjajaja
14:40:01gramosok
14:40:17enebo enters the room.
14:40:30NoKarmagramos: I'll see if I can get your patch in
14:44:44michaellatta enters the room.
14:45:47moofbong enters the room.
14:45:55NoKarmagramos: it's in
14:46:13gramosNoKarma: great!
14:46:17gramosthanks
14:48:03mutle leaves the room.
14:48:28mutle enters the room.
14:48:32moofbong leaves the room.
14:48:45moofbong enters the room.
14:50:58michaellatta_ leaves the room.
14:56:18mediogre leaves the room.
14:56:58headius enters the room.
15:00:00thehcdreamer_ leaves the room.
15:02:40RyanTM leaves the room.
15:03:08thehcdreamer enters the room.
15:04:14wmoxam enters the room.
15:06:22fbuilesv leaves the room.
15:06:23RyanTM enters the room.
15:06:57RyanTM leaves the room.
15:07:07headius leaves the room.
15:07:38RyanTM enters the room.
15:08:46esfam_ enters the room.
15:08:47RyanTM leaves the room.
15:09:46michaellatta_ enters the room.
15:10:47RyanTM enters the room.
15:14:34yipstar enters the room.
15:15:50Skip leaves the room.
15:16:21Ski1 enters the room.
15:17:34michaellatta leaves the room.
15:22:40nemerle enters the room.
15:24:30headius enters the room.
15:26:24esfam leaves the room.
15:26:52headius leaves the room.
15:26:58headius enters the room.
15:28:33dctanner enters the room.
15:30:48michaellatta enters the room.
15:36:54michaellatta_ leaves the room.
15:37:53nemerle_afk leaves the room.
15:41:29pauldix enters the room.
15:49:50michaellatta_ enters the room.
15:51:50esfamrue: just got back.. thanks! it now compiles :D
15:55:03mapar enters the room.
15:57:10michaellatta leaves the room.
15:59:45gnufied leaves the room.
16:02:14brixenSpakman: ping
16:06:38esfam_ leaves the room.
16:08:52michaellatta enters the room.
16:11:24atmos enters the room.
16:13:49mapar leaves the room.
16:15:57wvdschel enters the room.
16:16:00michaellatta_ leaves the room.
16:16:34trythil___ leaves the room.
16:17:07headius leaves the room.
16:26:44enebo leaves the room.
16:29:55michaellatta_ enters the room.
16:36:45michaellatta leaves the room.
16:39:26ShayArnett enters the room.
16:45:51lopex enters the room.
16:50:49trythil enters the room.
16:54:11evanI looked at perl6 stuff again last night
16:54:17evanman they've jumped the shark.
16:54:57michaellatta enters the room.
17:01:21michaellatta_ leaves the room.
17:02:21dlee enters the room.
17:04:32mapar enters the room.
17:05:35dctanner leaves the room.
17:08:46dctanner enters the room.
17:09:39joachimm enters the room.
17:09:49JimMc leaves the room.
17:10:18ShayArnett leaves the room.
17:10:24trythil leaves the room.
17:11:37chris2 enters the room.
17:12:48fearoffish enters the room.
17:13:12mutle leaves the room.
17:13:59michaellatta_ enters the room.
17:19:47NoKarmaevan: How so?
17:19:52enebo enters the room.
17:20:53michaellatta leaves the room.
17:21:50evanthere are just SO many rules
17:23:32JimMc enters the room.
17:24:01kw enters the room.
17:24:42corp enters the room.
17:25:42joachimmevan: is this an acceptable way of testing vm->inspect() http://pastie.textmate.org/private/lfaswnllx7grmomo9mtg
17:26:01evanyeah, thats fine.
17:27:30shame leaves the room.
17:33:13trythil enters the room.
17:33:41joachimmevan: could you give me commit rights? here is one of my commited patches http://rubinius.lighthouseapp.com/projects/5089/tickets/551-patch-additional-tests-for-float-in-th e-cpp-branch
17:33:50evansure
17:33:56evansend me your ssh key
17:34:38joachimmevan: http://pastie.textmate.org/private/lkvxopubaw3wziwev7gscg
17:34:47dfg59 enters the room.
17:35:02michaellatta enters the room.
17:35:27evanok, added!
17:36:48joachimmevan: thanks.
17:42:08michaellatta_ leaves the room.
17:43:08dfg59 leaves the room.
17:43:45corp leaves the room.
17:45:06dfg59 enters the room.
17:46:19dfg59 enters the room.
17:48:43thehcdreamer leaves the room.
17:50:53wycats leaves the room.
17:52:17thehcdreamer enters the room.
17:53:39benburkert enters the room.
17:53:45fmardini_ enters the room.
17:54:04michaellatta_ enters the room.
18:01:27ShayArnett enters the room.
18:02:00wvdschel leaves the room.
18:02:11thehcdreamer_ enters the room.
18:02:28michaellatta leaves the room.
18:03:31thehcdreamer leaves the room.
18:07:51wycats enters the room.
18:08:36anteaya_ leaves the room.
18:09:18fmardini leaves the room.
18:09:28thehcdreamer_ leaves the room.
18:10:20corp enters the room.
18:10:26edwardam enters the room.
18:12:53anteaya_ enters the room.
18:13:45dctanner leaves the room.
18:15:06michaellatta enters the room.
18:19:06edwardam leaves the room.
18:21:42michaellatta_ leaves the room.
18:22:04caike_ enters the room.
18:22:45naeu enters the room.
18:29:16corp leaves the room.
18:30:01wycats leaves the room.
18:30:41wycats enters the room.
18:31:49yugui leaves the room.
18:33:01octopod leaves the room.
18:33:06corp enters the room.
18:34:12michaellatta_ enters the room.
18:39:05trythil leaves the room.
18:40:04edwardam enters the room.
18:42:29michaellatta leaves the room.
18:44:06TheVoice enters the room.
18:52:44charlenopires enters the room.
18:52:59enebo leaves the room.
18:55:11michaellatta enters the room.
18:56:52gramos leaves the room.
18:59:14trythil enters the room.
19:00:47michaellatta_ leaves the room.
19:01:50dbussinkevan: is there any reason that you added INTEGER as a macro for Integer* and not one for Bignum* ?
19:03:14evanwell
19:03:20radarek leaves the room.
19:03:31evanto match FIXNUM mainly
19:03:37evanI didn't add them for all types
19:03:44dbussinkbut there is FIXNUM and Fixnum* there too
19:03:44evanjust the ones that are immediates
19:03:51dbussinkah no, nvm
19:03:51evansure
19:05:04dbussinkhmm, typedef Fixnum* FIXNUM
19:05:13dbussinkso i don't really see the difference then
19:05:56kw leaves the room.
19:07:56dbussinkevan: i guess we want a full fledged float object in the new vm?
19:08:04dbussinkbecause it's also the source of the primitives?
19:08:09evancourse
19:08:13evanwe had that in shotgun too
19:08:24dbussinknot really, everything was in the primitives
19:08:29evanhuh?
19:08:34dbussinkat least all arithmetic operations
19:08:37evani don't know what ya mean
19:08:45dbussinkunlike fixnum / bignum
19:08:51evanyou mean that in shotgun, there were no float_* functions?
19:09:17dbussinknot in float.c like in bignum.c and fixnum.h
19:09:22dbussinkwas just structured a bit different
19:09:27evanah, ok.
19:09:28evanyeah
19:09:34evanthey should all be structured the same in the new vm
19:09:42evanbecause of how we're going to bind the primitives
19:09:43dbussinkthat was my idea too yeah
19:14:37trythil leaves the room.
19:17:58pauldix leaves the room.
19:18:01pauldix enters the room.
19:18:38atmos leaves the room.
19:18:45atmos enters the room.
19:19:10fmardini_ leaves the room.
19:21:12shame enters the room.
19:26:24enebo enters the room.
19:29:17headius enters the room.
19:31:34trythil enters the room.
19:36:50trythil leaves the room.
19:37:59joachimm leaves the room.
19:42:59_mk_ leaves the room.
19:56:31Salve enters the room.
20:07:33Salveso is rubinius just a new virtual machine?
20:07:54Salveso the garabage collection won't be such a pita?
20:08:37brixenrubinius is many things to many people. yes it's a vm and yes gc is better in it than in mri
20:08:45brixenit's also a bunch more ruby code than in mri
20:08:53brixenand better concurrency primitives
20:09:16brixenit's rather hard to give a tiny sound bite style comparison
20:09:45DefilerRubinius is the development branch of Ruby 2.0
20:10:56Salvedevelopment branch of Ruby 2.0
20:10:59Salve...
20:11:09dbussinkmri just doesn't know yet ;)
20:11:16brixenheh
20:11:48SalveI don't understand, so you guys are using the ruby source along with a new vm?
20:11:55DefilerNo, we are replacing MRI
20:11:58Salveso its like a new ruby package?
20:12:04Salve..googling..
20:12:24DefilerMRI is shorthand for 'ruby 1.8 and 1.9'
20:12:43dbussinkMRI == Matz' Ruby Interpreter
20:13:07Salveinterpreter,...
20:13:20Salvejust when I thought I knew what I was talking about
20:13:23dbussinkSalve: but we use ruby to write significant parts of the implementation itself
20:13:26TheProkrammerMRI is to Rubinius as GCC is to TurboC
20:13:36Salveahh
20:13:49Salvethat helps TheProkrammer
20:14:06TheProkrammeror Watcom is to TurboC
20:14:12Salvehrm,...still scratching my head here about what Rubinius would be
20:14:27dbussinkwell, for java you have sun's jvm
20:14:32dbussinkbut there is also blackdown
20:14:38brixenSpakman: you around?
20:14:41TheProkrammerYou can also use Mono vs .NEt as an example.
20:14:46dbussinkfor .net there's microsoft's implementation, but also mono
20:15:04SalveBecause with Ruby isn't there just a vm that runs ruby? aka turning scripting code into byte code?
20:15:19TheProkrammerRubinius goal, from my understanding, is to have as much as possible actually written in Ruby, where MRI use's quite a bit of C.
20:15:26TheProkrammerdbussink: jinx.
20:15:44dbussinkTheProkrammer: looks like it :)
20:16:43TheProkrammerSalve: the vm's are essentially the interpreter. With 1.8 you have the original VM that Matz wrote, 1.9 brings us YARV, rubinius is a seperate VM, IronRuby is ruby running on the CLI vm, JRuby is ruby running on the Java VM.
20:17:09twbray enters the room.
20:17:23SalveTheProkramer: Bingo!
20:17:46SalveOk,...so it is a VM,...but a seperate one?
20:18:06TheProkrammer1.8/1.9 is mostly written in C, IronRuby is mostly written in C#, JRuby is mostly written in Java, Rubinius aims to be written mostly in Ruby.
20:18:12SalveSo does it run with 1.9?
20:18:21DefilerIt replaces 1.9
20:18:29SalveOh cool now I get it
20:18:38SalveIts a vm written in ruby
20:18:48DefilerIt is a VM, kernel, and compiler
20:19:05NoKarmawell, the vm is written in c/c++, but all the other parts are written in ruby
20:19:06TheProkrammerIt aims to be 1.8.7 complient, as does most of the other alternative implemenations.
20:19:07SalveDefiler: why the distinction?
20:19:09DefilerThe VM is currently written in C++
20:19:31TheProkrammerIt's an implementation, including all the things Defiler mentioned.
20:19:51SalveSo is it written in c++ with ruby hooks?
20:20:36DefilerThe VM is written in C++, and doesn't know anything about Ruby
20:20:40SalveLike import kernel.so inside a ruby file
20:20:43DefilerEverything else (with some exceptions I am glossing over) is Ruby
20:21:03Defilerthe compiler (written in Ruby), takes ruby code and emits Rubinius bytecodes
20:21:06Defilertha