Index

Show enters and exits. Hide enters and exits.

00:02:24evanthe ones that unpack array literals are stupid.
00:02:39evanit's exactly the same as unpacking a variable
00:02:45evanwhy those exist, i don't know.
00:04:51radarek leaves the room.
00:09:50fbuilesvevan: can you update my key for the git repos? Last one apparently was banned due to the Debian issue.
00:10:12fbuilesvnew one's at http://pastie.caboo.se/private/qa6o1bx3cfnuefdeuais6q
00:10:20evansure
00:10:26fbuilesvthanks
00:10:36fbuilesvmail is federico.builes@gmail.com in case it's needed
00:11:14jnicklas enters the room.
00:11:18evanok, added.
00:12:50fbuilesvthat worked, thanks!
00:15:21jnicklasIs it okay to run 'rake install' after building?
00:15:43headius enters the room.
00:27:21jnicklas leaves the room.
00:39:51yipstar leaves the room.
00:39:53joachimm_ leaves the room.
00:40:33foysavas enters the room.
00:44:33tonga enters the room.
00:44:43tongaHi, all
00:58:25wmoxam enters the room.
01:03:46anteaya__ enters the room.
01:05:14TheVoice enters the room.
01:06:30anteaya_ leaves the room.
01:06:49headius leaves the room.
01:08:21headius enters the room.
01:09:30benburkert_ leaves the room.
01:10:18radarek enters the room.
01:10:56dctanner enters the room.
01:11:13krisps leaves the room.
01:11:39benburkert_ enters the room.
01:12:45benburkert_ leaves the room.
01:16:41radarek leaves the room.
01:25:34dctanner leaves the room.
01:25:55yipstar enters the room.
01:34:50imajes_ leaves the room.
01:35:51imajes_ enters the room.
01:39:32qwert666 leaves the room.
01:47:07Yurik leaves the room.
01:50:58crafterm enters the room.
02:00:25jtoy enters the room.
02:13:47sudoer enters the room.
02:15:19kw enters the room.
02:15:38lopex leaves the room.
02:16:31imajes_ leaves the room.
02:24:09jtoy leaves the room.
02:26:54MenTaLguY enters the room.
02:27:02MenTaLguYhowdy
02:27:53tarcierihttp://www.bileblog.org/2008/05/java-haters-gtfo/ lol
02:38:03boyscout3 commits by Evan Phoenix
02:38:04boyscout * Fix class variable setting; c06b3bc
02:38:05boyscout * Raise if 'raise()' is called when ?! is nil; d39042d
02:38:06boyscout * More rails fixes; 4139d09
02:38:22evanwith those last fixes, i was able to run the rpaste demo app on rubinius
02:41:19blakewatters enters the room.
02:42:54robin_dewd enters the room.
02:43:21blakewatters leaves the room.
02:44:17blakewatters enters the room.
02:45:31ezmobius leaves the room.
02:46:44Reason enters the room.
02:52:18VVSiz_ enters the room.
02:53:54rubuildius_ppcEvan Phoenix: c06b3bc94; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
02:54:51craftermevan: awesome mate :)
02:56:46evanyeah
02:57:06evandouble awesome is that the fixes i've made over the last 48 hours to run more and more rails have been small
02:57:12evanno huge missing holes
02:57:25evanor fundementally misarchitected features
03:00:34VVSiz leaves the room.
03:06:49yipstar leaves the room.
03:09:16imajes_ enters the room.
03:21:23lstoll enters the room.
03:22:43cored leaves the room.
03:30:30twbray enters the room.
03:32:08twbray leaves the room.
03:33:47twbray enters the room.
03:36:48RyanTM leaves the room.
03:38:35RyanTM enters the room.
03:43:28blakewatters leaves the room.
03:48:32gramos enters the room.
03:51:35gramosHi I am trying to start with rubynius and I have somes problems to build it
03:52:19gramosSomebody can help me?
03:53:01twbray leaves the room.
03:54:10obvio enters the room.
03:54:11gramosThis is the error message
03:54:23obvio171 leaves the room.
03:54:23gramosGenerating lib/zlib.rb...
03:54:23gramosreadline.c:8:31: error: readline/readline.h: No such file or directory
03:56:06mernenmaybe you're missing libreadline?
03:56:27mernenfor ubuntu, it's libreadline5-dev
03:56:52gramosokis, I am on debian testing
03:57:11mernenshould be the same package then :)
03:57:16mernenhttp://rubinius.lighthouseapp.com/projects/5089/installation
03:57:19gramosyes
03:57:37mernenthey suggest the following packages: libglib2.0-dev zlib1g-dev build-essential libreadline5-dev bison
03:59:55yugui enters the room.
04:00:09gramosI checked this link
04:01:09gramosbut i don't saw libreadline package in the list
04:02:27gramosI am trying rebuild now
04:04:20wmoxam leaves the room.
04:04:22gramosIt's work
04:04:23gramos!
04:04:24gramos:D
04:06:15dysinger_ leaves the room.
04:08:24dysinger enters the room.
04:08:45yugui leaves the room.
04:11:28yugui enters the room.
04:13:06mernen leaves the room.
04:13:28obvio leaves the room.
04:15:31knowtheory enters the room.
04:16:00dysinger leaves the room.
04:16:31robin_dewd leaves the room.
04:18:17robin_dewd enters the room.
04:21:26obvio171 enters the room.
04:25:37binary42 enters the room.
04:26:00robin_dewd leaves the room.
04:26:35wmoxam enters the room.
04:38:17obvio171 leaves the room.
04:52:45fowlduck enters the room.
04:53:45rubuildius_amd64Evan Phoenix: c06b3bc94; 2175 files, 7035 examples, 25543 expectations, 0 failures, 0 errors
04:58:05michaellatta enters the room.
05:02:24gramos leaves the room.
05:07:59MenTaLguY leaves the room.
05:10:40RyanTM leaves the room.
05:27:58knowtheory leaves the room.
05:28:36fowlduck leaves the room.
05:31:33crafterm leaves the room.
05:33:09michaellatta enters the room.
05:35:32michaellattaNow that Rubinius runs Merb, what is the relative performance and stability? Is there a target for a 1.0 production ready release?
05:36:18AndrewO leaves the room.
05:36:20DefilerNobody has done any performance testing with real apps yet, to my knowledge
05:36:25Defilerbut we are a lot slower than MRI right now.
05:38:01michaellattaOk. Some early micro-benchmarks were pretty competitive, or even faster. Has the speed dropped due to all the libraries being in Ruby, or was my first impression mistaken for the early micro-benchmarks?
05:38:31DefilerImpressive micro-benchmarks come much earlier than real performance in actual code, sadly
05:38:36crafterm enters the room.
05:38:42DefilerThey are necessary but not sufficient
05:39:57DefilerThere will be a different answer in 6 months
05:40:22michaellattaHere is hoping someone has the resources to put a real dynamic compiler under Rubinius I guess. That would be a fun project to take on.
05:40:35DefilerWe have those resources
05:40:55DefilerThough if someone wanted to contribute some code that would be neat
05:41:42michaellattaSo that is in the plan already? Not for 1.0 I am sure.
05:42:19Defiler1.0 is coming in a couple of weeks
05:42:22DefilerSo no. Heh
05:43:48DefilerThe goal is real deal JIT with all the trimmings etc
05:44:09DefilerOr something better
05:44:59michaellattaSo the 1.0 target is feature complete and slow, and a later realse like 2.0 having high performance? That would make sense, the problem as seen in many languages is that until the performance is there people do not use it for real work. Java did not really take off until 1.2 and the HotSpot JVM.
05:45:22DefilerRuby already has a working implementation, so we hopefully don't have that problem
05:45:52michaellattaYou can definitely do better than a JIT these days, heck Smalltalk and Self had better 10 years ago or more, possibly 20.
05:46:02DefilerIndeed
05:46:32DefilerVPRI's IS stuff is the way I think these things are headed. So awesome.
05:46:43DefilerHopefully we can do half as well
05:46:46michaellattaNo question Ruby is serious and alive. But, Rubinius will not get heavy production use until the performance is competitive at least.
05:46:56anteaya__ leaves the room.
05:46:58DefilerWell, we're doing what we can.
05:47:32DefilerIt's not a popularity contest, it's a war to the death and beyond.
05:47:44michaellattaI certainly appreceate the work to date, and want it to pay off for you.
05:48:51DefilerThanks. Hopefully it will pay off while I sleep, in the next few minutes. :)
05:51:36imajes_ leaves the room.
05:53:10imajes_ enters the room.
06:06:53mediogre enters the room.
06:12:14imajes__ enters the room.
06:19:53imajes_ leaves the room.
06:37:01trythil enters the room.
06:38:43binary42 leaves the room.
06:39:19elight leaves the room.
06:42:06rubuildius_amd64 leaves the room.
06:42:10rubuildius_amd64 enters the room.
06:45:39djwhittsorry about the error being sans pastie. I was traveling over the weekend so I didn't see it till just now
06:45:57djwhittit should be fixed now
07:01:05xhanjian_ enters the room.
07:13:38xhanjian leaves the room.
07:22:15crafterm_ enters the room.
07:24:38Skip enters the room.
07:27:52crafterm_ leaves the room.
07:27:54lstoll leaves the room.
07:27:58Maledictus enters the room.
07:31:32crafterm leaves the room.
07:39:21djwhittbrixen: just saw the bit about the number of expectations too. I'll look into that tomorrow ... err today I guess, but after I sleep ;)
07:40:10mkrauskopf enters the room.
07:52:04kw leaves the room.
08:06:30mediogre leaves the room.
08:12:50NoKarma enters the room.
08:13:35thehcdreamer enters the room.
08:13:58krisps enters the room.
08:17:06mkrauskopf leaves the room.
08:22:47headius_ enters the room.
08:22:47headius leaves the room.
08:25:30headius enters the room.
08:25:30headius_ leaves the room.
08:25:44ezmobius enters the room.
08:26:56headius_ enters the room.
08:28:28yugui leaves the room.
08:30:12headius leaves the room.
08:34:22JimMc enters the room.
08:45:18xhanjian enters the room.
08:52:19benny leaves the room.
08:58:04octopod enters the room.
08:58:13lypanovwakes up again
08:58:16xhanjian_ leaves the room.
08:59:49olabini leaves the room.
09:09:05NoKarmahey
09:17:49VVSizmorning! :)
09:18:18VVSizNoKarma: it's rainy outside, so I expect more specs from you :) heheheh
09:18:19Yurik enters the room.
09:18:28NoKarmaVVSiz: heya
09:18:38NoKarmaVVSiz: what, even more?!
09:18:49NoKarma;)
09:18:58VVSizyeah! until github explodes!!! (and my head trying to keep up)
09:19:13NoKarmagsoc hasn't even started yet
09:19:33NoKarmaAnd I don't want to have no work then ;)
09:22:38VVSiz:)
09:26:46olabini enters the room.
09:28:31benny enters the room.
09:42:25trythil leaves the room.
09:45:19_mk_ enters the room.
09:46:35ezmobius leaves the room.
09:48:16qwert666 enters the room.
09:53:37olabini leaves the room.
09:53:59olabini enters the room.
09:57:08xhanjian leaves the room.
10:04:01jammi leaves the room.
10:07:08jammi enters the room.
10:08:28headius leaves the room.
10:08:34krisps leaves the room.
10:12:09olabini leaves the room.
10:12:51olabini enters the room.
10:13:31sudoerdoes Exception.class catch everything and then I can override it to do different things when an exception happens in rubinius?
10:13:45sudoerit seems like you cant do this with mri
10:16:07radarek enters the room.
10:18:56imajes__ leaves the room.
10:19:38imajes_ enters the room.
10:19:45BlackEdder enters the room.
10:42:46imajes_ leaves the room.
10:45:57mkrauskopf enters the room.
10:49:39gnufied enters the room.
10:54:18radarek leaves the room.
10:54:18olabini leaves the room.
10:54:33radarek enters the room.
10:54:33olabini enters the room.
10:57:32olabini leaves the room.
10:57:32radarek leaves the room.
10:58:20radarek enters the room.
10:58:20olabini enters the room.
11:16:23olabini leaves the room.
11:16:23radarek leaves the room.
11:16:41radarek enters the room.
11:16:41olabini enters the room.
11:34:02sudoer leaves the room.
11:45:15robin_dewd enters the room.
12:02:32yugui enters the room.
12:26:28NoKarma leaves the room.
12:26:42NoKarma enters the room.
12:50:28RyanTM enters the room.
13:03:40mkrauskopf leaves the room.
13:06:38imajes_ enters the room.
13:12:41imajes leaves the room.
13:18:45RyanTM leaves the room.
13:19:43RyanTM enters the room.
13:33:53dodecaphonic enters the room.
13:47:33w1rele55 enters the room.
13:50:57pd leaves the room.
14:06:19gramos enters the room.
14:14:03blakewatters enters the room.
14:16:27AndrewO enters the room.
14:26:13dodecaphonic leaves the room.
14:51:39moofbong enters the room.
14:53:19BlackEdder enters the room.
15:02:24yugui leaves the room.
15:02:46yugui enters the room.
15:19:20enebo enters the room.
15:21:12yipstar enters the room.
15:28:17trythil enters the room.
15:31:57shame leaves the room.
15:59:10dctanner enters the room.
16:05:49jtoy enters the room.
16:07:19dsmalley enters the room.
16:09:37nicksieger leaves the room.
16:10:05nicksieger enters the room.
16:13:58ShayArnett enters the room.
16:14:29kli_ enters the room.
16:32:50dctanner leaves the room.
16:35:54jtoy leaves the room.
16:42:59w1rele55 leaves the room.
16:44:16shame enters the room.
16:48:11chris2 enters the room.
16:48:31dlee enters the room.
16:57:48lopex enters the room.
17:20:35benburkert enters the room.
17:20:52gnufied leaves the room.
17:22:12JimMc leaves the room.
17:24:07hoopy leaves the room.
17:25:43gramos leaves the room.
17:26:39yugui leaves the room.
17:42:27thehcdreamer leaves the room.
17:43:14enebo leaves the room.
17:43:58thehcdreamer enters the room.
17:47:05jeremydurham enters the room.
17:51:23jtoy enters the room.
17:51:50trythil leaves the room.
17:54:13benburkert leaves the room.
17:57:24jeremydurham leaves the room.
17:58:55jeremydurham enters the room.
18:01:55thehcdreamer leaves the room.
18:02:18RyanTM leaves the room.
18:02:48enebo enters the room.
18:03:04jtoy leaves the room.
18:04:14twbray enters the room.
18:05:04sambo enters the room.
18:08:43ShayArnett leaves the room.
18:10:11TheVoice enters the room.
18:14:15octopod leaves the room.
18:19:35TehLaser leaves the room.
18:25:36jeremydurham leaves the room.
18:25:50smparkes enters the room.
18:42:09thehcdreamer enters the room.
18:46:54headius enters the room.
18:47:11twbray leaves the room.
18:54:09Arjen_ enters the room.
18:57:44eventualbuddha enters the room.
19:00:27trythil enters the room.
19:00:30thehcdreamer leaves the room.
19:01:32dfg59 enters the room.
19:01:52sambo leaves the room.
19:01:59jeremydurham enters the room.
19:12:54twbray enters the room.
19:15:20twbray leaves the room.
19:20:27nemerle enters the room.
19:25:03TehLaser enters the room.
19:28:55headius leaves the room.
19:29:10headius enters the room.
19:37:03corp enters the room.
19:37:07nemerle_afk leaves the room.
19:39:28michaellatta_ enters the room.
19:43:00fbuilesv leaves the room.
19:44:06olabini leaves the room.
19:45:56michaellatta leaves the room.
19:49:17ShayArnett enters the room.
19:51:16benburkert enters the room.
19:56:15eventualbuddha leaves the room.
20:00:30michaellatta enters the room.
20:02:24ShayArnett leaves the room.
20:03:37dodecaphonic enters the room.
20:07:00michaellatta_ leaves the room.
20:07:54trythil leaves the room.
20:14:39pauldix enters the room.
20:16:22corp_ enters the room.
20:17:57twbray enters the room.
20:18:38gohanlon enters the room.
20:19:32michaellatta_ enters the room.
20:23:08twbray leaves the room.
20:24:01olabini enters the room.
20:26:02michaellatta leaves the room.
20:26:08corp leaves the room.
20:26:42twbray enters the room.
20:28:56dysinger enters the room.
20:30:30twbray leaves the room.
20:34:37RyanTM enters the room.
20:37:52twbray enters the room.
20:41:35pauldix leaves the room.
20:43:53dodecaphonic leaves the room.
20:45:44gohanlonhello all, i've got a general question: what's the absolute minimal ram requirements for a rubinius vm process? (i'm willing to make tweaks that would decrease the ram requirement as much as possible, for example, adjusting GC to reserve less ram)
20:48:04brixengohanlon: looks to me like on leopard I'm right about 20 mb just doing a p 1; sleep 0.5 in a loop
20:48:22brixenbetween 18-26 mb looking at activity monitor
20:48:32gohanlonouch!
20:48:49brixenthat's for RSIZE
20:49:08brixenwell, you wanted something to do, right? :P
20:49:21gohanlonfor MatzRuby I get around 1.2 mb
20:49:57brixennah, I get 29 mb RSIZE with mri
20:50:06brixenagain, in activity monitor
20:50:56gohanlonbrixen: are you using Leopard Ruby, or a debug build?
20:51:24brixenheh, wait, I had another ruby process running
20:51:27brixenle'me try that again
20:51:52brixenyep, about 1.3 mb
20:52:39gohanlonbrixen: cool. and indeed i am looking for something to do, but what i'm looking to do is put ruby on a microchip for an embedded project, and we're going to have about 3K for RAM and about 500K for code. i'm wasn't feeling optimistic about squeezing 1.2 MB into 3K. ;-)
20:53:44brixenheh, quantum compression :)
20:53:48gohanloni thought that perhaps most of that RAM was eagerly allocated since everything is optimized for speed and machines have gigs of RAM
20:53:49gohanlonlol
20:54:34gohanlonthanks for the infoo!
20:54:47brixenso, running the rubyspecs with mri, i get around 15.5 mb max
20:55:14brixengohanlon: sure thing, a mini-ruby on embedded chip would be really sweet
20:55:21brixenhope you try it
20:56:53gohanlonbrixen: i'm pretty sure it is possible, but i didn't want to have to take a jump up to super powerful embedded chips, since the applications for those chips is much more limited (larger and suck more power). i hope i try it, too!
20:57:21_mk_ leaves the room.
20:57:23brixengohanlon: are you a university student?
20:57:56gohanlonbrixen: used to be, and in a way, i am again ;-)
20:58:58brixencool. it's neat to hear about what folks are up to
20:59:12brixengohanlon: what uni if you don't mind my asking?
20:59:28gohanlonbrixen: sure, UC Santa Cruz
21:02:46brixenah cool
21:04:47corp leaves the room.
21:04:52ezmobius enters the room.
21:09:44imajes enters the room.
21:15:17drbrainlet's see what I can do about these continuous :recursives
21:15:58drbrainum, help!
21:16:12drbrainmspec and spec/frozen are marked as modified for me
21:16:32zenspidergot it
21:18:14twbray leaves the room.
21:18:54twbray enters the room.
21:20:51drbrainall better
21:21:14brixenyay
21:21:19brixendrbrain: what did you need to do?
21:21:43drbrainrm those two dirs, and a couple rake/rake:pull
21:21:46boyscout1 commit by Brian Ford
21:21:47boyscout * Updated MSpec submodule to c1d6eb4b; 3fc0616
21:22:25brixenyeah, I've found when in doubt, that's the ticket
21:22:57imajes leaves the room.
21:23:15imajes enters the room.
21:26:02anteaya enters the room.
21:26:11gramos enters the room.
21:26:19brixendrbrain: what's the failsafe way to know rubygems is loaded if I want to look at a class/module existing? Gem?
21:26:41zenspideryup
21:26:48brixenzenspider: thanks
21:27:37sambo enters the room.
21:28:36VVSizdrbrain: when I see that mspec or spec/frozen are marked as modified, I invoke: 'git submodule update'
21:28:57dysinger_ enters the room.
21:29:04zenspiderVVSiz: cool. good to know
21:29:09drbrainthat's too much to remember
21:29:52zenspiderthat's why I'm adding it to my rake task: rake ryan:update
21:30:01kli_ leaves the room.
21:30:17VVSizyeah, I just set my history to 100000 lines, and do search there when need to remember the command :)
21:30:20brixenand indeed, there is rake [m]spec:update that does that
21:31:09zenspiderbrixen: are those covered in the regular git:update?
21:31:27zenspiderif so, there is something more that needs to be done. I still need to nuke mspec once in a while
21:31:29brixenyeah
21:32:00brixenhmm, since we added spec:update before and after git:update runs?
21:33:36zenspiderI think so, tho it may have butted up against the update... I'll keep poking and see when it next happens
21:33:44brixencool
21:33:47brixenfood, bbiab..
21:34:04rubuildius_amd64Brian Ford: 3fc061699; 2175 files, 7035 examples, 25543 expectations, 0 failures, 0 errors
21:34:04imajes_ enters the room.
21:35:34kw enters the room.
21:36:49rubuildius_ppcBrian Ford: 3fc061699; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
21:38:53imajes_ leaves the room.
21:39:11gramosHi, all
21:39:39hoopy enters the room.
21:40:23gramosI am starting with rubinius and now I am runing the specs
21:40:59gramosMy plattform is: Linux Debian Testing 2.6.22-1-k7 #1 SMP
21:41:10gramosruby 1.8.6 (2008-03-03 patchlevel 114) [i486-linux]
21:41:26gramos1281 files, 5648 examples, 20891 expectations, 2 failures, 0 errors
21:41:55imajes leaves the room.
21:42:09gramos Module#__script__ {} at ./spec/frozen/1.8/core/dir/shared/glob.rb:212
21:42:16VVSizgramos: that's expected. a small number of specs do fail on MRI depending on the platform.
21:42:42VVSizit would be great to mark those as such and make sure everything is clean with at least MRI.
21:42:49zenspidergramos: cool. welcome.
21:42:58gramosWhat's MRI?
21:43:09VVSizMatz's Ruby Interpretery
21:43:10drbraincorundum: MRI?
21:43:11corundum*shrug*
21:43:15brixengramos: also, patchlevel 111 is still the standard. we're probably not going to use patchlevel 114
21:43:43VVSizbrixen: that's news to me ;) I tend to use pl 114
21:43:51Skip leaves the room.
21:44:11brixenVVSiz: well, what happened with the weirdness from rexml? was that resolved?
21:44:25gramosok, may should be install patchlevel 111?
21:44:25zenspiderbrixen: no, 114 is the current release
21:44:29corp enters the room.
21:44:35drbraincorundum: MRI is Matz' Ruby Interpreter, typically the 1.8 branch
21:44:36brixenVVSiz: and I don't recall the other issues, but it seemed like 114 had lots of regressions
21:44:36corundumokies!
21:44:45gramos114 is the current release for debian testing
21:44:53dfg59zenspider: any further progress on the File#join stuff? I was watching the ticket.
21:45:55VVSizbrixen, corundum: looking and git diff between pl 111 and pl 114, most of the changes (all) are in rexml and rss files (and one webrick change - security fix I think)
21:46:09brixenVVSiz: ok, good to know
21:46:12VVSizso for most part, it doesn't really matter much
21:46:17zenspiderdfg59: they're all passing.... wrt the latest message on the ticket I may just ignore it... they're non use-cases anyhow and frankly imo he's just documenting odd bullshit from MRI that nobody needs
21:46:21drbrainVVSiz: um, corundum is a bot
21:46:57VVSiz:)
21:46:58dfg59zenspider: yeah, i was getting that back and forth stuff too, either we need to update the specs or forget about it
21:47:46VVSizpastie: for brixen
21:47:58pastiebrixen: http://pastie.org/199796 by VVSiz.
21:48:08zenspiderdfg59: I told him he had one more chance to add to the specs, and I'll make em pass... but I'm not doing any back and forth anymore. esp not on bullshit turds like File.join("a", "/////")
21:48:15VVSizthat's the whole changelog between 111 and 114
21:48:28dfg59zenspider: yeah, understandable
21:49:05brixenVVSiz: hmm, do you remember who was working on rexml? because they ran into a bunch of problems with 114
21:49:38zenspiderbrixen: what's your policy on speccing/documenting the BS of mri (like above)?
21:49:57dfg59yeah, awesome question. i was thinking the same thing
21:49:59brixenzenspider: file a ruby bug
21:50:09VVSizI tought it was micheael klishin... Also, fbuilesv is supposed to provide the full set of specs as part of GSoC program :)
21:50:12zenspiderI couldn't care less about shit like the above, or File.join with recursive arrays. they're non-usecases imfo
21:50:12drbrainum, my rubinius is getting stuck in machine_error_reporter
21:50:26gramosDir.[] orders directory-based entries before files when a glob matches both FAILED
21:50:31drbrainI just did a rake clean; rake
21:50:33gramosthis is the failure
21:50:35brixendrbrain: can you tell corundum that "ruby standard" is http://rubyspec.org/wiki/rubyspec/Current_Standard
21:50:59drbrainbrixen: you can :)
21:51:13brixendrbrain: oh, I thought you said he only responds when you talk to him
21:51:13drbraincorundum: ruby standard is http://rubyspec.org/wiki/rubyspec/Current_Standard
21:51:14corundumcan do!
21:51:18drbrainbrixen: like so
21:51:23brixencool
21:51:26drbrainhe only responds when addressed
21:51:30brixenahh
21:51:59dysinger leaves the room.
21:52:56zenspiderbrixen: you in the generic, not drbrain specifically. :P
21:53:04drbraingah! rake spec works
21:54:46twbray leaves the room.
21:55:53drbrainis shotgun/rubinius working for everybody else?
21:56:04zenspiderheh. I have a prompt with shotgun/rubinius
21:56:14drbrainits my .irbrc
21:56:35VVSizbrixen: btw, MRI pl 114 passes *all* rexml rubyspecs at the moment. pl 111 fails one.
21:57:17brixenVVSiz: ok, we'll do 114 then :)
21:57:54VVSizalso, I see that pl115 was tagged a couple of days ago. (no rexml changes there, so if something was severely broken on XML, ruby-core seems to be either not aware of them, or they consider no problems exist there)
22:00:16zenspiderVVSiz: how's it do on the rest of the specs tho?
22:00:59VVSizzenspider: let me try that. I think I tried some time ago and it was the ususal 4-5 failures. But I'll re-run.
22:01:44drbrainhrm, it was the .irbrc.compiled.rbc that did it
22:01:51dysinger_ leaves the room.
22:02:07sambo leaves the room.
22:02:27sambo enters the room.
22:03:29dysinger enters the room.
22:03:54VVSizzenspider, brixen: MRI 1.8.6 pl 111 - 5 failures/2 errors. MRI 1.8.6 pl 114 - 4 failures/2 errors.
22:04:00corp leaves the room.
22:04:42VVSizpastie:
22:05:18dysinger leaves the room.
22:05:22pastiehttp://pastie.org/199805 by VVSiz.
22:05:53VVSizall the usual failures I'v seen before, except for one new one: File.lchmod changes the file mode of the link and not of the file.
22:06:04VVSizMRI just doesn't have File.lchmod implmented on linux! :)
22:06:33nicksieger leaves the room.
22:06:37Maledictus leaves the room.
22:07:05corp enters the room.
22:08:07atmos_ enters the room.
22:09:58anteaya leaves the room.
22:10:40gohanlon leaves the room.
22:11:09VVSizfor the reference, today's build of MRI 1.8.7 fails 79 specs.
22:12:53zenspiderouch
22:12:59NoKarmathat's bad
22:13:21VVSizsome are bugs, but many are backported fixes from MRI 1.9
22:13:32VVSizor should I say, new features
22:13:41evanthe bugs are backported?
22:13:43evan:)
22:13:45VVSizhehehe
22:14:23VVSizbut the good news is that some ruby-core folks are actually starting to look at the specs, and even fixed some
22:14:43evanyeah, i've heard akr is fixing some
22:14:49VVSizyep
22:15:03zenspiderbbiab
22:15:20evank.
22:15:21Arjen_ leaves the room.
22:16:03anteaya enters the room.
22:16:55radarek leaves the room.
22:17:27VVSiztrying to run todays MRI 1.9 aganist the specs. SCARY!!! hangs on two Hash (merge, update) specs, on Thread.stop. Crashes on time specs.
22:18:09NoKarmaVVSiz: MRI hangs on some of the IO specs here
22:18:47VVSiz1.8.6? weird.
22:19:33VVSizOK, here are the sad results of running MRI 1.9 against specs (excluding Hash, Time, Thread, Socket - these are hanging or crashin):
22:19:36VVSiz650 failures, 387 errors
22:20:31NoKarmawow
22:23:34imajes enters the room.
22:23:39evanwow.
22:24:17dfg59jebus
22:24:26tarcierihehe
22:25:17evantarcieri: hey there
22:25:27tarcierisup
22:25:35evantarcieri: do you know much about the architecture of BEAM?
22:25:57tarcieriI know some of the high level aspects of it
22:26:33evani'm curious if it uses a spaghetti stack or not
22:26:57evanand if not, what it uses to make the tasks lightweight and switch quickly between them
22:27:06tarcieriI have no idea how it implements the stack... in that regard I don't really know anything beyond that it's a register-based VM
22:27:10imajes leaves the room.
22:27:12imajes_ enters the room.
22:27:47evanok.
22:27:48evanjust curious.
22:28:03blakewatters leaves the room.
22:28:25tarcierithe only things I've really looked at in depth are the SMP scheduler and the various heap implementations
22:28:32VVSizthe funny part about MRI 1.9 crashing on Time specs is that it crases right after it sees: if defined? Rubinius && Rubinius::WORDSIZE == 32; end :)
22:28:48drbrainVVSiz: that's a bug
22:29:07drbrainit should be if defined?(Rubinius) && Rubinius::WORDSIZE == 32
22:29:14drbraindefined? has higher precedence than &&
22:29:48drbrainVVSiz: can you fix that spec?
22:30:08VVSizdrbrain: YAY! :) changing that stopped MRI from crashing :)
22:30:36VVSizas for the whole check, I think we have a better guard for that :wordsize in mspec. Will fix.
22:30:38drbraincool
22:30:44VVSizheheh
22:30:48drbrainthat bit me a bunch of times in RubyGems
22:30:50krisps enters the room.
22:30:51evansounds like 1.9's defined? can't handle expressions
22:31:08drbrainno, the defined? was wrong
22:31:20evanright, but 1.9 shouldn't crash.
22:31:20evan:)
22:31:21drbrainyou only want to check Rubinius::WORDSIZE if Rubinius is defined?
22:31:22tarcierievan: pretty much everything I learned about BEAM was just from talking to people in #erlang so you might ask them
22:31:26drbrainoh, but, yes
22:31:34evantarcieri: ok.
22:31:47evandrbrain: you working on the Dir.glob problem?
22:32:17drbrainevan: yes
22:32:28boyscout4 commits by Eric Hodel
22:32:29boyscout * Sort methods in Dir; e7f720c
22:32:35boyscout * Remove whitespace from dir.rb; fa05821
22:32:35boyscout * Reduce whitespace; eca21f9
22:32:35boyscout * Some style fixes for File::fnmatch and friends; 9c1fc09
22:32:51drbrainI have reproduced
22:33:16evanok.
22:33:30evanseems like around linke 216, it's building up new_pattern wrong
22:33:39evani narrowed it down to there
22:33:47drbrainyeah
22:34:05drbrainI still remember most of what I translated from C
22:34:12drbrainso it shouldn't be too long to fix
22:35:15evanthats what i'd hoped.
22:35:17evanawesome.
22:36:18drbrainah-ha
22:36:29drbrainI mean, not that I fixed it
22:36:50drbrainit goes from **/tasks/**/*.rake to **/**/*.rake
22:36:57evanyep
22:37:08evanthats what i'd found.
22:37:11VVSizdrbrain: done. http://github.com/brixen/rubyspec/commit/a0435d8f08d45581f9e5d5c9cbd940187b7167d0
22:37:40drbrainmuch better
22:40:49sambo82 enters the room.
22:42:28mapar enters the room.
22:43:18hoopy_ enters the room.
22:44:10rubuildius_amd64Eric Hodel: e7f720cad; 2175 files, 7035 examples, 25543 expectations, 0 failures, 0 errors
22:44:14headius leaves the room.
22:47:34samruby enters the room.
22:47:49rubuildius_ppcEric Hodel: e7f720cad; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
22:49:49samruby_ enters the room.
22:51:26samruby__ enters the room.
22:52:56MenTaLguY enters the room.
22:53:05samruby___ enters the room.
22:53:43evaneyes the multiple joiner samruby
22:53:55MenTaLguYhowdy
22:53:57hoopy leaves the room.
22:54:00evanMenTaLguY: hey there.
22:54:14MenTaLguYI've got a best practices question about specs
22:54:26tarcierisup MenTaLguY
22:54:27MenTaLguYif a spec needs to create temporary files, where should it put them?
22:55:41boyscout1 commit by Ryan Davis
22:55:42boyscout * Fixed guard for DelegatedMethod; e8995e7
22:55:44brixenMenTaLguY: there's a helper for that
22:55:57brixenMenTaLguY: use tmp('some_name.ext')
22:56:07brixenand it will put it in the temp dir for your platform
22:56:13samruby____ enters the room.
22:56:20MenTaLguYbrixen: hm, what about directories?
22:56:40brixenit just returns a string
22:56:45MenTaLguYok
22:56:49brixenDir.mkdir tmp('this')
22:56:53RyanTM leaves the room.
22:57:30RyanTM enters the room.
22:57:44MenTaLguYtarcieri: I think I've found a nice variant on active objects for Ruby
22:57:54MenTaLguYtarcieri: http://github.com/mental/minicomic/tree/minicomic-ng/lib/minicomic/util/async.rb
22:58:35tarcierihmm, that's a cool approach
22:58:40samruby_____ enters the room.
22:59:07tarcieriMenTaLguY: I eventually want to do this in Reia, and something like that would be super easy with smerl
22:59:16MenTaLguYnods
22:59:47tarcierigot this sort of thing going this weekend:
22:59:47tarcieri>> lists:seq(1,10).reverse()
22:59:47tarcieri[10,9,8,7,6,5,4,3,2,1]
23:00:51tarcieriI'd like all the non-primitives to be their own process, and in that case doing that sort of invocation would do RPC
23:02:00MenTaLguYer, each non-primitive its own process?
23:02:31tarcierithat's the plan
23:02:35sambo leaves the room.
23:02:50sambo enters the room.
23:02:57MenTaLguYI hate to say it but 300 bytes isn't much for a process or thread, but it's a LOT for a non-primitive object
23:03:19tarcieriheh
23:03:51MenTaLguYI don't think Erlang process are cheap enough to do that without hurting performance badly
23:03:53tarcieriWell, the experiment may be a failure :)
23:04:10fbuilesv enters the room.
23:04:15tarcieriWe'll see
23:05:18evantarcieri: does BEAM compile bytecode into machine code?
23:05:26tarcierievan: no, but there's HiPE which does
23:05:39MenTaLguYincidentally, the pluggable scheduler thing permits schedulers like: http://github.com/mental/minicomic/tree/minicomic-ng/lib/minicomic/ui/event-scheduler.rb
23:05:41evanwhich do people use more?
23:05:49headius enters the room.
23:06:11tarcieriwell, using HiPE is just adding some flags to the compilation and startup
23:06:24tarcieriHiPE is effectively built into BEAM
23:06:25evanthey're packaged together?
23:06:28tarcieriyeah
23:06:52mapar leaves the room.
23:06:58tarcieriHiPE imposes some... interesting compile-time restrictions
23:07:10tarcierifor exampile, if you plan on using the SMP scheduler, you have to tell it that at compile time
23:07:43tarcieriI don't really get how HiPE works either, just that it somehow compiles to native code somewhere at some point in time
23:07:44evaninteresting.
23:07:50evanok
23:07:53evanjust curious.
23:08:06evani was looking at the DLR api over lunch
23:08:14evanit's funny how similar to the LLVM api it is
23:08:25tarcieriheh
23:08:31evanwith both, you basically just build an expression tree
23:08:37evanand it takes care of turning that into machine code.
23:09:18sambo82 leaves the room.
23:09:22rubuildius_amd64Ryan Davis: e8995e700; 2175 files, 7035 examples, 25543 expectations, 0 failures, 0 errors
23:09:34AndrewO leaves the room.
23:11:17samruby_ leaves the room.
23:11:30tarcierievan: everything I'm doing is completely detached from the low-level compilation :/
23:11:47atmos leaves the room.
23:12:00tarcieriI'm walking my AST and transforming it into Erlang AST
23:12:05tarcierifrom there, ???
23:12:17samruby leaves the room.
23:12:28zenspiderprofit?
23:12:33tarcieriheh, sure
23:13:07moofbong leaves the room.
23:13:34samruby__ leaves the room.
23:16:56samruby___ leaves the room.
23:18:13sambo82 enters the room.
23:19:35headius leaves the room.
23:20:34samruby____ leaves the room.
23:23:59samruby_____ leaves the room.
23:24:26rubuildius_ppcRyan Davis: e8995e700; 2175 files, 7168 examples, 25756 expectations, 0 failures, 0 errors
23:24:28rubuildius_ppc leaves the room.
23:24:59ShayArnett enters the room.
23:25:01rubuildius_ppc enters the room.
23:25:58sambo leaves the room.
23:27:48qwert666 leaves the room.
23:33:32sambo enters the room.
23:39:47chris2 leaves the room.
23:39:54fbuilesv_ enters the room.
23:40:45sambo82 leaves the room.
23:45:44enebo leaves the room.
23:47:35dysinger enters the room.
23:48:29sambo82 enters the room.
23:51:21evantarcieri: http://user.it.uu.se/~kostis/Papers/hipe-sttt.pdf
23:51:23evanthats pretty interesting.
23:54:01sambo leaves the room.
23:56:14fbuilesv leaves the room.