Index

Show enters and exits. Hide enters and exits.

00:00:01evanbingo.
00:00:17mernen enters the room.
00:00:40naeu enters the room.
00:00:41anonuserAnd another trivial problem is solved.
00:03:28anonuserevan, does that warrant a commit bit?
00:03:43evansure
00:03:46evanwhats your ssh key?
00:04:39naeu leaves the room.
00:06:10anonuserhttp://rafb.net/p/wwFpJc78.html
00:06:15Yurik leaves the room.
00:06:49evanadded!
00:07:15anonuserAwesome :)
00:08:00brixenanonuser: what's your github username?
00:08:11naeu enters the room.
00:08:43anonusersbryant
00:09:17brixenanonuser: cool, added to rubyspec
00:09:34anonuserThanks
00:09:46brixennp, thanks for helping :)
00:09:59brixenevan: git-lh is slick
00:10:05anonuserI plan on helping out a lot if possible :)
00:10:20brixenheh, there's lots to do
00:10:21anonuserThis keeps me motivated to keep working, especially considered how burnt out I've been
00:10:30naeu leaves the room.
00:10:53naeu enters the room.
00:11:03brixenanonuser: yeah, nothing like a little bit of interesting ruby code to refresh oneself
00:11:49blakewatters enters the room.
00:11:50anonuserbrixen, yeah. Especially when I came from writing a distributed video authoring system in C++ with turn a rounds in as little as 2 hours.
00:12:00anonuser100+ hour weeks are the worst things ever.
00:12:07brixenugh
00:12:51rubuildius_ppcSean Bryant: ec6bf683f; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
00:13:10anonuserIt was a neat project, tried to leverage opensource technologies even wrote a few libs but the environment was the hellish.
00:13:16anonuserAnd the pay was laughable.
00:13:25rubuildius_amd64Sean Bryant: ec6bf683f; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
00:13:25naeu leaves the room.
00:15:17anonuserrake git:committer does what I think it does?
00:15:36evanyep
00:16:59brixenanonuser: and rake spec:committer also
00:17:07anonuseryep saw that
00:17:12brixenevan: only thing git-lh doesn't seem to do is close the ticket for you ;)
00:17:48evanha!
00:17:52evani was just thinking about adding that
00:18:15brixenthis is just too cool, I might have to hadd git-rm for redmine
00:18:25brixenoh wait, bad choice :P
00:18:36brixengit-red
00:18:39anonuserredmine is really cool by the way.
00:19:05brixenyeah, seems to be working ok so far
00:19:15lopex leaves the room.
00:20:03anonuserI might use it as a replacement for trac.
00:20:13anonuserSince I've switched to git and well trac is abysmal.
00:20:33brixenahh, it's a no-brainer as a replacement for trac
00:20:48brixenI'm still not sure rubyspec wouldn't be better in LH, though
00:21:06NoKarma enters the room.
00:22:45anonuserwhat's with the github flakiness :(
00:23:21jbarnetteanonuser: it's being overwhelmed by the proliferation of forks of dr-nic-magic-awesome
00:23:22anonuserbrixen, redmine seems to have a lot of features you don't use
00:23:35anonuserjbarnette, oh. Well that clears that up
00:23:51brixenanonuser: any in particular you think we should use?
00:24:22anonuserLet me find a conversation I had with a friend.
00:24:37anonuserWe were looking for a project management thing.
00:24:40ch0wda leaves the room.
00:24:52anonuserbrixen, on the otherhand LH is something people are already used to.
00:24:56evanif it's useful for anyone, if you do
00:24:58anonuserEspecially the rubspec people
00:25:02evanhttp://rubini.us/t/<number>
00:25:08evanyou'll be redirected to the LH ticket of that number
00:25:20evani can never remember the LH urls, so I added that.
00:25:30brixenah nice
00:26:23anonuserLet's hope Rick Olsen doesn't go changing the URLs
00:26:28brixenanonuser: the thing I don't like about LH is having to have 2 sites. I want to just send people to rubyspec.org and everything is there
00:26:35brixenLH should allow transparent proxying
00:26:37atmos leaves the room.
00:26:38brixenor something
00:27:08anonuserWith enough proxying and mod_rewrite you can fake it?
00:27:58eventualbuddha enters the room.
00:28:56evanooh
00:29:02evancommand of the day "git whatchanged"
00:29:20anonuserthat's different from git diff?
00:29:25evancheck it out.
00:29:45anonuseram doing
00:29:53anonuseroh that is awesome
00:34:21jp_tixgit log --stat?
00:34:54brixenthis is useful:
00:34:56brixen[alias] lc = log ORIG_HEAD.. --stat --no-merges
00:43:05anteaya enters the room.
00:43:37imajes enters the room.
00:43:57stepheneb enters the room.
00:59:26anteaya leaves the room.
01:04:01eventualbuddha leaves the room.
01:06:35Fullmoon enters the room.
01:13:03yipstar leaves the room.
01:13:03edwardam leaves the room.
01:17:45felipemesquita enters the room.
01:19:06MenTaLguY enters the room.
01:21:23hemulen leaves the room.
01:22:34twbray leaves the room.
01:23:14jbarnette leaves the room.
01:27:53wycats leaves the room.
01:28:19ezmobius_ enters the room.
01:47:03blakewatters leaves the room.
01:47:04jennyw leaves the room.
01:54:47blakewatters enters the room.
01:56:23michalw leaves the room.
02:03:35obvio171 leaves the room.
02:04:41twbray enters the room.
02:05:09headius leaves the room.
02:05:11twbray leaves the room.
02:06:06jtoy enters the room.
02:10:02edwardam enters the room.
02:13:19wmorgan leaves the room.
02:13:39wmorgan enters the room.
02:15:59ezmobius_ leaves the room.
02:15:59ezmobius leaves the room.
02:17:03twbray enters the room.
02:17:54obvio171 enters the room.
02:17:58yugui enters the room.
02:21:43atmos enters the room.
02:23:49ShayArnett enters the room.
02:25:58iamclovin enters the room.
02:28:22twbray leaves the room.
02:28:31stepheneb leaves the room.
02:33:33VVSiz leaves the room.
02:34:45jennyw enters the room.
02:36:06pauldix enters the room.
02:37:14VVSiz enters the room.
02:37:25edwardam leaves the room.
02:45:17oliver_s enters the room.
02:46:47oliver_s leaves the room.
02:49:52boyscout1 commit by Evan Phoenix
02:49:53boyscout * Fix rbconfig and rubygems paths; c03b8e5
02:52:03loincloth leaves the room.
02:52:08obvio171 leaves the room.
02:55:01oliver_s enters the room.
02:57:38wycats enters the room.
02:59:40anteaya enters the room.
03:00:01cremes enters the room.
03:03:18rubuildius_amd64Evan Phoenix: c03b8e5b4; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
03:03:22rubuildius_ppcEvan Phoenix: c03b8e5b4; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
03:03:57oliver_s leaves the room.
03:10:36sambo82 enters the room.
03:16:09evanDefiler: yeah, one sucky thing about MRI bt's is the way they show if you're in a block
03:16:16evanit can be pretty confusing
03:16:32DefilerI'm at a point where I'm just trying to understand the 'correct' behavior. Heh.
03:16:40DefilerSo that I can make it match
03:16:53Defilerbut it seems like it just picks random files from $: and sorts them
03:18:20headius enters the room.
03:23:02evanso, the problem is that shotgun/rubinius can't be used in a #! with bash, right?
03:23:26DefilerI've been dreading it, but I guess it is time for real deal caller specs =(
03:24:10sambo82 leaves the room.
03:24:17sambo82 enters the room.
03:24:47pauldix leaves the room.
03:26:19anteaya leaves the room.
03:28:32anteaya enters the room.
03:29:20sambo82 leaves the room.
03:29:21sambo82_ enters the room.
03:29:56atmos leaves the room.
03:30:07Defilerhttp://pastie.org/209959
03:30:16DefilerCould someone run that on the latest rubinius and paste the output?
03:30:28DefilerI want to make sure I don't have any leftover changes affecting it.. feeling paranoid. Heh
03:31:01brixensure, sec..
03:31:02anonuserDefiler, sure
03:31:51yipstar enters the room.
03:31:55vertiginous enters the room.
03:32:07djwhittclear
03:32:20djwhittoops, sorry about that
03:32:24TheVoice leaves the room.
03:32:57anonuserDefiler, I'll let brixen handle it because require 'file2' is failing horribly for me and I know the problem just lazy right now :)
03:33:59brixensomething went horribly wrong with rbconfig. I have a toplevel "$(prefix)" dir now
03:34:06brixenanyway, back on track...
03:35:18iamclovin leaves the room.
03:35:48brixenDefiler: minus the fancy arrows: http://pastie.org/209961
03:36:03brixenDefiler: do you want me to pp that?
03:39:45DefilerNo, it's cool. Thanks.
03:40:04brixennp
03:40:08DefilerI have the output matching MRI now, at least the first frame
03:40:16anteaya leaves the room.
03:40:17Defilerbut spec_parser_spec still fails, so it must not be right yet
03:40:23Defilerurhg
03:40:39DefilerOh well. Tomorrow.
03:41:23brixenrbconfig is a treasure trove of goodness: http://pastie.org/209954
03:41:35brixenobserve --ruby= in that paste
03:44:01DefilerEye of terror
03:44:43anonuserhe
03:48:26twbray enters the room.
03:50:13brixenhah! just installed treetop in rbx, generated the example ArithmeticParser from the .treetop file, and parsed "1+1"
03:50:23brixenwhere's nathan? heh
03:50:45anonuserbrixen, that is awesome
03:50:57anonuseri <3 treetop
03:54:04evanbrixen: woooo!
03:54:08evantreetop in the HOUSE.
03:54:26jennyw leaves the room.
03:54:33binary42 enters the room.
03:54:51brixenthe example arithmetic tests run without error, but all the lambda calculus ones error
03:55:09brixenevan: where will the gem bin stubs be? they're still going to lib/bin ?
03:55:56binary42_ enters the room.
03:56:01evanthey go in lib/bin, yeah
03:56:03evanso you can do
03:56:06evanrbx treetop ....
03:56:14brixenah right
03:56:14binary42 leaves the room.
03:56:37sambo82_ leaves the room.
03:57:12brixenhmm, probably looking for <file>.rb though
03:57:20brixentreetop includes a tt script
03:57:37evanyeah
03:57:40evani just realized that.
03:57:57evani need to have gems put them somewhere special
03:58:01evanand have loader look there
04:00:57twbray leaves the room.
04:05:00anonuserhpricot fails to install because it really wants rbx in a particular place
04:06:09anonuserany idea of where I should start?
04:10:54binary42 enters the room.
04:11:00binary42_ leaves the room.
04:12:50twbray enters the room.
04:15:19twbray leaves the room.
04:25:58ezmobius_ enters the room.
04:32:53Yurik enters the room.
04:34:30yipstar leaves the room.
04:36:36mernen leaves the room.
04:43:05Guest18944 leaves the room.
04:49:23RyanTM leaves the room.
04:50:56atmos enters the room.
04:55:29michaellatta_ enters the room.
05:00:29headiusanyone around?
05:01:12michaellatta leaves the room.
05:02:37rueJust us dustbunnies
05:03:21anonuserI'm around, not sure helpful that is.
05:04:33headiuswhere would I find out how rubinius sizes the generations?
05:04:52anonuserlook at baker.rb and marksweep.rb
05:05:00anonuserI'm guessing
05:05:07headiusok
05:05:35anonusererr
05:05:39anonuseri'm sorry it's baker.c
05:06:08anonuserand marksweep.c
05:06:47anonuserlooking into the generational garbage collector?
05:07:24stepheneb enters the room.
05:07:45twbray enters the room.
05:09:16jtoy leaves the room.
05:14:51atduskgreg leaves the room.
05:18:26anonuserevan, are there any requirements for the instrumenting profiler? I'm looking at Zen Profile. It'll add requirements of ruby inline as the profiling crap is inlined C.
05:20:55fbuilesv leaves the room.
05:23:46benny leaves the room.
05:23:52headius leaves the room.
05:26:18atmos leaves the room.
05:31:05atmos enters the room.
05:44:24jtoy enters the room.
05:50:23headius enters the room.
05:53:26thehcdreamer enters the room.
05:54:44jtoy leaves the room.
05:59:34iamclovin enters the room.
06:02:18benburkert enters the room.
06:02:32jtoy enters the room.
06:04:11stepheneb leaves the room.
06:19:00massdoes a lil jig
06:19:32benburkert_ leaves the room.
06:20:27yipstar enters the room.
06:21:54headius leaves the room.
06:23:43tizianobis enters the room.
06:24:27cezarsa leaves the room.
06:28:06headius enters the room.
06:30:39qwert666 enters the room.
06:46:42edwardam enters the room.
06:49:14JonathanT enters the room.
06:49:43thehcdreamer leaves the room.
06:55:34atduskgreg enters the room.
06:55:40michaellatta_ leaves the room.
06:56:12michaellatta enters the room.
06:57:14JonathanT leaves the room.
07:02:14twbray leaves the room.
07:09:25sholden leaves the room.
07:09:31sholden enters the room.
07:11:57mitchellvriley enters the room.
07:26:24nemerle_afk leaves the room.
07:45:31tizianobis_ enters the room.
07:58:29tizianobis_ leaves the room.
07:59:36naeu enters the room.
08:02:16tizianobis leaves the room.
08:08:23Maledictus enters the room.
08:15:33Skip enters the room.
08:16:47thehcdreamer enters the room.
08:16:51krisps enters the room.
08:17:05yipstar leaves the room.
08:19:15twbray enters the room.
08:26:53TheVoice enters the room.
08:31:05boyscout1 commit by Evan Phoenix
08:31:06boyscout * Modify loader to look for #{$:.first}/bin/<command> directly; e8505a5
08:31:27ezmobius_ enters the room.
08:34:35dbussinkmorning
08:44:10rubuildius_ppcEvan Phoenix: e8505a517; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
08:44:45twbray leaves the room.
08:45:07JonathanT enters the room.
08:46:44Yurik_ enters the room.
08:48:20rubuildius_amd64Evan Phoenix: e8505a517; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
08:50:18naeu leaves the room.
08:51:39bitsweat leaves the room.
08:53:34krisps leaves the room.
08:54:09mutle enters the room.
08:55:00krisps enters the room.
08:56:04Yurik leaves the room.
08:56:28octopod enters the room.
09:06:09krisps leaves the room.
09:08:19naeu enters the room.
09:19:41atduskgr_ enters the room.
09:21:04atduskgreg leaves the room.
09:29:00gnufied enters the room.
09:47:26xhanjian leaves the room.
09:49:02xhanjian enters the room.
09:52:02ezmobius_ leaves the room.
09:52:03ezmobius leaves the room.
09:56:48atduskgr_ leaves the room.
10:02:38headius_ enters the room.
10:13:15headius leaves the room.
10:17:29krisps enters the room.
10:24:02headius leaves the room.
10:28:33krisps_ enters the room.
10:38:48michalw enters the room.
10:39:28krisps__ enters the room.
10:47:24xhanjian leaves the room.
10:50:23krisps___ enters the room.
11:04:04gnufied leaves the room.
11:04:04octopod leaves the room.
11:04:04Yurik_ leaves the room.
11:04:04Maledictus leaves the room.
11:04:04michaellatta leaves the room.
11:04:04edwardam leaves the room.
11:04:04binary42 leaves the room.
11:04:04cremes leaves the room.
11:04:04wycats leaves the room.
11:04:04yugui leaves the room.
11:04:04Fullmoon leaves the room.
11:04:04shame leaves the room.
11:04:04thoughtfiz leaves the room.
11:04:04nicksieger leaves the room.
11:04:04lstoll leaves the room.
11:04:04rubuildius_amd64 leaves the room.
11:04:04foysavas leaves the room.
11:04:04maharg leaves the room.
11:04:04timo leaves the room.
11:04:04pastie leaves the room.
11:04:04manveru leaves the room.
11:04:04goodney leaves the room.
11:04:04matthewd leaves the room.
11:04:33Maledictus enters the room.
11:06:12maharg enters the room.
11:06:29jtoy leaves the room.
11:07:06thoughtfiz enters the room.
11:09:14krisps__ leaves the room.
11:09:15michaellatta enters the room.
11:09:41wycats enters the room.
11:09:53krisps___ leaves the room.
11:10:02rubuildius_ppc leaves the room.
11:10:14lstoll enters the room.
11:10:43rubuildius_ppc enters the room.
11:10:53Fullmoon enters the room.
11:11:10binary42 enters the room.
11:12:17cremes enters the room.
11:12:24yugui enters the room.
11:13:02edwardam enters the room.
11:14:22octopod enters the room.
11:14:28nicksieger enters the room.
11:17:18shame enters the room.
11:20:58foysavas enters the room.
11:26:01gnufied enters the room.
11:30:28manveru enters the room.
11:30:41matthewd enters the room.
11:30:44manveru leaves the room.
11:31:44manveru enters the room.
12:06:46imajes leaves the room.
12:14:37imajes enters the room.
12:15:55imajes leaves the room.
12:31:42NoKarma enters the room.
12:35:07stepheneb enters the room.
12:38:20stepheneb leaves the room.
12:44:01yugui leaves the room.
12:49:21RyanTM enters the room.
12:57:29goodney enters the room.
13:03:27robertpostill enters the room.
13:18:08blakewatters leaves the room.
13:27:05iamclovin leaves the room.
13:27:57qwert666 leaves the room.
13:36:31fbuilesv enters the room.
13:44:20ShayArnett enters the room.
13:50:54mutle_ enters the room.
13:54:41robertpostill leaves the room.
13:55:31michalw leaves the room.
13:55:49michalw enters the room.
13:56:01benny enters the room.
14:02:13blakewatters enters the room.
14:12:54mutle leaves the room.
14:14:59yasuhito enters the room.
14:18:36Fullmoon leaves the room.
14:19:09stepheneb enters the room.
14:19:15imajes leaves the room.
14:20:25imajes enters the room.
14:20:48Fullmoon enters the room.
14:21:51kAworu leaves the room.
14:22:13kAworu enters the room.
14:22:52Fullmoon leaves the room.
14:23:39moofbong enters the room.
14:28:33pauldix enters the room.
14:33:51yugui enters the room.
14:41:51AndrewO enters the room.
14:43:08edwardam leaves the room.
14:44:54headius enters the room.
14:50:20headius leaves the room.
14:54:32headius enters the room.
14:57:15stepheneb leaves the room.
15:03:35stepheneb enters the room.
15:06:56enebo enters the room.
15:22:36anteaya enters the room.
15:25:46w1rele55 enters the room.
15:35:15mutle leaves the room.
15:35:49mutle enters the room.
15:42:41chris2 enters the room.
15:45:14vertiginous enters the room.
15:57:54thehcdreamer_ enters the room.
16:00:48thehcdreamer_ leaves the room.
16:04:44nexcastellandrinks coffee.
16:06:26thehcdreamer leaves the room.
16:14:32dschn enters the room.
16:16:22AndrewO leaves the room.
16:31:57tizianobis enters the room.
16:32:04edwardam enters the room.
16:32:38massthis place is off the hook
16:35:12AndrewO enters the room.
16:39:47benny leaves the room.
16:43:36michalw leaves the room.
16:48:01naeu leaves the room.
16:49:03shame leaves the room.
16:49:13cout leaves the room.
16:49:28anteaya leaves the room.
16:49:45cout enters the room.
16:49:56dbussinkafternoon everyone
16:50:58nexcastellanMorning!
16:51:41anteaya enters the room.
17:00:01twbray enters the room.
17:02:45gnufied leaves the room.
17:08:02tizianobis leaves the room.
17:08:50scudco enters the room.
17:09:28Yurik enters the room.
17:09:52mutle leaves the room.
17:14:59dbussinkhmm, not much happened during the day :P
17:17:00JonathanT leaves the room.
17:20:14anonuserMorning
17:22:12moofbong leaves the room.
17:22:37moofbong enters the room.
17:22:47ch0wda enters the room.
17:24:03cschneid_ leaves the room.
17:28:54evanmorning.
17:29:14vincentlumorning
17:29:18dbussinklooking at llvm these days?
17:29:33atduskgreg enters the room.
17:30:26yipstar enters the room.
17:32:27Yurik_ enters the room.
17:32:31benstiglitz enters the room.
17:33:13cschneid_ enters the room.
17:34:51twbray leaves the room.
17:35:13stepheneb leaves the room.
17:40:46anonuserevan, I'm looking into porting ZenProfile into shotgun.
17:41:11anonuserNot sure how difficult it'll be considering most of ZenProfile is inlined C
17:41:19evani should have put requirements in that document
17:41:24anonuseryeah
17:41:24evani didn't mean using ruby_inline
17:41:27evanand it can't use rb_funcall
17:41:32vincentluevan: currently rubinius doesn't do tail optimization, right?
17:41:33evanit shouldn't use subtend.
17:41:37evanvincentlu: no.
17:41:41Yurik- enters the room.
17:41:41evanit does not.
17:41:54evananonuser: when I said port, I mean copy the logic, not the code.
17:42:01vincentluyeah, I see context chain growing.
17:42:03anonuserevan, gotcha.
17:42:17evani'm going to tackle it today myself
17:42:22anonuserevan, I figured as much when I saw the code.
17:42:25evanafter I get the sampling profiler fixed.
17:42:25anonuserThe profiler?
17:42:29anonuserOh
17:42:47evanyou're welcome to take a look though too.
17:43:24anonuserwell after lunch maybe, I'll probably just move on to something else since you know what ya want.
17:44:28dbussinkevan: going to put it in the new vm?
17:45:20evandbussink: no, going to put it into shotgun so we can use it in 0.9
17:45:54dbussinkah, new vm is a bit on hold for now?
17:46:10evanwell, i'm working on it in parallel
17:46:37dbussinkis there a release target date set for 0.9?
17:46:43evanJune 17th
17:46:54evani'm trying to get us into doing shorter, fixed releases
17:47:20dbussinkah, put a bit more pressure on getting things done i guess?
17:47:29evanyeah, a little.
17:47:41dbussinkanything special about june 17th? or just a random date?
17:47:42evanwe're still going forward with the new VM though.
17:47:47evanit's before ruby kaigi
17:47:57evanbut otherwise, it's just a particular line in the sand.
17:48:45dbussinkit's the date of legal gay marriage in california i just found :P
17:48:54DefilerOK, so.. I think I understand the 'caller' problem now
17:49:02Defiler(finally)
17:49:11evandbussink: ha! just a coincidince
17:49:14anonuserdbussink, excellent
17:49:22DefilerIt is that 'caller's output doesn't always have any damn thing to do with who called the running method, at all
17:49:26shame enters the room.
17:49:35evanDefiler: yep.
17:50:10DefilerAre there any other circumstances other than eval "caller" where that is the case?
17:50:40evannot that I know of.
17:50:48DefilerTime to see what block.call does when "thingy do caller end" set up the block
17:50:48atduskgreg leaves the room.
17:50:53evanso, adding a BlockContext on the top didn't do it?
17:51:19DefilerI managed to get the correct output for the first line, but the rest of the trace is into totally different files
17:51:51Defiler'def calling_hierarchy' is doing too much work, I think
17:52:03Defileror maybe I am setting up the top context wrong.
17:52:03evanwhere is that?
17:52:14DefilerThat is what caller calls to do the owrk
17:52:19Defilerit is like fill_from, but for caller
17:52:31Defiler(context.rb)
17:52:39evanlooking now.
17:53:08Yurik-_ enters the room.
17:53:12evanhow can it be doing too much work?
17:53:15evanit does hardly anything
17:53:20yugui leaves the room.
17:53:44DefilerI am wary of the ctx.kind_of?(BlockContext) part
17:54:21DefilerI still don't know what the correct output should be here, though, so I have a lot of work left to do
17:54:42evanlet me go look at what MRI does.
17:55:27enebo leaves the room.
17:55:49fbuilesvdo you guys know if your presentation at railsconf will be on video any time soon?
17:56:21brixenfbuilesv: afaik, it wasn't recorded :(
17:56:22evanthey didn't record it
17:56:23evanso no.
17:56:32fbuilesvdang
17:57:19DefilerThis is my test case http://pastie.org/pastes/209959
17:57:35DefilerGemStone says they have video of their Maglev talk, but I guess they recorded their own
17:57:46evanthey did
17:57:52evanit was one of the guys we had dinner with
17:57:58evanrecording off to the side
17:58:09Defileroh, right. I remember seeing him with a camera
17:59:42DefilerI don't understand why MRI only prints a single stack frame with this test case
17:59:59evanbecause stuff like eval and such don't have frames
18:00:02evanin MRI
18:00:12evanthe only that that has a frame is ruby methods.
18:00:31DefilerOK, so the 'caller' is simply what line the block started on, and what got us to that line
18:01:03octopod leaves the room.
18:01:36evanpretty sure, yeah.
18:01:45evanesp since we're using the block as the binding
18:01:59evanand thus the only ruby code that has run up to that point is file1, calling spec
18:02:03evanon line 3
18:02:52evanDefiler: add "puts caller(1); p :break" to the first line of spec
18:02:55evanand compare them.
18:02:56DefilerThis is my current diff.. I know the body of 'caller' is wrong, but I was just messing around to come to my current conclusion..
18:02:59Defilerhttp://supremetyrant.com/ruby/caller_fumblings.diff
18:03:45DefilerOk, yeah, identical
18:03:52evanyou're caller code basically throws away what you created with prepare
18:04:02Defilerright
18:04:04evanok, now put
18:04:09evan"1" on the first line of spec
18:04:18evanactually no
18:04:21evanthat wont change it
18:04:24evanbecause we're doing caller(1)
18:04:33DefilerYeah, it needs to pass the block to something else to expose a difference
18:04:56jacen_ enters the room.
18:05:04evanbut essentially block.binding is like a continuation
18:05:10Yurik- leaves the room.
18:05:16evanit's asking for the moment in time that the block as created
18:05:21Yurik leaves the room.
18:05:46Defilerhttp://pastie.org/210256
18:05:55Defilerif file2 looks like that, it shows what I assume you were trying to get me to see
18:06:20headiuscaller just walks back up the frames
18:06:36headiusand the binding contains a reference to a frame
18:06:45Defileryes, but which frame
18:07:03headiusthe frame where it was created, or where the proc was created
18:07:06evanheadius: in the case of Proc#binding
18:07:14DefilerAt my current level of grokkage, that's like saying "caller just returns an array"
18:07:16evanthe frame in question isn't just the one that created the block
18:07:18Defilertrue, but not enough :)
18:07:21evanit's also the position the frame was in
18:07:24evanwhen the block was created.
18:07:54headiusproc clones the frame in which it is created when it is created
18:08:15evanthat makes sense
18:08:20evanwe're trying to avoid that
18:08:22evanbut maybe we can't.
18:08:26Defileraha see? I knew that was the easiest way :)
18:08:36evanDefiler: you were right
18:08:40Defilerstill, prepare should do the right thing, right?
18:08:45DefilerIt looks correct to me, at least
18:08:54evanDefiler: i'd think so.
18:09:01headiusyou can't
18:09:02Defilermaybe the blockenv isn't in the right state either, and it is copying the wrong juice
18:09:22headiusbecause the proc could be called across threads, line numbers will be changing all over, etc
18:09:37headiusproc additional clones when it's activated, at least in JRuby
18:09:49headiusnot sure if that's necessary or not
18:10:01evanalso when taking a binding
18:10:03headiusI still only have about an 80% understanding of all that
18:10:23headiuswe don't do it the way MRI does anymore, so we're mimicking behavior
18:10:28headiusdeath to ITER_*
18:11:46Defilerwhoa, what is def self.new all about in kernel/core/proc.rb?
18:11:57DefilerWhy does it do that current.sender.block thing?
18:12:10evanAH
18:12:15evanthats crazy town behavior
18:12:16evancheck out
18:12:20evandef blah
18:12:22evanp Proc.new
18:12:23evanend
18:12:25evanblah { 1 }
18:12:30Defilerwhaaaat
18:12:33evanexactly.
18:12:41evani'm pretty sure it's from before ruby had
18:12:46evandef blah(&block)
18:12:49evanbefore that syntax
18:12:51Defilernfw
18:12:55Defilernfw
18:12:56headiusyou didn't know about that?
18:12:57evanso you'd use Proc.new to get the current block.
18:13:07nicksieger leaves the room.
18:13:11evanDefiler: you can see my comment on that behavior in the code.
18:13:13DefilerHaha yeah how could I have failed to learn about something so common and basic
18:13:55nicksieger enters the room.
18:14:09evanDefiler: well, i'm going to fix the sampler now.
18:14:19evanif you want to go ahead and dup the MC when the Proc is created, thats fine.
18:14:25evanwe can always come up with a better way later.
18:15:07headiushopefully as we continue to fill out the specs we'll be able to figure out a better way for both of us
18:15:20headiusI know we're cloning frames more often than we need, but it's hard to know what we'd break
18:15:30evanyeah
18:15:41evanwe need specs for crap like caller
18:15:44evanwith all these cases
18:15:49evanDefiler: can you put these into specs?
18:15:51evanthese cases.
18:16:06evanyou're going to have to be careful how the spec works though
18:16:16evansince rubinius always has more content in caller than MRI
18:16:26evanyou'll have to just like check out the first entry or something
18:16:35Defilerevan: Yeah, going to.. just been trying to figure out wtf the behavior was
18:16:38evank
18:16:55DefilerI'm just going to get this working however I can, and we can optimize it later if possible
18:18:03DefilerI am thinking of writing Proc.new like this, though: http://pastie.org/210262
18:18:39Yurik_ leaves the room.
18:20:06evanok
18:25:08michaellatta leaves the room.
18:27:15atduskgreg enters the room.
18:31:12wmoxam enters the room.
18:38:31Defilerheadius: Does JRuby hide frames in 'caller' that MRI wouldn't show, or does it show everything?
18:39:18dfg59 enters the room.
18:40:04headiuswe only populate what MRI would have
18:40:20headiussince we don't expose the Java call stack to Ruby
18:40:26headius(at the moment)
18:42:29Fobax enters the room.
18:45:22boyscout1 commit by Christopher Thompson
18:45:23boyscout * Set host in Makefile, add rubyhdrdir to include path.; 7d9b9c0
18:52:42nexcastellanMy girlfriend thinks you should rename the project "Rubilicious".
18:52:52nexcastellan:)
18:52:57evanheh
18:53:17Defilerevan: OK, I am going to write specs for exact caller output and pound on it until they pass
18:53:31DefilerWe can do whatever we want in 'awesome_backtrace'
18:53:43Defilerbut caller seems to be used programmaticalllllly more than I feared
18:54:21evanok
18:54:36evanhow are we going to deal with the extra lines we generate?
18:55:41DefilerWe may be stuck with some, but I'll see what we can elide
18:55:52evanwait
18:55:56evanyou're going to try and prune them out?
18:56:34DefilerYes
18:56:43DefilerAt least I was leaning that way
18:56:47evanhow?
18:57:15DefilerWell, for one I will try to omit any of the block first-classery that MRI doesn't have, at least in the caller output
18:57:38evanno
18:57:39evanhow?
18:57:42evanhow are you going to detect them?
18:57:59DefilerI'll start by skipping all BlockContexts in caller_hierarchy
18:58:03Defilerand see what happens
18:58:09evanok
18:58:17evanmy guess is it will be pretty wrong
18:58:34DefilerYeah. I'll give it an hour either way, though
18:58:43evani don't get what thats going to solve.
18:58:54rubuildius_ppcChristopher Thompson: 7d9b9c01b; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
18:58:58evanother than making caller more confusing
18:59:00evanbut go for it.
18:59:15DefilerI suspect just making the first two frames the same as MRI is going to be enough
18:59:29dary enters the room.
19:00:14rubuildius_amd64 enters the room.
19:01:04headius_ enters the room.
19:03:58dbussinkhmm, looks like something related to FFI is broken on linux
19:04:06dbussinklibiconv isn't working
19:04:24nexcastellanlibiconv stopped working for me today.
19:04:35evanodd.
19:04:43DefilerI suspect it is specific to libiconv in some way
19:04:43nexcastellanThat's not for Rubinius, though.
19:04:44enebo enters the room.
19:04:58nexcastellanTry recompiling with --with-iconv=shared or some such.
19:05:27boyscout1 commit by Wilson Bilkovich
19:05:28boyscout * Revise Proc::new to better explain wacky edge cases; d78e7c3
19:05:32dbussinkit's a shared lib on linux already, so it should work
19:05:38dfg59 leaves the room.
19:06:04dfg59 enters the room.
19:07:19dbussinkhmm, nvm, it doesn't seem to be installed
19:09:59Defilerevan: the 'check_args' field on Proc seems unused. Is that likely?
19:10:11evani thought it was
19:10:31evanhm.
19:10:32DefilerI don't see that field being accessed in shotgun, and it is never used in kernel
19:10:35evani guess it's not anymore.
19:10:43evanwe must have refactored it out
19:11:07DefilerCI passes without it, so that seems likely
19:13:52dbussinkhmm, the iconv functions are in libc on my ubuntu system
19:15:59JonathanT enters the room.
19:17:22headius leaves the room.
19:18:22rubuildius_ppcWilson Bilkovich: d78e7c36a; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
19:19:16boyscout2 commits by Christopher Thompson
19:19:17boyscout * Merge branch 'master' of git@git.rubini.us:code; 708ecfb
19:19:18boyscout * Add data type defines.; e071f93
19:19:59evannexcastellan: please use 'rake git:pull'
19:20:06rubuildius_amd64Wilson Bilkovich: d78e7c36a; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
19:20:10evanso we don't see your merge commits
19:20:16nexcastellanSure thing.
19:20:17michaellatta enters the room.
19:21:00wmoxam leaves the room.
19:27:54dbussinkh
19:28:09dbussinkiconv works if i remove the set_ffilib call on the latest ubuntu
19:28:18evanhrm. ok.
19:28:26evanwe'll have to figure out how to deal with that.
19:30:33qwert666 enters the room.
19:31:50anteaya leaves the room.
19:32:03evanrad
19:32:11evanadded GC slices to the sampler output
19:32:18rubuildius_ppcChristopher Thompson: 708ecfb2b; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
19:33:12brixenevan: sweeet!
19:33:19rubuildius_amd64Christopher Thompson: 708ecfb2b; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
19:33:26evanhm.
19:33:30evanlet me paste these results
19:33:33evantheir... interesting.
19:34:30anteaya enters the room.
19:35:59nexcastellanOur extension builds with Rubinius. One down, six to go. :)
19:36:05evan:D
19:36:41brixennexcastellan: heh, awesome
19:37:32brixenI guess depending on the extension, we would get MRI level perf for free (assuming it doesn't call a lot of ruby code)
19:38:10evanhttp://pastie.org/210335
19:38:29evanthats running 'eval "1"' in a loop 10000 times
19:38:48evanbasically, the compiler uses a Hash to
19:38:54evanand thats where most of the time is spent
19:39:29dbussinkis it replaceable with a lookuptable?
19:39:31brixenwe should use LookupTable if those keys are symbols
19:39:41brixenand provide a LookupTable class to bridge for MRI
19:39:41evani'll look.
19:39:50brixenLookupTable = Hash
19:39:52brixenif MRI
19:40:03brixendbussink: heh
19:40:18dbussinkbrixen: something with great minds think alike ;)
19:40:43brixendbussink: yeah, I was going to ask where all the #hash_entry came from, but evan answered it before I could type :)
19:41:33anteaya leaves the room.
19:41:49anteaya enters the room.
19:48:13atduskgreg leaves the room.
19:48:30nexcastellanhpricot 0.6 now builds under rubinius.
19:49:09imajes enters the room.
19:49:55evanYAY!
19:49:57evanthats great news!
19:50:00evancan you run the specs?
19:50:18tarcieriheh, does it work?
19:50:21evani want a tshirt with cezar's face on it
19:50:31tarcierilast I tried it under YARV it had horrible, horrible issues with... crashing
19:50:33nexcastellanI'll get to that later today hopefully. json and rmagick aren't building, and I'm testing two others.
19:50:38evanWOW
19:50:40evanrmagic.
19:50:43evanthat would be....
19:50:46nexcastellantarcieri, it compiles. Surely we can ship it.
19:50:47evanintense.
19:50:48nexcastellan:)
19:50:51Defilerevan: When I do: eval "caller", shouldn't "MethodContext.current.sender" inside 'def caller' be the MethodContext that was created for the CompiledMethod generated by eval?
19:50:51tarcieriheh
19:51:01Defiler(wow that sentence was involved)
19:51:09evanDefiler: it's a BlockContext
19:51:11evanwhen you run an eval
19:51:16Defileryes sorry
19:51:20evanMethodContext.current is a BlockContext
19:51:22Defilerbut same deal
19:51:22benstiglitz leaves the room.
19:51:25nexcastellanmechanize-0.7.6 compiles.
19:51:31DefilerMethodContext.current is the one for 'def caller', right?
19:51:38evani mean
19:51:42Defilerand sender should be the BlockContext created for the eval?
19:51:43evaneval "p MethodContext.current"
19:51:48evanis a BlockContext
19:52:06dfg59 leaves the room.
19:52:24Defiler#<BlockContext:0x31 main#__script__ (eval):1>
19:52:31Defilerso, wrong line number, but at least that part is correct
19:52:50atduskgreg enters the room.
19:52:53DefilerI think we must not be copying enough out of the binding in 'def eval'
19:53:16DefilerBecause no matter how hard I mangle up the binding context, the output is the same
19:53:20evancopying?
19:53:25Defiler'using'
19:53:49evanactually.
19:53:50evanno.
19:53:57evanwell
19:54:06evanyou need to go look at the methods for BlockContext
19:54:19evanbecause lots of them delegate to home (which is env.home)
19:54:22DefilerOK, here's a question..
19:54:25evanperhaps they're delegating too much
19:54:27Defilerthat might clear it up for me
19:54:32evanand we should be using some info in BlockContext
19:54:50DefilerProc.__from_block__ is passed a BlockEnvironment
19:54:57evanyes
19:55:05Defilerwhat is the relationship between that BlockEnv and the BlockContext that I need to keep around to set up the binding?
19:55:15brixenwow, amazon.com really is down. someone mentioned but I didn't believe it
19:55:35evanDefiler: i don't get what you're asking.
19:55:44thehcdreamer enters the room.
19:55:47DefilerA binding wraps a BlockContext or MethodContext
19:55:50Defilernot a BlockEnvironment
19:55:52evanyes.
19:56:13evana BlockContext knows it's BlockEnvironment
19:56:18DefilerWhen headius says he is 'copying the stack frame when a Proc is created', how would he access said frame from a BlockEnvironment passed to __from_block__ ?
19:57:04DefilerI need to make that BlockContext be in the @context ivar of the Binding that is created
19:57:07evanblockenv.home_block = blockenv.home_block.dup
19:57:12evanblockenv.home = blockenv.home.dup
19:57:19evaninside __from_block__
19:57:32Defilerhrm. thought I had tried that first, because that was my guess
19:57:41krisps enters the room.
19:58:12evanagain
19:58:20evandon't confusing a BlockEnvironment with a BlockContext
19:58:27evaneach time #call is sent to a BE
19:58:35evana BC is created to execute it.
19:58:53krisps leaves the room.
19:59:04krisps enters the room.
19:59:33DefilerI don't think I am confusing them, no
19:59:47evanok.
19:59:59evanjust to be clear
20:00:07evanwhat are you trying to make work
20:00:16evanthe first line of caller when Proc#binding is used?
20:00:17DefilerOK, so that doesn't fix it.. so I think eval isn't doing what it needs to to set up 'MethodContext.current.sender' properly for 'caller'
20:00:23DefilerYes
20:00:29Defiler(that is what I am trying to make work)
20:00:33evanyou can't setup .sender
20:00:36evandoesn't work like that
20:00:38evanyou can not change it.
20:01:34evanso there is nothing eval can do about it.
20:01:35DefilerDo I need a conditional in 'def caller' then, to choose a different starting frame if sender is a BlockContext?
20:01:38evannor should it.
20:01:53evanwhy?
20:01:55evanwhy would that help?
20:02:01chris2 leaves the room.
20:02:20evanthat would mean that
20:02:22evan1.times do
20:02:25evan puts caller(0)
20:02:25evanend
20:02:29evanworks differently than
20:02:37evandef blah
20:02:40evan puts caller(0)
20:02:40evanend
20:03:09DefilerThey already do
20:03:10Defilerin MRI
20:04:09evanperhaps, but i don't see how that helps with caller-start-on-proc#binding
20:04:25DefilerWell, here's how I see it..
20:04:38Defiler'caller' starts at the sender, and walks up the senders of those contexts
20:04:57DefilerIs that same technique going to work for eval "caller", some_proc.binding ?
20:05:32Defiler'sender' is some BlockContext generated by eval, and its sender is def eval itself
20:05:41Defilerbut def eval should not be anywhere in the caller output for this scenario
20:05:54evanwrong.
20:06:07evaneval "caller", proc.binding
20:06:12evanstarts at the context inside the binding
20:06:27evandoesn't it?
20:06:29evanoh wait.
20:06:32evani see, i see.
20:06:42evancaller always sees the eval created binding
20:06:49evanbecause thats how eval works no matter what
20:06:55Defilerright
20:07:06evanyou could use the eval detection trick
20:07:19evankind_of(BlockContext) and ctx.from_eval?
20:07:28Defilerctx.env.from_eval? but yeah
20:07:34evanyeah
20:07:42Defilerand that code would be in 'def caller', yes?
20:07:48evanyeah
20:08:07anonusernexcastellan, what did ya do to get hpricot to build?
20:08:41nexcastellananonuser: Nothing more than to get our internal extension to build. What that was was adding rb_cArray, rb_cString, etc. to ruby.h and the corresponding change to ruby.c.
20:09:02nexcastellanThat's commit e071f93f9700344c803ed7fc59585790ec150c76.
20:09:49Defilerhrm.. how do I get at the context I need in that scenario?
20:09:54DefilerI expected ctx.env.home_block to contain it
20:10:03anonuseroh, just added a bunch of subtend primitive support?
20:10:14evanDefiler: that seems right.
20:10:23evananonuser: not primitives.
20:10:34evananonuser: just fleshing out missing C stuff in subtend.
20:10:42anonuserevan, err. okay
20:11:32nexcastellanI'm hoping most of the rest of my fixes are going to be missing C stuff, but we'll see. I'm working on rmagick now.
20:11:40boyscout2 commits by Evan Phoenix
20:11:41boyscout * Whitespace cleanup; 555fbcf
20:11:42boyscout * Fix sampler to use pthread instead of setitimer; aeb8008
20:11:43elight enters the room.
20:12:51tarcieritry this file with Hpricot and see if it causes a crash:
20:12:53tarcierihttp://code.whytheluckystiff.net/hpricot/attachment/ticket/144/crash.xml
20:13:27evanzoinks
20:13:30evanthats quite the file
20:13:35evancan't we start smaller?
20:13:38evanmaybe www.google.com
20:13:39evan?
20:13:41evan:)
20:14:40anonusertarcieri, is it supposed to crash?
20:14:47anonuserIs it even valid markup?
20:15:07nexcastellanI encourage people to try out Hpricot (and mechanize) and to report any failures in Rubinius that are not in Ruby to the appropriate location. :) Most likely, they'll just blow up in all cases but hey, if things work for you, I'd love to hear that, too. :)
20:15:27DefilerHrm.. home_block and home take me too far, but ctx.sender.sender is just right for this test case
20:15:31DefilerInteresting
20:15:54Defilerhome_block takes me all the way back to "kernel/core/compiled_method.rb:379:in `as_script'"
20:15:56evanthat makes sense
20:16:01evanhome_block is where it's created from.
20:16:05evansender is where it was called from.
20:16:40DefilerThe block was created in file1.rb though, not in compiled_method.rb
20:16:59evanhrm
20:17:56anonusernexcastellan, I'm unable to build hpricot. Though I've not looked at why
20:18:38nexcastellananonuser: Did you get the latest commits?
20:18:49anonuseryep
20:18:59anonuserI pulled them in ran rake
20:19:11anonusererr let me try something
20:19:21nexcastellanHhhmm. What error are you getting? I'm wondering if it is bitching about headers or something.
20:19:27anonuserprobably
20:20:07anonuserrebuilding everything just to make sure.
20:20:37nexcastellanI'm doing my tests from a 'PREFIX=/foo rake build' 'PREFIX=/foo rake install' setup. Oh, and I also provide a symbolic link from install/bin/rbx to install/bin/ruby (where install is where you installed things to)
20:20:53nexcastellanThat's something that I didn't commit, of course.
20:21:10anonusernexcastellan, yeah.
20:21:13thehcdreamer leaves the room.
20:21:19anonuserhpricot for some reason really wants rbx to be there.
20:22:08anonusercc -dynamic -bundle -undefined suppress -flat_namespace -o hpricot_scan.bundle hpricot_scan.o -L. -L/usr/local/lib/rubinius/lib
20:22:08anonusercc: suppress: No such file or directory
20:22:08anonusercc: unrecognized option '-bundle'
20:22:26brixennexcastellan: are you on OS X?
20:22:30anonuserAre you developing for OS X?
20:22:41nexcastellanI'm on OS X, using gcc 4.0.1.
20:22:47brixennexcastellan: I noticed trying to build sqlite3-ruby gem, the rbx lib is not right
20:22:56nexcastellananonuser, can you do: which cc and cc --version?
20:22:58brixennexcastellan: it's expecting librubinius-0.8.0.static
20:23:09anonusersbryant@baloo:~/src/rubinius$ cc --version
20:23:09anonusercc (GCC) 4.2.3 (Ubuntu 4.2.3-2ubuntu7)
20:23:15nexcastellanbrixen, same problem with rmagick, any guesses why?
20:23:29nexcastellananonuser, that's weird. It shouldn't be sticking -bundle in there for you, then.
20:23:30brixennexcastellan: I think we need to tweak our rbconfig
20:23:43anonusernexcastellan, I'm guessing it's a problem with config.
20:23:51anonuseroi.
20:23:54brixennexcastellan: since we're not parsing autoconf output, we're going to have to put conditional as appropriate in configure.rake
20:24:30rubuildius_ppcEvan Phoenix: 555fbcfe3; 2189 files, 7320 examples, 26112 expectations, 0 failures, 0 errors
20:25:15nexcastellananonuser: The fixes look to be all in configure.rake. Easiest is for you to run out and buy a mac, please. :)
20:25:29anonusernexcastellan, I have a mac :)
20:25:38anonuserIt's an old powerbook g4
20:25:48anonuserIt just died, and I develop mainly on my desktop
20:25:50TheVoice enters the room.
20:26:03anonuserWhich won't run anything I want it to. So I'm stuck on ubuntu.
20:26:36nexcastellananonuser: A quick hack would be for you to edit configure.rake and change anything that looks suspiciously OSX'ish to something that would work on Ubuntu.
20:27:20nexcastellananonuser: Good time to buy a recent MacBook Pro. :)
20:27:38anonusernexcastellan, that's no excuse :(
20:27:47anonuseryeah it's in configure.rake alright.
20:27:50dbussinkevan: just tested the profiler with the tree script from http://antoniocangiano.com/2008/06/05/maglev-handles-trees-like-a-monkey/
20:27:53anonuserIt just flat out assumes -bundle works.
20:28:04evandbussink: hows it look?
20:28:06nexcastellananonuser: Yeap, should be an easy fix.
20:28:27dbussinkevan: http://pastie.org/210380
20:28:39anonusernexcastellan, yeah it is.
20:28:40anonuserlet me test
20:28:59evanzoinks!
20:29:07evani guess the GC cycle count is throwing it off.
20:29:08evan:)
20:29:13dysinger leaves the room.
20:29:31nexcastellananonuser, you wanna work on it? Should be easy to check the PLATFORM variable and add some conditionals in there. If you know how to build shared libs, etc. on Linux, at least.
20:29:42nexcastellanPLATFORM =~ /darwin/ would be a test for OS X.
20:29:53dbussinkevan: it's still a tad faster than mri :P
20:29:56anonuseryeah
20:30:04headiusstupid maglev numbers
20:30:05evandbussink: with the profiler on it's faster?
20:30:21evandbussink: there is a bad bug with the sampler right now with large data sets
20:30:22dbussinkevan: without it is, let me try with
20:30:24evani need to fix it.
20:30:24headiusnotice in the comments antonio posts fannkuch and it's not even 2x faster
20:30:38anonusernexcastellan, I'll see about getting it to work on linux
20:30:39rubuildius_amd64Evan Phoenix: 555fbcfe3; 2189 files, 7261 examples, 25964 expectations, 0 failures, 0 errors
20:30:46headiusgross amount of hype going on
20:30:57dbussinkheadius: ah, i see
20:31:01dbussinkheadius: probably is yeah
20:31:31nexcastellananonuser, probably not much I can do to help but yell if you want me to try.
20:31:49dbussinkheadius: he even left java out :P
20:31:52anonuserI'll play aroudn with it.
20:32:05mahargheadius: just ignore maglev until they actually release some version of their vm
20:32:21anonuserheadius, I wouldn't pay too much attention until they run all of ruby.
20:32:50headiusyes, the initial cycle should die down soon enough
20:32:51mahargit's pretty clear there are a lot of things they don't get about the ruby community, and helping to throw massive attention their way at this stage isn't going to help them 'get' those things
20:33:13headiuswe'll talk when they've implemented openssl
20:35:24dysinger enters the room.
20:35:28dbussinkevan: looks like it's pretty broken indeed, it a lot slower, like 5 times slower
20:35:46evanyeah
20:35:50evanit thrashes the GC
20:36:01evanlet me see if I can fix that now.
20:38:47twbray enters the room.
20:40:53bitsweat enters the room.
20:44:27macournoyer enters the room.
20:46:27headius leaves the room.
20:46:49headius enters the room.
20:49:28wycats leaves the room.
20:51:05nexcastellanI'd like some feedback on this small patch to configure.rake: http://rafb.net/p/iYODyg43.html
20:51:27headius leaves the room.
20:51:38nexcastellanWe don't build a static lib for rubinius afaik, so I changed some things. It won't work except on OS X, of course, but anonuser is looking into that already.
20:51:55evanwe shouldn't use .dylib or .so
20:52:00evanuse Rubinius::LIBSUFFIX
20:52:24kAworuplop
20:52:36evanfine otherwise
20:53:47brixenkAworu: splash
20:54:30anonuserokay I've got changes in configure.rake to get the gem built and installed!
20:54:42nexcastellanWay to go, anonuser. :)
20:54:49anonuserLet's see if it runs
20:55:08macournoyer leaves the room.
20:55:20anonuserbut basically configure.rake makes a lot of assumptions and assumes a darwin linker.
20:55:31evan