Index

Show enters and exits. Hide enters and exits.

00:03:52agile leaves the room.
00:07:29KirinDav leaves the room.
00:07:34anonuserPeach, time to leave work. 11 hours is enough.
00:07:59KirinDav enters the room.
00:08:46rueNaww
00:11:20wmoxam leaves the room.
00:14:49lstoll enters the room.
00:22:09binary42 leaves the room.
00:26:31lvmc leaves the room.
00:26:52kofno leaves the room.
00:31:05agardinerwoohoo! i just got step working!!!
00:40:33evanyay!
00:40:47agardinerwhat a bitch it was, too...
00:41:23brixenagardiner: awesome!
00:41:31agardinermy first commit on my local branch was the 31st of last month!
00:42:08agardinernow lets see if it fixes your problem...
00:42:45agardineryay!!! it works!
00:43:49brixensweet!
00:44:00graza leaves the room.
00:44:11brixenagardiner: at this rate, I'll have to buy you a small micro-brewery mate :D
00:44:41graza enters the room.
00:47:26agardinerhehe... that might have helped!
00:47:37macournoyer enters the room.
00:47:54agardinerman, it feels good to get this working
00:48:08benburkert leaves the room.
00:48:52jeremydurham enters the room.
00:50:08therealadam leaves the room.
00:55:51rueagardiner: Nicely done
00:55:59agardinerthanks
00:57:22macournoyer leaves the room.
00:58:08macournoyer enters the room.
00:58:09evani hate how long it takes UPS to deliver to my place
00:58:17evanmy RAM went out at 7am for delivery
00:58:29evanstill not here.
00:59:41yipstar leaves the room.
01:04:07claudio_ leaves the room.
01:12:17rueOurs varies, sometimes it is before 9:30 when decent people get up, usually sometime after 6
01:12:44tbmcmullen leaves the room.
01:14:14boyscout2 commits by Brian Ford
01:14:15boyscout * Added reporting to mspec-tag.; 58818d2
01:14:16boyscout * Fixed YAML spec to pass MRI. Added fails tag for rbx.; fd0ff43
01:14:52antaresevan, hi. a question here: is there anything beside Henry Baker's book that describes algorithm used by shotgun gc? I see cheney's algorithm and other copying GCs algorithms are close but probably not exactly. Trying to find some information before the book ships.
01:17:52evanmmm
01:17:55evanwell
01:18:14evanthe young generation is the cheney algorithm nearly exactly
01:18:22evanthe old is a simple mark/sweep
01:21:17lstoll leaves the room.
01:21:30antaresevan, could I use references to Cheney's algorithm in shotgun doc with a note?
01:21:49evanuse references how?
01:22:02evanput comments in the source to Cheney's algorithm?
01:22:19antaresevan, for instance like "see ... (link to say Wikipedia)"
01:22:23lstoll enters the room.
01:22:33antaresevan, or briefly describe what's going on
01:22:47evanwhere at though? in the source?
01:22:55antaresyup, in the source
01:23:11evansure
01:23:13evango for it.
01:23:17rubuildius_ppcBrian Ford: 58818d280; 1573 files, 5472 examples, 18929 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155133
01:24:04technomancy leaves the room.
01:25:09rubuildiusBrian Ford: 58818d280; 1573 files, 5472 examples, 18958 expectations, 0 failures, 5 errors; http://rafb.net/p/0zQGuV91.html
01:26:22drbraincan someone please, *please*, PLEASE fix rubuildius' failures
01:26:28drbrainit's been two days now?
01:26:47evanagreed
01:26:49evanwho is on linux?
01:27:51drbrainand who's on PPC darwin?
01:28:08evani can fire up my powerbook
01:28:12drbrainor, can we just go back in the history and figure out what caused the failures
01:28:34brixenthe cause of the failures is specs that are now being run
01:28:42brixenwe can tag them fails if you want
01:29:15drbrainbrixen: hasn't the getsockopt spec always been run?
01:29:20brixenno
01:29:26brixenall the library specs are now being run
01:29:55drbrainwhy aren't all the specs being run?
01:30:01drbrainor, are they now?
01:30:07brixenyeah, they are now
01:30:15brixenthey weren't because they were failing before
01:30:19brixenthey run fine on os x
01:30:29brixenwe need someone to work on linux and ppc
01:30:37brixenand the ppc failures may implicate ffi
01:30:54drbrainPPC failures implicate endianness
01:31:05brixennot the File.truncate ones
01:31:09djwhittamd64 also craps out right now too
01:31:14djwhittwith an Illegal Instruction
01:31:25brixenyep, intel 64 too
01:31:33evan*sigh*
01:31:33evangreat
01:31:36evanhow long has that been true?
01:31:49brixenjust noticed it on ey slice yesterday
01:31:53evanif people are going to offer to run the ci bots
01:32:13evanthen they typically are the ones that need to submit fixes when stuff is broken
01:32:17evanfor their bot
01:32:17djwhittI'm working on a merb based ci server right now
01:32:19defunkt enters the room.
01:32:34evanwell
01:32:40djwhittah, I see, yeah, won't really help with that end of it
01:32:42evani'm going to ignore it until someone shows me breakages
01:32:49drbraindjwhitt: we've already got the tinderbox architecture
01:33:15djwhittdrbrain: I'm not familiar with it
01:33:55drbrainhttp://firebrigade.seattlerb.org/
01:33:56brixenI can work on ubuntu linux, but I need help setting up my default route so NAT works on my parallels
01:34:19djwhittah, neat
01:35:03djwhittdrbrain: what would be involved in getting that setup for rubinius builds?
01:35:42drbraindjwhitt: there's a tinderbox gem that provides the runner, but is gem specific
01:35:48drbrainit sits atop a firebrigade-api
01:35:51drbraingem
01:35:53evanyay! the oreilly book with by blurb finally arrived!
01:36:09evanand strangly, mine is the only blurb....
01:36:32drbraindjwhitt: it may require some backend improvements
01:36:42agardinerooh! which book is that?
01:36:52evanThe Ruby Programming Language
01:36:58evanby David Flanagan and matz
01:37:04brixenahh, got that last week
01:37:12agardinerwow! impressive
01:37:13brixenI was impressed to see stuff about rubinius in it
01:37:15djwhittdrbrain: cool, I'll look into that then
01:37:25djwhittdrbrain: no sense reinventing the wheel
01:37:26evanbrixen: did ya notice me on the back?
01:37:32brixenheh yeah
01:37:32jeremydurhamevan: you have a blurb? I got the book a few days ago; not bad so far.
01:37:38wycats leaves the room.
01:37:45evanjeremydurham: yeah, on the back
01:37:49evani didn't realize it was in stores
01:38:15jeremydurhamI got it last tuesday, as a free gift for speaking. Total surprise.
01:38:33drbraindjwhitt: I have Plans for also adding an ec2 backend to it
01:38:42drbraindjwhitt: so it can fire up and test things at will
01:39:06jeremydurhamspelunking.. very well put :)
01:39:11dysinger enters the room.
01:39:16brixenit's got _why art work, too. pretty cool
01:39:28evanyay! my ram arrived too!
01:39:54drbrainevan: what's the difference between archive_get_file and archive_get_file2
01:39:57drbrain?
01:40:02evanumm
01:40:18evanoh
01:40:26evanget_file2 takes an archive handle
01:40:32evanoh the already open zip file
01:40:35drbrainok
01:40:44evanget_file() reopens the the zip file and closes it before it returns
01:40:50evanget_file2 is an optimization
01:41:06evanok, time to upgrade the ram
01:41:07evanbrb.
01:42:54twbray leaves the room.
01:43:09rueI see we have to update the coding style guidelines for this new appending-a-number schema :)
01:45:24jtoy enters the room.
01:52:23djwhittdrbrain: where is the repository for tinderbox?
01:52:39drbraindjwhitt: zenspider's p4 repo
01:52:54djwhittp4? as in perforce?
01:53:01drbrainyes
01:53:35drbrainhttp://www.zenspider.com/ZSS/Process/Perforce.html
01:57:18boyscout3 commits by Adam Gardiner
01:57:19boyscout * Hook-up new StepBreakpoint to new debugger step commands; 7698ec3
01:57:20boyscout * Initial implementation of StepBreakpoint; f192d65
01:57:21boyscout * Prepare to move step logic to Breakpoint; 5af7a41
01:57:49agardinerthere ya go...
01:58:17agardinerall new step/next logic!
01:58:30lstoll leaves the room.
01:58:32brixenwoot!
01:59:08antares leaves the room.
01:59:14agardinerstill needs a bunch of work, but there is a big improvement in the step logic
01:59:20evanok, back
01:59:22evanupgraded to 4G
01:59:24chris2 leaves the room.
02:00:34agardinerevan: that last commit also includes a nice little command to view send site info in the current method
02:01:22agardinervss shows all send sites, the current cached receiver/method/module, plus hits and misses
02:01:28evanhow do I run it?
02:01:42brixenoh, debugger command? sweet
02:01:43agardinerneed to get into the debugger, then just type vss
02:02:32agardinerbrixen: i might look at hooking the debugger up to a -d flag or something on shotgun/rubinius next
02:02:43nicksieger leaves the room.
02:02:54brixenagardiner: oh, did you solve the problem of inserting the bp right after the method starts?
02:03:02agardineryep
02:03:12agardinerthats needed for stepping in
02:03:19agardinerand its all now in place
02:03:19brixenawesome!
02:03:34agardinerthe step logic is much smarter now all round
02:03:50agardinerit handles stepping in, but also branches within the bytecode of a method
02:04:05brixeninsanely cool
02:04:10brixenmany thanks :)
02:04:11agardinerso before, if you said step 5 instructions, it just set a bp at ip+5
02:04:39evanso, whats the easiest way to break into the debugger now?
02:04:46agardinernow it actually looks at each instruction between ip and ip+5, and sets mini-breakpoints at each branch
02:04:47boyscout1 commit by Brian Ford
02:04:48boyscout * Use File.delete in YAML specs instead of rm.; f396bd7
02:05:00brixenevan: you can use the --spec-debug command :)
02:05:11agardinerprobably still just easiest to add the line "breakpoint" to a piece of code
02:05:28brixenevan: e.g. bin/mspec spec/ruby/1.8/core/array/append_spec.rb --spec-debug -S correctly
02:05:55brixenagardiner: should that use Kernel.debugger or breakpoint?
02:06:09agardinersame thing - one is aliased to the other
02:06:13brixenok
02:06:22brixenevan: http://pastie.org/154654
02:06:39brixenbut agardiner fixed the fault
02:07:38brixenthe ipaddr specs seem to have terrible performance under linux
02:07:45benny_ enters the room.
02:07:45benny leaves the room.
02:08:01agardineroh yeah, i keep thinking it has hung!
02:08:19brixenyeah, not sure what to do about it
02:08:34brixenit barely stops on os x, but will sit there for like 10 sec on ubuntu
02:09:06agardineryeah, bin/mspec ci took 191s for me
02:09:27brixenshould be ~30
02:09:29brixenthat sucks
02:09:43agardineri reckon ~160s was in the library/socket specs
02:09:47brixenyeah, just ran it, 203sec
02:10:11brixenI'm going to run with -fs to see which are the offenders
02:10:38agardinerok, i'm gonna grab a bite... bbiab
02:11:22binary42 enters the room.
02:11:25evanagardiner: the debugger help is a bit wide.
02:12:44lopex leaves the room.
02:15:54rubuildius_ppcBrian Ford: f396bd718; 1573 files, 5472 examples, 18929 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155152
02:15:55rubuildius_ppcAdam Gardiner: 7698ec385; 1573 files, 5472 examples, 18929 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155149
02:16:15rubuildiusBrian Ford: f396bd718; 1573 files, 5472 examples, 18958 expectations, 0 failures, 2 errors; http://rafb.net/p/WKfBAx91.html
02:16:16rubuildiusAdam Gardiner: 7698ec385; 1573 files, 5472 examples, 18958 expectations, 0 failures, 5 errors; http://rafb.net/p/dZxfDT88.html
02:16:25brixenthere's definitely an endian issue on ppc: http://pastie.org/155151
02:16:55brixenbut the File.truncate issue according to dbussink is FFI related
02:17:09brixensaid that the function was being called with arg value 0x0 iirc
02:19:32evanagardiner: you around?
02:21:21ezmobius leaves the room.
02:22:25mimetnet enters the room.
02:26:32djwhittdrbrain: running that last step require zenspider giving access correct?
02:26:51drbrainwell, one of them does
02:27:52djwhittwell, I guess I just wasn't clear if I could even do "pull" (or whatever that's called in p4 speak) without authorization
02:28:30crafterm enters the room.
02:28:33agardinerevan: just got back...
02:29:34agardinerre the debugger help being too wide, there are plans for that...
02:29:35graza leaves the room.
02:29:59graza enters the room.
02:30:00agardinerbasically, the Output class used for all debugger output will handle wrapping at the right point
02:30:38agardinerand will also gain support for column headings as well
02:31:49evanagardiner: why do you have code that does "self.<< arg"
02:31:53evanand not just "self << arg"
02:32:15agardinerwhere's that?
02:32:38evanset_columns in debugger/command.rb
02:32:43evanand the ones below it
02:33:15agardiner(just on the phone)
02:34:04boyscout1 commit by Brian Ford
02:34:05boyscout * Account for variable length array in Socket#getaddrinfo.; 69576ed
02:34:31brixendamn, that should say *spec
02:34:47brixenanyway, that's 4 of 5 rubuildius errors down
02:34:58brixenor should be
02:35:13seydar enters the room.
02:35:23hornbeck leaves the room.
02:38:00kofno enters the room.
02:41:23brixenanyone know of a standard, non uname, method of determining the endian-ness of a platform?
02:43:15zenspiderbrixen: I've got something... sec
02:43:23zenspider def self.big_endian?
02:43:26kamal_farizanybody got ./mspec/runner/mspec.rb:58: warning: instance variable @before not initialized when running ./bin/mspec ci against MRI?
02:43:30zenspider [1].pack('L')[-1] == 1
02:43:33zenspider end
02:43:40zenspiderdepends on pack('L'), but that's it
02:43:44agardinerevan: ok, back
02:43:55brixenzenspider: that's the problem, our pack specs are wrong
02:44:01zenspiderbrixen: http://rafb.net/p/HRpf0u17.html attribution goes to berger
02:44:02brixenand need to be guarded for endian-ness
02:44:08kamal_farizhasn't been keeping up with the project .. suddenly there are 400 new specs :)
02:44:12agardinerno real reason, feel free to change it if you want
02:44:25rueHoly shit.
02:44:27brixenkamal_fariz: what platform?
02:44:35brixenkamal_fariz: and can you show me the whole command?
02:44:42rubuildius_ppcBrian Ford: 69576ede3; 1573 files, 5472 examples, 18929 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155161
02:44:44kamal_farizbrixen: intel leopard
02:44:46rubuildiusBrian Ford: 69576ede3; 1573 files, 5472 examples, 18958 expectations, 0 failures, 2 errors; http://rafb.net/p/po9nbI20.html
02:44:54kamal_farizbrixen: ./bin/mspec ci -tr
02:45:22rueSorry about politics but I am hopping mad: O'Reilly later stated: "I don't want to go on a lynching party against Michelle Obama unless there's evidence, hard facts, that say this is how the woman really feels. ..."
02:45:40brixenkamal_fariz: yep, got it
02:45:41evanrue: url?
02:45:58ruehttp://www.huffingtonpost.com/2008/02/20/bill-oreilly-i-dont-w_n_87616.html
02:46:07agardinerooh, that's offensive
02:46:27gnufied leaves the room.
02:46:27rueOh boy is it ever.
02:46:37rue"Unless" even.
02:46:37VVSiz_ enters the room.
02:47:02agardinergeez, unbelievable
02:47:17brixenkamal_fariz: don't use bin/mspec ci with ruby
02:47:27brixenI'm going to remove that target
02:47:36brixenmspec ci runs more than just spec/ruby
02:47:41drbrainhaha, LICENSE is corrupt
02:47:59drbrain(minorly)
02:48:23brixenkamal_fariz: bin/mspec ci is the CI runner, we're not doing development on mri
02:48:39lstoll enters the room.
02:48:45kamal_farizbrixen: ok cool
02:49:11kamal_farizVVSiz: how is jruby running the specs
02:49:34agardinerso, it looks like -d as an option is already used...
02:49:46brixenkamal_fariz: they're only running spec/ruby
02:51:22kamal_farizbrixen: ah ok, so for the time being, while the ruby target is still in, i can do ./bin/mspec ci -tr spec/ruby
02:51:26_mutle enters the room.
02:51:50agardinershould i change the existing option to something else, or pick another letter for starting shotgun in the debugger?
02:52:29kamal_farizbrb
02:54:04seydar leaves the room.
02:58:30KirinDav leaves the room.
02:59:38boyscout1 commit by Brian Ford
02:59:39boyscout * Redo expectation in Socket#getaddrinfo spec.; 3d39fb3
03:02:03seydar enters the room.
03:03:44VVSiz leaves the room.
03:03:51nicksieger enters the room.
03:04:39perplexes leaves the room.
03:05:11ezmobius enters the room.
03:05:33rueWeeble
03:05:39mutle leaves the room.
03:06:23agile enters the room.
03:08:13rubuildius_ppcBrian Ford: 3d39fb35d; 1573 files, 5472 examples, 18934 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155165
03:10:03rubuildiusBrian Ford: 3d39fb35d; 1573 files, 5472 examples, 18959 expectations, 0 failures, 1 error; http://rafb.net/p/bGs6k869.html
03:13:25enebo leaves the room.
03:17:03mimetnethi all - i was wondering if rubinius was at a point were someone could begin looking at embedding it, and if so, where I might look?
03:20:38brixenmimetnet: you could take a look at shotgun/main.c
03:23:17mimetneti started too, as i'm currently going through the environment vs. machine code to actually understand what each is for. No real docs at the moment I'm guessing? Either way, far more readable than MRI.
03:25:30brixenyeah docs for that are pretty sparse, there are some folks working on docs and commenting the code
03:26:47agardineri recall a conversation on the channel here environment vs machine vs cpu... you may have some luck searching the logs at http://www.donttreadonme.co.uk/rubinius-irc/index.html
03:29:59rueThe thumb used to be one machine per process, one cpu per thread
03:31:12mimetnetthanks for the heads up
03:33:19seydarevan: for the question you asked me 4 hours ago, i'm running PPC Mac OS X
03:33:53boyscout1 commit by Brian Ford
03:33:54boyscout * Prevent mspec ci from using -t ruby.; bf2a75e
03:34:56KirinDav enters the room.
03:35:16seydarhas anyone had any trouble running 'rake build
03:35:50seydaron a PPC - Mac OS X? actually, can someone just tell me what version of GCC you all are using?
03:36:10ruePastie errors
03:36:21brixenseydar: seems to work for me on 4.0.1 ppc os x
03:36:40seydarhmmm
03:36:43brixenour pack/unpack specs fail MRI on ppc
03:37:05benburkert enters the room.
03:37:09brixenI guess the 2 options are use berger's method in ruby or have specs depend on a C prog
03:37:14seydarbrixen: same; 4.0.1
03:37:30brixenseydar: ok, can you pastie the errors you're getting?
03:37:35rubyconsumer enters the room.
03:38:02nicksieger leaves the room.
03:38:10mimetnet enters the room.
03:38:27_ADS enters the room.
03:39:16seydarsure. http://pastie.caboo.se/155173
03:41:10brixenseydar: what os version?
03:41:16seydarMac OS 10.4.11
03:42:05seydarwhat commands should I run after `git pull`? I'm running rake build
03:43:55brixenyeah, rake build is normally fine
03:44:10rubuildius_ppcBrian Ford: bf2a75e10; 1573 files, 5472 examples, 18934 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155176
03:44:50rubuildiusBrian Ford: bf2a75e10; 1573 files, 5472 examples, 18959 expectations, 0 failures, 1 error; http://rafb.net/p/nOP9CE90.html
03:45:52seydarman. why could this happen? It's giving me parse errors, so it's something with.... it's using cc, isn't it
03:46:10seydarhow can I change it to use gcc?
03:46:28rueDoes CC=gcc work?
03:46:43seydarit should - i've just never had to do that.
03:47:40seydarfork. another fail
03:49:29seydarmake is just a wrapper, so it couldn't be make thats causing this...
03:49:33mimetnet leaves the room.
03:50:40brixenrue: what say you about the endian guard approach?
03:51:40brixenheh, top hit on google for _STRUCT_MCONTEXT is our LH bug report
03:52:37ezmobius leaves the room.
03:53:44seydardude! estis hieraux!
03:53:58KirinDav leaves the room.
03:54:27seydarare you serious!??!! so why isn't this appearing in your version, brixen
03:54:49twbray enters the room.
03:54:53ruePaste
03:54:59brixenoh, because I'm on 10.5.2
03:55:00crafterm leaves the room.
03:57:13seydarbrixen: gr..... but still, why wouldn't it error? is _STRUCT_MCONTEXT definied somewhere in leopard?
03:57:29seydarthis is why I hate C
03:57:36ruegrep for it :)
03:57:43_ADSseydar: check signal.c
03:57:49_ADSi mean h
03:57:59seydark - what am I looking for?
03:58:08_ADSSTRUCT_MCONTEXT
03:59:11seydarwell, duh, but am i just looking for its existence?
03:59:50_ADSon cygwin, I found a definition that didn't match context.h
04:00:31_ADSso I got compile errors b/c state.h tried to reference members that didn't exist in the <signal.h> definition
04:00:32seydarshotgun/lib/machine.h, shotgun/subtend/nmc.h
04:01:02seydar_ADS: and the references should be there - and if we're all synced up with the repos, why are they diff?
04:03:51wycats enters the room.
04:04:13_ADSthe mcontext structure is defined in one of the system headers
04:04:31_ADSso it's not guaranteed to be in sync with the repo
04:05:09seydarah. well.... what can I do to fix this?
04:05:53_ADSdunno , I don't even know it that' your problem for sure.
04:06:03_ADSwhat's your last build error?
04:06:24seydarexactly that
04:06:41seydarhttp://pastie.caboo.se/155173
04:08:56boyscout1 commit by Adam Gardiner
04:08:57boyscout * Add -debug option to launch with breakpoint set; 6eb130d
04:09:01_ADSseems like __DARWIN_UNIX03 is defined where maybe it shouldn't be.
04:09:27seydari'm no C-progger, you'll have to hold my hand through this
04:10:04tongueroo leaves the room.
04:10:08evanseydar: you're on 10.5.2
04:10:08_ADStry changing the #ifdef in machine.h to just #if 0
04:10:23evandon't do that
04:10:26evanit will just crash then.
04:10:29_ADSon nevermind
04:10:51evanseydar: intel?
04:11:19seydarevan: 10.4.11
04:11:25evanthats why.
04:11:26seydarOOC
04:11:29seydarPPC*
04:11:45evanthat __DARWIN_UNIX03 shouldn't even be running for you
04:11:52evanapple seems to have busted something in 10.4.11
04:11:55seydarso... what should I do?
04:12:07evango ahead and change to #if 0
04:12:07seydari've gotten rubinius to build on this machine before
04:12:19evanbefore you upgraded to 10.4.11 maybe?
04:12:53seydarnopez
04:12:58seydarwell, i really doubt it
04:13:27evanis your 'ruby' 1.9 or 1.8?
04:13:45_ADSI think evan was right the first time, #if 0 will probably crash, unless you catch them all.
04:13:57evanno
04:14:04evanthat check is only for 10.5
04:14:10evanit shouldn't be true at all on 10.4
04:14:18evani don't know why his gcc is saying what it's saying.
04:14:36evantrust me
04:14:39evani wrote that code.
04:14:49seydarmaaaaaaybe
04:14:54_ADSok :)
04:15:04seydarevan i totally love you. marry me (it worked!)
04:15:51evansorry, i'm on day 731 of my current marriage
04:15:59seydarhooooooooooly shit
04:16:09seydarcongrats, though
04:16:19evanthanks
04:16:21agardineryou count the days? wow...
04:16:24evanhah
04:16:25evanno
04:16:30agardinerhehe
04:16:39seydari'm on like 731 of highschool
04:16:42evani had it on the top of my head because i made a joke to abby and used the number of days we've been married at dinner
04:16:56agardineri was a bit worried there for a sec...
04:17:07agardineri'm lucky to remember the number of years i've been married
04:18:05rubuildius_ppcAdam Gardiner: 6eb130db1; 1573 files, 5472 examples, 18934 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155186
04:18:25evanagardiner: just curious, how old are ya?
04:18:33agardiner38
04:18:47seydarover twice my age! whoa!
04:18:48agardiner(9 years married)
04:19:07agardinerseydar: that's depressing... :-(
04:19:28seydarhey evan - check this out: http://pastie.caboo.se/155187
04:19:43evanagardiner: at least your australian!
04:19:44rubuildiusAdam Gardiner: 6eb130db1; 1573 files, 5472 examples, 18959 expectations, 0 failures, 1 error; http://rafb.net/p/Pn8z0l18.html
04:19:58evanseydar: oh bugger.
04:20:00agardinerhehe... yeah!
04:20:02evanseydar: it's all confused
04:20:08seydari modded shotgun/lib/subtend/nmc.h and shotgun/lib/machine.h so that #if _DARWIN... is #if 0
04:20:11twbray leaves the room.
04:20:17seydarwhat should I do?
04:20:27evanum. i'll have to look into it
04:20:33evani need to go do my taxes
04:20:35evanbbiab.
04:20:35seydarhmph
04:20:40seydari'll do your taxes
04:20:44seydaryou do this code
04:20:59seydardeal?
04:22:01evanheh
04:22:11evanthanks, but i'm pretty sure i have to do mine
04:23:56seydargrr
04:23:58seydarfine
04:24:07seydari have to do some haskell now
04:24:18seydardont get shot!
04:24:30seydarwear a helmet 24/7 to be safe
04:24:35seydarcant have anything happening to you
04:25:34macournoyer leaves the room.
04:28:22KirinDav enters the room.
04:29:30hornbeck leaves the room.
04:31:15hornbeck enters the room.
04:35:11_ADSso I was wondering why benchmark.rb was in stdlib instead of lib/ruby...
04:35:20_ADSand I got this error: http://pastie.caboo.se/155181
04:35:42_ADSit seems like something is wrong with module#include at the file level
04:36:18_ADSbut I couldn't figure out how to write a spec for it.
04:38:57grazaseydar, i got the same thing (10.4.11, PPC)
04:40:17jeremydurham leaves the room.
04:40:42Defiler_ADS: Yeah, I agree. This produces incorrect results in rbx http://pastie.caboo.se/155199
04:41:08_ADSDefiler: right, I just tried a similar example
04:41:13grazai also changed line 34 of shotgun/lib/subtend/PortableUContext.h to not test the value of __DARWIN_UNIX03
04:41:13Defiler..but it works fine if you include 'Z' into a class, say
04:41:27Defilergraza: Oh yeah, I was meaning to do that. Thanks
04:41:49_ADSDefiler: exactly, when i tried it inside a spec body, it passed.
04:43:34DefilerHere is how you can write a spec for it: http://pastie.caboo.se/155201
04:44:06_ADSah, ok
04:45:58dysinger leaves the room.
04:51:53hornbeck leaves the room.
04:53:15srbaker leaves the room.
04:53:32srbaker enters the room.
04:55:38imajes leaves the room.
05:03:18graza_ enters the room.
05:03:19graza leaves the room.
05:03:30_ADSDefiler: I can't get that to work in the specs. Everything I've tried using TOPLEVEL_BINDING fails in MRI too.
05:06:01sfaxon enters the room.
05:19:03graza enters the room.
05:19:04graza_ leaves the room.
05:21:09tongueroo enters the room.
05:27:10_ADS_ enters the room.
05:32:37AndrewO leaves the room.
05:36:25_ADS leaves the room.
05:41:02graza_ enters the room.
05:41:03graza leaves the room.
05:41:44enebo enters the room.
05:48:09nicksieger enters the room.
05:48:29benburkert leaves the room.
05:56:48kamal_farizhow do i run what used to be ./bin/ci -i for running all the failing excludes with the new Tags thing in mspec
05:56:49graza_ leaves the room.
05:56:52graza enters the room.
05:58:28wycats leaves the room.
05:59:43benburkert enters the room.
06:01:28KirinDav leaves the room.
06:04:18brixenkamal_fariz: bin/mspec -g fails path/to/specs
06:05:26kamal_farizbrixen: thanks .. how do you plan to use the tags? i see an option to save successful specs as tags too
06:06:26brixenkamal_fariz: sure, they're pretty much open ended
06:06:59brixenfor example: incomplete(Needs to spec boundary conditions):Some#method does this
06:07:17brixenthe only 'reserved' tag right now is 'fails'
06:07:47brixenI plan on adding a 'benchmark' tag
06:08:02brixenwith a BenchmarkAction
06:09:07kamal_farizbrixen: so the different runners will define a set of tag filters that they will accept and run
06:09:25enebo leaves the room.
06:09:26brixenyou could, but I wasn't really planning other runners
06:10:01brixendo you have something particular in mind?
06:10:11agardinerbrixen: did you mention yesterday that you were thinking of releasing mspec as a gem? coz that would be awesome...
06:10:51brixenyep, I think so
06:11:06brixenwe'll see how it works a bit first
06:11:44kamal_farizbrixen: would a benchmark runner fit?
06:11:54agardinersure, no great rush
06:12:15brixenkamal_fariz: not really a need for a separate runner, checkout mspec/runner/actions
06:12:33brixenit would just be an action invoked with a command line opt to mspec-run
06:12:35tbmcmullen enters the room.
06:12:56kofno leaves the room.
06:14:10agardinerhmm... is there an equivalent to the old bin/ci -C flag?
06:14:35agardineror is it 'rake clean:rbc'?
06:14:54brixenheh, yeah the latter
06:15:00agardinerk, no prob
06:15:03brixenI want to add the -C command to the compiler
06:15:17agardinerooh, yeah probably not a bad idea
06:15:23brixenthe problem is, it was using *_spec.rb to locate the files, but with the shared files, there's no _spec.rb
06:15:37brixenso some stuff wouldn't get recompiled
06:15:53brixenI think it would be useful on the compiler, with the version flag stuff
06:16:00brixenyou could always force compile if you wanted
06:16:21GMFlash leaves the room.
06:16:30GMFlash enters the room.
06:16:48agardineryeah, i think that would be handy, since we don't store our compiled files in a separate tree that's easy to just blow away
06:17:01brixenit's sad that decades after the essential architecture of a cpu was established, we can't all agree on an endian standard
06:17:24agardinerendian wars... it's like the clone wars
06:17:25tarcieriwhy can't everyone agree little endian is fucking stupid
06:17:40brixenI dunno, but it sucks
06:18:05benburkert leaves the room.
06:18:05tarcieriI love how countless CDs were mastered on big endian Macs, and the masters are... little endian PCM
06:18:09brixenyou'd think after 'network byte order' we'd just quit with the differences
06:19:29brixenwell, I'm going to add endian guards to the specs based on zenspider's suggestion
06:19:41brixenI guess if you can't get pack('L') right, too bad :P
06:19:56brixen(we don't get it right on ppc atm)
06:20:51agardinerman, i hate how these ipaddr specs make ci run forever on ubuntu
06:21:12tbmcmullenah, thats what does it... I've been wondering...
06:21:18brixenagardiner: yeah, actually let's exclude them for now
06:21:19agardiner211s... makes me wanna cry, after getting used to <1m
06:21:33brixenIPAddr.new and one other seem to be the worst
06:21:37brixenle'me check again
06:21:49boyscout1 commit by Adam Gardiner
06:21:50boyscout * Convert Selector to use primitives instead of FFI; 3761c42
06:22:19agardinermaybe tag 'em with SLOOOOOW, and exclude that from ci runs? :-P
06:22:23brixenthe remaining spec failure reported by rubuildius is a bitch too
06:22:32benburkert enters the room.
06:22:33brixenheh
06:22:34kamal_farizone thing that i've not been clear about in the project is stdlib .. is everything in and spec-ed?
06:22:49brixenkamal_fariz: no, stdlib is an import from MRI
06:22:59brixenwe're trying to spec stuff and move it to out /lib dir
06:23:05kamal_farizor do they migrate from somewhere as needed, like when trying to get rubygems or rake to run
06:23:35kamal_farizbrixen: so the stuff in /lib is spec-ed, the stuff in stdlib is not?
06:23:37brixenstdlib is just here for convenience
06:23:42brixenbasically yes
06:23:48brixenprobably some exceptions
06:23:59brixenthe idea is to spec it and move it to /lib
06:25:15wifelette leaves the room.
06:26:25kamal_farizcool .. so if i were an overzealous contributor, i could look there for things to spec .. seems like rubinius has got most things well covered in specs that i don't have any examples to show people where to start
06:26:37kamal_farizi'm gonna talk about getting started contributing to rbx tonight at my local meetup
06:27:14brixenoh, sounds cool
06:27:27brixenone source for new specs of course is trying to run code
06:27:32brixenfind out what fails
06:27:33kamal_farizthe other alternative would be to run completeness, or run the fails tags
06:27:37brixencheck if there's a spec for it
06:28:12tbmcmullenright, jeremydurham talked about that at the Boston group last week
06:28:12kamal_farizi'll look at what gems i have installed
06:28:13brixenif you show people completeness, be sure to explain they should run it like: bin/completeness -tr
06:28:20kamal_farizhow do you invoke rake with rbx
06:28:21brixenkamal_fariz: so it runs against MRI
06:28:44kamal_farizbrixen: yup, i remembered that last time i asked you about it
06:29:19brixenkamal_fariz: http://blog.segment7.net/articles/2008/02/14/rubygems-in-rubinius
06:30:00kamal_farizthanks
06:30:05brixensure
06:30:15brixentbmcmullen: you in the boston area?
06:31:49wvl leaves the room.
06:32:05tbmcmullenbrixen: yep
06:32:14wvl enters the room.
06:32:19tbmcmullenbrixen: somerville specifically
06:32:56brixentbmcmullen: how's the ruby community there?
06:33:06rubuildius_ppcAdam Gardiner: 3761c420c; 1573 files, 5472 examples, 18934 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155229
06:33:34tbmcmullenbrixen: Quite active. Theres usually about 40 people at the Boston meetups. Theres a few outside the city as well. One at the MathWorks specifically.
06:34:22tbmcmullenbrixen: Not to mention a ton of Ruby related jobs...
06:34:29brixenwow, that's cool
06:34:38brixengood to hear
06:34:58rubuildiusAdam Gardiner: 3761c420c; 1573 files, 5472 examples, 18959 expectations, 0 failures, 1 error; http://rafb.net/p/1N2zmn94.html
06:35:13tbmcmullenYeah. Its pretty nice. :)
06:35:46brixenI visited boston first time last summer, really nice city I thought :)
06:36:20tbmcmullenRather be in California.. but hey.
06:36:31brixenreally? why's that?
06:37:06tbmcmullenI just was out there on vacation. It was snowing here... and it was 80 there. Thats really about it.
06:37:22brixenheh, weather, yeah I can relate
06:37:32tbmcmullenwhereabouts are you?
06:37:42brixenportland oregon
06:37:47tbmcmullenah, right
06:37:50brixenso much rain, although it was almost 60 today
06:37:55brixenamazing
06:38:11agardinerlater..
06:38:16brixenagardiner: later
06:38:26agardiner leaves the room.
06:38:51aotearoa enters the room.
06:47:55kamal_farizpastie
06:48:13pastiehttp://pastie.org/155232 by kamal_fariz.
06:48:40kamal_farizis drbrain around? I'm getting an invalid gem format error
06:48:46kamal_farizwhen installing rake from rbx
06:56:45boyscout2 commits by Brian Ford
06:56:46boyscout * Exclude Socket#unpack_sockaddr_in spec. See tag comment.; 65b4ed8
06:56:47boyscout * Exclude super slow IPAddr specs.; 9fbda05
07:01:56graza_ enters the room.
07:01:56graza leaves the room.
07:02:41cavalle enters the room.
07:07:55rubuildius_ppcBrian Ford: 65b4ed860; 1573 files, 5469 examples, 18919 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155233
07:09:42rubuildiusBrian Ford: 65b4ed860; 1573 files, 5469 examples, 18944 expectations, 0 failures, 0 errors; http://rafb.net/p/Uytt7c39.html
07:13:32brixenhm, why is rubuildius showing a pastie if there are no errors
07:18:48evankamal_fariz: are you using rubygems from svn?
07:19:03evanyay got my taxes done
07:19:03kamal_farizevan: no, from rubyforge
07:19:14evanyou need to use rubygems trunk from svn
07:19:36kamal_farizok, retrying
07:19:49brixenevan: turbo tax?
07:20:44evanyep
07:20:45evanit rocks
07:20:53evangot a $3500 refund
07:20:53brixenheh, yep
07:20:57brixennice!
07:21:00tbmcmullenwow
07:23:11lstoll leaves the room.
07:24:21graza enters the room.
07:24:22graza_ leaves the room.
07:29:17benburkert leaves the room.
07:31:53boyscout1 commit by Brian Ford
07:31:54boyscout * Added endian guards to MSpec.; a7f8247
07:35:51VVSizbrixen: hi there! A quick question: is there some trick to convert existing excludes to new tag files? Sure, I could write a script to do that, but thought that maybe you have something already ;)
07:35:55twbray enters the room.
07:38:02brixenVVSiz: heh, sorry! I just did the script thing
07:38:29brixenVVSiz: otoh, you can just use mspec-tag
07:39:27VVSizyeah, this option is very attractive. :) the only problem is that I wipe out all the excludes specific to MacOS :)
07:39:58brixenoh, you guys keep both?
07:40:16brixenoh linux, windows, osx
07:40:17brixengotcha
07:41:02nemerle leaves the room.
07:41:19VVSizwe actually have a single one that is a union of all failures on all platforms
07:41:47VVSizbut probably now we could start differentiate somehow (that's where tags could possibly be more flexible)
07:43:04rubuildius_ppcBrian Ford: a7f824775; 1573 files, 5469 examples, 18919 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155237
07:43:28brixenVVSiz: yeah, either separate tags or comments
07:43:50brixenalthough, separate tags is probably better, since comments won't be matched for deleting a tag
07:44:07brixencomments are meta-tags ;)
07:44:10VVSiz:)
07:44:14nemerle enters the room.
07:44:41rubuildiusBrian Ford: a7f824775; 1573 files, 5469 examples, 18944 expectations, 0 failures, 0 errors; http://rafb.net/p/6QXrsY26.html
07:49:32claudio enters the room.
07:51:23claudio leaves the room.
07:56:49tongueroo leaves the room.
07:57:16brixenjeez, the String#unpack specs are horrid
07:57:52brixenI dunno why they don't follow the format of the Array#pack specs
07:59:04tbmcmullenthose are rather painful...
08:00:17geekounet leaves the room.
08:00:49tbmcmullenquick question: If a class does not exist in lib/, but does exist in stdlib/, will it automatically be loaded from there?
08:01:04brixenyep
08:01:09brixenwe added stdlib to $:
08:01:15tbmcmullenAhhhh.
08:01:25tbmcmullenThrew me way off...
08:02:20brixenheh, yeah magic
08:02:43tbmcmullenWell, the thing is that I had a dev version of CGI in my lib/ directory, but I deleted it.
08:02:48tbmcmullenAnd magically... heh.
08:03:44boyscout1 commit by Brian Ford
08:03:45boyscout * Added little/big_endian guards to Array#pack and String#unpack specs.; 83ed716
08:04:47kamal_farizevan: with trunk rubygems and fresh rbx, i'm still hitting the invaild gem format for rake
08:04:51kamal_farizweird
08:04:54nicksieger leaves the room.
08:05:04evanok
08:05:08evani'll have to ask drbrain then.
08:05:31drbrainkamal_fariz: do the rubygems tests pass?
08:05:50kamal_farizdrbrain: how do i run those
08:06:32drbrainsince you don't have a working rake, run autotest in ruby, then edit that command line
08:06:47drbrainor, you can use a shell for loop
08:07:29kamal_farizno such file to load -- rubygems/gem_open_uri (LoadError)
08:07:40drbrainah, sec
08:09:17drbrainkamal_fariz: update and try again
08:09:32kamal_farizok
08:09:57rueNite
08:11:57drbraindamn debug rdoc
08:12:36kamal_fariztjere
08:12:49rubuildius_ppcBrian Ford: 83ed71617; 1573 files, 5477 examples, 18920 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155243
08:12:52kamal_farizthere's something else that requires gem_open_uri
08:12:59kamal_farizi think mirror_command.rb?
08:14:42rubuildiusBrian Ford: 83ed71617; 1573 files, 5477 examples, 18945 expectations, 0 failures, 0 errors; http://rafb.net/p/bA9t9l24.html
08:14:55drbrainkamal_fariz: yes, I missed it
08:16:13evanbrixen: did the exit code of mspec change maybe?
08:16:32evanis it returning non 1 all the time now? perhaps thats why rubuildius is thinking there was an error
08:16:49drbrainkamal_fariz: ok, I updated mirror_command too
08:16:56drbrainkamal_fariz: sorry for the breakage :(
08:17:22kamal_farizdrbrain: is this with the net/http hack? :)
08:17:35evandrbrain: downloading should be a little faster now
08:17:43evandrbrain: well, starting since last week
08:17:44drbrainkamal_fariz: it is with keep-alives
08:17:48evanwhen I made timeout.rb sane
08:17:58kamal_farizsweet
08:18:04drbrainevan: cool
08:18:18evanit doesn't spawn a watcher thread for EVERY time into timeout()
08:20:28drbraingreat, I found a bug in Gem::DependencyList
08:21:01kamal_farizyou guys on leopard and not using emacs, what terminal do you use? iTerm freezes on me once in a while ..
08:21:33kamal_farizdrbrain: i get 5F, 94E
08:21:47evanTerminal.app
08:21:47drbraincan you paste them?
08:21:49thehcdreamer enters the room.
08:21:54evanthats all I've ever used.
08:21:55drbrainTerminal.app
08:22:02drbrainsame
08:22:14evanI used GLTerm back in 10.0, 10.1 days some
08:22:17evanbecause it was SUPER fast
08:22:23kamal_farizis there a hack to switch between tabs with cmd-1,23?
08:22:25evanand rendered pcf fonts directly as GL textures
08:22:26evanit rocked.
08:23:02kamal_farizi remember GLTerm .. but only for the old console giggle factor
08:23:14brixenevan: I checked, it exits with 0 on success, so dunno...
08:23:50brixenVVSiz: I updated some docs: http://rubinius.lighthouseapp.com/projects/5089/specs-runners
08:24:00brixenVVSiz: le'me know what needs clarification
08:24:05ezmobius enters the room.
08:24:27kamal_farizpastie
08:26:49graza leaves the room.
08:26:51wycats enters the room.
08:29:07kamal_farizdrbrain: http://rafb.net/p/y9t9V439.html
08:29:44wycats_ enters the room.
08:30:07drbrainkamal_fariz: I strongly suspect something broke in IO
08:30:48drbrainbut, it could be anything
08:30:54drbrainI will look into it tomorrow
08:31:04kamal_farizok
08:32:49drbrainlater, I'm off to watch a hot but wooden Summer Glau kick ass in another episode of Terminator
08:33:15evanenjoy!
08:33:19evani've got a bunch of those tivo'd
08:33:22evani need to catch up
08:34:56brixenanyone know if MACHTYPE is standard on unix
08:35:00brixenenv var that is
08:35:06evanhm
08:35:07evanprobably not.
08:35:37brixenI see it on osx, gentoo, and ubuntu
08:35:57brixenI was hoping maybe that would be an independent way to get endian-ness
08:36:25evanooh
08:36:27evanmaybe...
08:36:39evanyou might need multiple ways
08:36:46evanMACHTYPE, the pack() hack
08:37:10brixenI'd prefer MACHTYPE over pack() hack
08:37:11evanwe could add a Rubinius:: constant that indicates it
08:37:19brixenI could cascade I suppose
08:39:50geekounet enters the room.
08:45:25wycats leaves the room.
08:56:37VVSizbrixen: thanks. most confusing for me were those action-tag and action-string options
08:57:11VVSizespecially "... will trigger the specified action". what's the action? - was not clear
08:57:49brixenVVSiz: ok
08:58:32VVSizoriginally, I thought that I could invoke some external command... :)
08:58:34evanwhere are the tags stored?
08:58:40evannot in spec/data anymore?
08:58:47VVSizspec/tags now
08:59:04brixenVVSiz: http://pastie.org/155255
08:59:09evanwhats critical.txt?
08:59:11brixenVVSiz: that's pretend mode
08:59:19maduyb__ enters the room.
08:59:33brixenevan: that was for stuff that segfaulted, to exclude it from bin/ci -c
08:59:39brixenwe can probably remove it now
08:59:42evanok
08:59:43VVSizbrixen: cool! :)
09:00:54brixenVVSiz: so spec has --dry-run, do you think --pretend is better or should I use --dry-run ?
09:01:46VVSiz--dry-run is much better. --pretend is unclear (to me, at least)
09:01:54brixenok
09:02:01brixenwasn't sure which was more clear
09:03:36evan--underachieve
09:03:48VVSiz:)
09:04:08brixenheh
09:04:14brixen--fake-it
09:06:35evanyou know, there are so many CS papers, but it's so hard to find readable code that actually implements the ideas of the papers.
09:07:09brixenheh, l.t.r. :)
09:07:17evanltr?
09:07:19kamal_farizbrixen: how do you use the action tag to drop into debug mode just as the spec is about to execute? i tried this: ./bin/mspec run -K fails --spec-debug spec/ruby/1.8/core/process/
09:07:19brixenthe profs need something for tests
09:07:23brixenleft to reader
09:07:27evan:)
09:07:27scooprweekly "implement a paper" challenge? =)
09:07:57brixenkamal_fariz: that should work, le'me check
09:08:29brixenkamal_fariz: -g not -K
09:08:37brixen-K is only for mspec-tag
09:08:53brixenor not, damn, now I'm confused :P
09:09:16kamal_fariznope .. it doesn't take g
09:09:55kamal_farizbrixen: but if i understand the options right, anything u specify in -K, you can do the anything you want specified in -S
09:10:16octopod enters the room.
09:10:21brixenkamal_fariz: this is how I did it earlier, maybe I broke it: http://pastie.org/154654
09:12:25tbmcmullen leaves the room.
09:12:50ezmobius leaves the room.
09:15:23kamal_farizbrixen: i think i might be misunderstanding the utility of the option above .. my assumption is that it the debugger will put a breakpoint just as i enter the spec and let me step through from there
09:15:27Arjen_ enters the room.
09:15:30brixenkamal_fariz: yep, I'll have to fix that. I changed ActionFilter base so TagAction could operate on any spec
09:15:33kamal_farizbrixen: but i seem to be in mspec's code most of the time
09:15:41kamal_farizand not in the implementation of the class under spec
09:16:09brixenkamal_fariz: you have to use s, the debugger starts in mspec's code
09:16:22brixenyou have to step to the spec
09:17:12kamal_farizah ok
09:17:21brixenthe debuger will start before any before blocks are run
09:17:34brixenso you can walk the whole process
09:18:31dewd leaves the room.
09:18:35pietia enters the room.
09:19:10kamal_farizbrixen: so i need to step in at the line that is "protect nil, spec"
09:19:16kamal_farizlooking for the right entry point
09:19:22brixenyep
09:19:26brixenfor the spec itself
09:19:56brixenor protect "before :each", @before
09:19:58brixenetc
09:21:26rudebwoy leaves the room.
09:29:33brixenkamal_fariz: agardiner is adding l output for each s, n soon, so that will make it much easier to use
09:29:36twbray leaves the room.
09:30:43ragnard enters the room.
09:31:21kamal_farizbrixen: cool .. i'm having difficulty getting into the actual spec still .. i've stepped in up to when it wants to to_a the spec
09:31:31kamal_farizi don't see how i can get into the .each block
09:32:13brixenhmm not sure, perhaps try setting a breakpoint in the method you want
09:32:20brixeni.e. something the spec calls
09:33:21kamal_farizalright i'll do that
09:33:47kamal_farizbtw, anyone have existing material so i can show n tell at the local Malaysia.rb meetup?
09:33:51kamal_farizabout rbx
09:34:54evanwhat kind of material are you looking for?
09:36:01kamal_farizevan: fifty foot view of the architecture which i'm not clear my self (blind leading the blind)
09:36:14kamal_farizand contributing to it .. this i can do a bit
09:36:53kamal_farizevan: i understand from parsing, to sexp, to building the ast
09:37:01kamal_farizthen it gets hazy
09:37:05kamal_fariz:D
09:37:14evangloss over it
09:37:15evan:)
09:37:23evango over the parts you know
09:37:40evanthats the best way to handle it
09:37:49evanif people have questions you can't answer, have them email me
09:38:01kamal_farizcool
09:38:31kamal_farizi have a question, so after the ast is built, how does rubinius execute it?
09:38:55kamal_farizit traverses down the tree, pushing and popping things into/from the stack?
09:40:16evanno
09:40:24evanthe AST is handed to the compiler
09:40:29evanthe code in lib/compiler/
09:40:42evanit generates a CompiledMethod object from an AST
09:40:57evanthe VM knows how to execute a CompiledMethod objectly, and does so
09:41:22evanthe VM itself knows nothing about how the CompiledMethod was created.
09:41:37evanonly how to execute the instructions contained with it
09:41:45kamal_farizCompiledMethod is pure ruby?
09:41:56kamal_farizi mean a pure ruby object
09:43:04evanit's first class
09:43:04evanyes
09:43:11evanit's a normal ruby object
09:43:49RyanTM leaves the room.
09:44:54kamal_farizevan: has any publisher approached you or anyone to write a book about rubinius? :)
09:45:27evanheh
09:45:51evanbrixen and I have talked with an oreilly publisher a couple of times
09:48:07kamal_farizthat's cool
09:49:16scooprRubinius-08: The Language and Its implementation
09:49:20kamal_farizevan: how do you decide what becomes primitives vs normal ruby code
09:49:22evanscoopr: ha!
09:49:25evanthats perfect.
09:49:44evanprimitives are things that you can't write in ruby
09:49:45brixenheh, that's really cool
09:49:45kamal_farizwhat animal would represent it though ..
09:49:46scooprcall it the redbook
09:50:04evanthere are a few exceptions to that rule, but general, thats the rule.
09:50:10evanthings like Fixnum#+
09:50:19evanyou have to drop down to a lower level to actually perform the addition
09:59:49dbussinkmorning!
10:01:00evanallo.
10:01:02evani'm about to head to bed.
10:02:54dbussinkah, well, have a good night sleep then
10:03:44boyscout2 commits by Brian Ford
10:03:45boyscout * Added --dry-run mode to MSpec. Available from mspec-run script.; 0fa8ca4
10:03:46boyscout * Fix ActionFilter and TagAction's handling of -K option.; 9664099
10:07:59aotearoa leaves the room.
10:13:07rubuildius_ppcBrian Ford: 0fa8ca4e7; 1573 files, 5477 examples, 18920 expectations, 0 failures, 28 errors; http://pastie.caboo.se/155272
10:14:51rubuildiusBrian Ford: 0fa8ca4e7; 1573 files, 5477 examples, 18945 expectations, 0 failures, 0 errors; http://rafb.net/p/oO0CZ141.html
10:15:47lstoll enters the room.
10:21:44w1rele55 enters the room.
10:27:53cavalle leaves the room.
10:30:54antares enters the room.
10:34:17zenspider leaves the room.
10:38:06chris2 enters the room.
11:02:41cavalle enters the room.
11:27:20langenberg enters the room.
11:30:06langenberg leaves the room.
11:36:46rudebwoy enters the room.
11:49:54d2dchat leaves the room.
11:50:09imajes enters the room.
12:14:57nitay_ enters the room.
12:18:10ctennis leaves the room.
12:29:34nitay leaves the room.
12:31:01chris2 leaves the room.
12:38:17kevwil enters the room.
12:38:39kevwil leaves the room.
12:42:04ragnard leaves the room.
12:48:06rao45 enters the room.
12:52:30ariekere enters the room.
12:52:47w1rele55 leaves the room.
12:53:08ariekere leaves the room.
13:04:15nitay_ leaves the room.
13:11:20moofbong enters the room.
13:14:02hornbeck enters the room.
13:15:42rubyconsumer leaves the room.
13:16:57jeremydurham enters the room.
13:39:37imajes leaves the room.
13:42:29kwatch enters the room.
13:49:17cremes enters the room.
13:55:19cremes leaves the room.
13:59:51be9 enters the room.
14:00:43wycats_ leaves the room.
14:01:56rao45 leaves the room.
14:04:38be9evening. hm, after pulling and rebuilding I get a "zsh: illegal hardware instruction bin/mspec ci"
14:07:14AndrewO enters the room.
14:08:52rubyconsumer enters the room.
14:10:23be9With this new mspec stuff, how can I run in verbose mode to see what spec it bumps on?
14:12:23Arjen_ leaves the room.
14:14:38dbussinkwhat platform are you on?
14:15:02be9linux 64bit, as usual
14:17:37jeremydurham leaves the room.
14:18:38rudebwoy leaves the room.
14:19:29dbussinkbe9: ./bin/mspec ci --help
14:21:17probablycorey enters the room.
14:22:45be9dbussink: thanks. so it's spec/ruby/1.8/library/socket/basicsocket/getsockopt_spec.rb
14:23:11dbussinkbe9: probably fails in a ffi call doesn't it?
14:23:38be9dbussink: given the strangeness of the error, seems so
14:25:36dbussinkbe9: no backtrace or whatsoever?
14:26:59be9dbussink: ran under gdb, doesn't help much. http://pastie.caboo.se/155370
14:27:25dbussinkbe9: and if you compile with rake build:vm:clean build:vm:dev ?
14:28:00be9dbussink: will do now
14:28:21gnufied enters the room.
14:30:05benburkert enters the room.
14:31:48_mutle leaves the room.
14:31:53mutle enters the room.
14:32:49be9dbussink: much bettah now http://pastie.caboo.se/155374
14:33:05dbussinkbe9: ah, yeah, ffi_call
14:33:40gnufiedokay folks, how do you apply a patch created using git-diff ?
14:33:54dbussinkgnufied: just use patch
14:34:03dbussinkgnufied: for git-format-patch you can use git-am
14:34:14gnufiedand git-apply ?
14:34:21be9gnufied: git apply also works
14:34:33gnufiedwhen using patch it prompts me for file name to patch
14:34:56gnufiedgit-apply totally balks out:
14:34:56gnufiedfatal: git-apply: bad git-diff - expected /dev/null on line 24
14:35:12dbussinkgnufied: patch -p1 < file.diff
14:35:35be9gnufied: git apply works for patches made by git
14:36:07gnufiedhmm, cool.
14:36:47langenberg enters the room.
14:36:55kwatch[Error] I can't compile rubinius on MacOS X (tiger). Does anyone succeed to compile on Mac?