Index

Show enters and exits. Hide enters and exits.

00:02:19rubuildius_amd64Brian Ford: f89bd8c6c; 2091 files, 6703 examples, 23549 expectations, 0 failures, 0 errors; http://rafb.net/p/JSMwlC89.html
00:05:33tarcieriw00t!!!
00:05:47tarcieriMongrel works with a "Hello world!" HttpHandler now
00:07:04GMFlash leaves the room.
00:07:45rubuildius_ppcBrian Ford: f89bd8c6c; 2091 files, 6706 examples, 23577 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187381
00:09:31rby leaves the room.
00:12:36obiejuan leaves the room.
00:12:45boyscout1 commit by Tony Arcieri
00:12:46crio leaves the room.
00:12:46boyscout * Rubinius.asm Mongrel parser fixed to actually work with Mongrel::HttpServer; 7d246b1
00:12:51obiejuan enters the room.
00:13:32obiejuan leaves the room.
00:14:51wycats_ enters the room.
00:16:06binary42 enters the room.
00:22:25rubuildius_amd64Tony Arcieri: 7d246b17e; 2091 files, 6703 examples, 23549 expectations, 0 failures, 0 errors; http://rafb.net/p/DWEE1i80.html
00:25:54rubuildius_ppcTony Arcieri: 7d246b17e; 2091 files, 6706 examples, 23577 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187394
00:27:33tarcierihmm
00:27:39tarcieriit's ~135x slower than the C extension
00:27:39tarcieriheh
00:27:49MenTaLguYbtw, what's the story on your VMActor rack adaptor thingy?
00:28:27tarcieriGuess I'll try that out now :)
00:28:52MenTaLguYI mean, how does it work?
00:28:59MenTaLguYin particular, how are requests passed across VM boundaries?
00:29:10tarcieriI was planning on effectively doing a prefork server
00:29:26tarcierimessage passing the fileno of the accepted socket between VMs (as an integer)
00:29:37MenTaLguYahh
00:29:42tarcierithen have Mongrel running inside each of the child VMs
00:30:31wycats leaves the room.
00:31:07trythilit
00:31:28trythil enters the room.
00:31:38trythilwhoops
00:39:57yugui enters the room.
00:41:17mkrauskopf leaves the room.
00:42:09ruetarcieri: You are still leveraging the base event loop though, right, not building something on top of it?
00:42:36tarcierirue: yep
00:42:45tarcierirue: it'd be a normal TCPServer accepting connections
00:45:05smparke1 leaves the room.
00:52:57smparke1 enters the room.
01:01:31MenTaLguYtarcieri: I've been answering some questions that aok had for a talk about rails concurrency
01:02:00MenTaLguYtarcieri: I'm going to CC you on one of my answers about this stuff, since I'm still a little shaky on some details
01:03:13tarcieriRails... concurrency? Isn't that an oxymoron? :)
01:04:04MenTaLguYrails and threads maybe
01:04:08MenTaLguYotherwise, no
01:04:38tarcierishould ask Ezra what he discovered about the giant lock in ActionPack
01:05:01MenTaLguYin this case the question was regarding how dispatching to Rubinius MVMs each with their own Rails instance might work
01:05:06tarcieriwhatever it was, I think it was ugly enough to motivate him to write Merb instead of trying to fix Rails
01:05:29tarcieriMenTaLguY: yeah, that's what I was thinking of
01:06:28MenTaLguYJosh Peak and some other folks have been working on untangling the Rails threading stuff actually
01:06:30MenTaLguYsince Ezra moved on
01:06:46MenTaLguYer, Peek
01:06:46tarcieriaah
01:06:56MenTaLguYI don't envy them
01:07:08MenTaLguYbut it sounds like they might have some traction
01:07:54MenTaLguYgiant locks are the worst though
01:08:07MenTaLguYone of the reasons I think CPython and CRuby are doomed concurrency-wise unless someone really motivated gets in there
01:08:24MenTaLguYwell, for values of concurrency which == threads
01:08:57tarcierisomeone did fine grained locking for CPython
01:09:06MenTaLguYsomeone really motivated :)
01:09:14MenTaLguYI hadn't known that actually
01:09:19MenTaLguYgoes to show how little time I spend in Python-land these days
01:09:25tarcierihttp://code.google.com/p/python-safethread/
01:10:25tarcierihttp://code.google.com/p/python-safethread/wiki/DeadlockFallacy
01:10:30tarcieri^^^ that's somewhat amusing
01:11:12hoopyso pp doesn't call to_s on objects?
01:12:07tarcierifutures or IoC solve what he's whining about re:
01:12:07tarcieriThis split between "immediate" and "eventual" calls makes deadlocks impossible, but it only does so by making large operations impossible.
01:12:39tarcierithat whole rant makes it pretty clear he doesn't have much experience with shared nothing concurrency
01:13:04MenTaLguYyeah ...
01:13:14tarcieribut THREADS RULE!
01:13:19MenTaLguYevidently he didn't really solve CPython with shared-memory threads either
01:13:24MenTaLguYso I go back to what I said before :)
01:13:31tarcieriHEH
01:49:11brapse enters the room.
02:01:44trythil leaves the room.
02:04:46kw enters the room.
02:10:16anteaya_ leaves the room.
02:11:53MenTaLguY leaves the room.
02:19:48binary42 leaves the room.
02:23:14TheWordNerd leaves the room.
02:24:01VVSiz_ enters the room.
02:26:08jtoy enters the room.
02:26:55d2dchat enters the room.
02:31:19VVSiz leaves the room.
02:35:05manverudbussink: ?
02:35:25manverudbussink: i don't even have rights yet :)
02:54:31brixensweet Io moved to github (about a month ago)
02:55:05jtoy leaves the room.
02:55:17brixenheh, and I just did a clone in about 30 sec instead of the minutes (hours) it took with darcs
02:55:43manveruDefiler: anything new about the extend thing?
03:06:53dblack enters the room.
03:09:13benny enters the room.
03:29:47antares leaves the room.
03:35:08manverubrixen: why is it that CompiledMethod#file returns a symbol?
03:39:10brixenmanveru: no idea :)
03:39:43manveruhrm
03:39:50manverui'm searching for a shortcut
03:39:56manverucm = String.method(:to_s).compiled_method; puts File.readlines(cm.file.to_s).slice(cm.lines.first.last - 1, cm.lines.size + 1)
03:40:08BluR enters the room.
03:44:45brixenmanveru: have you looked at the Debugger code? perhaps adam has some code for reading the source given a CM
03:45:37manveruhmh
03:45:45manveruwill take a look later
03:48:32cored enters the room.
03:49:19xxi leaves the room.
03:56:03RyanTM_ leaves the room.
03:57:45RyanTM_ enters the room.
04:01:32KirinDav enters the room.
04:04:27headius leaves the room.
04:05:36brapse leaves the room.
04:17:10dblack leaves the room.
04:20:21marnen enters the room.
04:24:42wycats_ leaves the room.
04:32:17ruebrixen: Damn you, now it is 11pm :P
04:32:39rueMaybe I will get enough sleep reserves so I can keep dbussink company the next couple weeks
04:33:20cored leaves the room.
04:41:24marnen_ enters the room.
04:46:20nicksieger enters the room.
04:51:32RyanTM_ leaves the room.
04:52:42yipstar enters the room.
04:56:51brapse enters the room.
04:57:53yipstar leaves the room.
04:57:53marnen leaves the room.
05:08:21boyscout3 commits by Marnen Laibow-Koser
05:08:22boyscout * Amplify a comment.; dc9f427
05:08:23boyscout * Get BigDecimal#+ working. I hate this algorithm, but it works without running out ...; b9776b9
05:08:24boyscout * Implement BigDecimal#coerce.; b87ff5c
05:09:42obvio leaves the room.
05:11:11marnengod, that was tough
05:11:28marnenand still only preliminary
05:11:46brixenheh
05:12:33brixenI have the most bizarre thing I've ever seen
05:12:46brixenI've replaced $stdin with a string obj that has a gets method
05:13:06brixenthe spec checks that #gets (Kernel) sets $_ correctly
05:13:23brixenit works with rspec, but with mspec it's returning the first line of the file the spec is in
05:13:37brixenheh
05:14:08brixenI cannot fathom how it is getting the first line of the spec file
05:14:23brixena standalone script that includes the helper obj works fine
05:14:29brixenoh, this is all in MRI
05:17:01dysinger leaves the room.
05:17:21rubuildius_amd64Marnen Laibow-Koser: dc9f427ec; 2091 files, 6706 examples, 23565 expectations, 0 failures, 0 errors; http://rafb.net/p/7ZQG4h23.html
05:19:27obvio enters the room.
05:22:55obvio leaves the room.
05:24:08obvio enters the room.
05:25:31rubuildius_ppcMarnen Laibow-Koser: dc9f427ec; 2091 files, 6709 examples, 23593 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187492
05:27:36boyscout1 commit by Marnen Laibow-Koser
05:27:37boyscout * Finish implementing #@- and #infinite?. Update spectags, of course; c114106
05:29:05boyscout1 commit by Marnen Laibow-Koser
05:29:06boyscout * Typo.; 56c0088
05:33:40marnen enters the room.
05:34:15dysinger enters the room.
05:35:39marnenok, time for sleep...got one more show to play tomorrow
05:37:19rubuildius_amd64Marnen Laibow-Koser: 56c0088f9; 2091 files, 6710 examples, 23583 expectations, 0 failures, 0 errors; http://rafb.net/p/103V0I35.html
05:40:07BluR leaves the room.
05:40:28marnen leaves the room.
05:43:45rubuildius_ppcMarnen Laibow-Koser: 56c0088f9; 2091 files, 6713 examples, 23611 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187496
05:43:46rubuildius_ppcMarnen Laibow-Koser: c11410654; build failed! http://pastie.caboo.se/paste/187494
05:44:27nicksieger leaves the room.
05:49:53wycats_ enters the room.
05:50:50marnen_ leaves the room.
06:42:13adamwiggins enters the room.
06:43:54hornbeck enters the room.
06:48:20adamwigginsAnyone able to do a quick code review on this patch? http://rubinius.lighthouseapp.com/projects/5089-rubinius/tickets/518-io-popen-read-write-pipe-now- passes-spec
06:53:38GMFlash enters the room.
06:59:41adamwigginsAnd another: http://rubinius.lighthouseapp.com/projects/5089-rubinius/tickets/519-io-write-empty-strings
07:00:51brapse leaves the room.
07:09:04dysinger leaves the room.
07:18:49mkrauskopf enters the room.
07:19:59wycats_ leaves the room.
07:28:52yipstar enters the room.
07:31:35yipstar leaves the room.
07:41:19Somebee enters the room.
07:42:14dysinger enters the room.
07:45:51adamwiggins leaves the room.
08:06:47dysinger leaves the room.
08:14:00dysinger enters the room.
08:21:01headius enters the room.
08:25:58qwert666 enters the room.
08:27:48NoKarma enters the room.
08:27:53NoKarmahey
08:27:58NoKarmaDefiler: ping
08:39:09d2dchat leaves the room.
08:55:20binary42 enters the room.
09:03:14DefilerNoKarma: yo
09:04:16wvdschel enters the room.
09:04:23qwert666_ enters the room.
09:11:59wyhaines leaves the room.
09:14:07NoKarmaDefiler: I tested the pure ruby mysql driver
09:14:22DefilerYeah?
09:14:23NoKarmaDefiler: seems to work perfectly
09:14:37DefilerYep. Got fixed a couple of days ago
09:14:47NoKarmaDefiler: oh, ok :)
09:15:03DefilerTurned out to be a bug in the buffered IO stuff
09:15:45DefilerIt's ultra late here, so I'm going to turn in. Enjoy!
09:18:00NoKarmaDefiler: night
09:20:41qwert666 leaves the room.
09:27:42dysinger leaves the room.
09:29:57dysinger enters the room.
09:31:20KirinDav leaves the room.
09:38:52wvdschel leaves the room.
09:46:19Maledictus enters the room.
09:54:02mentz enters the room.
09:56:56ixx enters the room.
10:04:22yugui leaves the room.
10:07:11GMFlash leaves the room.
10:12:41yugui enters the room.
10:30:32yugui leaves the room.
10:40:43NoKarma_ enters the room.
10:41:08NoKarma.
10:45:24NoKarma_ leaves the room.
10:58:45akshay enters the room.
11:09:53Rich_Morin_ enters the room.
11:15:07kw leaves the room.
11:17:44akshay leaves the room.
11:27:16squeegy leaves the room.
11:32:20chris2 enters the room.
11:34:07squeegy enters the room.
11:41:12akshay enters the room.
11:51:10akshay_ enters the room.
12:09:38akshay leaves the room.
12:10:06dysinger leaves the room.
12:16:20boyscout1 commit by Adam Wiggins
12:16:21boyscout * IO#write returns 0 when writing a blank string, to match behavior of MRI; 4a1f394
12:24:17dblack enters the room.
12:26:45VVSizfolks, anyone knows who is "Marnen Laibow-Koser" here?
12:27:21rubuildius_amd64Adam Wiggins: 4a1f39426; 2091 files, 6711 examples, 23584 expectations, 0 failures, 0 errors; http://rafb.net/p/iNrRbq75.html
12:27:21VVSizI'd like to discuss the recent RubySpec changes in BigDecimal with him
12:27:23joachimmVVSiz_: marnen, I think
12:27:43VVSizjoachimm: good. So, he's hanging out here from time to time, right?
12:29:56joachimmVVSiz_: seems so.
12:30:42rubuildius_ppcAdam Wiggins: 4a1f39426; 2091 files, 6714 examples, 23612 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187549
12:48:18dbussinkVVSiz_: marnen yeah, but lives somewhere in the us afaik, so not in atm
12:52:31joachimmseems sleep only reads the zero
12:52:54joachimmsorry wrong channel
12:53:28headiushttp://headius.blogspot.com/2008/04/promise-and-peril-for-alternative-ruby.html
12:53:42headiusfyi...feel free to comment
12:55:54dblackheadius: "death-grip" is a slightly ungenerous way to put it... :-)
12:56:49headiusI picture a 900-foot Matsumoto-san wrassling with a 900-foot tall Ruby-colored Godzilla
13:00:39dblackheadius: do you really mean "wither Ruby 2.0" or "whither Ruby 2.0"? :-)
13:00:49headiuswhither
13:00:54headiusthanks
13:01:05dblackis a compulsive proof-reader
13:01:12headiusI tried to look it up but got no help
13:01:35headiusfixed
13:01:50headiusyeah, me too, tom hates it
13:02:17headiusespecially when I go through stuff he's written and flip word sequences
13:02:38headius"we have mostly" vs "we mostly have" kinds of things
13:04:26lopex enters the room.
13:04:43NoKarmaheadius: another typo. search for "and and"
13:05:11headiusexcellent, thanks
13:23:13dblackheadius: i just sent you some corrections
13:23:26headiusgreat, thanks
13:24:04dblackheadius: i want to start Ruby 1.8.65 :-)
13:24:12headiusheheh
13:24:23headiusyes, I know a few others who would join you
13:24:57headiusFYI my understanding is that the japanese almost all use MatzRuby...and I used it because MRI is a little more obscure
13:25:23headiusbut I did add an update mentioning the MRI name, and I usually use "Ruby 1.8" in the rest of the article I think
13:26:41dblackabout this "MatzRuby" thing....
13:26:44dblackwell, see my email :-)
13:27:00dblackheadius: oh
13:27:06dblackheadius: sorry, you were already answering me
13:27:08dblackheadius: duh
13:27:12dblackreaches for the coffee
13:27:17headiusheheh
13:28:18rueBraaainnsss
13:28:39headiusthanks for the edits, amazing what you can miss proofing your own stuff
13:30:15dblackheadius: tell me about it
13:58:47d2dchat enters the room.
14:00:54wycats enters the room.
14:11:27hornbeck leaves the room.
14:11:38Skip enters the room.
14:15:56headius_ enters the room.
14:15:56headius leaves the room.
14:21:52jtoy enters the room.
14:32:06headius enters the room.
14:32:06headius_ leaves the room.
14:33:08headius leaves the room.
14:46:04aquasync enters the room.
14:51:11rby enters the room.
15:02:30RyanTM_ enters the room.
15:06:29dblack leaves the room.
15:07:51marnen enters the room.
15:08:43akshay_ leaves the room.
15:18:05jtoy leaves the room.
15:19:30jtoy enters the room.
15:28:01fbuilesv enters the room.
15:31:40wyhaines enters the room.
15:34:16wyhaines leaves the room.
15:36:10weepy enters the room.
15:36:27weepyany one seen this artivle ?
15:36:28weepyhttp://headius.blogspot.com/2008/04/promise-and-peril-for-alternative-ruby.html
15:41:33jtoy leaves the room.
15:43:49Vagabond enters the room.
15:45:46thehcdreamer leaves the room.
15:47:40dblack enters the room.
15:47:56Vagabond leaves the room.
15:47:57obiejuan enters the room.
15:48:10Vagabond enters the room.
15:52:16marnen_ enters the room.
15:54:57rueweepy: Have not read it yet but headius was just by a bit ago
15:55:15weepyah - quite in depth
15:55:36weepysuggests that rubinus performance might be a tough mountain to climb
16:04:57aquasync leaves the room.
16:05:08anteaya enters the room.
16:08:56marnen leaves the room.
16:09:56rueI dunno. The JVM does OK
16:13:37imajes enters the room.
16:19:41marnen_ leaves the room.
16:21:11octopod enters the room.
16:26:13chris2 leaves the room.
16:28:19rubbish enters the room.
16:28:46rubbish leaves the room.
16:30:56yugui enters the room.
16:32:35anteaya leaves the room.
16:32:53anteaya enters the room.
16:36:33GMFlash enters the room.
16:37:08trythil enters the room.
16:37:41MenTaLguY enters the room.
16:39:15hornbeck enters the room.
16:39:15Fullmoon enters the room.
16:39:49anteaya_ enters the room.
16:40:16anteaya_ leaves the room.
16:42:31KirinDav enters the room.
16:42:48Arjen_ enters the room.
16:48:45KirinDav leaves the room.
16:56:42thehcdreamer enters the room.
17:03:01rubuildius_ppc leaves the room.
17:03:04rubuildius_ppc enters the room.
17:04:02wycats leaves the room.
17:08:15KirinDav enters the room.
17:13:15jennyw enters the room.
17:20:18rubuildius_ppcAdam Wiggins: 4a1f39426; 2091 files, 6714 examples, 23612 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187625
17:26:58Defiler" and a good bit more Ruby code in the core libraries, making several of the core methods easier to understand, maintain, and implement in the first place."
17:27:06DefilerI find this sentence misleading
17:27:20Defiler'good bit more' than what? Who else even writes core libraries in Ruby?
17:28:13rueGood bit more than in the VM
17:28:26DefilerThat's not the context of the sentence, at least as I read it
17:31:42rueI do not see an object of comparison
17:32:11wycats enters the room.
17:32:13DefilerThe article is a comparison of implementations
17:32:26DefilerThe only thing that sentence can mean is 'a good bit more than' (the previous entries in the list)
17:33:17DefilerCuban food time. Back later. ;)
17:34:00yugui leaves the room.
17:34:00dblack leaves the room.
17:42:36EugZol enters the room.
17:42:46wycats leaves the room.
17:48:30marnen enters the room.
17:48:52octopod leaves the room.
17:52:54KirinDav leaves the room.
17:52:55marnen leaves the room.
17:53:28marnen enters the room.
17:55:54trythil leaves the room.
18:00:25smparke1 leaves the room.
18:01:24TheVoice enters the room.
18:01:45benburkert enters the room.
18:06:34defn enters the room.
18:07:12trythil enters the room.
18:08:07KirinDav enters the room.
18:08:39defn leaves the room.
18:09:10defn enters the room.
18:09:42rubuildius_ppc leaves the room.
18:09:45rubuildius_ppc enters the room.
18:11:00KirinDav_ enters the room.
18:12:39trythil leaves the room.
18:14:45marnen leaves the room.
18:25:58rubuildius_ppcAdam Wiggins: 4a1f39426; 2091 files, 6714 examples, 23612 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187653
18:26:48mentz leaves the room.
18:33:31KirinDav leaves the room.
18:38:20KirinDav_ leaves the room.
18:38:20weepy leaves the room.
18:39:24KirinDav_ enters the room.
18:41:23rubuildius_amd64 leaves the room.
18:41:28rubuildius_amd64 enters the room.
18:42:55yaroslav enters the room.
18:42:55rubuildius_amd64 leaves the room.
18:43:28headius enters the room.
18:46:06rubuildius_amd64 enters the room.
18:49:04mentz enters the room.
18:49:05rubuildius_amd64 leaves the room.
18:50:04joachimm leaves the room.
18:53:17joachimm enters the room.
18:58:35brapse enters the room.
19:13:05ctennis leaves the room.
19:26:40joachimm leaves the room.
19:27:07dctanner enters the room.
19:28:44imajes_ enters the room.
19:30:51Somebee enters the room.
19:32:07fbuilesvbrixen: ping
19:34:28brapse leaves the room.
19:35:40imajes leaves the room.
19:37:00VVSizfbuilesv: hi there :)
19:37:07fbuilesvVVSiz: hey, how's it going
19:37:57nicksieger enters the room.
19:38:21VVSizpretty good here. Just finished a new blog post on the value of RubySpecs :)
19:38:27fbuilesv:O
19:38:39dschn leaves the room.
19:38:39VVSizhttp://blog.emptyway.com/2008/04/27/the-value-of-the-rubyspecs/
19:38:49fbuilesvI was just discussing that with another student who also got in the GSoC. I think we should plan out a meeting to define our goals so we don't overlap efforts.
19:39:04fbuilesvmeeting sounds too serious
19:39:11fbuilesvmeetup in IRC or something :P
19:39:18headius#rubyspec
19:39:22VVSizyeah. I propose #rubyspec
19:39:25VVSizdamn :)
19:39:29headiusBAM
19:39:30headiusI win
19:40:00VVSizthat's a small little channel with those who interested in rubyspecs, including brixen, rue and us
19:40:04fbuilesvthat sounds good, now we only need to set up some schedule that works for everyone, I know there's someone from Germany, Italy and Japan so far working on the GSoC this year.
19:40:24fbuilesvhas brixen said anything about when's the project gonna go live on Github?
19:41:15VVSizmy understanding is that moving rubyspec into separate repo is under way, but I wouldn't really expect it to happen soon
19:42:03VVSizI'd say that if you don't have a commit bit to rubinius repo, it's a good first step. There is no need to be blocked by things like timeframe for rubyspec move to separate repo
19:42:27fbuilesvI have my commit bit already, and I know NoKarma does too, no idea about the other guys.
19:42:55headiusif there's no logging in #rubyspec there should be...then we can have running piecemeal meetings and catch up with the logs
19:42:57VVSizthen, it's all set! :) once the specs moved, we move to, but this won't block us
19:43:26VVSizheadius: agree, the logging there would be useful, the traffic is not that high, and reading the logs would be fast
19:44:22obvio171 enters the room.
19:44:23obvio171 leaves the room.
19:44:36obvio171 enters the room.
19:44:52VVSizheheh, so far, 73 folks contributed to the shared rubyspecs (spec/ruby/1.8)
19:45:05obvio leaves the room.
19:45:33benburkert leaves the room.
19:46:16headiusnice
19:46:32headiusI certainly agree with the loose commit bit for the specs
19:46:35headiusseems no reason not to do that
19:46:36VVSiz175 commits during the last 2 months
19:48:28yaroslav leaves the room.
19:49:00MaledictusVVSiz: can you also generate a graph to see the growth in the number of expectations? :)
19:49:36KirinDav_ leaves the room.
19:49:38VVSizMaledictus: that's tricky, especially since now there are empty tests/stubs that do nothing.
19:49:46Maledictusah ok
19:50:34Maledictushmm, but I think "expactations" only match actual .should? so not an empty it do ... end
19:51:03VVSizyeah, expectation numbers should be OK, but not examples numbers
19:51:16Maledictusyep
19:51:23VVSizfwiw, since 1 Jan 2008: 1851 files changed, 28093 insertions(+), 6865 deletions(-)
19:56:09obvio171 enters the room.
19:56:21ctennis enters the room.
19:56:39nicksieger leaves the room.
20:03:49obvio leaves the room.
20:04:37boyscout1 commit by Vladimir Sizikov
20:04:38boyscout * Added testcase for BigDecimal#infinite? for NaN.; 4b541ed
20:05:08headius leaves the room.
20:06:13Fullmoon leaves the room.
20:10:38boyscout1 commit by Vladimir Sizikov
20:10:39boyscout * More test cases for BigDecimal#finite? specs.; 80932d2
20:14:47obvio enters the room.
20:14:52benburkert enters the room.
20:14:53obvio171 leaves the room.
20:16:17joachimm enters the room.
20:23:59boyscout1 commit by Vladimir Sizikov
20:24:00boyscout * More checks for BigDecimal#abs specs.; 94ba088
20:26:25wycats enters the room.
20:33:30Somebee leaves the room.
20:33:38rubuildius_amd64 enters the room.
20:37:12imajes enters the room.
20:37:34dgtizedbrixen: you about?
20:38:32rubuildius_ppcVladimir Sizikov: 94ba0884c; 2091 files, 6715 examples, 23624 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187700
20:38:33rubuildius_ppcVladimir Sizikov: 80932d25c; 2091 files, 6715 examples, 23621 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187696
20:38:34rubuildius_ppcVladimir Sizikov: 4b541ed23; 2091 files, 6715 examples, 23613 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187695
20:39:15dgtizedbrixen: the DRb spec fix that you commited is not a fix, it's a hack. That spec demonstrates that we have a problem in our networking stack when cleaning up sockets, fix the problem not the spec
20:39:56wycatsoh snap
20:40:37wycatsyo Defiler
20:41:58VVSizdgtized, brixen: there is also a race condition in DRb itself that leads to cases when socket is not closed: http://jira.codehaus.org/browse/JRUBY-2347
20:42:21dgtizedVVSiz: then why doesn't it show up when running that spec on MRI?
20:44:13VVSizdifferent timings. race conditions are like that -- non-deterministic
20:44:31dgtizedVVSiz: also I just got the same error anyway
20:45:10VVSiz:) nice
20:45:52dgtizedanyway, I'm reverting that change because either way it's expected behavior
20:51:13octopod enters the room.
20:52:16imajes_ leaves the room.
20:54:50joachimm leaves the room.
20:56:04VVSizwell, you'll get the failures then, from time to time
20:57:16rueMeh mee
20:57:22VVSizsince in JRuby that test fails very reliably (like, 100% of cases), we just excluded it
20:59:04AndrewO enters the room.
20:59:59boyscout2 commits by Charles Comstock
21:00:00boyscout * DRb.start_service spec fails because of a timing bug in DRb; 5830380
21:00:01boyscout * Revert "Made DRb spec depend partially on PID so multiple runs don't clash."; 08695d9
21:00:24dgtizedI added a spec failure
21:00:51dgtizedThe fix doesn't work anyway because we don't run each spec instance in a different Process so the PID doesn't change anyway
21:01:13smparke1 enters the room.
21:01:52dgtizedAnyway, why is it no one is just submitting a patch to fix DRb if that's the source of the problem?
21:03:15VVSizI was hoping that MenTaLguY would take a look at it (in JRuby first) and then submit it to ruby-core ;)
21:05:43dgtizedyea that would be nice, maybe I'll take a look at it later, gotta go though
21:07:24rubuildius_amd64Charles Comstock: 583038089; 2091 files, 6711 examples, 23591 expectations, 0 failures, 0 errors; http://rafb.net/p/dtYNIc32.html
21:09:52headius enters the room.
21:13:10rubuildius_ppcCharles Comstock: 583038089; 2091 files, 6714 examples, 23619 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187710
21:14:58defn leaves the room.
21:15:33wycats_ enters the room.
21:19:37AndrewO leaves the room.
21:19:48octopod leaves the room.
21:28:16adamwiggins enters the room.
21:34:02yipstar enters the room.
21:34:26zenspiderheadius: oy...
21:34:45headiuswhat
21:35:03zenspiderheadius: there's no mkmf for jruby, right?
21:35:36VVSiznone
21:35:37headiusthere doesn't really need to be because we don't have native extensions...we've kicked around the idea of something equivalent for Java, but nothing right now
21:35:45zenspiderheadius: do you guys have a sqllite gem that works with jruby?
21:36:01headiusthere's one in progress that uses a third-party JDBC driver for SQLite
21:36:27headiusit has two modes, one using JNI and the native C library and another running a MIPS-compiled SQLite on a MIPS engine written in Java called NestedVM
21:36:51zenspiderk.
21:37:33wycats leaves the room.
21:37:56zenspiderlooks like we'll punt on that and write out yaml, have an mri that picks up the yaml and writes to sqlite
21:38:09headiusthis is for CI thing?
21:38:30zenspiderno, totally different project
21:38:33headiusok
21:39:15kw enters the room.
21:48:56NoKarma enters the room.
21:51:49jennyw leaves the room.
22:05:16Maledictus leaves the room.
22:10:46wycatszenspider: rbx can use do_sqlite
22:11:05wycatsit's a C extension but much cleaner than the other native gem
22:13:43adamwigginsWhat do you guys think about this: in MRI, defined? Object == 'constant', but in Rubinius, defined? Object == true
22:14:27adamwigginsObviously Rubinius' return value makes more sense, but is it intended that the 1.0 should replicate all behavior of MRI as exactly as possible, even when non-ideal?
22:22:05headiusdefined? is supposed to return what type of syntactic construct it is, if it's present
22:22:19headiusso you can differentiate between e.g. local variables and "vcalls"
22:22:20nicksieger enters the room.
22:22:47headius~/NetBeansProjects/jruby ➔ ruby -e "v = 1; p defined? v"
22:22:47headius"local-variable"
22:22:48headius~/NetBeansProjects/jruby ➔ ruby -e "def v; end; p defined? v"
22:22:48headius"method"
22:24:01olabini leaves the room.
22:29:18adamwigginsInteresting, thanks. Seems to me that this is a candidate for modification of the language spec in the future - I would expect any method ending with a ? to be boolean in nature.
22:29:35adamwigginsBut clearly current implementations need to stick to the existing definition, so I'll write up some specs for that.
22:30:18rby leaves the room.
22:31:34Rich_Morinadamwiggins: Should there be a way to tell rbx to be bug-for-bug compatible?
22:31:54Rich_MorinIf nothing else, that might be useful in testing.
22:32:08adamwigginsRich_Morin_: Hrm, well, you gotta be careful with the word "bug"
22:32:53adamwigginsI was thinking this was more of a design inconsistency, just one of those little things that rubs us aesthetic-sensitive Ruby types the wrong way. :)
22:33:08Rich_MorinHey, I spent several months documenting Leopard's variations from UNIX03 (and how to turn them off).
22:33:12adamwigginsBut yeah, it would be cool to flag things to be up for future review or something
22:37:29wycats_ leaves the room.
22:37:40headiushmm
22:37:46headiusI thought bigdecimal was going to be implemented all in ruby
22:38:02rubuildius_ppc leaves the room.
22:38:05rubuildius_ppc enters the room.
22:38:10hornbeck leaves the room.
22:38:10headiusoh, nevermind I think I see my problem
22:38:33hornbeck enters the room.
22:39:24Rich_MorinAccording to irb, "def a1;end" returns a value of nil. If someone writes a spec testing for that, rbx will fail that spec. Having a way to either turn of the spec or change the behavior seems appropriate and conducive to good process.
22:41:21headiusthat's another difference that bugs me
22:42:01headius~/NetBeansProjects/rubinius ➔ ruby -e "p(class Foo; def foo; end; end)"
22:42:01headiusnil
22:42:01headius~/NetBeansProjects/rubinius ➔ shotgun/rubinius -e "p(class Foo; def foo; end; end)"
22:42:01headius#<CompiledMethod:0x9 name=foo file=(eval)>
22:42:04headiusgoofy
22:43:07dgtizedoh that is kinda odd -- I don't think nil is right either, but one would expect to at least get the class back or something
22:43:52Rich_MorinWell, I DON'T want the ability to grab CM objects to go away. I'm having great fun looking at all the information they contain.
22:44:19dgtizedRich_Morin_: I agree, but it doesn't seem like the CM is the last return value
22:44:45dgtizedRich_Morin_: it seems like the class definition is the lass return value
22:45:04dgtizedRich_Morin_: so you should still get the def return value inside of the class, but I don't know about outside of the class
22:45:11dfg59 enters the room.
22:45:17Rich_Morintoo deep for me
22:45:42dgtizedRich_Morin_: class Foo; end => Foo makes sense to me
22:45:57dgtizedclass Foo; def foo; end; end => Foo also makes sense to me
22:46:51dgtizeddef foo; end => CompiledMethod(foo)
22:47:24benburkert leaves the room.
22:48:23Rich_MorinWell, there should prolly be a "whiteboard" somewhere to let folks say what they think Ruby behavior should be. Meanwhile, there needs to be a way to handle "minor" inconsistencies smoothly (IMHO :-.
22:52:34dgtizedevan: vm is missing the gen/ directory
22:54:40rubuildius_ppcCharles Comstock: 583038089; 2091 files, 6714 examples, 23619 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/187748
23:01:43RyanTM_ leaves the room.
23:02:51RyanTM_ enters the room.
23:04:17qwert666 leaves the room.
23:18:59aotearoa enters the room.
23:19:32Rich_MorinAnyone here familiar with the structure of CompiledMethod#lines? Like, why is it a tuple of tuples?
23:22:42headius_ enters the room.
23:22:42headius leaves the room.
23:22:58Rich_MorinI think I get it now. Each tuple defines the range of op codes for a given line in the Ruby code. Hmmmm.
23:24:42NoKarma leaves the room.
23:24:44benburkert enters the room.
23:25:34rueYes
23:25:54rueRich_Morin_: Have you read the VM documentation? http://rubini.us/doc
23:26:08benburkert_ enters the room.
23:26:10mentz_ enters the room.
23:26:10rueNeed to redo the source
23:27:17Rich_Morinthanks for the pointer
23:27:33mentz leaves the room.
23:29:46headius_ leaves the room.
23:29:56headius enters the room.
23:30:29obiejuan leaves the room.
23:31:04Rich_MorinWhere can I get info on Tuples - like, are they basically arrays, or what?
23:31:46imajes leaves the room.
23:32:31rueTuples are heterogenous, mutable, fixed-size containers
23:33:32Rich_Morinordered?
23:35:08headiusheh
23:35:11headiuspastie
23:35:27cremes__ leaves the room.
23:36:13pastiehttp://pastie.org/187769 by headius.
23:37:14headiusprobably about as much as I can do before providing hand-implemented primitives for each
23:37:29nkpart enters the room.
23:37:36headiusthe C code in them couldn't be mutated into valid Java because of the stack stuff
23:43:51radarek enters the room.
23:48:41agardiner enters the room.
23:49:23mentz_ leaves the room.
23:50:56benburkert leaves the room.
23:52:19dgtizedheadius: wait are you making a java vm for rubinius?
23:52:29headiusjust playing with the possibility
23:52:47headiusseems like if the primitives were implemented that's the bulk of it
23:53:01headiuseverything from there up is mostly ruby that depends on primitives
23:53:14dgtizedyea though clearly a bit more is needed in the cpp branch backend
23:54:13headiusyeah, I'm not sure how that's going to end up
23:54:20headiusthat's why this is mostly an experiment
23:54:43headiusof course the new C++ branch looks an awful lot like JRuby's structure, so it might map better to what we have
23:55:02dgtizedis that just cause it's classes, and some of the classes map better?
23:56:02headiuswell the structure, which has a C++ class for almost every Ruby core class, is basically like JRuby or IronRuby's structure
23:56:44headiusin theory getting this new VM in place will bring the layer between Ruby and native closer to matching JRuby
23:56:55headiusso potentially it would be easier to replace the C++ with JRuby/Java
23:59:52dfg59quick question: i'm ready to do my first commit after being granted commit rights. i've been having a conversation regarding the changes in a ticket. should i have a conversation with people in the channel before committing? is there anything else i need to do (besides running ci specs, rebasing, etc)?