Index

Show enters and exits. Hide enters and exits.

17:14:49logbot_dbl enters the room.
17:24:07headius leaves the room.
17:24:21joachimm leaves the room.
17:25:19SoreGums- leaves the room.
17:25:22SoreGums- enters the room.
17:38:11binary42 leaves the room.
17:40:58naeu enters the room.
17:42:22tongueroo enters the room.
18:03:36ryanlowe leaves the room.
18:09:09felgi enters the room.
18:16:45imajes enters the room.
18:18:20headius enters the room.
18:29:15libc_ leaves the room.
18:29:15kronos_vano leaves the room.
18:38:16libc_ enters the room.
18:38:23kronos_vano enters the room.
18:51:20mediogre leaves the room.
18:55:04naeu leaves the room.
18:56:54bitsweat enters the room.
19:11:17radarek enters the room.
19:11:25rueevan: http://blogs.sun.com/dave/entry/biased_locking_in_hotspot
19:13:05qbproger_ enters the room.
19:20:11headiusyay biased locking
19:23:42qbproger leaves the room.
19:31:39imajes leaves the room.
19:39:41tongueroo leaves the room.
19:53:31rueLooks like Laurent is going full LLVM
19:53:56headiusyes, I gathered that from mentions of a new vm
19:54:10slavawhat is laurent
19:54:21headiusLaurant = lrz = macruby creator
19:54:28headiusa/e
19:55:23brixenwould make sense, since llvm is under the same roof
19:55:44brixenrue: btw, evan linked to this last week http://wikis.sun.com/display/HotSpotInternals/Synchronization
19:56:44binary42 enters the room.
19:58:12rueYeah. The blog post is from 2006, seems there is little embellishment on the original
19:58:34boyscoutReworked Array#rassoc to stop the search as soon as possible. - 810e724 - Tilman Sauerbeck
19:58:34boyscoutTurned Array::Packer::BASE_64_B2A into an array. - e3f8e9f - Tilman Sauerbeck
19:59:18headiusI don't know if either of those mention lock coarsening but it's another important technique on hotspot
19:59:31headiususually depends on inlining though
20:00:19boyscoutCI: e3f8e9f success. 1427 files, 6934 examples, 23104 expectations, 0 failures, 0 errors
20:01:38brixenah ha, found another bug running mongrel
20:01:41brixenI'm so close
20:04:41joearnold enters the room.
20:15:16ruetilman: Hee, found that FIX:? :)
20:15:42tilmanwhy is lib/rbconfig.rb in .gitignore when it's under version control!?
20:15:54tilmanor ratehr: why is it in version control, when rake generates it?
20:16:00brixenevil people ?
20:16:05tilmanrue: the one for Array#rassoc? yes
20:17:29tilmanbrixen: who's right? the build system or .gitignore? :p
20:17:36brixenthe build system
20:17:41ruetilman: That is why it is not in the repo :P
20:17:46rueIt is *generated(
20:18:14brixenwell, yeah, it is a .rb file, so git ignores it
20:18:19brixenvia .gitignore
20:18:38brixenif it's in the repo, it should be removed
20:19:08tilmanit's in the repo. that's what i meant by saying 'under version control'
20:19:19brixenyeah, should be removed
20:19:21tilmanok
20:19:23rueYep, gotcha, cannot read
20:19:35brixenI realized what you were saying after
20:20:18tilmani might be misinterpreting git log's output
20:20:48rueDoes not look like it is
20:20:48evantilman: where do you see it under version control
20:20:54evanjust had the final inspection on the condo
20:20:57brixenit's not
20:21:02tilmani ran 'git log' on it, which gave me a log
20:21:02brixenI just checked
20:21:10brixenevan: sweet! when do you sign?
20:21:13tilmani think that used to only be the case for files under version control
20:21:14evanmonday!
20:21:18brixenawesome
20:21:19tilmanbut apparently it's not anymore
20:21:19evanmonday we sign at escrow
20:21:26rueevan: Congrats :) (And condolences)
20:21:31evanheh
20:21:37evanit's a good time to buy
20:21:44evanand we've managed to get some family help
20:21:49evanso it's affordable and possible
20:21:54ruetilman: Yeah, it shows any history there is for it. You can use ls-tree
20:22:15tilmanrue: i only knew ls-files, whose default behaviour is strange :D
20:22:34rueOh, yes, that too
20:22:47evanyeah, i guess even though rbconfig.rb has been deleted
20:22:56evangit log will still find the old commit info
20:23:37rue(As it should)
20:23:51brixenah h to the e ll, yeah
20:23:52rueA note about not existing currently might be nice, though
20:24:01brixenjust served through mongrel
20:24:33brixenhah, served the file list
20:24:35brixensaweeet!
20:24:54evanbrixen: WORD yo!
20:24:58rueCool
20:24:58brixenyep, I can just browse away on my file system
20:25:06headiushey, congrats
20:25:17brixenheadius: thanks!
20:25:32ruebrixen: Grab Rack and see if the handler works (or did you already?)
20:25:45evanhopefully rack works fine
20:25:50evanit being so simple.
20:25:55rueIt should
20:26:46tilmanrakelib/configure.rake does set CONFIG['DLEXT'] to 'bundle' atm (hard coded). Rubinius::LIBSUFFIX would be '.bundle' on darwin -- is the leading dot a problem? rakelib/configure.rake uses Rubinius::LIBSUFFIX earlier and only overwrites that DLEXT later
20:27:19brixenthe leading dot isn't a problem
20:27:22brixenbut the hardcoding is
20:27:30evanyeah, needs to be unhardcoded
20:27:30brixenfeel free to fix on linux
20:28:16tilmani don't have an extension handy to test build though
20:28:51boyscoutEnsure immediates are given global capi handles. - d0a476b - Brian Ford
20:29:06brixenwith that you should be able to: bin/rbx gem install mongrel
20:29:09brixenon os x
20:29:14rueActually, .so should work everywhere if it is just hardcoded
20:29:18brixenand run the example in the mongrel readme
20:29:36brixengrabbing lunch, bbl...
20:29:53rueMm, nice to not be on a permanent branch again
20:30:02evanyep.
20:30:26evani've started to use tabs in Macvim a little
20:30:37evani think i've figured out a good use pattern
20:30:43slavaemacs, evan
20:30:45slavaemacs
20:31:12evanmm
20:31:13evannah.
20:32:11rueevan: Really? I just use minibufexplorer
20:32:31evani used mbe for a while, i don't anymore
20:32:37evani use normal bufexplorer
20:32:42rueAnd windows on the few occasions I remember to when I have a screen big enough
20:32:49evanthat + splits is the normal pattern
20:32:50boyscoutCI: d0a476b success. 1427 files, 6934 examples, 23104 expectations, 0 failures, 0 errors
20:32:57evanbut i've finding it's nice to have a tab with it's own splits
20:33:08evanie, for when i'm working on say, 2 sets of files
20:33:10evanone tab per set
20:34:09rueMm, yeah, that would work
20:34:19rueBut I never remember to split anyway
20:35:59rueSemi-aside, it is somewhat pointless to build with ggdb if optimisations are enabled
20:36:33evantrue, but it's nice to have
20:36:39evanbecause it includes all the symbol names
20:36:45evanso that our internal backtrace generator works
20:36:55evanabit in a limited capacity, but it last it's something
20:39:29felgi leaves the room.
20:51:54imajes enters the room.
21:05:33brixenthe symbols were essential to finding that bug
21:05:38brixenthat's why I added them
21:08:25dbussinkbrixen: was just testing data objects gem, but i've got a slew of methods :)
21:09:09dbussinkhttp://gist.github.com/87206
21:13:34brixendbussink: cool!
21:13:37rueI seem to have an issue with gem not getting the latest_spec thingy
21:13:43brixenthat's not a slew btw :)
21:13:51dbussinkbrixen: did have a lot of timed errors before it got here
21:14:01ruebrixen: Yes, the -O2 is the offender in dev mode
21:14:02dbussinkand i had to use --no-rdoc because that failed for a gem
21:14:25brixenI have --no-ri and --no-rdoc in my .gemrc
21:14:34brixenwho needs rdoc? :P
21:15:01brixendbussink: do you mean timed errors installing the gem or running mongrel?
21:15:07dbussinkbrixen: during gem install
21:15:21brixenyeah, I saw a couple
21:15:32brixenI think that's rubygems repos tho
21:15:47dbussinkwell i don't see it that often with mri
21:16:01brixennot that often, but I get it
21:16:24dbussinkme too, but with rbx now it was 4 or 5 failures before a success
21:16:35imajes leaves the room.
21:16:40dbussinkthat's a failure rate way higher than mri usually has for me
21:17:16brixenso, what gem is that pastie from?
21:17:29dbussinkbrixen: btw, data_objects uses rb_thread_select to be non blocking on running queries, dunno how well that will work with rubinius
21:17:31dbussinkdo_postgres
21:18:07brixenoh, hm, I don't know anything about DO
21:18:16brixenit has it's own pg adapter?
21:18:24brixenwhy not use the regular one?
21:18:53dbussinkbecause data_objects actually has a unified api for database drivers
21:19:06dbussinkand stuff like the regular mysql driver is not very actively maintained
21:19:22dbussinkand data_objects also does typecasting etc. in the c part
21:19:34brixenI see
21:19:35dbussinkso a db date comes out a ruby date
21:19:55dbussinkwe would like to see it as perhaps a new standard
21:20:11brixenok, I'll take a look
21:20:11dbussinki know sequel already has a driver for it too, we're also looking for making one for activerecord
21:21:05brixenI'll probably get the mysql and pg one working first though
21:22:11dbussinkwell, that's your call of course :)
21:22:18dbussinki'm just biased because i maintain these :P
21:22:20brixenI'm loving virtualbox btw
21:22:34brixensure, but I need to get rails running :)
21:22:47brixennot that you aren't as important to me :D
21:22:53brixener.. nvm
21:23:03brixenI wasn't comparing you to rails :P
21:23:28headiusvirtualbox is super nice
21:23:32headiusand you can't beat the price
21:23:48brixenindeed
21:23:57brixenthey have significantly improved it
21:24:24brixenI briefly tried it pre-2.x
21:24:54headius64-bit guests helps a lot...since I really don't want to maintain separate boxes for 64-bit linux and windows
21:25:21headiusonce I requisition a stack of memory for our CI box I'm going to host several platforms on it
21:25:23brixenyeah, that is a huge win
21:26:34ruebrixen: Did you install fastthread?
21:28:19brixenyes
21:28:23brixenit's smart now
21:28:27brixenjust makes a stub
21:38:17headius leaves the room.
21:40:58rueMeh, Wireshark is being a bastard too.
21:50:09rueBut no meaningful issues in the gem install?
21:58:01rueHmh, restarting the WiFi seems to have cleared my initial error at least
22:02:01rueNope
22:30:40evandbussink: getting rb_thread_select working is easy
22:42:24rueNo luck, keeps "timing out"
22:45:53evanwhat ya looking at?
22:46:51rueNothing yet.. had it trying to install mongrel
22:50:09naeu enters the room.
22:57:37rueOMGz0r
22:59:34rueRe-enabling sysread made quite a difference. Pushing in a moment
23:00:47tongueroo enters the room.
23:00:53evanre-enabling?
23:02:04naeu leaves the room.
23:02:23radarek leaves the room.
23:02:41rueI disabled it in Net whenever it was I got it working the first time in the fall, since sysread was not implemented
23:02:56evanah.
23:03:02evangotcha
23:03:08evani'm implementing rb_thread_select right now
23:03:09evanit's trivial
23:03:15evanjust writing a few specs for it
23:03:52boyscoutFixed VMMethod comment. - 16be65b - Eero Saynatkari
23:03:52boyscoutFix VariableScope to stackfulness (for .local_variables at least.) - 6ba2778 - Eero Saynatkari
23:03:52boyscoutUse IO#sysread again in Net. - e0d95f8 - Eero Saynatkari
23:04:06ruebrixen: ^^ Made a crazy difference in gems here
23:04:17evanspeed install wise?
23:05:33SoreGums- leaves the room.
23:05:39SoreGums- enters the room.
23:06:16rueirb completion works again too
23:07:12evanyou mixed real changes with indentation changes
23:07:13evan:(
23:08:06evandid you change completion.rb at all?
23:08:08evani can't tell
23:08:57rueOh, hm, I split it on stackfull. No, it is unchanged
23:09:14boyscoutCI: e0d95f8 success. 1427 files, 6935 examples, 23105 expectations, 0 failures, 0 errors
23:09:31rue`time bin/rbx gem install --no-ri --no-rdoc -r mongrel` is 13 secs
23:09:37evanso it's just an indent change?
23:09:44evanreally?
23:09:45evanwoo!
23:09:46evantries.
23:10:10rueevan: Yeah, the completion proc was completely unreadable, so the change itself is good
23:10:23rueForgot to split it from the rest on master, sorry
23:10:23evanah
23:10:27evanno prob
23:10:34evanjust makes it hard to figure it out
23:10:44rueIronically enough :)
23:10:46evanditto with indenting execute_specialized
23:12:53evanwow
23:12:55evanit was pretty fast
23:12:59evannot 13 seconds
23:13:04rueYeah, I should have re-checked the commit. The only change in execute_specialized is passing the CompiledMethod to prepare
23:13:17evanbecause i didn't have rubygems already compiled
23:13:22rueI did run off rbcs, yeah
23:18:32tongueroo leaves the room.
23:26:25SoreGums- leaves the room.
23:26:42SoreGums- enters the room.
23:33:25boyscoutAdd rb_thread_select and specs - 8124344 - Evan Phoenix
23:37:08brynary leaves the room.
23:37:54boyscoutCI: 8124344 success. 1427 files, 6935 examples, 23105 expectations, 0 failures, 0 errors
23:44:39krawek leaves the room.
23:46:55rueSeems Autoload still has some issues