Index

Show enters and exits. Hide enters and exits.

00:08:49vborjahey guys, checkout this cool t-shirt. it shows what is Typhon all about: snakes on rbx-head http://twinsrpnt.com/blog/?p=180
00:08:51vborja^^
00:19:48brixenvborja_: nice
00:19:50brixenand freaky
00:20:12rueHm
00:20:39rueThe rbx logo ought be a dia de los muertos skull with rubies in the sockets
00:37:34evani'm back.
00:38:15brixenwb evan
00:38:25brixenevan: the website does not appear to be auto-updating
00:38:35evanit's just serving what is in web/_site
00:38:41evanit is not running jekyll
00:38:49brixencremes pushed some stuff, I see the file in _site
00:39:00evandid boyscout report on it?
00:39:03brixenbut it's not displaying when I access rubini.us or dev.rubini.us
00:39:50brixenhrm, not sure
00:40:58evan...
00:41:00brixenit's here https://github.com/evanphx/rubinius/blob/master/web/_site/doc/en/tools/memory-analysis/index.html
00:41:39brixener wait, it's not
00:41:40brixen:/
00:42:16evanexactyl.
00:42:21evanthat looks empty to me.
00:42:26evanwell, "empty"
00:43:02brixenyeah
00:43:12brixenI thought he regen'd it when he asked me about it
00:43:12evanalso
00:43:15evanwtf is PeerPong.com
00:43:27evanand why the fuck is repeating a tweet over and over
00:43:39brixenthere's a bunch of stupid tweet bots
00:44:09brixenit seems to pick up Lazy Twitter tweets
00:44:31brixenor lazyweb
00:44:38evanbut it just repeats them
00:44:49brixenyep, stupid
00:44:50evantargetting people
00:45:03brixenthis is the original tweet: 'Dear lazyweb: is there a switch to redirect Rubinius bytecode ".rbc"s into a separate tree? If not, want one?'
00:45:07brixenabout .rbc files
00:46:02evanyeah
00:46:05evani found it
00:46:06mahargI report those accounts as spam
00:46:18evanit's been repeating dozens of times
00:46:32mahargvborja: so you know, I'm stormbrew
00:47:24jakedouglasi would like to work on storing the compiled code in a db…unfortunately time is scarce these days :(
00:47:42evani've some ideas
00:47:49evani'll put it on the "near term" list.
00:48:22mahargthe problem with storing them elsewhere is it's really easy as is to get rid of stale files, they usually go away withe the original file when you delete it
00:48:36mahargbut I definitely would like to not have .rbc files all over the place
00:49:23evanmaharg: yep
00:49:26evanthats why we haven't change it
00:49:35maharg.rbx/file.rbc would probably do well enough for me
00:49:43evanpeople have certainly lightly complained about it
00:49:53evanbut there have not been any ideas to change it
00:49:55evanbrixen and I have some
00:50:00evanI haven't implemented it yet.
00:50:16jakedouglasisn't a db the ideal solution, provided there was time to do it?
00:50:32brixenyeah, db is the best
00:50:37evana db is not a solution
00:50:44evana db is a mapping of keys to values
00:50:47mahargwhat kind of db? Would kind of suck if you needed sqlite to run rubinius
00:50:48evanwhat the keys are
00:50:50evanand what the values are
00:50:53evanis the important part.
00:50:54brixena different tree is a total non-starter for me because there's no cleaning up the files
00:51:11jakedouglasevan: sorry, i don't understand. i'm pretty sure you know what i'm talking about
00:51:18jakedouglasevan: do i need to be more explicit?
00:51:24mahargit's kind of too bad resource forks never caught on as a concept
00:51:36brixenevan: how do you mean "a db is not a solution" ?
00:51:38mahargfile.rb/rbx-compiled heh
00:52:57jakedouglasthat's like saying
00:53:13brixenI'm using "db" extremely generically
00:53:34brixenit could be a single flat file concatenation of .rbc's with an index
00:53:44jakedouglas"a ruby vm is not a solution, the code fed into it and the result of executing it are the important parts"
00:53:57jakedouglasor something
00:58:36evanwhat I mean is that the statement "just put them in a database" is to avoid most of the difficult questions
00:58:42evan1) where is the database?
00:58:46evan2) is there more than one?
00:58:52evan3) what if you can't read the database?
00:59:04evan4) how are the rbcs stored in the database?
00:59:10evan5) how do you access a stored rbc?
00:59:19evanthats all i meant before.
00:59:45jakedouglasare you suggesting that i'm trying to avoid those questions?
00:59:45jakedouglasor that none of them occurred to me or something?
00:59:45jakedouglasrelax
01:00:11evanchillaxes
01:00:17evanjakedouglas: didn't mean to be short
01:00:26evanjust getting your creative juices flowing
01:00:42Zoxcdrinks all the creative juice
01:00:46evanby trying to say that deciding to use a database is the first of many steps
01:00:49evanrather than a final step.
01:02:17evaneverybody chill out. it's the holidays!
01:02:23evandoes the "it's the holidays!" dance
01:02:28jakedouglasright - it's not like the software that other people build all day is somehow a one step process or does not require exploration. i hope that i'm at least somewhat familiar with the process by now.
01:02:35rueCreative juices: tequila and a punch in the face
01:03:51evangives jakedouglas a hug
01:04:02evanbreaks out the cookies
01:04:52jdsiegelall this rbc talk has reminded me I need to do some banking. *shakes fist at Royal Bank of Canada*
01:05:57jakedouglasi'll bring a 25 page proposal in PDF format next time
01:06:43evanjakedouglas: i'm sorry.
01:06:51evani didn't mean to sound like you were a dummy.
01:06:54evanyou're not.
01:06:57evancool?
01:07:12evancurses IRC
01:07:23jakedouglasyea whatever i just need to cool off
01:07:28jakedouglasttyl
01:08:07evancurse text only communication.
01:08:31evanhuman communication is too rich to be confined to such a lossy channel.
01:08:40rueHaha, puny humans.
01:23:55evanbrixen: are you going to commit an updated _site?
01:33:15mahargso what exactly is the call custom protocol (re. call_custom) -- it doesn't seem to be explained in the docs so far and the code for the instruction itself doesn't really make its purpose obvious (that I saw)
01:35:21evanmaharg: it's still an experiment
01:35:36evanbut the idea is the ability to have programmable method dispatch
01:35:47evanlike the JVMs invokedynamic
01:36:32evanI need to do more work on it
01:37:33mahargpython's method dispatch is kind of insane. a.blah is actually a.__class__.__dict__['blah'].__get__(a, type(a))()
01:37:39mahargwould it help with that? :P
01:38:09evansure, lets see.
01:38:52evanpart of the idea is the ability to do caching
01:39:07evanwithout out caching though
01:39:23evanyou'd call blah on a with call_custom
01:40:08evanthen, you'd setup CallUnits that call a method to find the method for :blah
01:40:20evanthen one to invoke the return value
01:41:17evanthe CallUnit for a method invokes the method as a static invocation
01:44:43rueCallOfCthulhuUnit
01:46:56evanCallUnit is a crappy name honestly.
01:46:59mahargI'm not sure how much you could cache that information, tbh. Even if you could assume the descriptor (the thing __get__ is called on) for the method stayed the same between invocations, you couldn't necessarily rely on __get__ being invariant
01:47:14mahargbut I haven't put a lot of thought into it yet
01:48:21evanwell
01:48:55evanbecause some of the invocations would be static rather than using ruby's dynamic dispatch it might at least still speed it up
01:49:32evanbut I really need to work on it and think about this situation as an example
01:54:45brixenthere has got to be a better way to do --auto in jekyll
01:54:52brixenthis is retarded
01:55:13brixenit's like regen'ing the entire site when I change the stylesheet or something
01:55:19brixen30 sec of 100% cpu
02:06:45brixenhah, totally chuckling over jbarnette's commit
02:06:56brixenthat must have been driving him totally crazy
02:27:00boyscoutMerged atom feed changes from weppos. - 8f2d35f - Brian Ford
02:27:00boyscoutBlog post index and explicit feed link. - a76eb9a - Brian Ford
02:27:00boyscoutDon't load twitter feed JS unless displaying a feed. - 6f83499 - Brian Ford
02:27:00boyscoutRegenerate the site. - 420727d - Brian Ford
02:38:52boyscoutCI: rubinius: 420727d successful: 3523 files, 15835 examples, 44084 expectations, 0 failures, 0 errors
03:58:45boyscoutString#tr treats descending ranges differently between from and to. - 0fdcbd5 - Matthew Draper
03:58:45boyscoutDescending ranges in String#tr's to_str aren't empty; fixes #620. - aa67709 - Matthew Draper
04:10:41boyscoutCI: rubinius: aa67709 successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
05:19:39boyscoutAdded Disqus commenting to blog and docs. - 40c5f27 - Brian Ford
05:19:39boyscoutSome docs on communicating and purpose of Ruby topic. - 5da3f00 - Brian Ford
05:19:39boyscoutRegenerate site. - ce94c7d - Brian Ford
05:23:29boyscoutFixed typo. - 3222754 - Brian Ford
05:23:36brixendamn typos
05:31:30boyscoutCI: rubinius: ce94c7d successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
05:55:17boyscoutBump to dev version - 6208c85 - Evan Phoenix (multiverse)
05:55:17boyscout- step 1 of bootstrapping spanish transalation done. - b148614 - Gaston Ramos (multiverse)
05:55:17boyscout- step 2, 3, 4 and 5 of bootstrapping spanish transalation done. - e116028 - Gaston Ramos (multiverse)
05:55:17boyscout- step 6 and 7 of bootstrapping spanish transalation done. - beb5ac0 - Gaston Ramos (multiverse)
05:55:17boyscout- bootstrapping spanish transalation done. - 1c73dff - Gaston Ramos (multiverse)
05:55:17boyscout- some last semantic fixes - 1956de5 - Gaston Ramos (multiverse)
05:55:17boyscoutMerge branch 'master' of git://github.com/evanphx/rubinius - f1ada1f - Victor Hugo Borja (multiverse)
05:55:18boyscout34 more commits
06:07:06boyscoutCI: rubinius: 3222754 successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
09:55:19hosiawakI'm getting a weird issue when trying to build rbx-head using RVM https://gist.github.com/752789
09:55:52hosiawakit's trying to use 1.8.6 to install rbx and for some reason Date.parse fails in RVM's 1.8.6
09:56:07hosiawakis 1.8.6 required to build rubinius or could I use 1.8.7 ?
10:01:16dbussinkhosiawak: 1.8.7 or 1.9.2 should both work
10:01:26dbussinkhosiawak: no need for 1.8.6, dunno if that still works even
10:02:01dbussinkhosiawak: you should just be able to do rvm install rbx
10:02:09dbussinkhosiawak: please make sure you have a recent rvm version though
10:03:27hosiawakdbussink: yeh, manual install with ./configure and rake using system's 1.8.7 works fine so I'll take a closer look at rvm
10:07:50fbuilesvhosiawak: had a similar issue yesterday, removing RVM and reinstalling fixed the issue (although I was using a really old version)
10:08:28fbuilesva really old version of 1.8.6 I should clarify.
10:09:20hosiawakfbuilesv: yeah, I just ran rvm get head and now rvm install rbx-head is compiling happily, thanks
10:11:30hosiawakoh, it just died when trying to compile oniguruma with: libtool: Version mismatch error. This is libtool 2.2.4, but the definition of this LT_INIT comes from libtool 2.2.10. You should recreate aclocal.m4 with macros from libtool 2.2.4 and run autoconf again - Gentoo fun :)
10:40:06dohzyahi
10:40:31dohzyahas someone already started a French translation of the doc ?
10:41:09dohzya(I'm a better English reader than writer, I promess :p)
10:42:25boyscoutTranslating memory_analysis.markdown to spanish. - 1241c60 - Federico Builes
10:42:25boyscoutMinor improvements to spanish version of memory-analysis - 2e0be01 - Victor Hugo Borja
10:47:02vborjadohzya: no one afaik, feel free to start it ! :)
10:47:23dohzyaalready done :-)
10:48:51fbuilesvvborja: thanks for the improvements :)
10:49:11vborjafbuilesv: thanks to you for submitting the patch ;)
12:23:21raluxgazaI must be doing something very wrong
12:23:38raluxgazaI can't seem to view my changes to the doc on my box
12:23:52raluxgazai run the command 'rbx -S jekyll --server --auto'
12:24:00dbussinkraluxgaza: in the right directory?
12:24:08raluxgazayup
12:24:10dbussinkraluxgaza: do you see files being modified / added etc.?
12:24:44raluxgazadbussink: yes i do
12:24:57raluxgazaif you are referring to git that is
12:25:27raluxgazai am getting a rubinius crash error
12:25:49raluxgaza# signal SIGSEGV
12:26:17raluxgazai have rubinius added to my list of rubies in rvm
12:26:28raluxgazausing rbx as my ruby vm
12:26:34raluxgazaand i get that crash report
12:28:24dbussinkraluxgaza: if you're working on the docs, it's probably better not to use rvm
12:28:49dbussinkraluxgaza: if you can try without using rvm and see if you still have the issue?
12:28:50raluxgazaokay
12:29:08dbussinkotherwise we can't know if it's rvm doing something or really an issue with rbx
12:30:20raluxgazaaright, let me try that
12:31:43raluxgazaAn exception occurred running ~/.rvm/gems/ruby-1.9.2-p0/bin/jekyll
12:31:44raluxgaza undefined method `kqueue' on EventMachine (Module) (NoMethodError)
12:32:00raluxgazais the error i get without rvm now
12:32:36raluxgazaalthough i see that the gem is stored under rvm, i must have configured it wrongly dbussink
12:32:57dbussinkraluxgaza: i think it's pretty confused about the gems etc.
12:33:07dbussinkraluxgaza: please try with disabling rvm completely
12:33:12dbussinkraluxgaza: rvm use system
12:33:16dbussinkraluxgaza: then build rbx
12:33:21dbussinkand follow the instructions
12:33:28raluxgazaokay, thank dbussink
12:59:55raluxgazadbussink: that works, thanks
13:01:12dbussinkraluxgaza: the problem is that when you have an rvm ruby active, is that there is a GEM_PATH and a bunch of other environment variables set
13:01:32dbussinkraluxgaza: but that means rbx picks up those variables too and tries to run native extensions compiled for for example 1.9.2 inside rbx
13:01:36dbussinkwhich of course blows up
13:02:01raluxgazaokay, so that means when working on rbx, rvm should be disabled
13:02:17raluxgazathanks again
13:05:56dbussinkraluxgaza: it will make your life a lot easier ;)
14:28:59raluxgazahow do I commit to the docs, will that be in my fork?
14:51:59dbussinkraluxgaza: if you don't have commit access to the main repo, then you commit to your fork and send a pull request
14:52:51raluxgazadbussink: thanks for answering all my noob questions, my first contribution to an oss project
14:54:33dbussinkraluxgaza: no problem, everyone needs to learn at first :)
14:55:30raluxgazadbussink: this is one reason I love the ruby community! ;)
17:03:37brixenraluxgaza: hi
17:03:50brixener bye
17:13:41evanMORNING
17:13:49brixenMORNING
17:14:45evanhow's everyone doing this morning
17:14:51evanI see i need to give fbuilesv a commit bit
17:14:52evan:)
17:15:18brixenyes, fbuilesv needs his commit bit back
17:15:31evanhm, what is his github user name?
17:15:40brixenfbuilesv: ^^^
17:15:51brixenI thought fbuilesv was his github user
17:15:55brixenlet's see...
17:16:03evanfound it.
17:16:12evanfebuiles
17:16:16brixenyep
17:16:53rueCold as all hell
17:16:59brixenum...
17:17:03brixenyou are in the wrong hell
17:17:07brixenit's supposed to be ho
17:17:09brixent
17:17:10brixen:)
17:17:12rueI think this is a special one
17:17:34evanyou're own private Hell.
17:18:28rue"Feels like" temp is -22ºF :/
17:18:58evanzoinks.
17:19:05evanthats some fucking weather right there.
17:19:20evanthats "be careful not to breath the air" weather
17:24:27hosiawakhey guys, I've translated Getting Started to Polish, do I need to commit web/_site as well or just web/doc ?
17:25:02hosiawaknever used Jekyll, guessing _site will get regenerated anyway ?
17:25:45brixenhosiawak: I'm just about to push better howto - write documentation
17:25:58brixenhosiawak: short answer is, we need the generated files in _site committed
17:26:31brixenhosiawak: we don't run jekyll on the server and we need those docs available to people that clone and can't or haven't built rbx yet
17:27:07hosiawakbrixen: right, ok
17:27:34brixenhosiawak: btw, thanks for working on a polish translation!
17:28:33hosiawakbrixen: np. it's a pleasure actually (not having to code for a while :)
17:28:53hosiawakbrixen: I should have some time during the break so expect more Polish translation
17:29:04brixenawesome!
17:29:09brixenhosiawak: do you live in Poland?
17:29:11hosiawakbrixen: I'm learning Rubinius btw
17:29:17brixenyay
17:29:27brixenthat's my hope, the docs help more people learn about it
17:29:43brixenhosiawak: I'm curious what the Ruby community is like in Poland
17:29:47hosiawakbrixen: yeah, I live in the UK and Poland
17:29:52brixencool
17:29:55hosiawakbrixen: now in Polan
17:30:00rueYou must be a very large man!
17:30:08brixenare most folks into Rails, or doing other stuff in Ruby?
17:30:08hosiawakrue: hehe
17:30:11brixenhosiawak: ^^
17:30:32boyscoutImproved howto - write documentation docs. - 05e6519 - Brian Ford
17:30:33boyscoutRegenerate site. - f04ddb3 - Brian Ford
17:31:01hosiawakbrixen: the community is quite active I think
17:31:13brixenevan: do you have spell checking on in gvim?
17:33:30hosiawakbrixen: https://github.com/hosiawak/rubinius - Polish docs started - do I need to send a pull req via github ?
17:34:53brixenhosiawak: you can, I'll take a look now
17:36:52hosiawakbrixen: I didn't add "pl" to the nav yet but if you think it's acceptable to publish please let me know when it's up there, I'll ask folks to contribute, it'll be faster with crowdsourcing :)
17:37:19brixenhosiawak: awesome, will do
17:38:36brixenok rubini.us, why aren't you updating?
17:41:03gramosbrixen: is there a way to figure out when a translated doc is out of date?
17:41:52rueI thought of that yesterday, a version of some sort
17:42:09brixengramos: I'll add a script for that
17:42:35brixengramos: it's really easy, the mtime on the translated doc is earlier than the source en doc
17:43:08brixenbut I'm thinking of inserting an attribute, like we have 'review: true' to signal that on a page when viewed
17:43:38ruePerhaps it would be good to auto-tag the doc page, a little box "Has not been checked for most recent doc updates"
17:43:50brixenrue: yes
17:43:55brixenthat's what I meant
17:43:55evanbrixen: still not updating?
17:44:00rueYep, that works
17:44:07brixenevan: let me check again
17:44:15brixenI'm sure I pushed the content this time
17:44:22rueCache?
17:44:29brixenevan: http://rubini.us/doc/en/how-to/write-documentation/
17:44:35sferikbrixen, evan: any update on the bug I reported a couple days ago?
17:44:37evanI see CI running
17:44:52brixenevan: that page should have numbered steps under adding and editing sections
17:45:10brixenrue: other than telling the browser to refresh, I don't know
17:45:12rueI see 6 steps at bottom
17:45:19brixenyeah, that's the old version
17:45:39evanwell
17:45:39brixensferik: no chance to work on it yesterday, I promise I will today
17:45:41evanon the server
17:45:44evani see some steps
17:45:55sferikbrixen: okay, ping me if you have any questions along the way
17:45:57brixenevan: there should be 2 sets of steps, not just one
17:46:04brixensferik: thanks, will do
17:46:08sferikbrixen: or any updates :)
17:46:14brixensferik: heh, ok
17:46:35evanbrixen: ok, there is not 2 lists in _site on elle
17:46:44brixenweird
17:47:02evanoh
17:47:03evani know wh.
17:47:05evanwhy.
17:47:08brixenit updated last night because the pages have comments now
17:47:20evanit's because you did commits in quick succession
17:47:25evanthe first kicked off CI
17:47:27evanwhich did the update
17:47:31evanthe later ones were ignored.
17:47:34evanbecause CI was already running.
17:47:36brixenhmm
17:47:45brixenI committed everything at once
17:47:53brixenit was 2 commits, but one push
17:48:07evanbut there were earlier ones
17:48:12vborjayay someone is doing polish translation of docs.
17:48:13evanscroll up and see when the CI annoucements were
17:48:43brixenI only see my commits
17:48:54evanoh, maybe it's hung.
17:49:18evanyeah, it's hung.
17:49:27evankilled it.
17:49:30brixenpoor CI, RIP
17:50:15evanweird
17:50:18evanelle is acting oddly.
17:51:11evannow it says it's compiling...
17:52:48evanok, killed cijoe and restarted
17:53:14evanbrixen: check the URL now.
17:53:17brixenk
17:53:27brixenyep, all good
17:53:39evanok
17:53:48evani'll need to set something else up
17:53:56evanto checkout the latest web/_site
17:54:06evanso it isn't tied to CI
17:54:19brixenthat would be cool, yeah
17:54:34brixenhosiawak: are you @hosiawak on twitter?
18:01:13hosiawakbrixen: yep
18:01:35brixenhosiawak: ok, cool
18:01:46brixenhosiawak: I'm merging stuff in now, will push in a bit
18:01:51brixenhosiawak: thanks for the help!
18:02:19hosiawakbrixen: cheers
18:03:19rueShould I just "stake ground" and toss in a translated index with everything else WIP?
18:04:07brixenrue: you certainly can, yes
18:04:23brixenrue: if you could hit getting started and what is Rubinius, that would be super helpful
18:04:34brixensince that's the foot in the door for folks
18:04:55rueYep. The index is the first translation hurdle
18:07:29brixenhosiawak: do you have a sec to give me translations for: Languages, Previous, Next, Up ?
18:08:35brixenrue: if you don't have time to translate the whole ToC, that's ok
18:08:46brixenrue: having what is rbx and getting started would be more useful
18:09:37boyscoutCI: rubinius: f04ddb3 successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
18:09:44hosiawakbrixen: Languages - Język , Previous - Wstecz , Up - Do góry , Next - Dalej
18:10:59brixenhosiawak: sweet, thanks
18:11:28shtirlicwhat is the normal compilation time of rubinius on something like MBP (~15 min?) ?
18:11:34hosiawakbbl
18:11:46brixenshtirlic: should be ~5 min unless you are building llvm
18:12:11brixenshtirlic: could you run ./configure --show and gist that for me?
18:13:01shtirlicbrixen: installing new version via rvm upgrade rbx-1.2.0 rbx-1.1.0
18:13:21brixenshtirlic: what version of rvm do you have?
18:14:19shtirlicbrixen: 1.1.6
18:15:10shtirlicbrixen: but i did 'rvm cleanup all' before, maybe this is the problem
18:18:06evanshtirlic: are you on 10.6?
18:19:11brixenyeah, sounds like it's building llvm
18:26:47boyscouttranslated Getting Started to Polish - a12414f - Karol Hosiawa
18:26:47boyscoutIntegrate Polish translation. - b3b68f2 - Brian Ford
18:26:47boyscoutRegenerate site. - d3f70bc - Brian Ford
18:38:38boyscoutCI: rubinius: d3f70bc successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
18:55:11shtirlicbrixen: sorry, I was dropped what was you last message?
18:56:20brixenshtirlic: evan was asking if you are on 10.6
18:56:37brixenshtirlic: I'm curious if you could look at the build logs for rbx and see if it's building llvm
18:56:58brixenshtirlic: you *should* be getting a prebuilt on OS X, so we need to figure out what's happening there
18:57:10brixena prebuilt llvm, that is
18:58:58shtirlicbrixen: yeh on snow and with latest upodate fox xcode
19:00:23brixenshtirlic: could you look in ~/.rvm/log/rbx-head/rake.log and see if rvm is building llvm?
19:00:52brixenshtirlic: or look in ~/.rvm/log/rbx-head/configure.log and see if the prebuilt was used successfully
19:01:04brixeninstead of downloading the llvm src tar
19:02:40shtirlicbrixen: https://gist.github.com/5f0f66da895de1020ea7
19:02:46shtirlicbrixen: of rake.log
19:03:07brixenshtirlic: erg, ok, it's not getting a prebuilt
19:03:15brixenshtirlic: could you pls gist configure.log
19:05:15evanwhat?!
19:05:16evanit's not?!
19:05:43shtirlicbrixen: https://gist.github.com/2ad14349e41c28d22f2b config.log
19:06:17evanwait
19:06:24evanwhat MBP is this?
19:06:30evanyou're on 32bit 10.6
19:06:38evanso it must be a very old MBP.
19:06:38shtirlicevan: it's virtual snow leopard
19:06:46evanhuh?
19:06:48evanthere is such a thing?
19:06:52shtirlicevan: yeeh clean for testing
19:07:06evanwell, thats why
19:07:14evanwe've never had anyone using 32bit 10.6
19:07:21evanyou'll have to build LLVM
19:07:33evanit takes about 30 mins on a real machine
19:07:36evanvirtually it could take a while.
19:07:42shtirlicevan: but prev version of rubinius1.1 was much fast on this box
19:07:44evanbut we can have you package it up and we'll host it
19:07:51evanso you don't have to do it again
19:08:02evanshtirlic: we upgraded LLVM
19:08:10evanmaybe there was a prebuilt for 32bit 10.6 before
19:08:18shtirlicevan: ah i see
19:08:25evani'm checking..
19:08:40evani see a 32bit version for 10.5
19:08:52brixenshtirlic: what are you using to run 10.6 virtually?
19:08:53evanbut we've never had a 32bit version for 10.6
19:09:07evanok, i have to take care of a few HOA things
19:09:08shtirlicbrixen: vbox4
19:09:10evani'll brb.
19:09:21evanshtirlic: i didn't know that was even possible.
19:09:34evanthats a configuration we've never seen before certainly.
19:10:03shtirlicevan: yeh waiting for new mbp, old have broken screen((
19:10:39evanif you build the 32bit LLVM
19:10:42evanwe'll host it
19:10:45evanso you don't have to do this again.
19:10:52shtirlicevan: ok
19:10:56evanoh, you're using rvm though
19:11:03evanso you'll have to dig in to create the prebuild
19:11:25shtirlicevan: they clean it after buld process?
19:12:03evanit's just in a weird directory
19:13:11brixenshtirlic: you should probably find it in ~/.rvm/src/rbx-head
19:13:26brixenwe have a package:llvm task for rake
19:13:44brixenso try rake package:llvm in that dir
19:17:55shtirlicbrixen: brixen hm it can't run on 1.9.2, but after switch to 1.8.7 it stops on can't run perl script ...
19:18:28shtirlicbrixen: it think i should do after succefull build
19:19:18brixenshtirlic: did the build not finish?
19:19:23shtirlicbrixen: yep;)
19:19:34brixenok, let's see
19:21:25brixenshtirlic: do this: cd to ~/.rvm/src/rbx-head; ./configure; rake -q package:llvm
19:22:33shtirlicbrixen: it's ok to do it, while building in progress?
19:25:42brixenshtirlic: well, wait for it to finish
19:27:10shtirlicbrixen: ok, why llvm is so huge to build?
19:29:14brixenit's just a big project
19:29:42brixenit's a whole compiler library with IR, optimizations, instruction generation
19:29:49brixenvery complex stuff
19:30:08brixenand gcc is slow :)
19:30:14shtirlicbrixen: for windows you have also prebuilt libs?
19:30:22brixennot yet, but we will
19:31:47FobaxWhen are you going to switch to compiling with LLVM?
19:32:17brixenFobax: well, you can already set configure --cc and --cxx to clang, clang++ respectively
19:32:25brixenbut there are a few issues last time I tried that
19:32:49brixenI expect this year many folks will start switching to using clang instead of gcc
19:33:39shtirlicbrixen: for speed or for new c++ features?
19:34:00brixenboth
19:34:24Fobaxwhat c++ features does it support that g++ goes?
19:34:28Fobax*doesn't?
19:34:53brixendunno :)
19:35:13brixenit supports incredibly better error messages for template compilation failures
19:35:23mahargI was under the impression that its C++ implementation was, if anything, kind of incomplete
19:35:23brixenlike, unbelievably better
19:35:40brixenmaharg: huh?
19:35:48brixenyou mean, in like 2005 or something?
19:36:02mahargdid clang even exist in 2005?
19:36:11brixenllvm compiles itself now afaik
19:36:23brixenI mean, clang++ can compile llvm
19:36:31Fobaxyeah, it definitely can compile itself
19:36:45brixenthat's not, if anything, kind of incomplete
19:36:47brixenfo sho
19:37:09boyscoutFixed display of comment counts. - 694e047 - Brian Ford
19:37:10Fobaxwell, LLVM doesn't necessarily use any of the newer features
19:37:20dbussinkbrixen: didn't really have an issue last time i tried
19:37:22brixenclang++ will compile all of rbx
19:37:26dbussinkbrixen: although libffi can't be compiled with clang (yet)
19:37:42dbussinkbecause it uses some gcc specific inline assembly stuff
19:37:45brixendbussink: I had an issue compiling a C ext
19:38:01dbussinkhmm, haven't seen that
19:38:06shtirlicbrixen: yeeh it's done as evan said 30minutes, trying to package
19:38:14brixenshtirlic: ok, cool
19:39:03mahargbeing able to compile itself doesn't mean it's a complete and conformant implementation. Rubinius could compile itself long before it could claim that status
19:39:20shtirlicbrixen: Computed MD5 to llvm-2.8-i686-apple-darwin10.0.0.tar.bz2.md5
19:39:45mahargI'm probably behind on its status, I'll grant that. Just wasn't aware it had become something people could really use. I knew clang had, just not clang++
19:40:08dbussinktheir c++ implementation is pretty recent
19:40:22dbussinkthe announcement they could compile llvm wasn't too long ago
19:40:37dbussinkat that point it could actually also compile rbx (the vm that is)
19:40:46brixenshtirlic: ok, great
19:41:17brixenshtirlic: when evan returns he'll tell you where to send the .tar.bz2 and the .tar.bz2.md5
19:41:24brixenevan: ^^^
19:42:29dbussinkbrixen: btw, on #623, i think we should actually advice against using rvm if people want to contribute to rbx itself
19:42:40dbussinkbrixen: because it then definitely brings more pain than gain
19:42:42brixendbussink: yes, indeed
19:42:49shtirlicbrixen: why?
19:42:56brixenshtirlic: why what?
19:43:25shtirlicbrixen: sorry, about advice against using rvm
19:43:37brixenshtirlic: I'm not advising against it per se
19:43:51brixenshtirlic: did you read my comment on your ticket?
19:44:26brixenI do not support recommending RVM in *our* docs
19:44:34brixenmainly because we don't control RVM
19:44:43brixenour docs are about what we do control
19:44:54brixenif you are developing on rbx, you definitely don't want to use rvm
19:45:02dbussinkshtirlic: when people want to contribute it's easiest to have your clone somewhere
19:45:03brixenit makes your life very unpleasant
19:45:17dbussinkif you then have rvm active, rbx will pick up on it's env vars etc.
19:45:18brixenrvm is for *running* multiple ruby implementations
19:45:22shtirlicbrixen: ah, I see your point
19:45:27brixenthat's its purpose
19:45:32dbussinkand you load the wrong gems etc., stuff blows up, etc.
19:46:19brixenok, I'm gonna take an early lunch with my gf, see y'all later :)
19:46:50shtirlicbrixen: see ya, have a nice lunch.
19:47:11brixenshtirlic: thanks! just poke evan when he returns about the llvm prebuilts
19:47:17brixenshtirlic: and thanks for building them
19:47:44brixenshtirlic: or if you have a dropbox link or something, upload them and paste the link here
19:48:06shtirlicbrixen: sure, will be done
19:49:00boyscoutCI: rubinius: 694e047 successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
19:57:56shtirlicevan: brixen http://dl.dropbox.com/u/1997036/llvm-all-prebuilt.tar
19:59:22NewbieJoefyi, I just build 1.2 and succeeded on first try - months ago I failed on 1.0 - good job guys! :)
20:02:53dbussinkgood to hear :)
20:03:47NewbieJoehowever, all my projects are 1.9 already, so no rbx testing yet...
20:03:52evanshtirlic: did you gzip it?
20:04:19evanshtirlic: um.
20:04:27evanyou should have a tar.bz2 and a .md5
20:04:29evanwhere are those?
20:04:46dbussinkNewbieJoe: what kind of 1.9 features are you using?
20:05:21NewbieJoemostly encoding and some of the new methods in core modules
20:05:27evanshtirlic: i've got it
20:05:43evani don't know why there is 4 files there
20:05:46evanthats odd.
20:12:02shtirlicevan: should be two as i think
20:12:25evani put them in place
20:12:33evanthere was an llvm--i686... set too
20:12:40evanI just delete it.
20:13:45shtirlicevan: so it's prebuilt libs of llvm for 10.6.x 32bit
20:13:55evanyep.
20:16:39NewbieJoecall me stupid, but I can't find in the docs any hint on how to run rbx in 1.9 mode (as far as implemented)?
20:17:21dbussinkNewbieJoe: there's only a 1.9 mode in the multiverse
20:17:25dbussink+ branch
20:17:31dbussinkbut that's heavily work in progress
20:17:39dbussinklot's of stuff needs to happen there
20:17:53NewbieJoeah, ok
20:26:10dbussinkevan: could it be that there's a lot of unused code in vm/gc/gc.cpp?
20:27:06dbussinkevan: everything still works fine after this change: https://gist.github.com/d2d0dac201ba583edf2a
20:37:22evanomg
20:37:23evanfuck no.
20:37:25evando NOT do that.
20:37:55evanha
20:37:55dbussinkevan: i'm just curious
20:37:55evani see.
20:38:01dbussinkevan: because it's not called at all
20:38:02evani'm not using unmark_all anymore
20:38:08evanand the only user of visit_call_frame was it.
20:38:15evani guess i'd forgotten about that.
20:38:35dbussinkevan: i'm not going in and just removing stuff :p
20:39:09evansorry, i reacted badly.
20:39:12evani guess we can delete that.
20:39:25evani read it wrong and thought you had walk_callframe
20:39:27evannot visit_
20:39:30evanthats why i was confused.
20:39:55dbussinkevan: hehe, well, i wouldn't have gisted this if everything wouldn't still work :)
20:40:25evanyeah, silly me.
20:40:26evananyway
20:40:28dbussinkevan: there's a slight bit more btw, i'll update the gist
20:40:29evansure
20:40:31evanremove it.
20:40:34dbussinkok, cool
20:40:37evanif we need it again
20:40:43evanwe can go back and get it.
20:42:51boyscoutRemove unused code from GC - 715cecf - Dirkjan Bussink
20:54:51boyscoutCI: rubinius: 715cecf successful: 3523 files, 15837 examples, 44086 expectations, 0 failures, 0 errors
20:58:29evandamn, imperator isn't around
20:58:43evanwe should just give him a commit bit and get him to fix these simple bugs.
21:00:51jakedouglasevan: hey, sorry i got upset last night. i sometimes just have a hard time reminding myself that you aren't intending to be rude and that it's usually a misunderstanding.
21:01:03evanno problem.
21:01:07evani have the same issue.
21:01:08evanIRC sucks.
21:01:21jakedouglasyea
21:02:16dbussinkjakedouglas: just read a few lines back ;)
21:02:18dbussinkevan: ;)
21:02:26evanexactly.
21:02:31jakedouglasheh
21:03:36jakedouglasevan: i think i'm going to be deploying that app next week, on rubinius, assuming i dont run into any hangups.
21:03:54evanwoo!
21:04:11evantesting has been going fine?
21:05:35jakedouglasi actually haven't had time to work on it much recently, it isn't what i work on full time. but no, i have yet to run into any problems on rbx, other than that it's really slow in development mode so i use MRI for that.
21:06:32evangotcha
21:06:50evanI think that in the near term i'm going to dedicate some time to profiling and improving rails dev mode.
21:06:58evanperformance wise.
21:07:10evani've put time into that on and off
21:07:16evani think it's time to allocate a bit of time again.
21:07:36jakedouglasin production mode though, once it warms up, it seems to do as good or better in many cases.
21:08:24jakedouglasif i can figure out a sane way to configure it systems-wise, i might run one instance of MRI/REE alongside one instance of rbx. that would be interesting
21:10:04jakedouglasiirc i couldn't get new relic to work with rbx though. it didn't blow up, but i never got any data in new relic. didn't have time to dig into what was happening.
21:11:11evanoh?
21:11:11evanok
21:11:24evanI've been approached by new relic before about rbx support
21:11:28evanso I should ping them
21:11:55jakedouglasyea
21:12:08jakedouglasthis was like…whenever we first talked about this, maybe a month or two ago
21:12:13evank
21:12:15jakedouglasso, recent versions of rbx and new relic.
21:12:23jakedouglasand rails 3
21:15:04vborjaguys, who made the rubinius logo?
21:15:15jakedouglasdo you have ideas about how to fix your IC contention problem?
21:16:00evanvborja: shane becker
21:16:05evan@veganstraightedge on the twitter
21:16:13evanjakedouglas: yeah
21:16:19vborjaevan: k, thanks
21:16:29evanbasically, the contents of the cache need to be packaged up as a single object
21:16:40evanand the cache itself points to the current contents as just a pointer
21:16:58evanand when the cache is read, the pointer is read onto the stack, and then consulted.
21:17:04evanthat way it's consistet while reading.
21:17:11jakedouglashmm
21:17:15evanand updating the cache means simplying changing the pointer
21:17:18evanwhich is an atomic action.
21:17:27jakedouglasmakes sense
21:17:33evan(you can't get half the bytes of one pointer and half the other)
21:18:11jakedouglasyea
21:18:18dbussinkevan: the indirection probably has a very small impact right?
21:18:26evansure
21:57:52PierreYhi guys !
21:58:11PierreYthanks for the christmas gift ;-)
21:58:36PierreYhum... where is the doc directory gone ?
21:59:05evandoc/
21:59:05evan?
21:59:40PierreYnope, there is no doc directory in the current master tree
22:00:13evanoh, web
22:00:15evanweb/doc
22:00:29PierreYarf, thank you
22:01:41vborjaFYI i've just registered #typhon-rbx channel.
22:02:39evan:D
22:02:42Zoxcnot just #typhon?
22:03:06Zoxcalso did you buy the humble indie bundle?
22:05:01vborjaZoxc: #typhon was already taken .. don't even know what it is about.
22:07:46PierreYlibffi don't compile with clang (unix64.S)
22:08:12evanPierreY: yep, i've heard that.
22:12:21PierreYi'll play with the standard gcc compiled version ;-) merry christmas guys
22:14:34tarcieriwhat's typhon?
22:14:49tarcieripython for rbx?
22:14:49tarcieriheh
22:15:04vborjatarcieri: yeah. http://github.com/vic/typhon
22:15:08tarciericool