Index

Show enters and exits. Hide enters and exits.

01:56:28boyscoutMake directories if they don't exist - 619743f - Evan Phoenix
01:58:21boyscoutCI: 619743f success. 3002 files, 11431 examples, 35502 expectations, 0 failures, 0 errors
10:34:42boyscoutOMG, Evan introduced a trailing white space - e6cd4b0 - Dirkjan Bussink
10:36:34boyscoutCI: e6cd4b0 success. 3002 files, 11431 examples, 35502 expectations, 0 failures, 0 errors
12:52:15boyscoutChange MemoryPointer#read_float, write_float to work with floats instead of doubles. - 5f82f9d - Mikko Perttunen
12:52:15boyscoutUpdate VM specs for memorypointer_write_float, memorypointer_read_float - 47a9c7b - Mikko Perttunen
12:52:15boyscoutAdd MemoryPointer#read_double, write_double and associated primitives - 0fc43fb - Mikko Perttunen
12:52:15boyscoutWhitespace fixes. - 70c636e - Mikko Perttunen
12:52:15boyscoutAdd VM tests for memorypointer_read_double and memorypointer_write_double - 00ca070 - Mikko Perttunen
12:52:17boyscoutAdd specs for MemoryPointer#read_double, write_double - 62059fd - Mikko Perttunen
12:54:17ruecyndis: Working toward FFI compatibility?
12:54:59cyndissomewhat, i started a FFI-opengl wrapper and i want it to work on both mri and rbx
12:55:06cyndissince the opengl extension doesn't compile on rbx atm
12:58:31dbussinkcyndis: ah, cool, curious as how it work out :)
12:59:02boyscoutCI: 62059fd success. 3004 files, 11431 examples, 35502 expectations, 0 failures, 0 errors
13:00:40cyndis:)
13:01:03cyndisi ran out of pure-ruby projects to test on rbx so i had to move on to the numerous opengl projects :p
13:01:15dbussinkcyndis: hehe, they all already work?
13:01:31cyndisnope :p
13:01:54cyndisbut i have my scenegraph engine with nice rotating triangles working
13:02:12dbussinkcyndis: cool :)
13:02:28dbussinkcyndis: how's the ffi performance of rbx and mri if you compare them?
13:02:40dbussinkcyndis: or not really noticeable?
13:02:53cyndishaven't done any benchmarks, but both certainly run at realtime on my machine
13:03:21rtubiogood morning
13:03:23cyndisdoing these simple opengl things
13:03:27cyndishello
13:03:32rtubiohello cyndis
13:03:49dbussinkcyndis: last time i played with opengl was in c++, so that was more suffering :P
13:04:01dbussinkdid have a nice scenegraph though with some animation stuff etc.
13:04:04cyndisyes :p
13:04:09rtubioI would like to know where can I get the last specification for the bytecode that rubinius runs
13:04:29dbussinkcyndis: but my quaternion knowledge is really rusty these days :P
13:04:50cyndismine is nonexistant, just copypasted some code :P
13:05:14rtubiook
13:05:19rtubioanybody?
13:05:34dbussinkrtubio: you can read doc/vm/rbc_files.txt but i have no idea whether it's still current
13:05:41rtubiook
13:07:30rtubiodoc/vm/intro.txt it seems to be a good introduction
13:07:34rtubiothank you very much
13:08:14rtubioI am trying to evaluate the feasibility of implementing a hardware processor that runs the bytecode of ruby
13:08:15rtubiothus
13:08:19rtubioa hardware virtual machine
13:08:23dbussinkrtubio: ah, cool :)
13:08:36rtubiohas anybody heard about a similar project?
13:08:37dbussinkrtubio: you want to implement the full instruction set?
13:08:45rtubioyes... that's the main idea
13:08:56dbussinksince it's a pretty large set
13:09:00rtubiothe problem with ruby, in the end, for embedded applications
13:09:10rtubiois its lack of real time code execution
13:09:11rtubio...
13:09:31rtubioan option to solve this problem is trying to implement the virtual machine
13:09:33rtubioin a soft-core
13:09:37rtubiothat can run in an FPGA
13:09:42rtubio...
13:09:46rtubioI know it is a lot of work
13:09:53rtubiobut I am working in a University as a teacher
13:10:03rtubioand I can get lots of MsC Students for doing this
13:10:12rtubio(as their MsC Thesis)
13:10:15rtubio...
13:10:20dbussinkif you can get it to work, i'd love to run it too, work at a hardware / software company so i could get my hands on some chips to try things :p
13:10:40rtubioxd, you might have to wait a little...
13:10:41rtubioxd
13:10:45dbussinkhehe
13:10:58rtubiowell, it is quite complicated but it is very interesting
13:11:18rtubiosince the most of the real time applications nowadays
13:11:23rtubioare still developed in C
13:11:24dbussinkwhich university?
13:11:25rtubio...
13:11:30rtubioUniversity of Vigo, Spain
13:11:41dbussinkyeah, i know, we recently built a new device, which runs embedded c stuff
13:12:16rtubioI think that to develop a processor able to run code written in a high-level language
13:12:18rtubiolike ruby
13:12:32rtubioeven it is not easy
13:12:52rtubioit would be very helpful for easing the development of embedded applications
13:12:53rtubio...
13:13:00rtubioI am currently working with ESA
13:13:05rtubioin an space project
13:13:14rtubioand they have to develop all their code in C/C++
13:13:15rtubioand
13:13:24rtubioeven though they are trying to move to Java RealTime
13:13:24dbussinkyeah, it would make embedded life a lot easier
13:13:50rtubioRTSJ specification does not fulfill certain hard real-time requirements...
13:14:07rtubiosince it depends on the execution of the Java bytecodes over a software virtual machine
13:14:09rtubio...
13:14:46rtubiowell,,, i will try to begin this project
13:14:55rtubioand will connect regularly to this irc channel
13:15:05rtubiomy mail is "rtubio@xatcobeo.com"
13:15:21rtubioif you ever know about a similar project, please let me know
13:15:24dbussinkrtubio: evan or brixen are usually the ones who know all the internal details about stuff
13:15:31dbussinkthey work on rubinius full time
13:15:32rtubioajam
13:15:44rtubiook, wonderfull
13:17:27rtubiodo they use to join frequently this irc channel?
13:17:29dbussinkrtubio: btw, this details more on the current compiler internals: http://www.engineyard.com/blog/2009/improving-the-rubinius-bytecode-compiler/
13:17:37rtubiothanks
13:17:44rtubioengineyard? ok, another web page
13:17:45rtubio...
13:17:46rtubiothank you
13:17:48dbussinkrtubio: yeah, but they're both at the us west coast
13:17:53rtubiook
13:17:57rtubiocalifornia?
13:18:13dbussinkone is on LA, the other in Portland
13:18:22rtubiook, thanks for the information
13:18:24dbussinkbut that's usually a good indication when they're available
13:18:29rtubioyes
13:18:41dbussinkbut it's rubyconf these days, so not a lot of people online
13:19:00rtubioI've been working in SLO (San Luis Obispo) past year, for a couple of weeks
13:19:21rtubioand there are 9 hours of difference between spain and US's west coast
13:19:32rtubiook
13:19:36dbussinkyeah
13:19:39rtubiorubyconf? a conference about ruby?
13:22:58rtubiowell, thank you very much for your help dbussink
13:23:06rtubiobut, in spain, it is lunch time
13:23:10rtubioand I have to go
13:23:15rtubiosee you on monday!
18:40:32rtubiohello
18:43:20rtubiobrixen, are you there?
19:16:48binary42In the spirit of conference driven development, I did a pull today and manage to consistently segfault rbx on some ruby code.
19:17:18dbussinkbinary42: if you can create a github issue out of it that would be great
19:17:40binary42Yeah. Will be trying sans llvm.
19:17:47dbussinktogether with the ruby code to reproduce it
19:18:08binary42Yes. It's just using my remix-stash project's benchmark.
19:18:21binary42It seems to be good at stressing certain things.
19:19:00dbussinkbinary42: hehe, is it on github?
19:19:12binary42binary42/remix-stash
19:20:06binary42Okay. Still fails w/o llvm
19:21:10dbussinkwhat's the command to cause it to crash?
19:21:36binary42miro:remix-stash brian$ ~/Projects/rubinius/bin/rbx benchmarks/get_set.rb
19:22:12binary42I do have rubygems and the other libs commented out in harness.rb
19:28:05dbussinkbinary42: getting it here too
19:29:10binary42http://github.com/evanphx/rubinius/issues/#issue/80
19:38:26dbussinkbinary42: hmm, the bench is still running for me if i disable llvm
19:38:36dbussinkbinary42: and the segfault is occuring inside llvm parts
19:39:05binary42I did configure w/o llvm and get a segfault.
19:39:13binary42Unless the build system doesn't clean properly?
19:40:24dbussinkbinary42: you did a rake clean after disabling it?
19:41:11binary42I am trying a new build. It probably wasn't.
19:45:31binary42I definitely still get a failure w/o llvm
19:48:06dbussinkbinary42: hmm, what do you see if you run it with gdb and no llvm?
19:48:31binary42Trying one more build to be sure but I'll try that next.
19:51:30binary42Hmm. Seems like rake clean is broken for me. git clean worked.
19:51:37binary42Now it's running... slowly.
19:53:40dbussinkit's slow without llvm yeah :)
21:45:53yakischlobayo
21:46:35yakischlobalooks like UDPSocket.open is supposed to exhibit the same behavior with a block as IO.open?
21:55:17boffbowshhey guys, i'm brand new to rbx dev. When i run bin/mspec spec/frozen/core i get a segfault. what's the next step for me trying to fix this?
21:56:36brixenboffbowsh: do you have a gist?
21:57:18boffbowshatm i'm trying to work out what spec is making it segv
21:57:22yakischlobabrixen: are changes to rubyspec supposed to be made upstream and you pull them back in or something?
21:57:58brixenboffbowsh: ok, you can do bin/mspec -fs spec/frozen/core
21:58:04brixenboffbowsh: add -V to get a file name
21:58:16brixenyakischloba: do you have a rubyspec bit?
21:58:23yakischlobano
21:58:29brixenok
21:58:44brixenideally, you would make changes to rubyspec and I merge them into rbx
21:58:58brixenbut I'm ok merging them back to rubyspec if it's like fixing an rbx bug or something
21:59:09boffbowsh http://gist.github.com/240301
21:59:22brixenbut if you commit to rbx spec/frozen, please make sure that's a separate commit
21:59:26brixenyakischloba: ^^^
21:59:42yakischlobaseparate commit from the associated bug you mean?
21:59:46brixenyakischloba: also, you can have a rubyspec commit bit since you are commiting here
21:59:49brixenyakischloba: yep
21:59:52yakischlobaok
22:00:53brixenyakischloba: I added you to rubyspec
22:00:59yakischlobathanks
22:01:08yakischlobaso it should be pushed there?
22:01:10brixenboffbowsh: ack
22:01:26brixenyakischloba: well, um...
22:01:42brixensure, if it's the exact same commit
22:01:58brixencus when I merge back to rubyspec, git will realize it's already committed
22:01:58yakischlobawell should I just push it to rubyspec and figure you'll merge it back in to rbx at some point?
22:02:16brixenyeah, I update spec/frozen from rubyspec
22:02:24brixenwill do that before RC1 is released
22:02:45yakischlobaok. ill just push any updates to rubyspec then and let you handle merging them
22:02:54yakischlobaless confusing
22:03:00brixenboffbowsh: I gotta head back to the conf, can you put the backtrace and all your system info in a github issue, please
22:03:10brixenyakischloba: ok, sounds good
22:03:18yakischlobabrixen: are you in one of the talks right now?
22:03:26boffbowshyeah i'll see if i can get some more deets first
22:03:27brixennah, I'm chillin in my room
22:03:34yakischlobaah ok.
22:03:37brixenyakischloba: are you at the conf?
22:03:41yakischlobayea
22:03:47brixendoh!
22:03:47slavahi brixen
22:03:48yakischlobabright green jacket, you cant miss me ;)
22:03:52brixenwe gotto poken
22:03:54brixen:)
22:03:56yakischlobahah
22:03:56brixenhey slava
22:03:59yakischlobapoken is busted
22:04:03brixendang!
22:04:23brixenyakischloba: ok, I'm headed to room 2
22:04:50brixenfades
22:04:54yakischlobak
22:05:23yakischlobaim actually gonna stay in here for nutters talk
23:11:57boffbowshAnyone able to review a patch, check I've done it right before adding to a ticket?
23:12:03boffbowshhttp://gist.github.com/240330 http://github.com/evanphx/rubinius/issues/#issue/81
23:21:57boffbowshsubmitted patch to ticket anyway
23:24:59boyscoutAllow IO.open to define block behavhavior for it's inherited classes - d417a53 - Jake Douglas
23:26:54boyscoutCI: d417a53 success. 3004 files, 11431 examples, 35502 expectations, 0 failures, 0 errors