#ruby-core on 2012-12-27 — irc logs at freenode.irclog.whitequark.org

00:10 xibbar has joined #ruby-core

00:14 marcandre has joined #ruby-core

00:28 drbrain has joined #ruby-core

00:34 carlosantonio has joined #ruby-core

00:36 drbrain has quit [Ping timeout: 252 seconds]

01:17 carlosantonio has quit [Ping timeout: 252 seconds]

01:28 drbrain has joined #ruby-core

01:36 drbrain has quit [Ping timeout: 260 seconds]

02:06 asarih has joined #ruby-core

02:22 asarih has quit [Quit: ["Textual IRC Client: www.textualapp.com"]]

02:28 drbrain has joined #ruby-core

02:36 <_ko1> drbrain: I got http://www.atdot.net/sp/view/zh4ofm

02:36 <_ko1> rdoc bugs\?

02:36 drbrain has quit [Ping timeout: 244 seconds]

02:44 <zzak> _ko1: ffi gem?

02:44 <zzak> which version of rdoc?

02:44 <zzak> i will try to reproduce

02:44 <_ko1> zzak: trunk :)

02:49 xibbar has quit [Remote host closed the connection]

02:51 charliesome has joined #ruby-core

02:58 <zzak> _ko1: yes, i get the same thing

02:59 <zzak> it is a bug in the changelog parser

02:59 <zzak> i will submit a ticket to eric on github/rdoc/rdoc

03:01 <zzak> https://github.com/rdoc/rdoc/issues/165

03:01 <zzak> thanks koichi-san

03:03 The_8473 has joined #ruby-core

03:03 The_8472 has quit [Ping timeout: 246 seconds]

03:06 <zzak> _ko1: btw, i was reading about objspace: #reachable_objects_from https://gist.github.com/4357176

03:06 <zzak> very cool!

03:10 xibbar has joined #ruby-core

03:27 <_ko1> zzak: thank you!!

03:29 drbrain has joined #ruby-core

03:35 <zzak> _ko1: make anything cool with debug_inspector or tracepoint yet?

03:35 xibbar has quit [Remote host closed the connection]

03:35 <_ko1> yet :(

03:36 <_ko1> nobu implemented profiler.rb w/ TracePoint

03:36 xibbar has joined #ruby-core

03:37 drbrain has quit [Ping timeout: 251 seconds]

03:37 <zzak> i saw

03:37 <zzak> but it is private api

03:38 <_ko1> debug inspector?

03:38 <zzak> i wonder

03:38 xibbar has quit [Remote host closed the connection]

03:38 <_ko1> ok, i'll try it today

03:39 <_ko1> after I can check rails+threading bug

03:40 <_ko1> i heard that rails (master) + ruby (trunk) make Deadlock error

03:40 <_ko1> at ConditionalVariable

03:40 <_ko1> this is why i try gem update :)

03:40 <_ko1> and got above bug

03:40 <zzak> Profiler uses TracePoint instead of set_trace_func

03:41 <_ko1> yes

03:41 <charliesome> what do you guys think of normalising method names to utf-8

03:41 <zzak> it uses TracePoint#enable for Profiler#start_profile and #disable for #stop_profile

03:41 <zzak> i think Profiler should use #enable(&blk) somewhere, that would be cool feature

03:42 <_ko1> i'm not sure about that

03:42 <_ko1> charliesome: method names?

03:42 <charliesome> basically at the moment method names are encoding oblivious

03:42 <_ko1> zzak: maybe it will be new feature

03:43 <charliesome> _ko1: see http://eval.in/5351 and http://eval.in/5352

03:43 <_ko1> charliesome: generally, method name should use 7bit ascii IMO

03:44 <_ko1> charliesome: if someone want to use other than ascii-7 bit, then

03:44 <_ko1> charliesome: a person has responsibility

03:44 <_ko1> i think it is overkill to enforce utf-8

03:44 <charliesome> i see what you mean

03:45 <zzak> encoding hell

03:49 <zzak> i am starting on iseq in my free time, but its tough

03:49 nari has joined #ruby-core

03:51 <zzak> hope to improve docs, and add rdoc for new api in r38076 and r38085

03:51 <zzak> i dont know vm programming, so learning as i go

03:54 <_ko1> r38076 was reverted :(

03:55 <_ko1> BTW, http://eval.in/ is cool which charliesome used

03:56 <charliesome> _ko1: thanks :) it's my site actually

03:56 <_ko1> cool!

03:56 <charliesome> it has ruby 1.0 which imo is a killer feature

03:56 <_ko1> how to make sandbox?

03:56 <zzak> _ko1: is there a ticket for that? i still want to add rdoc, i see it's not reverted but commented out

03:56 <_ko1> ah, yes. MRI 1.0 awsome!

03:56 <charliesome> _ko1: i use a modified version of the sandbox from here: http://www.ucw.cz/moe/

03:57 <charliesome> it is ptrace based

03:57 <_ko1> zzak: secret :)

03:57 <_ko1> (no ticket)

03:57 <_ko1> charliesome: so you filter all system calls using ptrace?

03:57 <charliesome> _ko1: user submitted code runs on an isolated vm running an old version of ubuntu with an old kernel (that sandbox doesn't work on newer or 64 bit kernels)

03:57 <charliesome> yes

03:57 <charliesome> i allow interpreters to make any syscall until they open the source file

03:57 <charliesome> then it locks down

03:58 <_ko1> charliesome: i'm confusing. isolating VM seems not need to use ptrace.

03:59 <charliesome> i don't start a new vm for each submission

03:59 <charliesome> so i need to ensure the integrity of the evaluator vm with a ptrace sandbox

03:59 <zzak> _ko1: have any use case or toy for iseq i can reference?

03:59 <_ko1> > while true; fork; end

04:00 <_ko1> > Forbidden syscall clone

04:00 <_ko1> :)

04:00 <charliesome> :)

04:00 <charliesome> the main reason i'm using a separate vm is because the kernel on my main server is too new for the sandbox

04:00 <_ko1> zzak: ISeq#path are only for debugging, i can imagine

04:01 <_ko1> charliesome: ah, i understand. there is two level isolation. (1) VM isolation (one vm for all submission) (2) ptrace (each submission)

04:01 <charliesome> yes

04:02 <_ko1> s/there is/are/

04:02 <_ko1> cool

04:02 <_ko1> > puts open('/etc/hostname'){|f| f.read}

04:02 <_ko1> > OK (0.016 sec real, 0.041 sec wall, 7 MB, 53 syscalls)

04:03 <_ko1> how to choose 'good' syscall and 'bad' syscall?

04:03 <_ko1> choose by hand with your experience?

04:03 <charliesome> _ko1: the sandbox i am using a modified version of has 3 levels of syscall filtering

04:03 <charliesome> always block, never block, inspect arguments

04:04 <charliesome> for example, open() has its arguments inspected to decide whether to allow it or not

04:05 <_ko1> i see

04:06 <zzak> _ko1: ありがとう

04:06 <_ko1> > Thread.new{}.join

04:06 <_ko1> > Forbidden syscall clone

04:07 <_ko1> But MRI2.0 needs to make timer thread

04:07 <_ko1> you change something on it?

04:07 <_ko1> ah

04:07 <_ko1> it's my mistake

04:07 <_ko1> MRI 2.0 (1.9.3 or later) don't need to make timet thread until single threading

04:08 <charliesome> _ko1: oh does it? that's problematic :|

04:09 <zzak> haha! tenderlove wrote a scheme compiler with fiddle and iseq

04:09 <zzak> https://gist.github.com/3536045

04:09 <charliesome> oh wow, nice

04:09 <_ko1> awesome

04:11 <_ko1> it may not work now...

04:11 <zzak> you mean because of fiddle changes?

04:12 <_ko1> iseq changes

04:12 <charliesome> and VM changes i presume

04:12 <charliesome> ah

04:12 <_ko1> instructions are changed

04:12 <_ko1> more correctly, operand of opt_* was changed

04:16 <zzak> interesting

04:17 <_ko1> ah,

04:17 <_ko1> it also my mistake

04:18 <_ko1> ah

04:18 <_ko1> not my mistake

04:18 <_ko1> ../trunk/test.rb:39: [BUG] Segmentation fault

04:19 <zzak> !

04:29 drbrain has joined #ruby-core

04:37 drbrain has quit [Ping timeout: 244 seconds]

05:30 drbrain has joined #ruby-core

05:38 drbrain has quit [Ping timeout: 265 seconds]

06:07 <_ko1> http://www.atdot.net/sp/view/tbeofm bundle failed...

06:30 drbrain has joined #ruby-core

06:38 <_ko1> after "gem install bundler --pre", it success

06:39 drbrain has quit [Ping timeout: 256 seconds]

06:39 <charliesome> so rails does not work out of the box on 2.0.0?

06:43 <_ko1> i'm not sure

07:17 davidbalbert is now known as davidbalber|away

07:23 The_8472 has joined #ruby-core

07:26 The_8473 has quit [Ping timeout: 255 seconds]

07:30 drbrain has joined #ruby-core

07:39 drbrain has quit [Ping timeout: 260 seconds]

07:49 <_ko1> rails test is too slow... not stopped yet

08:07 <_ko1> http://www.atdot.net/sp/view/07jofm

08:07 <_ko1> Rails master with trunk

08:08 <_ko1> many many warnings

08:31 drbrain has joined #ruby-core

08:39 drbrain has quit [Ping timeout: 256 seconds]

09:06 cyri_ has joined #ruby-core

09:31 drbrain has joined #ruby-core

09:34 carlosantonio has joined #ruby-core

10:09 _whitelogger has joined #ruby-core

10:32 drbrain has joined #ruby-core

10:40 drbrain has quit [Ping timeout: 260 seconds]

11:24 xibbar has joined #ruby-core

11:32 drbrain has joined #ruby-core

11:37 drbrain has quit [Ping timeout: 265 seconds]

11:43 <znz_v> biff: [ruby-changes:26577] glass:r38628 (trunk): * ext/stringio/stringio.c (strio_ungetc): raise IOError instead of RuntimeError - http://mla.n-z.jp/?ruby-changes=26577

11:47 xibbar has quit [Remote host closed the connection]

11:53 travis-ci has joined #ruby-core

11:53 <travis-ci> [travis-ci] The build was broken. by @glass: See http://travis-ci.org/ruby/ruby/builds/3839141

11:53 travis-ci has left #ruby-core [#ruby-core]

12:05 cyri_ has quit [Quit: cyri_]

12:09 <znz_v> biff: [ruby-changes:26578] charliesome:r38629 (trunk): * vm_core.h (VM_DEFINECLASS_TYPE): explicit cast to enum type to avoid 64->32 - http://mla.n-z.jp/?ruby-changes=26578

12:21 <znz_v> biff: [ruby-changes:26579] glass:r38630 (trunk): * ext/stringio/stringio.c (strio_getline): fix not to raise TypeError - http://mla.n-z.jp/?ruby-changes=26579

12:32 travis-ci has joined #ruby-core

12:32 travis-ci has left #ruby-core [#ruby-core]

12:32 <travis-ci> [travis-ci] The build was fixed. by @glass: See http://travis-ci.org/ruby/ruby/builds/3839466

12:33 drbrain has joined #ruby-core

12:35 <znz_v> biff: [ruby-changes:26580] keiju:r38631 (trunk): * lib/irb/ruby-lex.rb: allow to handle recursive heredocs on - http://mla.n-z.jp/?ruby-changes=26580

12:38 drbrain has quit [Ping timeout: 264 seconds]

12:45 <charliesome> does ruby intend to support pre-C89 compilers anymore?

12:47 <nurse> yes, CRuby requires C89

12:48 <nurse> doesn't support K&R

12:48 <charliesome> excellent, thanks

12:49 <nurse> You may know it shouldn't require C99, for example // comment

12:49 <nurse> it is because to suppot Visual C++ 6 and so on

12:49 <charliesome> there is a bug open related to ruby's stdarg.h/varargs.h checking

12:50 <charliesome> should i go through and make everything use stdarg.h?

12:51 <nurse> such change may break some environment like Solaris, AIX, HP-UX

12:51 <nurse> you shouldn't do unless you are sure that it is ok for such environments

12:51 <charliesome> i see, i thought there may be an issue like that

12:52 <charliesome> i don't have access to those environments so i will leave it be

12:53 <nurse> yeah, it seems better

12:54 <charliesome> it seems some people are having problems: https://gist.github.com/4172330

12:58 <nurse> is there a ticket for this on redmine?

12:58 <charliesome> yes: http://bugs.ruby-lang.org/issues/7478

12:59 <nurse> if HAVE_STDARG_PROTOTYPES is true, it won't fail.

12:59 <charliesome> if some files already unconditionally include stdarg.h, then i think it may be ok to remove the varargs.h code

12:59 <nurse> why this fail?

12:59 <charliesome> that's the strange part

12:59 <charliesome> it looks like the owner of that gist is using a custom gcc install on mac os x

12:59 <charliesome> i'm installing the same one to try and reproduce this

13:00 <charliesome> perhaps the check in configure.in could be fixed

13:00 <nurse> such macro is defined on .ext/include/<arch>/ruby/config.h by configure

13:01 <charliesome> i think there is a problem with the code configure uses to check

13:01 <nurse> the code which generates it is configure.in::977

13:01 <nurse> yeah, i think so

13:01 <charliesome> yep i'm looking at that right now

13:12 carlosantonio has joined #ruby-core

13:26 cyri_ has joined #ruby-core

13:34 drbrain has joined #ruby-core

13:38 drbrain has quit [Ping timeout: 265 seconds]

13:44 justincampbell has joined #ruby-core

14:18 carlosantonio has quit [Remote host closed the connection]

14:20 charliesome has quit [Quit: Textual IRC Client: www.textualapp.com]

14:30 nari has joined #ruby-core

14:34 drbrain has joined #ruby-core

14:38 drbrain has quit [Ping timeout: 252 seconds]

14:48 <znz_v> biff: [ruby-changes:26581] keiju:r38632 (trunk): * lib/irb/ruby-lex.rb: make lex_state to EXPR_END when next token - http://mla.n-z.jp/?ruby-changes=26581

14:50 carlosantonio has joined #ruby-core

14:53 travis-ci has joined #ruby-core

14:53 travis-ci has left #ruby-core [#ruby-core]

14:53 <travis-ci> [travis-ci] The build was broken. by @keiju: See http://travis-ci.org/ruby/ruby/builds/3841075

14:53 nari has quit [Read error: Operation timed out]

15:00 <zzak> おはよう！

15:35 drbrain has joined #ruby-core

15:38 drbrain has quit [Read error: Operation timed out]

16:16 <znz_v> biff: [ruby-changes:26582] jeg2:r38633 (trunk): * lib/csv.rb: Added more Hash methods to CSV::Row. - http://mla.n-z.jp/?ruby-changes=26582

16:16 <znz_v> biff: [ruby-changes:26583] svn:r38634 (trunk): * 2012-12-28 - http://mla.n-z.jp/?ruby-changes=26583

16:25 travis-ci has joined #ruby-core

16:25 <travis-ci> [travis-ci] The build was fixed. by @jeg2: See http://travis-ci.org/ruby/ruby/builds/3842433

16:25 travis-ci has left #ruby-core [#ruby-core]

16:28 davidbalber|away is now known as davidbalbert

16:30 kosaki8 has joined #ruby-core

16:30 travis-ci has joined #ruby-core

16:30 <travis-ci> [travis-ci] The build was fixed. by @svn: See http://travis-ci.org/ruby/ruby/builds/3842506

16:30 travis-ci has left #ruby-core [#ruby-core]

16:35 drbrain has joined #ruby-core

16:40 drbrain has quit [Ping timeout: 244 seconds]

16:46 <znz_v> biff: [ruby-changes:26584] naruse:r38635 (trunk): * variable.c (rb_mod_remove_const): fix segv caused by r38558. - http://mla.n-z.jp/?ruby-changes=26584

16:47 knu has quit [Remote host closed the connection]

16:52 knu has joined #ruby-core

17:04 carlosan_ has joined #ruby-core

17:08 carlosantonio has quit [Ping timeout: 256 seconds]

17:28 carlosan_ has quit [Remote host closed the connection]

17:36 drbrain has joined #ruby-core

17:41 drbrain has quit [Ping timeout: 260 seconds]

17:48 <indirect> _ko1: Bundler will work with 2.0.0 when it is released. Right now Bundler 1.3.0.pre.3 works with Rubygems 2.0.0.preview.2.2 and Ruby 2.0.0-preview3

17:50 <indirect> the CI build matrix builds ruby-head and then tests with rubygems-master: https://travis-ci.org/carlhuda/bundler

18:00 carlosantonio has joined #ruby-core

18:23 carlosantonio has quit [Remote host closed the connection]

18:36 drbrain has joined #ruby-core

18:40 drbrain has quit [Read error: Operation timed out]

19:13 __BigO__ has joined #ruby-core

19:17 carlosan_ has joined #ruby-core

19:36 carlosan_ has quit [Remote host closed the connection]

19:37 drbrain has joined #ruby-core

19:42 drbrain has quit [Ping timeout: 264 seconds]

19:47 carlosan_ has joined #ruby-core

20:03 carlosan_ has quit [Remote host closed the connection]

20:04 cyri_ has quit [Quit: cyri_]

20:16 carlosantonio has joined #ruby-core

20:18 cyri_ has joined #ruby-core

20:19 enebo has joined #ruby-core

20:19 headius has joined #ruby-core

20:23 carlosantonio has quit [Remote host closed the connection]

20:35 carlosantonio has joined #ruby-core

20:37 drbrain has joined #ruby-core

20:40 carlosantonio has quit [Remote host closed the connection]

20:42 drbrain has quit [Ping timeout: 250 seconds]

21:00 enebo has quit [Quit: enebo]

21:00 headius has quit [Quit: headius]

21:10 cyri_ has quit [Quit: cyri_]

21:13 davidbalbert is now known as davidbalber|away

21:17 davidbalber|away is now known as davidbalbert

21:18 cyri_ has joined #ruby-core

21:38 drbrain has joined #ruby-core

21:42 drbrain has quit [Ping timeout: 244 seconds]

22:11 <znz_v> biff: [ruby-changes:26585] naruse:r38636 (trunk): * def/id.def: use split(/^/) instead of String#lines to support - http://mla.n-z.jp/?ruby-changes=26585

22:28 kosaki8 has quit [Quit: Leaving...]

22:39 drbrain has joined #ruby-core

22:47 drbrain has quit [Ping timeout: 252 seconds]

23:11 __BigO__ has quit [Remote host closed the connection]

23:24 carlosantonio has joined #ruby-core

23:39 drbrain has joined #ruby-core

23:46 carlosantonio has quit [Remote host closed the connection]

23:47 drbrain has quit [Ping timeout: 255 seconds]

23:50 cyri_ has quit [Quit: cyri_]