Index

Show enters and exits. Hide enters and exits.

00:00:15benburkert leaves the room.
00:05:28anteaya enters the room.
00:12:23stepheneb enters the room.
00:15:42rubuildius_ppc enters the room.
00:19:00qwert666 leaves the room.
00:26:49cypher23 leaves the room.
00:28:30rubuildius_ppcMenTaLguY: ae738f219; 1995 files, 6495 examples, 22615 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180167
00:31:33brainopia leaves the room.
00:38:49probablycorey leaves the room.
00:44:46cored enters the room.
00:48:42jlindley enters the room.
00:55:51EugZol leaves the room.
01:03:01Cosmos95Hi - I see in FAQ that the typical platforms folks are developing on are OS X, Linux and FreeBSD... has anyone gotten a good build on Solaris by chance?
01:03:14KirinDav enters the room.
01:04:38DefilerI have heard people talk about it various times, but I haven't paid enough attention to know whether they are succeeding
01:04:52Cosmos95I've been trying... :)
01:05:14KirinDav leaves the room.
01:05:58Cosmos95I submitted one patch... although, I keep finding things that are just build related. I'm not sure its something silly I'm doing.
01:06:18KirinDav enters the room.
01:06:57jlindley leaves the room.
01:07:00DefilerAlmost certainly every problem on that platform is a makefile bug. Heh
01:07:12DefilerSo feel free to correct dumb things we are doing/assuming
01:07:24Cosmos95yeah. that was the one submitted.
01:07:47Cosmos95The other I've found is the shotgun/config/run script...
01:08:21Cosmos95Should I put all these on one ticket? or open separate ones?
01:08:41Defilerseparate is good unless you find that two things are caused by the same issue, etc
01:08:50Cosmos95ok. Thanks.
01:09:05DefilerThanks for trying it on Solaris
01:10:04Cosmos95heh... Its necessarily cause I like this platform... but in some cases I have to use it.
01:10:17Cosmos95not*
01:15:33Flamefork enters the room.
01:17:19lstoll enters the room.
01:18:34jlindley enters the room.
01:25:10KirinDav leaves the room.
01:28:48stepheneb leaves the room.
01:31:16KirinDav enters the room.
01:35:11xif leaves the room.
01:39:08KirinDav leaves the room.
01:44:04probablycorey enters the room.
01:44:38cremes leaves the room.
01:45:03benny leaves the room.
01:51:36benny enters the room.
01:52:46probablycorey leaves the room.
01:58:11stepheneb enters the room.
01:58:30lopex leaves the room.
01:58:35headius enters the room.
02:04:48seydar enters the room.
02:10:28seydar leaves the room.
02:17:23cored_ enters the room.
02:17:55stepheneb_ enters the room.
02:20:40benburkert enters the room.
02:22:06cremes enters the room.
02:22:32benburkert leaves the room.
02:23:09probablycorey enters the room.
02:28:24ezmobius enters the room.
02:31:57stepheneb leaves the room.
02:32:58cored leaves the room.
02:37:02jlindley leaves the room.
02:37:33cored_ leaves the room.
02:46:12_VVSiz_ enters the room.
02:46:59jtoy enters the room.
02:52:09headius leaves the room.
02:54:36VVSiz_ leaves the room.
03:00:36gnufied enters the room.
03:04:28seydar enters the room.
03:04:31KirinDav enters the room.
03:11:51seydarcan someone tell me why there is no accessible io/wait?
03:13:11wombleseydar: Perhaps it just hasn't been implemented yet?
03:13:33seydarbut it _has_
03:13:38seydari'm looking at it now
03:13:41seydarbut i can't require it
03:17:18wombleseydar: Oh. That's *weird*.
03:17:31seydarmega weird
03:22:22seydarrepulling and rebuilding
03:23:17seydarwhen i can get an mspec run to actually _finish_, i will also identify the common tie between 8921 bajillion failures/error
03:23:37probablycorey leaves the room.
03:25:48KirinDav leaves the room.
03:28:25hornbeck leaves the room.
03:29:35mernen enters the room.
03:45:55Cosmos95 leaves the room.
03:46:44Cosmos95 enters the room.
03:49:46fbuilesv leaves the room.
03:56:56wycats_ leaves the room.
04:03:02seydari'll find out tomorrow morning
04:03:06seydar leaves the room.
04:05:26obiejuan enters the room.
04:05:53srbaker leaves the room.
04:06:03sudoer enters the room.
04:06:09srbaker enters the room.
04:08:51seydar enters the room.
04:08:56seydaroh my god
04:09:04seydarwho submitted a patch in the past day
04:09:14seydarthey fixed like everything on Tiger/PPC
04:09:28seydaroh, except for that one test that just failed
04:09:38seydarbut seriously, all the problems are like..... gone
04:10:00seydar1 failures, 2 errors
04:10:06seydargonna fix these two homeboys soon
04:10:10seydar leaves the room.
04:14:53jtoy leaves the room.
04:26:27benburkert enters the room.
04:26:37mernen leaves the room.
04:27:39wycats_ enters the room.
04:28:52KirinDav enters the room.
04:28:55srbaker leaves the room.
04:29:04obiejuan leaves the room.
04:29:10srbaker enters the room.
04:33:14srbaker_ enters the room.
04:33:31srbaker leaves the room.
05:00:03RyanTM_ leaves the room.
05:03:02anteaya leaves the room.
05:13:20gnufied leaves the room.
05:14:23obvio171 enters the room.
05:19:34wmoxam enters the room.
05:25:44stepheneb enters the room.
05:32:00obvio leaves the room.
05:41:45srbaker leaves the room.
05:43:56stepheneb_ leaves the room.
05:47:59KirinDav leaves the room.
05:48:50srbaker enters the room.
05:52:37evanhuzzah! figured out the libev problem.
05:54:14srbaker leaves the room.
05:57:56moofbong_ leaves the room.
06:02:16ezmobiusyay
06:02:24ezmobiusyou coming up here on tuesday evan ?
06:02:36agardinerwhat libev problem?
06:03:03evantomorrow
06:03:11evanshould be at the office by noon
06:03:39evanagardiner: libev was behaving oddly wrt timers
06:03:47evanwriting some test code for the new VM
06:03:53evandoing event intergration now.
06:04:03agardinerooh! nice!
06:04:25agardinerplanning any functionality changes to what was there before?
06:04:42evani'm restraining myself from functionality changes right now
06:04:54agardinerhehe, fair enough i guess
06:04:58evanthat being said, the event code is completely reorganized
06:05:12evaninto C++ classes so that I can properly test them
06:05:19ezmobiuscool
06:06:03agardinersounds great. i think the timing for the rewrite was pretty optimal
06:06:45agardinerin that nearly all the functionality needed was there in shotgun, so a rewrite can remove all the cruft and greatly simplify things
06:07:39evanthats my thinking.
06:08:19agardinerthere is one thing i'd like to stick on your wishlist though... :-)
06:08:20evani feel like if i'd waited much longer, it would be too late to do this work
06:08:41agardineryeah, really had to be done pre 1.0
06:08:50Jzalae leaves the room.
06:09:10be9 enters the room.
06:09:53evanwhats the thing for the wishlist?
06:09:54agardinerso, my wishlist item: some way to indicate a task should yield to the debugger if an exception is raised
06:10:29evanok
06:10:32agardinerideally, it should be after we have the exception details, but before we've started unwinding the stack looking for a handler
06:10:45evansure
06:10:51agardinergreat!
06:11:05evani think that the new Task organization will make that kind of thing much easier to implement
06:11:35agardineryeah, i'm looking forward to reviewing the next set of changes you push! :-)
06:12:50agardineryou've kept the flag to yield on context creation (although the old name wasn't ideal), i hope?
06:14:40evani'm not at that phase yet
06:14:42evanbut yes
06:14:45evanthough i still dislike it.
06:15:14agardinerhehe, i know, but i can't think of a better way...
06:15:36agardinerthere is no way to set a flag on the context itself, since it doesn't exist at the time i need it
06:16:01evani'll leave it
06:16:18evanbut still want to revisit it, by have a Task wide flag that puts the entire Task itself into debug mode.
06:16:55agardinerhmm... that's an interesting idea
06:17:33KirinDav enters the room.
06:17:39agardinerit would probably impose additional overheads on the task itself, but that's certainly a fair trade-off when debugging
06:21:15agardinerare you thinking this to enable more functionality for debugging on the task in question? e.g. tracepoints/watchpoints?
06:23:05ezmobius leaves the room.
06:23:32evanno clue
06:23:37evanhaven't thought that far ahead
06:23:39evantry not to
06:24:18evanbut I do know that being able to put a Task in debug mode, where it's normal methods performed additional duties (sending channel data, for example) would simplify the code
06:24:32agardineryeah, it would probably help to have the use case first
06:24:34evanbecause you wouldn't have random code stuck it at random places
06:24:50agardineryeah, i know what you mean
06:25:00KirinDav leaves the room.
06:26:57agardineri think it might be advantageous to have more than one flag though... e.g. yielding on context change is only needed if we are stepping in
06:27:32agardineryielding on exception probably really ought to be possible at both the global level and the task level
06:28:05evanwhy?
06:28:37agardinerwell, lets say you've got some code that runs and chucks an exception, but you don't know where its coming from
06:28:51evanit can't go across Tasks though
06:29:04agardineryou'd want to be able to run that in the debugger and have it break when the exception occurred
06:29:08evanwe wont be adding any more of these kinds of flags until the new VM is in
06:29:36agardinerno, i'm not touching anything in shotgun
06:29:52agardinerand i'm a ways off needing it for the debugger, just thinking ahead
06:34:16agardinertell you what, forget about adding any new flags (including for exceptions) for now
06:34:31Gerardo enters the room.
06:34:33Gerardohi
06:34:49agardineri'll write up some use cases for the missing functionality so that we can then argue/discuss
06:34:58agardiner:-)
06:35:17Gerardoin shotgun/Makefile I found a reference to libexecinfo
06:35:25Gerardo# Support *BSD libexecinfo for backtrace etc.
06:35:25Gerardoifeq ($(findstring BSD,$(UNAME)),BSD)
06:35:25Gerardo LIBS+=-lexecinfo
06:35:25Gerardoendif
06:35:30evanyeah, it should be removed
06:35:32Gerardooops
06:35:32evanit's old.
06:35:53Gerardook, I'll send the patch
06:43:18Gerardoanother Solaris' related issue:
06:43:26brainopia enters the room.
06:43:28Gerardold: warning: option -o appears more than once, first setting taken
06:43:30KirinDav enters the room.
06:43:38Gerardowhen building libbstring
06:43:43Gerardothe actual command line is:
06:43:53Gerardoecho LINK libbstring.0.32.so;/usr/sfw/bin/gcc -shared -o libbstring.0.32.so bstraux.o bstrlib.o -Wl,-soname,libbstrings-0.32.so
06:44:08GerardoI suppose ld sees a -o in -soname
06:44:19evanmust
06:44:44Gerardoafter that, I get:
06:44:46Gerardold: fatal: file libbstrings-0.32.so: open failed: No such file or directory
06:45:03Gerardoand then:
06:45:03Gerardold: fatal: File processing errors. No output written to libbstring.0.32.so
06:45:03Gerardocollect2: ld returned 1 exit status
06:45:22KirinDav leaves the room.
06:48:19evansorry, i don't have time to help you debug that right now
06:48:30Gerardosure, no problem, I'm looking at it
06:48:44GerardoI'm trying to help! :) not to waste your time
06:49:00GerardoI'll try to fix it and send the patch
06:49:55Gerardothe -Wl,-soname,libbstrings-0.32.so
06:50:03Gerardothinking aloud
06:50:15Gerardocan be omitted
06:50:23Gerardoand works
06:51:04Gerardohmm, I got it... platform specific logic in Makefile
06:52:21aotearoa leaves the room.
06:53:59Gerardohmmm, Solaris' ld can't interpret "-soname"
06:54:42GerardoI may add an ifelse, but... seems dirty to me
06:54:51Gerardoalthough there's already a test for Darwin ...
06:56:48Gerardothe same happens for the rest of the libraries in external_libs
07:01:12dlee leaves the room.
07:02:19Gerardosee you guys, it's late here
07:02:23Gerardo leaves the room.
07:09:02benburkert leaves the room.
07:11:05obiejuan enters the room.
07:11:10obiejuan leaves the room.
07:14:21aotearoa enters the room.
07:14:38mentz enters the room.
07:16:18aotearoa leaves the room.
07:22:18KirinDav enters the room.
07:26:17KirinDav leaves the room.
07:27:48KirinDav enters the room.
07:28:56xif enters the room.
07:30:47obvio171 leaves the room.
07:42:53cypher23 enters the room.
07:44:20KirinDav leaves the room.
07:45:29jnicklas enters the room.
07:49:18jnicklas leaves the room.
07:57:18boyscout2 commits by Adam Gardiner
07:57:19boyscout * Refactor Debugger interface into a CmdLineInterface class; 4ae1638
07:57:20boyscout * Breakpoint clean-up; e612414
08:05:25wmoxam leaves the room.
08:06:18aotearoa enters the room.
08:10:39rubuildius_ppcAdam Gardiner: 4ae163810; 1995 files, 6494 examples, 22614 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180350
08:11:31agardiner leaves the room.
08:12:19thehcdreamer enters the room.
08:14:48_VVSiz_ leaves the room.
08:17:32Maledictus enters the room.
08:29:52trythil leaves the room.
08:32:40aotearoa leaves the room.
08:45:55Arjen_ enters the room.
08:47:39octopod enters the room.
08:54:32rby enters the room.
09:02:52mutle enters the room.
09:07:46ttmrichter enters the room.
09:12:38antares enters the room.
09:12:57Skip enters the room.
09:21:37rby leaves the room.
09:23:13GMFlash leaves the room.
09:25:05rubuildius_ppc leaves the room.
09:25:25ttmrichter leaves the room.
09:28:33qwert666 enters the room.
09:40:58naeu enters the room.
09:47:11BlackEdder enters the room.
09:52:13stepheneb leaves the room.
10:09:38MaledictusHow can I undo a modification to a file in git which I have not comitted?
10:09:58wombleMaledictus: git reset
10:12:22halorgiumgit checkout
10:12:25octopodgit checkout path/to/file
10:12:37halorgiumwomble: you like reset too much :)
10:12:54halorgiumgit checkout -- path/to/file is more explicit
10:13:12womblehalorgium: I *love* reset
10:13:19Maledictusah, checkout is good. reset is for already comitted stuff I think
10:13:58halorgiumreset is for the whole working directory
10:14:03halorgiumcheckout is for single files
10:14:33Maledictusok
10:14:35womblehalorgium: You can specify paths to reset.
10:14:43wombleputs on his "reset rulz" T-shirt
10:15:03halorgiumwomble: not with --hard
10:15:22halorgiumshreds womble's T-shirt
10:15:33MaledictusDo you all use the way described in the git howto on LH? even for small changes?
10:15:35womblehalorgium: I didn't say to use --hard this time. <grin>
10:15:40halorgiumwomble: :P
10:15:59dewd enters the room.
10:16:22Maledictusseems a little heavy for me to branch for single line changes to HEAD
10:17:24sholden enters the room.
10:26:20antaresyou can use reset to unstage changes to index: http://novemberain.com/2008/3/29/how-to-unstage-changes-to-index-in-git
10:26:38antaresbut git always has a plenty ways to do anything :)
10:32:52w1rele55 enters the room.
10:37:02benny leaves the room.
10:39:30radarek enters the room.
10:41:17imajes leaves the room.
10:42:04dctanner enters the room.
10:42:12chris2 enters the room.
10:51:08imajes enters the room.
11:00:03imajes leaves the room.
11:21:44jtoy leaves the room.
11:21:47dctanner leaves the room.
11:24:12dirkjan enters the room.
11:27:14brainopia leaves the room.
11:31:53dbussink leaves the room.
11:36:38antares leaves the room.
11:42:45webmat enters the room.
11:43:27radarek leaves the room.
11:45:53probablycorey enters the room.
11:45:59probablycorey leaves the room.
11:57:18dctanner enters the room.
11:58:28Cosmos95 leaves the room.
12:16:34wdperson enters the room.
12:24:47naeu leaves the room.
12:34:18rubuildius_ppc enters the room.
12:38:53naeu enters the room.
12:39:57sholden leaves the room.
12:41:28sholden enters the room.
12:41:51sholden leaves the room.
12:43:41dbussink enters the room.
12:50:03rubuildius_ppcAdam Gardiner: 4ae163810; 1995 files, 6494 examples, 22614 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180427
12:50:18dctanner leaves the room.
12:54:32chris2 leaves the room.
12:55:36thehcdreamer_ enters the room.
13:00:45antares enters the room.
13:01:56zimbatm enters the room.
13:02:23RyanTM_ enters the room.
13:05:25ctennis leaves the room.
13:06:45thehcdreamer leaves the room.
13:07:33zimbatm leaves the room.
13:09:52zimbatm enters the room.
13:10:14brainopia enters the room.
13:16:36hornbeck enters the room.
13:19:29jlindley enters the room.
13:33:49yugui enters the room.
13:47:40mutle leaves the room.
13:48:01mutle enters the room.
13:50:44Cosmos95 enters the room.
14:01:58benny enters the room.
14:05:22agile leaves the room.
14:07:22riffraff enters the room.
14:08:00agile enters the room.
14:10:04skaar enters the room.
14:11:13peeja leaves the room.
14:11:15EugZol enters the room.
14:13:18jtoy enters the room.
14:14:13jtoy leaves the room.
14:17:34GMFlash enters the room.
14:19:45daesan enters the room.
14:24:02headius enters the room.
14:25:29macournoyer enters the room.
14:27:26jero5 leaves the room.
14:28:39srbaker enters the room.
14:29:08daesan enters the room.
14:29:32daesan leaves the room.
14:29:55jtoy enters the room.
14:31:50smparke1 leaves the room.
14:40:27brainopia leaves the room.
14:47:16madsimian leaves the room.
14:48:24AndrewO enters the room.
14:56:06imajes enters the room.
14:57:37BlackEdder enters the room.
14:58:32headius_ enters the room.
15:01:04headius__ enters the room.
15:01:23obvio171 enters the room.
15:04:03moofbong enters the room.
15:05:10gnufied enters the room.
15:07:30stepheneb enters the room.
15:07:54rype enters the room.
15:09:15therealadam leaves the room.
15:13:58headius__ leaves the room.
15:14:02headius leaves the room.
15:14:55d2dchat enters the room.
15:15:04yaroslav enters the room.
15:17:12EugZol leaves the room.
15:17:23headius_ leaves the room.
15:17:58smparke1 enters the room.
15:18:10peeja enters the room.
15:27:41obvio171_ enters the room.
15:29:59enebo enters the room.
15:31:44yaroslav leaves the room.
15:35:11be9_ enters the room.
15:37:37smparke2 enters the room.
15:40:48smparkes leaves the room.
15:41:14agile leaves the room.
15:41:41divoxx enters the room.
15:45:28obvio171 leaves the room.
15:45:41jlindley leaves the room.
15:47:37qwert666_ enters the room.
15:48:31yipstar enters the room.
15:50:48be9 leaves the room.
15:50:58jlindley enters the room.
15:54:49dctanner enters the room.
15:55:05peeja leaves the room.
16:00:17peeja enters the room.
16:01:18therealadam enters the room.
16:05:35brainopia enters the room.
16:07:12qwert666 leaves the room.
16:08:03dlee enters the room.
16:08:22benstiglitz enters the room.
16:09:55trythil enters the room.
16:16:07KirinDav enters the room.
16:16:52rype leaves the room.
16:18:02peeja leaves the room.
16:27:41Maledictusis fixtures/classes.rb automatically included in the tests?
16:29:11obvio171 enters the room.
16:29:28obvio leaves the room.
16:33:18agile enters the room.
16:33:52thehcdreamer_ leaves the room.
16:40:08Fullmoon enters the room.
16:40:27Maledictusrue: git-1.5.3.8 is working fine here on freebsd 7.0
16:40:27Fullmoon_ leaves the room.
16:40:44yugui leaves the room.
16:44:39dctanner leaves the room.
16:45:08w1rele55 leaves the room.
17:03:01jero5 enters the room.
17:03:46wmoxam enters the room.
17:09:34RyanTM enters the room.
17:10:37RyanTM_ leaves the room.
17:10:47KirinDav leaves the room.
17:18:03Maledictusis bin/completeness currently broken?
17:19:39mutle leaves the room.
17:19:53Skip leaves the room.
17:21:10djwhittnot sure, I know it was recently moved to mspec/bin/completeness (the one in bin just calles the mspec/bin one)
17:21:25Maledictusok
17:22:08djwhittdoing a rebuild right now then I'll try it on my system and see what happens
17:22:17Maledictusthanks
17:38:35madsimian enters the room.
17:38:46naeu leaves the room.
17:39:30djwhitthmm... seems to be broken for me as well
17:39:50djwhittMaledictus: I'd ask brixen next time he's around
17:40:11peeja enters the room.
17:40:27vborja leaves the room.
17:42:00binary42 leaves the room.
17:42:06Skip enters the room.
17:47:03naeu enters the room.
17:49:02peeja leaves the room.
17:50:17dlee leaves the room.
17:54:34dewd leaves the room.
17:57:03yaroslav enters the room.
17:59:34dewd enters the room.
18:01:18brixenMaledictus: you need to require fixtures/classes.rb explicitly, but you probably realized that by now
18:01:36brixenMaledictus: bin/completeness is being deleted today. one bit of the functionality will be moved into mkspec
18:01:38wycats_ leaves the room.
18:08:14Maledictusbrixen: ok, thanks
18:08:27KirinDav enters the room.
18:08:57TheVoice enters the room.
18:10:10octopod leaves the room.
18:15:10dbussinkbrixen: you still there?
18:17:16brixendbussink: yeah, sup?
18:17:36dbussinkbrixen: what's your opinion on sleep / timing issues with regards to specs?
18:17:45obiejuan enters the room.
18:17:53dbussinkbrixen: because i was speccing some of the FIle::Stat stuff and i kinda need that there
18:19:49brixenwell, timing dependecies would be really bad
18:20:19brixenwe have specs for #sleep, so we use that, but anything more than a very small duration kind of sucks for overall spec performance
18:20:36brixenyou mentioned something about a second resolution?
18:21:55dbussinkwell, the dependency just is there, there's no way to work around it reliably
18:22:14dbussinkbut when i was working on it, i also found a problem with sleep
18:22:23dbussinkand that is that it doesn't sleep the time it should
18:22:34dbussinksleep 1 only sleeps for like 0.6 seconds
18:22:46dbussinkso i actually need a sleep 2 for the spec to actually work
18:24:26brixenyeah, but nothing guarantees the duration of sleep, so I don't want to go there
18:24:51brixen#sleep returns a number, that's about it for the spec
18:25:00brixenwhat's the timing dependency?
18:25:13dbussinki need two files with different mtimes
18:25:31dbussinki can work around it by assuming that i create one file and for example / is always older
18:25:34brixencan those not be created through the OS or stdlib?
18:25:40dbussinkbut that feels really clunky
18:25:55brixenwhat method are you spec'ing?
18:26:03dbussinkFile::Stat <=>
18:26:09dbussinkthat thing is weird in itself though
18:26:46dbussinkit does a <=> on mtime inside the File::Stat
18:27:33brixenwell, in sane OS's that can be set, dunno about windows
18:27:38brixenI wouldn't depend on fixtures for that
18:27:40Maledictusjust an idea: you could use `touch foo` or something like that
18:27:43brixenjust create two files and set them
18:27:48brixenMaledictus: right
18:28:27brixendbussink: but you don't need to depend on sleep for that
18:28:37dbussinkwell, i currently create all files etc. through ruby's own stuff
18:28:37dbussinkso no dependency on tools etc.
18:28:37dbussinkbut i could do that too
18:28:42brixeni.e. don't 'touch foo'; sleep; 'touch foo1'
18:28:50Maledictusshould be faster also, but platform_is_not :mswin ;)
18:29:33dbussinkbrixen: well, afaik i can't change the mtime through ruby's methods
18:29:41dbussinkbrixen: so hence the current sleep
18:30:18brixenhrm, no #mtime= ? what's this world coming to
18:31:34Maledictus`touch -m 04140000 foo1; touch -m 04140001 foo2`
18:32:25dbussinkMaledictus: yeah, i know you can do that, but i'd like a ruby only implementation
18:32:35Maledictussure, me would too :)
18:32:46brixenFile.utime
18:33:05brixendbussink: ^^^
18:33:59dbussinkbrixen: ah, i overlooked that one :)
18:34:11brixendbussink: yeah, the naming just jumps right out at ya :)
18:34:33brainopia leaves the room.
18:35:14dbussinkalso a nasty mapping of c like behavior on ruby :(
18:35:14brainopia enters the room.
18:35:17lopex enters the room.
18:35:33brixenyeah
18:35:50brixenhow hard would mtime= be :P
18:38:39dbussinkyeah, well, maybe we can add that some day :)
18:39:08brixenindeed
18:42:25trythil leaves the room.
18:42:55naeu leaves the room.
18:43:17naeu enters the room.
18:44:55boyscout3 commits by Dirkjan Bussink
18:44:56boyscout * Remove tag for fixed File#lstat; e5aa89f
18:44:57boyscout * Remove unneccary spec placeholders for File#stat / File#lstat; 21cd4a1
18:44:58boyscout * Specced different behavior between File.stat and File.lstat; eacb4f8
18:46:40MaledictusAnyone dislikes this? http://pastie.org/180592
18:46:52boyscout1 commit by Dirkjan Bussink
18:46:53boyscout * Fix specs for a bunch of File::Stat methods and implement File::Stat#<=>; 130e4bd
18:48:57trythil enters the room.
18:49:09brixenMaledictus: what do you need to do with the result?
18:49:13jtoy leaves the room.
18:50:02Maledictusbrixen: http://pastie.org/180593
18:50:50Maledictushad a return there first, but thought other people could use this behaviour too
18:51:20brixenwell, problem is you need a platform_is :mswin in there too
18:51:33brixenso, you need to explicitly assign result in self.get_env
18:51:49MaledictusDidn't return from block behaviour change from 1.8 to 1.9?
18:52:11MaledictusI have no clue about mswin, so I don't know howto implement that
18:52:29tarcierithe behavior of procs vs lambdas changed... lambdas now behave like blocks, iirc
18:52:46Maledictusbut you're right of course with result in get_env
18:53:07brixenMaledictus: http://pastie.org/180595
18:53:30brainopia leaves the room.
18:53:32brixenMaledictus: the change to the guard isn't bad, but unless there's another use case besides this, I'd prefer not
18:53:32VVSiz enters the room.
18:53:59Maledictusok.
18:54:48benny leaves the room.
18:55:40peeja enters the room.
18:56:38OnO enters the room.
18:57:15OnOhowdie... I just came across Rubinius and want to tell you it is awesome !
18:58:06Maledictushehe, sure it is :)
18:59:25OnOI was running benches... this ratio is how rubinius is faster to ruby right ?
18:59:43OnOis there any page or anything stating what's missing to be compatible with 1.8 ?
19:00:12OnOlike todo list... I've seen progress bar >80% but no details
19:00:36brixenOnO: there's a couple things you can do to see work that needs to be done
19:00:40brixenOnO: rake todo
19:00:47dbussinkOnO: basically everything that hasn't been specced needs review
19:00:47dbussinkand everything where the specs fail of course
19:00:47dbussinkOnO: the stdlib/ contains stuff that hasn't been tested yet
19:00:54brixenOnO: or bin/mspec -g incomplete --dry-run spec/ruby
19:01:07rubuildius_ppcDirkjan Bussink: 130e4bdb1; 1995 files, 6503 examples, 22631 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180603
19:01:08rubuildius_ppcDirkjan Bussink: e5aa89ff1; 1995 files, 6496 examples, 22618 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180601
19:01:19dbussinkover 6500 specs :)
19:01:30dbussinkonto 7000
19:01:41brixendbussink: woot :)
19:01:44OnOwow
19:01:51OnOthat's outstanding
19:02:48brixenOnO: make that: bin/mspec -g incomplete -fs --dry-run spec/ruby/
19:03:00brixenif you don't add the -fs, the output isn't that enlightening :)
19:03:25OnOtrying
19:03:51dbussinkOnO: but not enough :P
19:04:53OnOI'm in love with Ruby, however aside of that I'm missing Ruby to be ready for packaging comercial apps... I think Rubinius is right way with RBA
19:05:19djwhitthmm... what do you mean by "packaging commercial apps" ?
19:05:24dbussinkOnO: you mean like war style deployment i guess?
19:05:31OnOexactly
19:05:49djwhittOnO: you can actually do that with JRuby now
19:06:07OnOyeah... but frankly I don't like Java ;)
19:06:23djwhittyeah, I hear ya on that one
19:06:33OnOhave you guys seen LLVM ?
19:07:03brixenOnO: yeah, lots of folks interested in what llvm might do for the vm
19:07:23brixenOnO: the C vm is currently being rewritten in C++, might even make llvm easier to use
19:07:53OnOyeah... that would basically give VM to be JIT too
19:08:04OnOso it would run at full speed :>
19:08:12djwhittwell, kind of...
19:08:49djwhittjust running the VM on top of LLVM wouldn't really give much benefit
19:08:55brixenyeah, we still need the machinery to decide what to jit
19:09:02brixenetc
19:09:03tarcieriOnO: Closures on LLVM are problematic because they require eplicitly managed stack frames
19:09:08tarcieriexplicitly...
19:09:33OnOtarcieri: this is what I thought as possible implementation problem
19:10:10OnOdjwhitt: I was more thinking of using LLVM to generate intermediate LLVM code that gets compiled to native machine code on the fly by LLVM guts
19:10:21scooprI've lately shown some interest in llvm too, though it starts to get out of my area of expertise
19:10:35djwhittOnO: right, that would be cool, but pretty difficult I think
19:10:37OnOdjwhitt: something more like .NET, JIT instead of VM
19:11:20djwhittOnO: I don't really know that much about it though. I'm mostly just repeating previous conversations in here
19:11:38MaledictusHehe, and I would be interested in compiling ruby directly to machine code as far as possible :)
19:13:26OnOwell LLVM is aimed to be kind of compiler construction toolkit, so you don't need to know assembler neither processor architecture, you just write kind of meta-assembler (LLVM code) looking like FORTRAN
19:14:48MaledictusI would simply compile to C or C++, another compiler can do the asm part then. to ugly ;)
19:16:44benburkert enters the room.
19:16:55rby enters the room.
19:17:43boyscout2 commits by Marius Nuennerich
19:17:44boyscout * Use EnvSpecs where possible; d72c609
19:17:45boyscout * Use EnvSpecs module for platform dependent stuff; 8ccdf2d
19:17:48AndrewO leaves the room.
19:19:08OnObtw. does rubinius support real threading (running in parallel) ?
19:19:24Maledictusafaik there is still a giant lock
19:19:24scoopryou can compile llvm-ir as c too :P
19:20:02brixenOnO: rubinius has green threads, like 1.8
19:20:12djwhittyou can run multiple VMs though
19:20:15djwhitteach in it's own thread
19:20:19djwhitt(I think)
19:20:23brixenyeah
19:20:41OnOhmm... I was thinking once how to avoid giant lock.... wouldn't be simpler to make a mutex for each object for state modifications ?
19:21:00OnOyou don't need lock when you read, just when you write
19:21:26OnOso one lock per object when object state such as instance variable or methods are modified
19:21:57OnOsimple object such as ints or floats can be modified with cpu thread safe instructions
19:22:23AndrewO enters the room.
19:22:56MaledictusWhy wouldn't you need a lock for reading?
19:23:23djwhittI thought one of the major selling points of stacklessness was that you don't need a giant lock, but I'm a bit out of my depth talking about such things so I could be completely wrong
19:24:59OnOanyway stack is just used for local execution and local variables ? isn't it ? plus, it is separate stack per thread
19:25:05MaledictusHmm, where is rubuildius_amd64 ?
19:25:22djwhittMaledictus: down for the time being. I'm having trouble with extremely long CI runs again
19:26:00Maledictushmm, ok. ci time here is fine though. any tutorial available how to setup a rubuildius?
19:27:08djwhitthttp://github.com/djwhitt/rubuildius/tree/master
19:27:19djwhittsorry, README is a little outdated
19:27:26Maledictusnp, thanks
19:27:40cremesMaledictus: djwhitt is refining rubuildius so we can have master & slave bots; when done all the slaves will report to one (or more) master(s) so we don't
19:27:48cremesoverrun this channel with buildbot results
19:27:59Maledictusah, cool
19:28:06cremesdjwhitt: right?
19:28:22djwhittcremes: yeah, haven't had much time to work on it recently, but that's the goal
19:28:28cremesdjwhitt: i don't mean to sign you up for something you don't plan to do...
19:28:39cremesah, good
19:29:13djwhittcremes: you should be able to upgrade to the most recent code now if you like
19:29:26cremesdjwhitt: you pushed it to master?
19:29:41cremesor do i need to grab it from your git repository?
19:29:52djwhittcremes: grab a copy from my git repo
19:29:57cremeswill do tonight
19:30:06djwhittcremes: cool, I'll be around if you have any trouble
19:30:12rubuildius_ppcMarius Nuennerich: d72c609ce; 1995 files, 6503 examples, 22643 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180619
19:30:15cremesgreat, thanks!
19:30:48Maledictuspuh :)
19:31:05djwhittoh, and if anyone else has time to work on rubuildius get a github account, submit a patch and I'll give you access to the repo
19:31:38thehcdreamer enters the room.
19:31:54peeja leaves the room.
19:39:16peeja enters the room.
19:39:28dgtized enters the room.
19:40:32peeja leaves the room.
19:41:31dodecaphonic enters the room.
19:44:04imajes enters the room.
19:44:42benny enters the room.
19:44:44MaledictusHmm, what do I have to do so I can assign myself to tickets/close them?
19:45:04trythil leaves the room.
19:45:08benny leaves the room.
19:45:36trythil enters the room.
19:45:49wmoxam leaves the room.
19:46:22wmoxam enters the room.
19:46:50djwhittyou probably just don't have access to do that
19:46:50brixenMaledictus: what email are you using on LH?
19:47:01Maledictusmarius@nuenneri.ch
19:47:09brixenk, one sec..
19:47:36d2dchat leaves the room.
19:47:44brixendang, application error from LH trying to add a member :P
19:48:07Maledictusdoh. doesn't want me :)
19:48:16brixenheh
19:48:20brixenjust sent it
19:49:30Maledictusoh, we have a few more Enumerable methods...
19:50:01djwhittI wonder when the mythical LH 2.0 release is going to happen
19:50:01Maledictusbrixen: thanks :)
19:50:10brixensure thing
19:50:26brixendjwhitt: in about a man month :)
19:52:20DefilerMan, the weather is painfully awesome here today
19:52:43brixenDefiler: good to know, rain here :P
19:53:01Maledictusrain for weeks here
19:53:18Defilerhttp://www.wunderground.com/cgi-bin/findweather/getForecast?query=32312
19:53:25DefilerIt looks like a postcard outside. Crazy
19:53:34brixenDefiler: although, for our 3700 vertical foot hike yesterday (including snowshoing at the top and some 4-point scrambling) it was *beautiful*
19:54:08DefilerNice
19:54:27brixenlost the trail out of the snow and went bushwhacking for about an hour
19:54:37brixendown steep terrain. that was.. fun :)
19:58:25peeja enters the room.
19:59:01wmoxam leaves the room.
20:02:30yaroslav leaves the room.
20:02:52naeu leaves the room.
20:03:02naeu enters the room.
20:06:03Maledictusbrixen: do you delete bin/completeness?
20:09:43brixenMaledictus: not yet, but soon
20:09:56Maledictusthen I'll assign you to the ticket :)
20:09:59brixenfixing tags for windows
20:10:04brixenheh, ok
20:11:28brainopia enters the room.
20:15:51OnO leaves the room.
20:20:07josbtarcieri: you here?
20:24:53brixenVVSiz: how are you running the specs on windows? using cmd or something like cygwin bash?
20:25:42VVSizbrixen: with cygwin it was possible to run specs for some time... lately I been running them with CMD
20:26:27brixenok, it's running for me in cygwin but not cmd. what's your command look like?
20:27:29VVSizbrixen: ruby mspec/bin/mspec -t jruby.bat spec/ruby/1.8/core/array
20:27:54thehcdreamer leaves the room.
20:28:18brixenVVSiz: ok, hmm
20:29:07brixenI'm doing ruby mspec/bin/mspec -tr spec/ruby/1.8/core/array and it runs because I get the ruby -v line, but no specs
20:29:23brixenmakes another notch in his hate windows stick
20:29:51MaledictusHmm, how do I apply a patch so that the From is maintained and I just sign-off-by ?
20:29:54DefilerWe're probably not doing the right thing, when it comes to locating the 'engine' passed to -t
20:30:11DefilerYou can use 'git am' if the patch is in mail format already
20:30:24VVSizbrixen: there are some quirks I can't explain (yet). If I run mspec itself with MRI, I have to specify jruby as jruby.bat
20:30:28brixenMaledictus: git am -s
20:30:41VVSizbut if I run mspecs with jruby, the target must be specified as jruby (no .bat)
20:30:51brixenVVSiz: seems to not be globbing _spec.rb correctly
20:31:08brixenVVSiz: i.e. if I specify the actual file, it runs specs. otherwise, it doesn't run anything
20:31:20Maledictusthanks!
20:32:11VVSizbrixen: weird, it works for me
20:33:18imajes leaves the room.
20:33:25d2dchat enters the room.
20:33:44boyscout1 commit by rnicholson
20:33:45boyscout * Solaris should use 'gmake'.; 97188f4
20:35:34imajes enters the room.
20:37:00headius enters the room.
20:37:01nicksieger leaves the room.
20:38:32wmoxam enters the room.
20:41:30brixenVVSiz: ugh, you can't mix \\ and / in a path
20:41:55VVSizwell, I'm happy that I can use '/' in windows paths at all! :)
20:42:12brixenVVSiz: I was actually doing ruby mspec\bin\mspec -tr spec\ruby\1.8\core\array
20:42:29brixenand that was getting concated with /**/*_spec.rb
20:42:47VVSiznice
20:43:04brixenI suppose I could use #join on that instead
20:44:13brixenhah, nope
20:44:22brixenFile.join uses /
20:44:53rubuildius_ppcrnicholson: 97188f499; 1995 files, 6503 examples, 22643 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180667
20:44:53brixendecree: use / in paths that are args to mspec runners :D
20:45:30DefilerYeah, just use / always
20:45:40DefilerMRI converts them to backslashes correctly on Windows, as must we
20:45:46djwhittactually '/' works in most cases in Windows paths
20:45:52djwhitteven via the command line
20:46:24Defileryah
20:46:43Defilerthey should switch. It's not like there are THAT many command-line programs in Windows that people still use
20:46:58Defiler..and most of those already take -x style options anyway
20:47:07Defilerit's tracert -d google.com, not /d
20:47:44djwhittyeah, I imagine there are plenty of batch files out there that still rely on / style options though
20:48:07brixenanyone know the story about how msdos/windows ended up with \ in the path anyway?
20:48:32brixenit's not like either of those os-ish things were breaking ground
20:50:06djwhitthttp://blogs.msdn.com/larryosterman/archive/2005/06/24/432386.aspx
20:50:09djwhittthe internet knows
20:50:19brixenheh
20:50:36Defilerrelated, and also fun: http://blogs.msdn.com/michkap/archive/2005/09/17/469941.aspx
20:52:24Defilerhttp://www.thocp.net/biographies/bemer_bob.htm
20:53:45DefilerI had never really wondered about the backslash and how it got on the keyboard.
20:54:26DefilerThough I am totally going to call it 'reverse solidus' now
20:55:53headius_ enters the room.
20:56:12brixenI wonder if it needs to be all caps, to really give the sense of solidus
20:56:37wubo leaves the room.
20:56:54DefilerInteresting point
20:57:08DefilerSOLIDUS would be a sweet vanity license plate
20:58:15brixenRVRSLDS
20:58:21brixen:) and with that, lunch
21:00:27headius_ leaves the room.
21:08:50nicksieger enters the room.
21:11:15peeja leaves the room.
21:11:34headius leaves the room.
21:16:33ezmobius enters the room.
21:17:27be9_ leaves the room.
21:18:28tokengeek enters the room.
21:19:05anonuser enters the room.
21:22:17RyanTM leaves the room.
21:23:13Arjen_ leaves the room.
21:26:43RyanTM enters the room.
21:30:30MaledictusIs ltmain.sh autogenerated?
21:33:28peeja enters the room.
21:36:48imajes leaves the room.
21:37:38macournoyer leaves the room.
21:37:53jlindley leaves the room.
21:39:21boyscout1 commit by rnicholson
21:39:23boyscout * Fix syntax to be more portable; 19dfce0
21:39:57imajes enters the room.
21:42:03dlee enters the room.
21:43:16ctennis enters the room.
21:43:32wubo enters the room.
21:43:40seydar enters the room.
21:46:34rueMorning, kids
21:46:56seydarrue, aren't you on the east coast?
21:47:03rueMaybe
21:47:40Form enters the room.
21:48:07seydarits almost 5 here
21:48:17seydarman i wish i had your night life
21:48:18djwhittrue is on his own time I think
21:48:40peeja_ enters the room.
21:49:15peeja leaves the room.
21:50:54rubuildius_ppcrnicholson: 19dfce0f4; 1995 files, 6503 examples, 22643 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180705
21:52:31seydarso.... gnu lightning...
21:52:34seydarwas used for FFI?
21:53:12djwhitta while ago
21:53:27djwhitteverything was switched to libffi
21:53:39seydarisn't it used for JIT
21:54:30DefilerWe don't do JIT
21:54:49boyscout1 commit by Gerardo Santana
21:54:50boyscout * libexecinfo reference removed; 289bc77
21:56:15seydarwhy no JIT? isn't it possible, though? like maybe not in the current form of shotgun...
21:57:12DefilerSure, it is possible
21:57:25DefilerIt is just hard
21:57:31DefilerFeel free to add it :)
21:57:32seydarwhy so hard?
21:58:11brixenseydar: we have sendsite, we need some machinery to figure out traces to jit
21:58:13tarcieriSounds like something good to add after evan's happy about shotgun's architecture
21:58:14tarcieriheh
21:58:19Defilercross-platform compilation to high-quality machine code?
21:58:24brixenseydar: you don't just git everything
21:58:26DefilerIs pretty high-effort
21:58:31brixenseydar: er jit rather
21:59:00djwhittI'm pretty sure there are other opts that would make sense to do first too
21:59:15seydarI was reading PyPy's docs on JIT, and they made it seem pretty simple
21:59:39seydarshotgun is eventually going to be rewritten into Cuby, right?
22:00:32DefilerNot any time soon
22:00:44Gerardo enters the room.
22:01:15seydarwell isn't it just the fact the dialect isn't invented yet?
22:01:48djwhittwell, you could invent the dialect, but you still have to rewrite everything in it and write a compiler for it
22:02:32seydarand isn't that Evan's eventual plan?
22:03:00DefilerSomething like that, yeah. Probably closer to the OMeta / IS concept
22:03:09DefilerAt least, that's what I would vote for
22:03:23seydarwhats OMeta? and whats IS?
22:03:36DefilerFirst we have to be 100% MRI compatible before worrying about these things
22:03:56Defilerhttp://www.cs.ucla.edu/~awarth/ometa/
22:04:50DefilerYou can read about IS here: http://www.vpri.org/pdf/steps_TR-2007-008.pdf
22:05:09DefilerThat just seems like the right way to bootstrap your way into not needing C, to me
22:05:13seydarDefiler, I need a life. but thats not going to happen anytime soon. same with me learning C to any degree to be useful. But i can at least get some foundations for Cuby
22:05:47seydari was thinking about linking it in to LLVM. is there anything wrong with that idea?
22:05:50rubuildius_ppcGerardo Santana: 289bc77d9; 1995 files, 6503 examples, 22643 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/180710
22:07:03DefilerI personally think it is too complicated
22:07:09Defilerbut people seem to like talking about it
22:07:30seydarI looked at the OCaml bindings - they look pretty sweet and simple
22:07:56DefilerSure, I meant that LLVM itself is complicated
22:07:59Gerardohi
22:08:26seydarhey Gerardo
22:08:47Gerardoifeq doesn't have an 'elseif', does it?
22:08:54Gerardoehm, in a Makefile, I mean
22:09:03GerardoGNU Make
22:09:31GerardoI'm planning to modify external_libs/*/Makefile to accomodate SunOS
22:10:11Gerardowhich, by the way, looks _dirty_ to me
22:10:19Gerardobut it's already done for Darwin
22:10:29seydartry it out
22:10:33Gerardoshouldn't it be a job for configure?
22:10:48DefilerProbably. Feel free to move it
22:11:16Gerardook, I'll give it a thought
22:13:45seydarwhat restrictions would be reasonable in Cuby?
22:13:52seydarcan we call is Spherey?
22:14:09DefilerActually, we have been calling it 'Garnet'
22:14:15Defilermore recently
22:14:24seydarSpherey is my contribution
22:14:31DefilerThe restrictions should be driven by the output
22:14:53DefilerAs in, what can we let the programmer write and still get simple, static machine code out of the other end
22:15:14seydarso basically, a wrapper for C. new syntax
22:16:19DefilerWe don't have anything that can turn C into machine code
22:16:19seydaractually i'm lying
22:16:25DefilerSo it has to be more than that
22:16:31seydaris a C compiler not allowed?
22:17:04wycats leaves the room.
22:17:19ezmobius leaves the room.
22:17:31DefilerSomething that generated C code would be interesting, in a trivial sort of way, for rewriting the VM
22:17:33MaledictusGerardo: I'm not sure, but isn't ltmain.sh generated?
22:17:39wycats enters the room.
22:17:52Defilerbut I don't think it would be good enough to let us implements parts of the kernel in it. Maybe I am wrong
22:18:28seydari have to go. we'll talk later
22:18:32seydarare you going to GoRuCo?
22:18:35DefilerYep
22:18:42DefilerNeed to figure out a hotel, though
22:18:54DefilerHaving trouble finding one that isn't $400/night. Hah
22:18:58seydari'll see if i can smuggle myself in somehow.... shhh, don't tell
22:18:58skaar leaves the room.
22:19:08seydarDefiler: sleep in your car!
22:19:18GerardoMaledictus: the Makefile? not for libbstring at least
22:19:18seydarshower at the gym
22:19:43Maledictusltmain.sh, the patch in ticket #493
22:19:59seydar leaves the room.
22:20:10GerardoMaledictus: ah
22:20:10wycats leaves the room.
22:20:34tarcieriLLVM is the ultimate panacea to all woes
22:20:34tarcieriheh
22:20:37GerardoMaledictus: it is, from autotools I suppose, but I didn't found the original macros file
22:20:44wycats enters the room.
22:20:46GerardoMaledictus: and ltmain.sh is in the repository
22:21:02GerardoMaledictus: that's the reason of the patch against ltmain.sh
22:21:25brixentarcieri: yeah, I keep search the api docs for the doeverythingformekthxbye function
22:21:38brixentarcieri: haven't found it yet
22:21:51Maledictusok. I do not really know libtool. But I thought maybe it would be better to patch the source of that file...
22:21:55tarcieriheh
22:22:00GerardoMaledictus: that patch fixes libtool, which is generated from ltmain.sh
22:22:36GerardoMaledictus: indeed. It surprises me that it is not in the git repository. I suppose nobody planned to made modifications to it.
22:23:19josbtarcieri: see my email about rev's EventMachine emulation layer?
22:24:24tarcierijosb: nope, where did you send it?
22:24:36josbtarcieri: let me check my outbox
22:25:18josbtarcieri: tony at medioh.com. I basically replied to your initial email.
22:26:08rueDefiler: A C-generating backend to the compiler would not be too hard, assuming a restricted-semantics Ruby subset
22:26:32Defilerrue: Yeah, I'm just not sure it would be that exciting. Maybe I am missing the value
22:26:51rueDefiler: On the other hand, it might not be too much more cumbersome to generate an IR
22:26:56DefilerIt seems to me that we want to be able to generate machine code to implement a simple Ruby method at runtime
22:26:57tarcierijosb: hmm, I don't think I got it :/
22:27:32josbtarcieri: I can resend it, what email address should I use?
22:27:51tarcierioh weird
22:27:53tarcierithere it is
22:27:53tarcieriuhh
22:28:08tarcieriwonder why I didn't see it before
22:28:26josbtarcieri: I cc'ed Kirk, too.
22:28:29tarcieriyeah
22:28:58tarcieriIf you want I can give you commit access to the Rev repo and you can just work on it from there
22:29:04tarcieriadding timers should be pretty simple
22:29:25josbtarcieri: I haven
22:29:40tarcierioh man, Swiftiply uses Deferrables
22:29:43tarcierithose will be fun
22:29:59josbtarcieri: I haven't made it through all the Rev and EM docs yet but I'm working on it...
22:30:22josbYeah. It looks complicated :-/
22:30:49tarcieriWhat I'd like to do is get the Reactor core in place and hopefully we'll be able to pull in all those kind of features from the pure Ruby EventMachine
22:31:27tarcieriThis should almost have its own project
22:31:28Gerardohello rue, do you know why the file from wich libbbstring/ltmain.sh was generated is not present in the git repository?
22:32:22josbtarcieri: that sounds like a great plan.
22:33:05tarcieriso you're patching Swiftiply?
22:34:06josbActually, I'm just using your eventmachine.rb and the idea was to extend it with the bits swiftiply needs. Once that works I figured I can add the bits to swiftiply that puppet needs.
22:34:21josbThe patch mostly fixes some small 1.9 compat issues.
22:34:25tarcieriOh
22:34:26tarcieriOk
22:34:40tarcieriSo do you have a new version of the Rev compatibility layer, or that was just an assessment of what's missing?
22:34:51josbThere may be more in the code but these are the ones I've seen so far.
22:35:04tarcieriI can add timers pretty easily
22:35:13josbThe latter, sadly, so far. I'm still looking at the docs
22:35:21tarcieriOk
22:35:37josbIt means having to grok swiftiply, EM and rev at the same time so it's a big job
22:35:46tarcieriHeh, yeah
22:35:56tarcieriI haven't really looked at Swiftiply's code
22:36:02tarcierihttp://rev.rubyforge.org/svn/contrib/revem/revem.rb
22:36:09tarcierithere's the Rev/EM compatibility layer in svn
22:36:14josbGood news is that I have a test setup in which I can test things easily.
22:36:47josbYeah, that's what I'm using for eventmachine.rb in swiftiply.
22:36:59ezmobius enters the room.
22:37:16josbJust copied it into my 1.9 $RUBYLIB as eventmachine.rb.
22:37:51tarcieriaah
22:38:01josbI also figured I'd just try 1.9 instead of mucking with 1.8. Figured that 1.9 is the main development target of Rev, not 1.8.
22:38:07tarcieriWell, I'll commit a new version later today with timers
22:38:25tarcieriYeah, I'm trying to support everything on 1.8 but it's kind of an afterthought
22:38:27josbGreat! I'll test that right away and will let you know.
22:38:31tarcieriRev will always be much slower on 1.8 than 1.9
22:38:37antares enters the room.
22:38:52josbRight. So I don't mind using 1.9, looks like the right thing to do long-term anyway.
22:39:02Maledictus leaves the room.
22:39:03tarcieriRubinius is the right thing in the long term :)
22:39:21tarcieriAnd hopefully I can keep the Reactor largely API compatible with Rev
22:39:27josb(As an intermediate step between running it under Rubinius :-) )
22:39:37josbYep.
22:39:37tarcieriEnough that the EventMachine compat layer can work on both
22:39:49josbpecks slowly
22:40:08tarcierithe Reactor I'm trying to do for Rubinius, that is
22:40:16tarcieribased on Rev
22:40:34josbThat would be sweet. Maybe we can convince wyhaines to port swiftiply to Rev...
22:41:01tarcieriHeh, well as cool as that be, I think an EM compatibility layer is the better way in the long run
22:41:06josbYes. This will be uber-cool, once all the pieces are in place.
22:41:07tarcieriThere's already a lot of EventMachine projects out there
22:41:33josbtarcieri: yeah, that's true. EM compatibility will be good.
22:41:56josbThanks for working on this.
22:42:46tarcieriYeah, np
22:43:44naeu leaves the room.
22:44:54peeja leaves the room.
22:48:56wmoxam leaves the room.
22:51:42stepheneb leaves the room.
22:53:50antares leaves the room.
22:54:18seydar enters the room.
22:54:36moofbong leaves the room.
22:57:24Gerardo leaves the room.
23:03:38wycats_ enters the room.
23:05:22brainopia leaves the room.
23:08:11agile leaves the room.
23:08:55Fullmoon leaves the room.
23:11:16dysinger enters the room.
23:18:48qwert666_ leaves the room.
23:19:54AndrewO leaves the room.
23:23:09mentz leaves the room.
23:26:34seydar leaves the room.
23:27:35madsimian leaves the room.
23:28:14jlindley enters the room.
23:32:29cypher23 le