Da
Show enters and exits. Hide enters and exits.
| 00:00:30 | Arjen_ leaves the room. | |
| 00:02:31 | dctanner enters the room. | |
| 00:03:28 | imajes leaves the room. | |
| 00:05:02 | fabiokung leaves the room. | |
| 00:05:29 | fabiokung enters the room. | |
| 00:11:10 | fabiokung leaves the room. | |
| 00:11:31 | fabiokung enters the room. | |
| 00:11:44 | fabiokung leaves the room. | |
| 00:12:10 | atduskgr_ enters the room. | |
| 00:14:46 | atduskgreg leaves the room. | |
| 00:17:42 | chris2 leaves the room. | |
| 00:18:08 | dctanner leaves the room. | |
| 00:19:00 | ch0wda leaves the room. | |
| 00:19:27 | enebo leaves the room. | |
| 00:19:44 | twbray enters the room. | |
| 00:20:44 | AndrewO leaves the room. | |
| 00:23:28 | michalw leaves the room. | |
| 00:25:40 | Yurik leaves the room. | |
| 00:30:50 | evan | brixen: you around? |
| 00:32:27 | Yurik enters the room. | |
| 00:35:18 | brixen | evan: yah |
| 00:35:32 | evan | open up shotgun/lib/array.c |
| 00:35:36 | brixen | k |
| 00:35:37 | eventualbuddha enters the room. | |
| 00:35:52 | evan | line 8 |
| 00:35:59 | evan | you added the count < 8 ? ... thing |
| 00:36:00 | brixen | yes |
| 00:36:03 | evan | recall why? |
| 00:36:24 | brixen | yep, because a zillion arrays were created an then promptly promoted to ~8 elements |
| 00:36:28 | brixen | watching specs run |
| 00:36:46 | evan | hrm. ok. |
| 00:36:52 | brixen | but that was before lookuptable |
| 00:37:00 | brixen | hmm, nm |
| 00:37:12 | brixen | Hash uses Tuple, not Array |
| 00:37:19 | evan | yep. |
| 00:38:37 | brixen | it would be great to get our dtrace hooks into the new vm so we could profile this stuff |
| 00:38:47 | evan | we have dtrace hooks in the current one |
| 00:38:50 | brixen | I never got a chance at RC to work on that with crafterm |
| 00:38:50 | evan | you can do it now. |
| 00:38:56 | brixen | yeah, I know |
| 00:39:02 | brixen | but I want them in the new one :) |
| 00:39:05 | evan | heh |
| 00:39:12 | evan | just pondering splat a little. |
| 00:39:18 | brixen | ahh |
| 00:39:26 | evan | esp. for a case like |
| 00:39:30 | evan | def call(*args) |
| 00:39:34 | evan | @block.call(*args) |
| 00:39:35 | evan | end |
| 00:39:50 | brixen | yeah, some way to pass it though, huh? |
| 00:40:05 | evan | in the new VM, i'm pretty sure it will be easy. |
| 00:40:26 | evan | because there is the Message C++ class that can just lazily pass args through without copying them. |
| 00:40:31 | evan | in the current one |
| 00:41:01 | evan | i was curious how much memory a splat took up |
| 00:41:13 | evan | and found that it's at least 8 slots in the tuple |
| 00:44:11 | brixen | do we have any measure of how frequent a splat occurs in code? |
| 00:44:40 | brixen | we need a query env for statistics from the vm during execution :) |
| 00:45:03 | brixen | select * from splat where class = Array; :D |
| 00:45:10 | brixen | s/splat/splats/ |
| 00:45:35 | brixen | we should throw execution stats into sqlite during a run |
| 00:45:43 | brixen | query-able offline |
| 00:46:07 | brixen | evan: you talked about opening a write-out port to do this once |
| 00:46:33 | evan | i did. |
| 00:46:45 | evan | in a similar way that the jvm has one |
| 00:46:56 | brixen | but does jvm use sqlite? :) |
| 00:47:04 | evan | it would pump data out a socket |
| 00:47:16 | brixen | sure, and an adapter could read and dump to sqlite |
| 00:47:19 | brixen | that'd be fun |
| 00:47:25 | evan | yeah |
| 00:47:28 | brixen | ruport and gruff, here we come |
| 00:47:36 | evan | it's super easy to do |
| 00:47:44 | evan | we just have to decide what data we want. |
| 00:47:52 | brixen | yeah |
| 00:48:05 | brixen | obj creation/gc would be awesome |
| 00:48:31 | brixen | accumulate sendsite stats at intervals |
| 00:48:47 | evan | probably have to accumulate gc stats too. |
| 00:49:19 | thoughtfiz leaves the room. | |
| 00:49:35 | thoughtfiz enters the room. | |
| 00:50:17 | drbrain | ko1_away: added total malloc'd memory and total malloc alloctations to 1.9 (must be compiled-in) |
| 00:50:51 | evan | yeah |
| 00:50:56 | evan | those are easy |
| 00:51:15 | bitsweat | drbrain: got a patch up? |
| 00:51:17 | evan | we could write those to a socket everytime gc runs |
| 00:51:20 | evan | before and after |
| 00:51:28 | evan | so you get figure out stats on how well the gc is working too. |
| 00:51:39 | drbrain | oops, I didn't mean a ':' there |
| 00:51:48 | bitsweat | oh heh |
| 00:52:10 | bitsweat | total = currently malloced |
| 00:52:15 | bitsweat | not cumulative |
| 00:52:44 | drbrain | but, here is your patch: http://svn.ruby-lang.org/cgi-bin/viewvc.cgi/trunk/gc.c?r1=17018&r2=17017&diff_format=u |
| 00:53:14 | drbrain | (written by ko1_away) |
| 00:53:21 | evan | brixen: http://www.iunknown.com/2008/06/ironruby-at-tech-ed-2008.html |
| 00:53:30 | evan | brixen: sees our PositiveExpectation lives on! |
| 00:53:34 | evan | seems |
| 00:54:02 | bitsweat | ya, using that in ruby-prof now |
| 00:54:09 | bitsweat | if only other shit wasn't broken, it'd rock |
| 00:58:52 | Yurik leaves the room. | |
| 01:00:12 | brixen | heh |
| 01:00:41 | evan | brixen: he wrote that on stage... it looks so familiar... :D |
| 01:01:06 | brixen | indeed it does |
| 01:02:13 | dfg59 leaves the room. | |
| 01:02:16 | brixen | here's the original: http://pastie.org/198028 |
| 01:02:47 | evan | we called them PositiveExpectation and NegativeExpectation at one point too though, didn't we? |
| 01:03:21 | brixen | yeah |
| 01:04:00 | NoKarma leaves the room. | |
| 01:21:29 | boyscout | 4 commits by Eric Hodel |
| 01:21:30 | boyscout | * Descriptions for gems: tasks; d31ad8d |
| 01:21:31 | boyscout | * Wire in stub gem installation into the development environment.; b3179ef |
| 01:21:32 | boyscout | * Add stub fastthread gem and architecture for additional stub gems.; e2e87a0 |
| 01:21:33 | boyscout | * Death To Libtool\!; 2a6ee41 |
| 01:22:01 | AndrewO enters the room. | |
| 01:25:20 | dfg59 enters the room. | |
| 01:28:52 | loincloth enters the room. | |
| 01:30:55 | wyhaines leaves the room. | |
| 01:35:52 | rubuildius_ppc | Eric Hodel: d31ad8d93; 2377 files, 7769 examples, 26855 expectations, 0 failures, 0 errors |
| 01:36:05 | drbrain | (now you can install mongrel out-of-the-box! |
| 01:36:15 | tarcieri | orly |
| 01:36:28 | tarcieri | is that because there's stock dependencies for things like fastthread? |
| 01:37:17 | rubuildius_amd64 enters the room. | |
| 01:37:55 | drbrain | yes, for fastthread |
| 01:38:07 | drbrain | it's the only one we've added |
| 01:38:59 | tarcieri | ok |
| 01:40:48 | rubuildius_amd64 | Eric Hodel: d31ad8d93; build failed! http://rafb.net/p/i0qDmJ68.html |
| 01:41:13 | drbrain | if there are others we need, they are easy to add, look in the gems/ directory |
| 01:41:28 | drbrain | rubuildius_amd64: that's really strange |
| 01:41:51 | drbrain | can somebody run rubuildius_amd64 with -t? |
| 01:42:57 | djwhitt | drbrain: sure, just a sec |
| 01:43:20 | drbrain | djwhitt: thanks, it's odd that _ppc worked, but _amd64 didn't |
| 01:43:54 | twbray leaves the room. | |
| 01:44:38 | mernen | djwhitt: btw, regarding the two failures on amd64, the first one is too platform-dependant and is out of the specs; the other I'll investigate a bit more later |
| 01:44:55 | djwhitt | mernen: cool, thanks for taking a look at it |
| 01:45:15 | mernen | it really shouldn't happen, unless for some reason the iconv library on gentoo 64-bit is different from ubuntu 32 |
| 01:47:29 | djwhitt | drbrain: http://pastie.org/212701 |
| 01:48:13 | drbrain | maybe it requires in a different order between the two |
| 01:50:31 | djwhitt | btw if rubuildius is ever misbehaving when I'm not around brixen has access to the machine as well |
| 01:50:31 | hemulen leaves the room. | |
| 01:50:53 | boyscout | 1 commit by Eric Hodel |
| 01:50:53 | AndrewO leaves the room. | |
| 01:50:54 | boyscout | * Ensure rakelib/rubinius is required so LIBVER exists; 6743a90 |
| 01:51:30 | drbrain | let's try that out |
| 01:52:59 | djwhitt | same result. looks like that's actually being defined in configure.rake |
| 01:55:25 | drbrain | oh, shit |
| 01:55:49 | rubuildius_amd64 | Eric Hodel: 6743a90fc; build failed! http://rafb.net/p/8HkTwQ66.html |
| 01:58:04 | krisps leaves the room. | |
| 02:02:18 | yugui enters the room. | |
| 02:03:41 | jbarnette leaves the room. | |
| 02:03:45 | rubuildius_ppc | Eric Hodel: 6743a90fc; 2377 files, 7769 examples, 26855 expectations, 0 failures, 0 errors |
| 02:03:50 | mapar enters the room. | |
| 02:05:21 | jtoy enters the room. | |
| 02:05:42 | mapar leaves the room. | |
| 02:15:38 | boyscout | 1 commit by Eric Hodel |
| 02:15:39 | boyscout | * I'm dumb. Pull LIBVER and friends into configure.rb so they can be required.; 10e41f1 |
| 02:16:00 | drbrain | ok, THAT should do it |
| 02:19:00 | roo enters the room. | |
| 02:19:32 | rue leaves the room. | |
| 02:20:12 | roo enters the room. | |
| 02:23:29 | dfg59 leaves the room. | |
| 02:27:30 | djwhitt | drbrain: seems to have worked |
| 02:27:56 | djwhitt | drbrain: spec runs are taking a long time on rubuildius_amd64 these days, but I can tell that the build succeeded |
| 02:28:06 | roo leaves the room. | |
| 02:28:49 | eventualbuddha leaves the room. | |
| 02:28:49 | rubuildius_amd64 | Eric Hodel: 10e41f15f; 2377 files, 7770 examples, 26834 expectations, 1 failure, 1 error; http://rafb.net/p/MD3Ygb83.html |
| 02:28:59 | drbrain | it would have failed right away, since just rake itself was borken |
| 02:29:05 | rubuildius_ppc | Eric Hodel: 10e41f15f; 2377 files, 7769 examples, 26855 expectations, 0 failures, 0 errors |
| 02:29:27 | drbrain | I think evan's mistyping is contagious |
| 02:35:57 | rue | There are dissertations |
| 02:38:09 | VVSiz_ enters the room. | |
| 02:40:52 | atduskgreg enters the room. | |
| 02:41:33 | nicksieger leaves the room. | |
| 02:42:15 | nicksieger enters the room. | |
| 02:43:42 | yugui leaves the room. | |
| 02:45:28 | VVSiz leaves the room. | |
| 02:46:40 | atduskgr_ leaves the room. | |
| 02:49:43 | binary42 leaves the room. | |
| 02:58:42 | mernen leaves the room. | |
| 03:16:00 | bburcham enters the room. | |
| 03:17:09 | ezmobius leaves the room. | |
| 03:23:29 | jtoy leaves the room. | |
| 03:27:31 | stepheneb_ enters the room. | |
| 03:29:20 | stepheneb leaves the room. | |
| 03:32:36 | twbray enters the room. | |
| 03:33:17 | yugui enters the room. | |
| 03:36:58 | obvio enters the room. | |
| 03:37:32 | obvio leaves the room. | |
| 03:42:35 | stepheneb_ leaves the room. | |
| 03:44:35 | rphillips enters the room. | |
| 03:45:50 | dfg59 enters the room. | |
| 03:46:29 | obvio171 leaves the room. | |
| 03:58:11 | jero5 enters the room. | |
| 04:05:28 | jacen_ enters the room. | |
| 04:05:28 | jazen leaves the room. | |
| 04:11:48 | wmoxam enters the room. | |
| 04:15:05 | Cosmos95 leaves the room. | |
| 04:25:35 | lstoll enters the room. | |
| 04:26:13 | lstoll leaves the room. | |
| 04:29:23 | wycats leaves the room. | |
| 04:32:48 | twbray leaves the room. | |
| 04:33:09 | lstoll enters the room. | |
| 04:37:23 | shame leaves the room. | |
| 04:38:53 | vitaliy_ enters the room. | |
| 04:41:09 | wyhaines enters the room. | |
| 04:42:24 | bburcham leaves the room. | |
| 04:53:15 | lstoll leaves the room. | |
| 04:53:37 | lstoll enters the room. | |
| 04:56:14 | mapar enters the room. | |
| 04:58:06 | bburcham enters the room. | |
| 04:59:16 | bburcham leaves the room. | |
| 05:08:33 | wycats enters the room. | |
| 05:08:42 | yugui leaves the room. | |
| 05:09:03 | yugui enters the room. | |
| 05:09:48 | atduskgreg leaves the room. | |
| 05:10:51 | atduskgreg enters the room. | |
| 05:10:53 | stepheneb enters the room. | |
| 05:11:54 | vitaliy | nuby question: |
| 05:12:08 | vitaliy | while looking at some other stuff noticed |
| 05:12:25 | vitaliy | that fixnum in rubinius is up to 2**28-1 |
| 05:12:37 | vitaliy | compared to 2**30-1 in MRI |
| 05:12:42 | Defiler | Yep |
| 05:12:51 | Defiler | That is by design |
| 05:12:56 | vitaliy | gotcha |
| 05:13:01 | Defiler | Luckily, people already can't rely on the Fixnum cutoff in MRI |
| 05:13:08 | Defiler | because of 64bit vs. 32bit builds |
| 05:13:08 | loincloth leaves the room. | |
| 05:13:20 | Defiler | JRuby also has a non-MRI cutoff point, I believe |
| 05:13:42 | Defiler | shotgun/lib/oop.h has a little diagram (or should) showing what the space is used for |
| 05:13:50 | vitaliy | thanks |
| 05:14:31 | Defiler | We wanted to leave space for 'custom' immediate objects that we hadn't thought of yet |
| 05:22:58 | binary42 enters the room. | |
| 05:23:06 | wmoxam leaves the room. | |
| 05:24:00 | dfg59_ enters the room. | |
| 05:24:31 | dfg59_ leaves the room. | |
| 05:27:02 | twbray enters the room. | |
| 05:29:57 | rue | vitaliy_: Might change in the future too, of course |
| 05:30:33 | shame enters the room. | |
| 05:36:36 | lchin enters the room. | |
| 05:38:45 | dfg59 leaves the room. | |
| 05:45:19 | dfg59 enters the room. | |
| 05:46:44 | stepheneb leaves the room. | |
| 05:50:01 | ezmobius enters the room. | |
| 05:51:52 | dfg59 leaves the room. | |
| 05:52:00 | dfg59 enters the room. | |
| 05:57:46 | blakewatters leaves the room. | |
| 05:58:19 | dbussink enters the room. | |
| 06:04:25 | xhanjian leaves the room. | |
| 06:08:51 | dfg59 leaves the room. | |
| 06:11:28 | dfg59 enters the room. | |
| 06:11:42 | dfg59 leaves the room. | |
| 06:12:06 | mediogre enters the room. | |
| 06:13:06 | scudco leaves the room. | |
| 06:13:37 | scudco enters the room. | |
| 06:22:58 | vitaliy_ leaves the room. | |
| 06:24:49 | dfg59 enters the room. | |
| 06:27:56 | dfg59 leaves the room. | |
| 06:28:01 | dfg59 enters the room. | |
| 06:30:40 | dfg59 leaves the room. | |
| 06:32:14 | dfg59 enters the room. | |
| 06:33:01 | sambo82 enters the room. | |
| 06:34:11 | dfg59_ enters the room. | |
| 06:34:41 | dfg59_ leaves the room. | |
| 06:34:47 | dfg59 leaves the room. | |
| 06:40:11 | sambo82_ enters the room. | |
| 06:41:00 | qwert666 enters the room. | |
| 06:41:37 | sambo82 leaves the room. | |
| 06:49:28 | krisps enters the room. | |
| 06:54:11 | JonathanT enters the room. | |
| 06:54:15 | dysinger leaves the room. | |
| 06:59:38 | random8r enters the room. | |
| 07:03:55 | JonathanT leaves the room. | |
| 07:07:39 | mediogre leaves the room. | |
| 07:07:55 | mediogre enters the room. | |
| 07:07:57 | cschneid leaves the room. | |
| 07:08:23 | cschneid enters the room. | |
| 07:26:29 | jtoy enters the room. | |
| 07:30:51 | Maledictus enters the room. | |
| 07:35:57 | yugui leaves the room. | |
| 07:51:24 | random8r leaves the room. | |
| 08:06:23 | yugui enters the room. | |
| 08:20:27 | krisps leaves the room. | |
| 08:22:22 | TheVoice leaves the room. | |
| 08:24:38 | imajes enters the room. | |
| 08:32:58 | NoKarma enters the room. | |
| 08:35:42 | olabini enters the room. | |
| 08:37:47 | dysinger enters the room. | |
| 08:39:29 | JonathanT enters the room. | |
| 08:40:58 | bitsweat leaves the room. | |
| 08:44:43 | w1rele55 enters the room. | |
| 08:44:47 | w1rele55 leaves the room. | |
| 08:44:53 | ezmobius leaves the room. | |
| 08:49:41 | Ski1 enters the room. | |
| 08:51:20 | mutle enters the room. | |
| 09:00:36 | naeu enters the room. | |
| 09:10:18 | twbray leaves the room. | |
| 09:16:09 | rue leaves the room. | |
| 09:19:15 | lstoll leaves the room. | |
| 09:41:36 | Fullmoon enters the room. | |
| 09:42:07 | mapar leaves the room. | |
| 09:59:00 | BlackEdder enters the room. | |
| 10:16:03 | michalw enters the room. | |
| 10:25:35 | yugui leaves the room. | |
| 10:25:56 | imajes leaves the room. | |
| 10:30:57 | BlackEdder leaves the room. | |
| 10:34:42 | BlackEdder enters the room. | |
| 10:40:47 | krisps enters the room. | |
| 10:41:35 | krisps_ enters the room. | |
| 10:49:20 | mediogre leaves the room. | |
| 10:51:00 | gnufied enters the room. | |
| 10:51:58 | jtoy leaves the room. | |
| 10:53:16 | chris2 enters the room. | |
| 10:59:14 | yugui enters the room. | |
| 11:03:03 | krisps leaves the room. | |
| 11:07:52 | dysinger leaves the room. | |
| 11:12:19 | krisps_ leaves the room. | |
| 11:42:59 | thehcdreamer enters the room. | |
| 11:54:26 | Cosmos95 enters the room. | |
| 11:55:37 | Cosmos95 leaves the room. | |
| 12:03:38 | qwert666 leaves the room. | |
| 12:04:44 | yugui leaves the room. | |
| 12:19:43 | krisps enters the room. | |
| 12:19:44 | Fullmoon leaves the room. | |
| 12:20:06 | krisps leaves the room. | |
| 12:21:09 | ch0wda enters the room. | |
| 12:21:30 | Yurik enters the room. | |
| 12:21:56 | olabini leaves the room. | |
| 12:21:57 | cezarsa leaves the room. | |
| 12:25:30 | naeu leaves the room. | |
| 12:29:40 | naeu enters the room. | |
| 12:30:05 | yugui enters the room. | |
| 12:30:38 | Fullmoon enters the room. | |
| 12:33:09 | naeu leaves the room. | |
| 12:34:17 | atduskgreg leaves the room. | |
| 12:37:19 | naeu enters the room. | |
| 12:43:03 | NoKarma leaves the room. | |
| 12:43:06 | Fullmoon leaves the room. | |
| 12:48:37 | cezarsa enters the room. | |
| 12:57:02 | chris2 leaves the room. | |
| 12:58:16 | wmoxam enters the room. | |
| 13:00:18 | gnufied leaves the room. | |
| 13:08:23 | vitaliy_ enters the room. | |
| 13:14:09 | wvdschel enters the room. | |
| 13:16:05 | wmoxam leaves the room. | |
| 13:16:26 | gnufied enters the room. | |
| 13:20:47 | binary42 leaves the room. | |
| 13:24:52 | wmoxam enters the room. | |
| 13:38:20 | obvio171 enters the room. | |
| 13:41:02 | brainopia enters the room. | |
| 13:42:26 | gnufied_ enters the room. | |
| 13:42:30 | gnufied leaves the room. | |
| 13:48:22 | obvio171 leaves the room. | |
| 13:50:43 | brainopia leaves the room. | |
| 13:51:23 | brainopia enters the room. | |
| 13:55:09 | mutle_ enters the room. | |
| 13:55:42 | mutle__ enters the room. | |
| 13:59:24 | olabini enters the room. | |
| 14:05:24 | brainopia leaves the room. | |
| 14:05:31 | AndrewO enters the room. | |
| 14:05:36 | brainopia enters the room. | |
| 14:05:40 | brainopia leaves the room. | |
| 14:06:20 | brainopia enters the room. | |
| 14:06:31 | brainopia leaves the room. | |
| 14:07:04 | brainopia enters the room. | |
| 14:07:07 | brainopia leaves the room. | |
| 14:11:11 | mutle leaves the room. | |
| 14:12:46 | mutle_ leaves the room. | |
| 14:14:02 | wvdschel leaves the room. | |
| 14:14:06 | yukito enters the room. | |
| 14:15:43 | blakewatters enters the room. | |
| 14:17:10 | hemulen enters the room. | |
| 14:18:20 | jazen enters the room. | |
| 14:20:21 | qwert666 enters the room. | |
| 14:24:46 | webmat enters the room. | |
| 14:26:44 | checkmark enters the room. | |
| 14:27:33 | robertpostill enters the room. | |
| 14:31:40 | robertpostill leaves the room. | |
| 14:34:57 | mutle enters the room. | |
| 14:35:36 | jacen_ leaves the room. | |
| 14:36:08 | olabini leaves the room. | |
| 14:39:36 | lchin leaves the room. | |
| 14:40:00 | moofbong enters the room. | |
| 14:42:35 | yukito leaves the room. | |
| 14:49:22 | wvdschel enters the room. | |
| 14:50:22 | mkrauskopf enters the room. | |
| 14:50:40 | pauldix enters the room. | |
| 14:50:50 | mutle__ leaves the room. | |
| 14:51:59 | pauldix leaves the room. | |
| 14:54:06 | pauldix enters the room. | |
| 15:17:57 | wmoxam_ enters the room. | |
| 15:18:19 | wmoxam_ leaves the room. | |
| 15:25:35 | twbray enters the room. | |
| 15:36:56 | twbray leaves the room. | |
| 15:39:39 | mkrauskopf leaves the room. | |
| 15:59:15 | enebo enters the room. | |
| 16:04:04 | nicksieger leaves the room. | |
| 16:05:45 | yukito enters the room. | |
| 16:06:43 | wvdschel leaves the room. | |
| 16:09:57 | nicksieger enters the room. | |
| 16:16:46 | yukito leaves the room. | |
| 16:16:54 | Wim__ enters the room. | |
| 16:17:41 | tmornini enters the room. | |
| 16:21:40 | lopex enters the room. | |
| 16:38:49 | twbray enters the room. | |
| 16:47:16 | twbray leaves the room. | |
| 16:52:40 | twbray enters the room. | |
| 16:53:55 | tmornini_ enters the room. | |
| 16:53:57 | tmornini leaves the room. | |
| 16:56:29 | gnufied_ leaves the room. | |
| 16:56:54 | twbray leaves the room. | |
| 16:59:28 | shame leaves the room. | |
| 17:00:09 | twbray enters the room. | |
| 17:03:15 | cschneid_ enters the room. | |
| 17:08:32 | anteaya enters the room. | |
| 17:10:35 | careo enters the room. | |
| 17:13:30 | mutle leaves the room. | |
| 17:15:33 | rue enters the room. | |
| 17:16:00 | brixen | wb rue |
| 17:16:08 | rue | Helloes |
| 17:16:18 | brixen | are you east of the pond again? |
| 17:16:28 | sambo82_ leaves the room. | |
| 17:16:57 | rue | Nah, just having to use the laptop |
| 17:19:09 | rue | Been trying a few different solutions for the platform problem |
| 17:22:35 | brixen | ahh |
| 17:23:29 | rue | Everything seems to have its downsides |
| 17:26:48 | rue | The least intrusive seems to be the dichtomy of startup-time vs. later additions |
| 17:27:30 | binary42 enters the room. | |
| 17:28:11 | krisps enters the room. | |
| 17:32:24 | acfoeken enters the room. | |
| 17:41:02 | thehcdreamer leaves the room. | |
| 17:41:18 | atduskgreg enters the room. | |
| 17:41:57 | stepheneb enters the room. | |
| 17:44:24 | twbray leaves the room. | |
| 17:45:55 | acfoeken leaves the room. | |
| 17:46:13 | Wim__ enters the room. | |
| 17:46:16 | acfoeken enters the room. | |
| 17:47:31 | NoKarma enters the room. | |
| 17:50:21 | yukito leaves the room. | |
| 17:51:35 | krisps leaves the room. | |
| 17:51:53 | twbray enters the room. | |
| 18:07:04 | naeu leaves the room. | |
| 18:09:13 | dbussink leaves the room. | |
| 18:11:26 | shame enters the room. | |
| 18:13:27 | mapar enters the room. | |
| 18:16:28 | stepheneb_ enters the room. | |
| 18:16:51 | squeegy leaves the room. | |
| 18:16:57 | skweegee enters the room. | |
| 18:19:29 | atmos_ enters the room. | |
| 18:21:05 | dlee enters the room. | |
| 18:21:32 | ch0wda leaves the room. | |
| 18:21:40 | xmlhacker leaves the room. | |
| 18:22:12 | ch0wda enters the room. | |
| 18:22:20 | xmlhacker enters the room. | |
| 18:27:32 | blakewatters_ enters the room. | |
| 18:27:44 | enebo leaves the room. | |
| 18:31:17 | stepheneb leaves the room. | |
| 18:32:13 | blakewatters leaves the room. | |
| 18:33:48 | twbray leaves the room. | |
| 18:36:59 | JonathanT leaves the room. | |
| 18:41:17 | Guest60574 leaves the room. | |
| 18:41:44 | cschneid leaves the room. | |
| 18:41:56 | cschneid enters the room. | |
| 18:46:42 | thehcdreamer enters the room. | |
| 18:51:08 | acfoeken_ enters the room. | |
| 18:52:07 | thehcdreamer leaves the room. | |
| 18:52:50 | Wim__ leaves the room. | |
| 18:55:43 | ezmobius enters the room. | |
| 18:57:19 | ezmobius leaves the room. | |
| 18:58:47 | ezmobius enters the room. | |
| 18:58:53 | acfoeken leaves the room. | |
| 19:02:00 | ezmobius leaves the room. | |
| 19:02:22 | mapar leaves the room. | |
| 19:02:22 | atmos leaves the room. | |
| 19:02:22 | anonuser leaves the room. | |
| 19:03:34 | atmos enters the room. | |
| 19:03:34 | anonuser enters the room. | |
| 19:04:09 | loincloth enters the room. | |
| 19:06:33 | dysinger enters the room. | |
| 19:12:50 | twbray enters the room. | |
| 19:13:39 | dfg59 enters the room. | |
| 19:13:48 | thehcdreamer enters the room. | |
| 19:17:19 | dfg59 leaves the room. | |
| 19:17:40 | dfg59 enters the room. | |
| 19:17:45 | wmoxam_ enters the room. | |
| 19:18:11 | wmoxam leaves the room. | |
| 19:19:06 | acfoeken_ leaves the room. | |
| 19:19:14 | djwhitt enters the room. | |
| 19:19:31 | acfoeken enters the room. | |
| 19:21:31 | anonuser leaves the room. | |
| 19:21:31 | atmos leaves the room. | |
| 19:24:21 | atmos enters the room. | |
| 19:24:59 | anonuser enters the room. | |
| 19:27:52 | krisps enters the room. | |
| 19:33:24 | anteaya leaves the room. | |
| 19:34:16 | chopmo enters the room. | |
| 19:34:28 | acfoeken_ enters the room. | |
| 19:34:32 | careo leaves the room. | |
| 19:35:39 | careo enters the room. | |
| 19:37:10 | chop3 enters the room. | |
| 19:39:40 | acfoeken leaves the room. | |
| 19:39:42 | tmornini_ leaves the room. | |
| 19:46:06 | dbussink enters the room. | |
| 19:46:12 | dbussink leaves the room. | |
| 19:46:13 | dbussink- enters the room. | |
| 19:46:22 | fbuilesv | What's the current policy on updating the libs in StdLibs? |
| 19:47:39 | drbrain | updating? |
| 19:47:46 | drbrain | replacing with better versions? |
| 19:47:49 | fbuilesv | drbrain: yes |
| 19:48:00 | drbrain | only when backwards compatibility will mostly be maintained |
| 19:48:48 | fbuilesv | yeah, it should be a minor update on rbyaml |
| 19:52:24 | tarcieri | it'd be fun to rewrite net/http so it doesn't suck |
| 19:53:23 | drbrain | fbuilesv: we've got several custom changes in our rbyaml |
| 19:53:39 | drbrain | so you'll probably want to diff the rbyaml tree against its import |
| 19:53:49 | fbuilesv | drbrain: Oh I see, let me check the log |
| 19:53:56 | drbrain | we are bad open source developers |
| 19:54:41 | fbuilesv | heh |
| 19:55:10 | anonuser leaves the room. | |
| 19:55:13 | fbuilesv | There's at least one test case from rbyaml's tree broken, I wonder if it was introduced in those changes or if it's just an older version |
| 19:55:16 | tarcieri | bad open source developers foist their code on the community then disappear |
| 19:56:13 | JonathanT enters the room. | |
| 19:56:33 | anonuser enters the room. | |
| 19:56:48 | evan | is there an official rbyaml tree? |
| 19:56:51 | evan | i'd be happy to sync them up. |
| 19:56:54 | fbuilesv | evan: yes, on google code |
| 19:57:04 | evan | lets discuss with ola to keep them sync'd then. |
| 19:57:15 | fbuilesv | http://code.google.com/p/rbyaml/ |
| 19:57:17 | fbuilesv | cool |
| 19:57:24 | TheVoice enters the room. | |
| 19:59:39 | dfg59 leaves the room. | |
| 19:59:39 | atmos_ leaves the room. | |
| 20:01:18 | dfg59 enters the room. | |
| 20:01:34 | flori leaves the room. | |
| 20:03:01 | anonuser leaves the room. | |
| 20:03:37 | fbuilesv | mm fatal: bad default revision 'HEAD' when I do a git log on lib/, any idea of what could be wrong there? |
| 20:04:10 | krisps leaves the room. | |
| 20:04:41 | enebo enters the room. | |
| 20:04:44 | evan | fbuilesv: what command did you run? |
| 20:04:51 | fbuilesv | git log |
| 20:05:07 | fbuilesv | http://pastie.org/paste/213154 |
| 20:05:39 | krisps enters the room. | |
| 20:06:07 | anonuser enters the room. | |
| 20:06:22 | VVSiz | fbuilesv: if you have no local changes, try 'git co master' |
| 20:06:45 | fbuilesv | by co you mean checkout I guess? |
| 20:07:03 | VVSiz | yeah, I have a shortcut for chekcout :) |
| 20:07:11 | fbuilesv | Already on "master" |
| 20:07:17 | ezmobius enters the room. | |
| 20:08:09 | chr1s enters the room. | |
| 20:08:11 | chr1s | hey all! |
| 20:08:33 | VVSiz | fbuilesv: how about this one: git reset --hard master |
| 20:08:44 | chr1s | is there a simple way I could see the bytecode generated by rubinius (in an ascii form) |
| 20:08:46 | chopmo leaves the room. | |
| 20:09:02 | fbuilesv | VVSiz: nope |
| 20:09:26 | fbuilesv | I'm running all of this in the main dir. btw |
| 20:09:36 | VVSiz | right |
| 20:12:13 | imajes enters the room. | |
| 20:12:14 | flori enters the room. | |
| 20:13:15 | VVSiz | fbuilesv: how about this: 1) git co origin/master 2) git co master |
| 20:13:50 | fbuilesv | VVSiz: nope |
| 20:14:08 | VVSiz | well, then backup your current repo, and clone a new one :) |
| 20:14:10 | fbuilesv | is lib working fine for you? I could just get it again and see what's wrong with it |
| 20:14:11 | fbuilesv | ya |
| 20:15:54 | evan | chr1s: sure. |
| 20:16:08 | evan | chr1s: if you already have it compiled, you can do |
| 20:16:15 | evan | er. this is easiest: |
| 20:16:18 | evan | meth = def a |
| 20:16:19 | evan | ... |
| 20:16:20 | evan | end |
| 20:16:23 | evan | puts meth.decode |
| 20:22:03 | boyscout | 1 commit by Eric Hodel |
| 20:22:04 | boyscout | * Add some RDoc to Rubinius classes that didn't have any. pair: Evan Phoenix; 70b9ce8 |
| 20:24:23 | chr1s | evan: could I also do something like '1+1'.decode |
| 20:24:36 | wycats leaves the room. | |
| 20:24:41 | chr1s | I saw the Compiler class on the wiki, but that doesn't seem to exist in my version of rubinius |
| 20:25:01 | evan | chr1s: it should. |
| 20:25:08 | evan | it's available everywhere. |
| 20:26:01 | chr1s | evan: I'll paste a short session |
| 20:26:04 | evan | ok. |
| 20:26:17 | evan | let me work up a short method for ya too. |
| 20:26:35 | chr1s | [chris@gamma] rubinius-daily → shotgun/rubinius |
| 20:26:35 | chr1s | irb(main):001:0> Compiler |
| 20:26:35 | chr1s | => nil |
| 20:26:46 | evan | nil? thats... od. |
| 20:27:01 | evan | d |
| 20:27:02 | evan | do |
| 20:27:06 | evan | require 'compiler/compiler' |
| 20:27:26 | chr1s | yes, that works. weird. |
| 20:27:32 | chr1s | it's today's build |
| 20:27:37 | loop | strangely doesnt say that it's missing, but that it's nil |
| 20:27:41 | evan | yeah. |
| 20:27:51 | evan | i did some work at one point to let the system load 2 compilers at the same time |
| 20:27:56 | evan | perhaps that introduced some weirdness. |
| 20:28:16 | loop | maybe 'Compiler' hardcoded somewhere? |
| 20:28:19 | evan | it's best to just use |
| 20:28:26 | evan | Compile.compile_string "1 + 1" |
| 20:28:34 | evan | that uses the kernel's Compile module methods |
| 20:28:39 | evan | which lets it manage the compiler for ya |
| 20:28:50 | evan | looks like I broke compile_string raw though. |
| 20:28:58 | evan | it's expecting an eval environment. |
| 20:29:09 | chr1s | evan: yes, it looks like it's broken. |
| 20:29:25 | acfoeken_ leaves the room. | |
| 20:29:29 | evan | I should fix taht. |
| 20:29:44 | evan | if you put that code into a file, it's easy though. |
| 20:30:01 | elight leaves the room. | |
| 20:30:12 | evan | cm = Compile.compile_file "path" |
| 20:30:14 | evan | puts cm.decode |
| 20:30:58 | evan | this is pretty easy to fix |
| 20:31:00 | evan | one sec. |
| 20:31:27 | chr1s | evan: thanks |
| 20:32:54 | chr1s | evan: that's great, this works like a charm. |
| 20:33:24 | chr1s | I'm going to write a simple compiler tutorial for the ruby folks out there |
| 20:33:41 | evan | cool! |
| 20:33:42 | rubuildius_amd64 | Eric Hodel: 70b9ce855; 2377 files, 7770 examples, 26834 expectations, 1 failure, 1 error; http://rafb.net/p/DFhJBM82.html |
| 20:33:59 | chr1s | I'm just going to do a very simple language and target rubinius |
| 20:34:05 | evan | oh fun! |
| 20:34:17 | evan | check out lib/compiler/compiler.rb |
| 20:34:28 | evan | if you construct a sexp, you can just feed it to the Compiler |
| 20:35:49 | boyscout | 1 commit by Evan Phoenix |
| 20:35:50 | boyscout | * Support compiling a string with no context; 3c9dbc9 |
| 20:35:55 | evan | chr1s: there ya go |
| 20:36:02 | evan | cm = Compile.compile_string "1 + 1" |
| 20:36:04 | evan | should work now. |
| 20:36:34 | chr1s | evan: that's awesome, thanks |
| 20:36:46 | chr1s | evan: I would rather do it bytecode based |
| 20:36:52 | evan | ok |
| 20:36:53 | chr1s | so feed the 'assembly' code |
| 20:37:17 | evan | you can just call methods on Compiler::Generator to build up the bytecode |
| 20:37:22 | evan | thats the easist way |
| 20:37:27 | chr1s | nice |
| 20:37:29 | evan | because it will track labels for you |
| 20:37:38 | chr1s | cool! that's handy |
| 20:37:52 | chr1s | well, I'm going to play around with it and see what I can come up with. |
| 20:37:57 | evan | ok. |
| 20:38:06 | chr1s | I'll let it know once it's done |
| 20:38:09 | chr1s | thanks for your help! |
| 20:38:13 | evan | no prob. |
| 20:38:19 | wmoxam_ leaves the room. | |
| 20:38:24 | wmoxam enters the room. | |
| 20:38:28 | chr1s | is away now. |
| 20:39:12 | evan | brixen: you around? |
| 20:41:55 | twbray leaves the room. | |
| 20:48:16 | boyscout | 2 commits by Eric Hodel |
| 20:48:17 | boyscout | * Basic RDoc for kernel/platform. pair: Evan Phoenix.; 57e2fc3 |
| 20:48:18 | boyscout | * Add doc:rdoc task; 1c015aa |
| 20:48:40 | rubuildius_amd64 | Evan Phoenix: 3c9dbc9c3; 2377 files, 7770 examples, 26834 expectations, 1 failure, 1 error; http://rafb.net/p/98L0mW68.html |
| 20:48:51 | dysinger leaves the room. | |
| 20:49:52 | ezmobius | evan: the -dc flags to rbx don;t seem to work anymore |
| 20:50:42 | acfoeken enters the room. | |
| 20:51:13 | elight enters the room. | |
| 20:52:10 | evanlight enters the room. | |
| 20:52:12 | imajes leaves the room. | |
| 20:52:17 | rue | Code dumping flag? |
| 20:52:24 | ezmobius | yeah |
| 20:52:32 | ezmobius | if you do "rbx -dc irb" |
| 20:52:39 | ezmobius | it says [Compiler debugging enabled] |
| 20:52:48 | ezmobius | but it does not spit out the sexp or bytecodes anymore |
| 20:53:26 | imajes enters the room. | |
| 20:56:38 | dbussink | ezmobius: how about describe then? |
| 20:56:58 | ezmobius | how do i call describe? just rbx describe foo.rb? |
| 20:57:02 | chop3 leaves the room. | |
| 20:57:02 | dbussink | yeah |
| 20:57:05 | Maledictus | Anybody else seeing the Dir.glob spec failing? |
| 20:57:07 | ezmobius | cool |
| 20:57:54 | dbussink | Maledictus: which one exactly? |
| 20:58:08 | Maledictus | Dir.glob orders directory-based entries before files when a glob matches both |
| 20:58:43 | rubuildius_amd64 | Eric Hodel: 57e2fc37f; 2377 files, 7770 examples, 26834 expectations, 1 failure, 1 error; http://rafb.net/p/lDVd4082.html |
| 20:58:43 | dbussink | i've seen that one too on certain linux systems |
| 20:58:47 | chopmo enters the room. | |
| 20:59:08 | acfoeken | dbussink: singstar ... |
| 20:59:35 | dbussink | acfoeken: what's up with that? :) |
| 20:59:38 | elight leaves the room. | |
| 20:59:53 | acfoeken | dbussink: I have to endure it right now! :'( |
| 21:00:02 | olafski | sounds fun |
| 21:00:25 | Maledictus | dbussink: yeah, I thought that is OS dependant. Otherwise rubuildius would fail too |
| 21:00:45 | dbussink | Maledictus: just like those nasty iconv specs |
| 21:02:19 | dbussink | Maledictus: probably need to explicitly order them |
| 21:02:28 | dbussink | Maledictus: mri does work correctly? |
| 21:02:32 | ShayArnett enters the room. | |
| 21:02:39 | Maledictus | oh, I'll check that |
| 21:04:04 | rue | ezmobius: I think I see where -dc got disabled, although describe pretty much does the job anyway |
| 21:04:22 | anonuser leaves the room. | |
| 21:04:22 | TheVoice leaves the room. | |
| 21:04:22 | hemulen leaves the room. | |
| 21:04:22 | rubuildius_ppc leaves the room. | |
| 21:04:22 | jp_tix leaves the room. | |
| 21:04:22 | rudebwoy leaves the room. | |
| 21:04:22 | ko1_away leaves the room. | |
| 21:04:29 | ezmobius | no biggy, maybe we shoudl remove -dc from the rbx --help command |
| 21:04:52 | rue | Maybe should use an actual option parser even :) |
| 21:04:59 | dbussink | hmm, describe is failing for me |
| 21:05:07 | dbussink | No method 'parse_flags' on Compiler (Class) (NoMethodError) |
| 21:05:40 | rue | Weird error |
| 21:05:56 | Maledictus | dbussink: I forgot howto test against mri... what's the quickest way? |
| 21:06:15 | dbussink | Maledictus: ./bin/mspec -tr spec/file.rb |
| 21:06:25 | rue | dbussink: Although I do not see that method anywhere |
| 21:06:46 | TheVoice enters the room. | |
| 21:06:46 | hemulen enters the room. | |
| 21:06:46 | rubuildius_ppc enters the room. | |
| 21:06:46 | jp_tix enters the room. | |
| 21:06:46 | ko1_away enters the room. | |
| 21:06:46 | rudebwoy enters the room. | |
| 21:06:59 | elight enters the room. | |
| 21:07:25 | dbussink | rue: http://pastie.org/213215 |
| 21:07:41 | rue | Maybe in the move from describe.rb to describe |
| 21:08:07 | Maledictus | ah, ok. so mri fails too |
| 21:08:16 | rue | Plus it got reverted to the less complete describe at some point too |
| 21:08:44 | dbussink | Maledictus: then the spec is probably wrong :) |
| 21:08:50 | Defiler | We have all the infrastructure to trick it out again though |
| 21:08:55 | Defiler | It just needs doing |
| 21:09:14 | Defiler | We've always had too many little tools |
| 21:09:23 | Defiler | We should probably focus on a couple and polish them up instead |
| 21:09:42 | Maledictus | dbussink: yep |
| 21:10:02 | Maledictus | ok, and I get some ivonc errors with mri. But they are not counted in the stats |
| 21:10:14 | anonuser enters the room. | |
| 21:10:19 | rue | Defiler: Well, many little tools is fine.. many little tools, some of which do the same thing, in several different places is what we have had problems with :) |
| 21:10:25 | dbussink | Maledictus: hmm, same errors as on rubinius? or others once again? |
| 21:12:39 | Maledictus | I'll check |
| 21:15:28 | Guest25538 leaves the room. | |
| 21:16:34 | Maledictus | dbussink: ok, mri fails because iconv wasn't installed. now it works |
| 21:17:29 | Maledictus | rbx gives an error but a different one then the latest from the latest rubuildius: |
| 21:17:32 | Maledictus | Iconv.charset_map returns nil when given an unknown encoding name ERROR |
| 21:17:32 | Maledictus | NoMethodError: No method 'charset_map' on Iconv (Class) |
| 21:17:57 | naeu enters the room. | |
| 21:18:08 | bitsweat enters the room. | |
| 21:21:00 | dbussink | Maledictus: yeah, that method is simply missing on some platform, but mri implements in the extension itself |
| 21:21:47 | imajes leaves the room. | |
| 21:22:24 | imajes enters the room. | |
| 21:25:10 | Maledictus | ok, thank you |
| 21:33:45 | imajes leaves the room. | |
| 21:34:25 | imajes enters the room. | |
| 21:38:23 | twbray enters the room. | |
| 21:38:47 | acfoeken leaves the room. | |
| 21:45:20 | seydar enters the room. | |
| 21:45:36 | Maledictus leaves the room. | |
| 21:45:59 | seydar | so please /smack me if this is a dumb question, but what are people planning on doing to improve rubinius performance without writing everything in C? |
| 21:48:22 | djwhitt | jit |
| 21:48:26 | djwhitt | inlining |
| 21:49:15 | ezmobius | llvm |
| 21:49:24 | seydar | ah, llvm. that was the big one. |
| 21:49:25 | dbussink | seydar: someone starting working on a inliner written in ruby |
| 21:49:26 | djwhitt | llvm will help with jit and dispatch I think |
| 21:50:00 | seydar | what makes ruby/rubinius method dispatch take so long? |
| 21:50:32 | ezmobius | it has to do a lot of work to climb up the ancestor chain to find methods |
| 21:50:32 | djwhitt | heh, it doesn't take "so long" |
| 21:50:45 | djwhitt | oh Ruby in general you mean? |
| 21:50:51 | djwhitt | sorry, misread |
| 21:51:30 | dbussink | seydar: in rubinius it's a lot cheaper then mri afaik |
| 21:51:36 | djwhitt | when I said dispatch above I meant bytecode dispatch btw ... perhaps there's another term for that |
| 21:51:41 | seydar | is it wrong to represent object as pretty low level structs? with methods just hash tables, and subclasses having the ancestors hashtables "pointered" in? |
| 21:51:49 | dbussink | seydar: but not doing it at all is still faster of course |
| 21:52:00 | seydar | so is dispatch at all a problem in rubinius? |
| 21:52:39 | dbussink | not that big of a problem |
| 21:52:46 | seydar | sweet sauce |
| 21:53:00 | seydar | so what is rubinius' current large bottleneck? just the VM is slow? |
| 21:53:56 | seydar | or is this something _I_ could find out? |
| 21:54:03 | djwhitt | Rubinius has more Ruby code so it has to execute Ruby quite a bit faster in order to beat them in overall performance |
| 21:54:06 | djwhitt | that's one thing |
| 21:54:14 | dbussink | you can run some of the micro benchmarks that test various things |
| 21:54:15 | djwhitt | them being other implementations |
| 21:54:22 | dbussink | strings are still pretty slow |
| 21:54:29 | dbussink | ffi is really slow too |
| 21:54:49 | seydar | i really hope rubinius stays in ruby. I love being able to look things and see ruby |
| 21:54:57 | dbussink | of course they will |
| 21:55:14 | dbussink | optimization by turning stuff into c / c++ is definitely not the way we want to go |
| 21:56:08 | nexcastellan | Strings in rubinius are always exactly the requested length, correct? Would be worth looking into capacity vs. length as per MRI. I haven't checked Rubinius's underlying array, though, to see if that allocates more than was strictly requested. |
| 21:57:29 | imajes_ enters the room. | |
| 21:57:46 | dbussink | nexcastellan: dunno, but if you want to play with things, there is a benchmark in benchmark/rubinius/bm_string.rb that you can run |
| 21:57:55 | dbussink | it also shows a huge difference between mri and rbx |
| 21:58:15 | nexcastellan | Thanks, dbussink. I'd love to get strings faster. Busy with other things at the moment, but I'll add it to my list of things to look at. |
| 21:58:56 | hoopy | congrats on the rails run |
| 21:59:06 | hoopy | didn't know you guys were that far along |
| 22:01:19 | ezmobius leaves the room. | |
| 22:03:58 | imajes leaves the room. | |
| 22:07:48 | seydar | how long does bin/mspec ci take for people? |
| 22:08:46 | djwhitt | rubuildius_amd64 seems to be coming in at around 248 seconds these days |
| 22:08:54 | fbuilesv | seydar: around 100s, ~45 if using -j |
| 22:09:22 | djwhitt | rubuildius is building from scratch though, so that includes compiling the specs |
| 22:09:32 | seydar | ok |
| 22:09:34 | seydar | well last i checked |
| 22:09:35 | fbuilesv | true |
| 22:09:50 | blakewatters_ leaves the room. | |
| 22:09:55 | seydar | it was like 600+ sex for me |
| 22:09:58 | seydar | sex* |
| 22:10:04 | seydar | sec** |
| 22:10:08 | seydar | freudian slip |
| 22:10:10 | loop | :-) |
| 22:12:12 | yugui leaves the room. | |
| 22:13:28 | JonathanT leaves the room. | |
| 22:13:42 | wvdschel enters the room. | |
| 22:15:26 | seydar | btw, congrats! because rubinius so far hasn't had a single error on Tiger/PPC, the bastard child which nobody likes |
| 22:25:19 | seydar | yea, mspec just took 655 seconds for me. something is weird. |
| 22:25:47 | brixen | seydar: are you using the --turbo option? |
| 22:25:54 | shame leaves the room. | |
| 22:26:01 | seydar | ... no... should I? |
| 22:26:16 | brixen | oh wait, we don't have a --turbo option, my bad :P |
| 22:26:34 | djwhitt | seydar: what speed is your machine |
| 22:27:07 | brixen | seydar: you can run it like so: bin/mspec ci -V and let us know which files are particularly slow |
| 22:27:22 | brixen | the bigdecimal specs crawl in a few places |
| 22:27:33 | seydar | top of the line 1.25 GHz PPC G4, with 512 RAM |
| 22:27:45 | wmoxam leaves the room. | |
| 22:27:59 | djwhitt | seydar: might not be anything wrong then |
| 22:28:13 | seydar | ooh, thats a low blow |
| 22:28:43 | djwhitt | seydar: well, just sayin' rubuildius_amd64 is a 2.1GHZ dual core machine with 2GB of RAM |
| 22:28:56 | djwhitt | seydar: and the specs aren't exactly speedy on there |
| 22:29:37 | seydar | should i upgrade my box? is it worth it? |
| 22:30:24 | djwhitt | speaking of rubuildius, I wonder what happened to ppc ... |
| 22:30:46 | djwhitt | the bot is in the channel, but I haven't seen any runs from it in a while |
| 22:32:17 | djwhitt | cremes, cremes__: you might want to check on the bot sometime seems like it hasn't reported in in a while |
| 22:34:33 | seydar | EY should hold another hackfest in NY soon |
| 22:34:42 | seydar | and provide laptops for people who dont have one |
| 22:35:59 | imajes_ enters the room. | |
| 22:37:56 | AndrewO leaves the room. | |
| 22:39:18 | cremes | djwhitt: going home now... i'll take a look when i get there |
| 22:39:34 | djwhitt | cremes: cool, thanks |
| 22:41:59 | dbussink leaves the room. | |
| 22:50:46 | imajes leaves the room. | |
| 22:52:51 | seydar | have a good day everyone |
| 22:53:05 | seydar | din din + scouts awaits |
| 22:53:26 | vitaliy_ enters the room. | |
| 22:54:02 | kw enters the room. | |
| 22:57:58 | thehcdreamer leaves the room. | |
| 23:00:35 | moofbong leaves the room. | |
| 23:01:58 | kw leaves the room. | |
| 23:02:54 | cezarsa | evan: hi evan |
| 23:03:02 | evan | hi there |
| 23:03:08 | cezarsa | evan: i'm trying to figure out how rb_proc_new would look like in rubinius, any clues? |
| 23:03:28 | evan | hrm. |
| 23:03:37 | evan | it takes a function pointer, doesn it? |
| 23:03:49 | cezarsa | yes, and returns a Proc |
| 23:04:27 | evan | it's going to take a little work |
| 23:04:30 | evan | but doable. |
| 23:04:51 | chopmo leaves the room. | |
| 23:05:26 | dgtized leaves the room. | |
| 23:05:41 | evan | off, hand, the most hacky way to do it is: |
| 23:05:50 | evan | given the function pointer, create a NativeMethod object. |
| 23:06:01 | evan | the same kind of object created when rb_define_method is used. |
| 23:06:29 | evan | then, create a Proc, and install the new NativeMethod in the #call slot of it's metaclass |
| 23:06:46 | evan | something like: |
| 23:06:50 | webmat leaves the room. | |
| 23:06:55 | evan | def Proc.from_native(nm) |
| 23:07:00 | evan | pr = allocate() |
| 23:07:13 | evan | pr.metaclass.method_table[:call] = nm |
| 23:07:16 | evan | return pr |
| 23:07:16 | evan | end |
| 23:08:03 | evan | the only trouble might be argument handling. |
| 23:09:30 | evan | ok, looks like it's called like a splat. |
| 23:10:53 | cezarsa | hm, it seems to be more complicated then I thought :) |
| 23:11:02 | cezarsa | I'll read more code on NativeMethod and try to handle this |
| 23:11:03 | evan | yeah, it's going to be |
| 23:11:09 | evan | ok. |
| 23:11:14 | evan | i can help to. |
| 23:11:15 | evan | too. |
| 23:11:54 | evan | it's more difficult because, for the most part, the subtend code is organized around the idea that you hook up a C function as a method |
| 23:12:13 | evan | but rb_proc_new lets a C function be used in an entirely different context. |
| 23:12:20 | Defiler | We should just expose native functions as procs in general. Heh |
| 23:12:30 | evan | nah |
| 23:12:35 | evan | that would be very wasteful. |
| 23:12:53 | Defiler | We just need cheaper gas and that will be fine |
| 23:13:17 | cremes | djwhitt: looks like rubuildius_ppc is okay; maybe it hasn't reported because no one has committed anything |
| 23:13:22 | evan | can't we just run it on peanut oil? |
| 23:13:34 | Defiler | That smells yummy |
| 23:13:53 | djwhitt | cremes__: ok, I could have sworn there was a build that amd64 ran that ppc didn't pickup, but I could be wrong |
| 23:14:09 | rubuildius_ppc leaves the room. | |
| 23:14:25 | rubuildius_ppc enters the room. | |
| 23:15:17 | cremes | i'm looking back through the buffer here and i see reports from boyscout that did not get processed by rubuildius_ppc; just restarted it for kicks |
| 23:16:13 | djwhitt | cremes__: k |
| 23:16:21 | djwhitt | cremes__: cron job is still there too right? |
| 23:17:21 | cremes | djwhitt: yep |
| 23:17:53 | djwhitt | cremes__: ok, I guess we'll see next time someone commits. thanks for checking |
| 23:18:02 | cremes | djwhitt: every 2 minutes it gets a USR2 signal |
| 23:18:59 | thoughtfiz leaves the room. | |
| 23:22:50 | Defiler | isn't rb_proc_new just an rb_funcall to a flavor of attach_function, really? |
| 23:23:04 | Defiler | That does some extra work by creating a proc to wrap the method call |
| 23:25:11 | ch0wda leaves the room. | |
| 23:25:39 | Arjen_ enters the room. | |
| 23:25:54 | ezmobius enters the room. | |
| 23:37:02 | nexcastellan | Patch for review: http://users.nexopia.com/uploads/3233/3233577/rubinius/0001-gem-extension.patch (more) |
| 23:37:26 | nexcastellan | http://groups.google.com/group/rubinius-dev/browse_thread/thread/d2efc177d44f36a0 explains what I'm looking for. :) |
| 23:40:35 | Defiler | Whoa, has_value? |
| 23:40:41 | Defiler | Which thing needs that? |
| 23:40:50 | Defiler | So far this looks totally right to me, though |
| 23:41:06 | nexcastellan | json, most likely. :) json and rmagick were the two real pain-in-the-asses. |
| 23:41:09 | Defiler | In the case of rb_path2class, there are some Ruby methods in the kernel that might be easier to funcall |
| 23:41:11 | pauldix leaves the room. | |
| 23:41:16 | Defiler | than to implement the logic there |
| 23:41:59 | explody enters the room. | |
| 23:42:15 | Defiler | const_path_defined? is what I was thinking of, and I guess it doesn't really expose a path2class-compatible API |
| 23:42:38 | nexcastellan | It wasn't clear to me when it was best to do rb_funcall and when to reimplement the functionality. If you can give me a rule-of-thumb, I'd appreciate it. I'm thinking the rule is going to be more or less "when you can", but I didn't know that when I started. |
| 23:43:00 | Defiler | I think 'whichever is easier to implement' is the rule of thumb for the moment |
| 23:43:16 | Defiler | Since we don't have any performance data it is hard to worry about that |
| 23:43:59 | nexcastellan | Hahaha. Premature optimisation and all that. :) |
| 23:46:49 | naeu leaves the room. | |
| 23:49:54 | vitaliy_ leaves the room. | |
| 23:50:29 | Defiler | http://www.iht.com/articles/2008/06/11/business/11suits.php |
| 23:50:39 | Defiler | This is pretty interesting, though totally unrelated :) |
| 23:53:29 | shame enters the room. | |
| 23:53:51 | blakewatters enters the room. | |
| 23:54:11 | dfg59 | want to get back involved a bit, where's the best place to look right now for helping with upcoming milestones? just took a look at lighthouse |
| 23:54:23 | cezarsa | nexcastellan: still there? |
| 23:54:47 | Defiler | dfg59: We could use a bunch of testing of the binaries and installers |
| 23:55:07 | nexcastellan | cezarsa, yeap, for a few more minutes. :) |
| 23:55:10 | Defiler | dfg59: Making sure that "rbx blah" does the right thing after you install a "blah" gem with a binary stub etc |
| 23:55:22 | Defiler | That seems like something we just can't trust automated tests for and it is pretty new |
| 23:55:23 | stepheneb_ leaves the room. | |
| 23:55:33 | dfg59 | Defiler: hrmm, ok, is there an open ticket where i can look for more info? |
| 23:55:37 | cezarsa | nexcastellan: it seems like RHASH isn't copying data back to subtend |
| 23:55:57 | Defiler | dfg59: I suspect it is now a closed ticket, now that those features have been implemented.. but let me find it |
| 23:56:05 | cezarsa | so if some extension change rhash->tbl to another place subtend will never know about that |
| 23:56:06 | nexcastellan | cezarsa, you are right, nice catch. Bonus points if you send me a patch to fix it. ;-) |
| 23:56:22 | dfg59 | Defiler: sweet, and are there gems in particular that rubinius can run that we're interested in? |
| 23:56:38 | drbrain | dfg59: rubinius runs rake |
| 23:56:49 | dfg59 | drbrain: gotcha |
| 23:56:56 | cezarsa | i could work out on this, but not now, it would be a lot like RSTRING and RARRAY |
| 23:56:56 | drbrain | rbx rake clean; rbx rake spec would be a good test :) |
| 23:57:17 | Defiler | dfg59: thse two.. |
| 23:57:19 | Defiler | http://rubinius.lighthouseapp.com/projects/5089/tickets/591-gem-stubs-aren-t-executable |
| 23:57:24 | Defiler | http://rubinius.lighthouseapp.com/projects/5089/tickets/599-cleanup-bin-shotgun-rubinius-etc |
| 23:57:37 | Defiler | both now resolved, but shows you how recent it is |
| 23:57:40 | dfg59 | drbrain: does rbx remove the -S argument? no need to do rbx -S rake |
| 23:57:48 | dfg59 | Defiler: thanks |
| 23:57:52 | Defiler | We don't support -S yet I don't think |
| 23:57:56 | Defiler | Which may be a problem |
| 23:57:58 | dfg59 | k |
| 23:58:07 | drbrain | rbx rake should not look at shebang |
| 23:58:15 | dfg59 | i see |
| 23:58:17 | drbrain | you should be running rubinius rake at that point |
| 23:58:19 | nexcastellan | dfg59, with my uncommitted patch (http://groups.google.com/group/rubinius-dev/browse_thread/thread/d2efc177d44f36a0) adds support for a number of extensions. |
| 23:58:37 | nexcastellan | Six or so. :) |
| 23:58:58 | nexcastellan | Any fixes to that patch or any more extensions you can get working is definitely appreciated. :) |
| 23:58:59 | Defiler | dfg59: You could try testing that, as well. |
| 23:59:08 | Defiler | I guess it just feels like we need a round of QA before the release |
| 23:59:26 | dfg59 | sure, that's exactly what i wanted to know |
| 23:59:34 | dfg59 | i'll take a look at some of these things when i get off work tonight |