Show enters and exits. Hide enters and exits.
| 15:30:29 | roja | Hi guys |
| 15:31:37 | roja | I'm currently on annual leave from my phd and was looking arround for a project to commit to |
| 15:32:38 | roja | wondering if there is anything i could aid rubinus with, afterall if it reaches it's potential it may well make for a significantly better ruby world :) |
| 15:34:35 | rue | All kinds of stuff to do, depending on where your interest lies |
| 15:35:18 | roja | well |
| 15:35:36 | roja | i thought i would get my feet wet with something ruby-ish |
| 15:35:50 | rue | Can try running some libraries and fixing whatever bugs they expose, or look at failing specs and fix there |
| 15:35:59 | roja | i am a capable c++ coder but woulden't be happy jumping in there really |
| 15:36:50 | rue | Plus kernel/ is all Ruby, so just examining the code there can be enlightening as well as expose potential areas of improvement |
| 15:36:58 | roja | fair enough |
| 15:37:03 | roja | :) |
| 15:37:51 | roja | is there a document anywhere outlining prefered coding style? |
| 15:38:05 | rue | kernel/ ;) |
| 15:38:30 | rue | Actually I think there is something in doc/ |
| 15:38:46 | rue | Dunno how complete it is. Pretty basic insofar as the Ruby is concerned |
| 15:38:51 | roja | good good i will have a read :) |
| 16:08:11 | roja | bin/mspec :library segfaults :S |
| 16:49:00 | rue | That would be a good bug to try find ;) |
| 17:15:35 | evan | morning. |
| 17:19:13 | brixen | morning |
| 17:21:01 | brixen | adobe ads are laughable.. free shipping? well for spending $1000s, that's the least you could do |
| 17:21:28 | brixen | they should deliver in satin boxes with a chocolate on top |
| 17:23:36 | evan | hehe |
| 17:23:37 | evan | so true. |
| 17:23:53 | evan | esp. since it seems like people rebuy that $1000 thing every couple years too |
| 17:24:06 | brixen | yeah |
| 17:24:24 | brixen | creative suite is a few thousand |
| 17:43:54 | BrianRice-work | hey guys. so I wrote a little version of rbc for slate ("slc", of course), but I had to go about it in a different manner, as a serializer of tree-like data structures, with specialization for common collections, literals, and syntax tree nodes. I'm still debugging a few things but it seems sound |
| 17:44:47 | Defiler | What's slate? |
| 17:44:57 | evan | BrianRice-work: what does the Slate VM execute? |
| 17:44:59 | evan | bytecode? |
| 17:45:06 | BrianRice-work | SSA bytecode |
| 17:45:13 | evan | so why not put that in the slc? |
| 17:45:17 | BrianRice-work | Defiler: just some smalltalk dialect |
| 17:45:23 | brixen | Defiler: http://slatelanguage.org/ |
| 17:45:25 | BrianRice-work | evan: that'll be round two of this effort |
| 17:45:44 | BrianRice-work | and it's a little trickier because of some other design decisions |
| 17:45:57 | Defiler | cool |
| 17:46:16 | BrianRice-work | mainly that we don't define file source as a single compilation unit yet (that is, we parse and compile/run each top-level statement in turn) |
| 17:47:25 | Defiler | what advantages does that offer? |
| 17:47:33 | BrianRice-work | "that"? |
| 17:47:35 | Defiler | files seem like a natural unit of compilation |
| 17:47:47 | BrianRice-work | none. that's just how we wrote it a long time ago. |
| 17:48:00 | BrianRice-work | although I can think of one potential |
| 17:48:01 | evan | gotcha. |
| 17:48:23 | BrianRice-work | which is that it's easier to say that the methods in a file don't need the file's "top-level method" to close over, if it just isn't there. |
| 17:50:15 | BrianRice-work | all methods and blocks in slate are closures, although the compiler can do some basic analysis to determine that that's not necessary sometimes. |
| 17:50:58 | evan | ah |
| 17:51:10 | brixen | methods being closures is interesting |
| 17:51:17 | evan | you don't want every method to be closing over the toplevel |
| 17:51:38 | brixen | I'd think not |
| 17:51:40 | BrianRice-work | right. for us, the top-level is a namespace scope which makes it easier |
| 17:52:03 | BrianRice-work | I'll probably figure out a way to do this, but first I want the speed gain. |
| 17:52:26 | BrianRice-work | and serialization schema |
| 18:32:32 | BrianRice-work | actually, what I'll probably do (first) is to describe a file as a series of methods compiled into the namespace scope idea we have now. |
| 18:33:00 | BrianRice-work | </rant> anyway, just wanted to relate the outcome of the previous discussion |
| 19:07:02 | brixen | complexity complexifies everything it touches |
| 19:07:46 | brixen | you can't wall off complexity |
| 19:07:57 | brixen | it's like that radioactive green goo that eats through walls |
| 19:11:23 | Defiler | it makes the walls complex and now the next thing over is touching something complex |
| 19:12:14 | brixen | exactly |
| 19:13:37 | evan | speaking of which |
| 19:13:46 | evan | i'm trying to iron out our broken metaclasses |
| 19:13:49 | evan | complicated. |
| 19:14:05 | brixen | hmm |
| 19:14:30 | brixen | where are we broken now? |
| 19:15:29 | brixen | is it related to #37? |
| 19:21:12 | dwaite | wait, what do you mean metaclasses are complex? |
| 19:21:16 | evan | the metaclass one? |
| 19:21:37 | evan | yeah, it is #37. |
| 19:21:44 | evan | i might have to remove the MetaClass class entirely. |
| 19:21:48 | evan | to make this work the way it's supposed to. |
| 19:21:56 | evan | because the class of a metaclass is not MetaClass |
| 19:22:06 | brixen | hmm |
| 19:22:13 | evan | the class of a metaclass is the metaclass of the metaclass's superclass |
| 19:22:16 | evan | follow that! |
| 19:22:23 | brixen | oh I have |
| 19:22:24 | brixen | heh |
| 19:22:35 | evan | which means there is no where for MetaClass to live |
| 19:22:44 | dwaite | oh wait, now I see what you mean about metaclasses being complex |
| 19:23:03 | dwaite | it burns |
| 19:23:04 | brixen | evan: I still have that phrase in my notebook from the last time you explained it |
| 19:23:09 | brixen | dwaite: it does burn |
| 19:23:37 | evan | brixen: and I'd forgotten myself. |
| 19:23:41 | dwaite | perhaps you need a new type, to cut across the complexity? |
| 19:23:44 | dwaite | like EmoClass |
| 19:23:50 | brixen | oh lord |
| 19:23:51 | brixen | haha |
| 19:24:02 | brixen | evan: there's no escape! |
| 19:24:33 | evan | RUN |
| 19:24:37 | brixen | I blame Rock Band |
| 19:24:45 | brixen | damn video games |
| 19:25:02 | evan | GET OFF MY LAWN! |
| 19:25:06 | brixen | hehe |
| 19:26:11 | evan | i think the coffee i bought was secretly decaf. |
| 19:26:18 | evan | it doesn't say, but i swear it is. |
| 19:28:32 | dwaite | just make two pots, drink them both, and see if you start getting jitters |
| 20:00:36 | evan | GAH |
| 20:00:47 | evan | we put attach_method on Metaclass |
| 20:01:00 | evan | but i have to change the class of a metaclass so that MetaClass isn't in there anymore |
| 20:01:04 | evan | le sigh. |
| 20:01:58 | brixen | :( |
| 20:04:17 | rue | If there is an advantage to having a MetaClass class, then you can proxy it like IncludedModule |
| 20:04:29 | rue | But I am not sure it is |
| 20:05:22 | evan | well, i think i've got a handle on it |
| 20:05:25 | evan | but i need a haircut. |
| 20:28:23 | rue | At least a wash to get rid of the handle |
| 22:16:10 | rue | Mm. Glög with some quality white rum |
| 22:16:19 | evan | haircut. |
| 22:16:22 | evan | achieved. |
| 22:17:37 | rue | No more handle? |
| 22:18:02 | evan | handle bar? no, still have a beard. |
| 22:18:26 | rue | Handlebar hair would be sweet |
| 22:21:31 | slava | hi evan |
| 22:21:44 | evan | hello slava darling |
| 22:22:30 | slava | how's rbx? |
| 22:22:50 | evan | oh good |
| 22:22:56 | evan | working out some obscure bugs in our MOP |
| 22:31:28 | brixen | I'm working on some bugs in our BROOM |
| 22:32:41 | evan | finally the KITCHEN will run! |
| 22:32:41 | evan | :D |
| 22:32:44 | brixen | alas, I cannot devise a good meaning for that acronym |
| 22:32:48 | brixen | heh |
| 22:34:14 | evan | hm |
| 22:34:55 | slava | google chrome is pretty slick |
| 22:35:12 | Defiler | yes, I need a vimperator plugin for it |
| 22:35:13 | brixen | I should try it |
| 22:35:18 | Defiler | chrome is pretty boss though |
| 22:35:41 | slava | I like it more than safari |
| 22:35:51 | slava | safari 3.x was good but 4.0 sucks |
| 22:37:02 | evan | is there vimperator for safari? |
| 22:39:56 | rue | Chrome sucks and Google is evil |
| 22:40:14 | evan | but chrome plated goggles rock |
| 22:40:16 | evan | how ironic. |
| 22:56:48 | dwaite | is google evil now? |
| 22:57:24 | BrianRice-work | no, only on the top half of the hour :P |
| 22:57:33 | slava | heh, hi BrianRice-work |
| 22:57:39 | BrianRice-work | hey slava |
| 22:57:40 | evan | yep, they're good for another 3 minutes |
| 22:57:52 | evan | oh, maybe i've got it backwards. |