Da
Show enters and exits. Hide enters and exits.
| 01:44:00 | postmodern | i compiled rubinius 1.0.0-rc3, but bin/rbx -v just exits |
| 01:44:43 | brixen | postmodern: what's your platform |
| 01:44:43 | evan | it doesn't print out the version then exit? |
| 01:44:52 | postmodern | brixen, gentoo linux amd64 |
| 01:45:07 | brixen | needs to install gentoo |
| 01:45:09 | postmodern | evan, it doesn't seem to respond to any arguments |
| 01:45:21 | evan | i'd suggest a rake clean && rake |
| 02:01:06 | postmodern | evan, ah that did the trick |
| 02:01:56 | postmodern | holy cow, im running irb ontop of rubinius |
| 02:02:01 | postmodern | it's really that easy |
| 02:05:10 | postmodern | so i see that MRI is still faster than rubinius for some things |
| 02:05:25 | postmodern | which areas is rubinius still lacking in? |
| 02:05:30 | postmodern | or types of operations |
| 02:06:11 | brixen | depends |
| 02:06:24 | brixen | try running your code |
| 02:06:43 | brixen | you can see tagged specs (ie, ones that do not pass) with bin/mspec tag --list-all |
| 02:07:07 | brixen | but we can't say how complete rubyspec is, so that's just a guess for "what is missing?" |
| 02:09:39 | manveru | anybody looked at my ticket yet? |
| 02:09:46 | brixen | manveru: what #? |
| 02:10:17 | manveru | http://github.com/evanphx/rubinius/issues#issue/198 |
| 02:11:19 | manveru | i talked with yui naruse at the tokyo ruby kaigi, he said he's gonna try getting it into MRI |
| 02:11:56 | brixen | cool |
| 02:12:08 | brixen | I'll remind evan about it, he needs to make the final decision |
| 02:12:18 | manveru | thanks |
| 02:12:27 | brixen | n/p, thanks for making the patch |
| 02:12:56 | manveru | now i just need 1.9 syntax :) |
| 02:13:11 | brixen | coming soon! |
| 02:13:24 | manveru | anything i can help with? |
| 02:13:49 | brixen | basic approach, diff 1.8 and 1.9 MRI grammar |
| 02:14:03 | brixen | manually apply that to rbx grammar.y in the one_nine branch |
| 02:14:05 | manveru | where is all the 1.9 stuff? |
| 02:14:10 | manveru | oh |
| 02:14:25 | brixen | then fix the compiler |
| 02:14:41 | brixen | and the VM where it needs to understand different arg semantics |
| 02:20:43 | brixen | manveru: what 1.9 syntax do you use? |
| 02:20:56 | manveru | hmm |
| 02:21:02 | brixen | I personally find this horrible: ruby1.9 -v -e 'def foo(a) p a end; foo a: :b, c: :d' |
| 02:21:14 | manveru | missing end? |
| 02:21:29 | brixen | a: :b <- my eyes see this and stumble all over themselves |
| 02:21:36 | manveru | lol |
| 02:21:44 | manveru | i cannot stand => anymore |
| 02:21:57 | brixen | what is missing end? |
| 02:22:10 | manveru | oh, nm |
| 02:22:15 | brixen | :a => :b is much better than a: :b IMO |
| 02:22:27 | brixen | what is an a: in ruby? nothing |
| 02:22:28 | manveru | i even patched pretty inspect to output : when suitable |
| 02:22:32 | brixen | what is an :a, a symbol |
| 02:22:43 | brixen | horrid I say |
| 02:22:51 | brixen | where can you use a: ? |
| 02:23:01 | manveru | lol |
| 02:23:10 | brixen | seriously, it's really bad |
| 02:23:10 | manveru | you probably don't use -> either |
| 02:23:28 | brixen | cognitively you have to maintain this concept for "symbol in a method call" |
| 02:23:28 | manveru | or def foo(*a, b) |
| 02:23:34 | brixen | that looks nothing like a symbol |
| 02:23:40 | brixen | that's cognitive load for nothing |
| 02:23:44 | brixen | no gain all pain |
| 02:24:05 | manveru | it's easier to write |
| 02:24:09 | manveru | it reads nicer |
| 02:24:11 | brixen | bah |
| 02:24:13 | brixen | no |
| 02:24:17 | manveru | i find it very useful |
| 02:24:32 | brixen | that's like: def foo(a) ... end; oof(1) |
| 02:24:38 | brixen | a: is just plain backwards |
| 02:24:54 | brixen | :a => :b .... ooh soo hard to type :'( |
| 02:24:58 | brixen | haha |
| 02:25:09 | manveru | so you stare at %() for half an hour to figure out that it's a string? |
| 02:25:21 | manveru | and, omg, %r() might be a regex |
| 02:25:24 | brixen | no, it's an interpolated string |
| 02:25:41 | brixen | I would stare at (blah)% for quite a while though |
| 02:25:56 | brixen | especially if it were (blah)% %foo |
| 02:26:19 | brixen | %r is a regexp, btw, no might about it |
| 02:26:20 | manveru | http://github.com/manveru/ver/blob/refactor/lib/ver/buffer.rb#L11-45 |
| 02:26:48 | brixen | that's not any more clear to me than => |
| 02:27:06 | manveru | i don't say it's "more" clear |
| 02:27:09 | brixen | the : following the identifier is a special concept in a call or hash that isn't anywhere else |
| 02:27:12 | brixen | I dislike it |
| 02:27:30 | manveru | ok |
| 02:28:02 | manveru | anw, i use it :) |
| 02:28:11 | brixen | and I won't :) |
| 02:28:19 | brixen | what else do you use from 1.9? |
| 02:28:50 | brixen | btw, you should ask github to add syntax highlighting for foo: : |
| 02:28:53 | manveru | http://github.com/manveru/ffi-tk/blob/master/lib/ffi-tk/widget/text.rb#L198 |
| 02:28:53 | brixen | er :) |
| 02:29:23 | brixen | ok, 1.9 arg passing |
| 02:29:29 | brixen | anything else? |
| 02:29:40 | manveru | stabby proc? |
| 02:29:45 | brixen | ok |
| 02:29:47 | manveru | but not in any library code |
| 02:29:52 | brixen | all the really ugly stuff, great :P |
| 02:29:54 | brixen | hahah |
| 02:29:57 | manveru | that's just for my personal amusement :) |
| 02:30:01 | brixen | heh |
| 02:30:21 | brixen | well, the big thing with one_nine atm is that we need to merge in the recent changes from master |
| 02:30:21 | manveru | foo a: ->{ bar}, b: ->{ duh } |
| 02:30:29 | brixen | since we're still fixing a lot of bugs |
| 02:30:38 | manveru | yeah |
| 02:30:41 | brixen | the only advantage you will have on rbx from 1.9 is the syntax |
| 02:30:55 | brixen | we will have all the speed and goodness on 1.8 |
| 02:30:59 | brixen | the same as 1.9 |
| 02:31:07 | manveru | but it's like 13k LoC |
| 02:31:17 | manveru | i really don't wanna translate all that |
| 02:31:29 | brixen | what is 13kloc? |
| 02:31:34 | manveru | ffi-tk/ver |
| 02:31:38 | brixen | oh yeah |
| 02:31:47 | manveru | and i use lots of encoding stuff |
| 02:31:50 | brixen | well, just saying, we are close to getting 1.9 support |
| 02:31:57 | brixen | that's the huge thing, encoding stuff |
| 02:32:10 | manveru | yeah |
| 02:32:19 | brixen | but that doesn't require syntax support |
| 02:32:20 | manveru | a huge part of the old ruby-tk was about encoding |
| 02:32:31 | brixen | evan already hacked in magic comments iirc |
| 02:32:39 | manveru | nice :) |
| 02:33:41 | manveru | ok... |
| 02:33:52 | manveru | so you plan to do encoding in ruby too? |
| 02:34:25 | manveru | i really have no idea how most of it works... |
| 02:34:39 | brixen | git show d0f736008 |
| 02:35:07 | brixen | encoding support will mostly be figuring out how to back String and making the methods aware of that |
| 02:35:38 | brixen | one idea is to add a CharArray that uses a different immediate type rather than a ByteArray (which uses bytes) |
| 02:35:48 | manveru | yeah |
| 02:36:14 | manveru | you said that before... but that part is in C, so i can't really help you there |
| 02:37:30 | manveru | if we cover ascii and utf-8 first, we'll be running like 99% of apps out there without problems |
| 02:37:43 | manveru | well, and binary, but that's the current string already |
| 02:38:16 | brixen | cool |
| 02:38:50 | brixen | we could also just build it on ByteArray |
| 02:39:20 | brixen | perhaps make ByteArray encoding aware, so you could index by characters instead of bytes |
| 02:39:25 | brixen | that would be mostly in Ruby :) |
| 02:40:14 | brixen | most of String methods should just be agnostic (ie do things with "characters") |
| 02:40:44 | evan | postmodern: --enable-llvm is redundent, it's the default. |
| 02:40:49 | evan | ok, shower time. |
| 05:14:01 | manveru | thinking of giving Module#prepend implementation a try... |