Index

Show enters and exits. Hide enters and exits.

06:21:31Defilerleavengood: Yeah, temporary situation
06:22:15leavengoodit is mainly for rake, right?
06:22:23Defileryeah
06:22:45Defilerand generating some config
06:23:28leavengoodOK
06:23:50DefilerMRI works fine on Haiku though, right?
06:24:46leavengoodwell somewhat
06:25:24leavengoodit's annoying but I'm trying to run the alpha 1 release but apparently I need some fixes from our Haiku trunk build to run MRI 1.9.1
06:25:32leavengoodpthread stuff
06:25:40DefilerAah, sleazy
06:25:50leavengoodyeah
06:26:12leavengoodI can compile my own Haiku no prob, but updating an existing machine is a bit of a pain
06:26:22leavengoodplus I'm trying to 'dogfood' it by running the alpha
06:26:25leavengoodanyhow... ;)
06:26:39DefilerHeh, that sounds like a long weekend
06:26:53leavengoodnah, not that bad really
06:26:59leavengoodI'm just lazy
06:27:26leavengoodmy other fun project is WebKit, boy does that take decades to compile
06:28:15leavengoodI think it is mostly just GCC4 being slow, but Haiku probably has room for optimizing
06:28:51leavengoodI'm definitely looking at Go after these latest slow-ass compile sessions
06:28:57DefilerI can't wait until porting LLVM to your new OS is enough to get you a bunch of working software
06:29:06leavengoodyeah for sure
06:29:27leavengoodseems kind of silly all the porting we do on all systems
06:29:35leavengoodWebKit is just crazy with all the platform code
06:29:45leavengoodonce you get into GUI-land forget it
06:30:05leavengoodat least Rbx doesn't have to worry too much about that
06:32:44leavengoodof course part of my interest in rbx is to create a nice Ruby wrapper around some Haiku APIs, plus a nice GUI builder app for some true RAD
06:33:00leavengoodof course the same idea would be cool on other OSes
06:33:24danlucraftleavengood: does Java run on Haiku?
06:33:28leavengoodsort of like MacRuby I guess
06:33:46DefilerAt least ports are easy these days, relatively speaking
06:33:51leavengooddanlucraft: not yet...we have a few people working on porting OpenJDK
06:34:10Defileryou don't have to write a bunch of fresh assembly for your new computer
06:34:19danlucraftthat will also get you a lot of software for free.
06:34:20leavengoodDefiler: that is true at least
06:34:47DefilerI haven't had a lot of luck getting it built, but MacRuby sure has come a long way
06:34:53leavengooddanlucraft: indeed...though like any system long term we want native apps that take advantage of Haiku APIs and features, blah blah
06:35:21leavengooddanlucraft: we also have a surprisingly decent QT4 port that was recently released
06:36:05leavengoodso any comments on Go? :)
06:36:20leavengoodI'm not a big fan of the syntax, but the compile speed sure seems nice
06:37:13danlucraftsomeone suggested a dojo on Go this Thursday. it was not a popular suggestion
06:37:13leavengoodjust joined the #go-nuts channel...holy crap 468 people in there!
06:37:25danlucraftthink we're going with Ioke instead
06:37:41leavengooddojo? loke?
06:37:59DefilerGo looks very interesting, but I wish it had pointers
06:38:13DefilerI'd like something between Go and ooc, with more ruby-like syntax
06:38:35danlucraftleavengood: dojo == group exercise (at work), Ioke is Ola Bini's v. dynamic JVM lang
06:38:57leavengooddanlucraft: ah, OK
06:39:06DefilerWhat I really want is a language that lets me express higher-level relationships between modules
06:39:17leavengoodDefiler: seems like one day with LLVM we could each get custom languages to our own specs
06:39:28Defiler"I'm handing off this thing to be modified, and I promise not to modify it before the call returns"
06:39:31Defilerfor example
06:39:35leavengoodyeah
06:39:51leavengoodseems like programming hasn't advanced all that much over the years
06:40:06DefilerI think it is embarassing that we are still using 1950s coding styles
06:40:08Defileryeah
06:40:25leavengoodyup
06:41:19leavengoodhah, no wonder my Google-Fu was failing me
06:41:29leavengoodit is ioke, not Loke :)
06:41:35leavengooddamn fonts
06:44:34danlucrafthah
06:46:03leavengoodhmmm looks a bit too Lispy for my tastes
06:46:23leavengoodwish I could afford to put more time in trying other languages though...I think it expands the mind
06:50:26DefilerI'm going to write something in ooc, it looks keen
06:52:14leavengoodwow ooc is even on BeBits
06:52:37leavengood(the old BeOS software site...Haiku is a BeOS recreation)
06:56:53leavengoodthough apparently it doesn't really work that well since there is no Java
06:56:55leavengoodoh well
07:00:01leavengoodOK well it has been fun chatting...need to get to sleep
07:00:24leavengoodI'll probably drop back in here another day when I get Rubinius going on Haiku (hopefully I can...)
07:00:34leavengoodlater
17:00:58evanmorning
17:02:50DefilerYo
17:03:08Defilerooc is a pretty interesting language.
17:03:42evanoh?
17:04:06DefilerYeah. Translates to C, has a ton of nice ideas
17:04:16evanoh yes, this thing.
17:04:19Defilerlike a This type
17:04:30Defilerand the ability to 'cover' a C native type and add methods to it
17:04:50DefilerSo you can easily make "foo" reverse()
17:05:02Defilerinstead of converting a char* to something first and then blah blah
17:05:25evanright
17:05:30evango has that too I think
17:05:35Defilersimple type inference, so you can just do x := 5.0 and wham, Float
17:05:36evanthats an idea i've head a few times as well
17:05:37Defileryeah
17:05:39evanseems obvious
17:05:50DefilerIt's an old one, but things like Java decided not to implement it
17:05:58evanyeah
17:05:59DefilerSo as usual we had to go another decade or two without it
17:06:01evanit's quite cool
17:06:13evanthe question is always "what is this inside those methods?"
17:07:14Defilerooc has a pretty good solution to the old 'subclasses returning subtypes' mess as well
17:07:45DefilerBeing able to add methods to 'Pointer' without actually changing the fact that you are working with a bare pointer is pretty cool
17:09:36evantrue
17:27:18rue"Translates to C" != "interesting"
17:27:41evanyes, one problem with "translates to c" is "debugging sucks."
17:28:00evansucks more that writing straight C.
17:28:09evanfor references, please try and debug a yacc parser.
17:30:11danlucrafthow are you guys liking having rubinius docs etc. in another branch of the main repository?
17:30:18danlucraftis that working out?
17:31:25evanwe're doing that?
17:32:08danlucrafthttp://github.com/evanphx/rubinius/tree/gh-pages ?
17:32:23evanthats the website
17:32:27evanworks fine.
17:32:28danlucraftlooks like your website
17:32:54danlucraftok
17:36:35rueI have a vehement dislike of abusing branches that way, but it keeps gaining momentum.
17:37:01rueOn the other hand, if it encourages documentation, it is not a complete loss.
23:16:09tarcieriis translates to Erlang more interesting than translates to C? :)
23:16:24tarcieriyour guys Channels are asynchronous, right?
23:16:31tarcierijust noticed they're synchronous in Go...
23:16:46tarcierimuch like Stackless Python
23:17:44DefilerTranslates to anything is pretty cool in general, though rarely outright useful
23:29:10tarcieriDefiler: it's useful cuz Erlang is ooglaay
23:29:11tarcieri:)
23:49:49DefilerI had no idea this was the behavior: http://gist.github.com/237387
23:49:52DefilerThat is strange
23:50:16DefilerI thought it didn't matter WHEN you did the include, that the direct superclass chain always had the same result
23:51:52ezmobiusi thnk thats because it copies the chain when you do the include. and doesnt update the chain if you reopen the included module later?
23:53:03DefilerOhhh right yeah yeah