00:00:00 --- log: started retro/13.08.20 04:29:36 --- quit: Sgeo (Read error: Connection reset by peer) 04:58:27 --- quit: tangentstorm (Quit: zzz) 09:29:34 --- join: beretta (~beretta@74.135.121.153) joined #retro 09:52:40 --- join: Mat2 (~claude@91-65-144-133-dynip.superkabel.de) joined #retro 09:52:44 hello 09:53:10 hello 09:53:22 hi beretta 09:54:15 what are you doing currently ? 09:54:33 working on some forth code. 09:56:22 I'm still working on my AOT compikler for retro. After chatting with tangentstorm I decided to port it to Nimrod from C (good decision). The sources are now much better factored 09:56:39 working on this: https://sites.google.com/site/cocoboot2/ 09:58:05 I think this is a nice project for the coco. How's your progress ? 09:58:57 not bad. I'm working on reading the os9 filesystem now. 09:59:06 what's AOT ? 09:59:21 Ahead-Of-Time compilation 09:59:46 :) 09:59:52 like early binding ? 09:59:57 exactly 10:00:08 lol 10:00:25 it's the opposite strategy to Just-In-Time compilation 10:00:36 right :) 10:01:09 you can read my thoughts on JIT on that link: EW!! 10:01:37 although can't FORTH be looked at as a JIT :) ? 10:02:55 not really, a JIT compoiler traces the virtual execution-state of the interpreter for dynamic code optimization 10:03:09 ^compiler 10:03:40 I have adapted some of Retro into my thinking... some functional programming things, and the outer interpreter's simplicity. 10:04:56 which was to be expected, your project has many similarities with retro 10:07:50 the my outer interpreter does not process lines, it parses by white-space. 10:08:43 odd at first, but simpler to do away with buffer copies, and quite handy ( even if there's no traditional "OK" prompt. 10:10:09 yes, I think white-space limited seperation is easier to implement for a forth like environment 10:12:28 I found not resetting the state of the compiler after no finding a word pretty handy too. 10:17:19 well, for my future developments I try to limit states to the minimum 10:18:03 (because that limits concurency at some point) 10:18:44 sadly a complete state-less forth is not possible 10:20:09 functional programming is definately not my forte so I'll assume that's true. 10:21:12 I must admit I don't understand stateless programming :( 10:23:43 as reading your blog, your approach to multitasking seems very functional oriented for this statement (I mean you localize the interpreter state) :D 10:31:23 oh... I guess I'm smarter than I think. 10:32:34 what I do not understand is why you need ... 7 stacks ? 10:42:11 oh.. only SP and RP are "real" stacks. the others are just software stacks 10:43:02 ah ok (I wondered because as I know the 6809 has 2 stacks and only 2 index registers) 10:43:22 indeed. 10:44:32 but if the auxilarary stacks needed to be faster I would create more "hardware" stacks in the VM 10:45:39 I just tend to see everything as a stack now. 10:47:22 stack caching for the first 10 or 20 dictionary entries speed up dictionary searches significant 10:48:20 hmmm... I'll put that to some thought. 10:48:50 I've found a simpling hashing to help tremenously too. 10:50:58 agree (it's only memory expensive) 12:42:35 --- quit: beretta (Quit: Leaving) 13:03:32 ciao 13:03:35 --- quit: Mat2 (Quit: Verlassend) 15:22:38 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro 16:33:48 --- join: Sgeo (~quassel@ool-ad034ea6.dyn.optonline.net) joined #retro 18:09:32 --- quit: tangentstorm (Ping timeout: 245 seconds) 18:25:05 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro 20:11:51 --- join: roarde (~roarde@pdpc/supporter/active/roarde) joined #retro 21:04:20 --- quit: roarde (Quit: Leaving) 21:05:40 --- join: roarde (~roarde@pdpc/supporter/active/roarde) joined #retro 23:59:59 --- log: ended retro/13.08.20