00:00:00 --- log: started retro/12.10.10 00:26:11 --- quit: ncv (Ping timeout: 246 seconds) 00:52:24 --- join: tangentstorm (~michal@108-218-151-22.lightspeed.rcsntx.sbcglobal.net) joined #retro 00:53:10 i'm trying to go back and put the docstrings into the kernel, but I don't know where to do it 00:54:10 i can't tell for sure, but it doesn't look like t: definitions have a header yet 00:54:32 and they don't seem to affect 'last' 01:04:57 i'll put them in a separate file. 03:20:14 okay. just pushed docstrings for all core functions (except @each, which had a huge one)... "make helpful" from my branch builds a new image with all the docstrings 03:21:00 --- join: ncv (~quassel@79.114.104.81) joined #retro 03:21:00 --- quit: ncv (Changing host) 03:21:00 --- join: ncv (~quassel@unaffiliated/neceve) joined #retro 03:40:14 --- join: hurfadurf (~lol@don.gs) joined #retro 05:08:59 --- quit: ncv (Ping timeout: 246 seconds) 05:13:25 tangentstorm: merged, looks promising 05:17:05 thanks :) 05:18:25 I did fix an issue with chwords: if you used it on a non-chain, it would corrupt the dictionary. 05:19:26 http://bazaar.launchpad.net/~crc-x/retro-language/retro/revision/382 05:21:04 ouch 05:26:03 I'm kind of thinking about putting a bunch more stuff like this in... like maybe something that shows the stack as you go the way colorforth does, mixing in the tab completion code you have... maybe even adding some menus and stuff 05:28:55 i'd planned to put the docs inline but i couldn't figure out how to do it... but it kind of worked out because now you can choose whether to build a big fancy image or a tiny bare bones image 05:29:42 so.. i'm hoping i won't be in your way on the embedded side? :) 05:54:31 tangentstorm: I'm considering options for the embedded side. it may pay off to have a separate image for embedded applications 05:57:57 crc: just out of curiosity, what do you embed it in? 06:08:24 I know of one person using it on an arduino board, and I use it on an mbed.org (arm-based) board 06:08:38 and it runs on retrobsd for pic32 microcontrollers 06:12:25 cool. i have a box full of electronics i've taken apart... including a bunch of motors. one of these days i'm going to make a little cnc mill or router i think 07:46:33 --- quit: yiyus (*.net *.split) 07:54:20 --- join: yiyus (1242712427@je.je.je) joined #retro 09:27:28 --- join: ncv (~quassel@unaffiliated/neceve) joined #retro 09:30:50 --- join: kumul (~kumul@cvx-ppp-66-50-130-239.coqui.net) joined #retro 09:57:04 --- quit: hurfadurf (Quit: sleep) 12:10:25 --- join: Mat2 (~quassel@91-64-133-197-dynip.superkabel.de) joined #retro 12:10:32 hello 12:35:05 --- quit: ChanServ (shutting down) 12:35:48 --- join: ChanServ (ChanServ@services.) joined #retro 12:35:49 --- mode: hubbard.freenode.net set +o ChanServ 12:42:06 . 12:45:08 howdy 13:03:20 hi docl ! 13:03:36 what's new ? 13:17:45 tangentstorm has been adding some awesome new features like a help command and colors library. 13:17:59 I've been playing with proto-regexp stuff 13:24:28 also there's a guy who was giving out free shells at http://soupwhale.com/, but he has stopped for a few days until he gets his server upgraded. I guess the thing to do is idle in the irc until the upgrades come in. 13:25:06 I got mine a few weeks ago when he pitched it on lesswrong 13:26:07 colour support is good, now I can port my block editor 13:26:56 :) 13:28:10 nice link, thanks, I will try to get an account 13:29:42 it's a lot nicer to not have to keep logging in and out of irc all the time :) 13:29:53 yes 13:31:55 I progress in porting retro to my vm and implement some code optimisations besides 13:32:09 awesome 13:33:12 most changes should be backportable 13:33:34 the others are changes of language conceptions 13:35:14 I have implement a hashed dictionary for example 13:36:29 neat 13:36:41 that should make things faster, I would imagine. 13:38:08 a lot but would broken retro's vocabulary concept which build upon the traditional linked list approach 13:42:57 seems I must write a compatibly layer :) 13:43:19 crc is considering breaking that in the next version anyway. it might be worth experimenting with now :) 13:44:27 you're right 13:45:16 next major version anyway 13:47:48 hased access compute at 0.74 dispatches for my vm, that is much better than mean 0.9 dispatches for the linked list version which is used in retro 13:48:06 --- quit: kumul (Quit: WeeChat 0.3.9) 13:49:34 the ngaro values should be around 2 for hashed access and somewhat between 3-4 for an optimisated linked list approach 13:50:19 tangentstorm: I fixed "chains" in help.rx to show only chained vocabulary names, and fixed a reference to "emit" in the vt' vocabulary 13:51:38 per iteration 13:52:08 2 what? bytes? 13:52:22 mean vm dispatches 13:53:02 for one access 13:53:54 the value is < 1 for my vm because it executes opcode bundles inside a simple software pipeline 13:58:46 ok, after implementing the hash algorithm in retro it executes with 18 dispatches for a single access 14:03:19 ah, http://en.wikipedia.org/wiki/Dispatch_table 14:03:25 ~43 for seaching the impemented linked list 14:04:06 -> a hashed dictionary access would be much faster in both cases 14:12:20 docl: for interpreters like ngaro which implement some kind of bytecode a jump-table is of advantage. The great switch statement of the ngaro sources are compiled into one (very inefficient handled) 14:13:55 like a real cpu performance related to the number of dispatches needed for a given functionality 14:18:11 crc: how can I define an inline defination in retro ? 14:25:40 --- join: kumul (~kumul@cvx-ppp-66-50-130-239.coqui.net) joined #retro 14:55:13 ciao 14:55:19 --- part: Mat2 left #retro 15:06:17 hi all 16:01:15 hi erider 16:01:24 hi docl 16:39:02 ciao 16:39:09 --- quit: erider (Quit: ChatZilla 0.9.89 [Firefox 15.0.1/20120907231835]) 16:51:46 --- join: arescorpio (~arescorpi@222-206-17-190.fibertel.com.ar) joined #retro 17:27:29 --- quit: ncv (Ping timeout: 246 seconds) 18:49:09 mat2: no inlining support exists, outside of single bytecode functions 20:44:08 --- quit: arescorpio (Read error: Connection reset by peer) 21:16:07 crc: oh, yeah, that should have been vt:emit ... i wrote that before i knew about clear and putc. i'll clean that whole thing up soon. 21:19:11 crc: thanks for fixing the chain stuff. :) i wanted to do that but i'm not quite there yet. 21:50:43 --- quit: kumul (Quit: WeeChat 0.3.9) 23:59:59 --- log: ended retro/12.10.10