00:00:00 --- log: started forth/04.09.09 00:09:08 man copyrights are too damned long lol 00:13:12 --- join: mur (~mur@kyberias.uiah.fi) joined #forth 00:15:49 --- quit: mur_ (Read error: 60 (Operation timed out)) 00:52:50 --- join: Serg_penguin (~z@212.34.52.140) joined #forth 01:04:37 --- join: crc (crc@62-pool1.ras11.nynyc-t.alerondial.net) joined #forth 01:31:25 --- join: mur_ (~mur@smtp.uiah.fi) joined #forth 01:42:13 --- quit: mur (Read error: 110 (Connection timed out)) 02:33:27 --- quit: Serg_penguin () 03:03:11 --- join: mur (~mur@mgw2.uiah.fi) joined #forth 03:15:45 --- quit: mur_ (Read error: 110 (Connection timed out)) 04:05:17 --- join: mur_ (~mur@uiah.fi) joined #forth 04:10:25 * crc is still trying to fix literals in the 16-bit RetroForth 04:15:57 --- quit: mur (Read error: 110 (Connection timed out)) 04:20:51 --- join: mur (~mur@mgw2.uiah.fi) joined #forth 04:33:23 --- quit: mur_ (Read error: 110 (Connection timed out)) 04:53:42 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 05:25:40 --- join: I440r_ (~mark4@216-110-82-59.gen.twtelecom.net) joined #forth 06:01:11 --- join: crc_ (crc@64-pool1.ras11.nynyc-t.alerondial.net) joined #forth 06:01:35 --- quit: crc (Nick collision from services.) 06:01:37 --- nick: crc_ -> crc 06:04:53 --- quit: I440r (Nick collision from services.) 06:04:55 --- nick: I440r_ -> I440r 06:05:03 --- join: I440r_ (~mark4@216-110-82-59.gen.twtelecom.net) joined #forth 06:05:04 --- mode: ChanServ set +o I440r 06:24:23 --- join: madwork_ (~madgarden@derby.metrics.com) joined #forth 06:25:19 --- join: zruty (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth 06:34:15 * crc sighs 06:35:05 crc, cheep up... here is some code for you: 06:35:06 : ['] ' postpone literal ; immediate 06:36:15 :-) 06:36:51 I'm trying to port the routine that handles literals from 32-bit RetroForth to 16-bit RetroForth 06:36:58 It's not working :-( 06:37:10 UGH 06:37:13 no 06:37:17 Hmm. What's the problem? 06:37:21 --- nick: madwork_ -> madwork 06:37:39 : ['] compile literal ; immediate 06:37:46 or maybe ' literal alias ['] immediate 06:37:47 hehe 06:37:47 I think I figured it out... testing... 06:37:58 got it :-) 06:38:07 I440r, I like postpone. :P 06:38:12 Now the 16-bit compiler is half finished 06:38:15 postpone is evil :P 06:38:17 nuff sed! 06:38:20 (mostly because that's what I implemented :P) 06:38:35 and the alias method above makes the definition for ['] take up no code space hehe 06:39:01 err no i fergot the tick grrr 06:39:21 so : ['] ' compile literal ; immediate 06:39:34 mental note - think about what your coding first! 06:39:35 lol 06:39:39 Heh. 06:40:39 Either way, I'm still using postpone. :P 06:40:55 crc do you understand what "postpone/compile" are doing there ? 06:41:08 * crc does not use postpone 06:41:09 Here's another one for you: 06:41:12 : (is) >body ! ; 06:41:12 06:41:12 : is 06:41:12 compile? 06:41:12 if 06:41:12 ' postpone literal postpone (is) 06:41:14 else 06:41:16 ' (is) 06:41:18 then 06:41:20 ; immediate 06:41:31 crc, you use compile and [compile] ? 06:41:33 crc : compile? state @ ; 06:42:02 compile and [compiele] are best 06:42:10 I have 'forth' (runtime/compiled words) and 'macro' (compile-time only) vocabularies 06:42:25 Ahh. 06:43:07 : is ' >body state @ if compile %!> , else ! then ; immediate 06:43:54 the ' >body state @ part in my actual code is factored out because the next few words use it too 06:44:03 but thats basically my def heh 06:44:15 madwork, explain does> :) 06:44:27 %!>? 06:44:56 --- join: mur_ (~mur@kyberias.uiah.fi) joined #forth 06:45:18 %!> stores the value at the top of the stack to the address pointed to by the next execution token 06:45:19 My does> is a primitive, can't be 'explained' in Forth. 06:45:52 does> is one of the most difficult words to explain i think 06:46:15 mine uses ;code 06:46:18 : ;code 06:46:18 r> last @ name> 1+ dup>r 4+ - r> 06:46:18 ! ; 06:46:33 But basically, it sets the CF to DODOES which takes the address of the code to execute from the param field and enters it. 06:46:43 : does> 06:46:43 compile ;code ,call dodoes ; immediate 06:53:08 I still have to port does> to the 16-bit system 06:53:17 * crc wonders how long that will take 06:54:34 my definition would work on a 16 bit forth, i STOLE it from one lol 06:54:44 --- quit: mur (Read error: 110 (Connection timed out)) 06:54:46 lol 06:54:53 f83/fpc 06:55:11 crc:shouldddddddnt take long 06:55:47 arke do you want it done "right now" or do you want it done "right" ???? :) 06:56:32 heh 06:56:41 later 06:56:42 going to school 06:59:18 Ok, does> has been ported and seems to work 06:59:25 cool 06:59:28 what definition ? 06:59:30 try 06:59:36 : constant create , does> @ ; 06:59:45 5 constant five 06:59:50 6 constant 1 06:59:51 hehe 07:01:03 }:) 07:02:22 That works 07:02:42 Though '6 constant 1' is sure to confuse some people :-) 07:02:51 :) 07:04:34 I should have this finished by the weekend. All of the hard stuff is done now :-) 07:04:55 6 constant 1 works? 07:05:01 yes 07:05:18 : defer create ['] noop , does> @ execute ; 07:05:25 so it overrides the "if number, push" logic 07:05:38 ? 07:05:50 I *always* check for words first. 07:06:07 If the input isn't a word, it trys to convert it to a number 07:06:15 if that fails, it ignores it and goes on :-) 07:06:40 i've noticed that, how it ignores non-words 07:06:40 ignores ? 07:06:54 yes 07:06:58 is an interesting idea. any way one could trap the unknown word case? 07:07:07 like an "unknown word" hook? 07:07:14 That's not hard 07:07:30 could do something like perl's AUTOLOAD that way. muahaha 07:07:36 :-) 07:07:47 I440r: yes, it ignores things it can't find 07:07:57 it should abort" ?: 07:08:06 no ? 07:08:10 FIND 07:08:18 If you're compiling, it spits the name of the word, then a ? 07:08:31 If you're interpreting, it just ignores it silently 07:08:42 Either way, execution continues 07:09:37 If you do : foo 10 repeat crr until ; 07:09:41 And 'crr' isn't defined 07:09:44 just noticed retro4games ... shame i wiped out windows here :( 07:09:45 it says 07:09:50 crr ? 07:09:56 And compiles the rest of the word 07:10:00 :-) 07:10:02 should work on my gf's NT4 box tho 07:10:07 I'll do a linux version sometime... 07:10:19 using SDL? 07:10:22 * crc is looking into using libSDL to provide a framebuffer 07:10:23 Yes 07:10:43 yay 07:10:59 That's the next step in retro4games 07:11:04 then i can start on my almighty ForthHack ;) 07:11:08 :-) 07:11:46 I use Allegro + Forthy, works well. 07:11:58 I don't like Allegro 07:12:06 It doesn't seem to like working for me... 07:12:15 allegro focuses mostly on DOS 07:12:23 linux support is kind of an afterthought 07:12:24 Just give me a nice, linear framebuffer and I'll be happy 07:12:28 Allegro does not focus on DOS. 07:12:36 It used to focus on DOS 07:12:41 It's been Win32/Linux oriented for years now. 07:12:50 It started as a DOS library. 07:12:54 hm. should give it another look i guess 07:13:04 madwork: I've never managed to get it working well under Linux 07:13:05 i used to use a VNC client based on allegro 07:13:21 which vnc client ? 07:13:23 for an "embedded" device, translux LED picturewall namely 07:13:24 and whats allegro 07:13:31 www.allegro.cc 07:13:38 http://alleg.sourceforge.net 07:13:46 worked so well, the translux sales guy recommended it over their own software 07:13:53 to their next client that is 07:14:01 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 07:14:01 --- mode: ChanServ set +o Herkamire 07:14:12 Yea, I like it. Higher level than SDL, if you need that. 07:14:23 SDL is more streamlined. 07:14:33 Which is fine. They're different libs, different purposes. 07:14:38 All I need is a linear framebuffer 07:14:48 allegro is a little crufty and weird, partly the age of it i guess. i do like it tho 07:14:53 SDL is great 07:15:14 * zruty <- crufty and weird 07:15:20 well-designed, simple fast 07:15:45 really someone should write a protocol-oriented binding to SDL or allegro 07:15:56 * crc used to have a stripped down SVGALIB that did nothing but set up the framebuffer and return the address of it :-) 07:16:03 so any process could just write to a socket or pipe and have the display server do the graphics 07:16:08 sorta like X but without all the crud 07:16:19 crc: on some systems you can just mmap /dev/fb0 07:16:33 I know 07:16:40 like mine :) 07:16:50 But this worked better on my hardware at the time 07:16:56 I did that in herkforth just for fun 07:17:07 Cool :-) 07:17:37 I have code to fill the screen with a mandelbrot :) 07:17:48 crc, here's a screenshot of my Forthy Allegro console: http://www3.sympatico.ca/ppridham/misc/project/deathfist/console.png 07:18:23 madwork: there a homepage for Forthy? 07:18:28 I'm having trouble getting herkforth to bood 07:18:30 madwork: nice, but I don't like the shadows under the text 07:18:30 or Forthy+Allegro? 07:18:45 What kind of trouble? 07:19:14 under pearpc it runs for a while, and then jumps to address 0x300 (on a memory load instruction of all things... I think the emulator is broken) 07:19:29 zruty, there's just my sourceforge project page... http://sourceforge.net/projects/forthy/ 07:19:35 Odd 07:19:37 when I tell OpenFirmware to boot it directly, it says "claim failed" and prints the size of the elf file 07:19:52 * crc is looking forward to running Herkforth under PearPC 07:19:57 crc, the shadows help when the text is overlaid on the various editors in the background 07:20:28 madwork: I use a separate window for entering code/text :-) 07:21:12 Yea, but then I couldn't type and see the action simultaneously. 07:21:17 i hate sourceforge - i had isforth on sf for a year and got not one thing accomplished 07:21:30 spent too much time babysitting the account and trying to learn cvs 07:21:35 madwork: split the into two halfs 07:21:43 a month after deleting the account isforth was up and running 07:21:58 I gave up on SF after a month 07:22:03 I440r, I just use TortoiseCVS, works nice. I have no troubles with SF. 07:22:11 i cant even stand the LOOK of sf.net - ever page is sooooo visually cluttered. 07:22:20 crc, sure. I like the way i have it though. 07:22:26 mad too much too much on their pages 07:22:26 yeah, sf is a mess 07:22:45 I440r, I just use it for CVS since I don't have my own server set up currently. 07:22:46 * crc notes that bespin provided a much nicer environment, though I don't use CVS 07:22:57 madwork ever looked at subversion ? 07:22:58 awefull nice of them to offer free hosting though 07:23:12 CVS is ok when you're the only one using it 07:23:17 it rots in almost every other way 07:23:24 I440r, yea... might consider it when I do get a server up. Also DARCS. 07:23:37 thankfully there's also tortoiseSVN now too 07:24:03 Yes. The Tortoise clients are super handy 07:24:05 madwork, i spent months trying to learn cvs - i have a subversion server up and running here on a machine thats on my local network and tathi uses it for isforth-ppc 07:24:18 I can't keep them straight, but I got very frustrated with both darcs and arch 07:24:32 it took me like 2 days to figure out how to install and manage a subversion server 07:24:47 well. a little longer than 2 days for some of the things 07:24:48 * crc should learn to use a real version control system 07:24:49 I440r, I'll probably use Subversion on my own server. 07:25:01 CVS wasnt always easy to set up either 07:25:03 crc, what's wrong with backup.bat? ;) 07:25:17 madwork: I use FreeBSD, .bat files don't work 07:25:21 Heh. :P 07:25:25 crc sure they do 07:25:26 backup.pl then 07:25:30 #! /bin/bash 07:25:35 and chmod +x them :)\ 07:25:37 * crc doesn't touch perl 07:25:43 backup.py 07:25:44 backup.f 07:25:46 I440r: that's what I'm doing now 07:25:46 WHATEVER 07:25:47 :P 07:25:57 I don't understand why it would say claim failed and give the size of my elf file 07:26:04 I'm not trying to allocate that much memory. 07:26:09 the ELF header says to allocate 1MB 07:26:18 herk 1 meg of bss ? 07:26:19 Though mine is a little better than a simple backup script... 07:26:32 herk are you constructing the elf file yourself ? 07:26:57 make sure your bss address's last three hex digits are identical to the last 3 digits of the bss sections file offset 07:27:00 I don't know anything about bss. it just has one load in it, it says to load the file at the begining of a 1MB allocation 07:27:08 yes, I'm making it myself. 07:27:10 personally i like perforce 07:27:18 works great under linux. I changed the load address for OF 07:27:45 Maybe I'll just finish writing my version control script 07:27:47 We user perforce here 07:27:51 use 07:28:02 perforce is so simple small and fast that i could easily use it as a wiki backend 07:28:13 Hmm, actually... the "demo" version gives you 2 accounts, I think. Maybe I'll just use it. 07:28:25 I'm not aware of anything to do with BSS. I have one load section in my elf of type PT_LOAD 07:28:30 the windows gui for perforce is really buggy tho 07:28:42 zruty, how? Seems to work fine here. 07:28:58 madwork: always seems to get out of sync 07:29:36 Yea, mostly when other people edit files. :P 07:29:37 commit changes, and they dont get reflected in the client until i stop and restart it 07:29:44 granted this was more than a year ago 07:29:46 Huh... never seen that/ 07:29:47 so maybe it's fixed now 07:35:23 Is there a word for SWAP OVER? 07:36:01 tuck? 07:36:24 At least, that's what I use :-) 07:36:24 : tuck swap over ; 07:36:36 herk is this in linux or fbsd ? 07:36:37 Heh, yes. 07:36:40 do you have a data section? 07:37:03 I440r: herkforth runs under Linux 07:37:11 and are you allocating with a brk syscall ? 07:37:14 I440r: herkamire is working on a native version 07:39:57 how are you allocating that memory ? 07:49:08 --- join: jdavidboyd (~user@wbar5.tampa1-4-11-125-246.dsl-verizon.net) joined #forth 07:50:06 I think that the memory is allocated in the ELF header 07:50:15 then thats bss 07:50:27 or his file is 1 meg in size 07:50:35 I see 07:50:46 err - he cant make is memory size larger than his file size for his .text i do not believe 07:53:42 The binary is ~161k under Linux 07:53:55 eek! thats huge! 07:54:10 isforths kernel is 16k lol 07:54:18 extended its like 60 or something 07:55:31 --- join: arke|school (apache@11.198.216.81.dre.siw.siwnet.net) joined #forth 07:57:30 That has the kernel, editor, metacompiler, fractal generator, and a lot of other stuff though 07:57:38 Really it's a cool environmentr 07:58:02 I'm not 100% comfortable with it yet, but it is very nice 07:58:24 * crc has stolen several ideas from Herkforth since he started using it 08:01:47 :) 08:01:58 The latest rev of herkforth has 1,158 words :-) 08:02:03 isforth will grow to be about 100k i think 08:02:24 RetroForth will stay around 5k 08:02:58 Though I'm also working on a bigger Forth (Maybe 15-20k) 08:03:08 --- join: yome (~rewt@132.204.12.120) joined #forth 08:03:47 I440r: linux 08:04:09 I440r: I have one section in the elf file 08:04:16 no brk or anything 08:04:24 I just use the memory that the ELF file specifies 08:04:36 herk are you setting your memory size larger than your file size? 08:04:39 that wont work 08:04:42 yes 08:04:44 why not? 08:04:47 it works under linux 08:05:02 well - it wont work in freebsd :) 08:05:12 I don't care 08:05:26 well - its not SUPPOSED to work in linux 08:05:34 and you cant guarantee its working 100% correctly 08:05:42 I don't see why it wouldn't work 08:05:55 isn't that the point of sys_brk? 08:06:30 yes but brk operates on the data section of the executing process - tho in linux you can also brk your .text section 08:06:40 maybe OF is trying to use the p_addr instead of p_vaddr (like it's loading without virtual addresses 08:06:52 of ? 08:06:58 open firmware ? 08:07:06 yes. I'm trying to get it to boot 08:07:11 this all works fine under linux 08:07:19 maybe openfirmware's elf reader is more correct 08:07:35 less complete more likely 08:07:40 I'm reading through specs 08:07:46 try making your file size and memroy size the same 08:07:50 and see if it will load then 08:08:06 worth a try 08:08:14 I need more memory than that though 08:08:29 if the program runs you will prolly need to add a brk syscall to break to 1 meg 08:08:32 thats what isforth does 08:08:34 I'm going to try to set the real (not virtual) load address 08:08:55 I can't use syscalls when I'm writing the system 08:09:01 there is no system. 08:09:04 just OF 08:09:18 just set your memory size and your program size to the same thing in your program header 08:09:25 and possibly section headers too - i ferget 08:09:37 or dont you have any section headers ? 08:09:43 theoretically they are optional 08:09:50 it's just in one place 08:10:17 ill bet you it loads with filesize = memory size 08:11:21 quite possibly. 08:11:25 * I440r just had to go through all this with freebsd :/ 08:11:34 had to create a bss secion of ONE BYTE 08:11:37 it said "claim failed" and printed the filesize 08:11:55 what load address are you specifying 08:12:01 maybe someone owns that address 08:12:14 0x200000 08:12:19 I took it from some working examplecode 08:12:27 I may have to change a bunch of headers 08:12:48 my code actually maps the rest of the file (everything except the headers) to that address, plus the length of the headers 08:12:51 can you paste your elf header here ? 08:13:31 http://jason.herkamire.com:3/svn/herkforth/kernel.s 08:13:44 that's basically it 08:13:57 I changed the load address 08:14:12 (linux loads at 0x10000000 and OF loads at 0x200000 08:14:43 any reason why your taking a 64k alignment ? 08:14:48 instead of a 4k alignment ? 08:14:58 no 08:15:10 try making your flags = 5 :) 08:15:18 it's aligned on a GB boundary. I don't see how it matters 08:16:06 that doesnt i dont believe 08:16:37 try find out if OF insists you have section headers 08:16:43 or did the example not have any either 08:17:50 I just read through the requirements for the ELF file, and I men them. 08:17:53 met them 08:18:03 except maybe the load address 08:18:25 my elf header example works under linux 08:18:42 the example boot code I have uses ld to create an elf header (with a custom ldscript) 08:19:10 linux isn';t exactly stict about elf headers :) 08:19:31 no 08:19:43 linux is very forgiving of malformed elf files :) 08:19:53 :) 08:20:17 why is virtual address 0x100000054 and physical address 0 ? 08:23:30 --- part: yome left #forth 08:28:37 my example thing says that the physical address is unused. 08:28:44 which I guess is not nessesarily true of OF 08:29:02 depends on whether the real-mode? variable is set 08:29:47 yeah, I don't have any section headers 08:29:59 I didn't see anything in the spec about needing them though 08:32:45 ok, just dissasembled the working bootcode, and it has a different memory size than filesize 08:33:35 but it doesn't have this stupid thing where we don't load the headers (that is, where the load address xs 0x....54 and it loads 0x54 into the file) 08:33:43 and it sets the physical address 08:33:55 the spec says they are optional 08:34:04 well... and it has 7 section header entries 08:34:43 brb 08:34:57 kc5tja: good morning :) 08:35:31 re 08:35:41 * kc5tja sighs 08:36:17 kc5tja: whats wrong? ;/ 08:36:21 * kc5tja does not want to work graveyard next week. But, at least, it'll let me attend Aikido on a regular basis. 08:36:52 But still, it'll ruin my sleep schedule -- I'll now have to sleep during the hottest time of the day (approaching 100 degrees). 08:37:05 hehe ouch 08:37:13 ick 08:37:15 To say nothing of the fact that I'll probably be falling asleep at work for the next two weeks. 08:37:25 :) 08:37:43 AND I'll only have two days off instead of three. 08:38:25 --- quit: zruty ("leaving") 08:38:33 kc5tja: graveyard shift isn't that bad :-) 08:38:57 That depends. 08:39:40 * crc has been working graveyard shift for three years now 08:43:26 * crc looks at the time and realizes that he will have to go to bed soon 08:43:44 Cool, but see, I prefer to get my things done during the day. 08:44:25 I see 08:44:45 That's part of the reason I now stay up till noon most days 08:45:12 I am growing tired of working at night though 08:46:03 In my case, I won't be able to do that -- to get any kind of sleep, at reasonable temperatures, I need to sleep ASAP from work, but wake up some hours earlier. 08:46:36 Ahh 08:46:51 I understand that 08:48:56 Fuck, even then, a good 8-hour sleep is still going to bring me clean into 4PM range. :( 08:49:20 * kc5tja will need to perhaps explore the possibility of building a solar chimney or something to get some natural cooling in the house. 08:50:16 It'll look uglier than sin, but if it works, then that's all I really care about. 08:52:55 :-) 08:53:19 :) 08:56:06 crc, man, thats just gotta fuck with your growth hormone production 08:56:33 going to bed at roughly 9pm is very important to the natural circadian rythm and thats when the maximum amount of growth hormones are produced by the body 08:56:43 when ppl fail to go to sleep around that time, they don't get the full benefit 08:57:11 altho you could counter that thru exercise, getting protein with every meal, and whey protein powder 08:58:48 thin: I haven't gone to bed anywhere near 9pm in 10+ years 08:58:56 :) 08:59:08 yeah most computer guys tend not to 08:59:12 same here :) 08:59:21 arke: thats why you're fat 08:59:24 :P 08:59:25 * crc also eats very little 08:59:44 * crc drinks a LOT of water, fruit juice, and tea though 08:59:58 well going to bed at 9pm is only one part of the picture of growth hormone production 09:00:00 * arke|school punches thin 09:00:13 thin: I stopped growing years ago 09:00:32 but if i was doing some serious bodybuilding/weightifting/strength training, i will definitely focus on following each of the principles 09:00:37 About the same time I began riding a unicycle... 09:01:53 crc: growth hormones is just testosterone and some of the other stuff, higher levels basically increase the strength of your immune system, builds muscle, and reduces your adipose (fat) tissue 09:02:04 ok 09:02:20 --- quit: arke|school ("CGI:IRC (EOF)") 09:02:40 Then my sleeping habits could partially explain some of my recent health problems... 09:02:42 if you could manage to have high growth hormone levels, you could end up quite strong without much exercise :) 09:02:54 :-) 09:03:04 thin: Well, this is why I fully intend on returning to a regular schedule of Aikido, which is the first time I can pull this off in over three years. 09:03:20 kc5tja: cool :) 09:03:58 * kc5tja is also intending on starting to jog again; gahh, I'm such a fat ass. I can't even jog across the parking lot of where I work. Granted, it's two buildings worth, but still! 09:04:08 Plus it'll help my aikido immeasurably. 09:06:08 apparently it only takes 10 to 15 minutes of jogging to get the full metabolic effect, past 15 mins its diminishing returns 09:06:20 which means when i go jogging i only do about 10 mins of it 09:06:37 --- quit: crc ("Time for bed... Goodnight!") 09:06:47 Word for SWAP DUP... UNDER ? 09:07:33 i tend to go for random bike rides and totally knacker myself, heh 09:07:33 * madwork is starting back at the gym tonight for weights and the track, and Judo next week. 09:07:40 madwork: You mean SWAP OVER? 09:07:47 SWAP DUP is just SWAP DUP. :) 09:08:04 SWAP OVER is TUCK, right? 09:08:10 I figured SWAP DUP would be UNDER. :) 09:08:20 But, I dunno. 09:08:27 define UNDER another word to remember ;) 09:08:45 unless UNDER already exists.. 09:09:18 undefined in Win32Forth anyway 09:10:21 no swap OVER maybe heh 09:10:29 thats a tuck 09:10:54 swaop dup rot ? 09:11:02 i cant even remember what under is supposed to do 09:11:12