00:00:00 --- log: started forth/02.12.23 02:42:20 --- quit: lament ("mental mantle") 03:45:15 --- quit: proteusguy (forward.freenode.net irc.freenode.net) 03:45:15 --- quit: TreyB (forward.freenode.net irc.freenode.net) 04:39:48 --- join: FML (~FML@p50805FDB.dip.t-dialin.net) joined #forth 04:39:54 'lo 04:42:08 [13:40:47] 04:43:10 hi 04:43:16 hey 04:47:02 where is the src of enth.bin? 04:56:48 + i wonder from where it actually was metacompiled from... 04:59:45 onetom: hey nice 2 meet you 05:00:22 'morning 05:03:25 Speuler_: 'morning. just woke up? 05:04:43 not so long ago. just had a fast break 05:06:30 head still feels muddy 05:15:35 fresh air! 05:16:33 sorry, out of stash 05:31:55 --- join: ham (ham@stef.tvk.RWTH-Aachen.DE) joined #forth 05:31:59 hi 05:40:01 --- join: edrx (~user@200.155.19.36) joined #forth 05:40:08 hi 05:40:20 hi 05:40:47 are you familiar with forth? 05:40:50 do you remember the name of the colorforth channel? 05:40:55 I used to be 05:41:19 question? 05:41:22 heh no sorry i just wanted to know what makes forth special and what is it good at 05:42:32 edrx: afaik there is not irc channel for colorforth 05:42:49 edrx: so few users and colorforth doesn't have an irc client 05:42:56 it's incredibly funny to use - I mean, when you're using a system that you understand... and very often problems that would seem to take weeks to implement in other languages can be programmed in forth in a VERY short time 05:43:28 ham: simplicity, predictibility, factoring 05:43:53 from what i know is stack based isnt it? 05:44:06 it is stack based. 05:44:08 such that it makes it good for fast processing on most cpus 05:44:40 not really. writing a fast forth compiler for RISC cpus with huge register files is farily tricky 05:44:49 XeF4: hm, ok... I'm still waiting for a version that can be booted by grub, but I don't have much time to work on that myself right now 05:45:00 there is one that can be booted by grub 05:45:37 can't recall the name and it was of such quality that I never used it again after trying it 05:45:38 XeF4: really? d'you have a url, or keywords for searching for it? 05:45:44 :( 05:46:16 do you happen to know something about CSP? 05:46:39 CSP? no. 05:47:06 aha, ForthOS - http://www.forthos.org/ 05:47:40 ham: I think that the most important point is saving programmer time, and making people regain total control over their machines. CPU cycles are not so important - or shouldn't be. But there are many writers of Forth systems that care only for execution speed 05:48:31 edrx: who are these writers? I want to see a forth written only for execution speed 05:51:13 ham: I guess that this is one of the main intentions behind isforth... also, gforth is quite fast, but I find it almost impenetrable... and marcel hendrix seems to care a lot about speed in his Forths. Chuck's Forths seem to be very very fast, but he has lots of other concerns too 05:51:51 so you think that forth is still worth a look yes? 05:52:04 do you think that there are some applications for ai? 05:54:41 ham: what do you call "applications"? btw, are you in the windows world or in the unix world? 05:54:45 I think forth is worth a look. Otherwise I wouldn't use it. 05:55:12 unix :) 05:55:26 ham: I think that Forth is one of the very few really interesting programming languages 05:55:55 I thought isforth's goal was to be a simple, tractable, reasonably fast Forth for Linux 05:56:48 the others, IMO, would be Icon, Emacs Lisp (because of Emacs), ML, and Tcl, because of Expect 05:57:03 lisp is fine imho 05:57:23 ive heared of icon what does make this so special? 05:58:25 XeF4: I have to confess that I don't know much about isforth, I just carry some old prejudices against it, because of I440's attitudes :S 05:58:25 emacs lisp is nothing too special 05:58:35 but it is "good enough" and it comes with emacs 06:00:04 XeF4: yep - the good thing about it is that Emacs is ridiculously open and extensible, you can even put emacs lisp in arbitrary text files and execute them with C-x C-e... I use that as hyperlinks, and to invoke debuggers, etc 06:01:11 re: i440r's attitudes. yes, they are a bit abrasive at times. =< 06:02:15 ham: go to http://www.cs.arizona.edu/icon/ftp/doc/ and download one of the versions of _The_Icon_Programming_Language_ (the "blue book") 06:02:24 as for the code-in-buffers Enth/Flux can do that. 'tis a shame the the rest of the system is so flakey 06:04:14 XeF4: I took note of that, I'll take a look when I have time 06:04:19 and fpc for dos can do that 06:06:58 and I was working on something like that for a colorforth, but I've not touched the sources for 3 months so don't count on it being ready too soon. 06:08:05 XeF4: I have used (or tried to use) fpc for a while a long time ago - I had used HS-Forth rather heavily before that, and HS-Forth had an environment that I considered brilliant, that was much simpler, and non-windowish. It was inspired in the environment of the old ROM Basic/ gwbasic for PCs... 06:08:43 XeF4: you could walk around the screen and edit text anywhere. If you pressed , that line would be executed. 06:08:58 like CBM basic then 06:09:05 XeF4: yes 06:09:26 never tried HS-Forth 06:09:30 and emacs has a similar feel to me 06:09:59 emacs has too much separation between the Lisp system and the editor for my taste 06:10:22 but it could be much worse (C) 06:10:24 XeF4: it was commercial, unfortunately... I still have it, but most of the documentation is only in paper form. 06:12:06 XeF4: I have a small library for Emacs that blurs that distinction, but I never polished its docs enough to make it become an official part of emacs... and once I met RMS and we started to talk about it, and he said something that pissed me off - that "users should never be forced to see lisp" 06:12:20 I'm not sure about what's his current position, though 06:12:59 I daresay he's too busy with the political side to have a position on such trivial matters. 06:14:09 XeF4: oh, I don't think so, he has become again the official maintainer for Emacs, and he spends a lot of his free time working on Emacs 06:14:35 when did this happen? I've been out of touch with most things GNU since May last year or so 06:15:11 I think that around the time that Emacs21 was officially released 06:18:26 aha. there were unofficial prereleases of emacs21 circulating just as I took an extended holiday from such things 06:18:33 so July last year or so I guess 06:19:34 XeF4: I have to go... my home page is , btw, and I'm going to work on it now, it's a mess... thanks for the infos and the inspiring conversation :) 06:19:37 anyhow, what does this library do? 06:19:54 later 06:19:59 XeF4: it does two very simple things... 06:20:06 XeF4: I can explain that now :) 06:21:42 one, it implements a series of functions with names like `find-xxx' - for example, (find-forthosfile "README.ForthOS" "Output via TTY") can works as a hyperlink 06:22:53 (find-node "(gawk)Array Example" "five") is a hyperlink to an info node, (find-man "gforth" ...) a hyperlink to a manpage, etc 06:24:13 and it also implements a command called "eev", that lets you mark a block and save it to a temporary script file, that you can invoke from a shell; so you can also put arbitrary shell code inside text files, and execute it conveniently 06:25:15 and there are many variations on these things. There's a sort of a tutorial at 06:26:58 later - I'm not sure if the package is clean enough, etc - the whole eev project is napping while I finish my PhD thesis 06:27:11 bye 06:30:34 --- part: edrx left #forth 06:37:01 --- join: TreyB (~trey@cpe-66-87-192-27.tx.sprintbbd.net) joined #forth 07:32:31 hrm, time to switch to a forth with up-key memory :) 07:34:02 iangreen: write your own :) 07:36:34 --- quit: FML ("bb") 07:40:23 soon i will be ready, Speuler_ :) 07:40:36 once i write a fairly decent sized program 07:41:45 --- join: Serg_Penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 07:41:52 yeah, I'm giong to have to though.. I can see tha talready :) 07:42:02 hi 07:42:21 hey 07:42:45 what are u talkin about ? 07:47:44 --- quit: Serg_Penguin () 07:48:43 damn u i44or! make up-arrow buffer 07:51:14 --- join: XeF4_ (xef4@lowfidelity.org) joined #forth 07:53:21 --- quit: XeF4 (Read error: 104 (Connection reset by peer)) 07:54:37 what exactly does leave do ? 07:54:40 in if/then 07:55:09 ahh, got it i think. 07:57:25 iangreen, : break out of a do ... loop 08:01:40 yah, i was thinking it broke out of the word for some reason 08:01:51 should I avoid using the return stack outside of a do loop 08:02:16 eg... foo >r 5 0 do (get access to foo) loop r> drop ; 08:02:38 seems yucky, i'd have to r> r> >r 08:02:48 guess i should use datastack 08:05:20 or is i different from r@ ? 08:05:35 seems I thought r@ should be == i 08:05:44 but in isforth, r@ doesnt seem to give i 08:12:43 yes, you should avoid that 08:13:40 loop index and/or limit are not required to reside under r 08:14:39 k. 08:14:51 it's not bad that I am using dup a whole hell of a lot is it ? :) 08:15:26 nope, it's use is quite normal 08:15:59 and consistent with the idea "called words use up their arguments" 08:20:59 right. 08:22:12 damn, its so easy to just throw here on the rstack and not have to worry about stack manip :) 08:24:13 I suppose it would be ok if i did a >r .. r> outside of my loop, yes? 08:24:30 i've got 4 numbers going on here, seems 1 too many 08:27:46 iangreen: if you do an address calculation, consider to loop through addresses, rather than by loop index incremented by one 08:29:12 Speuler_: hmm what's a nice wy to do that? 08:29:32 like bla here 10 cells bounds do dup i ! cell +loop drop 08:29:38 or 08:29:49 ahh, cell +loop ? 08:29:52 neat 08:29:54 bla here 10 bounds do dup i c! loop drop 08:29:59 whats bounds 08:30:00 ? 08:30:08 whats c! ? stores into here ? 08:30:21 calculates end+1 start from start range 08:30:41 erm 08:30:50 * iangreen thinks / looks up 08:30:52 c! is byte store 08:31:21 speuler: so I can do 5 allot 1 c! 2 c! 3 c! ? 08:31:23 etc 08:31:29 oh, no 08:31:41 cos its 5 cells, which are 4 bytes to me 08:31:47 why would i want to store bytes ? 08:32:00 Also, C! does not increment anything. 08:33:52 iangreen, some data is often arganized as bytes. think of character strings 08:34:10 so i can write cells chars per ... cell allotted? 08:34:16 4 chars per cell? 08:34:21 in my case 08:34:29 iangreen, on 32 bit systems, yes 08:34:31 i've been writing one to each cell :) 08:34:32 hehe 08:34:35 doh :) 08:34:40 on 16 bit, it's 2 chars per cell 08:34:48 iangreen: Unicode :-) 08:34:58 bah i dont need it :) 08:35:04 but *nod* 08:35:30 so how does c! know where to point to if its not incremented ... hrm 08:35:52 C! takes two parameters from the data stack: the value and the address 08:36:00 ah - 08:36:25 create values 1 , 2 , 3 , <--- creates 4 cells? 08:36:40 so that could be rewritten to take up 25% the space? 08:36:43 on a 32 bit system ? 08:38:22 right 08:38:29 c, 08:38:33 It takes at least three cells. The spec leaves the dictionary structure undefined. 08:38:39 sweet 08:39:15 if , stores to 1 cell, that's c, do then ? 08:39:33 i feel like im asking silly questions sometimes... but i kinda feel like im getting somewhere :P 08:39:38 It "commas" in a byte. 08:39:51 ok, that's a tad counterintuitive 08:40:03 i'd think of c for cell 08:40:09 C == Character == Byte 08:40:10 Id think c, would write to a cell and , would write one byte 08:40:15 ahh 08:40:16 ;) 08:40:23 fair enough! 08:40:30 iangreen: it comes from VERY old Forth practice. 08:40:35 understandable 08:43:21 Forth uses the undecorated operators for cell-wide values because mostly you deal with cells. 08:46:17 cheers 08:47:23 --- quit: ham ("what makes a man become neutral?") 08:50:30 * Speuler_ also tends to represent data in an uniform manner, using cell when bytes could do 08:51:23 but my strings consist from char-size quantities 08:51:27 of 08:51:42 * Speuler_ feeds 08:51:58 char-size quantities? 08:52:07 I'm trying to see a reason why I should waste 3 bytes per char :) 09:00:35 --- join: I440r (~mark4@sdn-ap-031tnnashP0306.dialsprint.net) joined #forth 09:00:44 ok, i added block files :) 09:00:57 you can have up to 256 different block files open at one time :) 09:04:04 I440r: what do you think, does it seem wasteful to you to store only 1 char in a cell? 09:05:28 the larger the cell, the more wastefull :P 09:05:36 ya 09:05:56 the fix is to use c@ and c! :) 09:06:02 trying to understand c! now : "666 b c!" "b @ ." 154 09:06:08 why 154.. 09:06:13 char address c! 09:06:15 .whut am i doing wrong 09:06:31 0 to 256 only 09:06:33 oh, 666 isnt a valid char is it 09:06:34 :) 09:06:38 well 09:06:41 c! doesnt know that 09:06:44 hang on 09:06:44 *nods* 09:07:07 666 is 029a in hex 09:07:18 c! will only store the 9a part 09:07:23 word 09:07:38 0x9a is 154 09:07:49 Some of us want to store Unicode glyphs, though. 09:07:52 isforth has c@ c!, w@ w! and @ and ! 09:08:42 what's w@ and w! do? 09:08:58 "word" or 16bit sized, usually. 09:09:30 yup 09:09:30 i dint invent the idea heh 09:09:49 its an obvious choice of names tho 09:10:19 It certainly follows the common Forth practice. 09:10:43 * TreyB half-heartedly suggests 16@ and 16! 09:11:11 i.e. obviousness :) 09:11:42 no. i dont even like 2blah etc for double size 09:11:57 d@ and d! would be preferable in a 16 bit system 09:11:57 instead of 2@ 2! <-- very bad names 09:12:12 2swap should be dswap too 09:13:09 another thing, it seems that i != r@ in your do loops, is that correct? 09:13:21 i gotta write a block file editor now 09:13:23 I read in a (rather outdated) book that i is sometimes written as r@ 09:13:33 yes i and r@ are not the same 09:13:36 i noticed that doesnt seem to be the case in isforth 09:13:39 in SOME forths yes 09:13:44 ok 09:13:48 where do you store the index? 09:13:54 if not on the rstack 09:13:58 just out of curiosity 09:14:06 None of Your Business (tm) 09:14:20 :-) 09:14:27 but not in isforth. you have to examine (do) to understand why tho 09:14:27 it IS on the return stack 09:14:27 but its had some fudge factor applied to it 09:14:48 hrm, ok, more than i want to know at this point :) 09:14:51 LOOP will cause the overflow flag to be set when the loop reaches its limit 09:15:10 if you do 10 for blah blah nxt 09:15:20 you can get the for/nxt loop index by doing r@ 09:15:22 not i 09:22:14 grr, im having way too much difficulty juggling n and here in the beginning of : makestring ( n - addr ) 09:22:57 show me the code for makestring 09:23:08 im (re)writing it as we speak 09:23:20 i keep rewriting it 09:23:21 hehe 09:23:33 its good practice :) 09:23:51 yup 09:23:57 look at this now: 09:23:57 : makestring ( n - addr ) 09:23:57 dup here dup >r swap allot 09:23:57 0 do 09:23:59 looking rather 09:24:12 then gonna getkey and store in address as i loop 09:24:36 and then "r>" at the end I guess 09:24:50 thers dup>r as one word btw 09:24:52 seemed bad initially as i'd want to acoid mesing w/ return stack, but since do loop shouldnbt use it 09:24:57 k 09:25:04 that makes me feel better about it 09:25:09 that looks ok tho ? :) 09:25:25 but if you put an item on the stack outside a do loop you cant easilly grab it inside the do loop 09:25:38 ahh right. 09:25:51 i can't just do r@ in my loop 09:25:53 hmm 09:25:58 no point then 09:26:16 guess I can throw "n" on the return stack, allocate, grab n back 09:26:24 you could do 2r> r@ -rot 2>r but thats very very kludgy heh 09:26:47 whats happening inside the do loop exactly ? 09:26:54 2r> = r> r> ? -rot = ? 09:27:01 I'm passing the address to store the char in 09:27:08 : -rot rot rot ; 09:27:16 ok. 09:27:17 curraddr getkey 09:27:22 : getkey ( addr - ) 09:27:22 key dup emit dup 10 = if dup swap c! then drop leave ; 09:27:34 which may be broken :) 09:27:48 dup swap ? 09:27:52 *shrug 09:28:01 need to rewrite this 09:28:12 ( n1 --- ) dup ( n1 n1 --- ) swap ( n1 n1 --- ) 09:28:13 heh 09:28:22 haha 09:28:25 indeed :) 09:28:30 must have meant dup rot ;0 09:28:37 * iangreen goes back to tinkering 09:28:48 drup rot or swap over :) 09:29:45 or just swap :) 09:35:26 why is this never finishing? 09:35:26 : makestring ( n - addr ) 09:35:26 here swap 2dup allot 09:35:26 bounds do 10:00:42 anyone here have a source for a block file editor ? 10:00:44 :) 10:02:11 --- join: FML (~FML@pD9E4EA7F.dip.t-dialin.net) joined #forth 10:02:19 hi fml :) 10:02:26 you a forth coder by any chance ?? :) 10:02:43 newbie 10:02:50 newbies welcome :) 10:03:22 met forth couple of days ago 10:03:30 cool! 10:03:34 which forth ?? :) 10:04:08 I440r: thx for the welcome. 10:04:36 anyone who comes in here is always welcome. forth coder or NOT :) 10:04:51 were not snobs in here :) 10:04:57 except when it comes to our coding standards :) 10:07:13 was looking for suitable langs to express myself. read the description of gforth.deb. a strong guess told me immediately, yes this it. what i was looking for. was always very unhappy 'bout those major langs. 10:07:40 do you code assembler ? 10:07:52 a tiny bit 10:08:14 could grow by necessarity 10:09:10 why? ;-) 10:09:51 because understanding assembler will help you understand forth 10:09:56 even if you only get the basics 10:10:14 and i strongly feel that anyone wanting to be a GOOD coder should be an expert at assembler 10:10:35 ok=understand+agree 10:10:49 true! 10:11:10 getting the feet back on earth... 10:12:09 getting a grip 10:13:12 what os do you use ? 10:13:29 i dont know any assembler :/ 10:14:10 learn it then 10:14:12 :) 10:16:27 debian knoppix mandrake redhat suse m$win are currently permanent (no no, not in order of preference... ;-) ) 10:16:43 heh 10:16:49 debian is my primary os 10:16:58 i also have m$wincrap too 10:17:10 bsd and hurd are still missing, unfortunately 10:17:33 i have bsd installed on one machine here so i can get my forth to work there eventually 10:17:40 but thats way down the road - if ever :) 10:19:18 wincrap blue screen twice a week (NOT kidding) oh yeah ;-) :-) 10:19:48 heh 10:19:50 twice a week? with regular desktop use? that is very odd 10:20:03 y 10:20:35 because wincrap often bluescreens 3-4 times/day with any real use 10:20:48 y. 10:20:55 I wish I new 10:20:57 knew 10:21:27 nt series is stabler though 10:22:25 or in english: more stable... ;-) 10:23:33 what's wrong with this .. arg.. 10:23:33 : makestring ( n - addr ) 10:23:33 dup>r here swap allot 10:23:33 here >r bounds do key dup emit swap c! loop ; 10:23:46 vakaampi :-) 10:24:23 oops, r> 10:25:05 iangreen: what do you intend makestring to do? 10:26:45 allocate n, store string there, return addr 10:26:57 and size, or use null string ending 10:27:10 dont tell him about expect, he is learning forth :) 10:27:17 Store what string there? 10:27:29 iangreen: keep going. forget you ever heared me say "expect" hhe 10:27:33 TreyK: from key 10:27:39 I440r: done and done :) 10:27:42 Ah. 10:27:55 TreyB: im doing a bit too much in one word i know 10:28:05 does my loop and c! look right? 10:28:18 they dont seem to be :) 10:28:26 does your forth keep thins in pad/just above HERE? 10:28:29 things, even 10:28:38 it it does, you'd better not do it that way 10:28:50 um.. i wouldnt know :) 10:29:02 oh, i should use allocate maybe 10:29:10 or is that just allot 10:29:14 oh wait, you're alloting 10:29:15 nevermind. 10:29:26 yeh 10:29:32 dynamically creating this string basically 10:29:45 from user input... return the addr (and a size, I guess) 10:29:51 tho i think ill just use \0 10:31:32 also you should dup here instead of invoking it after allot because here of course changes with allot 10:31:39 iangreen: I don't think you can count on "10 here swap allot here -" to leave 10 on the top of the stack. 10:31:39 so you are building the string to the wrong addr 10:31:53 oops 10:32:07 oh LOL 10:32:08 no kidding 10:32:12 good point :) 10:32:30 stupidity just keeps creeping back in to this word :) 10:32:39 iangreen: make it shorter 10:32:54 reading enth docs, i guess the kernel src code is metacompiled from win32forth? or is it compliant AND 'metacompiles'/recompiles itself? 10:33:20 and you should do i c! 10:33:36 "here dup>r over allot 10:33:40 that'll be much better :) 10:33:52 how does this even execute without crashing? 10:33:56 XeF4_: it doesnt! 10:33:59 sorry, im learning 10:34:05 let me try some more 10:34:20 ah yes 10:34:26 i can get this damnit 10:35:01 iangreen: on which forth are you learning right now? 10:35:24 isforth 10:36:04 written in which assembler? 10:36:37 nasm, I believe.. ask I440r, it's his forth 10:37:21 running standalone? 10:37:45 nasm 10:37:47 it's linux,... all i/o is impl w/ linux kernel 10:37:48 On top of linux. 10:38:02 it treats linux as its bios heh 10:38:27 has it a relation to forthos? 10:38:39 by structure? 10:40:14 btw does anyone know of 10:41:16 ? 10:41:20 an assembler written in forth and then compiled for a standalone binary 10:41:48 ? 10:42:14 err isforths compiler kernel is coded in nasm assembler. 10:42:43 er, sort of. bigforth can be used as a standalone binary and it has an assembler 10:42:52 so if you ignore other features than the assembler, there is. 10:50:45 this one is crashing too, but it looks more correct: : makes swap 2dup allot bounds do key i c! loop ; 10:50:59 5 makes ... 10:51:32 what should it do? 10:52:27