00:00:00 --- log: started forth/13.01.16 00:13:20 --- quit: ASau (Ping timeout: 272 seconds) 00:18:15 --- join: epicmonkey (~epicmonke@188.134.41.175) joined #forth 00:18:15 --- mode: ChanServ set +v epicmonkey 01:10:48 --- join: dessos (~dessos@c-174-60-176-249.hsd1.pa.comcast.net) joined #forth 01:10:49 --- mode: ChanServ set +v dessos 01:36:51 --- quit: john_metcalf (Ping timeout: 248 seconds) 01:39:00 --- join: john_metcalf (~john_metc@211.67.125.91.dyn.plus.net) joined #forth 01:39:00 --- mode: ChanServ set +v john_metcalf 01:51:32 --- quit: Nisstyre (Ping timeout: 248 seconds) 02:04:06 --- join: Nisstyre (~yours@oftn/member/Nisstyre) joined #forth 02:04:06 --- mode: ChanServ set +v Nisstyre 02:04:19 --- quit: tgunr (Ping timeout: 240 seconds) 02:09:34 --- quit: epicmonkey (Ping timeout: 245 seconds) 02:50:30 --- quit: proteusguy (Remote host closed the connection) 03:04:13 --- join: proteusguy (~proteusgu@ppp-58-8-88-123.revip2.asianet.co.th) joined #forth 03:04:13 --- mode: ChanServ set +v proteusguy 03:11:47 i just got my begin until loop working in my implementation of Forth :D 03:17:32 --- join: epicmonkey (~epicmonke@host-224-58.dataart.net) joined #forth 03:17:32 --- mode: ChanServ set +v epicmonkey 03:35:41 --- join: newcup (newcup@peruna.fi) joined #forth 03:35:41 --- mode: ChanServ set +v newcup 03:58:23 --- quit: newcup (Remote host closed the connection) 04:33:47 --- quit: Bahman (Remote host closed the connection) 04:34:08 --- join: JDat_ (9f9411b5@gateway/web/freenode/ip.159.148.17.181) joined #forth 04:34:08 --- mode: ChanServ set +v JDat_ 04:44:51 someone explain the value of the else statement 04:45:21 if 1 else 2 then =======> if 1 then 2 04:53:26 if number in parameter stack is true then 1 else 2 04:53:28 --- quit: Nisstyre (Ping timeout: 252 seconds) 04:54:27 c style if (PSP_top !=){ 1; } else{ 2; } 04:54:40 c style if (PSP_top !=0){ 1; } else{ 2; } 04:59:33 explain how it is useful, please 05:00:27 i am writing an interpretter....and i could just make else act as if and ignore the next if at compile-time to implement else 05:00:32 interpreter* 05:00:42 make else act as then*** 05:00:51 and ignore the next then*** 05:00:53 uhg lol 05:01:05 i am writing an interpretter....and i could just make else act as then and ignore the next then at compile-time to implement else 05:01:08 there. 05:02:20 oh i think i see now 05:02:46 nope...changed my ming 05:02:47 mind* 05:07:56 --- join: Nisstyre (~yours@oftn/member/Nisstyre) joined #forth 05:07:56 --- mode: ChanServ set +v Nisstyre 05:10:16 sorry, I am forth noob 05:10:58 I think that IF ELSE THEN can be written with ?BARNCH or something similar 05:13:17 makes sense that that word exists....as a generalization of several things 05:22:58 --- quit: beretta (Quit: Leaving) 05:29:54 --- join: impomatic (~digital_w@211.67.125.91.dyn.plus.net) joined #forth 05:29:54 --- mode: ChanServ set +v impomatic 05:34:46 --- join: tgunr (~davec@cust-66-249-166-11.static.o1.com) joined #forth 05:34:46 --- mode: ChanServ set +v tgunr 05:50:30 --- join: newcup (newcup@peruna.fi) joined #forth 05:50:30 --- mode: ChanServ set +v newcup 06:50:59 impomatic: my Forth implementation now has begin/until loops :D 06:52:03 :-) 06:54:05 impomatic: mine implicitely allows recursion...which i like, but i may change so people can define things in terms of old definitions with the same name 06:54:25 impomatic: i might change that just so i can use existing code 07:01:55 Mine Forth also allow recursion. At some point I'll change it and add the RECURSE word 07:03:56 or alternatively, you could add an `old' word 07:04:15 as a compiler directing word that compiles the last thing by that name 07:04:44 : cr old cr cr ; 07:04:48 oops 07:04:59 : cr old cr old cr ; 07:05:18 to make cr do the old cr twice 07:06:54 i think recursion should be by default...but i would like to be able to use standard code 07:07:11 what do you think of `else'...it seems useless to me 07:08:44 if i implement else, i would make it behave like if at compile-time...then ignore the next then 07:09:29 behave like then at compile-time*** 07:11:16 if i did the `old' thing...i could just run a perl script over existing code easy enough 07:11:55 Hmmm... else is useful. 07:13:28 STATE @ IF .... ELSE .... THEN <- example from Forth outer interpreter, do one thing or the other depending if state is compile or interpret 07:14:33 1 while s/(:\s++(\S++)[^;])*$2([^;]*;)/${1}old $2$2/g 07:14:41 something like that to fix things 07:15:01 $2$3 *** 07:15:26 oo damn that would add olds over and over...meh it can be done 07:16:08 impomatic: but that could be written as STATE @ IF .... THEN .... 07:16:19 impomatic: i think else is syntactic sugar 07:17:35 found another problem in my regex :/ 07:19:21 impomatic: oh...my example would still do the second thing if true...i see :) 07:30:24 --- join: fantazo (~fantazo@213.129.230.10) joined #forth 07:30:24 --- mode: ChanServ set +v fantazo 08:07:41 --- join: RodgerTheGreat (~rodger@71-13-215-242.dhcp.mrqt.mi.charter.com) joined #forth 08:07:41 --- mode: ChanServ set +v RodgerTheGreat 08:23:45 --- quit: I440r (Ping timeout: 255 seconds) 08:29:52 --- join: Nisstyre-laptop (~yours@oftn/member/Nisstyre) joined #forth 08:29:52 --- mode: ChanServ set +v Nisstyre-laptop 08:42:10 --- join: I440r (~zhiming@167.sub-70-194-68.myvzw.com) joined #forth 08:42:10 --- mode: ChanServ set +o I440r 08:43:22 --- quit: epicmonkey (Ping timeout: 252 seconds) 08:46:33 --- join: ncv (~quassel@89.35.216.197) joined #forth 08:46:33 --- quit: ncv (Changing host) 08:46:33 --- join: ncv (~quassel@unaffiliated/neceve) joined #forth 08:46:33 --- mode: ChanServ set +v ncv 08:48:44 --- quit: fantazo (Ping timeout: 246 seconds) 08:50:37 --- quit: karswell (Read error: Connection reset by peer) 08:51:30 --- join: karswell (~user@93-97-29-243.zone5.bethere.co.uk) joined #forth 08:51:30 --- mode: ChanServ set +v karswell 08:55:49 --- join: jdavidboyd (~user@72.185.97.240) joined #forth 08:55:49 --- mode: ChanServ set +v jdavidboyd 09:06:53 I've scanned the gforth manual, but I don't see how to load my own file of code at startup automatically every time, like emacs does with .emacs. Is this something that is possible, or do I just have to run 'gforth stuff.fs' every time, and go from there? 09:08:13 --- quit: JDat_ (Ping timeout: 245 seconds) 09:12:06 --- join: MayDaniel (~MayDaniel@unaffiliated/maydaniel) joined #forth 09:12:06 --- mode: ChanServ set +v MayDaniel 09:21:36 --- join: Inode (~inode@time.uk.chromedpork.net) joined #forth 09:21:36 --- mode: ChanServ set +v Inode 09:22:36 jdavidboyd: i think you just put the name of the file after it 09:22:45 jdavidboyd: like gforth FILENAME.fs 09:23:13 jdavidboyd: yes, that is correct :) 09:23:25 jdavidboyd: where .fs is part of the filename 09:23:34 jdavidboyd: so i guess gforth filename 09:23:49 that part I have no problem with. I want to know if there is a 'magic' file that loads every time gforth starts, so I can put some stuff in there, and have it always available? like a definition for binary mode, etc. 09:23:52 jdavidboyd: oh sorry...didn't read your full question 09:24:36 from the man page 09:24:39 --path path 09:24:40 Uses path for searching the image file and Forth source code files instead of the default in the environment variable GFORTHPATH or the path specified at instal- 09:24:42 lation time (typically /usr/local/lib/gforth:.. A path is given as a :-separated list. 09:54:51 --- join: epicmonkey (~epicmonke@188.134.41.175) joined #forth 09:54:51 --- mode: ChanServ set +v epicmonkey 10:06:58 --- quit: Nisstyre-laptop (Ping timeout: 252 seconds) 10:16:30 still not quite what I'm looking for. I'm wondering if gforth has a default file that it loads at startup every time, like emacs does with .emacs, or vim does with .vimrc. 10:17:40 --- join: Nisstyre-laptop (~yours@oftn/member/Nisstyre) joined #forth 10:17:40 --- mode: ChanServ set +v Nisstyre-laptop 10:18:32 jdavidboyd: I notice that my gforth installation has a blank file called "siteinit.fs" and I notice that the last thing "startup.fs" does is require it 10:24:38 jdavidboyd: hrm 10:24:52 I tried fiddling with those and they don't appear to do anything. :/ 10:25:19 there must be some other change necessary 10:36:34 RodgerTheGreat: my Forth implementation now has begin/until loops and if/then conditionals :D 10:37:29 RodgerTheGreat: and it supports recursion implicitely, which i may change later 10:37:44 RodgerTheGreat: this works right now : addtoten 1 + dup 10 = 0 = if addtoten then ; 10:40:36 RodgerTheGreat: ooo shit found a something weird...recursing like that defers all the thens haha...so any code after the then would get repeated 10:53:56 protist: what does your forth do with this if you do 3039 TEST ? : test 10 /mod dup if test else drop then 48 + emit ; (or this version without else) : test 10 /mod ?dup if test then 48 + emit ; 10:56:34 impomatic: i don't have /mod yet....but it is possible to overflow the return_stack...if that is what you are wondering 10:57:37 protist: actually it's a recursive word to display unsigned decimal integers 10:58:42 : test 10 /mod ?dup if recurse then 48 + emit ; <- standard forth version 10:58:43 i would have to implement /mod and ?dup before i could try that out :D 11:00:28 : ?dup dup if dup then ; 11:02:21 well ?dup is working now...i will need /mod 11:05:59 Hmmm... it's probably best to do /mod in asm. In 8086 it's pop bx /pop ax / cwd / idiv bx / push dx / push ax 11:06:27 i am using AT%T...i will see if that translates :) 11:06:37 protist: neat 11:06:49 my forth makes definitions available immediately like that 11:06:58 i just made `literal' 11:07:07 although I am increasingly of the opinion that an explicit "recurse" is cleaner 11:07:34 doing what we do makes it harder to "wrap" existing procedures 11:07:47 RodgerTheGreat: i would prefer an `old' word for when i DON'T want to recurse 11:07:56 which would be simple enough 11:08:06 and if you change the name of a recursive procedure you have to update the name in at least two places 11:08:42 "recurse" is a little more DRY 11:08:59 --- quit: Nisstyre-laptop (Quit: Leaving) 11:09:54 impomatic: i am 32 bit as well 11:09:56 can anyone come up with an argument for why the name "recurse" is preferable to the terser "recur"? 11:10:21 impomatic: my stack would probably get corrupted by the 16 bit instructions 11:10:28 the latter of which is actually an english word 11:10:54 RodgerTheGreat: never though about it 11:11:10 "recurse" is how it's done in gforth at least 11:11:19 seems like a poor name 11:11:32 impomatic: i will try popl %ebx popl %eax cwd idiv %ebx pushl %edx pushl %eax i guess 11:14:52 impomatic: i got a floating point exception....my asm is probably wrong 11:23:12 well you divided by zero 11:24:42 you could always just test with ": /mod 2dup mod >r / r> ;" 11:29:57 i don't have mod or 2dup yet haha 11:30:05 lots of words to put in 11:33:15 protist: : 2dup over over ; : mod 2dup <= if begin swap over - swap 2dup < until then drop ; 11:33:16 :) 11:34:55 ok i got /mod implemented :D 11:35:09 actually a better mod would be : mod 2dup / * - ; 11:35:42 RodgerTheGreat: nice definitions :)...i am missing too many words to use those, though 11:36:15 protist: you don't have "over"? 11:46:59 RodgerTheGreat: nope...what does that do? 11:47:19 it's like dup for the second stack element instead of the first 11:47:29 it's a pretty basic primitive 11:47:47 oh :) 11:48:28 I'm kind of surprised that you're halfway through writing a forth without knowing about "over" 11:50:12 haha 11:51:24 --- quit: ncv (Ping timeout: 252 seconds) 11:52:34 impomatic: are you sure the non-else version of test is correct....i tried it and i get a `3' then three messed up unicode things or something 11:52:46 s/correct/correct?/ 11:52:47 protist: I'd tell you about "swap" but it might blow your mind 11:52:56 RodgerTheGreat: lolol 11:53:44 RodgerTheGreat: i also had to check just now to see what nip and tuck do...i usually only use swap, rot, dup, and drop 11:54:27 I usually only use dup, drop, swap, over, r> and >r. 11:55:21 ah i can see how over is handy haha 11:55:48 you might find yourself using fewer rot 11:55:51 instead of swap dup rot swap 11:56:02 just over haha 11:56:08 jesus 11:56:18 hey it was good practice ;D 11:56:25 I guess? 12:05:20 --- join: ncv (~quassel@89.35.216.197) joined #forth 12:05:20 --- quit: ncv (Changing host) 12:05:20 --- join: ncv (~quassel@unaffiliated/neceve) joined #forth 12:05:20 --- mode: ChanServ set +v ncv 12:07:22 --- join: JDat (JDat@89.248.91.5) joined #forth 12:07:22 --- mode: ChanServ set +v JDat 12:09:37 i now have `[' `]' and `literal' 12:10:06 : ' get_lexeme lookup ; 12:11:31 * kulp jealous 12:11:36 you're going to make me work on mine ! 12:11:49 kulp: :D 12:13:59 protist: just tested with gforth. Also works there. 12:14:44 ' this : dothis10times 0 >r begin literal execute r> 1 + dup r> 10 = until r> drop ; 12:14:52 impomatic: hmmm 12:16:27 impomatic: does your recursion defer computation?...or is yours a tail-recursion? 12:16:46 impomatic: mine defers, and i assume yours does too...:/ 12:16:57 protist: what does that code do? 12:17:18 protist: yeah I don't understand what you're trying to do there 12:17:33 impomatic: makes a word `dothis10times' that does `this' 10 times 12:17:46 well the short version is that won't work 12:17:46 impomatic: i got the execution pointer with ' 12:18:03 RodgerTheGreat: maybe my words are different 12:18:07 At compile time, my recursion just compiles a call to the word (just like any other word). At execution time, it just calls the word (like any other word) 12:18:08 i'll test it 12:18:23 is "literal" immediate? 12:18:48 i didn't think literal made sense as a text word 12:19:05 RodgerTheGreat: i suppose so...i am not fluent in forth, i thought i saw literal in Starting Forth 12:19:06 i thought you just had an execution token for it 12:19:14 it seems sort of odd to have a dictionary definition which takes an xt as an argument at compile time 12:19:49 as far as I can tell that's what you're doing 12:19:58 ooo after dup i meant >r 12:19:59 maybe protist doesn't have a separate compile and execution time ? 12:20:28 kulp: then what the hell does ' this do in that code? 12:21:02 RodgerTheGreat: well what does literal do :P 12:21:13 either way i am confused. but i am no forth expert either 12:21:19 ahg something is off 12:21:38 I know what those words would do in a normal forth and I do not think they would do what protist means for that definition to do 12:21:39 i just thought LITERAL was something that appeared in an xt stream but didn't have interpretation semantics 12:21:56 kulp: my literal takes the thing of the top of the stack at compile time, and compiles it as a constant 12:22:05 so it's immediate 12:22:07 kulp: so i can do 3 : three literal ; 12:22:27 or : three [ 2 1 + ] literal ; 12:23:05 hmm okay well maybe i have the wrong idea about it 12:23:18 --- join: Nisstyre-laptop (~yours@oftn/member/Nisstyre) joined #forth 12:23:18 --- mode: ChanServ set +v Nisstyre-laptop 12:28:08 --- quit: Nisstyre-laptop (Quit: Leaving) 12:30:11 well this is working...but my example of what i thought would isn't ' this : another literal execute ; another 12:30:27 i think my begin/until loop needs a little work 12:30:32 protist: dothis10times works on Itsy, but gforth didn't like it. 12:30:43 impomatic: interesting haha :) 12:31:30 : dothis10times 0 begin >r [ ' this ] literal execute r> 1+ dup 10 = until drop ; <- gforth was happy with this code though 12:34:03 ok i fixed my begin/until...now to try dothis10times 12:34:16 "[ ' this ] literal execute is a 12:34:25 fairly involved way of saying "this" 12:34:37 impomatic: odd that gforth is so finicky 12:34:56 protist, when you get to do loops your going to have problems unless you understand all 3 of the items that DO puts on the return stack :) 12:35:15 and how the run time of (DO) and compile time of DO work :P~ 12:35:30 i have a shorthand for [ 123 ] literal 12:35:34 [ 123 ]# 12:36:17 I440r: three? I thought it was just i=current, i'=max 12:36:53 nope the exit point of the loop is also pushed onto the stack so we can EXIT out of the loop using leave 12:37:07 : leave 2r> 2drop exit ; except its not done in a : definition 12:37:45 hm. That isn't how I've defined DO and LOOP in the past. 12:37:49 DO compiles (do) followed by a dummy cell. the dummy cell will be backpatched with the address of the XT following the TWO xt's compiled by loop 12:38:03 (loop) is a conditional branch back to the start of the loop 12:38:13 I440r: i do have separate compile time and runtime behavior 12:38:21 so : foo 10 0 do i . loop ; would be compiled as ..... 12:38:25 foo: 12:38:29 (lit), 10 12:38:33 (lit) 0 12:38:38 I440r: or i would have had massive trouble getting as far as i have 12:38:39 yet another reason I prefer FOR...NEXT as a looping construct I guess 12:38:46 (do) xyzzy 12:38:51 i dot 12:38:57 (loop) abcde 12:39:00 xyzzy: 12:39:02 exit 12:39:15 abcde: shuld be before i dot 12:39:38 RodgerTheGreat, theres an advantage to do loops. you can count backwards -1 +loop etc 12:39:47 0 10 do ... -1 +loop 12:39:54 or do increments of more than 1 12:39:56 yeah 12:40:01 10 0 do i . 2 +loop 12:40:10 0 2 3 4 6 8 12:40:23 but for loops are preferable 12:40:28 i also have rep 12:40:34 : blah 10 rep foo ; 12:40:37 repeat foo 10 times 12:40:43 hm 12:40:44 i can also do 12:40:48 how do you define rep? 12:40:49 10 rep foo 12:40:54 in interpretive mode 12:41:04 --- join: kylert__ (~kroot@c-65-96-213-133.hsd1.ma.comcast.net) joined #forth 12:41:04 --- mode: ChanServ set +v kylert__ 12:41:14 (rep) uses for loops 12:41:28 the behavior of rep is state dependant 12:42:10 : (rep) swap for dup>r execute r> next drop ; 12:42:45 : dorep param (rep) ; \ will show def for param in a sec 12:42:59 : rep state @ if compile dorep else ' rep then ; immediate 12:43:20 rep sort of reminds me of how chuck prefers IF A EXIT THEN B to IF A ELSE B THEN 12:43:45 my ?: is more efficient than CM's method 12:44:01 : foo some-test ?: foo bar fud ; 12:44:13 if sometest returns true FOO is executed and bar is skipped. 12:44:23 if it returns false foo is skipped and bar is executed 12:44:30 either way execution continuses with fud 12:44:47 ?: executes one or other of the 2 xt's following it then continues with the 3rd 12:45:14 thus *forcing* you to factor your conditional bodies 12:45:15 i can define min/max in as few cells as chuck moores every efficient but BAD versions using ?: 12:45:41 one of the reasons i prefer it to if/else/then but sometimes if/else/then is needed 12:46:09 it also compiels in way fewer cells than if/else/then because the if is 2 cells. followed by 2 more at the end of the if block to branch round the else block 12:46:28 then also compiles DOTHEN which is a noop - this is simply an aid for the decompiler 12:46:50 my case: constructs also FORCE you to factor because i have NO endof 12:46:52 you cant do 12:46:59 I440r: but you also pay for the dictionary entries associated with the two halves of the branch if they aren't reused elsewhere 12:47:02 case 10 of 1000 lines of code endof 12:47:16 no you dont. turnkey discards them :) 12:47:26 i dont count head space as being part of any app 12:47:48 each definition still costs you a return, which is one cell 12:48:09 so you only break even if the definitions are not reused 12:48:19 you save space if they're reused 12:48:37 if noop else noop then is compiled as a ?brahch + a vector, noop, brahch + vecror, noop, dothen 12:48:49 my ?: would be ?: noop noop 12:48:50 3 cells 12:49:11 space is not the issue. factoring is 12:49:25 plus i think my ?: executes faster too 12:50:25 even if your going to use IF ELSE THEN i say factor the IF part and ELSE part into separate words 12:50:43 well... where possible 12:50:45 cant always do that 12:50:57 I'd say it depends on the size and complexity of those branches 12:51:08 if they're repeated code, absolutely factor them 12:51:10 like when doing if r@ blah else r@ fud then where you have put something on the return stack beforehand 12:51:43 even if you never repeat them. factoring them out is preferable for the sake of readability 12:52:09 : fud 1+ swap ; <-- even if i only use fud ONCE iwould still factor out probablh 12:52:11 just like loop bodies it often makes more sense to have a body that is a few words long and *most* of the code is factored out 12:52:55 oh theres another reason why DO loops are sometimes preferable 12:53:10 with for loops one implementation will count the loop UP, others will count it DOWN 12:53:31 whereas with do loops YOU decide which direction the loop counts 12:53:44 I would argue that something like " : .table 255 for i calc-val . next cr ; " would make less sense if you factored the i and . into calc-val 12:54:11 you cant factor I out. its a primitive 12:54:21 no reason to factor it out 12:54:49 you can factor it out but it would be brittle 12:55:21 i takes the top two items off the return stack and adds them and puts the result on the parameter stack 12:55:33 you cant do 12:55:41 : do-i i ; 12:55:51 or it fetches the top element of the rstack or it fetches the top element of the loop stack, depending on your forth 12:56:00 because do-i has a return address burying the for loop parameters 12:56:20 I fetches TWO items off the return stack and adds them togehter 12:56:27 it doesn't have to 12:56:36 if your implementation of DO and I dont work this way i think they are probably broken 12:56:37 impomatic: i figured out why dothis10times isn't working for me....begin is putting it's return address on the same stack at compile time 12:56:48 impomatic: i could change that around 12:57:05 RodgerTheGreat, can you do : fud 0 10 do i . -2 +loop ; 12:57:07 i just needs to push the current loop index; that is all 12:57:11 and get the expected results? 12:57:12 impomatic: or [ swap ] 12:57:22 i.e. with your implementation of do/+loop etc 12:57:24 protist: My version above might work for you then? 12:58:25 proteusguy, BEGIN does not do anything except push HERE onto the stack. the repeat/again/until words use the item at the top of the stack to say where to loop back to 12:58:31 : dothis10times 0 begin >r [ ' this ] literal execute r> 1+ dup 10 = until drop ; <- I removed a couple of return stack instructions, might work for you now? 12:58:36 : begin here ; 12:58:37 impomatic: the [ ' this ] literal execute one? 12:58:41 is a good definition for begin 12:58:42 impomatic: probably :) 12:58:47 impomatic: i will try it 12:59:11 : again ( address --- ) compile (again) , ; immediate 12:59:32 or rather compile branch , ; 12:59:44 impomatic: yes, that works :D 13:00:01 protist, you need to learn how >mark >resolve I440r: i wanted to use a stack so i can nest conditionals and loops 13:02:43 protist: you already have one y'know 13:02:55 RodgerTheGreat: i know :) 13:03:41 phone 13:08:54 --- join: kumul (~Mo@173.215.194.228) joined #forth 13:08:54 --- mode: ChanServ set +v kumul 13:16:39 --- join: Onionnion (~ryan@adsl-68-254-169-75.dsl.milwwi.ameritech.net) joined #forth 13:16:39 --- mode: ChanServ set +v Onionnion 13:26:00 --- quit: bjorkintosh (Ping timeout: 252 seconds) 13:32:11 just had a phone interview. dont know if ill get the position because he has one more to interview 13:32:30 what's the position? 13:32:45 contract position in MO, they do medical gizmos 13:52:47 --- join: ASau (~user@46.115.98.84) joined #forth 13:52:47 --- mode: ChanServ set +v ASau 14:05:13 : constant get_lexeme dp @ dup >r write dp ! r> ep @ ! ep @ 4 + ep ! 3 dp @ ! dp @ 4 + ! dp @ ep @ ! dp @ 8 + dp ! ep @ 4 + ep ! ; 14:05:17 worked first try :D 14:06:16 protist: what does it do 14:06:47 RodgerTheGreat: defines constant such that i can say 3 constant three .....and now `three' will put 3 on the stack 14:07:11 RodgerTheGreat: my definition for variable will be similar, only it will return the address instead of the value 14:07:45 I'm thinking that maybe you could factor that a bit more 14:08:03 RodgerTheGreat: arrows keys don't work in my interpreter lol 14:08:14 RodgerTheGreat: i will factor it later :) 14:08:23 RodgerTheGreat: for now i am testing things 14:09:22 i will define +! and things will be smoother haha 14:09:57 4 constant cell 14:10:12 : cells cell * ; 14:10:25 protist, arrow keys in linux dont return a character. they return an escape sequence 14:10:35 the exact sequence depends on which terminal 14:10:53 brb, gotta start packing. will dc 14:10:58 kk cya :) 14:13:16 --- quit: MayDaniel (Read error: Connection reset by peer) 14:14:57 protist: what is "3 dp @ !" doing? 14:15:19 --- quit: I440r (Ping timeout: 240 seconds) 14:18:08 it really is a lot easier to write and understand forth code if you factor as you go rather than making tangles and then trying to discover repeated patterns 14:19:02 this is almost impossible to follow without knowing about the precise implementation of your dictionary 14:19:22 you should write words for manipulating the dictionary more abstractly 14:23:22 RodgerTheGreat: 3 dp @ ! is compiling the `constant' data type 14:24:09 RodgerTheGreat: yeah i know.....3 at the beginning of a definition tells execute to push the number in the next data location 14:24:52 wait so why does execute care about the implementation of constants? 14:25:15 execute is recursive and goes down until it finds a primitive or a constant 14:25:53 my variables will be implemented as constants that push an address value 14:25:59 what type of threading does this forth use? 14:26:11 i have no idea 14:26:18 i tried not to look at other implementations haha 14:26:23 hrm 14:26:47 maybe someone here should make a colorforth for unix 14:26:59 instead of reimplementing yet another forth :) 14:27:23 kumul: my code is all x86 gas assembly 14:27:29 kumul: 32 bit 14:27:45 kumul: oh you want colorforth haha 14:28:06 it doesnt seem that difficult though 14:28:17 kumul: I don't really trust languages that need a new text editor 14:28:23 it doesnt 14:28:27 this is part of why I don't use smalltalk 14:28:31 its forth with a tag 14:29:10 and considering terminal have color codes, thats just half the battle done 14:29:14 but im with you 14:29:44 oh, and there is a version that uses fonts instead of colors, for the colorblind 14:30:00 colorforth replaces punctuation and some of the state of the compiler with colors 14:30:18 so you must by definition enter code using the colors 14:30:31 displaying colorforth on a terminal is easy, yes 14:30:44 how many colors are there? 14:31:04 I think 6? 14:31:17 let's see if i can remember 14:31:19 really? i've only seen 2 14:31:22 red and green 14:31:25 red for colon defs, 14:31:29 green for compiled code 14:31:33 yellow for immediate code 14:31:40 grey for comments 14:31:57 protist: this works for Itsy : variable create 0 , ; 14:32:09 and then I think there are colors which change the base interpretation of numbers 14:32:19 --- quit: kylert__ (Ping timeout: 240 seconds) 14:32:41 that really simplifies the compiler you know... doesnt seem bad 14:33:04 it does simplify the compiler slightly but the editor has to do more 14:33:29 I'm unconvinced that losing the ability to use a lot of existing tools is worth it 14:33:55 it already highlights syntax, what more does it have to do? 14:34:24 if you're just coding on metal in a block editor that only runs on the pentium 2 you've had for years then that's fine, but what if you want to store your source in files and use a version control system? 14:34:55 or grep through your source files 14:35:11 or use almost any unix utility to help manage or inspect your code 14:35:59 impomatic: i dont have a create word yet 14:36:05 impomatic: or `,' 14:36:25 impomatic: that is very clean, though :) 14:36:28 it makes sense if you don't care at all about the underlying OS and don't intend to use it for anything, but if you want to interact with unix it's a waste 14:36:38 again, think about smalltalk 14:37:20 each word has a tag(byte) at start, so { grep $colorX^word file } would work i guess 14:38:17 if you give colorforth a text-based serialization format that adds the colors back in as text you haven't simplified the compiler anymore- arguably you have ADDED complexity 14:38:48 for all that fucking around you've only saved two variables worth of state 14:48:35 --- join: kylert__ (~kroot@c-65-96-213-133.hsd1.ma.comcast.net) joined #forth 14:48:35 --- mode: ChanServ set +v kylert__ 15:02:01 --- quit: kylert__ (Quit: leaving) 15:26:48 --- join: I440r (~zhiming@167.sub-70-194-68.myvzw.com) joined #forth 15:26:48 --- mode: ChanServ set +o I440r 15:27:12 --- quit: I440r (Client Quit) 15:30:39 --- join: I440r (~zhiming@167.sub-70-194-68.myvzw.com) joined #forth 15:30:39 --- mode: ChanServ set +o I440r 15:37:12 --- quit: JDat () 15:40:02 --- quit: ncv (Remote host closed the connection) 15:42:02 --- join: kumool (~Mo@173.215.194.228) joined #forth 15:42:02 --- mode: ChanServ set +v kumool 15:42:05 --- quit: kumul (Ping timeout: 246 seconds) 15:47:12 --- quit: kumool (Quit: Leaving) 16:38:12 --- join: kumul (~Mo@173.215.194.228) joined #forth 16:38:13 --- mode: ChanServ set +v kumul 17:02:52 --- quit: epicmonkey (Ping timeout: 272 seconds) 17:26:30 what is it that forth calls a size of two cells? 18:03:06 --- join: bjorkintosh (~bjork@ip68-13-229-200.ok.ok.cox.net) joined #forth 18:03:06 --- mode: ChanServ set +v bjorkintosh 18:09:05 --- quit: john_metcalf (Ping timeout: 246 seconds) 18:09:28 --- quit: impomatic (Ping timeout: 264 seconds) 18:10:31 --- join: john_metcalf (~john_metc@211.67.125.91.dyn.plus.net) joined #forth 18:10:31 --- mode: ChanServ set +v john_metcalf 18:17:51 double? 18:18:18 D is for double no? 18:21:13 ah possibly! 18:21:14 i 18:21:18 i'll test it 18:22:14 meh 18:22:33 i want something that works like `cell' ...only to times that 18:22:40 but i want to use the standard word 18:37:03 '2 cells' ? 18:38:27 ah 18:38:50 thank you :) 18:39:01 in GFORTH, is there any way to list a subset of words? I've tried "xxxx words" and "words xxxx", but neither seems to make any difference. 18:43:48 protist, use the ans standard, there's a nice zip available 18:44:10 jdavidboyd: you mean, as if you were using a pager? 18:46:05 it seems like he wants all words to have some sort of tag 18:46:22 darn you RodgerTheGreat 18:47:00 i still think ansi escape codes arent that big a deal 19:03:59 dzho: no, I mean like a filter of the words. so 'xxx words' would only show words that have the text string 'xxx' in them. 19:04:32 dzho: I don't want to have to see ALL the words ALL the time when I am looking for something 19:08:09 kumul: ill look into that :) 19:13:06 kumul: what 19:13:47 RodgerTheGreat, exactly 19:14:01 ... 19:14:10 jdavidboyd: that doesn't sound to hard to make 19:15:48 RodgerTheGreat: i now have variable as : variable create 0 , ; like impomatic 19:15:57 RodgerTheGreat: it made a lot of sense after i defined create 19:16:08 that's good 19:16:12 what does create look like? 19:16:16 sec 19:16:33 : create get_lexeme here dup ep, write dp ! here ep, con_type , here cell + , ; 19:16:57 that's much better 19:17:00 where : ep, ep @ ! 4 ep +! ; 19:17:04 :D 19:17:10 MUCH better 19:17:23 yeah im factoring it down :) 19:17:37 i shall...i was having fun lost in references, though ;) 19:17:44 thank you 19:17:50 I suspected there was a , like factor that could be extracted from your original version but I didn't understand what you were doing well enough to find it 19:18:13 ep, is a good factor 19:18:22 ep is my entry pointer 19:18:26 although I would recommend replacing the 4 with "cell" 19:18:29