00:00:00 --- log: started forth/03.12.29 01:39:40 --- join: Serg (~z@212.34.52.140) joined #forth 01:39:58 --- join: schihei (~schihei@p5085D89E.dip.t-dialin.net) joined #forth 01:48:47 --- join: aktnot_ (~ident@233.80-202-65.nextgentel.com) joined #forth 01:58:59 --- quit: aktnot (Read error: 110 (Connection timed out)) 02:18:36 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 03:31:26 --- join: aktnot (~ident@233.80-202-65.nextgentel.com) joined #forth 03:48:33 --- quit: aktnot_ (Read error: 110 (Connection timed out)) 03:59:33 --- quit: OrngeTide (Remote closed the connection) 04:01:33 * warpzero is back (gone 09:26:06) 04:01:49 --- part: MrReach left #forth 04:07:43 --- quit: aktnot (Read error: 110 (Connection timed out)) 04:07:50 --- join: aktnot (~ident@233.80-202-65.nextgentel.com) joined #forth 04:14:53 --- quit: Serg () 04:21:00 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 04:57:29 --- join: crc (~crc@ACAFFECF.ipt.aol.com) joined #forth 05:05:16 --- quit: hovil ("Leaving") 06:09:49 --- join: fridge (~fridge@dsl-203-33-163-107.NSW.netspace.net.au) joined #forth 06:22:41 --- quit: chrisrw (Remote closed the connection) 06:27:51 --- quit: crc ("I was using TinyIRC! Visit http://www.tinyirc.net/ for more information.") 06:34:55 --- join: chrisrw (~chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 06:38:58 --- quit: tathi ("leaving") 06:57:56 how do i test 3 values to be either all equal, or all different, with just the stack and without roll/pick... 06:57:57 :) 07:00:26 i now made variables for them... which makes it quite easy 07:00:27 : match-numbers ( -- flag ) \ see if n1 and n2 and n3 are all equal or all different 07:00:27 num1 @ num2 @ = if num2 @ num3 @ = if -1 -1 else 0 then else 0 then \ if (num1 == num2 == num3) { -1 -1 } else { 0 } 07:00:27 inv if \ if !(num1 == num2 == num3) {} 07:00:27 num1 @ num2 @ <> if num2 @ num3 @ <> if num1 @ num3 @ <> if -1 else 0 then else 0 then else 0 then \ if (num1 != num2 != num3 != num1) { -1 } else { 0 } 07:00:29 then 07:00:31 ; 07:08:45 --- quit: aktnot (Remote closed the connection) 07:10:05 --- join: aktnot (ident@233.80-202-65.nextgentel.com) joined #forth 07:20:32 --- quit: fridge (Remote closed the connection) 07:24:02 why is it illegal to use >r if the stack isnt empty? 07:24:32 "1 >r" results in access violation crash.. 07:28:29 is >r causing the crash, or are you forgetting to pop it off before return? 07:28:45 its >r 07:28:51 i made a word to do 3dup 07:29:10 i double checked it and the return stack should be the same as when the word started 07:30:05 "1 >r r> ." does work indeed 07:30:33 ok... which forth? gforth? 07:30:36 winforth 07:31:14 http://www.rafb.net/paste/results/O2722190.html 07:33:38 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 07:40:42 --- quit: schihei (Client Quit) 08:21:35 --- join: aktnot_ (ident@233.80-202-65.nextgentel.com) joined #forth 08:27:30 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 08:30:08 --- quit: aktnot (Read error: 60 (Operation timed out)) 08:33:13 how about XOR OVER XOR = 08:35:29 nice :) 08:35:33 for comparing 3 numbers. 08:35:48 yeah, I just checked the logs 08:42:26 nah, no good. 08:42:48 if the second two numbers are the same you get a positive result, even if the first one is different 08:43:26 nxor is = 08:43:30 Yea, hmm. 08:45:31 more fun with SET :) 08:46:24 ^^ 08:46:35 any idea why this crashes? http://www.rafb.net/paste/results/O2722190.html 08:47:37 num1 @ num2 @ = if num2 @ num3 @ = if 1 else 0 then else 0 then 08:47:41 for comparing btw 08:51:45 qFox: it crashes because you left a bunch of crap on the rstack 08:51:49 : 3dup 2dup >r >r rot dup >r rot rot r> r> r> ; 08:52:16 qFox: the last to >R should be R> ? 08:52:19 i did? as far as i can see i'm popping as many items as i'm pushing 08:52:30 yah, your last _two_ >r should be r> 08:52:38 qFox: you're pushing 6 and poping 2 08:52:41 oops 08:52:52 i have a form of dyslexia 08:53:25 i double checked it, counted it, and didnt see it 08:54:13 tathi> i tried to avoid pick roll and rot :) otherwise i'd just do : 3dup 2 pick 2 pick 2 pick ; 08:54:34 well, if you're working with three stack items you might as well use rot 08:55:26 the point of not using rot/pick/roll is to avoid using more than 2 stack items (and when you do have them in order already) 08:55:36 oh 08:55:43 ok i missed that bit :) 08:55:57 but hey i did it without anyways 08:56:00 :) 08:56:05 :) 08:56:31 i dont suppose you can swap the return stack? :) 08:56:37 I get the impression most people avoid pick/roll, but that only the hardcore minimalist people really avoid rot. 08:56:59 I still think it can be done with XORs. :) 08:57:19 well i'm actually on the edge of having an attitude that says screw you all, there's pick and roll so i'll use them if it makes it easier... :) 08:57:27 madgarden: yeah, I'm still trying to figure out a way to do that. 08:57:36 qFox: it's your program, go ahead :) 08:57:44 I almost had it working, but then the case of 1 2 1 failed. 08:57:52 xor to see either x = y = z ? 09:00:48 PASSES MY DRIVING TEST!!! W00T!!! 09:00:55 I sorta think that XOR would only work with base 2 or 4 09:02:14 chrisrw: congrats :) 09:02:27 XOR works easily for even argument count. 09:02:40 I failed my driver's test the first time. 09:02:47 :) 09:02:55 first two for me 09:03:25 tuck xor rot xor = 09:03:26 but from what I can tell, PA almost automatically fails boys the first two times ;) 09:08:02 qFox: Fails for the 2 2 1 case. 09:08:06 ya 09:08:11 xxy 09:08:16 aab 09:08:25 this this that 09:08:26 :P 09:09:03 --- quit: chrisrw (Remote closed the connection) 09:09:18 i dont think you can do it with just xor 09:09:35 --- join: chrisrw (~chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 09:09:41 so i'll agree with Herkamire :) 09:11:47 i made it i made it omg omg omg omg 09:13:29 gtg, more set fun tomorrow :p 09:13:56 set? 09:13:56 laters then 09:14:26 the card game SET, he's writing one in Forth 09:14:54 Oh, neaters. ciao qFox. 09:15:00 cu :) 09:30:00 --- quit: chrisrw (Read error: 110 (Connection timed out)) 09:34:28 grr. sometimes I really dislike make. 09:41:34 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth 09:41:34 --- mode: ChanServ set +o kc5tja 09:42:08 * kc5tja has officially met his stated objective, and ahead of schedule too. 09:42:19 FS/Forth can both print text to the Linux console and receive text input from it. 09:43:23 My next milestone is substantially more ambitious unfortunately, and by necessity. 09:44:01 I now need to implement the source code interpreter, which means I must include the Forth wordlist in the ELF executable. 09:44:26 that's always fun :( 09:45:06 Yeah, it'll be a learning experience, for sure. 09:45:07 ahhh...make removes implicit intermediate targets. *that's* what's going on. 09:45:21 At least the cross-compiler can be used to make real Linux applications now, though. :) 09:48:40 * kc5tja made a simple interactive Hello World program that compiled to a grand total of 921 bytes, including ELF header overhead. 09:48:40 : buffer r> ; 09:48:40 : buf buffer [ 80 allot 09:48:40 : name cr ." Please enter your name: " buf 80 accept 1- buf swap ; 09:48:49 : hello ." Hello, " type ." !" cr ; 09:48:59 : main [entry-point] name hello bye ; 09:52:30 --- join: chrisrw (~chris@wbar8.lax1-4-11-099-104.dsl-verizon.net) joined #forth 09:52:45 kc5tja: guess what!? 09:53:01 What? 09:53:06 i got my license. 09:53:10 :D :D :D :D 09:53:20 Drivers license? 09:53:24 yuupy 09:53:29 Cool. 09:53:36 yep :D 09:53:36 :D 09:53:37 :D 09:53:44 * kc5tja met his stated objective this morning. 09:53:53 i can now drive myself as well as people 18+ 09:53:56 (or was it 21+?) 09:54:00 ooh 09:54:02 what was that? 09:54:25 FS/Forth's cross compiler produced an application which was an interactive hello-world program today. 09:54:34 CONGRATULATIONS!! 09:54:45 The total binary size was 921, including ELF header overhead. 09:54:50 like, "whats your name?" "Hello, %s" ?? 09:54:57 Yes, exactly that. 09:55:02 :)( 09:55:03 er 09:55:05 :) 09:55:06 awesome 09:55:11 I already posted the 5-line source here; check the logs to see it. 09:55:18 where are the logs? 09:55:32 http://tunes.org/~nef/logs/forth/ 09:56:47 : buf buffer [ 80 allot <--- ? 09:56:50 Yep. 09:57:04 i mean, i know what all of those do, but !!!??!??!?!?! 09:57:07 My Forth does not have CREATE. 09:57:15 oh 09:57:16 : buffer r> ; 09:57:21 : buf buffer [ 80 allot 09:57:32 does not have create at all? 09:57:36 or just for now? 09:57:44 When buf calls buffer, the return address points to the 80-byte buffer. 09:57:45 At all. 09:57:49 !!! 09:57:51 NO CREATE!!? 09:57:55 WTF!? 09:58:05 It's not needed. 09:58:26 At least, for the things I'll be using FS/Forth for. 09:59:22 ... 09:59:24 dude 09:59:38 at least implement it later on, even if you won't use it. 09:59:53 I have no immediate plans to implement CREATE. 09:59:57 I just don't need it. 10:00:21 a forth without CREATE or DOES> .... 10:00:23 man 10:00:31 that is just .... 10:00:32 ugh 10:00:36 : buf buffer produces exactly the same binary code that create buf would produce, and I don't even have to implement the infrastructure for it. 10:00:50 chrisrw: You are aware taht ColorForth also lacks CREATE and DOES>. 10:00:58 create buf is a bit more readable... ;) 10:01:05 No, not when you think about it. 10:01:06 create buf looks prettier though 10:01:11 piso: bastard, you beat me :P 10:01:12 : makes a new definition. 10:01:21 yep. 10:01:24 I'm defining buf to be a buffer, which is 80 bytes long. 10:01:44 : buf buffer [ 80 allot 10:02:03 theres no ;, I know its not needed, but it just doesnt look right without ; 10:02:08 are you deliberately omitting the ';' at the end of that line? 10:02:14 Very. 10:02:37 Note what buffer is doing: it's popping the return address off the return stack, because that address points to the buffer. 10:02:47 Hence, buffer never 'returns' to the buffer. No need for ;. 10:03:22 so its immediate without being immediate 10:03:23 lol 10:03:27 so why not just : buf r> [ 80 allot 10:03:30 What's immediate? 10:03:33 Nothing here is immediate. 10:03:46 [ <--- makes it immediate 10:03:53 piso: Because r> removes the return address, and then will attempt to execute the contents of the buffer. 10:03:57 chrisrw: No it doesn't. 10:03:58 actually, it goes out of xompile mode. 10:04:01 It just stops the compiler. 10:04:04 yes. 10:04:06 but 10:04:10 the word is immediate in a sense 10:04:14 No. 10:04:18 The word definitely is not immediate. 10:04:24 if you call buf, then it does an immediate action 10:04:27 If I execute buf, it has absolutely zero compile-time semantics. 10:04:34 chrisrw: No, it doesn't. 10:04:39 yes it does. 10:04:41 [ <--- 10:04:43 No, it doesn't. 10:04:50 [ does not make a word immediate. 10:04:55 yes 10:04:57 No 10:05:04 in this case [ is just ending the definition 10:05:05 but anything after a [ is immediate 10:05:07 I don't know where you got that idea from, but it's patently false. 10:05:14 chrisrw: No!! 10:05:32 ... 10:05:36 By that logic, every time I execute 'buf', it'd allocate a new 80 byte buffer in the dictionary. That is NOT what's happening. 10:06:12 kc5tja is right, but I'm not sure it's worth the obscurity 10:06:19 There is no obscurity. 10:06:34 FS/Forth is a native code compiler. 10:06:40 It produces raw x86 machine code. 10:06:52 My code fragment above produces *exactly* the same code that CREATE and DOES> would produce. 10:07:04 But my code actually exposes the operation, not hides it. 10:07:19 yes, but one can't know that without being familiar with your particular implementation 10:07:20 yes 10:07:21 And I don't even need to implement the core logic for CREATE or for DOES>. 10:07:34 however, that would mean they wopuld have to know how it works. 10:07:39 piso: Dude, it even works in direct threaded ANSI Forth. 10:07:41 and forth is all about factoring anyway. 10:07:51 CREATE and DOES> is much more factored than that. 10:07:56 Forth is all about knowing hwo things work. 10:08:29 Factoring is a tool to make that happen, but if you look on comp.lang.forth, there's isn't a month that goes by when someone isn't thoroughly and utterly confounded by how DOES> works. 10:08:38 This makes the whole thing very, very, very explicit. 10:08:51 : array cells r> + ; 10:08:59 : myArray array [ 80 cells allot 10:09:05 : yourArray array [ 160 cells allot 10:09:11 5 myArray @ . 10:09:16 15 yourArray @ . 10:09:41 does> is so easy to get once you do get it. 10:09:49 the thing is that nobody explains it well. 10:10:11 the way i finally understood was when you explained, and even then i was utterly confused 10:10:22 Precisely. 10:10:25 DOES> is a mind-trap. 10:10:26 (at first) 10:10:33 however 10:10:38 It is not easy to understand, and is inherently considered "Advanced Forth." 10:10:41 nobody explains it the way it should. 10:10:47 And nobody will ever. 10:10:47 kc5tja: as a point of information, your code doesn't work under gforth (for example) 10:10:55 well, it is easy to understand if it is explained right. 10:11:09 piso: Well, because : and ; have to be balanced due to enforcing the control stack. 10:11:23 chrisrw: Dude, even Leo Brody couldn't explain DOES> easily. 10:11:29 And he's the author of Starting Forth. 10:11:42 well, you have exceptional talent in explaining forth 10:11:52 because i understood your version, and it took me a while to get his. 10:12:44 i should search the logs for your explanations, and make a book out of it 10:12:45 I find my solution to be much easier to understand. 10:12:46 and make you rich 10:12:47 :) 10:12:59 Even if it does look awkward. 10:12:59 kc5tja: well, yes and no. 10:13:04 so in any case, I don't think your code is guaranteed to work in direct-threaded ANSI Forth 10:13:14 kc5tja: : ..... [ .... <-- is very VERY awkward code 10:13:15 piso: It works fine in Pygmy. 10:13:29 I think it will work sometimes, sometimes not. 10:13:30 yes, but its implementation specific 10:13:30 chrisrw: Hardly!! Chuck Moore does this all the time in his code. 10:13:35 Consider the evaluation fo constants: 10:13:38 : foo [ ... ] literal ; 10:13:41 I mean, what the heck? 10:13:47 This isn't rocket science. :) 10:13:57 :) 10:14:05 : constant create , does> @ ; 10:14:14 easy to understand, produces good code. 10:14:20 chrisrw: No, hardly the same thing. HARDLY the same thing. 10:14:24 Produces HORRIBLE code. 10:14:37 Because the evaluation of a constant involves no less than TWO subroutine calls. 10:14:46 The code it produces depends on the optimizations performed by the compiler. 10:14:56 why no less? 10:14:58 Meanwhile, the use of LITERAL involves at most one, and easily inlined by an optimizing compiler to produce zero subroutine calls. 10:15:00 I see 1 call. 10:15:15 well, actually, you're right, 2. 10:15:25 but for constant, you can do other things 10:15:27 You see 1 call because you're not familiar with how DOES> is implemented. :) 10:15:34 : constant literal ; immediate 10:15:41 No. 10:15:44 That won't work. 10:15:45 well, theres the call to @ :) 10:15:57 well, something along the lines of that. I'd have to mess with it. 10:16:03 DOES> has to be implemented as a double-branch in order to get the body address for the word. 10:16:04 lemme start up pygmy really quick. 10:16:17 er!? 10:16:27 Look at how Pygmy implements DOES>. 10:16:38 You'll find it's not as "clean and efficient" as you make it out to be. 10:17:13 Anyway, I need to get to work. 10:18:05 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 10:18:19 --- join: Robert (~snofs@c-7e5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 10:20:55 hi Robert 10:21:06 Hi 10:22:52 where were you the last few days? 10:23:08 No internet.. 10:23:12 * Robert hates his ISP 10:23:21 ugh 10:23:25 even with the new ISP? 10:23:28 that sucks. 10:23:30 poor Robert 10:23:32 :) 10:23:38 * chrisrw got his drivers license today 10:24:09 Congratulatoins 10:24:13 -tions* 10:24:54 thanks 10:24:56 hehe 10:24:59 im driving to work today 10:25:00 yay 10:28:00 FWIW we had somebody in here just yesterday who was very confused about CREATE/DOES> 10:28:17 This time it wasn't me. :) 10:28:40 heh 10:28:48 CREATE DOES> is such an easy concept. 10:28:56 its just blown out of proportion. 10:29:01 thats why n00bs get confused' 10:29:17 chrisrw: why the hell do you keep saying that? you admitted yourself that after it being explained to you very well TWICE you were still confused 10:29:40 because its blown out of proportion, and thats why i was confused 10:29:48 if it confuses everybody, then it's obviously not an easy concept 10:30:05 --- join: snowrichard (~richard@adsl-068-209-159-248.sip.shv.bellsouth.net) joined #forth 10:30:14 took me years to figure out how it worked, a few minutes to figure out how to use it 10:30:15 well, if you explain it like "Sets the compile-time semantics of the word being created" few people are gonna get it. 10:30:52 that chapter in the brodie forth book was very deep 10:30:56 if you say that "the DOES> part is what happens when you use the new word in your code" its much easier to grasp 10:30:57 confusing 10:31:17 chrisrw: that's a little better sorta... 10:31:22 that makes sense somewhat 10:31:31 problem is it's not even that simple, because CREATE also sticks some code into your definition 10:31:45 DOES> is a bastard to implement too 10:31:45 yes, but no. 10:31:46 lol 10:32:01 I don't have those yet in my system 10:32:17 "DOES> is provided (as the top of stack) the item last allocated in the stack in the CREATE part. 10:32:23 I like kc5tja's trick of poping the rstack in ARRAY 10:32:32 I implemented does> on my system, but I've never used it yet 10:32:40 Herkamire: yes, that is neat, but ugly still :) 10:32:41 Robert: :) 10:32:45 Maybe once twice.. But not more. 10:33:02 In many pages of Forth code. 10:33:02 chrisrw: it's not ugly 10:33:18 : buf buffer [ 80 allot 10:33:26 it's perfectly strightforeward and clear 10:33:29 yes, that is not only ugly but also implementation specific 10:33:39 to anybody who understands forth reasonably well. 10:33:49 in michael> 80 VAR buffer 10:33:51 itll work on pygmy and FS/Forth, but not on lets say gforth 10:34:19 : array create allot , does> @ ; 10:34:21 I don't really care about compatibility with different forths, and I doubt kc5tja does 10:34:30 well, he probably doesn't. 10:34:38 but basic concepts are alike in all forths 10:34:38 he's writing his own forth. if he just wanted to be compatible, he'd just use an ANS forth and be done with it. 10:34:48 such as CREATE, DOES>, DUP, SWAP, OVER, etc. 10:34:57 ANS forth sucks tho ... 10:35:01 mine is not compatible and not intended to be 10:35:17 once i get around to mine, it wont be. 10:35:17 many forths do not have DOES> 10:35:25 and a bunch don't have CREATE either 10:35:26 but it will be very much like many others. 10:35:31 Herkamire: name some. 10:35:35 (like say _Chuck Moore's_ forth) 10:35:40 mine 10:35:44 kc5tja's 10:35:47 cmforth doesn't!? 10:36:10 NO 10:36:14 kc said that earlier 10:36:26 let me read my code and see if I can figure out a cleaner way to do it. bbl 10:36:26 DOES> is an ugly hack 10:36:44 I think this ": array create allot , does> @ ;" is ugly 10:37:07 I can't even see that it would work 10:37:28 --- quit: snowrichard ("Leaving") 10:37:40 what's with the comma? 10:38:01 hmmmm. I think it's terribly broken 10:38:24 you mean like this right? : array create allot does> swap cells + @ ; 10:39:06 vs kc's : array cells r> + @ ; 10:40:12 I like: : my-array array [ 80 allot 10:40:22 it doesn't introduce any new syntax 10:41:13 the only thing funny looking is if you just look at the def for my-array you don't see the end of the definition, but if you look into array, you will see that it ends the def 10:42:47 qFox: you still around? I had a fun idea for matching cards in your SET game: 10:43:20 this word checks two attributes (from different cards) and returns what the third should be 10:44:25 : check2 ( card1-attr card2-attr -- correct-card3-attr ) 2dup = if drop exit then 3 swap - swap - ; 10:46:49 then you can do: : check-colors card1 get-color card2 get-color check2 card3 get-color = if ." colors good" exit then ... ; 10:49:18 everybody: what should I name a word which subtracts the stack item under TOS from TOS ( a b -- b-a ) 10:50:21 PPC instruction is subf 10:50:26 (subtract from) 10:50:29 swap- 10:51:33 piso: yeah :) I might just do that 10:51:38 you know what I just thought 10:51:55 rlwinm completely eliminates any shifting, rotating, or ANDing :) 10:52:25 unless the shift/rotate amount is not known until runtime 10:52:37 oh 10:52:38 hrm 10:52:39 :( 10:52:56 well, then make a all-register rlwinm :) 10:53:01 rlwm :P 10:53:13 5 registers :) 10:53:34 there is one that is less constant let me see... 10:55:28 there's rlwnm which uses a register for the rotate. but the mask is still constant 10:58:15 wow, there's also rlwinmi (rotate register left by immediate then insert) 10:59:04 itstead of just sticking the masked bits into the destination register, it just sets those bits in the destination register (leaving the other bits in the destination register (not hit by the mask) alone) 11:01:04 odd. 11:02:24 err, no. 11:02:34 --- join: OrngeTide (orange@rm-f.net) joined #forth 11:03:32 rlwimi takes a bitfield from one register and inserts it into another register. 11:04:35 oh, that's what you were trying to say, never mind. 11:04:44 I was trying :) 11:04:55 im reading logs from a few months back 11:04:57 man 11:04:59 anyone knows how i can remove the frames in Deluxepaint IV (amiga), what keycombination or mouse path? 11:05:00 was i stupid 11:05:01 lol 11:05:30 --- quit: Hipo (Read error: 104 (Connection reset by peer)) 11:05:30 chrisrw: live in the now! ;) 11:05:36 : 11:05:37 :) 11:05:59 lunchtime. bbiab 11:06:09 bye 11:06:34 --- join: rO| (rg@pD95459E6.dip.t-dialin.net) joined #forth 11:11:00 Moo :) 11:11:47 moo yourself 11:12:51 ... 11:12:52 Moo. 11:12:57 Cheese. 11:13:20 bah. go do something useful 11:13:37 yeah 11:13:48 I am. 11:13:52 I'm sowing up my pants. 11:13:53 look up with google how i can remove tool bars in Deluxe paint IV on amiga 11:14:01 because there's a big hole where my crotch is. 11:14:09 mur: I got a link for you. 11:14:33 mur: http://www.google.com/search?q=remove+toolbars+Deluxe+paint+VI 11:15:01 chrisrw: what does that page say, i dont trust urls on irc ;) 11:15:21 chrisrw: learn to spell 11:15:22