00:00:00 --- log: started forth/03.04.04 00:06:46 --- join: mur (jukka@baana-62-165-187-95.phnet.fi) joined #forth 00:07:42 HEllo. 00:31:47 --- join: gilbertdeb (~gilbert@fl-nked-ubr2-c3a-29.dad.adelphia.net) joined #forth 00:31:53 terve! :) 00:31:58 terve! 00:32:10 hyvaa something or other? 00:32:21 * mur woudlnt be awake by now unless he had been so tired he had fallen asleep right away when he came home 00:32:29 hyvää päivää = good morning 00:32:35 hyvää huomenta = mornting 00:32:39 päivää = day 00:32:44 iltaa = evening 00:32:59 yötä = night, good night sleep well 00:33:09 all in objective 00:33:24 that i dont remember the real name for the word class 00:33:47 I should be sleeping. 00:34:30 * mur went sleep last day about 6 pm to 7 pm 00:34:36 woke up today 9 am 00:34:46 and felt asleep and woke up again 10 am 00:35:01 i slept monday-tuesday only 1 hour or less 00:36:11 why? 00:38:03 well first there was a bit hassle (some ship drivers had drunken a lot) 00:38:13 and i coudl not sleep because my bed was all wet 00:38:23 becasue they have tried the fire thing a bit 00:38:34 then i felt asleep on time on the bed 00:38:42 and then i had duty starting 2.30 to 6.00 00:38:47 and then i could nto sleep anymore 00:38:53 becsuae we had work to do 00:39:08 --- quit: Robert (brunner.freenode.net irc.freenode.net) 00:39:17 but now i'm in home :) 00:39:45 --- join: Robert (~snofs@h138n2fls31o965.telia.com) joined #forth 00:39:55 * mur was placed reserve (i was in conscription service, no volunteer at all!) special man (graphician) in naval hq 00:40:15 has finland ever been attacked? 00:41:35 whats mitakuluu? 00:45:35 gilbertdeb: ! 00:45:40 gilbertdeb: Go read some history. 00:45:51 gilbertdeb: The Soviet Union attacked them in 1939. 00:46:02 oh. 00:46:06 gilbertdeb: And then they attacked the Soviet Union in 1941 >:) 00:46:09 And they've been reacting since then? 00:46:13 They lost both times btw :P 00:46:19 I am quite sure. 00:46:37 Although they did quite a bit of resistance. 00:46:43 thats healthy. 00:47:28 ROOOBERY! :) 00:47:33 Yes, since more people get killed. 00:47:36 gilbertdeb, several times 00:47:39 That's..uh..healthy. 00:47:45 mur :)) 00:48:19 gilbertdeb, we coudl have get soviet union capital even, but we didnt' want to do it because they had nice forces that woudlhave done bad things after that 00:48:20 * Robert just slept 12 hours.. mmm. 00:48:25 * mur too :) 00:48:26 Well, I'm going to have a shower, brb 00:49:05 gilbertdeb, already at viking times finnish kings were commonly talked in brittish pubs accordingly to viking's stones 00:49:12 rhyme stones? 00:49:27 giu mitä kuuluu? = what's up? 00:49:34 giu= gilbertdeb 00:52:54 gilbertdeb, as a note, when we fought for our independence, we had 10 times more men on the opponents side and modern equivment. still we wont 00:52:56 won 00:53:27 i mean there were 10 times more men on the russians side 00:53:32 attacking us 00:53:44 and they had modern equivement 00:53:51 (of the time) 00:54:06 independence from who? 00:54:07 norway? 00:54:13 sweden? 00:54:17 or russia? 00:54:30 russia 00:56:13 before independence we had almost independent society with own currency and moneytary system, full legistimated governmental system (which gave all the laws) and own post system, police and eductaion suystem to mention 00:56:59 and finnish was independent language, btw 00:57:01 no 00:57:05 official not independent 00:57:12 well independent too 00:57:16 blarg :) 00:57:50 finland was long time the most advanced part of russia 00:58:01 very beyond other parts 01:10:33 --- quit: gilbertdeb ("Monk has left the building") 01:36:47 Hah. 01:36:54 He didn't like finnish history :) 01:43:36 me neither 01:48:48 :D 01:48:54 That's why you always tell us about it? 01:49:23 i like to bore you to death >:) 01:51:41 :~( 01:52:13 * mur licks Robert's tears 01:52:29 dont be sad, be scared 01:52:32 * mur stops 01:52:43 execution "evil.scd" done 01:52:58 ScriptDaemon> _ 02:33:39 Robert, hello 02:38:59 --- quit: deluxe (Read error: 104 (Connection reset by peer)) 03:07:06 real or fiction, the bible sais magic is bad 03:07:06 hmmm 03:08:11 somehow i think that making RPG just became harder 03:08:50 why is that? 03:09:27 RPG's often have spells 03:09:29 :) 03:09:38 not necessity, though 03:09:48 depends on the role 03:34:47 --- join: gilbertdeb (~gilbert@fl-nked-ubr2-c3a-29.dad.adelphia.net) joined #forth 03:35:03 'morning 03:35:13 terve Speuler, fridge ja gilbertdeb 03:35:14 moin moin 03:37:14 hehe. I just read through that old 'linux is obsolete' flamefest. 03:37:28 worse always seems to be better for some reason. 03:38:04 everything which doesn't break is obsolete 03:38:30 that's why space ships blow often 03:38:57 you rob yourself of the chance of replacement against innovate products 03:39:35 but then nothing is ever done! 03:39:51 simply by doing anything in sw/hw one renders it obsolete! 03:40:00 but I don't think thats the case. 03:40:09 'can it grow' should be the most important question. 03:40:39 "can we put bugs in" :) 03:41:30 v 3.x, comes now with the most innovative bugs every coneived... 03:41:46 hehehe. 03:42:09 Bugs are the software's way of saying 'umm guys I am making a self reference and trying to write my own code'. 03:42:26 giving this product a properly designed upgrade path, to protect your investment 03:43:02 I still say Bill Gates and Sergei Bubka are good friends :D 03:43:17 rescuing from the claws of stagnation 03:43:31 I do'nt think stagnation is all bad. 03:43:39 eg, I don't think there is new code for 'ed'. 03:43:43 t'is bad for sales 03:43:47 ah yes. 03:44:00 the animal that forced the Gnu child into existence. 03:44:22 most geeks are terrible business folk. 03:44:30 s/geeks/people 03:46:21 I say bring back the mainframes and install the old programs with better interfaces. 03:46:30 and _dont_ let any one 'improve' them. 03:54:05 --- nick: mur -> murAFK 03:58:16 netbsd seems pretty bloat free 04:00:25 --- quit: murAFK ("reboot") 04:04:07 --- join: mur (ammu@baana-62-165-187-95.phnet.fi) joined #forth 04:07:31 --- nick: mur -> mur__Poissa 04:57:31 --- join: skylan_ (sjh@Rockcliffe4.tbaytel.net) joined #forth 05:09:21 --- join: PoppaVic (~pfv@s106.waters.gtlakes.com) joined #forth 05:11:22 --- quit: skylan (Success) 05:23:54 hmm 05:47:07 --- nick: mur__Poissa -> mur 05:57:37 hi PoppaVic 05:58:03 hiya 05:58:37 what's wrong with this: VM->ip++; 05:58:37 VM->ip += VM->ip; 05:58:37 return NULL; 05:58:37 ? 05:59:05 damnfino - looks wrong, though 05:59:46 inc listp, listip+= listp seems "wrong" 06:00:35 invalid operands to binary + in VM->ip += VM->ip; 06:00:49 I don't know what you are doing, or where 06:01:13 intention: add branch offset to ip 06:01:14 char* xbranch( vm *VM ) 06:01:16 { 06:01:18 VM->ip++; 06:01:20 VM->ip += VM->ip; 06:01:22 return NULL; 06:01:24 } 06:01:55 ok.. let me think a sec. 06:02:23 oh - you want the next cells' VALUE as the new ip. 06:02:33 nice forth ;) 06:02:47 almost. want to add it to VM->ip, because it would be an offset 06:02:58 VM->ip += *VM->ip; 06:03:01 mur: it IS forth :) 06:03:12 it's C... a til-core 06:03:24 oh.. right 06:03:46 * Speuler slaps himself with a 12-ton steam roller 06:05:00 you are such stalin 06:05:10 stalin means steel man, super hero of kind 06:05:19 I've been twiddling my stack-code.. I like "pretty" ;-) 06:05:29 yes that works 06:05:48 of course ;-) 06:05:50 been starting with some stack primitives too. i use a stack cache for tos 06:06:15 right moment to do so .. previous machine didn't use one 06:06:26 hi stepan 06:06:39 the code I tar'ed & emailed had the rudiments of the datastack for you. 06:06:52 found it, used it 06:06:56 good. 06:07:04 mur: Hey.. I'm back now. 06:07:07 thanks 06:07:11 mur: What was it you wanted? 06:07:12 hey guys.. :) 06:07:21 howdy 06:07:21 Hi Stepan, Speuler, PoppaVic.. etc. 06:07:22 Stepan: meet PoppaVic 06:07:30 hi Robert 06:07:43 PoppaVic: Had a look at the code you sent to speuler.. it's really nice 06:07:43 * PoppaVic doffs his non-hat 06:07:46 PoppaVic: Stepan is founding member of #openbios 06:07:54 Cool - glad it was useful. 06:08:37 PoppaVic: our current forth engine keeps the actual VM that steps through the dictionary in docol.. which is not that good of an idea.. 06:08:47 I'd have been a lot happier using funcs, but *sigh* the goto's did the job. 06:08:55 basically the same mistake i did with my very first try feval 06:09:18 why did you not use funcs in the end? 06:09:21 vm is in a function? 06:09:45 in the CFA code for a colon definition, yes ;) 06:10:25 bad design[tm], but it worked like a charme.. only really nasty in some cases 06:10:41 I had a bad feeling about the calls - I fell back to using those goto's and everything "clicked" - certainly is going to be faster, but - sure.. At this point, the inner-interp can now be mad a set of funcs. 06:11:07 this seems to work too: VM->ip += *(++VM->ip); 06:11:29 pre/post-incrementing: in this case, the effect is the same. 06:12:02 ..it's when you use the "stacks" that [++p] and [p++] are going to make a diff 06:12:39 ok. compiler dependent inc-before/inc-after write, i presume 06:13:06 so, there's no reason not to pre-inc in the same statement, reading at *ip, in your NEXT implementation ? 06:13:36 * Speuler goes trying it out 06:13:37 hmm? What I meant was, in the case of such devices as a stack you have the issue of "pre-incrementing" (starting at -1) or "post-incrementing", (starting at zero) 06:15:37 the stack% stuff I just finished IN forth demonstrates it clearer. 06:15:48 same code size.. looks like it doesn't make a difference, concerning the generated code 06:16:03 it won't matter in that case. 06:16:18 so i just leave it the way it was, better readability 06:16:38 it would matter in cases where "A[B++]=C" 06:20:40 Stepan: I am presuming this is TIL is forming some "core" in that 'openbios' thing - what's the scoop? 06:22:55 Speuler: btw - I fixed that damned GForth prompt-shit last night.. 06:23:31 PoppaVic: It's the native core for openbios, yepp 06:23:51 OK. 06:23:51 on top of that all forth language constructs and the open firmware specific words will be implemented 06:23:57 this is intended for eprom, I presume? 06:24:01 yes. 06:24:15 ok - I think I get it. 06:24:19 our current implementation can be booted from grub for testing 06:24:27 (and it should work with linuxbios from flash) 06:24:45 I dunno' "linuxbios" - afaik, I have an award 06:24:59 Therefore i want it to be small, well readable and portable 06:25:08 Certainly. 06:25:28 linuxbios is an attempt to write a firmware from scratch. it started out by putting an IPL and a Linux Kernel to flash, thus the name 06:25:43 the linux kernel in flash knew how to read a new kernel from a hd for example 06:26:11 Yeah, anything that can play with dos/doze - but not cater to the bastards - is a plus. 06:27:28 well, if you want - I've also started a load of .fs extensions. Not sure exactly if they are applicable. 06:28:12 Also, I've a C lib of utilities I've been extending and adding to for the last year or so. 06:31:33 sounds interesting 06:32:15 what we definitely need is a vocabularies implementation, but Speuler started some stuff here some time ago.. i will try this as soon as I got the VM right 06:32:42 Yes, I happen to LIKE vocs/wordlist "namespaces" 06:35:55 --- nick: mur -> mur__Poissa 06:39:10 --- join: Herkamire (~jason@wsip68-15-54-54.ri.ri.cox.net) joined #forth 06:39:57 --- join: tathi (~josh@wsip68-15-54-54.ri.ri.cox.net) joined #forth 06:41:41 --- join: sifbot (~sifforth@ip68-9-68-123.ri.ri.cox.net) joined #forth 06:41:41 Type sifbot: (or /msg sifbot to play in private) 06:43:45 --- quit: fridge ("Time wasted on IRC: 8 hours 15 minutes 14 seconds") 06:46:18 sifbot: 06:47:00 sifbot: ." This is a test" 06:47:01 PoppaVic: 06:47:38 sif is BD... (bent-dick) 06:49:11 PoppaVic: ." is a compiletime only word. 06:49:23 use .( 06:49:26 sifbot: words 06:49:54 see ." 06:49:54 noname : 06:49:54 34 parse type ; 06:49:54 lastxt 06:49:55 noname : 06:49:57 34 parse POSTPONE SLiteral (compile) type ; 06:49:59 lastxt 06:50:01 interpret/compile ." immediate ok 06:50:07 I do, btw, use .( 06:50:37 ..used to work in the Long Ago, too 06:51:36 --- quit: sifbot (Remote closed the connection) 06:51:44 i added, for consistency, s( 06:52:00 s" is not guaranteed to work in interpret mode 06:52:02 --- join: sifbot (~sifforth@ip68-9-68-123.ri.ri.cox.net) joined #forth 06:52:02 Type sifbot: (or /msg sifbot to play in private) 06:52:13 sifbot: words 06:52:15 Herkamire: ." .( : if then do loop repeat until exit ... 06:52:28 s( foo) type complements .( bar) nicely 06:53:15 what is "s(" doing? 06:54:20 wait, is s" supposed to be compiletime or interpret time? 06:57:01 --- quit: tathi ("leaving") 06:58:49 PoppaVic: : s( [char] " word ; ( or something evry similar ) 06:59:18 mmmmmm.... like : s( [char] ) word ; 07:00:01 I'm just wondering what it's supposed to imply.... ".(" sez what it means. 07:00:32 to me it implys that it going to put the string on the stack (s) at runtime (() 07:00:50 s( reads string delmited by ) from input stream, puts addr and len on stack. like s", but interpret mode 07:01:08 s( relates to s" like .( does to ." 07:01:12 ahhh 07:02:35 found it somewhat inconstent, to use .( versus ." , but using s" at compile and interpret time 07:02:55 half-cooked 07:03:49 Yeah, It can become a pain quickly. 07:08:20 * Herkamire does some tests (and writes s( ) and remembers why he didn't have that word. 07:08:37 in my forth ever call to word overwrites the text from last time. 07:08:51 sif: s( bla bla ) type 07:09:21 : s( ..... somebuffersomewhere move$ ; 07:09:32 sifbot: s( bla bla ) type 07:09:33 Herkamire: type la 07:10:03 Speuler: good idea 07:10:29 i return the address in tib 07:10:36 not the ones at here 07:10:47 i presume your word moves string to here 07:10:57 mine doesn't 07:11:17 PAD 07:11:30 that my, my s( can also represent several strings in the tib. 07:11:44 but only as long no other line has been entered 07:12:12 if i need it longer, it do s( foo) push$ 07:12:51 Speuler: cool 07:12:58 if you move it to pad rather than here, why do you move it at all ? 07:13:07 I haven't played with strings that much in the forths I've written 07:13:21 did you have a look at my stringstack stuff ? 07:13:43 Speuler: because he was saying his accept/expect/parse/word was overwriting itself. 07:13:55 I'd just use a larger buffer. 07:15:48 PoppaVic: but that also happens if word overwrites the copied string at pad 07:16:08 I was hoping it would be larger than his in-buff seems to be. 07:16:40 perhaps it's just that sifbot needs a kick in the input/output routines ;-) 07:18:24 I didn't think it was important to have a string on the stack at interpret time. I figured if I even wanted that I could write it. 07:18:33 i think it doesn't matter how big a buffer is if you overwrite its contents 07:18:46 :) 07:18:51 i use it often 07:18:56 testing mainly 07:19:38 --- join: I440r (~mark4@1Cust102.tnt3.bloomington.in.da.uu.net) joined #forth 07:25:23 * PoppaVic needs to teach Speuler about those really NEW things: "file-attachments" ;-) 07:27:16 ahhh.. Good. I see you understood pre/post inc/dec for the datastack ;-) 07:31:00 heh 07:31:05 me ? 07:31:10 no - Speuler 07:31:15 oh - yea 07:31:20 speuler is cluefull :) 07:32:04 I love ++ and -- ...But they can really ruin yer day if yer not "seeing" what you want. 07:34:07 heh 07:49:28 --- nick: mur__Poissa -> mur 08:00:00 oops. 08:00:02 forgot 08:00:11 so'k - all fixed 08:00:11 modified foorforth.h 08:00:15 fooforth.h 08:00:23 added tos 08:00:29 which is the stack cache 08:00:35 contains top stack element 08:00:40 no prob - got it working fine - adding more for you. 08:00:43 don't cache anything 08:00:46 sure 08:00:54 stack caching is great 08:01:03 bad-bad. 08:01:13 there is no call for it in the code I saw. 08:01:14 why ? 08:01:53 what's the objective? it's already "cached" - it's in the array 08:01:57 should i rely on the compiler , to optimize tos access all out by itself ? 08:02:04 imagine @ 08:02:04 yeah. 08:02:10 tos=*tos 08:02:11 ok, so? 08:02:21 local var.. 08:02:24 no pop, fetch, push 08:02:38 unaries 08:02:42 1+ 08:02:45 tos++ 08:03:05 it's not necessary, I tell yah ;-) 08:04:14 even if you make a register variable from tos ? 08:05:02 so it is not in a local var anymore 08:05:10 no need to read from mem 08:05:37 if you need to do weird shit like that - for a func, a local var is ideal. 08:06:24 tos would be global 08:06:47 no reason to bother 08:07:06 how would the compiler know to keep tos in a register ? 08:07:35 Now, up in the header, perhaps the dp/rp should be "register" and maybe the VM* as well. 08:07:40 wouldn't i have to tell it somehow ? while tos on stack would be in an address which changes 08:08:20 the compiler "knows" - it'll optimize as it can, and you can add hints. But stop "optimizing".. It's like learning to fly the Shuttle, when you can't even walk yet. 08:08:21 i gotta go, ill be moving house today :/ 08:08:23 --- quit: I440r ("Reality Strikes Again") 08:09:17 i don't know about c . other impementations work a lot faster when caching tos 08:09:40 but there's no optimizing assembler, to do that all by itself 08:10:24 btw, i can "fly", just not in the c-written shuttle 08:10:55 it was a simile for any programming. 08:11:16 according caching tos, the shuttle should know how to fly itself :) 08:11:24 and i'm interfering, trying to steer it 08:11:31 day after dfay the dolts pop into #c and want to profile and optimize shit that was bad to start with. 08:12:18 boy, I wish you'd email from a reply-able account.. 08:12:32 relating to bashforth, i learned making a bicycle fly :) 08:12:46 Haha 08:13:27 throwing items dont solve anything! 08:14:16 incoming... 08:14:33 splash - "you.. have.. mail." 08:14:44 mur: bashforth actually works :) 08:15:53 got 08:17:25 i think i'm going to put conditional branches in, to enable me running a benchmark on cached tos vs stacked tos, to soo whether it makes any difference 08:17:35 see 08:18:38 can you define how to declare comments in forth? 08:18:57 btw, you can join #openbios if you like. some discussions of fooforth may make more sense there 08:19:34 quote "any forth written in c is not a forth" thus off-topic :) 08:20:21 you can define anything in forth ;-) 08:20:36 bah. 08:21:05 "any forth not written in forth ain't forth, (so yah can bugger off ;-)" 08:21:40 I've got 5 live tabs at the moment, adding another is gonna get too "busy" ;-) 08:21:54 how do you define : to start comment and ; to end. and [to start function and] to end? :) 08:22:58 you write a new function - I wrote a "/*" word - in a file, you can also - apparently - get away with (...) 08:26:27 mur: consider the following for ; : 08:26:34 make it a deferred word 08:26:50 default pointing to error word, complaining 08:26:54 BTW, I liked the idea of the byte (word|long) coded forth "binaries". 08:27:11 : code: anything: sets ; to another word 08:27:27 which ends the clause, and restores the previous vector 08:27:43 by that, you can use ; with anything you like 08:27:55 comment: bla bla ; 08:28:41 translate: foo 1 , 11 , 2 , 22 , ; 08:28:48 stop 08:29:00 comment may not be good to to with that ... 08:29:21 as end of comment is not a word wich gets executed, but just scanned for 08:29:39 * Speuler shuts up 08:30:12 i like something like code: foo opcodes ; 08:30:50 hm 08:30:53 PoppaVic: byte coded forth needs more than just a different data type to deal with address pointers 08:31:00 8 bit is a bit short to represent an address 08:31:28 you'd add some lookup table, to translate byte tokens to addresses 08:31:29 I lost you.. When did I mention dp? 08:31:41 Oh. Certainly you would. 08:31:56 "BTW, I liked the idea of the byte (word|long) coded forth "binaries"." 08:32:15 mixup here ? 08:32:24 yes, "byte-coded" sounds bad.. "token-coded" makes more sense. 08:33:07 openfirmware uses byte tokens. 08:33:09 I was trying to visualize how it could be particularly useful. 08:33:20 i presume openbios would use that too 08:33:32 ..and, I was "seeing" some neat ideas, since ANS is already a baseline. 08:33:43 as wider cell becomes, as more sense does it make to implement token threading 08:34:11 well, I was thinking of what you'd "pack into" those tokens.. 08:34:31 i'd think one could do: 08:34:55 bytes tokens 0..FE - 254 primitives 08:35:00 FF primitive is extension 08:35:16 nah - yer looking at it sideways for the nonce.. 08:35:17 reading next byte, allowing 16-bit primitives 08:35:29 for more 16-bit prims, 08:35:29 one could think of: 08:35:36 0...7F : byte tokens 08:35:55 80...FF -> 8000-> FFF0 word tokens 08:36:12 extend those, if not sufficient, to FFF00000 ... FFFFFFFF 08:36:13 yeah, that would be compact 08:36:42 the 7F*, remember 08:36:47 hmm.. 08:36:53 primitives 80...FF would read another byte, combine it to 16 bit token 08:37:00 * PoppaVic can see another email in the offing.. 08:37:01 no need to put that into interpreter 08:37:32 can be handled by the primitives, and interpreter would just execute byte tokens 08:37:46 right... lemme cogitate it more.. 08:42:37 * mur escapes from fire exit 08:42:45 so long, the bomb is set! 08:43:19 :bomb all select erase " buhaha" evil message send ; 08:43:59 --- quit: mur ("bomb") 08:55:08 --- join: deluxe (~deluxe@p508050D6.dip.t-dialin.net) joined #forth 08:59:11 --- join: rafe (~rafe@www.scinq.org) joined #forth 09:31:02 --- quit: rafe ("Client Exiting") 10:17:30 --- quit: deluxe (Read error: 60 (Operation timed out)) 10:22:07 --- join: Serg_Penguin (~z@ppp144-219.dialup.mtu-net.ru) joined #forth 10:25:14 --- quit: Serg_Penguin (Client Quit) 10:32:35 --- join: mur (ammu@baana-62-165-187-95.phnet.fi) joined #forth 10:35:28 Robert 10:35:34 gilbertdeb, people 10:44:44 Hi mur :) 10:44:55 lo 10:45:21 terve! :) 10:45:26 * mur is bored :( 10:45:36 and sad becasue of being bored 10:45:43 i dont know anyhting i coudl do 10:45:52 i cannot be without doing something 10:45:57 Hm. 10:46:16