00:00:00 --- log: started forth/02.04.28 00:10:20 --- quit: futhin ("bye") 00:19:14 --- join: Soap` (flop@210-55-149-23.dialup.xtra.co.nz) joined #forth 01:53:16 --- join: Fare (fare@samaris.tunes.org) joined #forth 02:03:33 --- join: Soap- (flop@210-54-75-41.dialup.xtra.co.nz) joined #forth 02:14:02 --- quit: joa (Read error: 110 (Connection timed out)) 02:15:25 --- join: joa (~james@59-125-237-24.anc-dial.gci.net) joined #forth 02:19:39 --- quit: Soap` (Read error: 104 (Connection reset by peer)) 02:36:31 --- join: rob_ert (~robert@h237n2fls31o965.telia.com) joined #forth 03:00:38 --- quit: Fare ("Connection reset by pear") 03:05:18 --- quit: Soap- (Read error: 104 (Connection reset by peer)) 03:16:26 --- join: Speuler (~l@a161161.upc-a.chello.nl) joined #forth 03:31:38 Hi 03:37:32 --- join: Soap` (flop@210-54-75-41.dialup.xtra.co.nz) joined #forth 03:37:48 Hello there, Soap` 03:38:23 Hi 04:51:29 hi the_rob 04:51:51 hi soap 05:56:39 --- quit: Speuler ("using sirc version 2.211+KSIRC/1.1") 06:07:32 --- join: Speuler (~l@a161161.upc-a.chello.nl) joined #forth 07:23:23 --- join: gilbertbsd (~gilbert@max2-82.dacor.net) joined #forth 07:29:07 Hey 07:29:17 hi rob_ert 07:31:31 --- quit: gilbertbsd (Excess Flood) 07:31:32 --- join: gilbertbsd (~gilbert@max2-82.dacor.net) joined #forth 07:31:51 wb, flooder :) 07:32:02 how did you know? 07:32:53 <-- gilbertbsd has quit (Excess Flood) 07:32:54 ;)) 07:33:04 ahhh 07:33:17 I didn't quit. xchat quit on my behalf. 07:34:02 Heh 07:34:09 OPN kicked you out :-) 07:34:28 oh OPN did that huh? 07:34:43 Are you learning forth for the fun of it? 07:36:45 Yup :) 07:36:57 What other reasons are there? ;) 07:37:04 to make CM happy. 07:37:13 CM? 07:37:15 What's that? 07:37:36 C. More! 07:37:39 Hah 07:37:44 Moore ;D 07:37:47 More. 07:37:51 No! No! 07:38:06 see why I simply said CM? 07:38:50 No. 07:38:51 .) 07:39:12 Well CM happens to be his initials 07:57:05 --- part: gilbertbsd left #forth 11:14:29 --- join: I440r (~mark4@1Cust208.tnt1.bloomington.in.da.uu.net) joined #forth 11:15:44 --- quit: I440r (Client Quit) 11:15:51 --- join: I440r (~mark4@1Cust208.tnt1.bloomington.in.da.uu.net) joined #forth 11:19:24 hi 11:20:46 hi 11:20:55 i ALMOST got vocabs working heh 11:21:01 cool 11:21:03 it runs. i can type buy but some words now crash 11:21:05 lol 11:21:15 and entering a blank line also segfaults now heh 11:21:24 well, that's not a shippable version @:^> 11:21:25 but i shud be able to fix that np 11:21:32 not quite :) 11:21:39 but i did say i ALMOST had it ready heh 11:21:49 glad to see yoiu decided to tackle that issue 11:22:05 i also discovered a new thing in fpc that had always realy realy confused me 11:22:09 I imagine it's kinda nasty in NASM 11:22:16 find called a word called prior_check 11:22:27 it is nasty - getting a meta compiler wil tidy it up 11:22:39 i could never figure wtf prior_check was for 11:22:42 heh, glad to hear you say that, too 11:22:48 only forth also also also also also also also also 11:23:05 you wuill be searching the forth vocabulary 9 times 11:23:09 for the same word 11:23:16 prior_check prevents this 11:23:17 on some forths, yes 11:23:28 ah! ok 11:23:30 if you say only forth also blah also forht alst blah 11:23:41 you WILL search both forth and blah twice 11:23:55 right ... for now anyway 11:24:27 hrm ... I'm almost afraid to ask ... 11:24:43 i think im going to make vocabs NOT add themselves to the top of the context stack if they are already in it 11:24:52 to ask what ? heh 11:25:09 rather the vocab word rotates its token to the top of the list? 11:25:47 that would work for me! 11:25:49 ANS did not specify how vocularies worked, buut they gave a set of primitives that would allow most current vocabulary schemes to be built 11:25:53 i hadnt thunked of that 11:25:56 did you use those primitives? 11:26:17 no i used the ones from fpc - but not verbatum 11:26:24 I'm pretty sure you didn't ... lemme find a URL for you 11:26:26 i still have issues that i wanted to resole 11:26:33 it's one of the more elegant solutions, IMO 11:27:02 :) 11:29:22 ok ... here's the discusion about *WHY* they did this ... 11:29:27 http://www.taygeta.com/forth/dpansa16.htm 11:30:01 on my way 11:30:47 is skip carter still alive ? 11:31:01 erm ... I think so 11:31:07 why would you ask that? 11:31:10 he used to come on opn you know 11:31:19 no, I didn't 11:31:23 yup 11:31:30 he has even been in here once :) 11:31:52 that page is loading in slowly 11:32:14 in fact it wasnt loading in. i restartd it 11:32:14 that's right, you're modem now 11:32:19 grr i hate my dial in :( 11:32:23 I can imagine 11:32:39 I'm spoiled, too 11:32:52 ijust reloaded my web email but that url you gave me isnt moving at all 11:32:53 grrr 11:33:23 the modem lights arent even blinking 11:33:33 sure reloads on mine, less than 1/2 second 11:33:48 * MrReach sticks out his tongue 11:33:52 i dont think i can resolve taygeta.com 11:34:00 hang on i saw the modem blink 11:34:16 jeesh - maybe theres a router down between me and him 11:34:28 and the net is too stupid to find an alternate route 11:35:11 anyway, im still not hashing yet :) 11:35:26 thanks, I was going to ask about that 11:35:27 vocabs have 16 threads each for now but all words are chained on the first thread 11:35:44 i may bunp to 32 threads 11:35:48 hashing is easy ... erm ... well, maybe not with NASM 11:35:50 fpc used 64 threads 11:35:58 it is impossible with nasm 11:37:04 or -it would be so fscking complex it would make even the best OOP coders heads spin 11:37:31 heh, that implies that OOP programmers write very complex code. 11:38:17 they do 11:38:22 they write sphagetti 11:38:31 write only code 11:38:52 they write naturally obfuscated code that they go out of their way to make even MORE obfuscated 11:38:54 :P 11:39:03 heh 11:39:44 that page wont load in at all 11:40:19 er i seem to recall having great difficulty reaching taygeta a cpl of weeks ago too 11:40:32 maybe he is dropping all requests that come from my domain ? 11:40:33 ok 11:40:36 earthlink i mean 11:40:53 grrr 11:40:56 do you have a copy of the dpANS on your local machine??? 11:41:00 this is pissing me off now 11:41:03 do i hell heh 11:41:09 what does the dp stand for ? 11:41:09 you ought to by now 11:41:18 Draft Proposed 11:41:34 the actual standard is like $400 from ANSI 11:41:43 lol 11:41:45 fuck that 11:41:54 the dpANS is the version submitted for final approval 11:41:56 theres another good reason NEVER to use ans :P 11:42:43 duh! it costs money to publish the book, and the ANS people get paid (and they should) to deliberate on the standards before publishing them 11:43:19 I think the ASCII standard is free now, if that any consolation @:^> 11:43:37 it doesnt cose me a penny to publish my isforth standard }:) 11:44:01 actually, I'm surprised that ANSI allows the draft to be so openly published 11:44:03 erm i mean - my isforth NON standard heh 11:44:09 right 11:44:29 do you have a copy of dpANS on your local machine for reference? 11:44:35 if they didnt i bet NOBODY but forth inc would use the standard 11:44:39 no 11:44:57 may I ask why not w/o a huge explosion? 11:45:15 i read parts of it but it got reeeeeel boring and reeeeeelll annoying so i didnt even read all of it heh 11:45:26 wel im not gona explode but 11:45:27 yes, it is boring 11:45:43 why should i retain a copy of a standard i have absolutely NO desire to follow ?> 11:46:08 www.taygeta.com refuses to load 11:46:36 ok, here's a brief rundown of the search order primitives that ANS encourages ... 11:46:48 k 11:47:14 that i did want to read - im not 'its ans so i dont want to know' - if it works for me ill use it :) 11:47:23 at the very base, it is predicated an WORDLIST GET-ORDER SET-ORDER GET-CURRENT SET-CURRENT 11:48:10 WORDLIST ( -- wid ) ... wid is a "wordlist" identifier 11:48:38 an opaque stack-wide number that the system implementor can use any way s/he sees fit 11:49:10 GET-ORDER ( -- widn ... wid1 n ) 11:49:21 get the current search order 11:49:50 SET-ORDER ( widn ... wid1 n -- ) ... set the current search order 11:50:04 ugh - if i use the methodology the OVERLY-VERBOSE-WORD-NAMES-FOO-BLAH-BAM will have to change :) 11:50:12 ohhh i like set-order and get-order ideas tho 11:50:26 SEARCH-WORDLIST ( c-addr u wid -- 0 | xt 1 | xt -1 ) 11:50:39 get-order puts the context stack on the PARAMETER stack and a depth count :) 11:50:50 which is just like SEARCH , but only searches a specific wid 11:50:55 i would prolly name then >order and order> 11:50:58 that's correct 11:51:11 and search-wordlist would HAVE to be renamed to FIND 11:51:25 : ALSO get-order over swap 1+ set-order ; 11:51:34 NO!!! 11:51:43 FIND searches all the wordlists, right?? 11:51:44 i do not like coders who use entire sentences for function names 11:51:48 no 11:52:01 actually find searches the ones in the context stack 11:52:09 oh this brings me to a question 11:52:11 words 11:52:16 you could also use ORDER@ and ORDER! 11:52:17 how should words work 11:52:25 thats good too!!! 11:52:41 i was thinking that if you type 11:52:51 brb 11:54:11 ok, may I paste part of the discussion that ANS has about search orders? 11:54:22 you've got to understand how they think about such matters 11:54:41 in the case of search orders, they refused to dictate how search orders were to be implemented 11:54:43 yes of corse 11:54:54 thats good :) 11:54:56 so instead they chose to write a set of primitives 11:55:13 mrreach never ask for permission to do something. just do it and we can discus it :) 11:55:21 the words they chose for the primitves must not conflict with systems already in existence 11:55:44 so they had to RENAME the words to something nobody would ever chose 11:55:56 that is why "entire sentences for word names" was used 11:56:07 that is correct 11:56:09 its still very un-forthlike :) 11:56:23 ok, here's the paste 11:56:46 Search-order specification and control mechanisms vary widely. The FIG-Forth, Forth-79, polyFORTH, and 11:56:46 Forth-83 vocabulary and search order mechanisms are all mutually incompatible. The complete list of 11:57:00 incompatible mechanisms, in use or proposed, is much longer. The ALSO/ONLY scheme described in a 11:57:00 Forth-83 Experimental Proposal has substantial community support. However, many consider it to be 11:57:13 fundamentally flawed, and oppose it vigorously. 11:57:14 . 11:57:20 it is flawed heh 11:57:39 Recognizing this variation, this Standard specifies a new primitive set of 11:57:39 tools from which various schemes may be constructed. This primitive 11:57:39 search-order word set is intended to be a portable construction set from 11:57:58 which search-order words may be built, rather than a user interface. 11:57:59 ALSO/ONLY or the various vocabulary schemes supported by the major 11:57:59 Forth vendors can be defined in terms of the primitive search-order word 11:57:59 set. 11:58:09 . 11:58:22 The encoding for word list identifiers wid might be a small-integer index 11:58:22 into an array of word-list definition records, the data-space address of 11:58:22 such a record, a user-area offset, the execution token of a Forth-83 style 11:58:36 sealed vocabulary, the link-field address of the first definition in a word 11:58:36 list, or anything else. It is entirely up to the system implementor. 11:58:52 . 11:58:59 [skip a section] 11:59:15 There has never been a portable way to restore a saved search order. F83 11:59:15 (not Forth-83) introduced the word PREVIOUS, which almost made it 11:59:15 possible to unload the search order by repeatedly executing the phrase 11:59:28 CONTEXT @ PREVIOUS. The search order could be reloaded by 11:59:28 repeating ALSO CONTEXT !. Unfortunately there was no portable way to 11:59:28 determine how many word lists were in the search order. 11:59:31 . 11:59:45 ANS Forth has removed the word CONTEXT because in many systems its 11:59:45 contents refer to more than one word list, compounding portability 11:59:45 problems. 11:59:58 [end quote] 12:00:38 does that make sense? 12:00:39 hrm - context was just an array that was treated SORT of like a stack 12:00:43 but yes 12:00:45 it does 12:01:08 yes, but that type of array is not used on all systems, so they chose to avoid it 12:01:16 yes 12:01:27 your free to use context, but they have nothing to say about it 12:01:45 (other than to say they aren't saying anything about it, deliberately) 12:01:55 heh 12:01:59 heh 12:02:49 : ALSO get-current over swap 1+ set-current ; 12:02:58 i dont understand that 12:03:17 oh 12:03:19 yes i do 12:03:31 ooops! I typo'd it! 12:03:44 : ALSO get-order over swap 1+ set-order ; 12:03:59 it's bad to typo example code ... VERY bad 12:04:24 lol 12:04:37 ok, let's try ... 12:05:26 actually, i dont like 'also' 12:06:02 i think the 'also' mechanism is part of the above mentioned FLAWS 12:06:10 : VOCABULARY create wordlist , DOES> @ >r get-order nip r> swap set-order ; 12:06:37 erm. let me digest that 12:06:38 in that vocab, the new wordlist replaces the wordlist in the top of the search list 12:06:45 what does 'wordlist' do 12:06:57 ( --- wid ) 12:07:13 creates a new wordlist, returns an "identifier" 12:07:25 not required 12:07:29 in your system, it'll probably be a pointer to the hash table 12:07:33 its an unneeded complexity 12:07:56 $10 says you've already got a word that does that 12:07:57 vocabularies are just WOARD with a specific function. there ID is the address of their body field 12:08:17 so i woukld do 12:08:24 hi 12:08:30 greets, Speuler 12:08:37 erm. no. 12:08:39 ugh 12:08:44 I440r: wordlists are not ONLY used for vocabularies 12:08:47 your compiling the id INTO the vocabulary 12:09:07 I440r: you'd like to be able to access them by an id 12:09:08 yes, this is to make it look like the F83 system 12:09:09 do you need to give each constant you define an ID ? 12:09:26 what about variables. lets give them ID's too heh 12:09:30 no, the wid can be handed around on stacks and never stored, if you wish 12:09:40 huh? 12:09:45 I440r: variable "id" may be their addresses 12:09:58 the only reason it's going into a word is because you have to access it, somehow 12:10:00 speuler a vocabularies ID can ALSO be their address 12:10:10 I440r: what about vars you can only refer to by name, not by address ... 12:10:22 I440r: yes. implementors choice 12:10:23 er, no such thing 12:10:28 I'm lost 12:10:31 absolutly NO such thing 12:10:43 if you have to refer to the variable by NAME you cant turnkey 12:10:44 * Speuler gone again 12:11:36 I440r: wordlists you'd like to be able to access by some id too, maybe their address. not just by vocabulary name 12:11:49 why 12:11:53 I440r: ok, I think you're doing vocabulary this way ... create 0 , 0 , 0 , ... 12:11:56 is that right? 12:12:04 sort of 12:12:14 not quite 12:12:16 but close ennuff 12:12:17 cause programs use wordlists for more than just to build vocabularies with 12:12:18 and then a does> clause at the end? 12:12:47 no theres more info in a vocabulary thatn just the thread array :) 12:13:04 each vocabulary also has a pointer to the previously defined one. 12:13:09 tho - i might dispense with that part 12:13:13 hwever. 12:13:22 ok, wordlist might look like this ... : wordlist ( -- wid ) HERE 0 , 0 , 0 , ... ; 12:13:25 each vocabulary will eventually have its own copy of HERE 12:14:03 the wids are what actually get stuck into the search order stack 12:14:16 er. wordlist is almost identical to array 12:14:35 its an array with no does> part - but it COULD HAVE a does> part! 12:14:43 so wordlist could simply be 'array' 12:14:50 however, words created by VOCULARY need to know about their own specific wordlist, and that's why it's CREATE ! 12:15:27 ... CREATE ! DOES> @ ... 12:15:36 this still doesnt make sense to me unless the word list array is compiled off in space smewhere 12:16:01 beg parden? 12:16:41 we dont NEEED to compile the address of the thread array into a vocabulary !!!!1 12:16:50 its IN THE VOCABULARY 12:16:54 its like 12:16:58 some_voc: 12:17:01 call do_voc 12:17:10 dd pointer-to-word-list 12:17:12 word-list: 12:17:15 dd 0,0,0,0,0 12:17:25 why the fuck have the pointer to data you KNOW the address of 12:17:34 i SERIOUSLY dont understand 12:18:47 you're doing the same thing I am 12:18:58 ok, coceptually ... 12:19:04 conceptually ... 12:19:20 a wordlist is known by the address of its hash table, right? 12:19:22 i am not storing an address of an array i KNOW THE ADDRESS OF inside the array itself 12:19:33 no 12:19:43 but close ennuff heh 12:19:53 ok, how do you identify one wordlist seperate from another? 12:20:00 some_voc: 12:20:04 call do_voc 12:20:12 dd pointer-to-previous-vocabulary 12:20:19 dd some-other-info-might-go-here 12:20:23 dd 0,0,0,0,0,0,0,0 12:20:27 ok, fine, but what does do_voc do? 12:20:49 : dovoc ( a1 --- ) current ! ; 12:20:54 basically 12:21:01 does it not pull the address off the return stack? 12:21:08 no 12:21:23 the call to dovoc left the body address of the vocabulary on the PARAMETER stack 12:21:30 oh, ok 12:21:50 so that's not TOO terribly different from the CREATE DOES> version, right? 12:22:30 erm i lied. its 12:22:40 : dovoc ( a1 --- ) context ! ; 12:22:41 heh 12:23:17 your version stored the address of the item following the address where you stored the address 12:23:30 you store a pointer at address a1 to address a1+1 12:23:31 ugh 12:23:53 now, here's the trick ... by isolating the WORDLIST and SEARCH-WORDLIST ... you can use vocabs completely outside of colon code create, etc 12:24:24 for example ... 12:25:35 : hidden-vocab ( -- ) WORDLIST dup (store wid somewhere ) CURRENT ! ( some type of interpreter loop here ) ... ; 12:26:04 rrrr 12:26:14 im TOTALLY LOST with this WID 12:26:17 totally 12:26:22 ok .. 12:26:35 it doesn't matter what the wid is 12:26:38 its like lets have an address inside all variable bodies pointing to the next address so we know where the data is stored 12:26:44 myvar @ @ 12:26:45 ugh 12:26:53 but you know that you can store it in CURRENT and that's where new defs will go 12:27:31 it might be completely unavailable other than as a number on the stack 12:27:31 context is an array of addreses to vocabulary words. each vocabulary word has a hash table within it. 12:27:40 thers no need for a ID 12:27:50 every word in the entire dictionary has an ID 12:28:06 right, but it would be several wids that are stored in the context stack ... follow? 12:28:13 err 12:28:14 no 12:28:17 i dont 12:28:30 they invented a name for that address (ANS did) 12:28:31 why store an ID - just store the address of the damned vocabulary ? 12:28:43 until now, that was called "the address of whatever" 12:28:56 that's exactly right, I440r 12:29:02 but your storing the ID inside the thing being identified 12:29:08 ugh 12:29:11 every word has an address 12:29:11 I did? when? 12:29:18 why store the address of that address inside that address!!!!!!!!!!!1 12:29:24 create here , blah 12:29:28 is basically what you are doing 12:29:37 is not 12:29:44 well, is and is not 12:29:53 ok 12:30:18 what this does is break the hash table out into a seperate entity in its own right 12:30:40 this allows the hash table to be used outside of the regular forth mechanism 12:30:42 compiles it off in space somewhere 12:30:55 featureitis 12:30:57 it might, or it might compile it just previous in the dictionary 12:31:21 im sorry but this ID mechanism is another thing you will never sell me on 12:31:38 im sold on the idea of the context stack being placed on the parameter stack 12:31:45 thats neat 12:31:56 the point is that by isolating the wordlists into WIDs, a whole bunch of different search order setups can be used ... even in the same system 12:32:06 but theres absolutely NO reason to give a vocabulary a 'special identifier' when it already has one 12:32:18 that's the point, I440r 12:32:38 they DO NOT SAY what a WID actually is! 12:32:52 every word in the entire dictionary has a special identifier. its called its execution token 12:33:03 only that you must pass it to SEARCH-WORDLIST to find your words 12:33:19 search-wordlist is (find) 12:33:25 they didnt need to rename that one 12:33:25 right 12:33:26 at all 12:33:31 yes, they did 12:33:34 no they didnt 12:33:54 because in every single forth ive ever seen (find) searches ONE VOCABULARY for the word 12:33:55 because they eforced ( adr len wid -- ... ) semantics on it 12:34:03 enforced 12:34:04 ugh 12:34:09 they "forced" 12:34:21 so they didn't DARE use a commonly used word 12:34:26 sorry mrreach - this wont work for me at all 12:34:35 well, it CAN work for you 12:34:47 in fact, I'd bet you're 90% compliant as it is 12:34:52 the way ans THINKS of things is TOTALLY non forth 12:35:02 * MrReach laughs. 12:35:28 i read a quote from chuck moore 12:35:29 ok, let's go back to get-order set-order ... 12:35:36 you like that, right? 12:35:46 the ans standard does not describe the forth language but a language with the same name 12:35:54