00:00:00 --- log: started forth/05.06.22 00:03:44 --- quit: warpzero (Excess Flood) 00:07:46 --- quit: Raystm2 ("User pushed the X - because it's Xtra, baby") 00:16:22 --- join: warpzero (~warpzero@wza.us) joined #forth 00:16:44 --- join: arke_ (f2@bespin.org) joined #forth 00:16:44 --- quit: arke (Read error: 104 (Connection reset by peer)) 00:45:40 --- nick: arke_ -> arke 01:23:14 --- quit: vitaminmoo ("Leaving") 01:25:38 --- join: vitaminmoo (~vitaminmo@dsl-94-128.peak.org) joined #forth 01:51:41 --- part: hakossem left #forth 03:42:54 --- join: nothingmuch (~nothingmu@CPE00b0d0cd759c-CM0012253ea774.cpe.net.cable.rogers.com) joined #forth 04:48:14 --- quit: PayphoneEd (Remote closed the connection) 05:29:14 --- join: PayphoneEd (~Ed@payphoneed.user) joined #forth 05:39:59 --- quit: Serg[ICQ] ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org") 06:51:18 --- join: sproingie (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth 06:51:22 --- quit: Robert ("brb") 06:52:14 --- join: PoppaVic (~pete@0-2pool236-177.nas22.chicago4.il.us.da.qwest.net) joined #forth 06:52:38 Mornin' 06:52:46 --- join: Robert (~snofs@c-f778e055.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:58:43 'lo 06:58:56 wassup? ;-) 07:01:07 Hi, hi 07:01:42 I wish Stepan would wake up occasionally ;-/ 07:04:44 --- join: true-grue (true-grue@ppp133.medlux.ru) joined #forth 07:08:10 --- part: true-grue left #forth 07:23:54 --- join: Robert__ (~snofs@c-f778e055.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 07:27:15 --- join: nothingmuch_ (~nothingmu@CPE00b0d0cd759c-CM0012253ea774.cpe.net.cable.rogers.com) joined #forth 07:27:21 --- quit: Robert (Nick collision from services.) 07:27:22 --- nick: Robert__ -> Robert 07:27:43 --- quit: nothingmuch (Read error: 145 (Connection timed out)) 07:30:17 is the S@ word in use by something? 07:30:43 maybe a string-stack? 07:31:06 if you can't find it, it's prolly free - what make-do? 07:33:06 nothingmuch_: I commend vocabularies/lexicons to you. 07:33:07 i see a ref of some sort saying it's the stack pointer 07:33:22 so my question is: 07:33:27 it may well be returning the stack-pos 07:33:58 nothingmuch_: listen again: consider vocs/lexicons 07:33:58 are there forths whick keep the stack in the heap? 07:34:17 no, the stack is almost always 180' away from the 'heap' 07:34:26 ..at least the FORTH "heap" 07:34:26 okay 07:34:49 so do you have a 07:34:55 'there' word? 07:34:59 almost every forth around grows the stack toward the botom of the heap and the heap toward tos 07:35:11 what the heck is 'there'? 07:35:29 i wasn't serious 07:35:40 there are some forths that use the heap for the data stack, yes 07:35:56 and the returns stack in those settings is just preallocated? 07:36:01 sproingie: well, I'd question 'heap' 07:36:12 far as i know, they're all pedagogical 07:36:36 they use the C heap 07:36:42 I get hollered at in ##C for using 'heap' as a term.. They all prefer magical-phrases now. 07:36:54 like what? 07:37:00 "dynamic memory" 07:37:03 hah 07:37:56 and "There is no 'heap' now." 07:37:57 the term heap is ambiguous. i guess some retards can't handle ambiguity 07:37:57 dynamic memory == brk? 07:37:58 which makes me laugh, but ok. Fine.. My mindset is very old 07:37:58 doesn't help that some heaps aren't even allocated using heaps 07:37:58 because they have a name for that already: dynamic memory 07:37:58 ns. The function is there 07:37:58 er some dynamic memory arenas that is 07:37:59 err, tired, sorry 07:38:03 i meant to say: data segment 07:38:10 more NS. 07:38:43 there is 'program space' - which has code/data segs and such. ANd there is "Everything else" - which is stack and heap 07:38:49 i say use nothing but segment names. which are the same on aout, elf, and PE i think 07:38:59 can someone please explain what exactly the create word does, btw? 07:39:01 no doubt someone on some alien architecture will flame you 07:39:12 like why the word variable is just 'CREATE 0 ,' 07:39:20 see create 07:39:20 : Create 07:39:20 header reveal dovar: cfa, ; ok 07:39:34 which forth is that? 07:39:43 because they are too lazy to create a new defining word; gforth 07:40:05 create defines the compile time behavior of a word 07:40:09 'create' should generate a header, just a header 07:40:19 and does> defines the runtime behavior 07:40:28 it SHOULD be the same as a variable 07:40:47 so when i invoke CREATE FOO 07:40:53 the dictionary linked list is prepended 07:41:04 huh?? 07:41:07 variables only take values, create can execute reserve arbitrary amounts of space and run arbitrary statements 07:41:11 with a pointer to 'here'? 07:41:15 just CREATE FOO is the same as a variable tho 07:41:22 it grabs the next 'word' as the name - creates a header, and it is done 07:41:53 by default, (everywhere I've been), it is the same as a 'variable foo' 07:42:36 what does "create a header" mean? a new dictionary entry 07:42:36 yep 07:42:44 okay, so that's what I sort of meant 07:42:51 and the dictionary entry points to 'here', right? 07:42:57 well, "sort of" in forthish can shoot yer foot 07:43:01 no 07:43:16 here is right after it anyway? 07:43:45 see variable 07:43:45 : Variable 07:43:45 Create 0 , ; ok 07:43:45 Yeah, GENERALLY 'here' follows it. 07:43:45 okay 07:44:03 However, be advised - this is all voodoo, because you can have headspace, dataspace and codespace, etc 07:44:06 so that allots some space, and writes out flags, and the name, and what not, and also includes a pointer to the preceding dictionary entry 07:44:23 assuming a very naive forth, where everything is in the heap, and the dictionary is a linked list 07:44:32 presuming it's a linked list, yep 07:44:39 dude, you need to check out cfas, lfas, etc 07:44:40 reva for instance uses a hashtable 07:48:12 i found a good explanation of create..does> a while back 07:48:28 oh, well - DOES> is lovely. 07:48:30 brodie's explanation in starting forth is pretty terrible 07:49:12 tho the rest of the book's good, so by that time, it might suffice for some 07:49:13 that would be a defining-word.. A word to create new words 07:49:50 I'm trying to meld this into my C module.. It took many revisions to get this far. 07:49:53 that's precisely what i don't like, that makes it sound so extra-magic. hell, : defines new words too. 07:50:03 no 07:50:20 i prefer the explanation of the split of compile/run behavior 07:50:24 : defines a word. does> yields DEFINING words 07:51:03 or.. 07:51:05 hmm 07:51:05 sure, along with create 07:51:11 --- join: Herkamire (~jason@c-24-218-95-147.hsd1.ma.comcast.net) joined #forth 07:51:11 --- mode: ChanServ set +o Herkamire 07:51:20 could think of it in OO terms 07:51:30 create is the definition, does> is the constructor 07:51:32 Better yet, 'does>' creates a vectored, reusable word-ref 07:52:03 --- join: nothingmuch (~nothingmu@CPE00b0d0cd759c-CM0012253ea774.cpe.net.cable.rogers.com) joined #forth 07:52:10 tho the oo comparison really starts to break down when you put extra behavior in that "constructor" 07:52:16 yep 07:52:24 for naive OO anyway 07:52:29 sorry.... lost conn again. CAan anyone link me to the backlog? 07:52:36 i tend to do a lot of RAII-style OOP, so i make instantiation do a lot of work 07:52:46 makes C++ very declarative 07:53:02 oh, talking abote does> 07:53:05 ? 07:53:06 all you can say of 'does>' is: it lets you create a new 'word' using the older "does>..." body. 07:53:47 --- quit: nothingmuch_ (Read error: 145 (Connection timed out)) 07:55:27 : foo create 1 . does> .s ; ok 07:55:27 foo bar 1 ok 07:55:27 bar <1> 208432 ok 07:56:40 It was sorta' "easier" when it was known as " where can I find logs forth #forth? 07:57:07 no idea 07:57:54 can someone msg me the lines from around "sproingie: reva for instance uses a hashtable"? 07:58:11 i'll do it 07:58:22 thanks! 07:58:26 good 07:58:43 --- quit: sproingie (Excess Flood) 07:58:58 hehe MSG< sproing 07:59:01 oh my, i've said a lot 07:59:24 nothingmuch_: okay, when i write a 0 using '0 ,' into the next space, when that is invoked, how does it know to put the address there instead? 07:59:26 [16:46] nothingmuch_: i mean, invoking a word doesn't just put the pointer to the words def on the stack 07:59:29 [16:47] nothingmuch_: does : run create and then set a bit in the header? 07:59:34 --- join: sproingie (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth 07:59:52 bah 07:59:56 i got /killed for flooding 08:00:14 hah 08:00:29 sorry 08:00:42 /msg is better ;-) 08:00:48 i did msg 08:00:54 dcc is better 08:00:54 At least you don't get waxed for it ;-) 08:01:04 firewalls screw up dcc tho 08:01:04 my dcc never worked right 08:03:11 PoppaVic? 08:03:30 huh? my xchats dcc doesn't want to work on macosx, that's all 08:04:05 So, I rely on email-attachments, msgs and rafb 08:08:47 anyone? 08:10:01 what? 08:10:36 when : invokes create, does it set some kind of flag to differentiate a variable and a word to be executed? 08:10:59 no 08:11:16 it compiles a func-ptr and an address-literal. 08:11:35 CREATE does? 08:11:36 the address can be 'here' or wherever does> came into play 08:12:03 and : overwrites that, by compiling the next words over that? 08:12:18 well, not : per se, but the fact that STATE is true? 08:12:33 I think you need a book or two and a LOT of code-reading 08:12:50 state is state - somefolks tweak it, others scream at that 08:13:04 i do 08:13:33 in fact, I can't recall offhand anymore, but I believe 'state' is "compile" versus "execute" 08:13:39 CREATE creates a now word. CREATE FOO would make a new word FOO which gives you the address of the heap 08:14:08 you can use ALLOT or , (comma) to make some space/data where FOO points 08:14:28 eg to make a 100 byte buffer named foo, you can do: 08:14:34 CREATE FOO 100 ALLOT 08:14:35 so : uses CREATE to make a new dictionary entry, switches to compilation sematics, and then... nothing? 08:14:50 no, : doesn't use CREATE 08:14:52 nothingmuch: look back at the def for "variable" 08:15:08 1 0 - see : 08:15:08 : : !CSP CREATE _TICK _NEST LAST _AT _STORE HIDE ] ; ok 08:15:09 : doesn't want the bit about having the new word return a pointer to the heap 08:15:31 that's where I got my impression 08:16:25 hmm 08:17:09 create doesn't reserve anything, either. Which is why I keep dissuadingyou 08:17:31 : variable create 0 , ; 08:17:56 PoppaVic: i didn't say it did 08:18:10 nothingmuch: just trying to insure you don't lose sight 08:18:28 ok, remember 'there' you joked about? 08:18:34 yes 08:18:59 : mark create does> release ; 08:19:04 mark there 08:19:18 DOES> is pretty confusing too 08:19:48 i know how to use does> 08:20:03 what I am trying to figure out is what exactly do CREATE, DOES> and : do 08:20:09 : foo create 0 , does dup @ 1+ tuck swap ! ; foo bar 08:20:14 oops 08:20:18 : foo create 0 , does> dup @ 1+ tuck swap ! ; foo bar 08:20:21 and by do i don't mean what you get from them, but how they work 08:20:51 so 'foo something' will create 'something' 08:20:58 right 08:21:05 something will jump to the code compiled after the 'does>' 08:21:07 everything after DOES> dosn't execute when FOO executes, instead it's coppied into the latest new word (bar) which should be created with CREATE. 08:21:13 dup duplicates the address allotted by '0 ,' 08:21:26 yeah, it's a code-vector 08:21:36 okay 08:21:39 Herkamire: copied or referenced 08:21:41 so now when you execute BAR it's going to first put that address to the part of the heap, then execute "dup @ 1+ tuck swap ! ;" 08:21:52 PoppaVic: whatever 08:21:54 right 08:21:58 without the ';' 08:22:11 more like dup @ 1+ tuck swap exit, right? 08:22:15 no, the ';' is a valid 'word' 08:22:21 referenced would be better 08:22:56 i know, but it's a compile time word that compiles a return to caller into the currently being compiled word, doesn't it? 08:23:07 nothingmuch: it just copies/references the compiled code that's the result of compiling "dup @ 1+ tuck swap ! ;" 08:23:37 okay, that's what I meant 08:23:37 if you are looking for state-smart (or restricted) words, THAT is a different set of beasts 08:24:29 see ['] 08:24:29 : ['] 08:24:29 ' POSTPONE ALiteral ; immediate compile-only ok 08:24:29 see ' 08:24:29 : ' 08:24:30 (') name?int ; ok 08:24:33 uh... i think i'm misrepresenting my questions 08:25:37 nothingmuch: yes, ; compiles a return to caller, and does some other stuff at compile time (changes STATE back to interpret, and (in ANS forths anyway) makes the latest new word findable in the dictionary 08:25:41 okay... 08:25:42 * nothingmuch puts on a hat saying "forth machine" 08:25:43 there's nothing to do right now, so /me reads a word from input 08:25:50 ":"... looking up in the dictionary 08:25:59 "oh boy, it's not compile-only, let's execute it" 08:26:03 instruction pointer = found word 08:26:11 jmp execute_loop 08:26:59 now, what's been compiled into :? 08:26:59 a word creating a new dictionary etnry 08:26:59 you missed at least one step 08:27:00 and then a word that does...? 08:27:01 okay 08:27:12 lookup is fine, then modes and flags 08:27:46 1) parser reads into next space ":" 08:28:00 2) looks up in dictionary.... found 08:28:08 well, since we're not in compile mode at the moment (startup) 08:28:08 then i assumed i can skip the STATE check 08:28:09 uh, pseudocode 08:28:12 3) not compile-only, so execute 08:28:13 PARSE-WORD FIND-WORD 08:28:13 DUP IMMEDIATE? NOT ( is the word regular? ) 08:28:13 STATE @ ( what is the state? ) 08:28:14 AND IF 08:28:14 , ( put the word pointer in our current word ) 08:28:16 ELSE 08:28:19 EXECUTE 08:28:21 THEN 08:28:29 4) : executes doing the following: 08:28:36 a) read in to the next space 08:28:39 it's #3 and #4 that will ruin yer day 08:28:43 in the ELSE it should check 'DUp COMPILE-ONLY? IF THROW_ERROR THEN EXECUTE ' 08:29:15 b) create a new dictionary entry the new name just read in from input "foo" 08:29:20 c) change STATE to compile 08:29:50 5) back to step 1 08:30:26 : foo dup * ; 08:30:31 1) read to the next space "dup" 08:30:42 rrr... changing that 08:30:48 : foo 3 * ; 08:30:58 1) read to next space "3" 08:31:05 2) look up in dictionary 08:31:21 3) not found, interpret as numer 08:31:33 4) success converting to number, STATE is compile, so compile a literal for that number 08:31:37 5) back to step 1 08:31:43 Yeah, forth 'state' is just an approximation and folks abuse it 08:31:44 1) read in to next space "*" 08:32:03 2) look up in dictionary... found 08:32:56 3) STATE is compile. If dup is immediate, execute it, otherwise compile a call to it. 08:33:03 4) back to step 1 08:33:11 1) read in to the next space ";" 08:33:19 2) look up in dictionary... found 08:33:31 3) state is compile, word is immediate, so execute it 08:33:38 4) executing ; does the following: 08:33:47 a) compile a return from the word 08:33:55 b) switch STATE back to execute 08:34:32 c) make the new word findable in the dictionary (: allocates space or whatever for the new word, but does not add it to the search lists or whatever) 08:34:34 --- join: nothingmuch_ (~nothingmu@CPE00b0d0cd759c-CM0012253ea774.cpe.net.cable.rogers.com) joined #forth 08:34:54 last message I got: [17:29] PoppaVic: it's #3 and #4 that will ruin yer day 08:34:57 brb, i found some backlogs 08:35:16 d) perhaps otherthings are needed, eg in herkforth, ; flushes the data/instruction caches 08:35:34 np. Yes, the statewise stuff is what I mean 08:35:56 "so compile a literal for that number" 08:36:11 its not that complecated 08:36:29 that means: "write the opcode to push a literal for that number into 'here', and allot as necessary?" 08:36:43 I didn't say it was complicated: I implied it was messy, and I said it was often abused 08:37:02 there are two states. interpret and compile. in interpret mode everything youtype is executed then and there 08:37:10 if your in compile mode its compiled 08:37:22 unless its an immediat word in which case its executed 08:37:30 right, messy 08:37:31 right 08:37:34 immediat words override state. they always execute 08:37:36 --- quit: nothingmuch (Read error: 145 (Connection timed out)) 08:37:42 --- nick: nothingmuch_ -> nothingmuch 08:38:16 but that's still not what I asked... what I asked is what's the difference between 'create 0 ,', which creates an entry, and writes an '0' into an allotted cell 08:38:32 for this example: : foo 3 * ; 08:38:32 1) parser reads into next space ":" 08:38:32 2) looks up in dictionary.... found 08:38:32 3) not compile-only, so execute 08:38:32 4) : executes doing the following: 08:38:33 and the ttype of creation that ':' does 08:38:33 and where do words get compiled into 08:38:34 a) read in to the next space 08:38:37 b) create a new dictionary entry the new name just read in from input "foo" 08:38:40 c) change STATE to compile 08:38:42 5) back to step 1 08:38:45 1) read to next space "3" 08:38:49 2) look up in dictionary 08:38:51 3) not found, interpret as numer 08:38:54 4) success converting to number, STATE is compile, so compile a literal for that number 08:38:56 5) back to step 1 08:38:58 1) read in to next space "*" 08:39:00 yeah, i found that in the logs 08:39:01 2) look up in dictionary... found 08:39:04 3) STATE is compile, word not immediate: compile a call to it. 08:39:05 4) back to step 1 08:39:08 1) read in to the next space ";" 08:39:10 2) look up in dictionary... found 08:39:13 3) state is compile, word is immediate, so execute it 08:39:15 4) executing ; does the following: 08:39:18 a) compile a return from the word 08:39:21 b) switch STATE back to execute 08:39:23 c) make the new word findable in the dictionary (: allocates space or whatever for the new word, but does not add it to the search lists or whatever) 08:39:26 d) perhaps otherthings are needed, eg in herkforth, ; flushes the data/instruction caches 08:40:55 err, that's more than 10 lines 08:40:58 HERKAMIRE @ INSTRUCTION_POINTER_OFFSET + DUP @ -10 + ! 08:41:00 e.g., go back to before your paste 08:41:26 We have 3 potential 'states': preprocessing (immediate), compiling and execution. The forth interactive-compiler tends to blur this. 08:41:48 right 08:41:56 preprocessing isnt PRE prosessing 08:41:57 but that's besides the point 08:41:58 its processing 08:42:24 you have to parse the input and then decide if your going to execute or compile 08:42:26 Further, some forth-variants and words will diddle the states OR they will use words that ONLY work during a specific state. 08:42:31 the data where '0' is written to in a variable is obviously not the same place compilation semantics write calls to words in 08:42:35 yep 08:42:51 yea some forths have "compile only" words 08:43:05 kinda the inverse of immediate 08:43:07 some even have "execute only" words 08:43:08 nothingmuch: words are compiled wherever you want. some forths compile them onto the heap, some have a seperate memory space for compiled code 08:43:34 let my try another approach 08:43:34 there is application for them, isforth has words that should only be used when compiling but i dont enforce it 08:43:50 i dislike calling it the heap 08:43:58 Sure, I already said the whole mess is a grey-area 08:44:24 : CREATE GET-WORD HERE LITERAL ; 08:44:37 err, open a dict entry too 08:44:38 This is why I am looking at a solution using 3 vectors/word 08:44:48 and use LITERAL's evil twin that is not compilation only 08:44:53 thats a bad definition of create 08:45:01 its too obfuscated lol 08:45:04 what forth is that from ? 08:45:06 yep 08:45:08 it's not 08:45:13 I440r: pseudo-code 08:45:15 i was making a guess 08:45:19 aha 08:45:24 you can't guess 08:45:28 try not to shout when you code :) 08:45:29 you CAN pseudocode 08:45:43 nothingmuch: it's not something you can guess. people define these words however they want 08:45:53 but, remember: compiling is all about doing and embedding 08:45:58 the standard doesn't say how the standard words should be defined, just what they do 08:46:04 Herkamire: what does "Joe Random, your typical forth implementation" implement it as 08:46:13 there is no such thing 08:46:14 he can't 08:46:18 they all do it differently 08:46:22 figforth was IT for decades 08:46:33 my forth is so weird you can't really do create 08:46:34 then ANS and shit got involved. 08:46:48 F83 and Fpc were also well received 08:46:58 the ans forth standard does not describe the forth language but a language of the same name 08:47:02 chuck :P 08:47:16 for me, in my universe, gforth is handy, and there are a few more that are sufferable 08:47:27 I440r: agreed 08:48:36 The only real "core" to forth is: 1) It's a TIL; 2) figFORTH was what was THE norm for decades 08:48:41 is my guess at CREATE sort of logical? make a new dictionary entry, write into the code space of the new word an instruction to put the value of 'here' (after the new dictionary entry has been allotted) on the stack? 08:49:08 nothingmuch: that's correct 08:49:14 Again, watch: 08:49:18 see create 08:49:18 : Create 08:49:18 header reveal dovar: cfa, ; ok 08:49:18 phe! 08:49:21 w 08:49:47 create a header, reveal it, compile SOMETHING - in this case 'dovar' and store that cfa 08:49:48 nothingmuch: now the code to accomplish that depends greatly on how your forth works, how you add things to the dictionary, and whether you compile code to the heap or not 08:49:50 i'm guessing 'header' is the 'new dictionary entry', and 'dovar' does the 'write into the code space ... ' 08:50:11 'reveal' is the 'make it findable in the dictionary' part? 08:50:13 yes, noth - and it varies over every single implementation 08:50:32 reveal vs hide 08:50:40 my "create" is slightly more complecated than that because i keep head and list space separate 08:50:47 sure 08:50:57 Herkamire: compiling to the heap, linked list dictionary... everything as simple as it could be 08:51:25 Anyone know if the FIG forth and F79 specs are online? 08:51:25 not really 08:51:28 yes, madwork the figforth.org site 08:52:38 nothingmuch: bear in mind - and this applies to forth or C - write modular shit. This way, you can change the lookup, the stepping, the - whatever. 08:53:05 you can find the 83 standard in various locations 08:53:07 im sure the 79 would be out there too 08:53:30 So far, I have avoided defining a datastack, but I am glaring at the code and trying to ascertain where to best shove it. 08:53:47 f79 was really not nice 08:54:05 f83 was quite decent, and they raged over it for a looong time after 08:54:35 we all of course love ANS forth ;) 08:54:35 * sproingie ducks 08:54:35 eww 08:54:54 I never managed to truly puzzle out ANS forth - it's even more crippled than ansi-C 08:55:52 * madwork likes F83 08:56:01 f83 (and fpc) had some really FINE ideas and some seriously powerful programmers/advocates 08:56:32 I always liked the only/also/previous shit, too 08:58:25 --- quit: sproingie (Remote closed the connection) 08:58:56 iirc, the huge debates over f83/fpc vs ans was the flow-control stuff. 09:00:45 I440r: where was yer site again? I should check it again. 09:01:28 isforth.clss.net 09:01:47 http://isforth.clss.net ? 09:03:28 hmm.. I see three "flavors". Is this not a compilable tarball? 09:08:51 ANyone know if the 'order' list is limited? I was going to use an array, but a LL seems more sensible now. 09:12:44 yes 09:13:02 it's a limited array then, eh? 09:13:06 Interesting 09:13:30 I think maybe Z might have been right, then 09:13:30 you download the tar for your specific system and assemcle and compile it 09:13:32 it probably has a prebuilt binary tho 09:13:53 I grabbed it, but I'm suprised there is no tarball. 09:14:39 as in "generic, universal source" 09:17:15 Sadly, I prolly won't probe it because of the format... I like to peer into tarballs first. 09:17:31 say i"m defining a compile time word 09:17:39 OK. 09:17:49 and I want it to compile, into the currently being compiled word, a call to another word 09:17:57