00:00:00 --- log: started forth/03.07.15 00:00:54 At any rate, Chuck's method of recycling dictionary space is overwhelmingly superior, I've found (remember, I've used PygmyForth [cmForth clone], gforth [pure ANSI Forth], and ColorForth -- ColorForth wins hands down). 00:01:25 The only thing it lacks is some means of restoring the previous MARK-point when some inner context is no longer needed. 00:02:10 inner context? 00:02:16 While not a problem for Chuck, it is a problem for what I want to do. 00:02:27 XeF4: Overlays within overlays. 00:02:56 kc5tja: : green ." ESC[0;32m" ; 00:02:57 : cyan ." ESC[0;36m" ; 00:03:34 * kc5tja nods to onetom 00:04:43 kc5: so you want to compile some temporary stuff to the dictionary and then forget it without losing other MARK points? I thought you could do that.. I suppose I don't understand what you want to do 00:05:20 XeF4: Consider my text editor. 00:05:43 I want my editor to exist as a self-standing application; hence, the first word its source executes is EMPTY, to reset the dictionary to its default state. 00:06:01 After loading the core definitions, it executes MARK, to set the empty-point for editor-specific overlays. 00:06:30 So when I execute the 'd' key, for example, it loads the delete-key block, which first EMPTYs to unload the previous editor overlay, and compiles its own definitions. 00:06:41 So, through time, what we see is this: 00:06:43 cccc 00:06:45 cccceeeeeee 00:06:49 cccceeeeeeeeddddddd 00:06:56 cccceeeeeeeerrrrrrr 00:07:00 cccceeeeeeeffffff 00:07:06 cccc 00:07:16 where 'd', 'r', and 'f' are different editor overlays. 00:07:32 'c' is the Forth core environment, and 'e' is the editor code. 00:08:05 Ignore the extra 'e' in the middle two examples: I can't count. :) 00:08:49 i was wondering about the very same problem last days 00:09:26 kc5: then I don't see the problem unless you want to do like ccceeerrdddff and then drop r without disturbing d and f 00:09:31 but theres nothin problematic w the above strucure, right? the ans MARKER is just suitable 4 it 00:09:32 The only requirement is that the editor must quit gracefully -- it must execute UNMARK (or whatever) to restore the previous MARK point. 00:09:58 XeF4: ColorForth doesn't support his mode of dictionary use. 00:10:09 kc5tja: flux does 00:10:10 XeF4: ColorForth supports MARK and EMPTY, but not UNMARK. 00:10:44 er, right. 00:11:09 onetom: The problem with MARKER is very simple: it's the same thing as FORGET. 00:11:24 It not only recycles up to the marker word, it also REMOVES the marker word. 00:11:37 but the markers are stored SOMEWHERE, so UNMARK oughtn't be more than a half dozen or so words, yes? 00:11:39 This means you MUST redefine it post-haste. 00:11:57 XeF4: MARK in ColorForth updates a global variable. There is no linked list of marks. 00:12:13 no voi perkele 00:12:22 XeF4: I'm sorry? 00:12:42 kc5: yes, that's silly. 00:13:32 could u define mark unmark empty marker & forget? 00:13:35 just roughly 00:13:48 onetom: I truely don't understand that last question. 00:13:57 : mark dp @ markpoint ! ; 00:14:45 : empty markpoint @ dp ! ; 00:14:46 : marker create lastxt , does> @ dp ! ; 00:15:25 : forget ' dp ! ; 00:15:30 See, marker has that fundamental flaw in it -- you need to know what to call the marker word, and it must be parsed from source, which precludes its use in a colon definition. 00:15:58 That, AND it removes the marker word in the process. 00:15:59 not in colorforth 00:16:01 Not useful. 00:16:13 onetom: marker doesn't exist in ColorForth. 00:16:19 editormarker 00:16:25 kc5tja: how about something like : mark here markpoint dup @ , ! ; assuming you never execute mark inside a definition 00:16:46 that way you can just follow the chain of markers back 00:17:13 XeF4: I'd use that, but I'd augment it with a singly-linked list of marks, just in case the dictionary pointer got weirded out somehow. 00:17:16 roughly in b&w4th : marker dp @ literal postpone dp postpone ! ; 00:17:55 onetom: That makes no sense. You're executing marker at compile-time, but with run-time semantics. 00:18:29 XeF4: That's basically how I was going to do it in FS/Forth, actually. 00:18:49 kc5tja: hmm... let me think... bte why should i compile a forget ? 00:19:32 onetom: You generally don't. But I've had need in PygmyForth for compiling FORGET, because I'm *utterly* sick and tired of having to do this: 00:19:36 s/bte/btw/ 00:19:39 : EMPTY 1000 LOAD ; 00:19:41 ( block 1000 ) 00:19:46 FORGET _EMPTY_ 00:19:51 : _EMPTY_ ; 00:20:26 It's a waste of a perfectly good block, and it requires an extra disk access everytime I execute EMPTY. 00:21:03 And heaven forbid I make the mistake of executing EMPTY before actually defining _EMPTY_ for the first time. Pygmy does NOT like that. 00:21:04 :) 00:21:05 huh... i cant understand. where do u use that empty whats defined as 1000 load? 00:21:30 onetom: One question or sentence at a time, please. I'm having a hard time understanding you. 00:21:36 * XeF4 must vanish now. 00:21:43 laters, XeF4 00:21:51 flux has a very good example of the above mentioned technique, and its absolutely not inconvenient 00:21:59 : editor 7 load ; 00:22:05 (block 7) 00:22:20 colored marker 00:22:22 ... 00:22:26 colored 00:22:48 I use EMPTY in my target compiler. 00:23:10 And I'll start using it in VIBE to implement deeper command hierarchies. 00:24:09 * kc5tja doesn't like that approach. 00:25:33 The flux method just doesn't make logical sense to me. 00:25:39 why? 00:26:20 I expect dictionary recycling to occur at the beginning of a program, not at the bottom. 00:27:14 why? 00:27:32 Because that's where my mind conceptually wipes the slate clean. 00:28:27 hmm.. it doesnt seem a very good practice 2 me 00:28:43 a module should cleanup after itself 00:28:49 * kc5tja points out that the original ColorForth didn't have MARK and EMPTY; it had REMEMBER, which is MARKER in ANSI terms. He switched away from it. :) 00:28:56 its a cleaner concept 00:29:55 Nothing . . . let me repeat this . . . NOTHING in Forth cleans up after itself. 00:30:04 Forth is just too primitive. :D 00:31:43 If I were to write my VIBE code like that, I'd have the following structure: 00:31:49 MARKER vibe 00:31:52 ...blah blah... 00:32:19 scr ! ed vibe 00:32:54 I suppose that works, but I'm just not happy with it. 00:33:58 why? 00:34:08 whats wrong w it? 00:34:48 * kc5tja just isn't happy with it. 00:35:07 It doesn't follow from the way I think. 00:35:39 When I see "EMPTY" in the source code, I know, right than and there, that it's actively resetting the dictionary to a known-good state. 00:35:53 MARKER vibe is just noise to me. 00:37:26 empty could also b considered a noise, coz its the default state when u start ur program... 00:37:41 The other thing is the top-level of the editor is "scr ! ed" -- that's it. That is the true end of the editor. Placing "vibe" after that is ... wrong to me. 00:37:53 onetom: Not always. 00:38:12 & what if its not? 00:38:13 It certainly is if you run the editor from a fresh boot of the system. 00:38:20 EMPTY resets the dictionary state. 00:38:25 Like "NEW" in Basic. 00:38:34 NEW 00:38:39 10 PRINT "HELLO WORLD!" 00:38:42 20 GOTO 10 00:38:47 RUN 00:38:52 In Forth: 00:38:53 EMPTY 00:39:08 : hw BEGIN ." Hello world!" CR AGAIN ; 00:39:09 hw 00:40:39 yes. and? 00:41:11 What do you mean "and?"? You're the one who asked me for my rationale. I gave it. 00:41:20 The ball is in your court now. 00:41:45 u can have empty, but w the modules, its a more convenient 2 make the selfcleaning 00:41:55 But it's *not*. 00:42:07 I would argue that it's different, but it's VASTLY more convenient to use EMPTY. 00:42:31 Consider how ColorForth works: when running a program, it wedges itself into the system state -- it sets up keyboard handlers, display tasks, background tasks, etc. 00:42:47 but empty is a hardwired thing isnt it? it always resets 2 the same position 00:42:51 When dropping to the command-line, these tasks are still running! 00:43:14 onetom: No. You can change where EMPTY empties to by using MARK. 00:44:08 aham. look: 00:44:23 : module mark ; : end empty ; 00:44:30 module ( editor) 00:44:31 ... 00:44:32 end 00:44:49 * kc5tja cries!!!!!!! 00:44:52 You just don't get it. 00:44:52 isnt it clear? 00:44:57 YES, it's clear. 00:45:03 yeah, that must b the problem :) 00:45:04 NO, it's not why I'm objecting to MARKER. 00:45:49 The point is, I can preemptively *terminate* a running program at *ANY* time, and LOAD another program block at a WHIM. 00:45:57 EMPTY is the **ONLY** solution that can gracefully handle that scenario. 00:46:00 btw, im gonna ask u a related question, so u can cry but dont run away plz. not yet @ least ;) 00:46:43 aham... 00:46:52 Please don't "ahem" -- it's very rude. 00:47:11 flux solve this problem w branches, i suppose 00:47:26 All Forths have branches. 00:47:32 I don't see hwo this is at all relavent. 00:47:46 kc5tja: its not ahem, its aham, what means i c i got it 00:48:09 its similar 2 the hungarian ühümm ;) 00:48:30 kc5tja: :) not that branch i was refering 00:48:53 onetom: Well, still, in English, it's rather rude. :) 00:49:06 It's a signal of indignation. 00:49:09 i cant fully explain how does flux branches work, but let me try: 00:49:14 OK 00:49:33 u say sg like : ' forth branch 00:49:57 no. hold on, im gonna chk the sources 00:50:50 It looks like a computed goto to me. 00:51:25 no no no, branch is a dict manipulation word like mark & things like that 00:51:27 hold on 00:53:07 http://sec.dunasoft.com:9673/wiki/forth/source.html 00:53:14 wget & links it! 00:53:27 & find branch 00:53:31 I just googled. 00:53:37 It's basically exactly the same thing as MARKER. 00:54:03 not really 00:54:05 The only difference is it doesn't physically recycle dictionary space. 00:54:12 it accepts an address 00:54:43 and limits searchg 4 word 2 definitions defined before that point 00:55:27 and limits searching 4 words 2 the definitions defined before that address 00:55:44 OK, but that doesn't solve the aforementioned problem. 00:55:46 Consider: 00:55:50 : ENDLESS BEGIN AGAIN ; 00:55:52 ENDLESS 00:55:55 * onetom considers ;) 00:56:00 yup 00:56:09 In nearly every Forth system I've used, ENDLESS will never, ever return. Only a reboot will break the loop. 00:56:16 (or some similar drastic action, like kill -9) 00:56:41 But in my Forth environment, the user interface always has priority over the running application. 00:56:51 F-keys are used to switch between applications at the user's whim. 00:57:19 If the above code was running, hitting F1 would *immediately* terminate the execution of ENDLESS and load a block containing the command-line interface for FS/Forth. 00:57:29 So, if we have: 00:57:32 MARKER endless 00:57:35 : ENDLESS BEGIN AGAIN ; 00:57:39 ENDLESS endless 00:57:50 we'd waste dictionary space -- the marker word never, ever, ever gets executed. 00:57:54 However: 00:58:04 if we precede each program with EMPTY, it is not an issue. 00:58:11 so, u cant continue a stopped task? 00:58:43 I just told the foreground task to switch to another program. The task never stopped -- it's still running -- in a different program. 00:58:43 u just have 2 exit from your commandline interface so u get back 2 ENDLESS 00:58:57 uhhh 00:59:02 uhh 00:59:04 uhhh 00:59:09 second.. 00:59:15 let me understand 01:00:33 i should see the full example 01:00:48 what 4th environment r u talkin about? 01:00:53 Do you understand how a CPU's cache works? 01:00:57 what os does it requires? 01:01:01 Does it matter? 01:01:06 sure 01:01:10 ColorForth is one, FS/Forth will be another when I'm done with it. 01:01:11 id like 2 try it 01:01:23 Chuck Moore's ColorForth is a real-world example. 01:01:36 I use XcolorForth under Linux. 01:01:41 There are variants available for Windows now too. 01:02:48 i use linux 01:04:04 * onetom tries XcolorForth 01:07:14 You're going to be in for some serious culture shock, especially if you're not used to Chuck's 27-key keyboard. :) 01:08:22 really, it knocked me off ColorForth 01:08:34 ayay... 01:08:51 i managed to run Win version, but i was compleetly lost at kbd 01:08:52 it has ruined my X 01:09:25 onetom: ??? 01:09:26 it doesnt runs remotely :) 01:09:30 run 01:09:37 but locally it works 01:09:51 Yes, it uses SDL for frame buffer access. 01:10:04 me c 01:10:08 Serg_Penguin: I like it actually. But then, I love Dvorak keyboard layout. 01:14:41 my thingers stuck to qwerty 01:15:04 now i play w/ speccy emulator, and have to look up the chart all the time 01:16:27 kc5tja: i cant start xcf in fullscreen, so i cant see the commandline 01:16:54 could u give me a kickstart 2 chucks4th? 01:17:32 onetom: Do you have the latest SDL libraries installed? 01:17:39 thinks so 01:17:43 i use woody 01:17:54 What screen resolution are you running at? You may have to edit the source and recompile to get a compatible screen size. 01:18:02 1024 01:18:05 It's set up for 1024x768 by default. 01:18:11 brb - food 01:18:24 but the windowmanagers taskbar.... 01:18:34 OK, I would reduce the size of the framebuffer by a small amount then. 01:18:42 On my system, I use 1000x700, for example. 01:19:01 k 01:19:14 Edit main.c -- there'll be a line that calls SDL_SetVideoMode(). 01:19:30 Change the resolution figures there. 01:19:55 :( it segfaults... :/ 01:20:00 Then, once that's done, edit the color.s file -- there will be two assembler macros called hp and vp, which have similar meaning. 01:20:12 What segfaults?? 01:20:21 ok 01:20:34 i havent changed hp vp yet 01:20:42 Ahh, that might be why. 01:20:59 Try that, and if it segfaults, then I'd talk to the folks on the ColorForth mailing list. 01:22:18 works, thx 01:22:34 * kc5tja wipes sweat off his forehead... *whew* :) 01:22:37 I was worried there for a minute. 01:23:36 gosh, im using hungarian kbd layout :) 01:26:20 It's actually Dvorak layout. 01:28:21 ok, i was just afraid it doesnt handle the keys by "scancode" 01:28:29 but its okay 01:28:42 * kc5tja nods 01:28:52 * onetom is lookin 4 chuck kbd tutorial 01:29:08 Yeah, it takes some getting used to. 01:30:17 but 01:30:27 until i master the layout 01:30:40 how would u solve the following problem 01:31:05 id like 2 build some words programmatically 01:31:22 & after that id like 2 forget the builder program 01:31:35 consider the following example 01:33:26 : cycles ( n-) create [ 4 ] literal compile * compile do-something ; 01:33:33 5 cycles 5c 01:33:40 20 cycles 20c 01:34:05 & id like 2 forget the generator word: cycles 01:34:19 just leaving the generated 5c & 20c words 01:35:55 You can't. 01:36:49 But you can do something similar with compiler words: 01:36:56 (using punctuation instead of colors) 01:36:58 MACRO 01:38:09 : cycles ( n- ) [ 4 ] literal compile * compile do-something ; 01:38:10 FORTH 01:38:19 : 5c [ 5 ] cycles ; 01:38:24 : 20c [ 20 ] cycles ; 01:38:34 But, then, you still have the defining word lying around. 01:38:42 SO, to fulfill your precise requirements, you can't. 01:38:43 yup 01:38:48 The definer word has to be there, always. 01:38:55 hm... 01:39:36 i was thinging of a separate marker 4 the forth & the macro "wordsets" 01:40:07 but probably i dont wanna define just macros 01:41:26 ColorForth imposes some pretty severe word count limitations too. Max: 128 defined macros, and 512 Forth words. :) 01:41:46 You can change that by recompiling of course, but that's how it's configured out of the box. 01:41:50 nice 1 :) 01:46:45 Well, I need to get to bed. 01:46:48 1:49 AM here. 01:46:52 haha! the RX-7 is victorious 01:46:59 a7r: What's up? 01:47:08 just went racing 01:47:16 At this god-awful hour of the night? 01:47:23 beat up on my friend's 2nd-gen eclipse 01:47:37 :) 01:47:48 kc5tja: good night&thx 01:48:54 So you're still not controlling the boost by controller yet -- you're still just collecting data, right? 01:51:20 Anyway, we can chat tomorrow. I have to get to bed. 01:51:27 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 01:58:39 --- quit: Serg_Penguin () 03:56:19 --- quit: Robert (Read error: 104 (Connection reset by peer)) 03:58:59 --- join: Robert (~snofs@h126n2fls31o965.telia.com) joined #forth 04:14:40 --- quit: a7r (Read error: 110 (Connection timed out)) 04:19:55 --- join: Serg_Penguin (Serg_Pengu@212.34.52.140) joined #forth 04:45:24 --- quit: Serg_Penguin () 05:11:23 --- join: Serg_Penguin (Serg_Pengu@212.34.52.140) joined #forth 05:27:30 --- join: w1k1_ (~w1k1@pD9EE18AE.dip.t-dialin.net) joined #forth 05:36:02 --- quit: w1k1 (Read error: 60 (Operation timed out)) 05:43:52 --- quit: Serg_Penguin () 05:56:07 --- join: mur (murr@baana-62-165-186-227.phnet.fi) joined #forth 06:02:42 --- join: Robert__ (~snofs@h126n2fls31o965.telia.com) joined #forth 06:03:38 --- quit: Robert (Read error: 60 (Operation timed out)) 06:42:28 --- nick: Robert__ -> Robert 07:09:33 --- join: flyfly (~marekb@ip164.ktvprerov.cz) joined #forth 07:30:00 --- quit: flyfly ("using sirc version 2.211+KSIRC/1.2.4") 07:32:26 --- join: xylzan (~root@bzq-200-115.red.bezeqint.net) joined #forth 07:51:39 --- quit: mur (Read error: 54 (Connection reset by peer)) 09:04:12 --- join: wossname (wossname@HSE-QuebecCity-ppp82077.qc.sympatico.ca) joined #forth 09:11:53 --- quit: xylzan ("xylzan has no reason") 09:24:11 --- join: flyfly (~marekb@ip164.ktvprerov.cz) joined #forth 10:04:24 --- join: mur (jukka@baana-62-165-189-64.phnet.fi) joined #forth 10:18:48 --- join: ASau (~asau@158.250.48.196) joined #forth 10:19:16 Dobryj veczer! 10:19:22 privet aleksei 10:19:45 what's the dorobla name of aleksej, like vanja is for ivan 10:19:45 ? 10:20:11 Ljosha. 10:20:31 sergei -> serjuska? 10:20:34 It's called "umen'shitel'no-laskatel'noje imja" 10:20:48 Serjozha. 10:20:56 near 10:21:04 no wonder i didn't remember the expression :) 10:21:18 Are you studiing Russian? 10:21:26 i was 10:21:31 been 2 years now, though 10:22:08 I remember. I mean, are you recalling your knowledge now? 10:22:47 A week ago I've found Czeck tutorial. 10:23:13 It's very interesting to read. 10:23:13 not really 10:23:17 ahaa 10:23:19 aga 10:23:22 gaf gaf 10:23:30 what was dog? 10:23:41 shto byl "gaf gaf" ? 10:23:42 :) 10:24:01 Especially in case when you're searching for Tatar one. 10:24:07 shto eto gaf gaf? 10:24:15 tatar one? 10:24:18 It's a dog, of course. 10:24:42 what's dog in russian? 10:24:59 i meant to ask po-russkij what's word for dog :) 10:25:06 Recently, I've run into one interesting discussion about Bulgaria (Volga one)... 10:25:08 not who says bark bark 10:25:09 :) 10:25:17 dog is "sobaka" or "pjos" 10:25:27 sobaka! da-da! 10:26:03 da, eto sobaka! sobaka snaju shto eto "gaf gaf"! 10:26:42 Now, when reading Caucasus map, I recall some Malqar words. :) 10:26:51 %) 10:27:40 shto bylo priglasenije? 10:27:46 excelent? 10:27:59 Knowing too many languages is very insteresting, now I see. 10:28:13 priglashenija? 10:28:19 Priglashenie = invitation. 10:28:23 ahhmmm 10:28:29 it just popped to my mind 10:28:41 2 years without russian does its tricks 10:29:17 Yes, you forget language faster than learn it. 10:29:52 what was party in russian? 10:30:04 what's invitation to party? 10:30:11 --- quit: Robert (Remote closed the connection) 10:30:18 knowing russian is actaully a bit useful 10:30:25 I try to find time for reading some german books, but can't. 10:30:33 you can understand quite ok slovak, checz and polish pages too 10:30:43 There's too many variants. 10:30:52 ? 10:31:00 Invitation to party can be: 10:31:31 a) priglashenije na veczer (also "veczjorka", rare word); 10:31:44 b) ... na prazdnik 10:32:23 Other variants are possible. 10:33:04 mur, You're quite a good liguist. 10:33:35 I know many people who can't understand how I read ceska and slovak. 10:35:10 Czech. "Ceska" is "Czech" in Czech. 10:35:55 --- quit: wossname () 10:36:00 Hm... 10:36:25 It seems that we have no straight equivalent of English "party". 10:36:54 veczer, eto harasho 10:37:11 OTOH, there is "sobranije" or "sborishche". 10:37:16 veczer is like "fest" 10:37:42 Fest is like prazdnik. 10:38:12 Veczer is more quite party. 10:38:21 It's like "a cup of tea". 10:38:33 ;) Or 10 cups. :))) 10:39:38 I've recalled one story I was told in train "Saratov--Moskva"... 10:40:04 --- join: wossname (wossname@HSE-QuebecCity-ppp81945.qc.sympatico.ca) joined #forth 10:40:08 saratov is in south right? 10:40:09 caucasia 10:40:13 or? 10:40:19 In Ivanovo region, in village, a man heard a phrase: 10:40:48 "O! Saratovskije prijehali. Sejczas czaj pit' budut." 10:41:03 Saratov is to south-east. 10:42:07 Saratov is just in opposite direction from Moscow than Leningrad. 10:42:17 St.-Petersburg. 10:43:34 what is czaj? 10:43:38 Tea. 10:43:42 yes 10:43:48 --- join: Robert (~snofs@h126n2fls31o965.telia.com) joined #forth 10:43:51 i had idea in brain that it coudl be tea or 7 10:44:05 "oh saratov went and tea came instead" 10:44:05 Privet, Robert! 10:44:27 robjushka! 10:45:54 "Oh. Saratov (men) has come. They are going to drink tea now, of course." 10:46:23 It is unable to omit unnecessary words in English. 10:46:33 saratovians 10:46:41 of course 10:46:41 It's the main difficulty for us. 10:46:46 i shoudl have paid more attention 10:46:53 (Russian accent. :) 10:47:07 Yes. Saratovians. 10:47:10 pit to drink 10:47:17 czaj tsai! 10:47:21 tsai! 10:47:21 yes! 10:47:26 now it makes more sense 10:48:01 It's too hard to transliterate Cyrillic into Latin. 10:48:21 Too little characters :] 10:48:50 how about kanji instead? :) 10:49:07 I can't understand how to transliterate "shch" better. 10:49:17 It's interesting idea. 10:49:31 hmm 10:49:33 Maybe I'll try to learn Kanji. 10:49:34 i have literation 10:49:36 here somewhere 10:49:44 kanji has only 200 characters :) 10:50:18 OTOH, what for do I need so many letters? 10:50:27 kanji has more than that :( 10:50:28 well 10:50:31 240 10:50:32 iirc 10:50:33 I'm pretty well with my 33/ 10:50:37 a few thousand 10:50:41