00:00:00 --- log: started forth/03.04.03 00:31:00 God bless loeliger ;-) 00:32:43 4559 bytes: and Speuler has a TIL ;-) 00:35:36 hahaha! stripped, the binary is only 4404 - Gofd, I love it ;-> 00:37:01 sifbot: words 00:37:02 PoppaVic: ." .( : if then do loop repeat until exit ... 00:37:39 sifbot: " : test 30 0 do ." +" loop ; 00:37:41 PoppaVic: Word not found: " 00:38:14 sifbot: : test 30 0 do ." +" loop ; 00:38:15 PoppaVic: 00:38:19 test 00:38:24 sifbot: test 00:38:26 PoppaVic: Word not found: test 00:38:58 sifbot: : test 30 0 do ." sifbot's an IDIOT" loop ; test 00:39:00 PoppaVic: sifbot's an IDIOTsifbot's an IDIOTsifbot's an IDIO 00:39:11 silly machine. 00:42:47 Speuler: my bad... delete the first "debug("}COLON ", VM);" on line 93 01:24:06 damn this places dies 01:38:54 --- join: fridge (meldrum@zipperii.zip.com.au) joined #forth 02:01:01 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth 02:32:45 --- quit: ramnull ("This isn't Happy Hour!") 02:44:31 * PoppaVic idolates 03:18:48 too early.. snooze-time.. 03:18:50 --- quit: PoppaVic ("(I don't need a reason)") 04:57:04 --- join: deluxe (~deluxe@pD9E4ED4D.dip.t-dialin.net) joined #forth 04:57:22 salut 04:57:33 hello sifbot 04:57:44 sifbot: hello 04:57:46 deluxe: Word not found: hello 04:59:36 sifbot: :hello ." welcome" . ; 04:59:37 deluxe: Word not found: :hello 04:59:45 sifbot: : hello ." welcome" . ; 04:59:47 deluxe: 04:59:53 sifbot: hello 04:59:55 deluxe: Word not found: hello 05:00:43 sifbot: : hello 1 2 3 + + . ; 05:00:44 deluxe: 05:00:56 sifbot: hello 05:00:57 deluxe: Word not found: hello 05:02:02 doesn't want to remember user defined words 05:02:38 or he doesn't like newbies ;-) 05:08:02 fridge: futhin set up the hp, right? 05:11:37 --- join: gilbertdeb (~gilbert@fl-nked-ubr2-c3a-29.dad.adelphia.net) joined #forth 05:15:21 onetom: For what? 05:15:49 hi Robert 05:16:21 Hi gilbertdeb :)( 05:17:44 whats news in #forth ? 05:17:49 futhin set up the hp, yes? 05:18:26 deluxe: I think so, yes. 05:19:27 because there's a .ca webmin mail adress... 05:19:55 :) 05:19:57 did you discuss the page recently? 05:20:11 Yes, we were talking about what futhin though about us :P 05:20:27 04:15:32 < Robert> "robert 05:20:27 04:15:33 < Robert> always on irc, very condescending, should be banned ;) website" 05:20:34 He's so nice to me! 05:20:45 awwwww 05:20:51 LOL 05:24:52 deluxe: were you supposedly doing something about a forth chip? 05:24:55 or did I read wrongly? 05:29:04 gilbertdeb: i'm very interested in this subject, yes. why? 05:29:16 ah how does one get them? 05:29:31 better yet, whats your interest? 05:30:16 gilbertdeb: x25 not available yet... 05:30:30 what will you do with it? 05:31:15 built an exciting new mobo 05:31:35 kewl 05:32:24 have you built a mobo yet? 05:32:33 or will this be your first project? 05:33:23 if it happens at all... (what about you?) 05:33:52 never delved in electronics... 05:33:57 but I am suddenly getting interested. 05:34:28 I was browsing through the 6502 pages when I realized with a few reads of TFM, I might actually build something 'interesting'. 05:34:49 more like surfing through. 05:36:26 gilbertdeb: never built a mobo b4, more dreaming about 05:36:58 gilbertdeb: and a forth newbie 05:39:35 gilbertdeb: but there's a strong potential 05:40:45 so when are you getting the x25? 05:41:12 gilbertdeb: but if we get 100000 bought cpu's together, dunno 05:41:32 what do you mean? 05:43:42 I mean, what do you mean by "100,000 bought cpu's" 05:44:02 a reasonable production minimum is 100.000 cpu's, if there are enough buyers, we could have the x25 in some months 05:44:25 oh. 05:44:37 thats gonna be tricky. 05:44:57 even if they each sold for a buck, there is no way I can fork over that much. 05:47:32 of course, several hundred forthers are needed... 05:50:26 if they're 1 buck each, i take 19 05:50:30 10 05:50:32 or 20 05:51:32 would pass one to poor gilb 05:52:19 Speuler: '10 or 20' good idea 05:52:55 but no way they're going to be in that price range 05:53:41 :) 05:54:06 I am interested. 05:54:11 a 25'er cpucluster per mobo? 05:54:13 as long as it is not too expensive :) 05:54:44 Speuler: have you worked with the 6809 and the amd 29k? 05:54:58 6809 yes 05:55:00 29k no 05:55:10 who made the 6809? 05:55:15 motorola 05:55:24 bah. should have seen that. 05:55:32 they made all the 68xx series 05:55:35 and the 68xxx 05:55:39 forgot all about that. 05:56:48 there are several emulators, running under linux 05:56:49 Speuler: how much do you estimate 1/100.000 x25 will cost? 05:57:13 estimate 50 $ 05:57:46 seriously? how come? 05:58:02 methinks under $10 05:58:04 because you asked me to 05:59:30 i think $5 is possible, but i'm a newbie 05:59:56 hmm. novix was put on gate array, not pulled up silicon. was (4000) gates about 100 $ . don't know about numbers though 06:00:15 x25 probably more complex to produce 06:01:00 the 100.000 mass make the price 06:01:09 f21 estimate was iirc "as slow as 2 $" if produced in quantity 06:01:19 as low ... 06:02:02 the novix wasn't produced in taiwan, or? 06:02:37 nope 06:02:54 in the $5 are 2 sample runs included 06:03:06 'more complex' 06:03:21 I was reading some place that the z80 and the 6502 cost all of 2-7 bucks. 06:04:37 i don't think cm will demand 'fat profit' 06:05:56 but those chips are very old so thats prolly why. 06:06:23 almost 22 years old!! 06:06:31 I suddenly feel young. 06:07:03 gilbertdeb: 'old' that's why i say 'fat profit' 06:07:26 I mean the z80,6502,6809 are old. 06:07:34 Speuler: what was the 6809 like? 06:07:39 risc? cisc? none of the above? 06:07:40 nice 06:07:48 cisc 06:07:57 hmmm. 06:08:05 I am getting a headache already. 06:08:06 best non-forth cpu to write forth in 06:09:37 interesting is, how to do the cpucluster on the mobo 06:10:00 deluxe: are you interested in building a cluster? 06:10:06 I think it would be a neat project!!! 06:10:11 'f course you need an x25 as keyboard controller. :) 06:10:12 a diy cdc 6600 :) 06:10:40 and one for serial i/o 06:10:56 wait I didn't get that? 06:11:07 you need a keyboard controller IC and a something else io? 06:11:26 don't think of a 'normal' cluster yet 06:12:03 i meant clustering cpu's on the mobo 06:13:02 Thats what I thought you meant. 06:13:10 what did you think I thought you meant?> 06:13:37 clustering boxes to a supercomputer... 06:15:25 is there somebody here, who has experience w/ mainframes? 06:15:25 Speuler: which available chip do you recommend for messing around with first? 06:15:39 * gilbertdeb points at speuler. he prolly does. 06:15:59 the A/390 ? 06:17:06 i started learning about computers on a ibm 370 06:17:31 but it didn't take long to become more interested in smaller machines 06:18:01 Speuler: about how fast would you say the 370 is? 06:18:57 it can multiply two numbers quicker than i can 06:19:09 heheh. 06:19:14 relative to machines of today. 06:19:27 --- quit: onetom (Remote closed the connection) 06:19:32 difficult to estimate 06:19:46 i'd say, 50...100 mhz pentium 06:19:55 :O 06:19:57 that _Fast_? 06:20:01 hm 06:20:09 thats not bad at all!!! 06:20:27 ibm 3900 could search primary numbers as quickly as novix could 06:21:29 Speuler: what would you say the advantage of mainframe arch is? 06:21:41 are there things modern desktop/deskside/whatevers can't do that the 370 still does well? 06:21:51 eg warming a small apartment? :) 06:21:58 touche deluxe 06:22:07 their rest value in metal weight is higher 06:22:22 hahaha. 06:22:23 LOL 06:22:48 meanwhile, general-purpose cpu have mainframe architecture 06:22:56 x386+ for example 06:23:01 Speuler: how much time did you spent on mainframes? 06:23:08 memory management unit 06:23:17 was inspired by maneframes 06:23:24 2 weeks 06:23:25 :) 06:23:34 hmmm. 06:23:57 but the 386 machine did not do as much heavy lifting as mainframes were designed to do! 06:24:08 so thats where the whole CISC thing comes from is'nt it? 06:24:38 yes there's something interesting about memory mgmt 06:26:05 how about buying old laptops from ebay and gutting them out? 06:26:12 reason why i didn't like learning about mainframes was their lack of interactiveness 06:26:21 batch processing? 06:26:22 risc is more misc as cisc, no? 06:26:36 john is more jane as joan, no? 06:26:45 and when i was exploring another machine, with built-in interpreter, i just forgot about the mainframes 06:27:08 deluxe: right 06:28:20 that's interesting for me, but how to specialize the cpucluster parts? 06:28:31 how? 06:28:43 think 'recursion'. 06:28:49 think if n == 0. 06:28:51 what next? 06:29:17 eeproms are too slow, no? 06:29:26 why are they too slow? 06:29:43 relatively... 06:29:45 ??? 06:29:55 'are they expensive for diy' is my question. 06:30:15 erm, diy=? 06:30:23 do it yourself. 06:31:01 --- quit: fridge ("http://lice.codehack.com") 06:32:11 gilbertdeb: in longterm probably not 06:32:25 Speuler: eeproms are too slow, no? 06:33:22 what is "too" ? 06:33:30 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 06:33:39 they are slower than ram 06:34:27 flash is quicker 06:34:39 and easier to write to 06:34:49 easier to erase 06:35:11 ignore the last 06:35:18 eeproms are easier to erase :) 06:35:40 well. 06:35:43 about the same 06:36:06 depends on what you want to erase 06:36:07 Speuler: flash can equal which max. ram speed? 06:36:20 yes, about, for reading 06:36:28 writing is way slower 06:36:28 deluxe: got a need for speed? 06:36:54 and you can't overwrite a location without erasing it first 06:37:08 (if you want to set a bit which was zero) 06:37:10 perhaps you might be interested in this n>1 Ghz machine with n>1 Gb ram module :) 06:37:51 a pentium 2 ghz with 4x256mg dram ? 06:38:10 n>=1 :) 06:38:23 I stubbornly refuse to buy a ghz machine. 06:38:25 4x256 + video ram 06:39:23 there are a few things i can't do with a slower machine 06:39:41 like video stuff? 06:39:45 right 06:39:54 get an sgi! 06:40:09 can't fit my 50$ framegrabber in ther 06:40:11 there 06:40:26 its a physical device? 06:40:41 yes. pci card 07:04:44 don't have hardware compression though 07:05:10 no real need with a reasonably quick cpu 07:05:31 that machine is still slow, compared to toadys standard 07:05:38 1.8 g celeron 07:06:15 compresses about does about 22 frames/sec 768x576 07:06:27 almost realtime 07:08:10 (company i come out for right now makes presentation systems. getting spots and presentations into the machine from cam is important) 07:08:58 everything is running under linux 07:09:23 i have liberated most machines here by now 07:09:45 and, guess, people really like it 07:11:04 that's actually the reason why i joined, when they were contemplating rebuilding their sytems, using linux 07:11:29 but of course i didn't stop at the presentation system 07:11:54 there's now exactly one computer remaining, running win 07:12:21 and that one runs with xserver software :) 07:16:42 which linux would you recommend for a 500mhz/300mb box? 07:19:14 debian 07:19:30 maybe gentoo 07:22:45 but to compile gentoo would take ages, no? 07:23:25 takes longer than installing debian, yes 07:25:45 never tried gentoo, someone said for compiling/installing gentoo there should be at least a 1ghz box there (for a reasonable install time) 07:26:27 compiling a kernel under debian takes about the same time 07:26:40 as compiling one under gentoo 07:27:11 but with gentoo, all apps you're installing are compiled too 07:27:33 but as a result, the system should run a bit quicker 07:34:19 yes. have you compiled gentoo already on a 500mhz box? 07:34:28 nope 07:39:10 --- quit: deluxe (Read error: 54 (Connection reset by peer)) 08:00:54 --- join: deluxe (~deluxe@pD9E4E602.dip.t-dialin.net) joined #forth 08:54:57 --- quit: deluxe ("Client exiting") 09:10:58 --- join: Serg_Penguin (~z@ppp143-186.dialup.mtu-net.ru) joined #forth 09:11:09 --- quit: Serg_Penguin (Client Quit) 09:31:11 --- join: I440r (~mark4@1Cust29.tnt3.bloomington.in.da.uu.net) joined #forth 09:53:39 --- join: tathi (~josh@wsip68-15-54-54.ri.ri.cox.net) joined #forth 09:54:41 hi 09:54:47 just released 1.12b 09:55:13 Cool :) 09:55:15 What's new? 09:55:28 fixed a realy silly bug in the text widnowing code that was making me set the cursor location prior to the output of EVERY SINGLE character 09:55:41 rewrote the args handling code. i now consider this to be sane :) 09:56:11 ive included an example forth script file "cat.f" that works sort of like the unix cat command 09:56:26 cat.f somefile writes all of "somefile" to stdout :) 09:56:36 cat.f file1 file2 file3 > file4 09:56:38 etc 09:56:47 very small :) 09:57:06 How do you invoke it? isforth - float cat.f file1 ... ? 09:57:16 no 09:57:18 ./cat.f blah 09:57:22 Ah 09:57:26 its a shebanged script file 09:57:31 OK :) 09:57:36 and we no longer use - fload 09:57:42 ./isforth -fload blah.f 09:57:45 :) 09:57:49 I'll go download it. 09:58:25 tell me if you have problems. last time i updated i forgot to also update the link to the new bzip heh 09:59:29 Uhm 09:59:30 404 :) 10:00:11 whats 404 - the bzip or the page ? 10:00:53 bzip 10:03:23 ok try now. i had a typo :) 10:03:42 #forth.tec.ru doesnt work for me 10:03:54 erm http://forth.tec.ru even 10:04:04 lol 10:04:15 htt://forth.tec.nu 10:04:20 That's what you wrote :) 10:04:26 Not http://forth.tec.nu 10:04:32 .ru != .nu btw :P 10:05:07 erm. i mean .nu 10:05:12 is it at .nu ? 10:05:18 Yeah. 10:05:20 thats what i have in my link and its not working :P 10:05:24 And http, not htt :) 10:05:38 that would be it :P 10:06:31 You should check that the page is correct... otherwise it's like releasing 1000 lines of totally untested Forth code. ;) 10:06:41 bleh :/ 10:06:44 i dont do html 10:06:45 lol 10:06:55 what is the exact url for this channels web page 10:06:59 i STILl aint gettin it 10:07:16 * Robert points at topic 10:07:21 the #forth website http://forth.tech.nu | native x86 linux forth http://isforth.clss.net 10:07:35 its .tecH 10:07:35 grrr 10:07:45 futhin emailed me forth.tec.nu dammit 10:08:35 Another reason you should try the URLs.. don't trust junkies like futhin 10:09:07 --- join: deluxe (~deluxe@p508050D6.dip.t-dialin.net) joined #forth 10:09:23 kc5 is listed in the user list on the page - is he back here as a regular now ? 10:09:28 * I440r hopes so! 10:09:31 No :( 10:09:46 I've seen him a few times here though 10:09:51 Remember why he left? 10:09:56 Robert always on irc, very condescending, should be banned ; 10:09:59 heh 10:10:23 Yeah... I'll get that bastard! 10:10:36 :) 10:14:30 --- quit: I440r ("Reality Strikes Again") 10:14:42 --- join: I440r (~mark4@1Cust29.tnt3.bloomington.in.da.uu.net) joined #forth 10:15:15 oopts 10:16:46 Wrong button? 10:17:02 :/ 10:25:53 ill bbl 10:25:55 --- quit: I440r ("Reality Strikes Again") 10:26:33 --- join: krish (KRISHNAKUM@61.1.220.144) joined #forth 10:26:44 hello! 10:27:23 Hej :) 10:29:52 --- part: flyfly left #forth 10:42:50 --- join: PoppaVic (~pfv@s88.waters.gtlakes.com) joined #forth 10:43:29 Well, what have I been missing. 10:45:05 'evening 10:45:06 PoppaVic: hi! 10:45:21 Speuler: hi! 10:45:32 krish ! 10:45:57 Hi.. Damnit, Speuler - I sent the solution.. But it bounced. 10:47:16 dcc Speuler 10:47:44 hmm 10:47:50 just looking at the mail log 10:48:04 there's none which has bounced 10:48:05 The source was 4456 chars, the stripped code was 4404 ;-) 10:48:25 well, do yer dcc - same file, sans chrotling note. 10:48:38 dcc won't work 10:48:43 shit. 10:49:03 firewall i guess 10:49:09 right 10:49:23 Speuler@forthfreak.net should do 10:49:40 or try Speuler@logilan.info 10:49:44 l@sbin.de 10:50:37 ah 10:50:44 there's one bounced mail 10:50:49 pfv 10:50:51 I'll try to reply to yer emai 10:50:57 that's you 10:51:03 well, sure - that's my id at the isp 10:52:05 incoming... 10:52:12 splash. 10:53:43 I believe that file should work fine for you - even the asm-output loked clean. 10:53:49 loOked 10:59:21 Well, did it get thru to any of those three? 11:05:33 --- join: Stepan (~stepan@likes.smoking.and.watching.spacenight.dk) joined #forth 11:05:36 hi there 11:05:57 Hi :) 11:05:59 lo 11:07:29 Speuler - yo! One of those three should have arrived. 11:07:40 --- join: Gnomon (~Gnomon@CPE000103c34069-CM014300001653.cpe.net.cable.rogers.com) joined #forth 11:08:43 i got two 11:08:58 cool 11:09:19 sandwich probably didn't work 11:09:41 lets look at some applied magic 11:09:56 as long as the code makes you smile ;-) 11:10:20 i recognize nothing anymore 11:10:33 no, I went for broke and made it work. 11:12:24 hmmm ... 11:12:28 segfault ... 11:12:52 whoa.. 11:13:02 IT shouldn't - built & compiled fine here. 11:13:14 ah 11:13:21 a.out form april 2 ... 11:13:26 from 11:13:43 oh.. "gcc -o foo fooforth.c" ;-) 11:14:59 pointers are not permitted as case values 11:15:06 is that error or warning ? 11:15:30 I ignored it - any warnings/errors that bug you, I'm sure I can resolve. 11:15:34 doesn't say error or warning but gives no executable 11:16:20 stop using warnings on the voodoo - just use the silly gcc above 11:16:41 no executable with -o name , no a.out without -o 11:17:00 what the heck os/compiler are you suffering? 11:17:01 Speuler: Poppa: what are you guys talking about ? 11:17:19 kirsh - an inner-interpreter for a TIL 11:19:37 Speuler: 11:19:40 gcc -Wall -o foo fooforth.c 11:19:40 fooforth.c: In function `interp': 11:19:40 fooforth.c:117: warning: label `DONE' defined but not used 11:19:40 fooforth.c:104: warning: label `RUN' defined but not used 11:19:40 fooforth.c:77: warning: label `TOKENS' defined but not used 11:20:13 PoppaVic: using the local labels mechanism ? 11:20:35 well, I didn't use the "computed goto's" feature, although I wanted to. 11:21:16 why ? 11:23:10 I had no need for them. It's not like I can call into that 'vector' from outseide. 11:24:09 guess i gonna downgrade my gcc again 11:24:54 The 3 series was always nothing but trouble for me. 11:25:19 i gto 2.95 on the machine too 11:25:26 2.95.3 was pretty good. 11:25:39 probably just the link gcc to gcc-version changed 11:25:53 symlinks are very handy ;-) 11:26:02 indeed. compiles with 2.95 11:26:13 * PoppaVic shutup goddamned lpr that way. 11:26:31 another eg. of the usefullness of pointers (gilbertdeb - rut ?) 11:26:33 --- join: kc5tja (~kc5tja@ip68-8-206-137.sd.sd.cox.net) joined #forth 11:26:34 I added the #include line, for memset 11:27:02 dont use the C library for string and buffer management 11:27:06 krish: hmm? 11:27:06 its lousy 11:27:07 Hi kc5tja :) 11:27:14 looks like nest unnest next all work 11:27:15 bah - spare me 11:27:16 the output indicates that 11:27:34 PoppaVic: have you seen djb 's code ? 11:27:42 yeah: it tests a primitive and calling secondaries from that prim. 11:27:56 can't say I have - what's a djb? 11:28:17 Robert: Howdy. :) 11:29:37 --- quit: Gnomon ("ChatZilla 0.8.23 [Mozilla rv:1.3/20030312]") 11:30:56 hmm.. Damn.. In some cases my comments are a bit out-of-date, Speuler - I'm sorry.. It was VERY early 11:31:17 i left it to you without any comments at all 11:31:22 that's worse :) 11:31:39 i'm going to study your code 11:31:42 I'll diddle what I see, so if you need a cleaner-copy, It'll be avail. 11:31:47 to get an idea how it works now 11:31:58 maybe i can even understand why it didn't work before 11:32:09 yeah - it was the steppers 11:32:46 that code is meant for #openbios 11:32:55 I'm not really "happy" with that inner-interp, but I couldn't see running three/four functions all adding to the stack-frame. 11:32:59 oh. i think i told you yesterday 11:33:05 yep. 11:33:39 I'd add support for a literal, and the basic math ops, and see where that leads, next. 11:34:20 yeah, once the vm works, the rest is just adding on 11:34:29 yah. 11:34:57 once you get branching done, you can cobble a loop up and leave it in forever-mode ;-) 11:35:37 PoppaVic: can I have a look @ the code ? 11:35:44 sure.. 11:36:12 dcc 11:36:33 par stack pop, to convert "star" to "emit" 11:37:30 essential diagnostic tool :) 11:37:41 sure. - but remember, you'll need to unbuffer stdout, then. 11:38:04 PoppaVic: djb => Dan J Bernstein - author of qmail 11:38:05 look at gforth & what they borrowed/stole for setting up for char i/o 11:38:10 ugh 11:40:31 well 11:40:34 PoppaVic: checkout http://www.fefe.de/djb/ 11:40:45 I remember qmail.. 11:42:16 downloading his lib now. 11:42:37 solid C code . i'm very impressed 11:43:50 OK, that's utterly retarded. 11:43:59 whazat? 11:44:13 http://www.fefe.de/djb doesn't mean the same thing as http://www.fefe.de/djb/ -- a violation of WWW RFCs. 11:44:20 i only understand 50 % of it 11:44:34 Speuler - the foo-file? 11:44:39 right 11:45:02 as I said, I had to dig up an ooold hardcover and look at pictures, myself ;-) 11:46:01 where are you getting "lost"? It's pretty straightforward, until you hit that one damned func - interp() 11:46:40 1. the struct vm_ is just a sort container, for the required variables and arrays ? 11:46:53 absolutely: I hate, HATE globals. 11:48:05 I don't mind globals as long as they are *all* contained in ONE module file to make them easy to maintain. That being said, I almost never have more than 6 globals in any of my programs. Forth programs are excepted, by the very nature of Forth. 11:48:35 It's better to stick them all in one struct called, coincidentally, "global" ;-) 11:48:54 I don't do that because it's not very compatible with how POSIX works. 11:49:06 It can be made to work, and I experimented with that technique before with good success. 11:49:15 But I find comprehension is better the more "traditional way". 11:49:16 I pass the VM around, since it makes for easier threading anytime, but they could all play with the global vm 11:49:54 Now, where I find this technique most useful is when you're writing your "application" 98% as a library, where only the main function is the real POSIX application. Then it works beautifully. 11:50:27 kc5tja: "it" => ? 11:50:31 Generally, though, I exercise OO design even when writing plain vanilla C, so I end up rarely ever using globals. 11:50:43 wrapping your globals in a data structure. 11:50:52 yeah, globals will rise on up and bite a chunk outta' yer ass. 11:51:08 * kc5tja nods 11:51:26 wheras, using one global - the struct can cure a lot, and passing around a "context" makes even more sense. 11:51:56 It also tends to make your code more "functional" too, and supports unit testing like no other technique can. 11:52:03 yeppers 11:52:05 (Yes, I practice XP religiously) 11:52:19 xp? 11:52:24 Extreme Programming 11:52:29 kc5tja: pair programming ? 11:52:36 * PoppaVic sighs 11:52:36 krish: Whenever I can, yes. 11:52:57 Why the sigh? 11:52:58 ah. the primitives i understand :) 11:53:09 kc5tja: how do you write tests ? 11:53:10 hehe 11:53:32 I write the tests *first*, before I write the core logic of the software I'm intending to write. 11:53:35 "XP" and such - I can live without. 11:54:04 hi-level also makes sense to me 11:54:06 kc5tja: so do you write tests in C/C++ ? 11:54:24 PoppaVic: While I'm not trying to convince you to switch over, I also would also ask you to keep an open mind, and that the results can't be argued with. 11:55:02 kc5tja: writing tests in C++ can be time consuming esp. with a slow compiler like GCC 11:55:08 krish: Yes. See http://cut.sourceforge.net for my own C Unit-testing Tool. It's a wiki site dedicated to supporting my CUT. I openly admit my documentation of it isn't world class, so I'm hoping that the Wiki can be used to help refine my documentation over time. 11:55:28 I don't need an "open mind" - anything "extreme" depresses me no end. 11:55:30 I find writing tests in C or C++ isn't any slower than, say, in Forth or Lisp or Smalltalk. 11:55:49 func-testing? 11:55:49 PoppaVic: Well, then it's your loss, not mine. 11:56:00 kc5tja: a much better alternative is to use a scripting language to write tests - SWIG is invaluable in this aspect 11:56:17 krish: That's good for acceptance tests, but not for unit tests. 11:56:31 The point of unit tests is to test production code at the *functional* level, not the application level. 11:56:33 a fine example is the GNUstep library 11:57:16 they developed bindings for guile. the unit tests for the entire framework is written in guile ! 11:57:24 Since unit tests are semantically and logically equivalent to formal methods, I question whether it's a superior alternative. 11:57:46 its easy to code in a dynamically typed language 11:57:51 Great, so now you have to test not only the core logic, but the test framework bindings as well. 11:58:04 It's very easy to code in a statically typed language too. 11:58:11 I do it daily. 11:58:13 SWIG takes care of marshalling and type conversion 11:58:25 OK, so you have to unit test SWIG too. 11:58:25 Any further luck, Speuler? 11:58:32 the switch statement in interp would have to be extended for dovar, dodefer and so on ? 11:58:37 yes. 11:59:19 Speuler: the GNU smalltalk VM is also worth studying 11:59:25 Frankly, I don't see any advantage at all of writing unit tests in an interpretted, dynamically typed language when the application language is perfectly suitable. 11:59:38 No, I don't believe so, Speuler - they are either primitives or secondaires. 11:59:47 "secondaries" 12:00:16 but, the next step is defining your dict-struct, vocabs and suchlike. 12:00:53 kc5tja: tried SWIG ? 12:01:18 Yes. 12:01:27 And I think it's woefully inadequate for testing purposes. 12:01:29 hmmm 12:01:32 Speuler: I'm suprised you guys haven't glomed onto the recent fig-forth release for the x86 stuff. 12:03:58 that was asm, not c 12:04:06 ahh 12:04:13 partly asm, mostly forth 12:04:34 still useful as example-source w/notes 12:05:02 ficl is another that seems lioke it might help. 12:06:02 --- join: futhin (~thin@198.162.22.58) joined #forth 12:06:06 hey all :) 12:06:18 lo. 12:06:22 futhin: hi 12:06:59 have you checked out the site? :P 12:07:04 got any suggestions? 12:07:08 ideas, etc 12:08:04 futhin: downloaded the archive. but not yet opened. 12:08:15 krish: I am still not going to do pointer arithmetics for a compiler :D 12:08:16 archive? 12:08:32 gilbertdeb: ha ha you wake up atlast 12:08:46 krish: archive? 12:09:25 are we talking about the http://forth.tech.nu site? 12:09:28 PoppaVic: i know forth interpreters from the inside out. but, my c is rubbish 12:10:00 well, that's why it talks back a lot 12:10:05 PoppaVic: i can implement interpreters in my sleep, using other languages :) 12:10:36 and then they end up being done over & over ;-) 12:10:37 preferably assembly lang 12:11:00 futhin: i meant the url 12:11:12 never done a c implementation. 12:11:29 (assumed it to be easier :) 12:11:43 * PoppaVic must get an international NO teeshirt.. over ASM 12:12:06 6809 vm in asm would take 7 lines of code 12:12:43 yeah, segfaulting in one line is easy ;-) 12:15:39 if it takes one line to segfault, i'd only have 7 to choose between 12:16:09 if my prog has 1000 lines, segfault potential is much bigger 12:16:39 or, bug potential, in general 12:17:42 But if you had 1000 lines, plus another 500 lines for testing (Forth makes this cake, since you can test each definition as soon as its compiled at LOAD-time), then the odds of segfaulting drops asymptotically towards zero. Of course, since it's an asymptote, it can never be truely zero except in the most trivial of cases, but I digress. :D 12:25:15 spammers should be spammed. 12:25:19 apropos of nothing. 12:26:03 * Speuler alias PhoodPhrenzy 12:26:34 kraephtig phuttern 12:31:24 EVIL!!! The heathens who registered localhost.net.au be sent to HELL!!! GRRRR!!! 12:31:24 --- quit: futhin (Connection reset by peer) 12:31:42 Now I have to type in 127.0.0.1 manually because THEIR DNS registration overrides my local definition for localhost! 12:32:06 does your resolv.conf say localhost first? 12:32:14 No, it's in my hosts file 12:32:20 hosts is supposed to be searched FIRST. 12:32:49 OK, I fixed it. Doggone DHCP program wiped my DNS settings... >:( 12:32:54 I just noticed isForth is listed on /. under freshmeat. 12:33:05 But damn them for choosing localhost.net as their base name anyway -- that's just WRONG. 12:33:06 yeah DHCP program knows best. 12:34:02 --- join: futhin (~thin@198.162.22.58) joined #forth 12:47:50 bye all! 12:48:12 --- quit: tathi ("leaving") 12:49:19 Later krish 12:49:51 --- quit: krish ("Client Exiting") 12:59:49 --- quit: futhin ("brb") 13:05:26 --- part: gilbertdeb left #forth 13:06:06 --- join: thin (~thin@acc-1-11.cariboo.bc.ca) joined #forth 13:06:13 --- nick: thin -> futhin 13:21:35 --- quit: futhin (Read error: 54 (Connection reset by peer)) 13:21:50 * kc5tja sighs -- Futhin just isn't having a good day today, is he? :D 13:23:07 --- join: thin (~thin@acc-1-11.cariboo.bc.ca) joined #forth 13:48:35 thin! 13:48:46 --- join: gilbertdeb (~gilbert@fl-nked-ubr2-c3a-29.dad.adelphia.net) joined #forth 13:49:00 hey delux 13:49:03 hey deluxe 13:50:19 life again ;-) 13:50:24 deluxe: how do you intend to order the x25 should the demand reach a critical mass? 13:50:44 hey thin :-) it was you who put up the new #forth hp? 13:50:47 yes 13:50:55 hp? 13:51:01 hp = home page 13:51:03 good work, thx 13:51:07 not a very common acronym ;P 13:51:14 oh. 13:51:17 where's the wiki ;-) 13:51:26 deluxe: good question :) 13:51:40 will wait for someone to code a forthwiki ;))) 13:51:50 :-) 13:52:13 btw do you know the wikibot? 13:52:27 or chumpbot? 13:52:38 thin: why didn't you put 4thsmith or some such? 13:52:49 or 4thwright :) 13:52:56 instead of top? 13:53:10 13:53:18 thin: i would help you, if you like... 13:53:27 because 4thwright/4thsmith is corny :P 13:53:33 no its not!!! 13:53:39 yeah it is!! :) 13:53:48 * gilbertdeb hasn't head 4th{smith,wright} outside of this #! 13:55:09 the website is more than just the channel's website 13:55:23 my fav link so far. -> http://www.eforth.com.tw/academy/ 13:55:31 it'll get bigger and every forther will check the site out ;) 13:55:48 --- join: crc (CharlesChi@ACA843B3.ipt.aol.com) joined #forth 13:55:51 maybe i should bug #forth people to write articles 13:55:57 hey crc 13:55:58 yes and then they'll identify themselves as 4th{smith,wright} 13:56:02 thin: yes you should. 13:56:03 crc: are you new? :) 13:56:06 yes 13:56:10 welcome. 13:56:13 you have mail. 13:56:14 :) 13:56:16 crc: do you code forth? :) 13:56:21 Yes I do 13:56:30 crc: yay, how did you find this channel? :) 13:56:50 I found it on Mark Manning's IsForth page 13:56:59 ah cool 13:57:11 deluxe: check this out: http://www.eforth.com.tw/academy/lab/forthchips.htm 13:57:16 welcome welcome 13:57:18 they may already be in production. 13:57:28 isForth is listed on /. today 13:57:34 under freshfish 13:57:45 freshmeat you mean 13:57:59 deluxe: it's a good idea to get everyone together to buy forthchips 13:58:02 yeah fresh{animal} 13:58:03 and i've figured out how to do it 13:58:36 thin: hey nice, i like to hear that! 13:58:42 crc: how long have you been playing with forth for? 13:59:04 about three years 13:59:12 wow, professionaly or for fun? 13:59:13 what we can do is get on CLF and start a forthchip production campaign.. and ask them to commit money towards the production of the forthchips 13:59:19 For fun mostly 13:59:25 cool. 13:59:32 but how do we get them to commit money? trust accounts ! :D 13:59:35 er 13:59:37 * deluxe is busy atm bbl 13:59:38 trust companies 13:59:43 thin: don't you suppose other c.l.* or c.s.* will be interested? 14:00:10 gilbertdeb: maybe, dunno :) 14:00:10 aftearall, there is no C on a chip or perl on a chip ;) 14:00:17 Wrong. 14:00:24 C on a chip? perl on a chip? 14:00:33 there is already a forthchip that you can buy for 15 bucks 14:00:35 do'nt say the whole x86 us C on a chip. 14:00:39 AT&T makes the Hobbit chip, or at least MADE the Hobbit, which is architecturally optimized for C in ways even RISC processors aren't. 14:00:40 thin: url? 14:00:46 from patriot science (http://www.ptsc.com 14:00:54 oh. kc5tja but thats a dead machine. 14:01:02 can be found lingering on ebay along with the newton :D 14:01:08 Well, technically, so is the P21 series. :) 14:01:40 and this too: http://www.eforth.com.tw/academy/chips/p24.htm 14:01:43 Only Patriot's stuff is still commercially selling to the best of my knowledge, with only limited production runs of RTX2000s for special-case, rad-hard designs. 14:04:11 hmmm, I don't see a link to the forth chip on ptsc 14:04:30 They're "Java chips" are really Forth chips. 14:05:14 s/They're/Their/ 14:05:19 * kc5tja hates English sometimes... 14:06:30 gilbertdeb: click the download button 14:06:35 http://www.ptsc.com/Download/index.html 14:07:36 --- quit: crc ("Leaving") 14:19:57 oy, geezus... thunder, sleet & snow 14:52:24 --- join: deluxe_ (~deluxe@p508050D6.dip.t-dialin.net) joined #forth 14:54:09 --- quit: deluxe ("bb") 14:54:46 --- nick: deluxe_ -> deluxe 15:28:42 Speuler: sending more mail.. 15:29:08 Splash. 15:32:36 re 15:32:52 lan migration at full speed for the night :-) 15:33:06 thin: still there? 15:33:31 yes 15:33:34 whats up? 15:34:13 i have some good ideas for the # hp 15:34:19 eh? 15:34:23 deluxe: we could get people to put their money into a collection for buying a production run. we could have a trust company that keeps their money, and if the deal falls through, they get their money back 15:34:34 deluxe: sure, give me the ideas :) 15:34:40 deluxe: don't say hp tho 15:34:44 say website :P 15:34:46 nobody understands hp 15:34:55 hp is hewlett-packard :P 15:35:19 oh, give them time - they got to calling tcp/ip a "stack".. 15:35:21 okok how to shorten website :) 15:35:32 url 15:35:36 uri 15:35:42 home/site/page 15:35:43 poppavic: look at topic 15:36:35 btw idea atm bot for acronyms :-) 15:37:00 sifbot: what is hp??? :/ 15:37:01 deluxe: Word not found: what 15:37:07 aww 15:37:14 heheh 15:37:23 sifbot isn't much of a bot 15:37:46 someone will write a better bot 15:37:51 most likely i440r 15:37:57 but it'll take him a year 15:37:57 heh 15:38:02 it's a very delighting beginning 15:38:38 i thought about rewriting the wikibot in forth 15:39:47 --- quit: thin ("Leaving") 15:39:57 --- join: thin (~thin@acc-1-11.cariboo.bc.ca) joined #forth 15:40:04 eww 15:41:20 apropos projecting x25, i'll be happy to join/contribute 15:41:37 better now than newer 15:41:49 *never* 15:41:58 a production run of x25 is unlikely 15:42:03 f21 is more possible 15:42:21 x25 hasn't been produced yet. it would have to be produced and it'll likely have bugs, etc 15:42:26 f21 has less bugs 15:43:49 thin: key q is, why are forth chips so few. or, what misses 15:44:35 Key answer: Forth chips are commodities -- it's not possible to sell them for a profit because they're so cheap to make, and so simple, that they offer no proprietary advantages to any one manufacturer to charge extra for. 15:44:56 No distinction in the market place means no compelling or unique selling point, which translates to reduced sales. 15:45:05 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth 15:45:44 kc5tja: i'm not sure, if you're kidding a bit... 15:45:50 I'm damn serious. 15:46:00 Any 5-year old can reverse engineer the F21's main CPU core. 15:46:22 Even I built a simple 4-bit, and next year an 8-bit compatible, processor using discrete component TTL. 15:46:42 This was...goodness...back 15 years ago. I'm 28 years old now, so that'd place me at 13 years old. 15:47:30 I never did well with ic's.. I could draw pcb & etch/plate/press/patch & install - but, damned if I could ever throw parts together. 15:47:42 Another problem that is often cited (often fallaciously) is that you can't take advantage of instruction-level parallelism. 15:49:04 There are two approaches to this so-called problem. One, put multiple execution cores on the same die -- hardware multithreading or true MIMD. But people don't know how to write well-factored programs, let alone parallelizable programs, so that wouldn't have worked in the longer term anyway. 15:49:23 * PoppaVic sighs 15:49:40 There's no point in going that ape anyway 15:49:42 The other is to utilize a technique similar (but not exactly a copy of) what's used in today's processors, where whole *patterns* of instructions are recognized as a group, and concurrently executed as a single meta-instruction. 15:50:13 PoppaVic: Why not? 15:50:16 this is for the x25? 15:50:26 A board with slots for stackable daughter-vards, and a decent bus let's you do all sortsa' freaky-deaky. 15:50:34 vards/cards 15:50:41 And a multiprocessor affects this how? 15:51:07 the card may be no more than a dimm/simm-style, but cheap..? oh hell yah 15:51:22 And a multiprocessor affects this how? 15:51:54 I'm really confused. 15:51:55 There are two ways to go: create a new chip - requiring a totally new programming paradigm & tools, or you use what is to better advantage. 15:52:20 I suspect the former would cost pennies to make - and billions in training & tools 15:52:36 Sorry, that just plain isn't true. 15:52:50 kc5tja: ic what you mean. on the other side, the market are the customers. if you're able to give them a decent system (at a good price), they'll take it. else are stones in the path, a common 'strategy'. so my q was 1)what is missing 2)where are the stones/bricks 15:52:56 * PoppaVic is scared shitless - imagining todays haqueers trying multiprogramming foo-foo 15:53:19 We already have compilers TODAY that take imperative-style code, rewrites it in functional notation, optimizes it functionally, and re-writes it back to imperative style code, slot-optimized for instruction dispatch for CPU's multiple execution units today. 15:53:56 Multiprocessors have been used in the past to *excellent* effect with functional decomposition of software. 15:54:13 Knock yerself out ;-) 15:54:23 * kc5tja sighs 15:54:30 /ignore PoppaVic 15:55:58 deluxe: I interpretted your question as being specifically targeted at stack processors. 15:56:06 I don't think anything is *missing* per se. 15:56:30 But the state of the art today is influenced solely by marketing, not by technicians. 15:56:30 eg. a forth distro? 15:56:55 While I was working for Hifn, our HIPP-I architecture chips were hardly being utilized to full capacity when we announced HIPP-II. 15:58:19 And now they've got HIPP-III on the market, which is ludicrous in my opinion. but, that's the market they're after, and whatever marketing says, it must be true. 15:59:19 let's hope marketing read the book "the innovator's dilemma" ;) 15:59:29 or not 15:59:51 kc5tja: would you say forth chips are dead? 16:00:20 kc5tja: a bit exaggerated... 16:00:38 deluxe: For something like desktop or server markets, they never had a chance. Accumulator, which led to register-register, architecture microprocessors all came from the 60s "big-iron" machines, which were also accumulator machines. 16:00:52 It follows naturally that the software those big-iron machines also will tend to run on the microprocessors. 16:01:07 The success of Linux, for example, is a clear demonstration of this concept. 16:01:30 For embedded or truely unique devices (I would vote for PDAs, but PalmOS has wiped that opportunity out), I would say they definitely have a chance. 16:01:46 As I understand it, RTX2000s are still used in satellite designs today. 16:02:21 worse is better = accumulator ? :) 16:03:27 The accumulator architecture is actually one of the better architectures. Logically, it's an SS1 stack architecture with a one-deep stack, if you were to use Phil Koopman's classification system. 16:03:41 Thus, it's a nice union of both register-register ideas and stack ideas. 16:04:23 The 6502, 65816, and 6809 make extensive use of its advantages to routinely outperform equally clocked 80286 processors (though the 80386 is faster). 16:05:56 That's sad when a 4MHz 6502 (8-bit data bus) can out-perform a 5MHz 8086 (16-bit data bus). See http://www.westerndesigncenter.com/compare.html 16:06:48 yeah, didn't the 8086 do 8 bit outside and 16 inside, or is that reversed - double-fetching iirc. 16:07:59 thin: about #forth website, eg. wiki/wikibot/... on forth.tech.nu or shall i open another? (currently not 24w/7d) 16:12:04 ? 16:12:49 say that again? 16:15:24 i can help you with the wiki/... @forth.tech.nu or i can start a new/other website (for #forth). thing is i can't provide 24h uptime atm 16:16:26 i don't really need a wiki yet 16:16:34 and it should be 24h uptime too ;) 16:16:52 there are already two forth wikis 16:17:10 sure, i could look for *free* hosts... 16:18:15 if you find a good free host let me know ;) 16:18:33 ok 16:18:41 --- quit: Speuler (Read error: 110 (Connection timed out)) 16:19:08 a wiki wouldn't hurt i suppose 16:19:16 but we could use one of the wikis available.. 16:19:23 there's 2 wikis 16:19:39 and i think onetom might have a wiki available.. 16:20:11 --- join: Speuler (~semtex@mnch-d9ba4dae.pool.mediaWays.net) joined #forth 16:22:28 re 16:23:20 wb 16:23:31 ha 16:23:52 "you have mail" 16:24:06 indeed. 16:24:12 vocs 16:24:30 i'd think vocs wouldn't have big, if any, impact on interp 16:24:42 vocs can be implemented high-level 16:24:49 you need an outer-interpreter to get the tokens for the lists 16:25:01 to speed up the switch statement, few word handlers would be good 16:25:42 the switch is in interp() and bothers no one else 16:25:43 outer interpreter can be pure high-level too 16:25:49 needs some primitves of course 16:26:14 as longer it is, as more time it takes to branch to the last selections 16:26:32 next is used often 16:26:51 docol / nest too 16:26:57 those should be first 16:27:17 some are optional 16:27:30 can be written high-level or as word handlers 16:27:37 dodoes may be useful 16:28:02 dodoes can replace many word handlers, by allowing to write defining words instead 16:28:32 rather than doconst, use : constant create , does> @ ; 16:28:40 doconst would be quicker 16:28:59 would mean, looking for a good compromise 16:29:08 vocs don't need word handler at all 16:29:19 those are compile-time active anyway 16:29:25 no need to be quick there 16:29:59 vocs used to contain words 16:30:01 headers 16:30:19 hardly ever does run-time access headers 16:31:10 i presume, but better make be sure, that it is ok for your code to fall under the gpl license ? 16:31:53 cause openbios is probably going to be gpl 16:31:55 doesn't bother me, no. 16:32:01 nice 16:32:03 hi stepan ! 16:32:12 I'd think that lgpl makes more sense, but - whatever. 16:32:20 stepan is founding member of openbios 16:32:49 --- part: deluxe left #forth 16:33:12 as a rule, I'd suggest a nice lgpl inner/outer interp - everything beyond that, sure: make it gpl. 16:33:29 * PoppaVic is tired of seeing everyone rewrite the whackiness over & over 16:33:32 lgpl is fine with me 16:34:32 well, I'll tinker in gforth - you can tinker w/ fooforth and send me an update, then.. I'm trying to solve a set of irritations to my code, before I tackle vocs/headers 16:35:01 in one forth, i used segmented memory for vocs 16:35:15 each voc had an own set of pointers to words 16:35:22 words like find 16:35:26 create 16:35:46 so by making another voc context, 16:35:47 yeah, trying to keep it all isolated strikes me as a "good thing" 16:36:05 ..in the case of things like embedded-controls & bios, particularly 16:36:24 the voc specific find was used 16:36:36 by default, all pointed to the same find 16:36:42 but you could hook up another one 16:36:45 old f-pc "words " ? 16:37:24 you could for example create a voc, which allowed , say, pattern matching searches 16:37:49 right 16:37:51 or you could set case-sensitiveness to specified vocs 16:39:13 it worked quite well, and still was downward-compatible 16:40:12 was important that, when applying special searches, the other vocs weren't affected 16:40:48 still, i don't see where interp would have to be affected by vocs 16:41:09 I'm thinking of tick'ing and building the lists mechanically. 16:41:52 in the same forth, i used an unusual wordlist structure as well. 16:42:08 that one allowed me to change header locations on-the-fly 16:42:26 or remove headers as desired, 16:42:31 relocate headers, etc 16:42:49 --- quit: ramnull ("This isn't Happy Hour!") 16:42:50 used that to save image with simple header organistion, 16:43:01 and after loading image, i reorganized headers 16:43:14 first, doing a binary search on those 16:43:30 easy to do because i could just plug in another find 16:43:52 it was for a slow microcontroller 16:44:25 Yeah, relocatable seems sensible to me - as is shlib-support., (but not down in a bios ;-) 16:44:26 being able to speed up searches, by just loading a file 16:44:49 i tend to put no body-to-header link in my systems 16:44:55 only header-to-body 16:45:10 result is, you can move headers around without patching up pointers 16:45:37 the few cases where you need to refer to a header by body address, 16:45:52 (which do not need to be quick), 16:46:14 i was just doing a find, comparing the pointer-to-body with the body address 16:46:23 match -> header found 16:47:05 that gave me great liberty, mistreating headers all the way i liked 16:47:39 also, i could eliminate the link field 16:48:24 didn't need it because the header length gives enough information to get to the next header 16:48:38 (provided they are in a kind of list structure) 16:49:41 stepan idle for 3 hours 16:50:04 stepan knows c much better than i do 16:52:30 i don't see the need to add vocs very early in a project 16:53:01 some forths i did came totally without vocs 16:53:49 as long as your list of headers has a begin and an end, it would fit into a voc scheme, added later on 16:54:23 basically, you prented you got vocs, but you only use one 16:54:32 pretend 16:55:27 and only if you really need another voc, you fit you only voc into the voc scheme you add onto the interpreter at that time 16:55:55 like you can do with a tasker, for example. 16:56:13 your running system is the only task, until you get to the point to add more 16:56:49 wohaa 16:56:57 there's an attachement to your mail :) 16:57:20 yes, we call those "file attachments" - very "extreme email" 16:57:35 almost missed it 16:57:54 was focusing on your message, and the voc thoughts 16:58:16 --- part: thin left #forth 17:02:08 vm went into .h, LFAnew and delete seem to be the main changes 17:02:10 HAH! I got the bastard workin' - that's cool. 17:02:39 CCODE instead of prim ? 17:02:50 yeah - easier to follow, I suspect 17:03:20 ah 17:03:44 and a real primitive cr 17:04:42 there's quite some code existing already for openbios btw 17:04:50 good. 17:04:55 was just busy with catch/throw 17:05:16 when the strange handling of return address of primitives hit me 17:05:17 I still hate 'exceptions'. 17:05:28 they are useful 17:05:38 localizing error handling 17:05:45 bah 17:05:45 rather than a system wide abort 17:06:02 an int or * return can do the same - but folks refuse to handle 'em. 17:06:43 makes it easy to throw no matter where from, regardless of par stack or return stack balance 17:06:50 no need to clean up yourself 17:07:01 I don't like 'throw' - which was my point. 17:07:34 if you throw from, say , three words nested deep, you may have a hard time to het to know how to clean up 17:08:13 catch just puts a fence on the stack, saying "restore stacks to this level" 17:08:47 i think there's nothing wrong with that 17:09:30 you're beyond 40 i guess 17:10:04 well, you gave me sufficient clue to allow me guessing that 17:10:46 i guess you're from somewhere between 1955 to 1965 17:10:59 43 17:11:03 1960 17:11:11 (soory, was beating this compiler-word) 17:12:06 * Speuler is of vintage 60 17:13:26 but still have a month to go to become 43 17:15:51 hah.. sooo cool ;-) 17:17:57 "vocs stacks structures ..." 17:17:57 structures >order words 17:17:58 stacks stack% stack%sii stack%sbp stack%sbz stack%swl ok 17:17:58 stacks >order words previous previous 17:17:58 2sdrop 2s> 2>s 2s@ sdrop s> >s s@ &s ?depth depth ok 17:20:13 10 newstack test test :: ?depth . cr 17:20:39 very nice namespaces-of-doom ;-) 17:21:27 &s ? 17:21:42 I didn't like sp@ 17:21:49 ah. ok 17:22:35 '&s' leave a ptr to the top of the stack 18:01:22 --- part: kc5tja left #forth 18:30:25 --- join: I440r (~mark4@1Cust188.tnt1.bloomington.in.da.uu.net) joined #forth 19:14:11 * PoppaVic finally got peeved, tweaked & is rebuilding gforth... 19:14:42 heh 19:14:54 use isforth not gforth :) 19:15:13 I hate the prompt - and, of course, that's something important they neglected to make deferred 19:15:39 gforth doesnt have defered words ? 19:15:50 but emit/key etc HAVE to be defered 19:15:54 tons - but NOT the friggin' prompt (idiots) 19:15:58 its just all wrong for them to NOT be defered 19:16:09 whats the prompt ? 19:16:44 heh.. I keep looking for a nice little > in the corner - what I get is the ANS "recommended" ' ok' THEN a cr 19:16:45 i made quit a defered word in isforth, and key, and emit, and "default" of corse 19:16:55 i have too many defered workds i think 19:17:18 thers no prompt in isforth, theres not supposed to be one 19:17:24 you see an "ok" except after an error 19:17:28 and you see the cursor 19:17:32 thats your prompt 19:24:11 you could ADD a prompt like that to isforth tho 19:24:17 real easilly methunks :) 19:34:25 --- quit: PoppaVic ("(I don't need a reason)") 19:42:36 --- quit: I440r ("Reality Strikes Again") 19:48:21 --- join: PoppaVic (~pfv@s70.waters.gtlakes.com) joined #forth 19:48:54 dead cars in the drive *sigh* 19:51:48 I think I fixed it - part of it was the prompt always followed the the query/interpret in (quit) - and part was the idiots using goddamned cr as a throw/catch test at the top. 20:00:34 --- quit: skylan ("EPIC Panic: EPIC4-1.0.1:Referential integrity failure: Channel [#emacs] on server [0] is connected to window [18] on server [0") 20:02:01 --- join: skylan (sjh@vickesh01-4731.tbaytel.net) joined #forth 20:33:19 yah.. nice. 20:57:19 --- quit: gilbertdeb ("Monk has left the building") 22:00:22 --- quit: PoppaVic ("(I don't need a reason)") 22:14:08 --- join: deluxe (~deluxe@p508050D6.dip.t-dialin.net) joined #forth 22:25:07 bon jour 22:50:21 --- quit: sifbot (Remote closed the connection) 23:20:49 --- join: fridge (~fridge@dsl-203-33-160-120.NSW.netspace.net.au) joined #forth 23:59:59 --- log: ended forth/03.04.03