00:00:00 --- log: started forth/03.04.27
03:09:15 --- quit: Speuler ("NO CARRIER")
05:29:48 --- join: Mur (murr@baana-62-165-188-13.phnet.fi) joined #forth
06:09:52 --- join: snowrichard (~cpu2@c66.190.103.147.ts46v-01.mrshll.tx.charter.com) joined #forth
06:10:07 <snowrichard> morning
06:11:14 --- part: snowrichard left #forth
06:13:55 --- join: snowrichard (~cpu2@c66.190.103.147.ts46v-01.mrshll.tx.charter.com) joined #forth
06:15:15 --- part: snowrichard left #forth
07:13:39 --- nick: Mur -> mur_AFK
08:04:39 --- nick: mur_AFK -> mur
08:08:39 --- join: snowrichard (~cpu2@c66.190.103.147.ts46v-01.mrshll.tx.charter.com) joined #forth
08:09:31 <snowrichard> hello
08:13:18 <mur> terve!
08:16:29 <snowrichard> oh i was in a browser
08:17:36 <snowrichard> I have a simulated vax running netbsd 1.6.1 with networking 
08:18:00 <snowrichard> I might try my hand at coding a vax forth again
08:19:45 <snowrichard> Robert I just took a look at your web site
08:19:57 <Robert> :)
08:20:00 <Robert> Did you like it?
08:20:21 <snowrichard> its fine.  I am a little older than you in that picture ;) 
08:20:49 <Robert> The average age in here seems to be 30-40...
08:21:05 <snowrichard> in lots of channels its in the 20's
08:21:11 <Robert> I'm very young in this channel. But almost old in the rest of the network :P
08:21:29 <snowrichard> i am 46
08:21:39 <mur> Robert?!
08:21:44 <snowrichard> or will be at the end of may that is
08:21:50 <Robert> :)
08:21:55 <Robert> That's a little bit older, yes
08:22:00 <mur> what are you blabbering about , Robert :)
08:22:08 <Robert> Nothing, why?
08:22:11 <mur> i dont trust with that :P
08:22:29 <mur> if the avg woudl be that small i doubt .. urhg
08:23:37 <snowrichard> well I feel old when I tell people I have used punch cards and paper tape and they go like WHAT?
08:23:48 <Robert> Haha
08:24:10 <Robert> That was obsolete long before I was born, so I guess I should look suprised too.
08:24:40 <Robert> But... 46 isn't THAT old.
08:25:10 <Robert> Not compared to all those veterans.
08:25:15 <mur> 70's punchcard computers weren't exactly like laptops
08:25:46 <snowrichard> well kind of hard to tote around a building and tons of airconditioning
08:26:07 * mur has a CS book of that age :)  it tells that in soon future there will be laser beam and ink bubble memories .. not seen yet
08:26:26 <mur> i think it might have been as late as 77
08:27:38 <snowrichard> the IBM 360 I was using in 1977 had punch card readers still.  By 80 or so they had phased out most of that for students and put in a terminal server they built with MP/M
08:27:41 <mur> yes 1977. no wonder some people have computerphobia if they have seen computers that time :)
08:28:25 <mur> this general "information technology" book contains the usual assembly and basic book contents
08:29:06 <snowrichard> now I can run Hercules emulator on this walmart PC and it is probably faster than the 360 was
08:29:34 --- join: Speuler (~Speuler@mnch-d9ba4bd3.pool.mediaWays.net) joined #forth
08:29:49 <Robert> Hi Speuler 
08:29:56 <Speuler> g'day
08:30:56 <snowrichard> do any of you remember discussing the harvard architecture of the vax (code space not writeable by default)? 
08:33:11 <snowrichard> I have figured out a way around that, I could define the core of the interpreter in native vax code, but in definitions store a byte code indexing a primitive function table.
08:34:51 <snowrichard> and put all the user dictionary in data space.
08:39:34 <Speuler> snowrichard: right. forth hi-level words are just address or token lists, to be considered as data
08:40:26 <snowrichard> I have a C version working like that, and it compiles on the VAX gcc as well.
08:41:14 <Speuler> 've written a PIC runtime which works about the same. PIC also uses a harvard architecture
08:42:43 <snowrichard> not strictly standard's compliant this is more or less my own language definition. 
08:42:52 <snowrichard> I call it Michael after my son
08:43:13 <Speuler> standard doesn't require a specific threading model
08:43:27 <snowrichard> well I don't have all the ansi words
08:43:57 <snowrichard> borrowed operators from C
08:44:00 <Speuler> i guess you got enough words to implement the missing ones
08:45:19 <snowrichard> I have not studied the standard in great detail yet actually.  I was more or less designing my own way of doing it
08:47:05 <snowrichard> but I do have all the C arithmetic operators and logical operations so I would be able to do most anything with 32 bit integers data and characters. 
08:48:40 <snowrichard> my var word is a little different   I stack the number of bytes I want allocated , then var name, instead of the initial value.
08:50:54 <snowrichard>  so 10 4 * var ptarray   could allocate an array of  10 ptrs. 
08:52:49 <snowrichard> ptrarray 3 ARR@    gets the 4th ptr contents
08:54:04 <snowrichard>  and if it were a pointer to integer an @ would get the actual value
08:59:59 --- quit: snowrichard ("using sirc version 2.211+KSIRC/1.2.4")
09:20:20 --- join: tcn (~r@tc1-login26.megatrondata.com) joined #forth
10:03:11 * tcn is away. Automatically set away [SZon]
10:09:49 --- quit: tcn ("That's it for today")
10:13:14 --- join: tcn (~r@tc1-login26.megatrondata.com) joined #forth
10:18:22 --- quit: tcn ("Leaving")
10:43:40 --- join: tcn (~r@tc1-login26.megatrondata.com) joined #forth
10:54:30 * tcn is away. Automatically set away [SZon]
11:11:55 --- nick: mur -> mur_OUT
11:13:51 --- quit: tcn ("That's it for today")
11:37:44 --- join: kc5tja (~kc5tja@ip68-8-206-137.sd.sd.cox.net) joined #forth
11:53:31 --- join: Serg_Penguin (~z@ts21-a142.Moscow.dial.rol.ru) joined #forth
11:54:58 <Robert> Hey Serg_Penguin 
11:55:12 <Robert> And kc5tja 
11:55:34 <kc5tja> re
12:14:54 <mur_OUT> hey Serg_Penguin!
12:15:14 * mur_OUT out
12:15:18 --- quit: mur_OUT ("MURR!")
12:16:06 --- quit: Serg_Penguin (Read error: 110 (Connection timed out))
13:12:08 --- join: a7r (~a7r@206.72.82.135) joined #forth
13:12:10 <a7r> hey..
13:12:32 <kc5tja> re
13:16:12 <Robert> Hi a7r 
13:16:31 <a7r> what's going on Robert?
13:16:44 <Robert> Not very much, to be honest :)
14:12:00 <Klaw> the world is spinning around a gigantic ball of fire
14:12:09 <Klaw> and you say nothing is happening
14:21:10 <Robert> Nothing new on the western front.
14:22:23 <Robert> ..or whatever the English title might be.
14:52:49 --- join: tcn (~tcn@tc2-login24.megatrondata.com) joined #forth
14:53:53 --- nick: kc5tja -> kc-food
14:54:29 <tcn> l
14:54:46 <tcn> i mean, hello :)
14:55:36 <Robert> Hi tcn :)
14:56:33 <tcn> i'm trying to figure out this IRC client.. god, they're all so LAME!!!
14:57:01 <tcn> but at least this one's only 28k
14:57:19 <Robert> :)
14:57:22 * Robert uses irssi.
14:57:27 <Robert> Anyway, time for bed. Night!
14:57:31 <tcn> see ya
15:00:14 <a7r> tcn: whch client?
15:00:18 <tcn> tinyirc
15:00:39 <a7r> werd.
15:02:43 --- quit: a7r (Remote closed the connection)
15:02:56 --- join: a7r (~a7r@206.72.82.135) joined #forth
15:03:01 <a7r> damn it
15:03:08 <a7r> crashing xchat
15:03:11 <tcn> heh
15:03:35 <tcn> tinyirc would be great, with backscroll and window switching
15:04:36 --- join: a7r_ (~a7r@206.72.82.135) joined #forth
15:04:49 --- quit: a7r (Remote closed the connection)
15:04:50 --- nick: a7r_ -> a7r
15:05:23 <tcn> if msg crashes you then I won't do it :)
15:05:38 <a7r> yeah, I've jumped into BitchX. ;)
15:05:47 <tcn> oh tha'ts better.
15:05:53 <a7r> that is xchataqua for OS X's Cocoa, and it has some rough spots.
15:21:36 <onetom> while(*dest++=*src++);
15:21:54 <tcn> yeah, it translates directly to asm :)
15:29:35 --- join: monk23 (monk23@mstr195175-33459.dial-in.ttnet.net.tr) joined #forth
15:29:45 <monk23> hi ! :)
15:29:50 <tcn> hey
15:32:02 <ianni> cocoa irc clients all suck
15:32:09 <ianni> hehe
15:32:13 <ianni> shadowirc is ok
15:32:15 <ianni> i use epic
15:35:58 <onetom> tcn: asm? what asm? anyway, it was just a misclick
15:37:05 <tcn> oh, a7r and I were talking about C just as you said that
15:39:37 <tcn> that snippet you showed really demonstrates how C is just a shorthand for assembler
15:39:42 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth
15:45:09 <onetom> ah, c :)
15:45:17 <onetom> tho, it was just an accident
15:45:25 <tcn> yeah.
15:45:33 <onetom> whats up anyway?
15:45:53 <tcn> I was on #c a little yesterday.. picked up some good tips which should help with Isforth amongst other things!
15:46:33 <onetom> like what?
15:46:35 <tcn> dlopen() for example.. 
15:47:18 <ramnull> Yeah, I've been curious about interfacing with C also.
15:47:21 <tcn> implement that in forth/asm, and you have access to all shared libs
15:47:53 <tcn> newer JED versions also have some kinda module system..
15:48:32 <tcn> ie, you take those .o files and link them into a running interpreter
15:48:44 <ramnull> Hmmm...I dont imagine putting a dlopen hook into a Forth would be too difficult.
15:49:24 <onetom> yeah, why should 1 code it again in any other lang?
15:49:26 <tcn> it's more complicated than i'd like, but it'd help ease the transition away from this bloatware
15:49:29 <onetom> 4 fun?
15:49:49 <ramnull> Wouldn't that force the Interpreter to compile around the the DLLs?
15:50:18 <onetom> but if u plan a transition, define own, simpler file format, so u dont exactly need dlopen
15:50:18 <tcn> around them?
15:50:27 <tcn> yeah
15:51:17 <ramnull> Hmmm...I suppose one could tell Forth to fork itself for the specific purpose of interfacing C DLLs.
15:51:26 <onetom> but, storing things precompiled is 1 thing what "bloats wares" .. ;)
15:51:56 <tcn> the problem is when you precompile code for every possibility
15:52:17 <onetom> running from src compiling it all the time seems a better idea
15:52:32 <onetom> anyway, i gotta go
15:52:39 <tcn> ok, see ya
15:52:41 <onetom> b back in half an hour
15:52:51 * onetom just changes location
15:53:16 <tcn> if you run something alot you should precompile it
15:53:33 <ramnull> What about hacking a C interpreter(they do exist) to recompile C source visa vi the Forth source? Know what I mean?
15:53:36 <onetom> no, if compilation time is very short
15:53:58 <ramnull> "On the fly" sort of.
15:54:08 <onetom> & if u modularize it well u dont even have 2 compile all the sources @ startup time
15:54:45 <onetom> ramnull: c compilation is a very expensive operation
15:55:07 <ramnull> Good point.
15:55:16 <Speuler> there's cint (c interpreter)
15:55:23 <onetom> ramnull: thats why on the fly C compilation is not a common practice
15:55:49 <tcn> i don't see why C is any slower than Forth to compile.. except for slow compilers like GCC
15:55:51 <onetom> Speuler: & there is awk which is also very similar...
15:56:23 <Speuler> hmmm
15:56:24 <onetom> tcn: because of the more complex syntax of c. mainly, ithink
15:56:33 <Speuler> awc and c are pretty different
15:56:37 <ramnull> I suppose one could make a specialized Forth linker, optimized for that task alone.
15:56:47 <onetom> Speuler: mmmm.. why?
15:57:01 <Speuler> i can program awk. i can't c :)
15:57:08 <onetom> Speuler: there r no pointers, but anyway they r pretty similar
15:57:14 <onetom> lol
15:57:35 <onetom> me 2
15:57:37 <tcn> I guess C is all about pointers :)
15:57:57 <onetom> and data types :/
15:58:04 <Speuler> and types, casting etc
15:58:05 <ramnull> That's one reason, in an earlier life, I used Ada over C, because I could compile in new code on the fly while the program was executing.
15:58:19 <onetom> wow
15:58:26 <onetom> i didnt know it about ada
15:58:56 <ramnull> Well, Ada uses seperate compilation for those tasks.
15:58:56 <tcn> anyway, whatever the cost of compiling C, it'd be good to have a small *simple* C compiler even if it's not the most efficient.. just so you can run C programs.
15:59:31 <ramnull> Sort of like Forth deferred words. Kinda.
15:59:42 <onetom> tcn: yeah, c is not an absolutely evil thing
16:00:07 <onetom> tcn: u can express things in it a more compact way
16:00:19 <tcn> and if you use a big C program that takes 10 minutes to compiler, you only want to do it once!
16:00:29 <ramnull> Heh. I outta write a generic "Interfacer" dll for Ada, Forth, C, etc...
16:00:34 <onetom> tcn: tho, awk, tcl, python all has this feature:
16:01:00 <onetom> tcn: u can express certain things in them more easily than in other langs...
16:01:18 <ramnull> Awk, tcl, python, and some versions of Lisp are all based on lower level precompiled C libs.
16:01:25 <ramnull> For the most part.
16:01:53 <onetom> tcn: but it wouldnt take 10mins if it were written in a very modular 4th
16:02:26 <onetom> tcn: just, say 1min & because probably it would run for a lot of minutes
16:02:27 <tcn> if you guys ever write some really big Forth programs you'll wish you could precompile them..
16:02:55 <onetom> tcn: that 1min shouldnt be spent on compilation in 1 term
16:03:00 <ramnull> tcn: Precompilation and run-time compilation can work hand in hand.
16:03:17 <onetom> tcn: chances r u can defer most of it 2 run time
16:03:22 <tcn> or you could pre-parse them..
16:03:25 <onetom> "run time"...
16:03:35 <ramnull> I see no dichotomy here.
16:03:43 <Speuler> tcn: the bigger ones i compiled
16:04:02 <ramnull> Of course, I'm drawing from my experiences with the GNAT Ada compiler.
16:04:50 <ramnull> Big source, and bigger binaries. But fast as hell. Heh.
16:05:02 <tcn> hmm.. yeah, it's ideal if you can precompile the program but extend it at runtime.
16:05:13 <ramnull> tcn: Exactly.
16:05:29 <onetom> tcn: the graphical interface of bigforth is pretty large
16:05:33 <ramnull> tcn: Think of precompilation as defining the condition the program should start out in.
16:05:45 <monk23> byes
16:05:57 <tcn> later, monk
16:06:06 <monk23> :)
16:06:10 <onetom> but i havent ever experienced any significant compilation time while i was using it
16:06:29 <tcn> how fast is your computer?
16:06:47 --- quit: monk23 ()
16:06:53 <ramnull> Bigforth tries to cram too much stuff into to small a space. At least in my opinion.
16:07:10 <onetom> i tried it on a p200 yrs b4
16:07:54 <tcn> it must not be THAT big
16:07:55 <onetom> woody:~/x/bigforth# time make
16:07:59 <onetom> real    0m2.192s
16:08:36 <onetom> tcn: probably not, but its a full development env.
16:09:16 <onetom> this 2secs is all the compilation of the C & 4th blks on an 1.7 celeron
16:09:27 <tcn> 1.7ghz???
16:09:42 <onetom> yup
16:09:55 <tcn> so it should take 30 secs on this machine, in theory :)
16:10:01 <tcn> that's unacceptable.
16:10:10 <onetom> i wont think..
16:10:17 <ramnull> tcn: Same here.
16:11:09 <ramnull> Alright, what would be the time penalty to have Forth generate an elimination tree for the DLL it wants to open?
16:11:13 <tcn> sure, I could get a new machine but it'd cost a lot and I'd prolly have to upgrade my kernel and everything to use the hardware..
16:11:17 <onetom> compiling only the 4th blks is:
16:11:21 <onetom> real    0m0.647s
16:12:21 <ramnull> i.e. With a C binary, is it possible to have Forth pull out only the functions it needs, and dump the rest?
16:12:32 <tcn> say you load libvga.so and put all the symbols into a "libvga" vocabulary
16:12:49 <tcn> quick hack :)
16:13:10 <ramnull> tcn: Not all the symbols, have a tree file eliminate the unwanted ones.
16:13:26 <tcn> it could be done, probably not with ELF though
16:14:05 <tcn> the old "ar" libraries could handle it.. when you compile a program, it just pulls out the functions you need since it's a static binary.
16:14:20 <onetom> look:
16:14:21 <onetom> woody:~/x/bigforth# echo `rm bigforth.fi ; make 2>&1 | grep '\.f' | awk '{print $1}'`
16:14:25 <onetom> mv: make: ./forthker assem486.fb stream.fb hash.fs fileint.fb errore.fs strings.fs savemod.fb tasker.fb idle.fs tasker.fb vt100key.fs output.fb disass.fb tools.fb except.fs startup.fb struct.fs random.fs environ.fs ./bigforth minos.fb locals.fs oof.fb lambda.fs sincos.fs points.fs qsort.fs string.fs minos.fb x11.fs xpm.fs opengl.fs minos.fb splines.fb minos.fb pixmap.fs minos.fb xstyle.fs edit.fb ediwind.fs edit.fb edit.fs edit.fb browser.fs login.fs ptt
16:14:32 <ramnull> tcn: Look at gnatelim if you get a chance. That's what it does.
16:14:42 <onetom> u dont need all these at once...
16:15:51 <tcn> ramnull: a static library, shared lib, executable, or '.o' object file?
16:15:58 <ramnull> I'm picturing a sort of "run-time" gnatelim lite.
16:16:05 --- nick: kc-food -> kc5tja
16:16:19 <tcn> wb, kc
16:16:22 <kc5tja> re
16:16:55 <onetom> tcn: so as this experiment show, most of the time is spent on compiling the C part...
16:17:08 <ramnull> tcn: It works on either executables or object files. But it uses source "headers"(theres really no such thing as a header in Ada) to generate the parse tree.
16:17:40 <tcn> ramnull: only Ada programs, then?
16:18:10 <ramnull> tcn: gnatelim is a part of the GNAT GCC package.
16:18:17 <onetom> & dont forget that loading precompiled things also need time
16:18:21 <ramnull> tcn: I speaking conceptually here.
16:18:37 <onetom> thats why shell scripts r so slow...
16:19:29 <onetom> so if u add the 1time compilation + all-the-time load time
16:19:31 <ramnull> tcn: It was designed for embedded projects in large internetworked systems.
16:19:57 <onetom> u can clearly see that we should only compete w load time in case of 4th
16:20:40 <ramnull> onetom: And having a pre-"compiled" elimination tree would cut the load time down drastically.
16:20:45 <tcn> remmeber, GCC is *really slow*
16:21:04 <ramnull> tcn: Yes. Hence gnatelim was developed.
16:21:20 <tcn> I was better off with Turbo C in DOS, except for that 16 bit limit
16:21:21 <ramnull> I'm not suggesting using gnatelim.
16:21:55 <tcn> well, something similar would help
16:22:02 <ramnull> Rather I'm suggesting looking at it conceptually and seeing what features could be used for a Forth/C "run-time" linker.
16:23:09 <tcn> well, are we talking Linux/Unix or something like RetroForth?
16:23:15 <onetom> time for (( i=0 ; i<500 ; i++ )) ; do awk ; done
16:23:19 <onetom> real    0m0.762s
16:24:02 <ramnull> tcn: It could be adapted to any environment. At least theoretically.
16:24:02 <tcn> becuase in Linux we want to use shared libs and they're not gonna like being cut up
16:24:26 <tcn> and a simple OS is gonna be single address space and won't like shared libs
16:24:40 <onetom> awk: 95k
16:24:52 <onetom> *.fs *.fb: ~2M
16:25:49 <onetom> so loading a 95k program 500times takes roughly as much time as compiling 2M noncolor forth src
16:26:46 <tcn> I envision using old-style "ar" libs, but loading & linking only the needed parts, at runtime
16:28:04 <ramnull> tcn: You'd have to look at the Ada toolchain. Theres absolutely no intermarriage between the Language and the OS. Forth, by it's very nature is like that(from my perspective).
16:28:09 <onetom> gcc --version
16:28:09 <onetom> 2.95.4
16:28:36 <onetom> the measured time was referin 2 this (the faster) version of gcc
16:28:56 <onetom> btw, is there any free 2 use C 4 linux?
16:29:05 <onetom> ibm or intel c compiler?
16:29:18 <tcn> LCC
16:29:19 <onetom> has any1 tried it 2 compile, say mplayer w it?
16:29:42 <onetom> or u have 2 adjust too much compiler option 4 such a test?
16:29:48 <onetom> tcn: LCC what is it?
16:29:52 <onetom> url?
16:29:54 <tcn> oh, there's also TCC.. tiny C compiler.. you need glibc2.1 though
16:30:06 <onetom> ramnull: what is that elimination tree?
16:30:08 <a7r> there's also Tendra
16:30:50 <onetom> tcn: lcc doesnt seem 2 b the part of debian (woody) :/
16:31:06 <onetom> tcn: but i guess, it wont b able 2 compile mplayer
16:31:27 <onetom> tcn: coz its not as featureful as gcc
16:31:31 <tcn> i dunno.  I couldn't get LCC to c.. to work
16:31:52 <onetom> tcn: btw, thats said if the featurefulness of a compiler is
16:32:07 <ramnull> onetom: Basically it looks at the sources of two program sections(as text files, sources) and goes down the line recording all the dependencies, and then generates a specific compile chain including only the functions that are interdependent.
16:32:13 <onetom> tcn: is rational w its overall general compilation speed :)
16:32:48 <ramnull> onetome: There is no excess code or data included in the resulting executable or object file.
16:32:57 <onetom> ramnull: aaand its would b useful 4 C or 4th?
16:33:16 <onetom> s/its/it
16:34:00 <onetom> ramnull: so the result would b some structure of references (offsets) 2 the source?
16:34:06 <tcn> onetom: oh, you mean proportional to compilation speed.. "rational" means something different in English :)
16:34:09 <ramnull> onetom: Well, I'm not sure, yet, if such a thing is possible with C(it is with Forth though), as C sources shackle themselves to each other through vague #includes and #defines.
16:34:21 <onetom> tcn: hehe, yeah proportional, sorry
16:34:22 <ramnull> onetom: Exactly.
16:34:59 <onetom> ramnull: k, i c now. & have u heard of similary preparsed 4th code?
16:35:03 <tcn> uhh.. it could be reallyt tough w/ forth1
16:35:15 <tcn> !!!
16:35:25 <ramnull> onetom: Not yet. But theres a first time for everything.
16:35:26 --- quit: Speuler (Connection timed out)
16:35:46 <onetom> ramnull: coz jeff fox has described such a system...
16:36:09 <onetom> ramnull: probably he has also realized it in aha
16:36:13 <tcn> hm.. say you parse C, Forth, everything into linked lists.. use this internal format to generate machine code, check depencencies, whatever..
16:36:15 <ramnull> onetom: Maybe Jeff Fox and Robert Dewar run into each other in a coffee shop. Heh.
16:36:38 <tcn> then at least the compiler and the checker would always concur :)
16:36:44 <onetom> ramnull: jeff was talkin about a preparsing editor
16:36:59 <ramnull> Well, the thing with Ada is ...
16:37:01 <onetom> ramnull: what stores a preparsed src
16:37:26 <onetom> ramnull: what can b compiled a magnitude faster than pure txt src
16:37:38 --- quit: CaffeineJunkie (Read error: 110 (Connection timed out))
16:37:54 <onetom> ramnull: so compilation cost has been shifted 2 edit time this way
16:38:00 <ramnull> The input and output for each and every procedure/function is described in minute detail, so the compiler/eliminator can do the preparsing job with ease. With Forth it would be a bit more complex.
16:38:22 <tcn> that's an understatement :)
16:38:33 --- join: CaffeineJunkie (~Speuler@mnch-d9ba46f8.pool.mediaWays.net) joined #forth
16:38:34 <onetom> certainly introducing more complexity in2 the editor
16:39:04 <ramnull> Yet, since Forth is so simple, as it should be, technically it should be much more efficient. I'm speculating.
16:39:27 <tcn> it's hard to extend a preparsing editor..
16:39:42 <onetom> tho, u dont have too ;)
16:39:58 <onetom> s/tho/but
16:40:20 --- join: Speuler (~Speuler@mnch-d9ba46f8.pool.mediaWays.net) joined #forth
16:40:49 <onetom> but its still an area 2 be researched
16:41:27 <ramnull> I think the idea has merits, and should be investigated.
16:41:41 <onetom> a preparsed color4th editor would b really fun
16:41:50 <tcn> sait also makes it tough to generate code from another program..
16:42:06 <tcn> s/sait/it/
16:42:13 <onetom> but c4th is already very easy 2 compile, so its not a real urge 2 write such a beast
16:42:49 <ramnull> One could always have the parser take the code from the elim tree and generate an equivalent forth word/definition. That might kill the load time though.
16:42:59 <tcn> but regular text utilities DON'T WORK on colorforth code!
16:43:14 <onetom> tcn: why?
16:43:21 <tcn> it's not text
16:43:26 <onetom> it is
16:43:42 <tcn> not ascii
16:44:05 <onetom> just [:space:]!='[ \t\r\n]'
16:44:09 <kc5tja> BAUDOT CODES!!  (sorry, had to interject that)
16:44:31 <onetom> but [\001-\030]
16:45:13 <onetom> kc5tja: ;p
16:45:53 <onetom> tcn: i dont really c 2 much difference... do u still really c?
16:46:31 <tcn> oh, i thought you meant Huffman coded.. but even what you're saying, still won't work good with text utils..
16:46:51 <tcn> you'd have to rewrite 'em all
16:47:08 <ramnull> Anyways, I'm starting to think Chuck Moore and the Ada guys were seperated at birth.
16:47:27 <onetom> all all all ... dont always think in the traditional *nix environment
16:47:43 <onetom> ramnull: :)))
16:48:34 <tcn> oh well, we could go on for hours
16:49:14 <onetom> tcn: most other langs running under unices had 2 interface w kernel & lib functions somehow... terrible.. brrrr
16:49:48 <onetom> but python tcl pike perl whatever has all these interface files
16:49:50 <tcn> the syscalls are easy
16:49:54 <onetom> tonns of megabytes
16:50:17 <ramnull> onetom: Sorry. I just see so much in common between the two.
16:50:31 <tcn> well, see ya guys
16:50:37 <onetom> & u r whinning about hacking grep & libregexp a bit?... mmm
16:50:51 <onetom> ramnull: yeah, i got the puchline
16:50:53 <tcn> it's about dnner time
16:51:05 --- quit: tcn ("TinyIRC 1.1")
16:51:08 <onetom> have good apetite
16:51:09 <onetom> a
16:51:12 <onetom> eeh :)
16:51:12 --- join: tcn (~tcn@tc2-login24.megatrondata.com) joined #forth
16:51:17 --- quit: tcn (Client Quit)
16:51:19 <onetom> tcn: have a good apetite
16:51:23 <onetom> shit
16:51:30 * kc5tja laughs
16:51:58 <onetom> i just wanted 2 express him how i like him otherwise
16:52:15 <onetom> not counting disagreement, but havent really managed 2 :)
16:53:02 <ramnull> http://www.openravenscar.org  You'll see what I mean.
16:54:06 <ramnull> Hmmm...To fork Forth or clone it. Hmmm.
16:54:46 <onetom> wow
16:55:32 <onetom> so, the main idea behind ada is the programatical rightness verification of any kind of code?
16:55:45 <ramnull> onetom: essentially, yes.
16:55:47 <onetom> s/is the /is the ability of
16:55:56 <onetom> essentially? what else?
16:56:05 <onetom> interactivity?
16:56:25 <onetom> or its not that interactive as 4th or smalltalk?
16:56:41 <kc5tja> It's about as interactive as compiled C is.  :)
16:56:45 <ramnull> onetom: All that. It's designed, from the ground up, so that you can strip out everything except what you specifically need.
16:56:52 <onetom> (ive only seen squeak as smalltalk, but dont know much about it)
16:57:12 <onetom> kc5tja: :)
16:57:55 <ramnull> onetom: You will find no two identical dialects of Ada(like Forth), using the RavenScar profile you can statically verify every branch you program will ever possible take. Excepting of course a silicon meltdown.
16:58:43 <ramnull> onetom: Using elimination trees, I've Object Databases engines <5kbytes.
16:58:49 <ramnull> I've seen.
16:58:58 <ramnull> Got three terminals going at once.
17:00:50 <ramnull> Essentially, you've got the perfect pair Ada at the Macro project level, interfaced with Forth at the Micro project level.
17:01:08 <onetom> aaaaha! got it
17:05:59 <ramnull> ... and a hush filled the room.
17:07:26 * onetom is just transporting himself by bike
17:07:27 <onetom> bbl
17:07:38 <ramnull> onetom: Did I say something dumb?
17:07:46 <onetom> no
17:07:58 <ramnull> Oh.
17:08:02 <onetom> i just have 2 change workplace
17:08:13 <ramnull> Heh. Laters then.
17:08:41 <onetom> i have 2 monitor the net traffic of a 20box cabinet
17:08:53 <ramnull> I see.
17:37:08 <onetom> back
17:38:50 <onetom> the doorman was a askin me very surprised: "now???" when i woke him up & said: "hello, i came 2 work" :)
17:39:01 <onetom> its 2:43am here
17:41:22 <kc5tja> Heheh
17:43:46 <ramnull> onetom: You work third shift too eh?
17:44:05 <ramnull> onetom: Where at?
17:49:35 <ramnull> bbl
17:49:38 --- quit: ramnull ("This isn't Happy Hour!")
17:50:30 <onetom> :)
17:51:27 <onetom> just there was a wedding last day so iwas sleeping during daylight time
17:51:38 <onetom> anyway its already monday here ;p
17:55:24 --- join: karingo (karingo@43.portland-08-09rs.or.dial-access.att.net) joined #forth
18:26:13 --- quit: karingo ()
18:37:22 --- join: karingo (karingo@43.portland-08-09rs.or.dial-access.att.net) joined #forth
18:49:02 --- join: kapple (karingo@80.portland-01rh15rt.or.dial-access.att.net) joined #forth
19:01:03 --- quit: kapple ()
19:07:08 --- quit: karingo (Read error: 113 (No route to host))
19:08:05 --- join: karingo (karingo@80.portland-01rh15rt.or.dial-access.att.net) joined #forth
19:14:37 --- join: snowrichard (~cpu2@c66.190.103.147.ts46v-01.mrshll.tx.charter.com) joined #forth
19:14:48 <snowrichard> hello again
19:15:14 <kc5tja> re
19:16:42 <snowrichard> started re-implimenting my little language a little differently.  I figured out a more efficient way to code the definitions.
19:18:47 <kc5tja> Cool.
19:19:04 <snowrichard> this way I could have primitive byte codes ranging from 00 hex to EF hex.  User defined words would be coded as FXXX where XXX is a 12 bit index into the user def table. 
19:19:12 * kc5tja was doing more work on his ATM ARNI specification and development site.  I think I'll move it to its own sourceforge site when I get working code.
19:19:29 * kc5tja nods
19:19:37 <kc5tja> That's pretty reasonable to me.
19:19:48 <kc5tja> I envision most bytecodes would end up as primitives anyway.
19:19:49 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth
19:20:18 <snowrichard> since I am writing this all in C for now, by primitive I mean "a C function" .
19:20:36 * kc5tja nods
19:20:55 <kc5tja> I have no problems with Forth systems written in C.  I may end up doing the same for my port of THE to Linux.
19:21:31 <snowrichard> working on the parser part now, but need to rest some tonight. 
19:22:03 * kc5tja nods
19:22:16 <snowrichard> I do my own version of extreme programming - top down development using stubs that basically say "this isn't done yet" and quit. 
19:22:21 <ramnull> Writing some new Forths eh?
19:23:07 <kc5tja> snowrichard: Heh, I don't even do that.
19:23:29 <kc5tja> My XP is to implement what I need, when I need it.  I really have little use for stubs overall.  Plus *LOTS* of unit tests.
19:24:22 <snowrichard> so I can use this stub idea in my little language, I could encode a definition that isn't done yet so that if it runs, it would tell me its not defined and ask if I want to run the editor. 
19:25:37 <snowrichard> then compile the file I just edited and try the test again
19:26:04 <ramnull> Aaarg. I can get most of the syscalls to work. getdents is being a bitch though.
19:26:30 <snowrichard> where could I find a reference on netbsd/linux syscalls? (I know that's two questions) 
19:26:47 <kc5tja> Good question.
19:26:49 <ramnull> snowrichard: /usr/include/asm/unistd.h
19:26:50 <kc5tja> I have no idea.
19:26:57 <kc5tja> What does getdents() do?
19:27:16 <ramnull> snowrichard: Then you can do man 2 <syscall> at the shell.
19:27:19 <snowrichard> it ran on my car when a  beer truck cut into my lane. 
19:27:54 <kc5tja> Heh
19:27:58 <ramnull> kc5tja: I gets directory entries from the filesystem.
19:28:08 <kc5tja> After arguing with the guy about the damage to the car, I would have asked him for a beer.  :D
19:28:16 <kc5tja> Ahh
19:28:33 <snowrichard> unistd.h has the syscall numbers? ok thanks (linux I guess) 
19:28:33 <onetom> is there anybody here who already has some experience w menuetos?
19:28:52 <snowrichard> he didn't stop
19:28:53 <ramnull> snowrichard: Yeah, should work for BSD too. Dont know about AIX or Solaris.
19:28:59 <kc5tja> Menuetos -- the FRESHMAKER!!  (joke)
19:29:10 <onetom> i think, 1 should write some color4th 4 it
19:29:35 <onetom> it seems 2 b a very convenient framework 4 such development
19:29:58 <onetom> it already has tcpip support too...
19:30:31 <snowrichard> the reason I stayed up so long was I finally got the simh networking to work right. I installed bsd on a simulated vax
19:30:48 <kc5tja> heh :)
19:31:18 <snowrichard> there is a little deal about attaching the simulated ethernet to the real one and I went back into the docs trying to find it again and couldn't 
19:31:30 <onetom> ramnull: menuetos.org?
19:32:46 <snowrichard> If I had lots of extra money I would pick up one of those sharp zaurus pda things for about $300.  but I don't .
19:33:07 <ramnull> onetom: I downloaded and played with it a couple times. Dont know much about it though.
19:33:20 * kc5tja nods
19:33:35 <ramnull> Damn. getdents keeps returning -1.
19:33:36 <kc5tja> I ended up getting a Sony Clie for half that price, with similar features.  
19:33:55 <snowrichard> but does it run linux or windows ce ?
19:34:42 <ramnull> I give it the file descriptor, the buffer, and the count. no luck.
19:36:09 <kc5tja> No, it runs PalmOS 4.1, which is still a nice environment, even if its memory management is a bit ... weird.  :)
19:36:55 <onetom> ramnull: sad... so, i left wo company :)
19:37:47 <snowrichard> I have never seen a PALM os machine it might be as nice as some of the other small machines
19:38:34 <ramnull> Anyone here running Linux?
19:38:42 <snowrichard> I  am running linux now
19:38:57 <snowrichard> I just pulled up that manual page a second ago
19:39:28 <ramnull> snowrichard: Is it telling me to read FROM the structure or INTO the structure?
19:40:11 <snowrichard> reads several dirent structures from the directory pointed at
19:40:11 <snowrichard>        by fd into the memory area pointed to by dirp.
19:40:12 <ramnull> I'm getting the file descriptor using fopen.
19:40:24 * kc5tja is running LInux
19:40:33 <ramnull> Well, I got it set up right then.
19:41:00 <snowrichard> open returns an fd, fopen returns a  FILE * 
19:41:05 --- quit: karingo ("you all shall join me in the end")
19:41:14 <ramnull> Just a sec...
19:41:21 <kc5tja> The lightbulb goes on...
19:41:23 <kc5tja> ;D
19:41:47 <ramnull> I'm going "0 apath <open> path !"
19:42:29 <kc5tja> Don't you need to use diropen() instead?
19:42:34 --- join: kitsune (foxchip@adsl-209-182-168-45.value.net) joined #forth
19:42:48 <kc5tja> OK, never mind.  diropen doesn't appear to exist on Linux.  :(
19:42:55 <kc5tja> re kitsune :)
19:43:04 <kitsune> hai
19:44:04 <ramnull> I just pass the file descriptor stored in path plus the buffer address plus how many bytes I want it to read.
19:44:05 <kc5tja> kitsune: I'm right pissed.  >:(  My car broke down on the way to aikido last Friday.  Turns out I was running off of the battery -- the alternator wasn't working.  So now my battery is dead, and the alternator needs to be replaced.
19:45:04 <snowrichard> and the labor is more expensive than the alternator itself usually the way some of those belts are
19:45:12 <ramnull> 3 141 syscall getdents
19:45:38 <ramnull> fchdir works fine.
19:46:29 <kc5tja> RX-7s are trivially easy to change alternators on.
19:46:37 <kc5tja> Labor isn't going to be that much.
19:47:03 <snowrichard> actually we have a shop here that does a good job on most cars for under $75 for the whole deal
19:47:32 <snowrichard> if you want the rebuilt one
19:47:40 <kc5tja> No, I always get the new ones.
19:47:44 <kc5tja> The one I have in here now is rebuilt.
19:47:51 <kc5tja> And this is why I never get rebuilt ones.  :)
19:47:55 <kc5tja> (I bought the car with it)
19:48:18 <kc5tja> It'll probably cost me $200 or so for a Mazda alternator, including labor.
19:48:28 <kc5tja> If I had the tools, I'd do it myself.
19:48:36 <kc5tja> (and the time)
19:49:03 <snowrichard> I am a mechanical klutz.  repairing computers is about the limit of my ability that way. 
19:49:19 <kc5tja> Heheh :)
19:49:25 <ramnull> Lets see here. Read is 1, Write is 2, Read/Write is 3?
19:49:42 <kc5tja> Well, I blew my car's sound system trying to replace the power antenna after it got damaged...does that count?  And I'm quite proficiant(sp?) with electronics too.  
19:49:51 <kc5tja> It's just that Mazda's wiring is ... rather nonstandard.  :D
19:50:21 <kc5tja> ramnull: rwx is the bit order for permissions.  read is 4, write is 2, read/write is 6. 
19:50:34 <kc5tja> ramnull: I'm *assuming* you're referring to Linux permissions.
19:50:42 <ramnull> Thats right, now I remember.
19:53:06 <ramnull> I'm opening the directory with read perms, but getdents still returns error.
19:53:27 <snowrichard> you doing this in forth?
19:53:32 <snowrichard> (duck)
19:53:33 <ramnull> I'll probably bug I440r about it. I'm doing something wrong.
19:53:53 <ramnull> Yeah. Heh.
19:53:59 <kc5tja> ramnull: Do you even have read permissions on the directory?  :)  (e.g., is the power cord plugged into the wall socket?)
19:54:28 <snowrichard> in the isforth?
19:54:29 <ramnull> kc5tja: Well, I'm using my homedir, so yeah.
19:54:34 * ramnull nods
19:54:39 <kc5tja> OK, Just Making Sure (tm).  :D
19:55:06 <kc5tja> I've done really stupid stuff in Linux like that before, spent hours and hours figuring out why it didn't work, only for it to be something as simple as a single permission bit on a directory.
19:55:07 <kc5tja> :)
19:55:11 <ramnull> I'm not opening it with write perms, so I should be safe.
19:55:20 * kc5tja nods
19:55:35 <ramnull> Or rather in write mode.
19:56:14 <snowrichard> did you check the status of the open to see if it did open it?
19:56:17 <ramnull> Hmmm...I just thought of something. Be back in a second.
19:56:36 <ramnull> snowrichard: It opens alright. fchdir syscall works too.
19:57:54 <ramnull> That didnt work.
19:58:15 <ramnull> Was checking to see if it could only see my CWD.
19:58:38 <snowrichard> well I need to be getting ready for bed when I finish that last cup of coffee
19:58:58 <ramnull> Try not to let that coffee keep you awake.
19:59:15 <ramnull> I know it's something simple.
19:59:19 <snowrichard> I think I have ADHD or something, drink enough and I get sleepy
20:00:03 <snowrichard> actually i am taking some medication that makes me sleep
20:01:47 <ramnull> Integers in Forth are assumed to be unsigned?
20:03:07 <snowrichard> why
20:04:09 <snowrichard> never mind... need to go now
20:04:13 <snowrichard> bye
20:04:15 --- quit: snowrichard ("using sirc version 2.211+KSIRC/1.2.4")
20:15:05 <ramnull> bbl
20:15:08 --- quit: ramnull ("This isn't Happy Hour!")
20:50:48 --- quit: onetom (sterling.freenode.net irc.freenode.net)
20:52:55 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth
20:58:38 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
21:06:33 --- quit: kitsune ()
21:34:27 --- quit: njd (Read error: 54 (Connection reset by peer))
22:01:05 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth
22:12:01 --- quit: ramnull ("This isn't Happy Hour!")
22:47:51 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth
22:48:07 <ramnull> You guys all lurking?
22:48:27 <onetom> no, me not
22:48:35 <onetom> im still workin
22:50:31 <Robert> Hey
22:50:40 <Robert> I'll go to school in ~20 min.
22:51:04 <onetom> i WAS in the school all the night ;)
22:51:38 <ramnull> I'm trying to figure out where IsForth put's it's errno value. The source code offsets it 5 from eax.
22:51:52 <ramnull> I gotta work tommorrow night.
22:55:23 <Robert> Hmmm.
22:56:34 <ramnull> Think I found it.
22:56:53 <ramnull> Well, the same number comes up everytime I try it.
22:56:57 <ramnull> Heh.
22:57:10 <ramnull> errno 8
22:57:17 <ramnull> endexec
22:57:33 <Robert> Hmm..
22:57:50 <ramnull> A nice little cryptic explanation "Exec format error".
22:57:55 <Robert> errno+5, that's the body of the word, right?
22:58:02 <ramnull> Yeah.
22:58:13 <ramnull> I'm just doing here 5 + c@
22:58:17 <ramnull> To get it.
22:59:09 <Robert> Hm?
22:59:23 <Robert> "here", where? :)
23:00:44 <ramnull> After executing 3 141 syscall getdents
23:01:01 <ramnull> errr...
23:01:11 <ramnull> 3 141 syscall getdents here !
23:01:27 <ramnull> Saves the same number in the buffer pointer too.
23:02:10 <onetom> pi syscall ;)
23:02:28 <ramnull> Now that you mention it. Heh.
23:02:46 <ramnull> Coincidence thats the only one I'm having a problem with?
23:03:34 <ramnull> I'm gonna go google for ENOEXEC
23:03:54 <Robert> Time for some school, bye.
23:04:59 <ramnull> Later Robert.
23:05:56 <ramnull> Alright. Does this make sense http://www.wlug.org.nz/ENOEXEC  ?
23:06:20 <ramnull> I mean, I'm using an interpreter.
23:06:28 <ramnull> IsForth.
23:11:06 <ramnull> Oh shit.
23:12:12 <ramnull> I gotta register the Forth interpreter with binfmt/register under /proc
23:12:52 <ramnull> Aint that some shit.
23:13:32 <ramnull> Anyone seen I440r?
23:32:44 <ramnull> Yup. I got it all figured out now.
23:35:17 <onetom> then tell us the solution
23:36:20 <ramnull> Well, I'm working on implementing a solution. I'll let yah know if it works. Gotta pass a magic number to the kernel for the IsForth interpreter.
23:36:31 <ramnull> Using binfmt_misc.
23:36:55 <ramnull> I could do it from a shell script I suppose.
23:37:38 <onetom> yup
23:37:46 <onetom> google can always help...
23:38:00 <ramnull> onetom: I already found the docs I need.
23:38:14 <onetom> & also an rgrep binfmt_misc /usr/src/linux/Doc*
23:38:31 <ramnull> Just finished reading it.
23:38:44 <onetom> & how have u found it?
23:39:01 <ramnull> http://www.tat.physik.uni-tuebingen.de/~rguenth/linux/binfmt_misc.html  Is the URL.
23:39:18 <ramnull> Simple. If a bit ugly.
23:39:49 <ramnull> You'll find it under /proc/sys/fs
23:40:24 <ramnull> Hence I suspect this problem is only limited to filesystem calls.
23:58:47 --- quit: ramnull (Remote closed the connection)
23:59:59 --- log: ended forth/03.04.27