00:00:00 --- log: started forth/05.03.20
00:05:41 --- join: hyrax_ (~chatzilla@adsl-64-219-100-33.dsl.lgvwtx.swbell.net) joined #forth
00:10:03 --- join: Topaz (~top@spc1-horn1-6-0-cust219.cosh.broadband.ntl.com) joined #forth
00:22:10 --- quit: hyrax (Read error: 110 (Connection timed out))
00:43:04 --- quit: Topaz ("Leaving")
01:09:47 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth
01:43:02 --- quit: zardon ("leaving")
01:58:15 --- quit: hyrax_ (Read error: 54 (Connection reset by peer))
03:03:03 --- join: aum (~aum@60-234-138-239.bitstream.orcon.net.nz) joined #forth
04:18:24 --- join: Baughn (~svein@cloud.brage.info) joined #forth
05:56:58 --- quit: aum ()
06:38:10 --- join: hyrax (~chatzilla@adsl-64-219-100-33.dsl.lgvwtx.swbell.net) joined #forth
06:38:33 <hyrax> join #gentoo-amd64
06:38:44 <hyrax> :)
07:14:02 --- join: BACbKA (~vassilii@CBL217-132-156-157.bb.netvision.net.il) joined #forth
07:14:16 <BACbKA> is it legal for : to put things on the stack during the compilation?
07:15:15 <BACbKA> (gforth does it, kforth doesn't do it)
07:15:27 <BACbKA> : foo [ .s ] ;
07:17:50 <hyrax> hmm...  the : itself?  your foo example is ok. I don't think : should.
07:18:22 <hyrax> Because it is common to put something on the stack before a word, and use it in the word.
07:18:41 <hyrax> well, maybe on common, but it is done
07:19:21 <BACbKA> Hyrax:I was used to doing this before, and suddenly I came to gforth which breaks it :( 
07:20:17 <BACbKA> just a moment I will switch compuiters
07:20:18 <hyrax> confusing... because then instead of LITERAL , you would have to say [ SWAP ] LITERAL , or some other nonsense.
07:20:19 --- quit: BACbKA ()
07:33:49 --- join: BACbKA (vassilii@res2.iGuide.co.il) joined #forth
08:03:24 --- join: docl (~docl@dpcbw098235.direcpc.com) joined #forth
08:57:30 <BACbKA> : chk-depth depth 0= IF ." sane" ELSE ." sick" THEN ; immediate
08:57:35 <BACbKA> : nop chk-depth ;
08:58:35 <BACbKA> Are systems that print "sick" standards-compliant? can somebody please give me a reference to the standard? (FYI: gforth 0.6.2 is "sick" as per the above)
09:01:23 <BACbKA> minforth is "sick"
09:01:37 <BACbKA> pforth yforth and kforth are "sane" :-)
09:18:33 --- join: tathi (~josh@pcp01375108pcs.milfrd01.pa.comcast.net) joined #forth
09:41:31 <crc> hi all
09:52:20 <BACbKA> OK, I got it. The thing on the stack is legal and is called "colon-sys" in DPANS94. Yuck....
09:54:04 <hyrax> eh? I've not seen that before.
09:54:41 <hyrax> well, I might have see reference to "colon-sys" but I've not seen compilers put it on the stack
09:55:59 <hyrax> : foo [ .s ] ;
09:56:11 <hyrax> : foo [ .s ] ;
09:56:25 <hyrax> <-Top  ok
09:56:39 <hyrax> nothing on the stack, that is with SwiftX
09:56:49 <BACbKA> : chk-depth depth 0= IF ." sane" ELSE ." sick" THEN ; immediate : nop chk-depth ;
09:57:06 <BACbKA> This one gives "sick" for gforth and minforth
09:57:14 <BACbKA> and "sane" for everything else I have tried
09:57:33 <BACbKA> But yeah, the sick ones are compliant nevertheless
09:57:58 <hyrax> : chk-depth depth 0= IF ." sane" ELSE ." sick" THEN ; immediate : nop chk-depth ; sane ok
10:01:04 <hyrax> I guess you could maybe do ": : >r ;"  and  ": ; r> POSTPONE ; ; IMMEDIATE" or some other convulated work around
10:01:46 --- quit: docl (Read error: 104 (Connection reset by peer))
10:03:04 <BACbKA> hyrax: this should be done with everyone doing colon-sys, and also you're assuming colon-sys 1 cell wide :-)
10:03:11 <BACbKA> hyrax: I'd say smth like
10:04:09 <tathi> BACbKA: when is this a problem, anyway?
10:04:26 <BACbKA> tathi: I was teaching somebody forth today
10:04:50 <BACbKA> tathi: showed : 1+! ( a -- ) dup @ 1+ swap ! ;
10:05:12 <hyrax> Well, you could do the sick/sane check, and determine how many cells colon-sys is, and then make the corrections in the overloaded : and ;
10:05:24 <BACbKA> tathi: then showed how to do a creator for words like this along the following i/face
10:05:25 <BACbKA> ' 1- modify!: 1-!
10:05:41 <BACbKA> tathi: first with create (which the student got right away)
10:05:55 <BACbKA> tathi: then I said -- watch this -- and wrote the following
10:06:13 <BACbKA> : modify!, ( xtModifier -- )
10:06:13 <BACbKA>     ['] dup compile,
10:06:13 <BACbKA>     ['] @ compile,
10:06:13 <BACbKA>     compile,
10:06:13 <BACbKA>     ['] swap compile,
10:06:14 <BACbKA>     ['] ! compile, ; immediate
10:06:19 <BACbKA> : 1+! [ ' 1+ ] modify!, ;
10:06:39 <BACbKA> then (again) I said "watch this"
10:06:41 <BACbKA> and wrote
10:07:13 <BACbKA> : modify!: ( xtModifier -- )
10:07:13 <BACbKA>     :
10:07:13 <BACbKA>     postpone modify!,
10:07:13 <BACbKA>     postpone ;
10:07:13 <BACbKA>     ;
10:07:22 <BACbKA> ' 1- modify!: 1-!
10:07:27 <BACbKA> and it crashed ;-)
10:07:41 <BACbKA> what works is : modify!: ( xtModifier -- ) >R
10:07:41 <BACbKA>     :
10:07:41 <BACbKA>     R> postpone modify!,
10:07:41 <BACbKA>     postpone ;
10:07:41 <BACbKA>     ;
10:08:33 <BACbKA> hyrax: yeah
10:08:53 <BACbKA> hyrax: pity there is nothing like limits.h in forth
10:08:56 <tathi> huh. ok.
10:12:17 <hyrax> BACbKA: well, the sick/sane check could figure that out, but the equiv of limits.h would be better
10:21:02 <BACbKA> hyrax: I can create smth like that myself:
10:21:15 <BACbKA> variable sizeof
10:22:05 <BACbKA> depth : TEST [ depth sizeof ! ] ; sizeof @ swap - emit
10:22:10 <BACbKA> s/emit/./
10:22:47 <BACbKA> and use this to generate forth code producing
10:23:04 <BACbKA> <system-dependant> constant colon-sys-cells
10:23:13 <BACbKA> etc for all other control structures
10:23:24 <BACbKA> and then instead of >R : R>
10:23:31 <BACbKA> one would do smth like 
10:24:29 <BACbKA> : my-word ... : colon-sys-cells roll  ...
10:24:49 <BACbKA> but a standard extension for those things would be great
10:26:07 --- nick: hyrax -> hyrax|afk
10:28:01 --- join: Zymurgy (thrud@TBay-ppp42352.sympatico.ca) joined #forth
10:36:08 --- quit: Zymurgy ("Leaving")
10:37:53 --- join: Zymurgy (zymurgy@TBay-ppp42352.sympatico.ca) joined #forth
10:47:41 --- part: Zymurgy left #forth
10:51:32 --- join: Sonarman (~snofs@adsl-64-160-165-139.dsl.snfc21.pacbell.net) joined #forth
11:06:08 --- join: Topaz (~top@spc1-horn1-6-0-cust219.cosh.broadband.ntl.com) joined #forth
11:21:42 --- quit: Topaz (Read error: 104 (Connection reset by peer))
11:37:28 --- join: Topaz (~top@spc1-horn1-6-0-cust219.cosh.broadband.ntl.com) joined #forth
11:45:44 --- nick: hyrax|afk -> hyrax
11:46:34 <BACbKA> hyrax: I have browsed some gforth-specific sources and the `info gforth' doc, and realized why the folks there don't suffer. They use to have a locals idiom to use in these cases, see the "Advanced macros Tutorial" in the info. Still insane imho.
11:47:54 <BACbKA> I wonder if there is an easy way to re-write complicated macros with EVALUATE?
11:48:19 <hyrax> I've used evaulte for such things
11:48:28 <hyrax> evaluate
11:49:13 <hyrax> but if create words in an evaluate, they can't be located
11:49:41 <hyrax> so I've used interpret
11:51:59 <hyrax> eh, no forths on native x86_64
11:52:11 <BACbKA> hyrax: say, how would you re-write modify!, ?
11:53:50 * hyrax is looking at modify!
11:54:22 <BACbKA> hyrax: INTERPRET is no longer there...
11:57:14 <hyrax> eh... no gforth on cygwin either
12:01:46 <hyrax> ok, your modify!: works as on the forth I'm using
12:02:39 <hyrax> So modify!: just makes the resulting word do a store, but runs the modifier first
12:15:12 <hyrax> I'm not sure of a clean way to do it where the modifier xt is passed in on the stack, rather than eaten after the word
12:15:51 <hyrax> well, clean, as long as colon-sys is not in the way
12:22:08 <Quartus> I would think it's a fairly common implementation technique to a) use the data stack as the control-stack and b) place colon-sys there.
12:23:07 <hyrax> not sure how I would do. Normally I don't write words like modify!:
12:23:47 <hyrax> how I would do it
12:55:06 --- nick: hyrax -> hyrax|afk
13:17:51 --- quit: ooo (Read error: 110 (Connection timed out))
14:35:53 --- nick: hyrax|afk -> hyrax
14:41:10 --- quit: qFox ("this quit is sponsored by somebody!")
14:47:29 <BACbKA> I am in an immediate word. What string do I have to give to EVALUATE as if I put ['] DUP COMPILE, into my immediate word execution semantics?
14:48:04 <BACbKA> (i.e., in order to have a call to DUP compiled into the currently compiled open definition)
15:03:04 <BACbKA> ok, I pull it off
15:17:43 <BACbKA> hyrax: want a generic macro generator that uses EVALUATE?
15:18:52 <hyrax> EH.. you read my mind. I was pulling a generic macro generator out of some stuff I did at work. (have to scrub it bit though, as it references too much stuff)
15:18:56 --- quit: Topaz (Remote closed the connection)
15:19:11 <BACbKA> is 25 lines flood OK here?
15:19:45 <hyrax> uh, use a paste site, that is a bit much, plus paste site is easier to copy
15:20:02 <hyrax> http://rafb.net/paste/
15:20:34 <BACbKA> hyrax: http://www.tarunz.org/~vassilii/modifier-maker.fs
15:21:35 <BACbKA> reload pls I have updated 2 comments
15:21:46 <hyrax> ok, that is better than mine. Mine evaluated it all on the HOST  (as I was using on a cross compiler). On the target it was compiled though.
15:26:09 --- join: ooo (~o@dsl-olugw3p33.dial.inet.fi) joined #forth
15:26:18 <hyrax> ok, COMPILE:, eats till ;
15:26:24 <BACbKA> aha
15:26:42 <BACbKA> see 1+!    is your friend to start with
15:27:36 <BACbKA> you can also put            2DUP ." EVALUATE(" TYPE ." )" CR      before the EVALUATE
15:30:48 <BACbKA> I would be happy to do the buffer stuff not in the dictionary space, but I don't know how (short of dynamic memory extensions usage) -- things like PAD get stomped when compilation progresses
15:35:31 <hyrax> yah I ran into that problem on mine too... Hmm, my forth is choking on COMPILE,:
15:35:43 <BACbKA> hyrax: how
15:36:22 --- join: aum (~aum@60-234-138-239.bitstream.orcon.net.nz) joined #forth
15:36:26 <hyrax> hold on, let me include instead of pasting it
15:36:37 <BACbKA> hyrax: compiling COMPILE,: or running it?
15:36:50 <hyrax> compiling it
15:36:55 <BACbKA> hmmmmmmm
15:37:22 <hyrax> it might be me, I'm going to include it and see
15:39:10 <hyrax> ok, it compiled now, I got a 1
15:39:54 <BACbKA> hyrax: good. Does SEE work on your system?
15:40:40 <BACbKA> hyrax: if yes, what do the SEE statements at the end produce? the thing similar to the 1st (commented out) line of the file?
15:41:11 <hyrax> not on this one, I'll have to copy a forth from a different computer to this one to get SEE
15:41:13 <BACbKA> hyrax: meanwhile I have discovered that yforth goes bananas on this file :) never gets out of compilation of COMPILE,:
15:42:18 <hyrax> compile:, does not seem to like literals
15:42:34 <hyrax> compile,: I mean
15:46:33 <BACbKA> hyrax: yeah, right now COMPILE,: W1 W2 W3 ; is just an alias to ['] W1 COMPILE, ['] W2 COMPILE, ['] W3 COMPILE,
15:46:54 <hyrax> yah
15:48:13 <hyrax> on my macro expander I modify SOURCE and let the compiler compile in the text
15:48:17 <BACbKA> minforth has surprised me. Had no problems with [IF] and [ELSE], but choked on [ENDIF]... oooh, it should have been [THEN] !!!!
15:48:36 <hyrax> same here, I had the change [ENDIF] to [THEN]
15:49:05 <hyrax> but mine is not generic, only works on a cross compiler like SwiftX
15:49:27 <hyrax> I was trying to make mine generic when you asked me if I a generic one
15:49:53 <BACbKA> hyrax: that is what I used to do on our own F83 system on the Hobbit computer... but I read DPANS as this is now kinda unwelcome :(
15:50:46 <BACbKA> that == modifying source
15:51:56 <hyrax> Mine plays games with 'TIB, >IN, and #TIB
15:54:46 <BACbKA> hehe... kforth didn't like it because it doesn't know HERE 
15:55:58 <hyrax> http://rafb.net/paste/results/MLjW4w10.html
15:56:34 <hyrax> But that is only partial
15:58:43 <BACbKA> I440r: I'm running Debian. Pity you don't have a package -- I would have had it installed and tried out already, as every other *forth* package :)
15:59:08 <BACbKA> I440r: aha, I see -- it's x86 only
15:59:14 <I440r> BACbKA, its not ready for packaging yet
15:59:20 <I440r> actually its been ported to ppc too
16:01:34 <BACbKA> The links session is impressive ;)
16:02:31 <BACbKA> the dragon is nice, too
16:02:49 <I440r> heh
16:03:11 <I440r> ive never looked at my page with links
16:03:14 * I440r looks
16:03:45 <I440r> it doesnt quite look right but its still ok :)
16:05:08 <BACbKA> I440r: I first began learning forth back in the 6th grade, ie 1988
16:05:23 <I440r> thats about when i learned it too
16:05:39 <I440r> but i didnt like f83, prefered the flat files used in fpc
16:06:02 <I440r> tom zimmer has featureitis tho and i cut is 284562456k f.com down to aobut 35k or something
16:08:50 <BACbKA> I440r our Hobbit forth (/3.5 Mhz Z80) beat FPC (/25MHz AT286) pretty much consistently back in 1990 :)
16:09:21 <BACbKA> we used subroutine threading and a lot of optimizations
16:09:36 <BACbKA> OK, I've finished the readme
16:09:42 <I440r> oh... im against that, isforth is 100% direct threaded
16:09:47 <I440r> sub threading breaks see
16:10:00 <I440r> and i consider see to be a neat debug tool
16:10:13 <I440r> also helps remind you how word xyz is defined
16:10:55 <BACbKA> I440r: remember -- these were the days w/o caches :)
16:11:25 <I440r> :)
16:12:31 <BACbKA> fetching nasm
16:14:00 <BACbKA> running
16:14:00 <hyrax> I440r: does isforth run on amd64 yet?
16:14:33 <BACbKA> I440r: you're a good boy, with >NAME NAME>
16:14:49 <BACbKA> BEHEAD :-)))
16:15:27 <BACbKA> ugh... no disasm of unnamed ones...
16:15:27 <hyrax> well, I installed nasm, so I'll try it
16:16:40 <BACbKA> What is "MIMIV" in the upper right corner? it's not a valid Roman numeral at least...
16:17:11 <hyrax> hmm...   ".   . ?"
16:18:16 <BACbKA> I440r: how do I source a forth file from the interpreter?
16:18:40 <BACbKA> (w/o defining a string const with its name in the dictionary)?
16:19:39 <hyrax> well, it does not compile on amd64 http://rafb.net/paste/results/HNKt7m78.html
16:19:51 <hyrax> ld: warning: i386 architecture of input file `../kernel.o' is incompatible with i386:x86-64 output
16:23:08 <BACbKA> IsForth -- another "sane" one (empty colon-sys) but YUCK!!! it doesn't accept uppercase :(
16:24:14 <hyrax> well, it does compile it, but I get segmentation fault when I go to run it
16:25:36 <BACbKA> hyrax: yeah, I have seen the readme
16:26:29 <hyrax> well, the one compiled for i386 works. Interesting, that is only i386 binary that runs on my machine. :)
16:27:14 <hyrax> I guess that is because it does not link to any libraries and only uses syscalls
16:28:48 <hyrax> ok, extend worked, now I have a working isforth
16:29:45 <I440r> BACbKA, fload
16:30:00 <I440r> fload path/to/file/blah.f
16:30:05 <I440r> or from the command line
16:30:12 <I440r> ./isforth -fload src/blah.f
16:30:28 <I440r> or you can put #! /path/to/isforth -sfload
16:30:44 <I440r> as the first line of the source file and chmod +x that file and execute it directly
16:31:43 <BACbKA> I440r: yeah, I had it figured out already from the extend script
16:31:54 <I440r> :)
16:32:05 <I440r> how fast a box do you run on ?
16:32:48 <hyrax> time ./extend > /dev/null
16:33:10 --- quit: tathi ("leaving")
16:33:24 <hyrax> real    0m0.089s ; user    0m0.083s ; sys     0m0.006s
16:33:37 <I440r> i was gona suggest running the window.f and itowers.f stuff
16:33:58 <I440r> assuming your not runiing from a screen session that is
16:34:11 <I440r> screen is seriously broken and im not fixing isforth to cater to it
16:34:53 <hyrax> running in an rxvt, but the box is on the network
16:35:03 <I440r> yea
16:35:07 <I440r> rxvt should be fine
16:36:54 <BACbKA> I440r: it chokes on my nasty code :)
16:37:02 <BACbKA> misses bl in the dictionary
16:37:04 <I440r> show me your nasty code ?
16:38:50 <BACbKA> http://rafb.net/paste/results/L31i2P28.html
16:41:22 <I440r> ok isforth doesnt use verbal pukeage like [char]
16:41:27 <I440r> [char] x
16:41:35 <I440r> use 'x' instead
16:41:38 <I440r> note... not 'x
16:41:54 <I440r> also s" will never be implemented in isforth :)
16:42:04 <I440r> nor will evaluate
16:42:22 <I440r> havent seen a postpone in there yet but if its there it wont work either
16:42:40 <I440r> yea. you have a postpone :/
16:43:00 <I440r> oooh i see an anti ans item in your code hehe
16:43:22 <I440r> you cant do r> at the start of a definition because on entry into ANY definition youc annot guarantee that the top item of the return stack is your return address
16:43:39 <hyrax> that is >r
16:43:51 <BACbKA> I440r I am afraid you've misread the code around that >R
16:43:52 <I440r> and if that was the only thing wrong with ans forth (which it isnt) i wouldnt use it
16:43:53 <I440r> oh duh
16:43:57 <I440r> im used to seeing
16:43:59 <I440r> : blah
16:44:02 <I440r>  code starts here ;
16:44:06 <I440r> not
16:44:18 <I440r> : blah code
16:44:22 <I440r>  starts here ;
16:44:27 <hyrax> : is in a colon word
16:45:48 <I440r> putting : inside a colon definition is perfectly valid :)
16:46:03 <I440r> ive even seen
16:46:05 <I440r> : : : ;
16:46:06 <I440r> lol
16:46:07 <crc> except in retroforth, where it defines a new entry point :)
16:46:47 <BACbKA> I'm afraid : : : ; is ambiguous as the 2nd inner colon creates an entry while the 1st has never been closed
16:47:12 <hyrax> no, the second does not
16:47:20 <I440r> no : startsx a new definition. : is the name of that new definition : is compiled into taht definition and ; ends it
16:47:24 <hyrax> it creates an entry when the word is run
16:47:25 <hyrax> ?
16:47:26 <I440r> nothing wrong with : : : ;
16:47:47 <I440r> hyrax i believe he means you can create an entry point in the middle of a word
16:47:51 <BACbKA> I440r oic, you say that : can be immediate
16:47:57 <I440r> : blah ...... : foo ...... ;
16:48:14 <I440r> blah executes both the code for blah and the code for foo. but you can also call just foo
16:48:17 <I440r> i consider that bad form
16:48:22 <I440r> foo should be factored out
16:48:27 <hyrax> oh, I've never seen : as immediate on any forth I've worked on
16:48:33 <I440r> : isnt immediate
16:48:53 <crc> retroforth has an immediate :
16:48:57 <I440r> : : : ;    first colon starts a new definition called :
16:49:11 <I440r> the third colon is compiled INTO the new definition and the ; finishes it
16:49:45 <hyrax> I440r: that code compiled and ran ok on SwiftForth
16:49:53 <I440r> er yea i was talking about retro forth above lol
16:50:12 <I440r> i consider retro's behaviour to be bad form
16:50:28 <I440r> hyrax only on retro would it be broken :)
16:50:39 <I440r> crc you should really change that btw
16:50:43 <I440r> its very bad form
16:51:02 <I440r> instead of creating an entry point into the middle of a definition you should do
16:51:07 <I440r> : foo ...... ;
16:51:12 <I440r> : bar ..... foo ;
16:51:14 <hyrax> One can always do : :: ; IMMEDIATE   :)
16:51:29 <I440r> not : bar ..... : foo ..... ;   <-- this is burying the definition for foo!!!!
16:51:40 <hyrax> er : :: : ; IMMEDIATE 
16:52:07 <I440r> i would just do ' : alias :: immediate
16:52:08 --- quit: aum ()
16:52:16 <I440r> and make the alias an immediate word
16:52:17 <BACbKA> OMG i have to teach in 9 hours... better go get some sleep
16:52:18 <crc> I440r: it allows me to do some things more efficiently
16:52:30 <I440r> crc but more obfuscated
16:52:43 <I440r> efficiency should always play second fiddle to readability
16:52:47 <BACbKA> good night everybody
16:52:53 <I440r> night BACbKA 
16:53:07 <BACbKA> bye
16:53:10 --- quit: BACbKA ("Leaving")
16:53:16 <hyrax> obfuscated forth? I'm shocked! How could anybody write obfuscated forth?!?  :)
16:54:05 <crc> I440r: it's not more obfuscated to me
16:55:44 <I440r> crc ok then :)
16:56:02 <I440r> : : : ;  <-- that would be totally obfuscated to someone who didnt know forth :)
16:56:54 <crc> yup
16:58:19 <hyrax> most forth is obfuscated to those you don't know forth. 
16:58:49 <I440r> :)
16:58:59 <I440r> which is why they go screamin for mamma :)
17:00:59 * crc has been doing some work on an interactive fiction game in retroforth
17:05:37 <warpzero> sup i440r
17:07:42 <I440r> warp! hi dood
17:12:33 <warpzero> d00d
17:12:38 --- quit: Baughn (Read error: 113 (No route to host))
18:06:24 <crc> goodnight
18:49:40 --- quit: I440r ("brb")
18:50:48 --- join: I440r (mark4@rrcs-24-242-160-169.sw.biz.rr.com) joined #forth
19:41:08 --- quit: saon ("Lost terminal")
22:22:29 --- quit: Sonarman ("leaving")
23:26:53 --- quit: Herkamire ("leaving")
23:59:59 --- log: ended forth/05.03.20