Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #59930 > unrolled thread

pforth return stack strangeness

Back to article view | Back to comp.lang.forth

Contents

  pforth return stack strangeness admin@simple.gopher - 2017-05-27 17:03 +0000
    Re: pforth return stack strangeness Julian Fondren <julian.fondren@gmail.com> - 2017-05-27 10:22 -0700
      Re: pforth return stack strangeness admin@simple.gopher - 2017-05-27 19:15 +0000
        Re: pforth return stack strangeness Julian Fondren <julian.fondren@gmail.com> - 2017-05-27 12:37 -0700
    Re: pforth return stack strangeness Roelf Toxopeus <rt4all@nothishetnet.nl> - 2017-05-27 21:10 +0200

#59930 — pforth return stack strangeness

Maybe there are some pforth users here..

QUADRATIC example from page 112 in Starting Forth book:
% pforth
  / compute a * X^2 + b * X + c
  : QUAD ( a b c X -- n) >R SWAP ROT I * + R> * + ;    ok
  2 7 9 3 QUAD . ==> 30 ok

The answer should be 48.  TRACE suggests stuff is getting drop/corrupted:

  2 7 9 3 TRACE QUAD .
  << QUAD +0      <10:4> 2 7 9 3           ||  >R                  >> 3    ok
  s
  << QUAD +4      <10:2> 2 7               ||  SWAP                >>    ok
  s
  << QUAD +8      <10:2> 7 2               ||  ROT                 >>    ok
  s
  << QUAD +12     <10:2> 2 0               ||  I                   >>    ok
  s
  << QUAD +16     <10:3> 2 0 1515870810    ||  *                   >>    ok
  s
  << QUAD +20     <10:2> 2 0               ||  +                   >>    ok
  s
  << QUAD +24     <10:1> 2                 ||  R>                  >>    ok
  s
  << QUAD +28     <10:2> 2 9               ||  *                   >>    ok
  s
  << QUAD +32     <10:1> 18                ||  +                   >>    ok
  s
  << QUAD +36     <10:0>                   ||  EXIT                >>    ok
  s Finished.
     ok

For reference, same thing in gforth:
  % gforth -e ": quad >r swap rot i * + r> * + ; 2 7 9 3 quad . bye"
  48

pforth documentation doesn't suggest limited return stack functionality; not
sure why it doesn't work..

[toc] | [next] | [standalone]

#59932

On Saturday, May 27, 2017 at 12:04:00 PM UTC-5, ad...@simple.gopher wrote:
> Maybe there are some pforth users here..
> 
> QUADRATIC example from page 112 in Starting Forth book:
> % pforth
>   / compute a * X^2 + b * X + c
>   : QUAD ( a b c X -- n) >R SWAP ROT I * + R> * + ;    ok
>   2 7 9 3 QUAD . ==> 30 ok
> 
> The answer should be 48.

The answer's also wrong under SwiftForth.

You can get the intended answer under both by accessing the top
of the return stack with R@ , rather than I :

  $ ./pforth_standalone 
  PForth V27-LE/64, built May 27 2017 12:18:53 (static)
  
  : quad >r swap rot r@ * + r> * + ; 2 7 9 3 quad . 48    ok
  Stack<10> 

[toc] | [prev] | [next] | [standalone]

#59943

>On Saturday, May 27, 2017 at 12:04:00 PM UTC-5, ad...@simple.gopher wrote:
>> Maybe there are some pforth users here..
>>
>> QUADRATIC example from page 112 in Starting Forth book:
>> % pforth
>>   / compute a * X^2 + b * X + c
>>   : QUAD ( a b c X -- n) >R SWAP ROT I * + R> * + ;    ok
>>   2 7 9 3 QUAD . ==> 30 ok
>>
>> The answer should be 48.
>
>The answer's also wrong under SwiftForth.
>
>You can get the intended answer under both by accessing the top
>of the return stack with R@ , rather than I :
>
>  $ ./pforth_standalone
>  PForth V27-LE/64, built May 27 2017 12:18:53 (static)
>
>  : quad >r swap rot r@ * + r> * + ; 2 7 9 3 quad . 48    ok
>  Stack<10>

Hey nice!  Seems to also work with 'ff'[1] forth:

% echo ": quad >r swap rot r@ * + r> * + ; 2 7 9 3 quad . cr bye" | ff
 ok
 48

Thanks ; it was probably a very novice question..

J.

[1] http://bitbucket.org/bunny351/ff

[toc] | [prev] | [next] | [standalone]

#59947

On Saturday, May 27, 2017 at 2:15:40 PM UTC-5, ad...@simple.gopher wrote:
> Hey nice!  Seems to also work with 'ff'[1] forth:
>
> % echo ": quad >r swap rot r@ * + r> * + ; 2 7 9 3 quad . cr bye" | ff
>  ok
>  48

Neat, don't think I've seen this one.

README:

    `2nip           ( x y -- y )`

  Drops the 2nd item on the stack.

README (later):

  2nip           ( x y z -- z )

core.f:

  : 2nip  ( x y z q -- z q ) rot drop rot drop ;

At least the implementation does what I'd expect :p

>
> Thanks ; it was probably a very novice question..
>

don't sweat it. First the code in the book is rotting a bit and this
is a natural question to have about it. Second, there aren't so many
novice questions around here that people are having to ask for
directions to the loony threads.

[toc] | [prev] | [next] | [standalone]

#59942

On 27/05/17 19:03, admin@simple.gopher wrote:
> Maybe there are some pforth users here..
> 
> QUADRATIC example from page 112 in Starting Forth book: % pforth /
> compute a * X^2 + b * X + c : QUAD ( a b c X -- n) >R SWAP ROT I * +
> R> * + ;    ok 2 7 9 3 QUAD . ==> 30 ok
> 
> The answer should be 48.  TRACE suggests stuff is getting
> drop/corrupted:
> 
> 2 7 9 3 TRACE QUAD . << QUAD +0      <10:4> 2 7 9 3           ||  >R
> >> 3    ok s << QUAD +4      <10:2> 2 7               ||  SWAP
> >>    ok s << QUAD +8      <10:2> 7 2               ||  ROT
> >>    ok s << QUAD +12     <10:2> 2 0               ||  I
> >>    ok s << QUAD +16     <10:3> 2 0 1515870810    ||  *
> >>    ok s << QUAD +20     <10:2> 2 0               ||  +
> >>    ok s << QUAD +24     <10:1> 2                 ||  R>
> >>    ok s << QUAD +28     <10:2> 2 9               ||  *
> >>    ok s << QUAD +32     <10:1> 18                ||  +
> >>    ok s << QUAD +36     <10:0>                   ||  EXIT
> >>    ok s Finished. ok
> 
> For reference, same thing in gforth: % gforth -e ": quad >r swap rot
> i * + r> * + ; 2 7 9 3 quad . bye" 48
> 
> pforth documentation doesn't suggest limited return stack
> functionality; not sure why it doesn't work..
> 

It's like Julian said, use R@ instead of I. Has to do with loop
implementation strategies (and standards). Use I as the current index
for the innermost loop. Use R@ to copy the top of the return stack to
data stack.

Both the second Starting Forth edition (p.100) and the online version
use R@ (it's now also called QUADRATIC).
: QUADRATIC  ( a b c x -- n )   >R SWAP ROT R@ *  + R> *  + ;

In case you run into more issues using the first edition and whatever
Forth, you could consult/try the online updated version:
http://home.iae.nl/users/mhx/sf.html
https://www.forth.com/starting-forth/

or just ask here of course.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.forth

csiph-web